Yes Africa Can Success Stories from a Dynamic Continent Punam Chuhan-Pole Manka Angwafo Editors 63431 Yes Africa Can Yes Africa Can SUCCESS STORIES FROM A DYNAMIC CONTINENT EDITORS PUNAM CHUHAN-POLE AND MANKA ANGWAFO © 2011 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved 1 2 3 4 14 13 12 11 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomina- tions, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permis- sion may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8745-0 eISBN: 978-0-8213-8746-7 DOI: 10.1596/978-0-8213-8745-0 Cover design: Naylor Design, Inc. Library of Congress Cataloging-in-Publication Data Yes Africa can : success stories from a dynamic continent / editors: Punam Chuhan-Pole, Manka Angwafo. p. cm. Includes bibliographical references. ISBN 978-0-8213-8745-0 — ISBN 978-0-8213-8746-7 (electronic) 1. Africa—Economic policy. 2. Africa—Economic conditions—1960- 3. Africa—Economic conditions— 21st century. 4. Africa—Social conditions. 5. Africa—Politics and government—21st century. I. Chuhan-Pole, Punam. II. Angwafo, Manka. HC800.Y47 2011 330.967—dc22 2011014918 CONTENTS Foreword ix Acknowledgments xi Abbreviations xiii Overview 1 Punam Chuhan-Pole and Shantayanan Devarajan PART I TRACKING SUCCESSFUL GROWTH EXPERIENCES Chapter 1 Tanzania: Growth Acceleration and Increased Public Spending with Macroeconomic Stability 21 David O. Robinson, Matthew Gaertner, and Chris Papageorgiou Chapter 2 Building on Growth in Uganda 51 Sarah Ssewanyana, John Mary Matovu, and Evarist Twimukye Chapter 3 Rapid Growth and Economic Transformation in Mozambique, 1993–2009 65 Antonio M. D. Nucifora and Luiz A. Pereira da Silva Chapter 4 Botswana’s Success: Good Governance, Good Policies, and Good Luck 81 Michael Lewin Chapter 5 Mauritius: An Economic Success Story 91 Ali Zafar Chapter 6 Cotton Dependence in Burkina Faso: Constraints and Opportunities for Balanced Growth 107 Jonathan Kaminski PART II POST CONFLICT SITUATIONS – BUILDING INSTITUTIONS AND GOVERNANCE Chapter 7 Postconflict Economic Governance Reform: The Experience of Liberia 127 Vishal Gujadhur Chapter 8 Decentralization in Postconflict Sierra Leone: The Genie Is Out of the Bottle 141 Vivek Srivastava and Marco Larizza Chapter 9 Transport Infrastructure and the Road to Statehood in Somaliland 155 Jean-Paul Azam v PART III LEVERAGING SECTORAL ADVANTAGES TO EXPAND EXPORTS Chapter 10 Growing Mali’s Mango Exports: Linking Farmers to Market through Innovations in the Value Chain 167 Yéyandé Sangho, Patrick Labaste, and Christophe Ravry Chapter 11 Economic Liberalization in Rwanda’s Coffee Sector: A Better Brew for Success 185 Karol C. Boudreaux Chapter 12 Cocoa in Ghana: Shaping the Success of an Economy 201 Shashi Kolavalli and Marcella Vigneri Chapter 13 Apparel Exports in Lesotho: The State’s Role in Building Critical Mass for Competitiveness 219 Mallika Shakya Chapter 14 The Success of Tourism in Rwanda: Gorillas and More 231 Hannah Nielsen and Anna Spenceley PART IV BOOSTING AGRICULTURAL EFFICIENCY AND OUTPUT THROUGH TARGETED INTERVENTIONS Chapter 15 Increasing Rice Productivity and Strengthening Food Security through New Rice for Africa (NERICA) 253 Aliou Diagne, Soul-Kifouly Gnonna Midingoyi, Marco Wopereis, and Inoussa Akintayo Chapter 16 Fertilizer in Kenya: Factors Driving the Increase in Usage by Smallholder Farmers 269 Joshua Ariga and T. S. Jayne Chapter 17 Malawi’s Agricultural Input Subsidy Program Experience over 2005–09 289 Andrew Dorward, Ephraim Chirwa, and T. S. Jayne Chapter 18 MoneyMaker Pumps: Creating Wealth in Sub-Saharan Africa 319 I. V. Sijali and M. G. Mwago PART V ENGAGING THE PRIVATE SECTOR TO UPGRADE INFRASTRUCTURE Chapter 19 ICT in Sub-Saharan Africa: Success Stories 339 Kaoru Kimura, Duncan Wambogo Omole, and Mark Williams Chapter 20 Mobile Payments Go Viral M-PESA in Kenya 353 Ignacio Mas and Dan Radcliffe Chapter 21 Independent Power Projects in Sub-Saharan Africa: Determinants of Success 371 Anton Eberhard and Katharine Nawal Gratwick PART VI IMPROVING HUMAN DEVELOPMENT OUTCOMES WITH INNOVATIVE POLICIES Chapter 22 Innovative Financing for Health in Rwanda: A Report of Successful Reforms 403 C. Sekabaraga, A. Soucat, F. Diop, and G. Martin Chapter 23 The Malaria Control Success Story 417 Anne-Maryse Pierre-Louis, Jumana Qamruddin, Isabel Espinosa, and Shilpa Challa Chapter 24 Health Extension Workers in Ethiopia: Improved Access and Coverage for the Rural Poor 433 Nejmudin Kedir Bilal, Christopher H. Herbst, Feng Zhao, Agnes Soucat, and Christophe Lemiere vi CONTENTS Chapter 25 Family Planning Trends in Sub-Saharan Africa: Progress, Prospects, and Lessons Learned 445 Mona Sharan, Saifuddin Ahmed, John May, and Agnes Soucat Chapter 26 Achieving Universal Primary Education through School Fee Abolition: Some Policy Lessons from Uganda 465 B. Essama-Nssah CONTENTS vii F O R E WO R D In April 2011, a study entitled “Hiding the Real Africa” doc- low; the production structure of African economies is umented how easily Africa makes the evening newscasts largely undiversified, heavily concentrated in primary com- and newspaper headlines in the West when a major famine, modities; and governance remains weak. pandemic, or violent crisis breaks. For example, over a five- To harness the recent growth and dynamism in the con- month period from May 2010, more than 245 articles on tinent to address these long-term challenges, we need to Africa published by the 10 most-read US newspapers understand what was behind Africa’s resurgence. This focused on poverty. Only five of them mentioned wealth book documents some of the success stories in Africa. The and growth. 26 case studies, which have been prepared by local and The tendency to dwell on Africa’s challenges, long- international academics, analysts, and practitioners, take standing problems, and failures rather than on its opportu- an in-depth look at economic and social development nities and successes is one of the continent’s most enduring achievements across countries, themes, and sectors. Some stereotypes. Despite an acceleration of economic growth of these are well known, such as Mauritius’s economic over the past 10 years and a growing army of African mid- growth experience and the information, communications, dle-class consumers, the narrative about Africa has and technology revolution in Africa. But others are not: remained one of poverty, disease, and conflict. the transformation of Rwanda’s coffee sector and the shift This suggests that changing the narrative on Africa to high-quality coffee production; Lesotho’s success in requires not only sustained economic progress in the con- boosting apparel exports; Mozambique’s rapid economic tinent, but also a collaborative platform for telling the story growth (averaging more than 8 percent a year from in a compelling way. 1993–2009); and the progress in combating malaria. In Economic growth has spurted, averaging 5 percent a year addition to documenting the nature of the successes, the from 1998 to 2008. Despite a sharp slowdown in the wake of case studies identify the reasons and draw lessons for other the recent global financial crisis, growth has rebounded and countries on the continent and elsewhere. is back on track. Africa’s private sector is showing signs of I expect Yes Africa Can: Success Stories from a Dynamic dynamism, attracting growing amounts of investment from Continent to stimulate the interest of both the media and domestic and foreign investors. Africa’s poverty rate declined filmmakers to seek out and tell more of Africa’s many success by about 1 percentage point a year, from 59 percent in 1995 stories; to help put Africa’s positive achievements on the map; to 51 percent in 2005. Primary completion rates are rising and to inspire a vigorous discussion on how Africans can do faster in Africa than anywhere else. More girls are in school, more to unleash the full economic potential that is needed to as educational opportunities for girls have expanded across transform one of the world’s fastest growing regions. the region. The death of children aged five or younger con- tinues to decline. Obiageli Katryn Ezekwesili Undeniably, much remains to be done: nearly 400 mil- Vice President, Africa Region lion people still live in extreme poverty; human capital is The World Bank ix AC K N OW L E D G M E N T S This volume was prepared by a team led by Punam Donald F. Larson, Hannah Messerli, Vincent Palmade, Chuhan-Pole under the general guidance and direction of Markus Moeller, Agnes Soucat, and Hassan Zaman. Discus- Shantayanan Devarajan, Chief Economist of the World sions with Jorge Saba Arbache, Christopher Paul Jackson, Bank’s Africa Region. Several of the case studies included in and Stephen Mink were also useful in selecting case studies. this volume were written by local researchers and practi- Laurent Wagner made a central contribution in the early tioners based in Sub-Saharan Africa. phase of the study. The team contributing to the book consisted of: The report benefited from input by peer reviewers: Rocio overview—Punam Chuhan-Pole, Shantayanan Devarajan, Castro, Vandana Chandra, Quy-Toan Do, Elena Ianchovichina, Manka Angwafo and Dana Vorisek; chapter 1—David Hannah Messerli, Celestin Monga, Blanca Moreno-Dodson, Robinson, Matthew Gaertner, and Chris Papageorgiou; chap- Claudia Paz Sepulveda and Hassan Zaman. Comments were ter 2—Sarah Ssewanyana, John May Matovu, and Evarist also received from Fabiano Bastos, Jaime Biderman, Maya Twimukye; chapter 3—Antonio Nucifora and Luiz A. Pereira Brahmam, Mukesh Chawla, Delfin Go, Monica Das Gupta, da Silva; chapter 4—Michael Lewin; chapter 5—Ali Zafar; Kenechukwu Maria Ezemenari, Errol George Graham, Jane chapter 6—Jonathan Kaminski; chapter 7—Vishal Gujadhur; Wangui Kiringai, Patrick Labaste, Karen Mcconnell Brooks, chapter 8—Vivek Srivastava and Marco Larizza; chapter 9— Moussa Diarra, Motoky Hayakawa, Maureen Lewis, Christine Jean-Paul Azam; chapter 10—Yéyandé Sangho, Patrick Lao Pena, Anne M. Pierre-Louis, Kofi Nouve, Rojid Sawkut, Labaste, and Christophe Ravry; chapter 11—Karol Boudreaux; Shahzad Sharjeel, Toni Sittoni, Yutaka Yoshino, and Ali Zafar. chapter 12—Shashi Kolavalli and Marcella Vigneri; chapter Participants at the April 2010 forum provided useful 13—Mallika Shakya; chapter 14—Hannah Nielson and Anna insights and perceptives that helped shape the output. Excel- Spenceley; chapter 15—Aliou Diagne, Soul-Kifouly Gnonna lent editorial review was provided by Dana Vorisek, Barbara Midingoyi, Marco Wopereis, and Inoussa Akintayo; chap- Karni, and Martha Gottron. Ann G. Karasanyi and Kenneth ter 16—Joshua Ariga and T.S. Jayne; chapter 17—Andrew Omondi provided administrative support throughout the Dorward, Ephraim Chirwa, and T. S. Jayne; chapter 18— project, particularly during the forum. The Web site featuring I.V. Sijali and M. G. Mwago; chapter 19—Kaoru Kimura, the stories was produced by Michael Matovina and Yohannes Duncan Wambogo Omole, and Mark Williams; chapter 20— Kebede. Mapi M. Buitano and Odilia Renata Hegba managed Ignacio Mas and Dan Radcliffe; chapter 21—Anton Eberhard the communication and dissemination activities. Book and Katharine Nawal Gratwick; chapter 22—C. Sekabaraga, design, editing, production, and printing were coordinated by A. Soucat, F. Diop, and G. Martin; chapter 23—Anne-Maryse Steven McGroarty, Rick Ludwick, and Denise Bergeron of the Pierre-Louis, Jumana Qamruddin, Isabel Espinosa, and Shilpa World Bank Office of the Publisher. Challa; chapter 24—Nejmudin Kedir Bilal, Christopher Finally, the production of this volume was made possible H. Herbst, Feng Zhao, Agnes Soucat, and Christophe Lemiere; by the World Bank Office of the Publisher. chapter 25—Mona Sharan, Saifuddin Ahmed, John May, and Agnes Soucat; and chapter 26—B. Essama-Nssah. Punam Chuhan-Pole Many people provided guidance in identifying potential Manka Angwafo success stories. Notably, Cecilia M. Briceno-Garmendia, Editors xi A B A P T E AT H R E E C H B R E V IR T I O N S ACT artemisinin-based combination therapy BPO business process outsourcing CPR contraceptive prevalence rate DHS Demographic and Health Survey ECOWAS Economic Community of West African States FAO Food and Agriculture Organization FMOH Federal Ministry of Health GDP gross domestic product GSM Global System for Mobile Communications HEP Health Extension Program ICT information and communications technology IDA International Development Association IT/ITES information technology/information technology–enabled service MOES Ministry of Education and Sports NERICA New Rice for Africa NGO nongovernmental organization PC personal computer PIN personal identification number PMA Plan for the Modernization of Agriculture SADC Southern Africa Development Community SMS short messaging system TFR total fertility rate UNFPA United Nations Population Fund USAID U.S. Agency for International Development WAEMU West African Economic and Monetary Union WHO World Health Organization (all dollar figures are U.S. dollars) xiii Overview Punam Chuhan-Pole and Shantayanan Devarajan WHY STUDY AFRICAN SUCCESSES? discourse on Africa’s economic development has shifted Over the past decade Sub-Saharan Africa has seen a remark- from whether the region will develop to how the region is able turnaround in economic performance. After years of developing. stagnation, economic growth has spurted—gross domestic Yet, there are still causes for concern. For one thing, per- product (GDP) grew from an annual average rate of less formance across countries varies substantially. Also worry- than 2 percent in 1978–95 to nearly 6 percent over 2003–08. ing is that, historically, Africa’s performance has been Inflation is half its level of the mid-1990s. Private capital volatile—with short periods of acceleration followed by flows have risen to $50 billion, exceeding foreign aid. Exports long periods of deceleration and decline brought on by are growing, as is private sector activity. The number of dem- recurrent crises (Arbache and Page 2007). Moreover, the ocratic regimes has risen and the security situation has pattern of progress underscores serious shortfalls in some improved. The poverty rate is falling by 1 percentage point a areas, notably, in the economic diversification of many year. Countries such as Ethiopia, Ghana, Mauritania, and countries and in the integration of African economies into Rwanda are on track to reach many of the Millennium the global economy (UNIDO 2009). African countries also Development Goals. Nine African countries have need to correct large infrastructure deficits, dramatically achieved or are on track to achieve the target for extreme expand the skills base of their labor forces, and improve poverty (World Bank and IMF 2011). Among other encour- their ability to absorb technical knowledge in the private aging trends are more fair and effective leadership, an sector to improve their global standing in the years ahead. improving business climate, increasing innovation, a more With the lingering specter of past failures, continued involved citizenry, and growing reliance on home-grown economic development challenges in Sub-Saharan Africa solutions. More and more, Africans are driving African have led to questions of whether improved performance in development. recent years is sustainable. This book attempts to answer This increased dynamism in Sub-Saharan Africa is evi- these questions by documenting and better understanding dent across a broad swath of countries. It has created opti- some of the impressive achievements that have occurred in mism that Africa’s favorable development performance will recent years. By systematically identifying and assessing pos- be long lasting and that it could dramatically transform itive outcomes, it is possible to draw out lessons regarding countries in the region. Along the way, the prevailing what has worked in practice and why. These lessons will in We wish to thank Manka Angwafo and Dana Vorisek for their excellent contributions. 1 turn inform our judgment about whether Africa’s growth is value terms in the 1980s, exports doubled in volume and sustainable. rose fivefold in value in the period 1994–2008. Nevertheless, After a review of the major recent economic develop- exports of goods (excluding oil) account for less than ments in Africa, this overview describes the approach and 20 percent of GDP (reflecting little change over the previous methodology used in the study of African successes. We then decades), and Africa’s share of world exports remains low, present a framework for understanding the 26 case studies, a after falling for more than half a century. Export growth has framework that is based on the balance between overcoming been led by extractive resource exports, fueled by a com- market failures and overcoming government failures. We modity boom (figure 4). Africa’s exports continue to be conclude that the way in which African governments have dominated by extractive and agricultural commodities. In been increasingly addressing government failures—with many cases, commodity dependence seems to have intensi- impressive results—bodes well for the sustainability of fied. Manufacturing exports have grown slowly, and much Africa’s decade-long growth. of the region has yet to break into the global industrial mar- ket (UNIDO 2009). The diversity and sophistication of exports has changed very little. Whereas the mostly upward A CHANGING ECONOMIC LANDSCAPE trend in extractive commodity prices has provided revenues After lackluster economic performance for decades, during to finance much needed infrastructure and social spending, which the gap between Sub-Saharan Africa and the rest of it also presents challenges of avoiding the “resource curse” the developing world widened, the region’s economies have and converting the revenues into sustained development. seen a visible turnaround that began in the mid-1990s. The Improving economic prospects have attracted foreign most visible evidence is the acceleration in output growth direct investment (FDI) to the region (figure 5). FDI grew by from under 2 percent in 1978–95 to nearly 6 percent in more than 17 percent a year between 1995 and 2008. Although 2003–08.1 Oil exporters have seen an especially steep rise, extractive industries account for the bulk of the FDI inflows, but growth has been widespread (figure 1). Indeed, 21 significant amounts—especially greenfield investments—are non-oil countries, home to over 40 percent of the region’s being recorded in the construction, communications, and population and accounting for nearly a quarter of the banking sectors. Emerging market countries such as China region’s GDP, have posted annual economic growth of and India are becoming important sources of FDI. more than 4 percent during 1995–2008. Many of these A notable feature of the transformation under way in 21 countries face the additional challenge of being land- Africa is the upsurge in agricultural output and productiv- locked. Within this overall improvement in the growth ity. Agricultural GDP growth averaged nearly 3 percent a trend, some countries are lagging, but their numbers are year over 1980–2004, but growth per capita averaged only small and declining. 0.9 percent, between one-third and half that of other devel- In a remarkable break from historical performance, not oping regions (World Bank 2008). This average hides con- only has growth accelerated, but the acceleration has been siderable variation across countries and over time. For sustained for a longer period than in the past. For example, example, per capita growth in agricultural output was neg- Mozambique, Tanzania, and Uganda have experienced ative in the 1980s and early 1990s, before turning positive. A growth rates above 5 percent in every year during 2001–08. recent study (Fuglie 2010) finds that the pickup in agricul- Arbache et al. (2008) find that growth accelerations were tural growth in the 1990s and 2000s resulted from expan- more frequent in 1995–2005, whereas growth decelerations sion of land under agriculture. Block (2010) finds that agri- were more common in the preceding two decades: the like- cultural total factor productivity growth declined in 1960s lihood of a growth acceleration was 0.42 in 1995–2005 but and 1970s, but productivity picked up over the last two 0.21 in 1985–94. decades. Signs of change in African agriculture are evident Strong output growth has lifted per capita incomes as in the successes documented in this book, such as the well. Per capita incomes fell in 1975–84 at an average annual expansion in maize production in Kenya that was driven by rate of 0.88 percent, and at an even faster pace (1.13 percent) liberalization of the local fertilizer market, in cotton in West in 1985–94. By contrast, per capita income expanded at Africa, and in cassava and rice production in many coun- nearly 2 percent a year in 1995–2008, in step with growth tries, boosted by improved varieties. In a region where two- rates in other developing regions (figure 2). thirds of the population is rural and 70 percent of the poor African exports are growing as well but remain highly derive their livelihood from agriculture, the favorable trend concentrated (figure 3). After stagnating in volume and in agriculture is encouraging. Robust growth in agriculture 2 OVERVIEW Figure 1 Average Real GDP Growth Rates in Sub-Saharan Africa: 1995–2008 Equatorial Guinea Angola Chad Sudan 31% of African Nigeria population Cameroon Congo, Rep. Gabon Liberia Rwanda Mozambique Uganda Ethiopia Sao Tome and Principe Cape Verde Burkina Faso Botswana Mali Tanzania 41% of African Ghana population Malawi Benin Sierra Leone Gambia, The Namibia Mauritius Mauritania Senegal Niger Zambia Madagascar Guinea South Africa Togo Kenya Lesotho Swaziland Seychelles 27% of African Cote d'Ivoire population Comoros Central African Republic Congo, Dem. Rep. Guinea-Bissau Eritrea Burundi Zimbabwe –5 0 5 10 15 20 25 30 percent Source: World Development indicators, World Bank. holds the promise of enhancing food security and facilitat- poverty rate. Several African countries are likely to achieve ing broad-based economic growth and poverty reduction. the MDG of halving income poverty by 2015 (Goal 1). As stagnation in African economies has given way to Overall, however, the region lags other regions in attaining growth, poverty has begun to decline. The incidence of Goal 1 of the MDGs. extreme poverty—the share of people living on less than Education and health indicators have also improved $1.25 a day—fell from 59 percent in 1995 to nearly 50 per- (figure 6). For example, primary school enrollment has cent in 2005, a decline of 1 percentage point a year in the jumped 14 percentage points, from about 59 percent in OVERVIEW 3 Figure 2 Average per capita income growth by country Figure 4 Trends in Commodity Prices and Terms of groups: 1975–2008 Trade in Sub-Saharan Africa 4.5 a. Commodity and oil prices 4.0 400 120 2000 = 100 using constant US$ 3.5 350 3.0 100 $/bbl, constant 2000 US$ 2.5 300 percent 2.0 80 1.5 250 1.0 200 60 0.5 0 150 –0.5 40 0 100 –1.5 20 50 a n e e e ric ra m m m 0 0 Af Saha co co co in in in b- 5 0 5 0 5 0 5 0 gh e w –7 –8 –8 -9 –9 –0 –0 –1 dl Su lo hi Jan id Jan Jan Jan Jan Jan Jan Jan m 1975–1984 1985–1994 1995–2008 metals and minerals agriculture crude oil brent, in $/bbl Source: World Development indicators, World Bank. b. Sub-Saharan Africa: Terms-of-trade index Note: Middle and low income includes Sub-Saharan Africa. 160 index, 2000 = 100 140 120 Figure 3 Trends in Export Volumes and Values, Sub-Saharan Africa, 1970–2010 100 a. Export volumes 300 80 80 85 90 95 00 05 10 250 19 19 19 19 20 20 20 in constant US$ billions 200 Source: World Development indicators, World Bank. 150 2000 to 73 percent in 2008—the fastest improvement of any 100 region. More girls are attending school. Some significant 50 gains have also been made in the area of health, where progress has admittedly been mixed. For example, child 0 deaths (under-five mortality rates) have declined from 181 70 74 78 82 86 90 94 98 02 06 10 19 19 19 19 19 19 19 19 20 20 20 per 1,000 births in 1990 to 132 per 1,000 in 2009, with some b. Export values of the poorest countries, such as Eritrea and Malawi, show- 600 ing remarkable progress. Maternal mortality rates are also 500 trending down, though not fast enough, and the region is in current US$ billions 400 beginning to make inroads in halting the spread of commu- nicable diseases. 300 Since the mid-1990s African countries have made strides 200 in the area of economic management and structural poli- cies. Combined with debt relief, these reforms have helped 100 to redress external and fiscal imbalances and rein in infla- 0 tion. Median inflation has fallen from double-digit levels in 70 74 78 82 86 90 94 98 02 06 10 the 1970s and 1980s to well below 10 percent in 1996–2007. 19 19 19 19 19 19 19 19 20 20 20 Source: World Development indicators, World Bank. In 1995 nearly a third of countries had inflation rates above 20 percent; in 2008 this figure had dropped to one-tenth 4 OVERVIEW Figure 5 Foreign Direct Investment in Sub-Saharan Africa 40 5 35 4.5 4 30 3.5 percent of GDP (%) billions of US$ 25 3 20 2.5 15 2 1.5 10 1 5 0.5 0 0 87 93 96 02 78 81 84 90 99 05 08 19 19 19 20 19 19 19 19 19 20 20 foreign direct investment, net inflows (BoP, current US$) foreign direct investment, net inflows (% of GDP) Source: World Development indicators, World Bank. Figure 6 Progress in Human Development in (figure 7). Exchange rates have also been maintained at Sub-Saharan Africa competitive levels, including the one-time devaluation of the CFA franc in 1994. a. Primary school enrollment rate Over the past several years, many African countries have 80 60 moved to liberalize trade, and reforms have brought down 70 40 tariff rates. Average most-favored-nation (MFN) applied percent millions 60 tariff rates more than halved from 1981 to 2009 and are 20 comparable to those of developing countries in other 50 regions (figure 8). But improvements in trade facilitation 40 0 have lagged. In a bid to enhance competitiveness, African countries have also made progress in recent years in imple- 99 01 03 05 07 19 20 20 20 20 menting reforms to support the investment climate. total primary-age children out of school Because of reforms, Rwanda is now one of the fastest places net primary enrollment rate in the world to start a business (11th overall according to the World Bank’s 2010 Doing Business report). There are b. Under-five mortality rate also indications that the quality of governance is improving 181 in some countries (figure 9). But, overall, weak governance 180 under-five mortality rate (per 1000) remains a problem in Africa, especially in fragile states and 155 resource-rich countries, and exerts a drag on long-term 130 132 development. 105 80 APPROACH AND METHODOLOGY 90 95 00 05 06 07 08 09 19 19 20 20 20 20 20 20 A number of major academic and policy studies pub- Source: World Development indicators, World Bank. lished in recent years have examined trends related to OVERVIEW 5 Figure 7 Inflation Patterns in Sub-Saharan Africa economic development in Sub-Saharan African countries. In Can Africa Claim the 21st Century?, released in 2000, 100 authors from four multilateral institutions conclude that 16.7 10.3 31.6 the emergence of three favorable trends—greater political percent of countries 80 36.0 33.3 participation (the basis for more accountable govern- 48.7 60 ments), a changing view of Africa in the wake of the end 28.9 40 of the cold war, and the increasing presence of globaliza- 52.0 tion and information technology—hold the promise of 50.0 41.0 20 39.5 accelerating development and moving Africa out of 12.0 poverty. At the same time, the book indicates that 0 1975 1985 1995 2008 strengthening governance, investing in people, increasing under 10% inflation 10–20% inflation competitiveness, and reducing aid dependence are crucial above 20% inflation to development paths in African economies. Source: World Development indicators, World Bank. Ten years later, Radelet, in his book Emerging Africa: How Note: The figure shows the share of countries in each inflation range. 17 Countries Are Leading the Way, finds that an improving governance environment, stronger economic policies, a Figure 8 Trends in Average MFN Applied Tariff Rates, changed relationship with the international donor commu- Sub-Saharan Africa, 1981–2008 nity, the increased accountability provided through new technologies, and a new generation of business leaders and 35 policy makers have been indispensable in bringing about percent, unweighted 30 progress in the 17 African countries where a turnaround has 25 been most evident. 20 Drawing on the existing knowledge of African develop- 15 ment presented in these and other publications,2 Yes Africa 10 5 Can: Success Stories from a Dynamic Continent takes an in- 0 depth look at 26 economic and social development successes in Sub-Saharan African countries—20 from individual 5 0 5 0 5 8 –8 –9 –9 00 –0 –0 81 86 91 –2 01 06 countries and 6 that occurred regionwide.3 The successes 19 19 19 96 20 20 19 manifest at the project, provincial, subnational, national, or Sub-Saharan Africa developing countries (134) regional level and cut across themes, programs, and sectors. Source: World Development indicators, World Bank. The overall goal of the research is to address how Sub- Saharan African countries have overcome major develop- ment challenges. The approach straddles the development Figure 9 Trends in Political Rights and Civil Liberties in debate that Cohen and Easterly explore in What Works In Sub-Saharan Africa Development?: Thinking Big and Thinking Small. That is, should one adopt a big-picture approach that focuses on the 7 role of the quality of institutions and macroeconomic and 6 structural policies or an approach that focuses on micro- 5 economic interventions such as conditional cash transfers 4 and bed nets, whose impact can be measured through 3 impact evaluations? 2 The case studies were chosen through a stocktaking 1 exercise that included consultation with experts inside and 0 outside the World Bank. To qualify for selection, each case 1974 1979 1984 1989 1994 1999 2004 2009 study had to meet several criteria. First, the achievement political rights civil liberties identified in the case study had to be observable and meas- Source: Freedom House. urable. Second, the achievement had to represent an outcome Notes: Political rights and civil liberties are measured on a one-to-seven scale, with one representing the highest degree of freedom and seven the perceived as desirable in the literature on development. lowest. Third, there must have been existing analytical work—impact 6 OVERVIEW evaluations, assessments (including against relevant regional traditional goods? How have some countries been able to or sectoral benchmarks), and reports—regarding the build capacity so as to respond to quantity and quality achievement. And fourth, the achievement must have been needs of the global market, including compliance with sustained over time. In light of the above criteria, and international standards? because of the potential lessons that can be learned, there was a focus on critically evaluating well-recognized success Boosting agricultural efficiency and output through stories in addition to those that are less well known. Where targeted interventions and reforms. The vast majority possible, the case studies attempt to identify achievements of Africa’s poor are concentrated in rural areas and are heavily that have the potential to be scaled up or that have elements dependent upon agriculture for their livelihoods. Spurring of transferability. agricultural development will be vital for achieving the The coverage of development successes is selective rather MDGs in Africa. Raising agricultural productivity is impor- than comprehensive. Indeed, some achievements were not tant for stimulating growth in other sectors as well. What are examined because of resource constraints and timely avail- some of the ways in which countries have succeeded in boost- ability of experts who could undertake the studies. A ing agricultural production, commercial agriculture, and notable example that was not studied in detail is the success productivity? of Kenya’s horticulture industry in becoming a major sup- plier of cut flowers to the European Union. Engaging the private sector to upgrade infrastruc- The 26 case studies presented in Yes Africa Can: Success ture. Africa has large unmet infrastructure needs: the latest Stories for a Dynamic Continent cover six broad topical areas: estimates point to a financing gap of more than $30 billion a year (World Bank 2010). Closing the infrastructure gap Successful growth experiences. Both a high rate of will require more private investment and operation. But the growth and the quality of growth are central to achieving region has been slow to mobilize private resources, partly poverty reduction. because of governance, institutional, and regulatory issues. How have some countries managed to attract private ■ Sustaining strong economic growth. High and sustained participation in delivering infrastructure services? One economic growth is needed if countries are to substan- important area where African countries have seen rapid tially reduce poverty and achieve the MDGs. A few African progress is in the mobile telecommunications segment of countries have grown at a brisk pace over many years. the information and communications technology (ICT) What explains these long periods of growth accelerations? sector. Deployment of this technology holds the potential What has been the role of a strong business environment for large benefits—for example, connecting people and in unleashing growth? Has growth been inclusive? firms to formal financial markets and linking farmers to ■ Quality of growth—industrial development. Empirical output markets. The African private sector has responded to evidence shows that countries with more diversified pro- competition in this market in unprecedented ways. What is duction and more sophisticated production sectors have behind the vibrancy of this market and what is the scope for stronger growth. Africa’s industrial development has scaling up innovative uses of ICT in African countries? been disappointing. Yet, there are glimmers of progress. How have some African countries achieved more broad- Improving human development outcomes with inno- based production? vative Policies. Sub-Saharan Africa lags on all the health and education MDGs. Yet, some countries have made sig- Postconflict situations: building institutions and nificant progress in improving health and education out- Governance. Institutions and governance are central to the comes. How have these countries achieved rapid progress? development process anywhere, but they are particularly What policies and approaches have worked and why? What vital in postconflict situations. How have some postconflict has been the role of public and private donors in improving countries established better governance and transparency? service delivery? Each case study includes: a description of the achievement Expanding exports. Increasing competitiveness and diver- and the elements that qualify the outcome as successful; an sifying exports is important if a country is to integrate in the assessment of the main policies, interventions, actions, and global economy and benefit from globalization. How have other factors that contributed to the positive outcome; a pres- some countries managed to diversify exports away from entation of the lessons learned and the contribution to the OVERVIEW 7 discourse on African development; and, where, possible The 26 case studies in this book—all policy success sto- insights on the transferability of the achievement and poten- ries from Sub-Saharan African countries—illustrate the tial for scaling up the interventions and actions. Individual many ways Africa has avoided or overcome market and gov- case studies also examine the role of the key stakeholders— ernment failures. That these countries have surmounted the government, donors, or private investors—in facilitating failures and generated economic and social progress bene- and promoting the achievement. fiting millions of poor people is testament to the innova- In examining government contributions to successful tion, dynamism, and spirit of the African people. But the outcomes, the focus is on the quality of economic policies case studies do more than inspire: they teach us about the and actions—that is, whether public action relieved con- nature of government and market failures, and how they straints to growth—whether they provided macroeconomic can be prevented or corrected. stability, improved the investment climate, enhanced com- To learn from the studies, it is useful to classify them into petitiveness or eased access to markets, promoted human four categories: overcoming or avoiding massive government development, or leveraged the global economy. How well failure, rebuilding or creating a government, rationalizing strategies and policies adjusted in response to changes in the government involvement in markets, and listening to the local and external economic environment is also addressed people (table 1). Regardless of which category each of the in many of the case studies. The narrative avoids being pre- case studies falls into, as a group, they demonstrate that scriptive because there are several policy combinations for the African landscape is dotted with success stories. The achieving desirable goals. sectoral and geographical diversity of the case studies illustrates that there are many ways to overcome Africa’s challenges. To paraphrase Tolstoy, African failures are all MAIN FINDINGS OF THE CASE STUDIES alike; but each success is successful in its own way. There has been no shortage of narratives about Africa’s many failures, giving rise to academic titles such as Africa’s Growth Overcoming massive government Tragedy, or The Bottom Billion, not to mention relentless failure or bad policies media coverage of poverty, conflict, disease, and famine on the continent. While the proximate causes of African countries’ In all the case studies in this group, existing policies were failure to thrive are quite different—hyperinflation in standing in the way of growth, either at the sectoral or Zimbabwe, civil war in Sierra Leone, desertification in the economywide level. In Ghana’s cocoa sector, Rwanda’s cof- Sahel, and HIV/AIDS in South Africa, to name a few—they fee sector, Kenya’s agricultural sector, Burkina Faso’s cotton can be boiled down to two main sources: market failures, such sector, and Tanzania’s wider economy, the willingness of as the common-property externalities associated with deserti- the government to allow liberalization has paid off. In fication, and government failures, often created by state inter- Uganda and Mozambique, postconflict reforms have paved vention to correct market failures. For instance, governments the way for impressive improvements in economic perfor- throughout the world have central banks to manage monetary mance. Across Sub-Sahara Africa, governments are grap- policy because markets alone cannot do that; but Zimbabwe’s pling with opening up the power sector to private sector central bank started issuing so much money to finance the fis- participation. Finally, the case study on Botswana proves cal deficit that the currency became worthless, resulting in that the resource curse can successfully be avoided even in hyperinflation and many years of hardship. an economy that is highly dependent on diamond mining. Market failures can be corrected by implementing a tax All of these cases show that the timing and nature of or subsidy or by generating a public good, such as improved reforms is key to their success. But the specifics of each case infrastructure or a better public health care system, for offer important lessons. which the government can take credit. Overcoming govern- Long an integral part of Ghana’s economy, the cocoa ment failures is more difficult, because the source of the industry faced near collapse in the early 1980s, battered by problem is usually powerful people who are benefiting from several forces. For one, Ghana was experiencing hyperinfla- the intervention. Correcting the problem involves under- tion and an overvalued exchange rate (the market exchange mining these people’s rents. In some cases, governments fail rate was approximately 44 times the official rate in 1983), to correct a market failure—another form of government making selling its cocoa at official rates almost worthless. As failure—because vested interests are benefiting from the much as 20 percent of Ghana’s cocoa harvest was smuggled distortion (think of politically connected monopolists). into Côte d’Ivoire for export. At the same time, an aging tree 8 OVERVIEW Table 1 Categorizing Successes: Overcoming Government and Market Failure Overcoming or avoiding massive Liberalization of the exchange rate and other reforms to revive the cocoa sector in Ghana government failure Removing barriers to trade and creating incentives for entrepreneurship in the coffee sector in Rwanda Liberalization of the fertilizer market in Kenya Liberalization of the cotton sector in Burkina Faso Facilitating private partnerships in the power sector across Sub-Saharan Africa through independent power producers Wide-ranging economic liberalization in Tanzania Reforming the economy in a postconflict environment in Uganda and Mozambique Good timing and good luck for diamond mining in Botswana Rebuilding a government or creating Rebuilding government following civil wars in Liberia and Sierra Leone a government where none existed Using traditional systems for collective action in Somaliland Rationalizing government Development of a system of air, rail, and road transport and cold storage to support mango involvement in markets exports in Mali Provision of textile and apparel industry infrastructure in Lesotho Catalytic government role in private sector development in Mauritius Using input subsidies to improve agricultural output in Malawi Provision of gorilla reserves to boost tourism revenues in Rwanda Shifting the government role in the ICT sector from monopoly provider to regulator across Sub-Saharan Africa Success in malaria control across Sub-Saharan Africa Listening to the people Performance-based financing in the health sector in Rwanda Abolishment of fees to achieve free universal primary education in Uganda Training and deploying extension workers to improve access to health care in Ethiopia Lowering fertility through family planning programs Developing new varieties of rice, NERICA, to increase yields and decrease food insecurity Using Moneymaker pumps to support innovation and diffusion of technology in the agricultural sector Using mobile phones to improve financial access in Kenya via M-PESA stock and the spread of plant diseases made Ghana’s cocoa In Rwanda, where the vast majority of people depend on crop increasingly unproductive and investment in the cocoa agriculture for their livelihoods, major transformation of sector unattractive. Ghana’s proportion of global cocoa pro- the coffee sector came about more quickly than it did in duction fell by half between the mid-1960s and the early Ghana’s cocoa sector. As of the late 1990s farmers were pro- 1980s, from 36 percent to 17 percent. Starting in 1983, ducing a small amount of mediocre-quality coffee that within the context of a wider economic recovery agenda attracted little attention from discriminating importers and that devalued the Ghanaian cedi and increased the farm gate consumers, while the economic destruction of the country prices paid to farmers for commodities (thus decreasing the during the genocide in 1994 remained a serious impediment incentive to smuggle cocoa out of the country), the govern- to markets. The first wave of reforms was introduced in ment undertook a specialized program intended to revive the late 1990s—several barriers to trade were removed, a the cocoa sector. Among other reforms, the program com- network of coffee-washing stations that allowed producers pensated farmers for replacing ailing cocoa trees (often with to shift from semiwashed to fully washed (higher-value) better-performing hybrid varieties), shifted responsibility coffee was implemented, and incentives to foster entrepre- for cocoa procurement to privately licensed companies, and neurship in the coffee industry were introduced. As a result, provided subsidies to increase the usage of fertilizer on both the quality and quantity of coffee produced in cocoa crops. Cocoa production and exports boomed. The Rwanda increased, allowing the government to launch, amount of cocoa produced per hectare doubled between in subsequent years, a successful strategy to boost Rwandan 1983 and 2006 as a result of increased fertilizer application producers’ participation in the international specialty- and better production practices. A greater share of cocoa coffee market. Between 2003 and 2008, the average export sales is now passed on to the 700,000-plus cocoa farmers, price of green coffee originating in Rwanda nearly doubled, reducing poverty and improving prospects for cocoa- from $1.60 to $3.10 a kilogram, leading to important producing households. improvements at the household level. The coffee washing OVERVIEW 9 stations, which had created 4,000 jobs as of 2006, also Efforts to overcome bad policies in the power sector have appear to have led to improved economic cooperation been effective in some Sub-Saharan African countries. The among people using them. case study on independent power producers (IPPs) focuses Like cocoa in Ghana, cotton has long been integral to the on the seven countries that have had the most experience economy of Burkina Faso. Cotton accounted for 60 percent with IPPs. As of the early 1990s, virtually all major power of the country’s exports between 1994 and 2004 and producers in Sub-Saharan Africa were publicly owned. They contributed to Burkina Faso’s good economic performance. had been performing poorly for decades, and public financ- A series of institutional reforms implemented in the cotton ing for new projects (from both domestic and international sector during the 1990s and 2000s—including the formation sources) was drying up. Governments needed to develop a of a national cotton union and the partial privatization of new tactic. Starting in Côte d’Ivoire in 1994, they experi- the national cotton parastatal, SOFITEX—have succeeded in mented with opening power generation to the private sec- opening the sector. The Burkinabe reform model is unique tor. Ghana, Kenya, Nigeria, Senegal, Tanzania, and Uganda, among Sub-Saharan African countries, because it addressed among other countries, followed suit shortly thereafter. government failures and local realities within the existing While the presence of IPPs has not solved the electricity institutional framework, adopted reforms using a cautious, deficit on the continent—only 25 percent of the population piecemeal approach, and built the capacity of producers and has access to electricity—IPPs have added several gigawatts upgraded institutions within the commodity chain while the of capacity to Sub-Saharan Africa’s electric grid, comple- government withdrew from most of its activities. But it is menting incumbent state-owned facilities. It is also clear also clear that Burkina Faso’s dependence on cotton makes that a sound policy framework (namely, adequate legisla- economic growth and performance in the country quite tion), a good investment climate, and local availability of volatile, because cotton earnings are dependent on interna- cost-competitive fuel can all contribute to the success of tional markets. In the years ahead, the country will need to IPPs. When these elements are not in place, IPPs are much seek avenues by which to diversify the economy. more likely to be untenable. In Kenya, as in Sub-Saharan Africa as a whole, fertilizer The case study on Tanzania addresses long-term market use traditionally has been quite low, meaning that overall reforms that have gone far beyond a single sector. Following productivity in the agricultural sector is also quite low. independence, the country’s economy languished for more Combined with the price volatility brought about by than two decades, held back by loss-making public enter- unpredictable weather and poor infrastructure, this low prises, deteriorating infrastructure, and mismanagement productivity has contributed to food insecurity in Kenya. of terms of trade and weather shocks, among other For decades, the Kenyan government addressed this food things, while poverty increased. Starting in 1986, a series of insecurity through direct, monopolistic involvement in the structural reforms focusing on liberalization was imple- markets for a wide array of agricultural products. This mented, many of which were similar to the reforms put in changed in the 1980s, when the government began relaxing place by Uganda in the same time period (see below). Trade its hold on agricultural markets, allowing the private sector and exchange rate regimes were liberalized; marketing to compete with state agencies and easing trade restrictions boards for agricultural products were dismantled; and on fertilizer and maize. A more complete liberalization of parastatals, the financial sector, and the civil service began the fertilizer market occurred in 1993. This change in policy, to be reformed. The reform cycle continued into the 1990s, coupled with liberalization of the foreign exchange regime supported by a significant amount of debt relief from inter- in 1992, encouraged the entry of a significant number of national donors, which freed budgetary resources for other private sector firms in importing, wholesaling, distribution, purposes. By the mid-1990s Tanzania was on the cusp of a and retailing of fertilizer. Government price controls and decade-long growth take-off, with real GDP growth averag- import licensing quotas were ultimately eliminated, and fer- ing around 6 percent from 1996 to 2008. Despite a difficult tilizer donations by donor agencies were phased out. Kenya external environment, Tanzania has maintained economic now stands as a notable departure from the Sub-Saharan stability in recent years. Aside from the fact that fiscal and African average in terms of fertilizer usage, which almost monetary authorities have succeeded in putting the econ- doubled between 1992 and 2007. Much of the increased use omy on much better footing than existed decades ago, two has been among smallholder farmers. Largely as a result of other factors—low reliance on exports for growth and low the increase in fertilizer usage, maize yields in Kenya levels of foreign exposure within the banking sector— increased 18 percent over 1997–2007. allowed Tanzania to weather the global financial crisis 10 OVERVIEW relatively unscathed. And while poverty among Tanzania’s more jobs (particularly outside the agricultural sector, population remains high, the country is on track to meet where most people are still employed) and widen the pro- several of the targets laid out under the Millennium ductive base in order to create an economy in which wealth Development Goals. is more broadly shared. In Uganda sustained, carefully sequenced macroeco- Policy choices in Botswana, like those in Tanzania, have nomic reforms in a postconflict environment have produced been key to long-term economic progress. At the time of remarkable outcomes. At the end of decades of political independence in 1966, Botswana’s per capita income was instability and civil war in 1986, Uganda’s economy was in among the lowest in the world. Many of its human devel- disarray and its population was struggling with a very high opment indicators were equally abysmal: life expectancy incidence of poverty. The reform agenda, begun in 1987 and was 37 years, and the primary school completion rate was implemented in stages, focused on stabilization, liberaliza- less than 2 percent. The country’s outlook was not encour- tion, and structural reforms: prices controls were loosened, aging. But Botswana’s economic performance in the years a floating exchange rate regime was adopted, marketing since has proven those expectations wrong: per capita boards were abolished, parastatals were abolished or priva- income growth, for example, averaged 7 percent a year tized, the civil service was reformed, and efforts were made between 1966 and 1999. Without a doubt, Botswana met to stimulate private investment. These steps were followed with a good bit of luck along the way—the discovery of by a wide-ranging poverty reduction plan that started in large deposits of diamonds just after independence. In the 1997. The performance of Uganda’s economy in the 1990s decades since, diamonds have fueled solid economic and 2000s suggests that the reform policies worked. Growth growth and a rapid increase in per capita annual income. averaged 7.7 percent a year over 1997–2007. Increased inter- But effective, transparent governance and good economic national confidence in the economy spurred substantial aid management have been crucial to Botswana’s success; they and foreign direct investment inflows and reversed capital helped Botswana avoid the resource curse experienced by flight. Poverty figures have also improved dramatically, so other Sub-Saharan African countries. In addition, the poli- much so that Uganda will be one of the few Sub-Saharan cies Botswana did not adopt also appear to have con- African countries to achieve the first Millennium Develop- tributed to its economic success. Botswana did not follow ment Goal of halving extreme poverty by 2015. At the same an import substitution policy, and it did not expand state- time, the Uganda case strikes a cautionary note: despite owned enterprises, which employ only a small percentage strong growth, income inequality has increased and the for- of the workforce. In the ensuing decades, economic and mal sector employment rates are very low, particularly social indicators in Botswana outperformed those of Sub- among youth. Saharan Africa as a whole by a wide margin. Life Having emerged from 16 years of devastating civil war in expectancy in Botswana was 60 years as of 1990, 10 years 1992 as one of the poorest countries in the world and with above the African average, while the under-five mortality the second-lowest Human Development Index score, the rate had fallen to about 45 per 1,000, compared with 180 economic situation in Mozambique was bleak. As in for Africa as a whole.4 Uganda, Mozambique implemented a series of macroeco- nomic and structural policy reforms shortly after the war Creating or repairing government ended, in addition to encouraging support from the interna- tional donor community. The turnaround in Mozambique The second set of case studies details how countries have between 1993 and 2009 was profound: growth averaged recovered from what is arguably the biggest government more than 8 percent a year (compared to 0 percent from failure of all—civil conflict. In Liberia and Sierra Leone, 1981 to 1992), the poverty rate dropped dramatically, and governments have been successfully rebuilt from the rubble social indicators improved. Double-digit growth has been of devastating wars. In Somalia, on the other hand, which observed in mining, manufacturing, construction, electric- has existed without an internationally recognized central ity, gas, and water. Sound macroeconomic management government since 1991, traditional power structures have allowed Mozambique to attract an increasing amount of been used in the breakaway region of Somaliland to foreign direct investment, which increased from an average strengthen infrastructure. of 1.5 percent of GDP in 1993–98 to 5.2 percent of GDP in A long, violent conflict that ended in 2003 left Liberia in 1999–2010. While progress has been impressive, in the years a state of collapse. Corruption was rampant, external debt ahead, Mozambique will need to determine how to create was approximately 800 percent of GDP, the government’s OVERVIEW 11 ability to manage public finances had essentially disinte- making transporting goods through and out of the country grated, and the Liberian people received among the lowest a risky ordeal. But Somaliland, which unilaterally declared amounts of per capita public spending in the world— its secession from Somalia the same year the central govern- approximately $25 a year. In the years since, the country’s ment collapsed, has made admirable progress in improving reform agenda has vastly improved the governance and eco- its infrastructure to foster trade—namely, in expanding the nomic environment. Liberia’s situation took a solid turn for port of Berbera, on the Gulf of Aden, and in ensuring the the better at the beginning of 2006. Under the direction of usability and security of the highway that leads to the port. the newly elected president, Ellen Johnson Sirleaf, and in Travel on the road is now reliable enough that Ethiopia conjunction with a host of international donors, Liberia transports a large portion of its exports along it before ship- embarked upon a broad-ranging recovery and reform plan. ping them from Berbera. Somaliland’s success in building Improving revenue collection and establishing an effective up its infrastructure stands in sharp contrast to the situation national expenditure process were the initial areas of focus. in the rest of Somalia, which remains tense and dangerous. A unique feature of the program was its cosignatory Factors contributing to Somaliland’s success were twofold: arrangement, under which no major public financial trans- use of the traditional social structures to control violence, actions could take place without first being scrutinized by and successful political cooperation in order to instill the both a Liberian manager and an international advisor. This level of security needed to allow development and operation element of balance was fundamental to the functioning of of the port. While Somaliland is an extreme case, it is the overall reform program. While the outcomes have been undoubtedly representative of scores of cases across Sub- generally positive—revenue collection and expenditure Saharan Africa in which local communities succeed in pro- management has improved dramatically—a final lesson is viding leadership the official government is unable or that a lot of capacity building still has to be done in the area unwilling to provide. of economic governance in Liberia. Neighboring Sierra Leone was in a situation similar to Rationalizing government involvement in markets that in Liberia. More than a decade of civil war in Sierra Leone came to a close in 2002, leaving the little infrastruc- The third set of case studies illustrates how governments ture the country had in shambles, measures of human have successfully intervened to correct market failures—and development among the worst in the world, and capacity for no more. In the cases of mangoes in Mali, gorilla tourism in economic and political governance virtually nonexistent. Rwanda, and apparel production in Lesotho, the respective The process of restoring (and indeed, improving) the system governments stepped in to provide the elements necessary of political governance in Sierra Leone was a crucial part of for the private sector to thrive. Responding to the possibil- the country’s postconflict reform program, because margin- ity of profits, private entrepreneurs took on the bulk of the alization of people living outside the capital, Freetown, was a work to develop the sectors in question. In Malawi meas- main cause of the conflict. A national reform program, ured government involvement in agriculture has produced begun in 2004, focused on devolving political, fiscal, and positive outcomes, while in Mauritius strategic government administrative power to local governments. In some respects, involvement in several sectors over the course of three the reform process has been quite successful—access to decades played a catalytic role in the country’s private sector quality health services has improved dramatically, for one, development. Across Sub-Saharan Africa, the government’s and previously marginalized groups, such as women and pullback from the power and information and communica- ethnic minorities, have benefited from the new space for tion technology sectors has allowed more people to take political participation. But in other areas positive results are advantage of those services. The ICT sector, especially, has less clear. In addition, recent developments within the cen- experienced a dramatic take-off. In all of these cases, the tral government indicate there may be increasing pressure governments provided services essential for private sector to pull back from the decentralization, potentially reducing success and then stepped back to allow private sector actors the future successes of the initiative. to carry out sectoral development. Even in difficult sectors Finally, the case study on Somaliland is an illustration of such as health, initiatives that balance private and public traditional structures playing the role of government where sector involvement have produced truly impressive results, no formal system of government existed. Following the col- as discussed in the case study on malaria control. lapse of the central government of Somalia in 1991, much of In the 1970s Mali, where the vast majority of the popula- the country fell into the domain of bandits and warlords, tion works in agriculture, began to explore opportunities to 12 OVERVIEW export its fresh mangoes. Malian mangoes quickly found a country and an important source of jobs. While interven- market in specialty stores in France and elsewhere in tion in the maize market, including input subsidies, has Europe. But Mali is landlocked, so traditionally, it either been a longstanding—though often contentious—feature depended on neighboring countries for road, rail, and port of government and international donor strategies to pro- infrastructure for export purposes or shipped goods from mote agricultural productivity and food security in Malawi, within its own borders via expensive air freight. By the early the 2005 program went further than past efforts in its scale 1990s, with demand for mangoes in Europe increasing and and scope. Vouchers for maize fertilizers were provided to countries such as Brazil (where shipping costs were lower more than half of Malawi’s farm households. Additional because of good ocean transport options) becoming bigger vouchers were distributed for improved maize seeds and producers, Mali grew increasingly uncompetitive in the tobacco fertilizers. The core objective of the program was global market for mangoes. Less expensive, more efficient twofold: improving farmers’ ability to achieve food self- methods of transportation were needed. Starting in 1995 sufficiency, and boosting farmers’ incomes through the government in partnership with the private sector— increased crop production. Without a doubt, Malawi’s input farmers, farmer groups, and other businesses—launched an subsidy program has had positive impacts: maize produc- innovative response: a multimodal supply chain overhaul tion is estimated to have increased by 26–60 percent under that used temperature-controlled containers to transport the program, while food availability has improved and real mangoes and improved road, rail, and sea links with neigh- wages in the agricultural sector have increased. But doubts boring countries. The level of government involvement in about the effectiveness and appropriateness of the program the mango sector turned out to be ideal, and the efficiency have also been raised—namely, that the fiscal costs of the improvements have led to a host of positive outcomes: a 150 program are high, as are its opportunity costs (in terms of percent increase in the price mango producers receive for crowding out other needed public investments), and that it their products, a 1,000 percent increase in the tonnage of may be impeding the development of sustainable commer- mangoes exported between 1993 and 2008, and a reduction cial agricultural input markets. in the average transit time for mangoes between Mali and With the specter of the genocide still looming in the back- Europe, from 25 to 12 days. ground in 1994, the Rwandan government embarked upon a Another landlocked country in Sub-Saharan Africa, national recovery strategy that included the redevelopment Lesotho, was also struggling to export in the 1990s, thanks of the tourism sector. The goal was to focus on attracting to its limited transportation infrastructure, underdeveloped high-end, conservation-oriented tourists interested in view- factor markets, and inadequate backward and forward ing Rwanda’s gorilla population. The strategy also included industrial linkages and technical expertise. The solution an international marketing campaign intended to improve came in the form of an international mandate, the U.S. Rwanda’s image in the world. Importantly, it also called for African Growth and Opportunity Act (AGOA). In the early near-complete privatization of the hotel and leisure sector. 2000s Lesotho launched a series of aggressive investment The results have been impressive. Tourism has emerged as and export promotion strategies that positioned it to capi- Rwanda’s top foreign currency earner, ahead of the coffee talize on the apparel industry benefits of AGOA. One and tea sectors. The number of visitors to Rwanda’s national notable aspect of Lesotho’s experience is the government’s parks has increased exponentially—from 417 in 1999 to decision to combine an apparel industry competitiveness 43,000 in 2008. Moreover, it is clear in the case of Rwanda initiative with a series of early-stage incentives for investors that initial actions by the government to revive the tourism (they were offered publicly owned factory shells at subsi- sector have encouraged active involvement of the private dized rents). The tactic has yielded positive results—namely, sector: 86 percent of all new tourism-related projects in the that Lesotho’s apparel exports to the United States, at $177 country are locally owned. per capita in 2009, are higher than in any other apparel- The case study on Mauritius discusses government producing country in Sub-Saharan Africa. involvement in markets not over several years but several In Malawi a government initiative in the agricultural sec- decades. Much has been made of the Mauritian “miracle,” tor has had positive results within just a few years. Launched whereby the island transformed itself from a poor, sugar- by the Malawian government in 2005 in response to severe dependent economy at independence in 1968 into what it is food security difficulties in the early 2000s, the agricultural today: a fast-growing, much more diversified economy with input subsidy program was intended to reverse the low pro- one of the highest per capita income levels in Africa. Along ductivity and high price of maize, a staple food for the the way, human development indicators such as life OVERVIEW 13 expectancy (73 years in 2008, compared to 62 years in 1970) the ability of people across Sub-Saharan Africa to commu- have registered impressive improvements. And unlike in nicate. Phone services are now affordable to the majority of many other fast-growing economies, income inequality in Africans rather than a privileged few. Private sector involve- Mauritius has declined solidly over the past 40 years. Aside ment in the ICT sector has also allowed innovative, mobile- from prudent fiscal, monetary, and exchange rate policies, based services in the banking, agricultural, and health sectors Mauritius’ economy has benefited from vibrant partnership to take hold quickly. between the public and private sectors. Among the most vis- ible of the results of these partnership efforts has been the Listening to the people establishment of export processing zones (EPZs) to push along development of the light manufacturing sector in the The final category of case studies examines policies and 1980s and policies supporting growth in the financial ser- innovations that were successful in part because end users vices and ICT sectors in recent decades. Another important were consulted in the process of developing them. Two of key to success in Mauritius has been the country’s ongoing the case studies in this section discuss how the provision of efforts to forge consensus between the Franco-Mauritian tangible products—human-powered water pumps and business elite and the Indo-Mauritian political elite. new varieties of high-performing hybrid rice—have con- In the health sector, too, an appropriate level of govern- tributed to a reduction in food insecurity across Sub-Saharan ment participation in markets has produced positive out- Africa. Four other cases studies detail important successes comes in Sub-Saharan Africa. One example is in the battle at the country level: abolishing school fees to achieve uni- against malaria. Governments, communities, donors, and versal primary education in Uganda; providing health ser- individuals are increasingly coordinating their responses to vices in underserved rural areas in Ethiopia; implementing the disease, reducing duplicative efforts, increasing the a performance-based financing system to improve health capacity to mobilize resources, raising awareness of the coverage in Rwanda; and reducing the cost of transferring problem, and creating a network of technical and imple- money in Kenya through M-PESA, a mobile-phone-based mentation experts. Efforts under the umbrella of the Roll electronic payments system. In some cases, similar innova- Back Malaria Partnership, for example, have been a key in tions had been previously attempted but failed. The differ- reducing the incidence of malaria in numerous Sub-Saharan ence this time around was the way governments went African countries, such as Eritrea, Ethiopia, Rwanda, and about policy making. Rather than assuming they under- Zambia by promoting the use of insecticide-treated bed stood what worked well for their people and implementing nets, managing malaria vector breeding sites, and provid- policy from the top down, governments elicited feedback ing widespread diagnostic and treatment services. In from the public in the course of formulating and imple- Eritrea, for example, malaria morbidity dropped from menting the policy. They adopted bottom-up approaches. about 100,000 cases in 2000 to about 8,000 in 2008. In As a result of being included in the policy-making process, Zambia the percentage of under-five children with people were much more likely to benefit from the final malaria parasite prevalence fell from 22 percent in 2006 to product or service. Finally, the case study on family plan- 10.2 percent in 2008; the number of deaths caused by ning programs illustrates that the responsiveness to users malaria declined 47 percent over the same years. These of family planning services is a prerequisite to reducing emerging successes notwithstanding, more remains to be fertility. done to control malaria in Sub-Saharan Africa. Founded in Kenya in 1991, KickStart International’s mis- Finally, the information and communication technology sion is to promote economic growth and employment cre- revolution that has occurred across Sub-Saharan Africa ation in Sub-Saharan Africa by developing and promoting since the late 1990s is the result of government moving out technologies that can be used to establish and operate prof- of the way of the private sector. With demand for mobile itable small-scale businesses, including its range of low-cost, phones increasing, most governments in Africa switched human-powered pumps that allow farmers to irrigate their from being the monopoly provider (as they were for land- land. In turn, pump users are able to increase their yields lines) to being only the regulator in the mobile phone (providing more food for their families), earn additional industry. As a result, the private sector was able to move in, income, and create agricultural jobs. An important part of increasing competition and reducing costs. The number of KickStart’s model is the participatory fashion in which mobile phone subscribers grew exponentially, from 4 mil- products are developed, with farmers advising on marketing lion in 1998 to 259 million in 2008, dramatically improving and, in the case of the SuperMoneyMaker pump, the design 14 OVERVIEW itself. This element has been key in attracting pump cus- rural poor in Ethiopia, where maternal and child mortality tomers and, as a result, improving agricultural returns. ratios were among the highest in the world and the number That participatory spirit also contributed to the success- of trained health workers was inadequate, desperately ful uptake of the New Rice for Africa (NERICA) hybrid needed better health services. Seeking innovative ways to varieties in the 1990s. The rationale behind developing the respond to this challenge, policy makers came upon the idea hybrids was quite clear. Though rice is the third-highest of creating a national network of health extension workers source of caloric intake and the fastest-growing food staple who could address the basic health needs of the rural popu- in Africa, up to 40 percent of rice is imported. From the lation. The program, launched in 2003, involved training start, the developers of NERICA, the Africa Rice Center and and deploying more than 34,000 (predominantly female) a consortium of partners, intended for it to increase food community health workers who provide essential services security and farmers’ incomes. Importantly, the team con- covering hygiene, disease control and prevention, family sulted extensively with farmers in the course of developing planning, and health education. Although full implementa- the new varieties. That extra step ensured active adoption of tion of the program was completed only in 2010, the out- the new varieties in a relatively short time frame. Initial comes thus far have been extremely promising. Childhood empirical evidence from five Sub-Saharan African countries vaccination coverage has increased dramatically: 62 percent indicates that NERICA has had positive impacts on yields— of children were fully immunized as of 2010, representing especially in Benin and The Gambia—and on household an average annual increase of 15 percent since 2006. Prena- income. Female farmers, in particular, have profited from tal and postnatal maternal health coverage has improved. cultivating NERICA varieties. Women who previously desired family planning services In Uganda free primary education was abandoned in the but were unable to find them now have access to those ser- 1980s because the government’s money was not reaching vices. Use of insecticide-treated bed nets increased 15-fold schoolchildren. Budgetary leakages were ubiquitous, and between 2004/05 and 2009/10, so that 68 percent of house- teacher performance abysmal. But after more than a decade, holds in malaria-affected areas are now protected by at least it was clear that charging school fees was not working either. one bed net or indoor residual spray; the reduction in When Uganda reintroduced free primary education in the malaria incidence in recent years is attributable to expanded 1997, it not only provided an incentive for parents to send coverage of these malaria interventions. their children to school but also publicized the amount of In Rwanda, too, listening to what people need has money each school district was receiving as a way for the cit- worked in the health sector. In addition to broad health sec- izens to hold the government accountable for the program. tor reform and fiscal decentralization, Rwanda made use of By quantitative measures, the universal primary education two innovative policy tools—community-based health initiative in Uganda has been a resounding success. Net pri- insurance and performance-based financing—to increase mary enrollment rose from 57 percent to 85 percent in a access to health services. The first policy made health ser- single year, from 1996 and 1997, and to 90 percent in 1999. vices more available in rural areas and available at a lower Importantly, poor children, girls, and rural residents have cost. Under the second policy, Rwanda pays doctors a bonus benefited disproportionately from the increase in access. depending on the number of children immunized or the Over time, the quality of primary education in Uganda has number of pregnant mothers examined, more closely link- also improved: between 2003 and 2010, for example, the ing rewards with performance. Rwanda’s innovative policy numeracy rate among students at the Primary 3 level rose tools have contributed to the profound improvements in from 42 to 72 percent, while the literacy rate increased from health indicators in recent years. Under-five mortality, for 36 to 58 percent. That said, there is still considerable space example, dropped from 196 to 103 per 1,000 live births for improvement in Uganda’s education sector, particularly between 2000 and 2007, and the maternal mortality ratio in reducing overcrowding in schools, ensuring adequate fell by more than 12 percent each year from 2000 to 2008. supply of teachers and materials, and reducing the number Rwanda is on track to reach the maternal health Millen- of over-age students in primary schools. nium Development Goal by 2015. Improvements among Cases in which governments acted with sensitivity to the the most vulnerable segments of the population have been people have also brought dramatic improvements in the particularly evident. health sector. In Ethiopia, a country of 80 million people, In crafting family planning programs, being responsive mostly in rural areas, more than half lived more than 10 to the needs of users is a prerequisite to reducing fertility. kilometers from a health services facility as of 2005. Yet the Sub-Saharan Africa has the highest total fertility rate in the OVERVIEW 15 world, but some countries in the region are undergoing From a political economy perspective, poor policies per- dynamic and unprecedented fertility transitions. In coun- sist because the preferences of those in power are not tries in which the greatest progress has been made, some of aligned with those of the public, and the former use bad or the key ingredients of success have been a high level of inefficient policies for political purposes or financial gain, political commitment, strong country-level institutions, or both (either politicians are beholden to interest groups or and effective family planning service delivery strategies. politicians create interest groups that are beholden to Family planning programs that have delivery points them). Individuals and groups who expect to lose economic throughout the country; provide a range of contraceptive rent from policy correction are likely to resist change or methods; ensure easy availability of contraceptives; adopt a weaken its effects. Accountability mechanisms—checks and reproductive health approach; and reach adolescents, men, balances—on policy makers are designed to reduce incen- and unmarried people are most likely to accelerate progress tives for opportunistic behavior and distortionary policies. toward fertility decline in Africa. Many countries in Africa But coordination and collective action problems prevent the have tried community-based distribution of contraceptives establishment of strong accountability mechanisms. The to extend family planning to hard-to-reach populations, weaker the constraints on politicians, the less likely reforms particularly in rural areas. Community depots, mobile clin- are to occur. Conversely, where political constraints are ics, women’s groups, and both paid and volunteer village strong, bad policies will not persist, so the scope for reform health workers are some modes of service delivery utilized to have a major improvement is low (Acemoglu et al. 2008). by such programs. However, policy change can occur in an environment of The phenomenal success of M-PESA’s mobile payments weak political constraints when there is a big change in system is the quintessential example of (the private sector, in political power.5 Such a change disrupts the existing status this case) providing a service that poor people in Kenya quo and shifts power from existing interest groups to new sorely needed—the ability to transfer small sums of money ones and possibly to a broader group of people. That is what to people in remote areas at a low cost. Built on a mobile seems to have happened in the case of several countries is phone platform, M-PESA filled that niche, enabling cus- the study. tomers to send money and store money through a simple For example, in Ghana there was a shift from military interface. The service has been wildly successful. Launched rule to democracy. Under military rule the government in 2007, the number of subscribers surpassed 9 million in had an incentive to keep agricultural and import prices late 2009. Recent figures indicate that M-PESA handles $320 low because its power base was the urban elite, including million in person-to-person transfers a month, or roughly the army. But when democracy came, politicians needed 10 percent of Kenya’s GDP on an annualized basis. the rural vote, so it was in their interest to liberalize the exchange rate and allow agricultural prices to rise. In Rwanda the situation was different. The liberalization of WHY DID REFORM HAPPEN AND WHY WAS the coffee sector was part of the postgenocide govern- IT EFFECTIVE? ment’s attempt at restoring growth and rebuilding trust in The above findings beg two important questions: what government among all the people. prompts change in policies, and why did policy reforms Likewise, in Uganda and Mozambique, both of which work? In all but one of the cases of correcting massive gov- had been devastated by decades of political instability and ernment failure, existing policies were standing in the way of civil war, major political reforms occurred in a postconflict growth. How then did change come about? How were the environment. In both cases the new stability allowed new obstacles to correcting government failure overcome? To governments to adopt broad-based reforms to reverse eco- address these issues one needs to understand why failed poli- nomic decline and boost economic development. cies were present to begin with. Central to this issue is polit- Botswana, on the other hand, represents a case where the ical context—that is, the political economy factors that country was able to avoid bad policies after independence, induced the prior bad policies. There is a vast literature despite the discovery of resource riches. One reason why the examining the political economy of poor policies, and study- elite did not capture resource rents was the presence of tra- ing it could help countries recognize when reforms could ditional institutions—the Tswana tribal tradition of work and how to design reforms. Although an assessment of consultation, known as kgotla—which emphasized that the the political economy dimension is outside the scope of this government exists to serve the people and promote develop- study, some observations can be made based on the findings. ment and is not the instrument of one group or individuals 16 OVERVIEW for the purpose of getting hold of the wealth. Tswana tradi- offset by reforms elsewhere as powerful interest groups tion also respected private property; the fact that many of opposed to the reforms attempt to thwart their effect. the tribal leaders who helped usher in modern government Simultaneous reforms that were complementary meant that were also large cattle owners may have reinforced this the effect of any one reform was likely to be enhanced. respect. These traditional institutions promoted respect for Another factor is the level of commitment to reform. property rights and the rule of law, reducing incentives for Here, it is important to consider what prevents the reform distortionary policies. from being reversed. The case studies on growth find that Although transformation or creation of functioning policy makers were generally able to commit to appropriate institutions takes time, Radelet (2010) finds strong evidence reforms—especially macroeconomic stabilization. One of improved governance and more accountable govern- could reasonably argue that a benign global environment of ments in African countries that have achieved steady eco- strong growth, rising commodity prices, and low interest nomic growth and lower poverty rates since the mid-1990s. rates provided an enabling environment for sensible macro- As noted above, stronger constraints on politicians reduce economic and structural reforms. incentives for inefficient policies.6 It would thus appear that strengthening institutions can be extremely effective in CONCLUDING REMARKS bringing about change. Many of these countries are also identified in our study as experiencing successful growth Whether viewed through aggregate indicators or the case episodes (table 2). studies in this book, Africa’s growth and development over the The range of reforms—exchange rate liberalization, past decade has been impressive. The question on everyone’s opening up of trade, reduction of barriers to market entry, mind is: Will it be sustained? By showing that progress has and liberalization of input and output markets—are not come about through a combination of policy reform and new. But why do these reforms work in some cases and fail active government interventions—balancing market failure in many other instances? Country context is one reason: it is and government failure—the 26 case studies included here central to understanding what works. A few insights are give cause for optimism. If, as seems to be the case, African provided by the case studies. One possible factor could be governments are reforming large, distortionary policies, and that the reform was part of a broader set of reforms. This replacing them with selective interventions where there is gen- meant that successful policy reform in one area was not uine market failure and, more important, intervening with the being largely offset or negated by reforms elsewhere—what feedback of the ultimate beneficiary to avoid government fail- Acemoglu et al. (2008) call a “seesaw” effect. The seesaw ure, then there is a good chance that the continent’s strong occurs when successful policy reform in one area is largely economic performance of the last decade will be sustained. NOTES Table 2 Countries with Strong Economic 1. Several studies find a shift in trend beginning in the Performance Saw an Improving Trend mid-1990s. in Institutions 2. World Bank 2005; Growth Commission 2008; IFPRI Freedom House Polity IV Score 1.0 = best score 10 = best score 2009; Center for Global Development 2007. 7.0 = worst score –10 = worst score 3. The National Bureau for Economic Research Africa Country 1989 2008 1989 2008 Project, which is ongoing, identifies and analyzes a large Botswana 1.5 2.0 7 8 number of economic development successes in the region— Burkina Faso 5.5 4.0 –7 0 35 cases have been already selected for in-depth study and Liberia 5.5 3.5 –6 6 40 are anticipated—to better understand the factors behind Mozambique 6.5 3.0 –7 6 the positive experiences and to evaluate the sustainability of Rwanda 6.0 5.5 –7 –3 Sierra Leone 5.5 3.0 –7 7 the successes and their transferability to other African coun- Tanzania 6.0 3.5 –6 –1 tries. The study covers four broad topics: macroeconomic Uganda 5.0 4.5 –7 –1 dimensions of growth; microeconomic aspects of growth; the intersection of health and growth; and cross-regional Source: Radelet 2010. Note: Polity IV measures qualities of executive recruitment con- comparisons. The full findings of the 4.5 year project straints on executive authority, and political competition. Freedom (2007–12), which is being supported by the Gates Founda- House measures political rights and civil liberties. tion, will be available in a few years. OVERVIEW 17 4. Life expectancy and health indicators have deteriorated Commission on Growth and Development. 2008. in the face of the HIV/AIDS epidemic. The Growth Report: Strategies for Sustained Growth 5. Politicians can also adopt reforms when their earlier and Inclusive Development. Washington, DC: World policies prove to be utterly ruinous to the economy and Bank. changing course seems to be the only viable option to avoid Devarajan, Shantayanan, David Dollar, and Torgny political change. Jones and Olken (2007) find empirical evi- Holmgren. 2001. Aid and Reform in Africa: Lessons from dence that assassination of autocrats affects institutional Ten Case Studies. World Bank, Washington, DC. change—specifically, a move toward democracy. IFPRI (International Food Policy Research Institute). 2009. 6. Where there is competition, inefficient institutions will Millions Fed: Proven Successes in Agricultural Develop- be eliminated ( Kingston and Gonzalo Caballero 2008). ment. Washington, DC. Jones, B., and B. Olken. 2007. “Hit or Miss? The Effect of Assassinations on Institutions and War.” NBER Working REFERENCES Paper 13102. National Bureau for Economic Research, Acemoglu, Daren, Simon Johnson, Pablo Querubin, and Cambridge, MA. James A. Robinson. 2008. “When Does Policy Reform Kingston, C., and G. Caballero. 2008. “Comparing Theories Work? The Case of Central Bank Independence.” of Institutional Change.” Journal of Institutional Econom- Brookings Papers on Economic Activity (Spring): ics 5: 151–80. 351–418. Radelet, Stephen. 2010. Emerging Africa: How 17 Countries Arbache, J., D. Go, and J. Page. 2008. “Is Africa’s Economy at a Are Leading the Way. Washington, DC: Center for Global Turning point?” In Africa at a Turning Point? Growth, Aid, Development. and External Shock ed. D. Go and J. Page, 13–85. World UNIDO (United Nations Industrial Development Organi- Bank. Washington, DC zation). 2009. Industrial Development Report 2009: Arbache, J. and J. Page 2007. “More Growth or Fewer Breaking In and Moving Up. Vienna: UNIDO. Collapses? A New Look at Long-Run Growth in Sub- World Bank. 2000. Can Africa Claim the 21st Century? Saharan Africa.” Policy Research Working Paper 4384. Washington, DC: World Bank. World Bank, Washington, DC. ———. 2005. Scaling Up Poverty Reduction: Learning Block, S. 2010. “The Decline and Rise of Agricultural Pro- and Innovating for Development: Findings from the ductivity in Sub-Saharan Africa Since 1961.” NBER Shanghai Global Learning Initiative. Washington, DC: Working Paper 16481, National Bureau for Economic World Bank. Research, Cambridge, MA. ———. 2008. World Development Report 2008: Agricul- Center of Global Development. 2007. Case Studies in Global ture for Investment. Washington, DC: World Bank. Health: Millions Saved.” Washington, DC. ———. 2010, Africa’s Infrastructure: A Time for Transfor- Cohen, Jessica, and William Easterly, eds. 2009. What Works mation. Washington, DC: World Bank. In Development?: Thinking Big and Thinking Small. World Bank and IMF (International Monetary Fund). 2011. Brookings Institution, Washington, DC. Global Monitoring Report: Improving the Odds of Achiev- Fuglie, Keith. 2010. “Agricultural Productivity in Sub-Saharan ing the MDGs. Washington, DC: World Bank. Africa.” U.S. Department of Agriculture, Washington, DC. 18 OVERVIEW PA R T I Tracking Successful Growth Experiences CHAPTER 1 Tanzania: Growth Acceleration and Increased Public Spending with Macroeconomic Stability David O. Robinson, Matthew Gaertner, and Chris Papageorgiou n the space of a few decades, Tanzania moved from I TANZANIA’S GROWTH TAKE-OFF colonialism to independence—including the union of Sub-Saharan Africa enjoyed relatively rapid growth over the two states (Tanganyika and Zanzibar)—to socialism to decade leading to the 2009 global financial crisis. After many a market-oriented developing economy. Each of these years of poor performance, it grew faster than developed stages involved significant change, with different economic countries between 1995 and 2005. institutions and economic incentives. Tanzania experienced sustained growth acceleration over Tanzania has emerged from this period of significant this period (figure 1.1). This section examines when and economic transition as one of the most rapidly growing why this acceleration occurred. economies in Sub-Saharan Africa. For the first time since Most low-income countries exhibit frequent phases of independence, it has broken out of the cycle of short- growth, stagnation, and decline (Pritchett 2000), even where lived accelerations in growth that has characterized many key policies and country characteristics have been relatively low-income countries, enjoying strong uninterrupted stable (Easterly et al. 1993). A technique developed by Bai growth since the mid-1990s. During the period between and Perron (1998) can be employed to identify structural 1992/93 and 2008/09, inflation remained in single digits; “up” and “down” breaks in economic growth trends, captur- the debt burden fell dramatically; the level of public ing this “stop and go” behavior of economic growth in low- spending increased significantly, permitting expansion income countries. A break in growth is identified as the of public services; and international reserves rose point after which the average growth rate is above or below sharply (annex figures 1.A1 and 1.A2). the previous average growth rate. Berg, Ostry, and Achieving and sustaining rapid growth while preserving Zettelmeyer (2008) extend this methodology to identify macroeconomic stability represents a major achievement, “growth spells” by modifying the procedure to determine and significant improvements have taken place in many sample-specific critical values when the time dimension is aspects of development. But Tanzania remains a low- 30 years or less, the typical time horizon in low-income income country, with per capita gross domestic product countries growth series. This methodology was used here to (GDP) of only $550 in 2009, and it is on track to meet only identify episodes of growth acceleration for all countries in about half of the Millennium Development Goals (MDGs). Sub-Saharan Africa for which sufficiently long data series The challenge is to harness the country’s enormous poten- were available (table 1.1). The results reveal a large number of tial to increase growth and thus to provide opportunities for countries with growth accelerations in the 1990s (11 of 25).2 all to enhance living standards.1 21 Figure 1.1 Real GDP Growth in Tanzania, 1970–2009 10 8 6 Percent 4 2 0 –2 70 75 80 85 90 95 00 05 09 19 19 19 19 19 19 20 20 20 Source: Tanzanian authorities. making during this period was to balance the need to create Table 1.1 Countries in Sub-Saharan Africa Experiencing “Growth Spells” a supportive environment for growth against the need to in the 1990s contain the potential vulnerabilities that have derailed eco- nomic booms in the past, in Tanzania and elsewhere. Tanza- Year of Country acceleration nia’s macroeconomic record has been remarkably successful Burkina Faso 1998 during a period that has seen both external shocks (sharp Cameroon 1994 fluctuations in global commodity prices, such as oil and Djibouti 1998 food, and a global financial crisis) and domestic shocks Equatorial Guinea 1994 Ghana 1997 (periodic droughts, bank failures, and governance scandals). Liberia 1994 Navigating such shocks without major macroeconomic dis- Mozambique 1995 ruption is a reflection of strong institutions and responsible Namibia 1998 Rwanda 1994 policy making. South Africa 1995 In addition, Tanzania’s move to a higher growth trajec- Tanzania 1996 tory came following a period of substantial economic Source: Authors’ compilation, based on data from Berg, Ostry, and reform. Three distinct phases in economic policy making Zettelmeyer 2008. can be identified (figure 1.2).3 The first, which began at the time of the Arusha Declaration in 1967, was the period of Ujamaa socialism, which created a one-party system with Tanzania’s growth take-off was spurred by several key state control of the economy and nationalization of all factors, including the significant structural changes that major enterprises. This period ended in the mid-1980s, with occurred as the basic institutions of a market economy—a attempts to gradually introduce key components of a private banking system, the unification of the exchange rate, market-oriented economy. and price liberalization—were introduced. Relative to other The pace and breadth of key reforms—both absolutely countries that have experienced growth accelerations, and relative to the rest of the world—can be highlighted Tanzania stands out in two respects. First, the extent of the by using newly constructed indexes of structural reforms reforms was much broader and larger than the average of developed by the Research Department of the Interna- other countries that saw growth accelerate. Second, there tional Monetary Fund (IMF) covering several kinds of was a long lag between key reforms and the realization of real (trade, agriculture, and networks) and financial the growth acceleration. The role of macroeconomic policy (domestic finance, banking, securities, and capital account) 22 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.2 Chronology of Transformation of Tanzania’s Economy, 1967–2009 1967–1985: Ujamaa Socialism 1986–1995: Liberalization 1996–: Macroeconomic and Economic Decline and Partial Reforms Stabilization and Structural Reforms • State control of the economy • Liberalization of exchange and and state ownership of all major trade regimes • Privatization and reform of enterprises parastatals • Liberalization of agricultural marketing • Exchange rate and pricing • Liberalization of financial sector system and domestic prices policies based on non-market • Creation of market-oriented POLICIES mechanisms • Initiation of financial system reform regulatory framework • Devaluations and • Trade reform, regional integration expansionary fiscal and • Initiation of parastatal and civil service monetary policies reforms • Reversal of fiscal dominance of monetary policy • Mismanagement of exogenous shocks (terms of trade and • Fiscal consolidation droughts) • Sizable financial assistance from • External trade and forex donors (debt relief, grants and controls concessional loans) • Low export and real GDP growth • But: large segments of the • Higher broad-based real GDP growth • Deterioration of physical economy still dominated by public • Inflation declined to single digits infrastructure monopolies • Strong growth in nontraditional exports • Loss-making state enterprises, and • Insolvency of large state-owned and turnaround in balance of payments large subsidies financed by directed banks and losses in other • Large increase in international reserves bank lending parastatals OUTCOMES • Composition of expenditure moved • Budget deficits financed by printing • Persistent weaknesses in toward more allocations for povert-reducing money budgetary management programs • Shortages of goods and high • Large fiscal deficits • Creation of efficient, competitive inflation • Continued accumulation of arrears banking system • Large external imbalances, and monetization of the deficits • Increased credit to the productive exhaustion of foreign reserves, • Elusive macroeconomic stability sectors of the economy and buildup of external arrears • Low growth for most of the period • Rising poverty Source: Nord et al. 2009. indicators. The indexes, which date back as far as the Tanzania and the limited means of communication, it took 1970s, are summarized in annex C and in Prati, Onorato, time to convey information on the reforms throughout the and Papageorgiou (2010). country and to translate the legislative reforms into actual There was a broad trend across much of the world change on the ground. Third, as discussed in Mwase and toward greater structural reform and liberalization during Ndulu (2008), failure to address key bottlenecks prevented the 1990s. Tanzania implemented significant reforms in the the realization of gains from other reforms. The exchange real and financial sectors beginning in the early 1990s and market was unified only in 1994, for example, removing a continuing into the late 1990s, reaching levels well above the severe obstacle to trade, including access to needed averages for Sub-Saharan Africa (an exception is the capital imports (figure 1.5). account index, which lagged the region as a whole). Clear standouts are current account and domestic financial SUSTAINING GROWTH FOLLOWING reforms (annex figure D1), which show extraordinary THE INITIAL ACCELERATION improvements in a very short period. Figures 1.3 and 1.4 document the association between reforms (especially in Tanzania’s economy has grown by 3.5–7.8 percent a year since trade and finance) and growth breaks. 1996, averaging 6.0 percent, well above the rates for Sub- Various factors may explain this pattern. First, given the Saharan Africa as a whole. These rates represent a substantial nationalization of all private property that occurred in increase over earlier growth rates in Tanzania and, unlike pre- 1967, a credibility gap probably existed regarding the irre- vious accelerations in growth, have been maintained over a versibility of reforms. Second, given the physical size of sustained period of time (figure 1.6). CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 23 Figure 1.3 Indexes of Real Sector Reform in Tanzania and Other Countries Experiencing Growth Acceleration a. Trade index 0.75 0.70 0.65 0.60 More Index liberalization 0.55 0.50 Less liberalization 0.45 0.40 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth break b. Current account restrictions index 0.90 0.80 0.70 0.60 0.50 Index More 0.40 liberalization 0.30 Less 0.20 liberalization 0.10 0 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth break c. Agriculture index 0.70 0.65 More 0.60 liberalization 0.55 Less Index 0.50 liberalization 0.45 0.40 0.35 0.30 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth break Average liberalization index Tanzania's liberalization index Source: Authors estimates using GDP data from IMF 2010d. Note: The number of countries used to compute each average varies across indexes, based on data availability. See annex B for a list of countries included in sample. 24 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.4 Indexes of Financial Sector Reforms in Tanzania and Other Countries Experiencing Growth Acceleration a. Domestic finance index 0.80 0.70 Index 0.60 0.50 0.40 More liberalization 0.30 Less liberalization 0.20 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth breaks b. Banking subindex 1.00 0.90 0.80 0.70 0.60 Index 0.50 0.40 More liberalization 0.30 0.20 Less 0.10 liberalization 0 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth breaks c. Capital account index 0.54 0.52 0.50 0.48 0.46 Index 0.44 More liberalization 0.42 0.40 Less liberalization 0.38 0.36 0.34 –5 –4 –3 –2 –1 0 1 2 3 4 5 Years before and after growth breaks Average liberalization index Tanzania's liberalization index Source: Authors estimates using GDP data from IMF 2010d. Note: The number of countries used to compute each average varies across indexes, based on data availability. See Annex B for a list of countries included in sample. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 25 Figure 1.5 Official and Parallel Market Exchange Rates in Tanzania, 1970–95 600 800 700 500 Market premium (percent) 600 Shilling/U.S. dollar 400 500 300 400 300 200 200 100 100 0 0 19 4 19 5 19 6 87 19 8 19 9 19 0 91 19 2 93 94 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 19 9 19 0 19 1 19 2 19 3 8 8 8 8 8 9 9 7 7 7 7 7 7 7 7 7 7 8 8 8 8 19 19 19 19 Parallel market premium Parallel market rate Official rate Source: Nord et al. 2009. Figure 1.6 Real GDP Growth in Tanzania, 1970–2008/09 8 7 6 5 Percent 4 3 2 1 0 5 5 5 6 7 8 9 –8 –9 00 –0 –0 –0 –0 70 86 –2 05 06 07 08 19 19 96 20 20 20 20 19 Tanzania Sub-Saharan Africa Sources: Tanzanian authorities and IMF staff estimates. Several patterns are evident from Tanzania’s growth per- ■ The key sectors contributing to growth have been ser- formance (figures 1.7 and 1.8 and table 1.2): vices and, to a lesser degree, industry; agriculture has not contributed to growth. Within the service sector, ■ Acceleration has been driven by domestic demand, not particularly rapid growth has been experienced in con- exports. Large increases in both consumption and invest- struction, telecommunications, financial services, and ment have been recorded, in both cases reflecting signif- mining—all subsectors that have been liberalized. The icant increases in public spending. limited contribution of agriculture—in a country where 26 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.7 Contributions to Real GDP Growth in Tanzania by Type of Expenditure and Production, 1999–2009 a. Contribution to real GDP growth in Tanzania by b. Contribution to real GDP growth in Tanzania by type of expenditure sector of production 20 15 16 12 12 Percentage points Percentage points 8 9 4 6 0 3 –4 –8 0 99 00 01 02 03 04 05 06 07 08 ) (e 99 00 01 02 03 04 05 06 07 08 ) (e 19 20 20 20 20 20 20 20 20 20 09 19 20 20 20 20 20 20 20 20 20 09 20 20 Consumption Investment Net exports Agriculture Industry Services Sources: Tanzanian authorities and IMF staff estimates. Figure 1.8 Contribution to Real GDP Growth in Tanzania of Labor, Capital, and Total Factor Productivity, 1995–2008 9 8 7 6 Percentage points 5 4 3 2 1 0 95 96 97 98 99 00 01 02 03 04 05 06 07 ) (e 19 19 19 19 19 20 20 20 20 20 20 20 20 08 20 TFP Capital Labor Sources: Tanzanian authorities and IMF staff estimates. about three-quarters of the population resides in rural ■ The acceleration in growth can be traced to a combi- areas, where poverty is concentrated—is a concern. It also nation of higher investment and increases in produc- represents an opportunity, because international experi- tivity. Growth decomposition is always subject to ence suggests that the sector can respond rapidly if the interpretation, particularly regarding the estimation of right incentives and supporting infrastructure are put the capital stock (and therefore the contribution of in place. investment and the interpretation of the residual as CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 27 stimulate agricultural production. Disruptions in the global Table 1.2 Factor Contributions to Real GDP Growth in Tanzania, 1986–2008 rice market in 2008 hit Zanzibar hard, contributing to infla- (percent) tion that reached 27 percent a year in September 2008 before declining to 5 percent a year later. Item 1986–90 1991–95 1996–2000 2001–08 Fairly broad restrictions on capital account transactions Real GDP growth 5.3 1.8 4.3 7.1 Labor force 2.2 2.5 1.7 1.7 remain in place—preventing, for example, foreign portfolio Capital 0.9 1.3 0.3 1.9 inflows to the domestic government securities markets or Total factor private capital outflows into neighboring securities markets. productivity 2.2 –2.0 2.3 3.5 Private capital inflows are nevertheless becoming an Source: Tanzanian authorities and IMF staff calculations. increasingly important source of financing for investment. Foreign direct investment, most of it in the mining sector, averaged 3.5 percent of GDP during 1996–2008. Before the productivity growth). That said, it is clear that the global financial crisis, loan syndications were being used to acceleration in Tanzania was not driven by greater use finance investment, mostly in the telecom sector. of labor.4 Studies of the agricultural sector during this Official capital inflows have been important sources of period show little or no improvement in yields for the financing for Tanzania since independence. Official devel- sector as a whole, with increased output coming from opment assistance peaked on average during 1986–92 (it an increase in land under cultivation. represented almost 25 percent of GDP in 1992) before declining through the 1990s and then rising again, to about Although exports did not lead the growth acceleration— 17 percent of GDP in 2007. The component of aid that goes indeed at just 25 percent of GDP in 2008/09, they remain at through the budget has been rising, albeit with significant a low level—the composition of exports has seen significant year-on-year variability (table 1.3). changes (figure 1.9). Traditional exports (cotton, coffee, and In addition to the absolute level of aid inflows, two other tea) have declined significantly in importance, partly reflect- aspects of aid have had macroeconomic impacts. First, the ing the reorientation of the sector toward meeting the composition of aid inflows has shifted. In line with the gov- consumption needs of a rapidly growing local population. ernment’s stated preferred modality, aid has increasingly At the same time, gold exports went from zero in 1999 to been provided in the form of general budget support—cash $1.4 billion (nearly 40 percent of export receipts) in 2009, that goes directly into the budget—rather than project sup- making Tanzania the fourth-largest gold exporter in Africa port. Such aid gives the government more control over the after South Africa ($6.3 billion), Ghana ($2.6 billion), and use of resources provided by development partners and can Mali ($2.0 billion). The service sector—tourism and other be used to better meet national rather than externally driven services, such as transport—has also grown at a steady rate. (foreign) priorities. However, the use of general budget sup- port also places a heavier burden on government processes to ensure the effective allocation of budgetary resources and The external environment maintain a broad dialogue with development partners on Tanzania’s growth acceleration coincided with a period of results and use of funds to ensure continued access to aid. both unprecedented expansion in the global economy and Its use also complicates macroeconomic management, mak- significant volatility in commodity prices and capital flows ing it critical to ensure that the spending of aid does not (private and official, including both aid and debt relief) that induce undesirable macroeconomic consequences, such as a have posed challenges for macroeconomic management. sharp appreciation of the exchange rate. An early evaluation Given its low level of exports, Tanzania has been insulated of the impact of general budget support in Tanzania noted from fluctuations in prices and demand for exports and the that it facilitated the government’s ability to implement boom-bust cycles such fluctuations have created in other policies to which it was committed without discernable commodity-dependent countries. However, significant adverse macroeconomic consequences, but that it was less impacts have been felt on the import side. The sharp effective in securing commitment to policies (Lawson increase in oil prices during the 2000s resulted in an oil et al. 2005). import bill of $1.7 billion in 2007/08, almost 9 percent of Significant aid has been provided in the form of debt GDP. Recent spikes in the price of fertilizer, most of which relief. Tanzania qualified for the Heavily Indebted Poor is imported, have constrained the government’s ability to Countries Initiative in 2001, receiving debt relief of about 28 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.9 Composition of Exports of Goods and Services in Tanzania, 1996/97–2008/09 100 80 60 Percent 40 20 0 7 8 9 0 1 2 3 4 5 6 7 8 9 /9 /9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 20 20 20 20 20 20 20 20 20 Gold Traditional Other goods Other services Tourism Source: Tanzanian authorities. Table 1.3 Gross External Program and Project Assistance in Tanzania, 1996/97–2007/08 Program assistancea Project assistance Year Grants Loans Grants Loans Debt service relief Total assistance 1996/97 1.7 0.9 3.0 3.2 0.0 8.7 1997/98 1.7 1.3 2.6 2.2 0.0 7.8 1998/99 1.7 1.2 2.5 1.9 0.0 7.3 1999/00 1.7 0.7 2.8 2.8 0.0 8.1 2000/01 1.6 0.5 3.2 2.1 0.4 7.8 2001/02 2.4 0.9 2.8 1.1 0.6 7.8 2002/03 3.2 1.3 2.3 1.1 0.6 8.5 2003/04 3.4 1.7 2.6 2.0 0.6 10.2 2004/05 4.2 1.3 2.7 2.1 0.5 10.7 2005/06 3.5 2.0 2.0 1.9 0.7 10.1 2006/07 3.3 1.7 1.2 2.2 1.2 9.7 2007/08 3.5 2.4 3.0 2.2 0.9 12.1 Source: Tanzanian authorities and IMF staff estimates. a. Includes both general budget support and basket funds. $3 billion. In 2006 the IMF, the World Bank, and the African making Tanzania a potentially attractive destination for Development Bank implemented the Multilateral Debt capital inflows. Cognizant of the difficulties that arose from Reduction Initiative, providing debt relief of an additional the very high debt burden, in 2004 lawmakers introduced a $3.5 billion. formal minimum concessionality requirement in external Debt relief has substantially reduced the debt burden in borrowing by the government, initially established at Tanzania, freeing up budgetary resources for alternate uses. 50 percent but subsequently reduced to 35 percent. It has also left the country with a low level of debt—in absolute and especially net present value terms, given the Fiscal policy highly concessional nature of much of the outstanding stock—generating a low risk of debt distress in conventional The most striking feature of fiscal policy between 1992/93 debt sustainability analyses (see, for example, the joint and 2008/09 was the expansion in public spending. After Bank/Fund debt sustainability analysis in IMF 2010c) and falling to 15–17 percent of GDP in the mid-1990s as inflation CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 29 was brought under control, spending increased sharply in ■ The tax system was simplified, through restrictions on 2001/02, reaching more than 25 percent of GDP in 2008/09 the range of taxes subnational authorities can impose (even higher figures were budgeted for 2009/10 and and the repeal of numerous low-yielding taxes. 2010/11) (figure 1.10). Government spending per capita ■ Relative stability was achieved in key tax rates. Following increased from about $40 in 1998/99 to $150 in 2008/09. initial reforms, the rates of the major taxes were left How was this spending financed without endangering largely unchanged, in part because they were linked to price stability or increasing debt sustainability? Where did regional tax harmonization initiatives.5 the money go? Expenditures are financed by increasing revenues, obtaining financing from abroad (as loans or Improvements in tax administration have been no less grants), or borrowing domestically. Tanzania has used all important. The first step in strengthening administration three sources (figure 1.11). came with the creation of the Tanzania Revenue Authority, Revenue performance has been striking. After an initial in 1998/99. The Tanzania Revenue Authority has been able dip in the second half of the 1990s, revenues increased to progressively modernize tax collections, by introducing strongly, from just 10 percent of GDP in 1999/2000 to self-assessment procedures and electronic filing and pay- almost 16 percent in 2008/09. The improvement reflects a ments and by carefully allocating its own human resources combination of tax policy and administration. into high-return areas. Indeed the Large Taxpayer Depart- Tax policy has focused on reorienting the tax system to ment introduced in 2003 now collects almost three-quarters reflect the changing economic structure, as parastatals were of domestic taxes, with the number of enterprises covered gradually privatized; trade (internal and external) was liber- rising from 98 to more than 370. alized; and a private sector, often informal, emerged. Key There is still scope for improving tax policy. In addition changes included the following: to its revenue-raising role, tax policy has been used as a component of the broader development strategy to pro- ■ A value added tax (VAT) was introduced in July 1998, vide incentives for specific economic activities in an providing about one-third of government revenues. The attempt to attract investment. Exemptions, such as tax hol- rate was reduced to 16 percent in July 2009, bringing it idays, are provided for qualified investments and for com- into line with most other East African countries. panies established in designated export-processing zones ■ Income taxes, which also provide about one-third of gov- or special economic zones. A special regime was also cre- ernment revenues, were progressively modernized, with ated for mining activities, based on a tax/royalty regime the maximum rates reduced to 30 percent, down from but with additional relief from VAT and other taxes, such 70 percent rate in 1991. as the fuel levy. The efficacy of such incentive regimes for attracting investment is the subject of a fierce debate and beyond the scope of this chapter. Recent budgets have also introduced VAT exemptions for specific industries or Figure 1.10 Government Expenditure in Tanzania as products, thereby eroding the tax base and complicating Percent of GDP, 1992/93–2008/09 tax administration. 30 Reliance on domestic financing—initially largely from the central bank, later from commercial banks and social 25 security funds—has decreased significantly, in part out of a 20 desire to contain the rate of monetary growth to bring down inflation. For several years in the mid-2000s, budgets were Percent 15 anchored on a zero net domestic financing target, which provided continued support for inflation stabilization, 10 eased interest rate pressures, and increased room for banks 5 to lend to the private sector. Where there were shortfalls in foreign financing, limited resort was made to domestic 0 financing in order to avoid disruptions in expenditure 3 5 7 9 1 3 5 7 9 programs.6 /9 /9 /9 /9 /0 /0 /0 /0 /0 92 94 96 98 00 02 04 06 08 19 19 19 19 20 20 20 20 20 Even with this dramatic improvement in revenues, Source: Tanzanian authorities. donor dependency has not declined. Government revenues 30 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.11 Sources of Fiscal Expenditure in Tanzania, 1992/93–2008/09 a. Sources of fiscal expenditure as b. Sources of fiscal expenditure as percentage of GDP percentage of expenditure 110 30 100 25 90 80 20 70 Percent Percent 60 15 50 10 40 30 5 20 10 0 0 –5 –10 3 5 7 9 1 3 5 7 9 3 5 7 9 1 3 5 7 9 /9 /9 /9 /9 /0 /0 /0 /0 /0 /9 /9 /9 /9 /0 /0 /0 /0 /0 92 94 96 98 00 02 04 06 08 92 94 96 98 00 02 04 06 08 19 19 19 19 20 20 20 20 20 19 19 19 19 20 20 20 20 20 Revenue Grants Foreign Domestic Source: Tanzanian authorities and IMF staff calculations. still cover only about 50–60 percent of government to about 95 percent, up from about 60 percent a decade expenditures—less than the level of recurrent spending earlier. But issues remain with unfilled vacancies for teach- (see figure 1.11). ers, particularly in rural communities; uneven financial Weak domestic revenue perpetuates reliance on foreign transfers; and concerns about the quality of education, support, potentially adding both uncertainty and volatility with pass rates in standardized examinations declining. In to the government’s ability to deliver its programs. Alterna- the health sector, a recent value-for-money study identi- tively, the low share of government revenues to expenditures fied sharp regional differences in activity levels and per- can be interpreted as an effort by the government to exploit formance, with no linkage to resource allocations the availability of cheap financing sources to expand deliv- (National Audit Office 2010). ery of public services beyond the level possible from domes- tic resources alone. The key is to ensure that efforts are Monetary and exchange rate policies under way to build a broad tax base capable of sustaining the desired level of public services in the event that donor The primary objective of the Bank of Tanzania is to main- funds diminish over time, that these efforts are not reduced tain domestic price stability conducive to balanced and by the presence of aid inflows, and that the additional public sustainable growth of the national economy. A secondary spending is used wisely. objective is to support the integrity of the financial system. The higher spending has appropriately been focused on With inflation reduced to single digits and a financial sys- priorities related to the National Strategy for Growth and tem that has grown rapidly in recent years, the Bank of Reduction of Poverty (known by its Swahili acronym, Tanzania appears on track to meet both objectives. MKUKUTA). The 2009/10 budget allocates 71 percent of It has not been an easy transition. In the late 1980s Tanza- expenditures to MKUKUTA priorities. At the sectoral nia had one of the least developed financial systems in the level, the two sectors receiving the largest budget alloca- world. Credit was allocated centrally; all of the major finan- tions are education and health, which together account for cial institutions were state owned; there were no capital or about 30 percent of budgetary spending. The impact of the money markets, so the Bank of Tanzania simply printed spending is difficult to gauge. In education, for example, money to finance the fiscal deficit and provide liquidity to substantial investment in construction of new classrooms insolvent banks; and foreign exchange was rationed, with a has supported an expansion in primary school enrollment substantial spread between the official and black market rates. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 31 Interest rates and foreign exchange markets were pro- banks were privatized, but the process was lengthy and the gressively liberalized starting in 1991, but the initial distor- former state-owned banks—CRDB (privatized in 1996), tions—the legacy constraints—took time to resolve and NBC (privatized in 2000), and NMB (privatized in 2005)— served to shape economic policy options for most of the continue to dominate the financial system. Indeed, only decade. In particular, with minimal foreign exchange since about 2000 has the private sector been able to access reserves, central bank interventions in the market were ini- credit from the banking system. tially targeted at accumulating reserves while smoothing Credit has grown rapidly since then (figure 1.13). Entry of seasonal volatility in the exchange rate. As the government new banks has accelerated in recent years, with 41 banks was able to rein in its domestic financing needs and donors licensed as of the end of June 2010. Access to financial ser- increasingly moved to providing budget support, the avail- vices remains limited, however—only 11 percent of adult ability of foreign exchange increased, allowing the Bank of Tanzanians held a bank account at end-2009—with financial Tanzania to accumulate reserves, reducing upward pressure services in rural areas increasingly shifting to mobile tele- on the exchange rate (figure 1.12). It did so at the cost of phone systems. assigning the burden of liquidity management to Treasury Financial sector supervision has evolved along with bill sales, however, resulting in rising domestic debt and the changing role of Tanzania’s financial institutions. The upward pressure on nominal interest rates. government and the Bank of Tanzania developed a risk- This unbalanced monetary policy mix, with high yields based supervisory framework that provides sufficient on government securities, low inflation, and exchange rate authority to the body responsible for supervising and stability, proved difficult to sustain, however, particularly regulating the banking system. A bank intervention and given the shallow domestic financial market with a small resolution framework is in place, supported by a deposit number of key players (Abbas, Ali, and Sobolev 2008). In late insurance scheme. These institutions facilitated the clo- 2007 the central bank issued a press release indicating that its sure of four private banking institutions, including two foreign exchange operations were limited to smoothing foreign-owned, without major disruptions to the banking short-term volatility and that the exchange rate could appre- system and with full compensation for individual deposi- ciate. Only thereafter did markets become more balanced. tors.7 The legislative foundation for supervision of the broader financial system—chiefly pension funds and Private banks insurance companies—has been put in place, but imple- mentation needs to be enhanced to guard against the Private banks were formally permitted in 1992, but none buildup of potentially large fiscal liabilities. began operations until 1994. All of the large state-owned Figure 1.12 Foreign Exchange Reserves in Tanzania, 1992/93–2008/09 3,500 3,000 2,500 Millions of dollars 2,000 1,500 1,000 500 0 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 9 /0 / / / / / / / / / / / / / / / 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 Source: Tanzanian authorities. 32 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.13 Private Sector Credit as Percent of GDP in Tanzania, 1992/93–2008/09 18 16 14 12 Percent 10 8 6 4 2 0 19 6 19 7 19 8 19 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 8 9 /9 /9 /9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 20 Source: Tanzanian authorities and IMF staff calculations. USING THE NEWFOUND POLICY SPACE: Despite initial concerns that contributed to a request COUNTERCYCLICAL POLICY RESPONSE for financing from the IMF under the exogenous shocks TO THE GLOBAL FINANCIAL CRISIS facility, Tanzania has managed to contain the impact of the global financial crisis through a macroeconomic policy The degree to which Tanzania, and many other countries in response supported by interventions at the microeco- Africa, have changed in terms of both economic structures nomic level, all laid out in an economic recovery plan and macroeconomic policy space was clearly demonstrated launched in June 2009. At the macroeconomic level, both during the global financial crisis. Past global slowdowns fiscal and monetary policies were eased to help sustain the have been very difficult for Africa: as documented in IMF growth momentum. On the fiscal side, easing of more than (2010a), in previous global slowdowns, Africa suffered both 2 percent of GDP was envisaged, including a reduction in a deeper slowdown and a more gradual recovery. This crisis the VAT rate from 20 percent to 18 percent, partial gov- has been different, with Sub-Saharan Africa as a whole ernment guarantees for loan restructurings in troubled seeing a large but short-lived dip in economic growth sectors, expanded agricultural input subsidies, expanded (figure 1.14). investments in energy and road sectors, and temporary Another key difference is that, for the first time, many exemptions from royalties for tanzanite and diamond countries in Sub-Saharan Africa were in a position to pur- miners. The interventions were specifically targeted to sec- sue countercyclical policies in the face of an adverse shock. tors expected to be hardest hit by the crisis. The fiscal eas- In response to previous shocks, African countries generally ing was financed in part by development partners, several pursued procyclical macroeconomic policies, largely out of of which were able to advance funding from future years, lack of choice: high debt burdens, limited sources of financ- but also in part from domestic sources, with a modest ing, and low levels of international reserves provided few increase in the volume of sales of Treasury bills and excep- options for easing fiscal or monetary policy in the face of a tional direct credits from the central bank. As elsewhere, growth slowdown. In addition, changes in the structure of commercial banks became more risk averse, with credit the global economy—specifically the increased role of growth to the private sector falling sharply and high China and India, both of which are key trading partners for demand for government obligations serving to push inter- much of Sub-Saharan Africa and which were able to contain est rates to record lows. Although the crisis is not yet over, the damage to their economies—mitigated the external preliminary estimates for 2009 point to growth slowing to demand shock experienced by countries in Sub-Saharan 6.0, rising to 6.5 percent in 2010. Africa from the global financial crisis. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 33 Figure 1.14 Real GDP Growth in Sub-Saharan Africa and the World, 1992/93–2008/09 a. Real GDP growth, average of past three economic cycles, 1992/93–2008/09 b. Real GDP growth, current cycle 2005–2013 6 8 5 6 4 4 Percent Percent 3 2 2 1 0 0 –2 t–4 t–3 t–2 t–1 t t+1 t+2 t+3 t+4 05 06 07 08 09 10 11 12 13 20 20 20 20 20 20 20 20 20 World Sub-Saharan Africa Source: IMF 2010d. Note: t = trough in the cycle. WHAT’S NEXT? reforms can drive a growth up-break, they are not sufficient in isolation to sustain one (Hausmann, Pritchett, and Tanzania has managed to achieve and sustain rapid growth Rodrik 2005). while preserving macroeconomic stability for the past Infrastructure is a constraint across Africa. But Tanzania 15 years. Despite this achievement, it is on track to achieve lags even other countries in Africa (table 1.5). only about half of the MDGs (table 1.4) (United Republic of Where will growth come from? Studies of potential Tanzania 2009b). What needs to be done in the years ahead? growth drivers for Tanzania (Mbelle et al. 2010, for example) What macroeconomic policies would be beneficial? highlight a combination of agriculture, tourism, transport, and mining, with a focus on developing manufacturing and value addition activities. The identified sectors largely Accelerating pro-poor growth reflect physical and geographical endowments, with existing Constraints to growth in Tanzania are similar to those seen activities in these sectors fairly nascent—only a small por- in most countries: inadequate infrastructure, regulatory tion of Tanzania’s natural resource endowments are being bottlenecks, skill shortages, and deficiencies in the legal exploited—and often characterized by low productivity. In environment. These weaknesses constrain the realization of many countries developing agriculture has been key to income opportunities at all levels—from the smallholder ensuring sustained progress toward the MDGs (World Bank farmer trying to get surplus crops to market to the large for- 2005). Tanzania has decided to emphasize agriculture, hav- eign investor. ing launched the Kilimo Kwanza (Agriculture First) cam- The past few years have seen little progress in several paign designed to accelerate reforms in the sector, including indicators of the business environment. Tanzania’s rank in expanding input subsidy programs and significant invest- the Doing Business Index has been slipping, reaching 131 in ments in irrigation. 2010, as other countries have taken more determined strides to make their economies more business friendly. Tanzania Meeting macroeconomic challenges scores particularly poorly on basic bureaucratic require- ments involving numbers of procedures and permits and Addressing the infrastructure deficit will be expensive and the time spent to receive them. The importance of address- will likely require funding beyond that likely to be available ing this slowdown in these broader areas of structural on the highly concessional terms at which Tanzania has reform is reinforced by a key result of the literature on been borrowing. The additional funds could come from growth spells: although macroeconomic and financial many potential sources, including donor-financed projects, 34 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Table 1.4 Projected Progress toward Selected MDG Targets by Tanzania Target 1990 2000 2008 2015 Likely to achieve target? Population below basic needs poverty line (percent) 39.0 36.0 33.6 19.5 No Under-five underweight (percent) 28.5 29.5 22.0 14.4 No Under-five stunted (percent) 46.6 44.4 38.0 23.3 No Primary school net enrollment rate 54.2 58.7 97.2 100.0 Yes Under-five mortality rate (per 1,000 live births) 191.0 153.0 112.0 64.0 Yes Infant mortality rate (per 1,000 live births) 115.0 99.0 68.0 38.0 Yes Maternal mortality rate (per 100,000 live births) 529.0 — 578.0 133.0 No Births attended by skilled health personnel (percent) 43.9 35.8 63.0 90.0 No HIV prevalence, 15–24 (percent) 6.0 — 2.5 < 6.0 Yes Access to potable water (percent of rural population) 51.0 42.0 57.1 74.0 No Access to potable water (percent of urban population) 68.0 85.0 83.0 84.0 Yes Source: United Republic of Tanzania 2009a. Note: — Not available. Table 1.5 Infrastructure Indicators in Selected Countries in Sub-Saharan Africa Households Mobile Households Access to Access to with fixed phones with electricity Roads improved improved telephone (subscribers (percent (km per sanitation water source Country or (percent of per 100 connected to 1,000 km2 (percent of the (percent of the country group households)a people)b network) a of land)c population)a population)a Ghana 8 32 44 187 10 80 Kenya 12 30 13 111 42 57 Rwanda 1 7 5 568 23 65 South Africa 27 92 63 300 59 93 Tanzania 10 21 11 62 33 55 Uganda 3 18 8 385 33 64 Zambia 4 21 20 50 52 58 Sub-Saharan Africa 7 16 29 — 31 58 Sub-Saharan Africa low incomed 6 19 26 — 15 25 Sub-Saharan Africa middle incomed 19 36 55 — 41 66 Source: World Bank Africa Infrastructure Country Diagnostic Database and World Bank various years. Note: — Not available. a. Data are from Demographic and Health Surveys, latest available year for 2001–08. b. Data are for 2006. c. Data are for latest available year 2001–08. d. Income groups are based on World Bank classification. local or international bonds, and public-private partner- littered with examples of projects in which the host gov- ships (Ter-Minassian, Hughes, and Hajdenberg 2008). ernment ended up either simply taking over the project or Countries in Sub-Saharan Africa are increasingly accessing having to renegotiate in order to provide better terms for international capital markets, in several cases in the context the private sector partner) (Gratwick and Eberhard 2008). of IMF-supported programs (Redifer 2010). Whatever the Fiscal pressures—not just from infrastructure spending source of financing, it will be critical to ensure strong gov- but also from growing demand for public services as a result ernment processes for project selection, planning, and of high population growth—could be eased by mobilizing implementation and effective debt management to maxi- additional domestic resources. Estimates of tax potential mize the probability of a strong economic return while suggest that revenues could reach 21 percent of GDP, almost containing macroeconomic risks. Additional capacity to 5 percentage points more than current collections (IMF monitor risks in public-private partnerships will also be 2010c).8 A big part of the revenue loss stems from exemp- important to avoid the emergence of potentially significant tions, which amounted to 30 percent of tax collections contingent liabilities (the track record of such partnerships (3.5 percent of GDP) in 2007/08, according to estimates by in Tanzania and more generally in Sub-Saharan Africa is the Tanzania Revenue Authority, making Tanzania’s VAT CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 35 one of the least efficient in the region (table 1.6). A possibly In the financial sector, Tanzania’s basic crisis resolution substantial upside is natural resource revenues: the mining framework, together with a deposit insurance scheme, was sector is expanding rapidly and, as exemptions are reined in able to handle failures of individual banks. That framework and those for the more established mines expire, the tax take should be continuously reviewed to ensure that it can meet can be expected to increase significantly. Close cooperation potential risks; there is room to further strengthen banking and transparency in the granting of exemptions will also supervision, including risk analysis and enforcement of need to be tackled in the context of regional integration prudential requirements, while addressing staffing con- efforts, in order to avoid a race to the bottom, with poten- straints (IMF 2010b). tial investors able to induce neighboring governments to The absence of effective oversight of pension funds is a compete against one another on the physical location of critical deficiency in the current regulatory framework. Lit- activities, resulting in the erosion of the tax base. tle information is available on the financial position of pen- The exchange rate may also pose challenges. To date, the sion funds, their role in capital markets, and potentially flexibility in the exchange rate, the apparent absence of large fiscal liabilities. resource capacity constraints in the economy, and produc- With the increasing integration of the financial system of tivity growth have all served to avoid the emergence of a the East African Community, there is also a need to ensure misalignment of the exchange rate (see Hobdari 2008). that a coordinated policy response or intervention can be Going forward, in the event of the planned scaling up of orchestrated in the event of, say, pressure on a systemic both public spending and private sector investment, emerg- institution. The initial difficulty in agreeing on a coordi- ing capacity constraints will need to be monitored carefully, nated response was perhaps one of the key features of the and it will be important to ensure that public spending is global financial crisis. channeled into productivity-enhancing investment to sup- More generally, a well-defined social safety net can help port the competitiveness of the economy. Similarly, the move protect the most vulnerable in the event of economic down- to a common currency for the East African Community— turns. All of these responses require policy flexibility. which could generate additional growth opportunities— Retaining the policy buffers that were used to such good will need to address the issue of how individual national effect during the global financial crisis is critical for the con- economies will adjust to localized shocks once they relin- duct of countercyclical policy and the ability to prevent an quish the ability to adjust through the exchange rate. adverse shock from derailing the economy’s growth momentum. Limiting vulnerabilities but preparing for the worst CONCLUSION Even with the best macroeconomic policies and monitoring systems, bad outcomes can happen. Timely policy responses Tanzania has seen unprecedented sustained growth accel- are often critical for containing the impact of bad outcomes. eration since 1996. Several major (real and financial) Being able to craft them requires having appropriate sys- reforms played a pivotal role in the take-off. Maintaining tems in place before the shock occurs. macroeconomic stability was a critical component in Table 1.6 VAT Revenue and Revenue Productivity in Selected Countries in Sub-Saharan Africa, 2009/10 Current standard VAT VAT revenue Revenue productivity Country rate (percent) (percent of GDP)a (percent)b Kenya 16.0 6.0 37.5 Malawi 16.5 6.0 37.0 Rwanda 18.0 6.0 33.3 South Africa 14.0 5.7 40.5 Tanzania 18.0 4.6 25.5 Uganda 18.0 4.0 22.2 Unweighted average 16.8 5.4 32.7 Source: IMF VAT database and staff projections. a. Projection for 2009/10. b. Total VAT revenue, as percent of GDP, divided by the standard rate. 36 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY sustaining growth. Navigating the post–1996 period has Is another 15 years of uninterrupted growth likely? involved the development of new policy instruments, as Untapped growth potential clearly remains, and the policy the nature of the economy changed and legacy constraints space that has been developed over the past decade or so can were eased. provide needed support. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 37 ANNEX 1.A MACROECONOMIC PERFORMANCE Figure 1.A1 Macroeconomic Performance in Tanzania and Sub-Saharan Africa, 1970–2009 a. Real GDP growth b. Inflation 8 35 7 30 6 25 5 Percent Percent 20 4 15 3 10 2 1 5 0 0 5 5 5 6 7 8 9 5 5 5 6 7 8 9 –8 –9 –0 –0 –0 –0 –0 –8 –9 –0 –0 –0 –0 –0 70 86 96 05 06 07 08 70 86 96 05 06 07 08 19 19 19 20 20 20 20 19 19 19 20 20 20 20 c. GDP per capita 700 600 500 400 US$ 300 200 100 0 5 5 5 6 7 8 9 –8 –9 –0 –0 –0 –0 –0 70 86 96 05 06 07 08 19 19 19 20 20 20 20 Tanzania Sub-Saharan Africa 38 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Figure 1.A2 External Sustainability Indicators in Tanzania, 1970–2009 a. Public sector external debt b. Public sector external debt 120 40 35 100 30 Percent of exports Percent of GDP 80 25 60 20 15 40 10 20 5 0 0 5 5 5 6 7 8 9 5 5 5 6 7 8 9 –8 –9 –0 –0 –0 –0 –0 –8 –9 –0 –0 –0 –0 –0 70 86 96 05 06 07 08 70 86 96 05 06 07 08 19 19 19 20 20 20 20 19 19 19 20 20 20 20 c. Foreign reserves 6 5 Months of imports 4 3 2 1 0 5 5 5 6 7 8 9 –8 –9 –0 –0 –0 –0 –0 70 86 96 05 06 07 08 19 19 19 20 20 20 20 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 39 ANNEX 1.B LIST OF COUNTRIES IN THE SAMPLE Table 1.B1 List of Countries in the Sample Low income Middle income High income Bangladesh Albania Australia Burkina Faso Algeria Austria Côte d’Ivoire Argentina Belgium Ethiopia Azerbaijan Canada Ghana Belarus Czech Republic India Bolivia Denmark Kenya Brazil Estonia Madagascar Bulgaria Finland Mozambique Cameroon France Nepal Chile Germany Nigeria China Greece Pakistan Colombia Hong Kong SAR, China Senegal Costa Rica Ireland Tanzania Dominican Republic Israel Uganda Ecuador Italy Uzbekistan Egypt, Arab Rep. of Japan Vietnam El Salvador Korea, Rep. of Zimbabwe Georgia Netherlands Guatemala New Zealand Hungary Norway Indonesia Portugal Jamaica Singapore Jordan Spain Kazakhstan Sweden Latvia Switzerland Lithuania Taiwan, China Malaysia United Kingdom Mexico United States Morocco Nicaragua Paraguay Peru Philippines Poland Romania Russian Federation South Africa Sri Lanka Thailand Tunisia Turkey Ukraine Uruguay Venezuela, R. B. de Source: World Bank classification. 40 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY ANNEX 1.C MEASURES OF REFORM Table 1.C1 Measures of Reform Coverage MIN # of MAX # of countries in countries in Measure Description Source Start year End year any year any year Real Indices Trade Openness Tariff Rates Average tariff rates, with missing values extrapolated Various sources, including IMF, 1960 2005 47 142 using implicit weighted tariff rates. Index World Bank, WTO, UN, and normalized to be between zero and unity: zero the academic literature means the tariff rates are 60 percent or higher, (particularly Clemens and while unity means the tariff rates are zero. Willamson, 2004) Current-Account An indicator of how compliant a government is with Quinn (1997), and Quinn and 1960 2005 50 65 Restrictions its obligations under the IMF’s Article VIII to free Toyoda (2007; forthcoming). from government restriction the proceeds from international trade in goods and services. The index represents the sum of two sub-components, dealing with restrictions on trade in visibles, as well as in invisibles (financial and other services). It distinguishes between restrictions on residents (receipts for exports) and on non-residents (payments for imports). Although the index measures restrictions on the proceeds from transactions, rather than on the underlying transactions, many countries in practice use restrictions on trade proceeds as a type of trade restriction. The index is scored between zero and 8 in half-integer units, with 8 indicating full compliance. Product Markets Telecom and Electricity Simple average of the electricity and telecom Based on legislation and other 1960 2003 106 108 Industries markets sub-indices, which are constructed, in official documents. turn, from scores along three dimensions. For electricity, they capture: (i) the degree of unbundling of generation, transmission, and distribution; (ii) whether a regulator other than government has been established; and (iii) whether the wholesale market has been liberalized. For telecom, they capture: (i) the degree of competition in local services; (ii) whether a regulator other than government has been established; and (iii) the degree of liberalization of interconnection changes. Indices are coded with values ranging from zero (not liberalized) to two (completely liberalized). (continued next page) 41 42 Table 1.C1 (continued) Coverage MIN # of MAX # of countries in countries in Measure Description Source Start Year End Year any year any year Real Indices Agriculture Given that developing countries constitute most of Based on legislation and other 1960 2003 96 104 our sample, the degree of regulation in agriculture, official documents. which continues to account for a large part of many of these economies, is an essential aspect of product market competition. Index aims to capture intervention in the market for the main agricultural export commodity in each country. As data limitations preclude coding separate dimensions of intervention, the index provides a summary measure of intervention. Each country-year pair is assigned one of four degrees of intervention: (i) maximum (public monopoly or monopsony in production, transportation, or marketing); (ii) high (administered prices); (iii) moderate (public ownership in relevant producers, concession requirements); and (iv) no intervention. Financial Indices Capital Account Qualitative indicators of restrictions on financial Abiad and others (2008), which 1973 2005 72 91 Opennes: Aggregate credits and personal capital transactions of residents follows the methodology in and financial credits to nonresidents, as well as the Abiad and Mody (2005). The use of multiple exchange rates. Index coded from original sources are mostly zero (fully repressed) to three (fully liberalized). various IMF reports and working papers, but also central bank websites, etc. Resident/nonresident-specific indices are based on Quinn (1997), and Quinn and Toyoda (2007). Capital Account Measures the extent to which residents Openness: Residents (nonresidents) are free from legal restrictions to (nonresidents) only move capital into and out of a country. Domestic Financial The index of domestic financial liberalization is an Liberalization average of six subindices. Five of them relate to banking: (i) interest rate controls, such as floors or ceilings; (ii) credit controls, such as directed credit, and subsidized lending; (iii) competition restrictions, such as limits on branches and entry barriers in the banking sector, including licensing requirements or limits on foreign banks; (iv) the degree of state ownership; and (v) the quality of banking supervision and regulation, including power of independence of bank supervisors, adoption of a Basel I capital adequacy ratio, and framework for bank inspections. The sixth subindex refers to the regulation of securities markets, including policies to encourage the development of bond and equity markets, and to permit access of the domestic stock market to foreigners. The subindices are aggregated with equal weights. Each subindex is coded from zero (fully repressed) to three (fully liberalized). Source: Prati, Onorato, and Papageorgiou (2010). 43 44 Table 1.C2 Description of Reform Indices Coverage MIN # of MAX # of countries in countries in Reform indices Description Source Start year End year any year any year Financial Sector Domestic Financial The index of domestic financial liberalization is an Abiad and others (2008), 1973 2005 72 91 Sector Liberalization average of six subindices. Five of them relate to following the methodology in banking: (i) interest rate controls, such as floors Abiad and Mody (2005), based or ceilings; (ii) credit controls, such as directed on various IMF reports and credit and subsidized lending; (iii) competition working papers, central bank restrictions, such as limits on branches and entry websites, and others. barriers in the banking sector, including licensing requirements or limits on foreign banks; (iv) the degree of state ownership; and (v) the quality of banking supervision and regulation, including power of independence of bank supervisors, adoption of Basel capital standards, and a framework for bank inspections. The sixth subindex relates to securities markets and covers policies to develop domestic bond and equity markets, including (i) the creation of basic frameworks such as the auctioning of T-bills, or the establishment of a security commission; (ii) policies to further establish securities markets such as tax exemptions, introduction of medium and long-term government bonds to establish a benchmark for the yield curve, or the introduction of a primary dealer system; (iii) policies to develop derivative markets or to create an institutional investor’s base; and (d) policies to permit access to the domestic stock market by nonresidents. The subindices are aggregated with equal weights. Each subindex is coded from zero (fully repressed) to three (fully liberalized). External Capital Qualitative indicators of restrictions on financial Abiad and others (2008), 1973 2005 72 91 Account Liberalization: credits and personal capital transactions of following the methodology in Aggregate residents and financial credits to nonresidents, as Abiad and Mody (2005), based well as the use of multiple exchange rates. Index on various IMF reports and coded from zero (fully repressed) to three working papers, central bank (fully liberalized). websites, and others. External Capital Indicators measuring the intensity of legal restrictions Based on the methodology in 1960 2005 50 65 Account Liberalization: on residents', respectively nonresidents', ability to Quinn (1997) and Quinn and Residents vs. move capital into and out of a country. Index Toyoda (2007), drawing on Nonresidents originally coded from zero (fully repressed) to 50 information contained in the (fully liberalized). Fund's AREAER. Real Sector Trade Liberalization Tariff Rates Average tariff rates, with missing values extrapolated Various sources, including IMF, 1960 2005 47 142 using implicit weighted tariff rates. Index World Bank, WTO, UN, and normalized to be between zero and unity: zero the academic literature means the tariff rates are 60 percent or higher, (particularly Clemens and while unity means the tariff rates are zero. Willamson, 2004). Current-Account An indicator of how compliant a government is with Based on the methodology in 1960 2005 50 65 Restrictions its obligations under the IMF’s Article VIII to free Quinn (1997) and Quinn and from government restriction the proceeds from Toyoda (2007) drawing on international trade in goods and services. The information contained in the index represents the sum of two subcomponents, Fund’s AREAER. dealing with restrictions on trade in visibles, as well as in invisibles (financial and other services). It distinguishes between restrictions on residents (receipts for exports) and on nonresidents (payments for imports). Although the index measures restrictions on the proceeds from transactions, rather than on the underlying transactions, many countries in practice use restrictions on trade proceeds as a type of trade restriction. The index is scored between zero and 8 in half-integer units, with 8 indicating full compliance. (continued next page) 45 46 Table 1.C2 (continued) Coverage MIN # of MAX # of countries in countries in Reform Indices Description Source Start Year End Year any year any year Product Markets Telecom and Electricity Simple average of the electricity and telecom Based on various existing studies 1960 2003 106 108 Industries markets subindices, which are constructed, and datasets as well as national in turn, from scores along three dimensions. legislation and other official For electricity, they capture: (i) the degree of documents. unbundling of generation, transmission, and distribution; (ii) whether a regulator other than government has been established; and (iii) whether the wholesale market has been liberalized. For telecom, they capture: (i) the degree of competition in local services; (ii) whether a regulator other than government has been established; and (iii) the degree of liberalization of interconnection changes. Indices are coded with values ranging from zero (not liberalized) to two (completely liberalized). Agriculture The index captures intervention in the market for Based on IMF commodities data, 1960 2003 96 104 the main agricultural export commodity in each various existing studies and country. As data limitations preclude coding datasets, and national legislation separate dimensions of intervention, the index and other official documents. provides a summary measure. Each country-year pair can take four values: (i) zero (public monopoly or monopsony in production, transportation, or marketing, e.g., export marketing boards); (ii) one-third (administered prices); (iii) two-thirds (public ownership of relevant producers or concession requirements); and (iv) one (no public intervention). Source: Prati, Onorato, and Papageorgiou (2010). ANNEX 1.D NEW DATASET ON STRUCTURAL REFORMS A discussion of structural reforms, their determinants, and financial transactions of residents and nonresidents, as well their economic impact requires solid data on structural as the use of multiple exchange rates. The index of domes- policies and how they have changed (IMF 2009). A new tic financial liberalization is an average of six subindexes, dataset was compiled by the Research Department of the covering credit controls, such as directed credit; interest rate IMF that brings together information on a variety of struc- controls, such as floors or ceilings; entry barriers in the tural reforms in both the real and financial sectors of the banking sector, such as licensing requirements or limits on economy. The dataset spans about 30 years (1973–2005) foreign banks; competition restrictions, such as limits on and 90 countries, selected on the basis of data availability. branches; the degree of state ownership; and aggregate The key advantage of these new indexes over those used in credit ceilings. All indicators were rescaled to range between previous work, such as IMF (2004), is that they cover both a zero and one, with higher values representing greater liber- long sample period and a broad range of countries, includ- alization. Differences in the values of each index across ing advanced and developing economies. countries and over time provide useful information on the Among the indicators of real sector reforms, two separate variation in the absolute degree of economic liberalization indexes measure trade openness. The first covers average within each sector. Instead, differences in the value of the tariff rates; the second covers restrictions on other current indexes across sectors are not a precise quantitative measure account transactions, including payments and receipts on of whether one sector is more liberalized than another exports and imports of goods and services.9 Two separate because of the different methodology used to construct each indexes measure product market liberalization. The first index. For instance, a positive difference between the trade covers the network industries, specifically the degree of lib- index and the financial index does not necessarily mean that eralization of the electricity and telecom markets. The sec- trade is more liberalized than the financial sector. ond covers the agriculture sector. It aims to capture public As illustrated in figure 1.D1, all indexes trend upward intervention in the market for the main agricultural export over time toward a high degree of liberalization. At a sectoral commodity in each country. level, the global liberalization of international trade, capital Among the indicators of financial sector reforms, the movements, and the domestic financial sector has been index of capital account openness measures restrictions on fairly steady and gradual over the past three decades, Figure 1.D1 Indexes of Structural Reform, 1973–2003 (All countries) 0.9 0.8 0.7 0.6 0.5 Indices 0.4 0.3 0.2 0.1 0 19 3 19 4 19 5 19 6 19 7 19 8 79 19 0 19 1 82 19 3 19 4 19 5 19 6 19 7 19 8 19 9 90 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 02 20 3 04 05 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 0 0 0 19 19 19 19 20 20 Year Agriculture Capital flows Domestic finance Networks Trade CA restrictions Source: IMF 2009. Note: Data are based on sample of 90 countries for which data were available. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 47 whereas product market liberalization started only about and services—captures only some nontariff barriers to 1990. There have been no global setbacks in the average trade. For other nontariff barriers, such as quotas and sub- degree of liberalization in any sector. Liberalization mea- sidies, there is not a sufficiently long time series to be used sures display significant differences across regions, but all in the analysis. sectoral indexes have converged across income levels, point- ing to imitation and “catch-up” effects. REFERENCES Abbas, S., M. Ali, and Y. Sobolev. 2008. “High and Volatile NOTES Treasury Yields in Tanzania: The Role of Strategic Bid- ding and Auction Microstructure.” IMF Working Paper 1. This chapter draws heavily on Nord et al. (2009). The 08/81, International Monetary Fund, Washington, DC. authors are grateful to Daehaeng Kim, Saul Lizondo, Roger Bai, J., and P. Perron. 1998. “Estimating and Testing Linear Nord, Laure Redifer, and World Bank reviewers for helpful Models with Multiple Structural Changes.” Econometrica comments. 66: 47–78. 2. To the authors’ knowledge, this is the first time Tanzania Berg, A., D., J. Ostry, and J. Zettelmeyer. 2008. “What Makes has been formally shown to have had an “up-break” in 1996. Growth Sustained?” IMF Working Paper 08/59, Interna- Previous studies (Hausmann, Pritchett, and Rodrik 2005; tional Monetary Fund, Washington, DC. Berg, Ostry, and Zettelmeyer 2008) failed to identify a break, Drummond, Paulo, A. Mrema, S. Roudet, and M. Saito. probably because they used shorter time series. 2009. Foreign Exchange Reserve Adequacy in East African 3. Excellent descriptions of the reforms undertaken during Community Countries. African Departmental Paper these periods from varying perspectives can be found in 09/01, International Monetary Fund, Washington, DC. Maliyamkono and Mason (2006), Mtei (2010), Mwase and Easterly, W., M. Kremer, L. Pritchett, and L. H. Summers. Ndulu (2008), and Nord et al. (2009). 1993. “Good Policy or Good Luck? Country Growth Per- 4. A World Bank (2007) study identifies increases in higher formance and Temporary Shocks.” Journal of Monetary education attainment as an important contributor to Economics 32 (3): 459–483. growth. Gratwick, K. N., and A. Eberhard. 2008. “An Analysis of 5. The tax system has not been fully predictable for investors, Independent Power Projects in Africa: Understanding because of fairly frequent changes in administration—often Development and Investment Outcomes.” Development designed to close loopholes—and the existence of a large Policy Review 26 (3): 309–38. number of tax exemptions granted to particular sectors, Hausmann, R., L. Pritchett, and D. Rodrik. 2005. “Growth subcategories of investors, and specific enterprises. Accelerations.” Journal of Economic Growth 10 (4): 303–29. 6. Government programs during this period, supported by Hobdari, Niko. 2008. “Tanzania’s Equilibrium Real Exchange the IMF (first under the Poverty Reduction and Growth Rate.” IMF Working Paper 08/138, International Mone- Facility, subsequently under the Policy Support Instrument), tary Fund, Washington, DC. were designed in such a way that a shortfall in foreign pro- gram financing automatically triggered the possibility of IMF (International Monetary Fund). 2004. “Advancing resort to high domestic financing subject to a pre-agreed Structural Reforms.” World Economic Outlook (April), upper limit; in the event that foreign program financing Washington, DC. exceeded targets, the government was unconstrained in its ———. 2009. Structural Reforms and Economic Performance ability to spend the additional financing received. in Advanced and Developing Countries. Occasional Paper 7. The bank closures were Meridien Bank (in 1995), 268, Washington, DC. Tanzania Housing Bank (in 1995), Greenland Bank (in ———. 2010a. “Sub-Saharan Africa: Back to High Growth?” 1999), and Delphis Bank (in 2003). Depositors were fully Regional Economic Outlook, April, Washington, DC. compensated in all cases, with the resolutions of both ———. 2010b. Tanzania: Financial System Stability Assess- Meridien Bank and Delphis Bank facilitated by the trans- ment Update. IMF Country Report 10/177. fers of deposits to other banks operating in Tanzania. ———. 2010c. Tanzania: Seventh Review under the Policy 8. The models predict tax potential through a regression of Support Instrument, Second Review under the Exogenous the tax to GDP ratio on observable structural characteristics Shocks Facility, and Request for a New Three-Year Policy of the economy, such as the size of the agricultural sector. Support Instrument. IMF Country Report 10/173, 9. This index—measuring how compliant a country is Washington, DC. with its obligations under the IMF’s Article VIII to free from ———. 2010d. World Economic Outlook, April. Washington, restrictions the proceeds from international trade in goods DC. 48 CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY Lawson, A., D. Booth, M. Msuya, S. Wangwe, and Pritchett, Lant. 2000. “Understanding Patterns of Economic T. Williamson. 2005. Does General Budget Support Work? Growth: Searching for Hills among Plateaus, Mountains, Evidence from Tanzania. Overseas Development Institute, and Plains.” World Bank Economic Review 14 (2): 221–50. London. Redifer, Laure. 2010. “New Financing Sources for Africa’s Maliyamkono, T. L., and H. Mason. 2006. The Promise. Dar Infrastructure Deficit.” IMF Survey July 21, Washington, es Salaam: TEMA Publishers. DC. Mbelle, A. V. Y., L. Rutasitra, B. Makenda, R. Lokina, Selassie, Abebe Aemro. 2008. “Beyond Macroeconomic A. Naho, and Y. Aikaeli. 2010. “Analytical Study on the Stability: The Quest for Industrialization in Uganda.” Drivers of Growth and Implications for Growth Strategy IMF Working Paper 08/231, International Monetary for Mainland Tanzania,” University of Dar es Salaam, Fund, Washington, DC. Dar es Salaam. Ter-Minassian, Teresa, Richard Hughes, and Alejandro Mtei, Edwin. 2009. From Goatherd to Governor. Dar es Hajdenberg. 2008. “Creating Sustainable Fiscal Space for Salaam: Mkuki na Nyota Publishers. Infrastructure: The Case of Tanzania.” IMF Working Paper Mwase, Nkunde, and Benno Ndulu. 2008. “Tanzania: 08/256, International Monetary Fund, Washington, DC. Explaining Four Decades of Episodic Growth.” In The United Republic of Tanzania. 2009a. “Millennium Develop- Political Economy of Economic growth in Africa ment Goals Report, Mid-Way Evaluation: 2000–2008.” 1960–2000, vol 2., Country Case Studies, ed. Ndulu, B, www.tzdpg.or.tz/external/mkukuta-mkuza-review/ S. O’Connell, J. P. Azam, R. Bates, A. Fosu, J. Gunning, studies-tors-cvs.html. and D. Njinkeu. Cambridge: Cambridge University Press. ———. 2009b. MKUKUTA Annual Implementation Report National Audit Office. 2010. A Performance Audit Report on 2008/2009, Success in the Midst of Turbulence. Dar es Health Care Centres Efficiency in Tanzania. www.nao.go.tz Salaam. Nord, Roger, Yuri Sobolev, David Dunn, Alejandro Hajden- ———. 2010. Public Expenditure Tracking Survey for Pri- berg, Niko Hobdari, Samar Maziad, and Stephane mary and Secondary Education in Mainland Tanzania. Roudet. 2009. Tanzania: The Story of an African Transi- Dar es Salaam. tion. International Monetary Fund, Washington, DC. World Bank. Various years. World Development Indicators. Ostry, J. D., A. Prati, and A. Spilimbergo. 2009. Structural Washington, DC: World Bank. Reforms and Economic Performance in Advanced and ———. 2005. Agriculture and Achieving the Millennium Developing Countries. IMF Occasional Paper 268, Development Goals. Report 32729-GLB, Agriculture and Washington, DC. Rural Development Department, Washington, DC. Prati, Alessandro, Massimiliano Onorato, and Chris Papa- ———. 2007. Tanzania: Sustaining and Sharing Economic georgiou. 2010. “Which Reforms Work and under What Growth. Country Economic Memorandum, Washington, Institutional Environment? Evidence from a New DC. Dataset on Structural Reforms.” International Monetary ———. 2009. Doing Business 2010, Tanzania. Washington, Fund, Washington, DC. DC: World Bank. CHAPTER 1: TANZANIA: GROWTH ACCELERATION AND INCREASED PUBLIC SPENDING WITH MACROECONOMIC STABILITY 49 CHAPTER 2 Building on Growth in Uganda Sarah Ssewanyana, John Mary Matovu, and Evarist Twimukye ver the past two decades, Uganda has seen a O Uganda’s per capita GDP at purchasing power parity remarkable turnaround in economic perfor- remains about half that of Sub-Saharan African as a whole. mance, with growth averaging about 7.7 percent During the decade following the end of political instabil- a year over the 1997–2007 period. Equally impressive has ity and civil war in 1986, Uganda’s economy grew at an aver- been the sharp decline in poverty rates, which fell about age rate of 7.7 percent a year. (Economic growth declined by 15 percentage points over this period. Improved macroeco- 1.4 percentage points between 2008/09 and 2009/10.) Initially, nomic management and economic reforms contributed to economic growth was driven by postwar recovery and recon- the country’s strong growth performance. struction. Since the early 1990s it has been driven by compre- Although growth and poverty reduction have been hensive macroeconomic and structural reforms. Investment impressive, Uganda has experienced worsening income dis- growth also remained strong, with private investment rising tribution; a decline in the relative importance of agriculture by an estimated 17 percent and public investment rising to overall gross domestic product (GDP); a growing youth 15 percent (table 2.2). Private investment growth was led by population, with increasing unemployment; and a low rate construction (AfDB 2009). of urbanization. There has also been limited structural Economic growth has been export led, with the share transformation of the economy, a reflection that growth has of exports in GDP rising over the past two decades. The come largely from the services sector, which employs the expanding regional market for Uganda’s food and manu- highly skilled, rather than from the agricultural sector, factured products has boosted exports during the past five which still employs 70 percent of the population. Notwith- years, a reflection of the dividend enjoyed by Uganda’s standing the considerable progress in diversifying its export neighbors, whose demand for Ugandan goods has base away from coffee, Uganda still remains a primary increased. commodity exporter, with limited value addition to its Growth has been driven by the services sector, which has major exports. accounted for almost half of GDP since 2001/02 (table 2.3). Before the global financial crisis, the key subsectors driving growth in the services sector included financial services, GROWTH EXPERIENCE transport and communications, public administration, and Uganda achieved impressive economic growth over the past defense (see annex table 2.A1). The agricultural sector grew two decades, with positive per capita GDP growth since more slowly than the other sectors between 2001/02 and 1987 and stronger growth than the continent as a whole 2009/10, accounting for about 24 percent of annual GDP on (figure 2.1 and table 2.1). Despite this improvement, average over the period (table 2.3). 51 Figure 2.1 Annual GDP Growth in Uganda, 1983–2007 12 10 Per capita GDP (dollars) 8 1,000 6 4 500 Percent 2 0 0 –2 –4 –6 –8 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 GDP per capita, (purchasing power parity in constant 2005 dollars) GDP growth (annual percent) GDP per capita growth (annual percent) Source: World Bank 2009. Table 2.1 Annual GDP Growth in Selected Countries in Sub-Saharan Africa, 1997–2009 Sub-Saharan Year Ghana Kenya Mauritius Mozambique South Africa Tanzania Tunisia Uganda Africa 1997 4.20 0.47 5.69 10.24 2.65 3.53 5.44 5.10 3.58 (1.59) (–2.23) (4.37) (7.26) (0.32) (0.84) (4.00) (1.94) (0.87) 1998 4.70 3.29 6.07 10.78 0.52 3.71 4.78 4.91 2.42 (2.15) (0.59) (4.96) (7.96) (–1.82) (1.12) (3.46) (1.77) (–0.24) 1999 4.40 2.31 2.61 8.12 2.36 3.53 6.05 8.05 2.46 (1.90) (–0.33) (1.31) (5.39) (–0.08) (0.97) (4.68) (11.11) (–0.19) 2000 3.70 0.60 9.03 1.09 4.15 5.10 4.70 3.14 3.53 (1.24) (–1.98) (7.96) (–1.52) (1.61) (2.46) (3.52) (–0.01) (0.87) 2001 4.00 3.78 2.57 11.9 2.74 6.24 4.92 5.18 3.62 (1.56) (1.11) (1.46) (8.93) (0.65) (3.53) (3.73) (1.92) (1.00) 2002 4.50 0.55 2.11 8.82 3.67 7.24 1.65 8.73 3.45 (2.07) (–2.03) (1.24) (5.92) (2.25) (4.46) (0.53) (5.31) (0.87) 2003 5.20 2.93 3.66 6.02 2.95 5.67 5.56 6.47 4.23 (2.79) (0.29) (2.59) (3.21) (1.65) (2.84) (4.94) (3.09) (1.67) 2004 5.60 5.10 5.75 7.88 4.55 6.73 6.04 6.81 6.19 (3.23) (2.40) (4.84) (5.08) (3.33) (3.88) (5.05) (3.39) (3.60) 2005 5.90 5.91 1.24 8.39 5.28 7.37 3.98 6.33 5.72 (3.58) (3.17) (0.44) (5.66) (4.09) (4.46) (2.98) (2.92) (3.15) 2006 6.40 6.32 3.95 8.68 5.60 6.74 5.66 10.78 6.17 (4.13) (3.56) (3.14) (6.04) (4.43) (3.79) (4.63) (7.23) (3.59) 2007 6.46 7.01 5.52 7.28 5.49 7.15 6.33 8.41 6.60 (4.24) (4.22) (4.87) (4.75) (4.34) (4.15) (5.32) (4.92) (4.02) 2008 8.43 1.55 5.09 6.74 3.68 7.44 4.64 8.71 5.26 (6.20) (–1.09) (4.41) (4.29) (2.54) (4.39) (3.60) (5.21) (2.72) 2009 4.66 2.59 2.14 6.33 –1.78 5.50 3.13 7.06 1.65 (2.52) (–0.08) (1.62) (3.96) (–2.83) (2.48) (2.10) (3.62) (–0.80) Source: World Bank 2010. Note: Figures in parentheses are per capita growth rates. 52 CHAPTER 2: BUILDING ON GROWTH IN UGANDA Table 2.2 Components of GDP Growth in Uganda, 2001–09 (percent, in 2002 prices) Component 2001 2002 2003 2004 2005 2006 2007 2008 2009 Consumption Private 6.3 6.1 2.8 3.4 8.1 12.3 2.9 8.2 11.5 Public 2.9 5.2 5.2 3.7 4.5 3.7 –1.5 1.7 –1.9 Investment Private 10 14.1 16.1 14.6 23.5 10.3 14.6 12.0 1.1 Public 0.3 –8.1 2.3 11.6 9.3 13.5 19.6 –15.3 28.8 Trade Exports 17.2 7 6.6 20.9 21.5 –6.3 53.8 45.0 –12.0 Imports 12.8 7.5 6.2 9.6 16.8 17.2 15.7 17.6 4.6 Source: Uganda Bureau of Statistics 2010. Table 2.3 Sectoral Contributions to GDP and Growth Rates in Uganda, 2001/02–2008/09 (percent, in 2002 prices) Item 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 Sectoral contributions to GDP (in current prices) Agriculture 23.1 22.1 21.1 20.2 24.1 22.3 21.4 23.1 23.9 Industry 22.0 22.6 22.8 24.0 22.8 25.1 25.8 24.7 24.6 Services 48.3 48.6 49.1 49.0 47.2 47.0 46.9 46.4 45.4 Growth rate Total GDP 8.5 6.5 6.8 6.3 10.8 8.4 8.7 7.2 5.8 Per capita GDP 5.1 3.1 3.4 3.0 7.3 5.0 5.3 3.8 2.4 Agriculture 7.1 2.1 1.6 2.0 0.5 0.1 1.3 2.5 2.1 Industry 7.4 9.5 8 11.6 14.7 9.6 8.8 5.8 8.9 Services 11.0 7.4 7.9 6.2 12.2 8.0 9.7 8.8 5.8 Source: Uganda Bureau of Statistics 2010. The contribution of industry to GDP—which ranged broad-based welfare gains in terms of consumption from 22.0 to 24.6 percent over the period—is well below the increases. The gains were not equally distributed across 35 percent benchmark for countries graduating from low- to social groups and spatially. Overall, the pattern of growth middle-income status (Bevan and others 2003). Growth in the has been skewed, with growth taking place in subsectors sector fell from 9.1 percent in 2007/08 to 5.8 percent in such as telecommunications and finance, which employ 2008/09, largely as a result of the global economic crisis, which highly skilled people, not the poor. reduced the remittances that had fueled a construction boom The mismatch between the contribution of the different in Uganda. The decline caused the share of the construction sectors to GDP and the proportion of the population that subsector in GDP to fall from 10.8 percent in 2007/08 to derives its livelihood from the sectors has serious implica- 3.7 percent in 2008/09 (see annex table 2.A1). The slowdown tions for the level of unemployment and underemployment was also a result of the increase in the costs of imported in Uganda. Although the contribution of agriculture to inputs arising from the depreciation of the Ugandan shilling. overall GDP has remained lower than that of services or The contribution of the manufacturing subsector to overall industry, its share in employment remains high. Services, GDP was well below that of the construction subsector. which account for almost half of GDP, employ only about 24 percent of the population. Employment in industry, Structural transformation of the economy which accounts for more than a quarter of GDP, is also very minimal, at about 8 percent of the population (table 2.4). Uganda’s economy has been growing rapidly, but growth has failed to create enough jobs for the ever-increasing labor Agricultural performance force. Evidence from the Uganda National Household Sur- veys shows that as a result of impressive economic growth Because more than two-thirds of Ugandans work in agricul- during the 1990s, the population experienced important ture, development of the Ugandan economy is closely linked CHAPTER 2: BUILDING ON GROWTH IN UGANDA 53 to transformation of the agricultural sector. Agriculture in then, Uganda has diversified its exports base to include Uganda is still characterized by low productivity, mainly as larger shares of flowers, fishing, and other agricultural a result of poor inputs, undeveloped value chains, and low exports (see annex table 2.A2). Revenue from noncoffee public and private investment in the sector. The lack of sus- exports increased by more than sixfold between 1997/98 tained agricultural growth and the slow process of diversi- and 2008/09, rising from $189.6 million to $1,199.6 million fication in agriculture pose serious threats to poverty (figure 2.3). reduction efforts. Trade liberalization was designed to reverse and even The reforms of the early 1990s, especially the dismantling eliminate the trade deficit by increasing export earnings of the agricultural public enterprises and liberalization of the and curtailing the demand for imports. Incentives for economy, led to greater participation of the private sector in export-oriented trade and market-determined exchange marketing agricultural produce. The agricultural reforms rate policies were expected to encourage both traditional implemented since the mid-1990s, however, have largely and nontraditional exports. Nevertheless, merchandise benefited only a small fraction of farmers, particularly richer exports continued to decline throughout the liberalization and better-educated farmers, who have been able to diversify period (1987–92), partly because the manufacturing sector their agricultural production. Once these efficiency gains had shrunk as a result of the economic mismanagement of were exploited, other innovations were needed to maintain growth in the sector (Okidi and others 2007). Only 50 percent of agricultural production in Uganda is Figure 2.2 Exports as Percentage of GDP in Uganda, sold on markets. With the exception of coffee farmers and 1998/99–2008/09 farmers engaged in other tradables or niche markets (vanilla, fruit, tomatoes), most smallholders in Uganda are 25 still engaged in subsistence farming. For the most part, 20 Ugandan farmers remain poor and out of mainstream eco- nomic activity. 15 Percent 10 Export performance 5 Exports as a share of GDP have increased over time in Uganda (see figure 2.2). Before the liberalization of the 0 economy and the emphasis on import substitution and 9 0 1 2 3 4 20 5 6 20 7 20 8 9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 98 99 00 01 02 03 04 05 06 07 08 export diversification in the 1990s, Uganda depended 19 19 20 20 20 20 20 20 mainly on coffee as its main export. This dependence on a Source: Bank of Uganda 2009. single commodity was a major constraint to terms-of-trade growth, especially when world coffee prices dropped, as they did in the mid-1990s. Figure 2.3 Coffee, Noncoffee, and Total Exports from To insulate the economy from adverse terms of trade Uganda, 1997/98–2008/09 and instability in export earnings associated with com- modity concentration, the government adopted a policy 3,500 shift in 1987 that sought to diversify the exports base to 3,000 Millions of dollars include nontraditional (mainly agricultural) exports. Since 2,500 2,000 1,500 Table 2.4 Distribution of Employment in Uganda, 1,000 by Sector and Gender, 2003 500 (percent) 0 8 9 0 1 2 3 4 5 6 7 8 9 /9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 Sector Women Men Total 97 98 99 00 01 02 03 04 05 06 07 08 Agriculture 76 62 69 Total exports Coffee Industry 5 10 8 Noncoffee Informal cross-border Services 19 28 24 Source: World Bank 2009. Source: Bank of Uganda 2008/09. 54 CHAPTER 2: BUILDING ON GROWTH IN UGANDA earlier regimes. Since then, the value of exports has Figure 2.5 Trade Deficit as Percentage of GDP in improved markedly. Uganda, 1998/99–2008/09 Alongside the increase in noncoffee exports has been a huge increase in the importance of informal cross-border 10 9 trade, which rose from 1.3 percent of total exports in 8 2002/03 to 50.2 percent by 2008/09—a significant develop- 7 ment in the wake of the global financial crisis, which was 6 Percent beginning to affect traditional exports, especially coffee. 5 Including informal cross-border trade, the share of indus- 4 3 trial products in total exports increased from 43.8 percent in 2 2007/08 to 54.9 percent in 2008/09 (figure 2.4). As new 1 emerging export markets within the region stabilize, it will 0 be difficult for Uganda to sustain the recent rate of indus- 19 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 20 7 20 8 9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 98 99 00 01 02 03 04 05 06 07 08 trial product export growth. 19 Most primary commodities earn a fraction of what Source: Bank of Uganda 2009. they would earn if they had been processed. Conse- quently, although the shilling depreciated against the U.S trading with postconflict economies in the region, includ- dollar in real terms (from U Sh 558 per dollar in 1987 to ing Southern Sudan, the Democratic Republic of Congo, U Sh 2,000 in 2008), implying higher domestic producer and Rwanda. As these countries stabilize, the trade deficit prices in Uganda shillings, the trade deficit ballooned, of Uganda with its major East African Community partners from $446.7 million in 1997/98 to $936.3 million in will have to be addressed by increasing exports to Kenya 2008/09. The increase in the trade deficit reflected the low and Tanzania. returns from exports caused by the deteriorating terms of trade and the large increase in the value of imports caused by the depreciation of the shilling. The trade deficit as a Impact of growth on poverty reduction and share of GDP has declined since 2005/06 (figure 2.5). other millennium development goals Uganda will potentially benefit from the expanded East Uganda is one of the few Sub-Saharan African countries African Common Market, estimated at 120 million people, to achieve the first MDG of halving extreme poverty which allows unimpeded movement of labor, capital, and before 2015 (annex table 2.A3). The proportion of the other services across borders within Burundi, Kenya, population living below the absolute poverty line declined Rwanda, Tanzania, and Uganda. However, trade with the from 56.4 percent in 1992/93 to 31.1 percent in 2005/06 largest economies in the union (Kenya and Tanzania) has and to 24.5 percent in 2009/10 (table 2.5).1 Income been one sided, with limited exports from Uganda to the poverty remains a key development challenge, however, two countries. Uganda’s trade has been boosted mainly by with the absolute number of poor people declining only marginally, from 9.8 million in 1992 to 8.4 million in Figure 2.4 Industrial Products as Percentage of Total 2005/06 and to 7.5 million in 2009/10, as a result of a pop- Exports in Uganda, 1997/98–2008/09 ulation growth rate of 3.2 percent a year. The majority of the poor live in rural areas, in particular in northern 60 Uganda (Ssewanyana and Okidi 2007; Uganda Bureau of 50 Statistics 2010). 40 The regional ranking of other poverty measures in Percent 30 Uganda is identical to that of the headcount index, although 20 there are growing differences in the poverty gap between 10 northern Uganda and the rest of the country. Overall, the 0 poverty gap dropped faster than the poverty headcount, implying rising mean consumption by Uganda’s poor. The 8 9 0 1 2 3 4 5 6 7 8 9 /9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 97 98 99 00 01 02 03 04 05 06 07 08 recovery in the agricultural sector, especially the food All exports Industrial formal Industrial informal crop subsector; return to peace in northern Uganda and Source: Bank of Uganda 2008/09. some parts of the eastern region; and resettlement of the CHAPTER 2: BUILDING ON GROWTH IN UGANDA 55 Table 2.5 Poverty Headcount and Income Inequality Estimates for Uganda, 1992/93–2005/06 Poverty headcount Gini coefficient Region 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 National 56.4 38.8 31.1 24.5 0.365 0.428 0.408 0.426 Rural 60.3 42.7 34.2 27.2 0.328 0.363 0.363 0.375 Urban 28.8 14.4 13.7 9.1 0.396 0.483 0.432 0.447 Central 45.6 22.3 16.4 10.7 0.395 0.460 0.417 0.451 East 58.8 46.0 35.9 24.3 0.327 0.365 0.354 0.319 North 73.5 63.0 60.7 46.2 0.345 0.350 0.331 0.367 West 52.7 32.9 20.5 21.8 0.319 0.359 0.342 0.375 Source: Ssewanyana and Okidi 2007; Uganda Bureau of Statistics 2010. Note: The poverty headcount is the percentage of people estimated to be living in households with real private consumption per adult equivalent below the poverty line for their region. internally displaced persons partly accounted for the involving careful sequencing and determined implementa- decline in poverty between 2005/06 and 2009/10. tion. The World Bank has referred to Uganda’s efforts as “the Although absolute income poverty fell during the past most far-reaching stabilization and structural reform pro- two decades, the distribution of income worsened. Income gram in Africa, and one of the most comprehensive reform inequality as measured by the Gini coefficient increased efforts in the world” (World Bank 2007, 4). from 0.365 in 1992/93 to 0.408 in 2005/06. Growing income The first reform was the Economic Recovery Program, disparities are evident between 2005/06 and 2009/10, when introduced in 1987, with support from the World Bank and the Gini coefficient reached 0.426. Growth rates for con- International Monetary Fund (IMF), which focused on sumption grew more rapidly among the richer quintiles, price stabilization and liberalization. Policies under this increasing inequality (Ssewanyana and Okidi 2007; Uganda program included currency reform, devaluation, liberaliza- Bureau of Statistics 2010). Income inequality between rural tion of domestic prices, and conversion to a floating and urban areas and between regions widened between exchange rate regime (in 1993). 2005/06 and 2009/10. The next set of reforms involved the adoption of the Increasing inequality has slowed the rate of poverty structural adjustment program that was meant to free up reduction in Uganda. For every 1 percent decrease in growth, markets and create price incentives, stimulate private the percentage of people living below the poverty line will investment, and encourage competition. Reforms under increase by 2 percent, holding income distribution constant this program included the abolition of marketing boards, (Ssewanyana 2009). This means that the economy has to the privatization or abolition of parastatals, and the grow by at least 7 percent and household consumption has establishment of the Uganda Investment Authority. This to rise at least 4 percent if Uganda is to avoid reversals in its period was characterized by sustained macroeconomic poverty reduction efforts. stabilization, adjustment, and structural reform efforts Uganda is likely to attain MDG 3 (promoting gender that affected almost all sectors of the economy. Policies equality and empowering women), MDG 6 (combating mainly involved the macroeconomic stabilization process, HIV/AIDS), MDG 7 (ensuring environmental sustainability), price liberalization, financial sector liberalization, public and MDG 8 (developing a global partnership for develop- enterprises reform, and civil service reform. In addition to ment). The stagnation in net primary school enrollment since changing and stabilizing the structure of the economy, 2003 at about 85 percent is a clear indication that intensified reorientation of pricing and marketing policies, restarting efforts are required if Uganda is to meet MDG 2, however, of economic growth, and strengthening the institutional and attainment of MDG 4 (reducing child mortality) and framework constituted the major cornerstones of the MDG 5 (improving maternal health) is unlikely even with program. To achieve these ends, the Economic Recovery improved policies, institutions, and funding (UNDP 2010). Program focused on ensuring macroeconomic stability; liberalizing the foreign exchange system, trade, prices, and marketing systems; improving the incentive structure REFORMS DRIVING GROWTH and business climate to promote savings mobilization and The past two decades have seen tremendous economic investment; and rehabilitating the economic, social, and transformation in Uganda, fueled mainly by good policies institutional infrastructure. 56 CHAPTER 2: BUILDING ON GROWTH IN UGANDA With the economy back on its footing, in 1997 the govern- was aimed at spurring growth as well as the level of produc- ment introduced the Poverty Eradication Action Plan (PEAP), tivity of assets of poor people. Policies undertaken under a multisectoral program aimed at reducing poverty. Policies in the framework of the PEAP included universal primary and this program included the Plan for the Modernization of secondary education, intended both to raise education indi- Agriculture (PMA), which sought to address agricultural cators and to remove the financial burden of education constraints to production and to turn agriculture commer- from parents as a means of reducing poverty. Efforts were cial. The plan has not been as successful as envisaged, mainly also made to increase health coverage, by constructing because it was too broad and in some cases ambiguous, with health centers in all subcounties and parishes. several programs having too many pillars when the focus Examination of the composition of the public budget for should have been on enhancing agriculture productivity. most of the 1990s and 2000s reveals that total government The National Agricultural Advisory Services, the flagship expenditure steadily increased, from about 18.6 percent of agricultural productivity enhancement program within GDP in 1992/93 to about 32.0 percent in 2008/09. Evidence PMA, was beset by inefficiencies and other implementation suggests that increased funding to social services has bene- problems that limited its impact on agriculture.2 fited the poor, especially in rural areas (Kappel, Lay, and Other sectoral reforms that contributed to the liberaliza- Steiner 2005). tion and stabilization of the economy included the With reasonable progress in the social sectors, in the late Medium-Term Competitive Strategy for the Private Sector, 2000s the government began addressing the infrastructural the Strategic Export Program, and the Strategic Export constraints that have dogged the country since indepen- Intervention Program. These policies were accompanied by dence. Since 2007/08 substantial resources have been com- important institutional reforms, such as decentralization mitted to the rehabilitation and construction of roads and efforts, the abolition of state-owned marketing boards, and hydroelectric dams. The government has prioritized the the restructuring of the public administration. building of roads in the medium and long term, with more The 1990s saw a substantial reversal in the decline of the than 20 trunk roads planned or in the process of being built economy that had characterized the 1970s and early 1980s, in the next 20 years (MFPED 2009). The share of total cen- suggesting that reform worked. Confidence in the economy tral government budget allocated to works and transport was restored, spurring substantial inflows of aid and foreign rose from 10.0 percent in 2005/06 to 18.4 percent in direct investment and a reversal of capital flight (figure 2.6). 2009/10, and the share of the budget allocated to energy and Most economic indicators rebounded, and by 1996 the mineral development rose from 3.5 to 10.3 percent. Taken economy had recovered to its nominal 1971 dollar per together, allocations for works, transport, and energy repre- capita GDP (World Bank 2007). sented almost a third of the 2009/10 budget—a substantial To sustain rapid economic growth, the government figure given Uganda’s historically low investments in infra- needed to reorient expenditures toward social sectors and structure. Although issues about absorption capacity and increase spending on infrastructure. Social policy spending quality of infrastructure remain, there is a perception that the government’s tightened focus on improving infrastruc- Figure 2.6 Official Development Aid as Percentage of ture will help reduce Uganda’s perennial infrastructural GDP in Uganda, 1998/99–2008/09 shortages, giving the economy a major boost. 14 CHALLENGES TO DEVELOPMENT 12 Policy makers need to address a variety of challenges in 10 reducing poverty and spurring growth. In addition to the Percent 8 challenges identified in this section, they need to ensure that 6 growth is inclusive. Failure to achieve equity is likely to 4 exacerbate unemployment and lead to social unrest. 2 0 Slow progress in reducing poverty and slowing population growth 9 0 1 2 3 4 5 6 7 8 9 /9 /0 /0 /0 /0 /0 /0 /0 /0 /0 /0 98 99 00 01 02 03 04 05 06 07 08 19 19 20 20 20 20 20 20 20 20 20 Although Uganda has already achieved MDG 1 (halving Source: Bank of Uganda 2009. extreme poverty), poverty remains high, especially in the CHAPTER 2: BUILDING ON GROWTH IN UGANDA 57 northern region. Poverty reduction interventions, especially constraints by increasing sector funding represent a move the Peace and Reconciliation Development Plan (PRDP), in the right direction and should be followed by prudent in this part of the country need to be restructured to ensure use of the resources by responsible agencies to avoid leak- that they reach and benefit vulnerable groups. ages, which have often derailed service delivery. Regional Even at the average GDP growth rate of 6.9 percent infrastructure also remains a key challenge, especially for attained during the 1990s, it would take Uganda about 20 boosting regional trade. years to double average per capita income. With population growth rate at 3.2 percent a year, one of the highest popula- Lack of sufficient private sector development tion growth rates in the world, it is going to be very difficult to reduce poverty (World Bank 2007). Private sector growth has been impressive since liberaliza- Uganda’s 3.2 percent annual population growth thwarts tion, but Uganda’s economy remains dominated by small development, especially in social services provision and firms that usually employ fewer than five people, making social outcomes. Unless efforts at reducing the high popula- it difficult to absorb the growing number of graduates and tion growth rates are intensified, many of the MDGs will be exacerbating the youth unemployment problem. In addi- difficult to achieve. tion, although the private sector has played a significant role in areas such as education and health services, it has not been prominent in other sectors that require signifi- Inadequate transformation of the cant investments, such as energy, and even in sectors in agriculture sector which the private sector is active, the impact of its activi- Land reform is a critical factor that will ensure sustainable ties on employment creation has been limited. Enlarging agricultural productivity and facilitate commerce and the role of the private sector calls for more active partici- agriculture. Such reform would involve the government pation of the government under private-public partner- buying out landlords in order to consolidate the small ship arrangements. plots on which many rural farmers depend to enable com- Government development programs and successive gen- mercial agriculture. The government could also try to erations of Country Strategy Papers have emphasized the draw the rural population to urban centers through more development of the private sector as a major goal for the planned urban development. country. But progress has been slow in this area, partly Other factors explaining low agricultural productivity because of weak human and institutional capacity, which include poor inputs, undeveloped value chains, and low limits private sector participation in execution of contracted public and private investment in the sector. The issue of low or self-initiated projects. Further limiting the growth of the productivity in agriculture will need to be addressed by, for private sector is the poor state of the financial sector, which instance, accelerating the creation of on-farm and off-farm limits access to credit. Equity finance is also lacking, and processing zones to add value to agricultural commodities, small businesses have limited access to commercial banking financing activities along the entire value chain as opposed to facilities. It is important that the government expedites the funding on-farm activities only, and land reforms that will recapitalization of the Uganda Development Corporation allow commercial agricultural production as articulated in the and the Uganda Development Bank to make private sector National Development Plan (Republic of Uganda 2010). access to credit easier. Weak infrastructure Narrow export base and terms-of-trade vulnerability Uganda’s infrastructure is among the worst in the world (Republic of Uganda 2010; World Bank 2007). The National Despite the diversification of its export base, Uganda Development Plan identifies weak infrastructure as one of remains heavily dependent on primary commodities. Diver- the key binding constraints in Uganda. Roads, power and sification of the export base is of paramount importance. railways are all below those of Uganda’s neighbors, with The factors that continue to constrain export diversification grave implications for the economy.3 The National Develop- include the primary and low-value-added nature of ment Plan identifies the most challenging infrastructural Uganda’s exports, poor product quality, and poor regulation impediments as power, transport, and access to and the high standards, which inhibit competition in marketing and cost of finance, in that order. Efforts to address infrastructure export of primary commodities. 58 CHAPTER 2: BUILDING ON GROWTH IN UGANDA Low tax revenue More needs to be done to improve governance in Uganda, where civil society organizations are weak and the Tax revenue in Uganda is low, the result mainly of untaxed legislature often fails to rein in Uganda’s powerful executive. sectors, especially informal businesses and some agricul- Unless governance is strengthened, even increased resources tural activities, and tax evasion (Sennoga, Matovu, and may fail to bring meaningful development to Uganda. Twimukye 2009). The lack of revenue has translated into an unending dependence on foreign aid, which accounted for about 32 percent of the budget in 2008/09. In 2007/2008 tax Weak human and institutional capability revenue amounted to about 12 percent of GDP. This figure Uganda’s Five-Year National Development Plan 2010–15 is low relative to Uganda’s neighbors (tax revenue was about identifies weak human and institutional development as one 17 percent of GDP in Tanzania and about 27 percent of of the economy’s key binding constraints. Most public sector GDP in Kenya for the same period). Lack of adequate tax departments are characterized by coordination failures, cor- revenues constrains government operations and weakens ruption, endemic malaise, and weak institutional linkages economic management. It also increases government among relevant stakeholders, including the Ministry of reliance on foreign aid to finance development, making eco- Finance Planning and Economic Development, sector line nomic management more difficult. ministries, and the private sector. As a result, budgeting The recent discovery of oil in Uganda could significantly processes are inefficient, and scarce investment resources are boost the resources mobilized domestically, widening not allocated rationally. Problems are particularly acute at resources available to finance Uganda’s development agenda. the local government level. The central government needs to The government could also try to expand the tax base, by address these problems as it devolves planning, delivery, and targeting sectors that are currently untaxed, especially the management of basic services to local governments (EPRC informal sector, and setting up regulation to reduce tax 2010). There is a need to increase human and institutional evasion.4 capacity to meet the growing needs of local governments as the scope of their functions broadens. Corruption and weak governance The government has made numerous efforts to strengthen CONCLUSION good governance, by allowing civil society organizations to participate more in planning and budgeting and by setting Uganda has experienced relatively strong growth and up institutions to strengthen accountability. Such initiatives poverty reduction since the 1990s. The growth period was included the creation, in 2008, of the anticorruption court characterized by sustained macroeconomic stabilization, to handle corruption-related cases and the value-for-money adjustment, and structural reform efforts, which had a large audit unit in the Auditor General’s office to check on gov- impact on most sectors. Efforts included price liberaliza- ernment spending at all levels. Uganda has also adopted a tion, financial sector liberalization, public enterprises national anticorruption strategy. reform, and reform of the civil service. These efforts notwithstanding, the perception remains Growth has been accompanied by rising inequality and that corruption is rampant, and the government has come very high unemployment levels, however, especially among under pressure from several quarters to show more political youth. Moreover, growth has been registered in sectors will to address the problem. Uganda’s ranking on the (particularly services) whose contribution to employment Ibrahim Index of African governance improved from 27th is limited. More needs to be done to translate impressive in 2007 to 19th in 2008. But in May 2009, Transparency performance at the macro level into improvements in International ranked Uganda the third most corrupt coun- the welfare of the majority of Uganda’s people, especially try in the world. And although there has been considerable the rural poor. progress in tax administration with the formation of the Looking forward, Uganda needs to address several issues Uganda Revenue Authority in 1991, tax policy and adminis- in order to enjoy equitable growth. First, policy makers need tration remain issues as they relate to revenue mobilization, to increase the pace of transformation of the agricultural partly because of corruption within the tax body. In 2010 sector and strengthen the sector’s weak link to industry, Transparency International ranked Uganda’s revenue where outputs from agriculture can be used as inputs. Given authority the second most corrupt tax body in East Africa that the sector employs 70 percent of the workforce, doing so (Transparency International 2010). would spur equitable growth and reduce high unemployment. CHAPTER 2: BUILDING ON GROWTH IN UGANDA 59 Second, policy makers have to find new ways to mobilize policy makers also need to address other key challenges, domestic revenues to finance the budget. To address the including weak human and institutional capability, a weak growing trade deficit, Uganda should continue diversifying private sector, poor infrastructure, poor governance, and its export base into higher-value-added products. Third, high population growth rates. ANNEX 2.A Table 2.A1 GDP in Uganda, by Economic Activity, 2001/02–2008/09 (percent, at constant 2002 prices) Item 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 GDP at market prices 8.6 6.6 6.8 6.3 10.8 8.4 9.0 7.2 5.8 Agriculture, forestry, and fishing 7.1 2.1 1.6 2.0 0.5 0.1 1.3 2.5 2.1 Cash crops 12.5 3.2 7.3 –5.5 –10.6 5.4 9.0 5.6 –2.9 Food crops 5.7 2.2 –1.5 –0.2 –0.1 –0.9 2.4 2.6 2.7 Livestock 4.0 3.5 4.7 3.0 1.6 3.0 3.0 3.0 3.0 Forestry 6.8 5.2 3.1 6.5 4.1 2.0 2.8 6.3 2.4 Fishing 13.8 –4.3 9.6 13.5 5.6 –3.0 –11.8 –0.7 2.6 Industry 7.4 9.5 8.0 11.6 14.7 9.6 9.1 5.8 8.9 Mining and quarrying 12.2 12.8 1.7 27.2 6.1 19.4 3.0 4.3 12.8 Manufacturing 6.7 4.4 6.3 9.5 7.3 5.6 7.6 10.0 5.9 Formal 7.7 4.6 8.3 11.8 7.8 4.9 9.2 12.0 6.1 Informal 4.5 4.0 1.7 3.6 6.0 7.7 3.3 4.4 5.5 Electricity supply –1.7 3.7 7.7 2.1 –6.5 –4.0 5.4 10.6 8.9 Water supply 3.0 3.9 4.2 3.9 2.4 3.5 3.8 5.7 2.7 Construction 10.1 14.6 10.0 14.9 23.2 13.2 10.8 3.7 10.9 Services 11.0 7.4 7.9 6.2 12.2 8.0 10.2 8.8 5.8 Wholesale and retail trade; repairs 7.4 5.1 6.3 7.2 12.3 10.4 14.7 9.7 –0.3 Hotels and restaurants 2.8 8.2 9.5 6.5 8.7 11.3 10.7 4.5 4.5 Transport and communications 17.8 14.9 15.8 9.8 17.1 17.7 21.3 14.3 15.1 Road, rail, and water transport 6.0 5.6 8.9 6.7 12.8 9.5 20.8 12.9 2.8 Air transport and support 0.5 5.8 13.8 19.4 6.9 13.8 17.8 –3.6 –1.2 Posts and telecommunication 76.5 40.4 28.6 11.8 26.2 29.1 22.6 19.8 30.3 Financial services 32.6 13.2 0.0 13.0 31.7 –11.9 17.1 25.4 21.1 Real estate activities 5.4 5.5 5.5 5.5 5.6 5.6 5.6 5.7 5.7 Other business services 11.6 7.6 7.0 9.2 12.5 8.0 10.8 12.4 10.4 Public administration and defense 20.4 3.6 7.7 –5.4 15.8 –6.3 12.1 5.5 3.9 Education 14.2 7.2 9.1 4.4 9.4 10.6 –6.5 4.3 –0.5 Health 18.0 13.7 0.9 5.6 12.9 2.7 –4.8 –3.2 11.0 Other personal services 8.5 8.5 16.1 15.0 14.1 13.4 12.8 12.3 11.8 Source: Bank of Uganda 2009. Table 2.A2 Composition of Uganda’s Exports, 1997/98–2008/09 (percent) Year Coffee exports Noncoffee exports Informal cross-border exports 1997/98 58.7 41.3 0.0 1998/99 55.9 44.1 0.0 1999/2000 42.0 56.6 1.3 2000/01 24.5 75.5 0.0 2001/02 18.1 81.9 0.0 2002/03 20.8 77.9 1.3 2003/04 17.0 78.4 4.6 2004/05 16.3 71.4 12.3 2005/06 16.6 67.3 16.0 2006/07 15.2 67.4 17.4 2007/08 13.4 45.4 41.2 2008/09 10.9 38.9 50.2 Source: Bank of Uganda. 60 CHAPTER 2: BUILDING ON GROWTH IN UGANDA Table 2.A3 Poverty Headcount, Poverty Gap, Poverty Severity, and Income Inequality Estimates, 1992/93–2005/06 Poverty headcount Poverty gap Severity of poverty Gini coefficient Region 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 1992/93 2002/03 2005/06 2009/10 National 56.4 38.8 31.1 24.5 20.9 11.9 8.7 6.8 10.3 5.1 3.5 2.8 0.365 0.428 0.408 0.426 Rural 60.3 42.7 34.2 27.2 22.6 13.1 9.7 7.6 11.2 5.7 3.9 3.1 0.328 0.363 0.363 0.375 Urban 28.8 14.4 13.7 9.1 8.7 3.9 3.5 1.8 3.7 1.6 1.4 0.6 0.396 0.483 0.432 0.447 Central 45.6 22.3 16.4 10.7 15.3 5.5 3.6 2.4 7 1.9 1.3 0.8 0.395 0.46 0.417 0.451 East 58.8 46 35.9 24.3 22 14.1 9.1 5.8 10.9 6 3.4 2.1 0.327 0.365 0.354 0.319 North 73.5 63 60.7 46.2 30.3 23.4 20.7 15.5 15.8 11.5 9.2 7.3 0.345 0.35 0.331 0.367 West 52.7 32.9 20.5 21.8 18.7 8.5 5.1 5.4 9.0 3.3 1.8 2.0 0.319 0.359 0.342 0.375 Source: Ssewanyana and Okidi 2007. Note: The poverty headcount is the percentage of people estimated to be living in households with real private consumption per adult equivalent below the poverty line for their region. The poverty gap is the sum over all individuals of the shortfall of their real private consumption per adult equivalent and the poverty line divided by the poverty line. Severity of poverty is the squared poverty gap (the sum over all individuals of the square of the shortfall of their real private consumption per adult equivalent and the poverty line divided by the poverty line). 61 Table 2.A4 Targets for and Status of Progress toward Millennium Development Goals and Poverty Eradication Action Plan (PEAP) Target possible with better Target policies, 2007/08 2013/14 2015 possible at institutions, PEAP PEAP MDG current and additional MDG/target/indicator 1990 2005/06 2009/10 target target target trend? funding? 1. Eradicate extreme hunger and poverty: Halve, between 1990 and 2015, the proportion of people whose income is less than $1 a day 1.1 Poverty headcount ratio (percent) 56 31.1 24.5 n.a. 28 28 Met Yes 1.2 Prevalence of child malnutrition (percent of children under five) 23 23 n.a. n.a. n.a. 12 No Yes 2. Achieve universal primary education: Ensure that, by 2015, children everywhere, boys and girls alike, will be able to complete a full course of primary schooling 2.1 Net primary enrollment ratio (percent of children 6-12) Boys 58 84 82.4 90 100 100 Yes Yes Girls 48 85 83.2 89 100 100 Yes Yes 2.2 Primary completion rate (percent of boys and girls) n.a. 56 n.a. 69 n.a. 100 No Yes 3. Promote gender equality and empower women: Eliminate gender disparity in primary and secondary education, preferably by 2005, and in all levels of education no later than 2015 3.1 Ratio of girls to boys in primary education (percent) 83 99 n.a. 100 100 100 Met Yes 4. Reduce child mortality: Reduce by two-thirds, between 1990 and 2015, the under-five mortality rate 4.1. Under-five mortality rates (per 1,000) 177 152 n.a. n.a. n.a. 53 No Uncertain 4.2 Infant mortality rate (per 1,000 live births) 98 88 n.a. 68 n.a. 32 No Uncertain 4.3. Immunization against DPT (percent of all children) 45 83 n.a. 90 n.a. n.a. n.a. 5. Improve maternal health: Reduce by two-thirds, between 1990 and 2015, the under-five mortality rate 5.1 Maternal mortality ratio (modeled estimate, per 100,000 live births) n.a. 505 n.a. 354 n.a. 126 No Uncertain 5.2 Deliveries in health care centers (percent of all deliveries) n.a. 24 n.a. 50 n.a. n.a. Met Yes 6. Combat HIV/AIDS, malaria and other diseases: Have halted by 2015 and begun to reverse the spread of HIV/AIDS 6.1 Prevalence of HIV, total (percent of adult population) 20 6.2 n.a. 5 n.a. <20 Met Yes 7. Ensure environmental sustainability: Integrate the principles of sustainable development into country policies and programs and reverse the loss of environmental resources (continued next page) 62 CHAPTER 2: BUILDING ON GROWTH IN UGANDA Table 2.A4 (continued) Target possible with better Target policies, 2007/08 2013/14 2015 possible at institutions, PEAP PEAP MDG current and additional MDG/target/indicator 1990 2005/06 2009/10 target target target trend? funding? 7.1 Forest area (percent of total land area) n.a. 24 n.a. 27 30 >24 7.2 Access to safe water (percent of population) Urban 45 86.8 92.3 100 n.a. 90 Yes Yes Rural 45 63.6 69.5 90 n.a. 90 Yes Yes 7.3 Access to improved sanitation (percent of population) Urban n.a. 65 n.a. 100 n.a. Rural n.a. 55 n.a. 80 n.a. 7.4 Tilled land (percent of land) n.a. 13 n.a. 17 25 8. Develop a global partnership for development 8.1 Debt service (percent exports of goods and services) 305 238 187 Yes Yes Source: 2004 PEAP; Uganda Bureau of Statistics various years. Note: n.a. is not applicable. PEAP is Poverty Eradication Action Plan. NOTES BIBLIOGRAPHY 1. Some reversals in the decline occurred during this period: AfDB (African Development Bank). 2009. “Uganda Eco- the share of poor people rose from 34 percent in 1999/00 to nomic Outlook 2009/10.” Africa Economic Outlook 38 percent in 2002/03 (Appleton and Ssewanyana 2004). 2009/10. Tunis. 2. The program targeted relaxation of marketing infra- Appleton, S., and S. Ssewanyana. 2004. “Poverty Estimates structure constraints, production and dissemination of from the Uganda National Household Survey of technology, removal of financial constraints, improvement 2002/03.” Economic Policy Research Centre, Kampala. of land tenure and policy, formation of farmers organiza- Bank of Uganda. 2009. Annual Report 2008/09. Kampala. tions, improvement of human resources and information, Bevan, D., C. Adam, J. Okidi, and F. Muhumuza. 2003. “Eco- promotion of on-farm and off-farm storage, and efforts to nomic Growth, Investment and Export Promotion.” reduce environmental degradation and mitigate the effects Paper prepared for Uganda’s Poverty Eradication Action of HIV/AIDS. Plan Revision, 2002/03. 3. Electricity coverage stood at 8 percent of the population Collier, P., and R. Reinikka. 2001. “Reconstruction and Lib- in 2009, far below coverage in other low-income countries, eralization: An Overview.” In Uganda’s Recovery, The which stands at 35 percent. At 190 meters per square kilo- Role of Farms, Firms and Government, ed. R. Reinikka meter, the density of roads in Uganda is much higher than and P. Collier, 15–48. Washington, DC: World Bank. in the average low-income country (where the average is 126 meters per square kilometer), but the roads are of poor Dijkstra, A. G., and J. K. Van Donge. 2001. “What Does the quality, with only 4 percent of the road network paved. Only ‘Show Case’ Show? Evidence of and Lessons from Adjust- 26 percent of the railway network in Uganda is functional. ment in Uganda.” World Development 29 (5): 841–63. Access to improved water stands at 51 percent of the EPRC (Economic Policy Research Centre). 2010. Corrup- population, far below the 64 percent for the average of a tion Trends in Uganda: Using Data Tracking Mechanism. low-income country (Republic of Uganda 2010). Kampala. 4. Taxing the informal sector could be implemented by Kappel, R., J. Lay, and S. Steiner. 2005. “Uganda: No More taxing small establishments based on characteristics such Pro-poor Growth?” Development Policy Review 23 (1): as turnover, profits, and location of business. 27–53. CHAPTER 2: BUILDING ON GROWTH IN UGANDA 63 MFPED (Ministry of Finance Planning and Economic ———. 2010. “Combating Chronic Poverty in Uganda: Development). 2009. “Budget Speech for Financial Year Towards a New Strategy.” EPRC Research Series 67, 2009/2010.” Delivered at the Fourth Session of the Eighth Economic Policy Research Centre, Kampala. Parliament of Uganda, June 11. Ssewanyana, S., and J. A. Okidi. 2007. “Poverty Estimates Okidi, J. A., S. Ssewanyana, L. Bategeka, and F. Muhumuza. from the Uganda National Household Survey of 2007. “Uganda’s Experience with Operationalising Pro- 2005/06.” EPRC Occasional Paper 34, Economic Policy Poor Growth, 1992–2003.” In Delivering on the Promise of Research Centre, Kampala. Pro-Poor Growth: Insights and Lessons from Country Transparency International. 2010. www.transparency.org Experiences, ed. T. Besley and L. J. Cord, 169–98. Wash- Uganda Bureau of Statistics. 2009. Statistical Abstract. Kampala. ington, DC: Palgrave Macmillan and World Bank. ———. 2010. Uganda National Household Survey 2009/10: Republic of Uganda. 2010. National Development Plan Socio-Economic Module, Abridged Report. Kampala. 2010/11–2014/15. Kampala. UNDP. 2010. Uganda MDG Progress Report, Kampala. Senoga, E., J. M. Matovu, and E. Twimukye. 2009. “Tax World Bank. 2007. “Uganda, Moving beyond Recovery: Evasion and Widening the Tax Base in Uganda.” EPRC Investment and Behavior Change for Growth.” Country Research Series 63, Economic Policy Research Centre, Economic Memorandum Volume II: Overview, Report Kampala. 39221-UG, Africa Region, Poverty Reduction and Eco- Ssewanyana, S. 2009. “Growth, Inequality, Cash Transfers nomic Management Unit, Washington, DC. and Poverty in Uganda.” Country Study 19, International ———. 2009. World Development Indicators. 2009. Wash- Policy Centre for Inclusive Growth, United Nations Devel- ington, DC. opment Programme, Brasilia. http://www.ipc-undp.org/ ———. 2010. World Development Indicators. 2010. Wash- pub/IPCCountryStudy19.pdf. ington, DC. 64 CHAPTER 2: BUILDING ON GROWTH IN UGANDA CHAPTER 3 Rapid Growth and Economic Transformation in Mozambique, 1993–2009 Antonio M. D. Nucifora and Luiz A. Pereira da Silva ozambique’s economic growth after 1992 and the M (e.g., job creation) in other sectors and how to effectively poverty alleviation achieved since then constitute employ these revenues, while improving governance and an extremely successful development take-off. accountability. Average real GDP growth rate soared from 0 percent This chapter provides a brief overview of Mozambique’s (1981–1992) to 8.1 percent (1993–2008), making profound economic transformation and highlights the Mozambique the fastest-growing non-oil economy in Sub- major drivers of change and the challenges ahead. It does Saharan Africa over the period (AfDB 2009) (figure 3.1). As not discuss the macroeconomic policies and performance a result, per capita GDP has doubled since 1992 (figure 3.2). that underpinned the economic transformation over the The poverty headcount dropped rapidly during the early past two decades (for a detailed discussion of macroeco- years of this process, falling from 69 percent in 1996 to nomic and stabilization issues in Mozambique, see Clement 54 percent in 2002. and Peiris 2008). The political stability brought about by the end of the armed conflict in 1992, a first wave of structural reforms GROWTH DRIVERS AND ECONOMIC coinciding with responsible macroeconomic policies, and TRANSFORMATION the support of the international donor community have been the bedrock of Mozambique’s transformation. Sound macroeconomic environment allowed donors to They have enabled the country to better face the develop- contribute substantial amounts of aid to Mozambique, ment challenges that lie ahead—in particular, how to tap averaging about 14 percent of GDP a year since 1993 (figure into the country’s resources (labor and land) and the entre- 3.3). That aid inflows have remained fairly constant as a per- preneurial drive of its people to promote strong, sustain- centage of GDP over the period, during a period when the able, and inclusive economic growth. Large mineral economy grew at an average annual rate of more than 8 per- resources and the recent growth in extraction activities is cent, reflects the rapid growth in aid inflows in U.S. dollar allowing Mozambique to expand further its development terms. These large inflows financed investments in educa- strategy. On the one hand this new revenue source can tion and health, as reflected in the rapid improvements in allow the country to become less dependent on foreign aid human development indicators (table 3.1).1 They also and define with more independence new elements of its financed substantial investments in rebuilding the country’s development agenda. On the other, it raises a number of roads, ports, and railways, which had been shattered by textbook challenges such as how to trigger externalities 16 years of war. 65 Figure 3.1 Real GDP Growth in Mozambique and Sub-Saharan Africa, 1981–2005 20 15 10 5 Percent 0 –5 –10 –15 –20 05 97 01 81 91 99 03 83 85 87 89 93 95 20 19 20 19 19 19 20 19 19 19 19 19 19 Mozambique Sub-Saharan Africa (median) Source: World Bank World Development Indicators various years. Figure 3.2 Per Capita GDP in Mozambique and Sub-Saharan Africa, 1980–2006 400 300 200 In constant (2000) US$ 100 0 –100 –200 –300 80 82 84 86 88 90 92 94 96 98 00 02 04 06 19 19 19 19 19 19 19 19 19 19 20 20 20 20 Mozambique Sub-Saharan Africa (median) Distance from median Source: World Bank World Development Indicators various years. Good macroeconomic management also attracted sub- inflows has funded large investment projects in the mining stantial foreign direct investment (FDI). FDI inflows sector, underpinning recent export performance in increased from an average of 1.5 percent of GDP in Mozambique. 1993–98 to an average of 5.2 percent of GDP in 1999–2010 From a supply-side perspective, capital accumulation, (figure 3.4). In 2009 and 2010 FDI reached an estimated higher quality–adjusted labor input, and positive aggregate $900 million, about 9 percent of GDP. A large part of these productivity performance were important determinants of 66 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 Figure 3.3 Gross Annual Aid Inflows to Mozambique, 1998–2010 2,000 20 Millions of dollars 15 Percent of GDP 1,500 1,000 10 500 5 0 98 0 99 00 01 02 03 04 05 06 07 08 09 te 0 jec 01 19 19 20 20 20 20 20 20 20 20 20 20 d) ro 2 (p Gross aid inflows (percent of GDP), righthand side Gross aid inflows (USD million), lefthand side Source: Bank of Mozambique, IMF, and World Bank. Table 3.1 Selected Social Indicators for Mozambique, 1990–2008 Latest single year Indicator 1990–1996 2000–03 2006–08 Primary school enrollment (net %) 43.000 56.000 96.000 Primary school enrollment (gross %) 61.000 84.000 105.000 Ratio of girls to boys in primary and secondary education (%) 72.000 75.000 85.000 Under-5 mortality rate (per 1,000 live births) 212.000 178.000 138.000 Infant mortality rate (per 1,000 live births) 145.000 122.000 93.000 Life expectancy at birth (years) 27.000 42.000 42.000 Physicians per 1,000 people 0.012 0.024 0.030 Inmunization, DPT (% of children under 12 months) 60.000 72.000 72.000 Inmunization, measles (% of children under 12 months) 58.000 77.000 77.000 Access to improved water sources (% of population) 39.000 42.000 48.000 Access to sanitation facilities (% of population) 22.000 27.000 31.000 Source: World Bank staff, based on data from Instituto National de Estatistica. Note: Latest single year means the latest year for which data are available in a given period. growth in Mozambique. Growth accounting exercises indi- butions are subtracted—was also important. Figure 3.6 sug- cate that physical investments partly associated with mega- gests roughly balanced growth of inputs and productivity.2 projects and significant improvements in education led to Productivity increases contributed about a third to overall growth dynamics heavily influenced by the accumulation of growth between 1993 and 2008. These results are broadly human (quality-adjusted) and physical capital between consistent with those of Jones (2006) and Vitek (2010).3 1993 and 2008. Figure 3.5 shows the average capital contri- Jones presents evidence suggesting a positive contribution bution to growth strengthening from 1999 onward, a result (crowding in) of public capital accumulation in the after- consistent with the timing of mega-projects. Aggregate pro- math of the civil war, which opened the door for private ductivity growth (total factor productivity)—measured as investment to play an increasingly important role in the residual of output growth after labor and capital contri- Mozambique’s growth. CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 67 Figure 3.4 Foreign Direct Investment in Mozambique, 1993–2010 1,000 10 800 8 Millions of dollars Percent of GDP 600 6 400 4 200 2 0 0 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 ro 2 9 te 0 0 jec 01 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 d) FDI (Percent of GDP), right hand side (p FDI (millions of dollars), left hand side Source: Bank of Mozambique. Figure 3.5 Decomposition of Growth in Mozambique by Factors, 1993–2008 14 12 10 8 6 Percent 4 2 0 –2 –4 –6 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Capital contribution Labor contribution (quality adjusted) Total factor productivity GDP growth Source: World Bank staff calculations based on World Development Indicators (World Bank various years). Figure 3.6 Average Factor Contributions to Growth in A level change in aggregate investment from mega proj- Mozambique, 1993–2008 ects triggered strong export and import activity from 1999 onward, with direct impact on aggregate demand. The expenditure components of output (figure 3.7) indicate that GDP and consumption have been trending up since the end of the civil war but that growth after 1999 has been 32% different. Strong investment performance starting in 45% 1999—which matured into strong export performance from 2001 onward—allowed for continued increases in 23% output and consumption at the same time that imports were growing. This pattern inaugurated a different growth model for Mozambique, whereby strong exports, though Capital Labor (quality-adjusted) offset by accompanying imports, reflected stronger private Total factor productivity sector activity. Source: World Bank staff calculations based on World Development Indica- More recently, as a result of sharp increases in the price of tors (World Bank various years). aluminum from 2004 onward, the pace of profit repatriation 68 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 Figure 3.7 Components of GDP Expenditure in Mozambique, 1993–2006 160,000 140,000 120,000 Millions in local currency 100,000 80,000 60,000 40,000 20,000 0 93 94 95 96 97 98 99 00 01 02 03 04 05 06 19 19 19 19 19 19 19 20 20 20 20 20 20 20 Consumption Investment Exports Imports GDP Source: World Bank 2010a. Figure 3.8 Gap between GDP and GNP in Mozambique, 1994–2007 14 180,000 160,000 12 140,000 Millions in local currency 10 120,000 Growth rate (%) 8 100,000 6 80,000 60,000 4 40,000 2 20,000 0 0 00 01 99 06 03 05 02 04 08 07 98 93 94 97 96 95 20 20 19 20 20 20 20 20 20 20 19 19 19 19 19 19 GDP growth rate (left hand side) GNP growth rate (left hand side) GDP level (right hand side) GNP level (right hand side) Source: World Bank 2010a. associated with mega-projects has widened the gap between abundant factors of production, particularly labor, land, and GNP and GDP (figure 3.8). This stylized fact is a reminder entrepreneurial drive. of a key challenge for the country: the need to compliment Although the private sector has spearheaded large mega- the current export drive based on mega–projects with a projects in the mining sector, its role in economic growth diversification strategy that activates a broader set of growth over the past decade should not be overemphasized. drivers and taps more effectively into Mozambique’s more Sonne-Schmidt, Arndt, and Magaua (2008) developed a CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 69 methodology to evaluate the direct contribution of three Indeed, the positive total factor productivity reflects key operational mega-projects (Mozal, Sasol, and Moma) the reshuffling of labor resources and a rapid transforma- to growth at factor cost. They conclude that the direct tion in the output structure of the economy. In a postwar effects of these mega-projects were an increase in the context—where reconstruction efforts, a first wave of average annual growth rate between 1996 and 2006 of structural reforms, and a stabilized macroeconomic envi- about 1 percentage point (average annual growth rate was ronment spurred private sector activity—the Mozambican about 8 percent). Moreover, as foreign-owned, capital- economy posted double-digit growth rates in mining, intensive, export-oriented companies benefiting from manufacturing, construction, electricity, gas, and water. fiscal exemptions, mega-projects made only a small contribu- Sectoral output shares experienced substantial changes tion to job creation, tax revenue, use of domestic intermedi- between 1993 and 2006, especially in agriculture, which ate inputs, and profit reinvestment in Mozambique. nearly halved its output participation, despite the positive Benito-Spinetto and Moll (2005) present similar results. contributions of agriculture to overall economic growth Using a computable general equilibrium (CGE) model to (figures 3.9 and 3.10). Sectoral output reflects a deep replicate the performance of the Mozambican economy, transformation to more productive sectors that generated they find that about half of the 8 percent average GDP positive composition effects in aggregate productivity.4 growth during 1994–2004 was driven by catch-up growth in Nevertheless, although economic growth has been strong the agricultural sector (as people returned to the fields fol- and has underpinned the process of economic transfor- lowing the end of the war), a quarter was driven by the mation, the structure of the economy still remains nar- growth in aid inflows (and the resulting investments in rowly based on subsistence agriculture and a few isolated, social and physical infrastructure sectors), and a quarter albeit large, mega-projects. Agriculture, which employed was equally divided between mega-projects and growth in about 78 percent of the economically active population, other private sector activities. Clearly, the initial bet on accounted for 25 percent of GDP in 2009. It was followed mega-projects, despite the usual uncertainty in quantifying by manufacturing (12 percent, two-thirds of which was their precise economic effects, rested more on their capacity accounted for by one large aluminum smelter); trade and to generate immediate revenue rather than truly create retail services (11 percent); transport and communica- downstream externalities (especially jobs and demand for tions (10 percent); financial services (7 percent); and industrial and services inputs). extractive industries (1 percent). Going forward, a key Figure 3.9 Sectoral Contribution to Growth in Mozambique, 1994–2006 20 15 10 Percent 5 0 –5 94 95 96 97 98 99 00 01 02 03 04 05 06 19 19 19 19 19 19 20 20 20 20 20 20 20 Agriculture Mining Manufacturing Electricity, gas, and water Construction Services GDP growth Source: World Bank various years. 70 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 Figure 3.10 Changes in Sectoral Output Share in Mozambique, 1993–2008 45 45 40 40 Real growth rate (percent) Sectoral share (percent) 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 re g n w and g rin in io tu in er ct tu ul , M as ru at ac ric ,g st uf Ag on ity an ic C M tr ec El 1993 Share 2008 Share Average annual real growth (93–08) Source: World Bank various years. challenge is not only to support the continuation of struction, and transportation sectors between 2003 and intersectoral (structural) transformations but to ensure 2006 (table 3.3).6 The dominance of large companies in for- within-sector productivity gains by addressing impedi- mal employment generation reflects the lack of growth in ments to microeconomic efficiency (World Bank 2008a). small and medium enterprises. The “entry margin” (cre- Economic transformation in Mozambique entailed a ation of formal jobs through new firms) also appears con- profound rebalancing of sectoral roles, with the (pri- siderably more stifled for micro and small firms than for vate) service sector absorbing labor at a particularly medium and large firms, perhaps a reflection of the fact that strong pace. Agriculture displays the lowest level of labor entry costs and overall business environment constraints productivity of all sectors (figure 3.11). Between 1996 impinge more severely on smaller producers. and 2003 its output contribution to GDP fell by 6 per- The concentration of the economy in a few large compa- centage points (to 27 percent), and its share of employ- nies is also reflected in export patterns. The export basket ment in the total labor force fell by 8 percentage points remains extremely limited, with only 15 products register- (to 82 percent) (figure 3.12). Although output shares ing exports of more than $1 million a year. More than half were gained primarily by the industry sector, employ- of Mozambique’s exports remain concentrated in the Mozal ment shares were gained by the private service sector. aluminum smeltering enclave (figure 3.13), virtually all of While the increase in the employment share of the pri- whose production goes to the European Union. Electricity vate service sector has benefited the poor (World Bank and gas are also important export products (bought largely 2008b),5 the challenge to generate an even greater number by South Africa). of jobs at higher levels of the value chain remains. The Overall, Mozambique’s exports increased significantly overwhelming concentration of employment (more than (in real terms), with average annual growth rates of 78 percent) remains in agriculture. In urban areas, retail 14.9 percent over 1995–99 and 19.8 percent over 2000–08. trade and vehicle repair employ a quarter of all workers; Most of this growth, however, was driven by exports from manufacturing industries employ 7.5 percent (table 3.2). the mega-projects; although excluding mega-projects, Formal employment generation has been very limited in exports grew at an average rate of 10 percent a year during recent years and is confined to a few large companies. Estab- 2000–08, which represents a reasonable performance. lishment-level data from a sample of 600 firms suggest that Progress in penetrating Southern Africa Development large firms (more than 100 employees) and some medium- Community (SADC) countries continues; Mozambique large firms (about 100 employees) were responsible for the should position itself to exploit non–SADC markets as majority of jobs created in the manufacturing, services, con- well. Mozambique exploits only 2 percent of the potential CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 71 Figure 3.11 Labor Productivity in Mozambique by Sector (Relative to Agriculture), 1996 and 2003 30 Average labor productivity normalized 25 20 to agriculture 15 10 5 0 Industry Services (private) Services (public) 1996 2003 Source: World Bank 2008b. Note: Figures are normalized to agriculture. Figure 3.12 Changes in Employment versus Output Shares in Mozambique, by Sector, 1996–2003 100 80 Change in output share between 60 1996 and 2003 (percent) 40 20 0 –20 –40 –60 –50 0 50 100 150 200 Change in employment share between 1996 and 2003 (percent) Industry Services (public) Services (private) Agriculture Source: World Bank 2008b. Note: Services (private) include trade, and transport. Services (public) include health, education, and public administration. The larger the bubble, the higher the labor productivity. bilateral flows for all the products it exports (Gillson A STRATEGY FOR BROADLY SHARED 2008).7 Although this is comparable to other countries in ECONOMIC GROWTH the region, it is far lower than the levels achieved by more Rapid economic growth was accompanied by significant advanced and better integrated economies (such as South strides in reducing poverty during the stabilization phase, Africa) (figure 3.14).8 after the peace accords and up to 2003.9 Household survey 72 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 Table 3.2 Distribution of Employment in Mozambique, by Sector, Gender, and Location, 2004/05 (percent) Sector Total Men Women Urban Rural Agriculture and fishing 78.5 68.0 87.3 40.1 92.7 Extractive industries 0.3 0.6 0.0 0.4 0.2 Manufacturing industries 3.1 5.4 1.2 7.5 1.5 Electricity, water, and construction 1.2 2.6 0.1 3.5 0.4 Retail trade and vehicle repair 9.2 11.7 7.0 25.9 3.0 Transport and communication 0.8 1.6 0.1 2.6 0.1 Financial services and rental 2.9 3.9 2.0 9.5 0.5 Administration 1.7 2.9 0.7 5.3 0.4 Education 1.6 2.3 1.0 3.5 0.9 Health and social work 0.7 0.8 0.5 1.6 0.3 Source: INE 2006. Table 3.3 Number of Full-Time Employees in Sample of 600 Firms in Mozambique, by Firm Size, 2003 and 2006 Growth rate in Change in number of employees number of between 2003 and 2006 employees between 2003 and 2006 Number of Number of In firms created In firms created Firm size (percent) employees in 2003 employees in 2006 before 2003 after 2003 Micro 63 408 666 190 68 Small 35 2,064 2,803 493 246 Medium 36 4,630 6,323 992 701 Large 60 2,922 4,685 772 991 Total 44 10,024 14,477 2,447 2,006 Source: World Bank 2010b. Note: Full-time employment in 2003 was used to determine firm size. For firms created after 2003, employment in 2006 was used to determine firm size. Figure 3.13 Mozambique’s Main Exports, 2008 1,600 1452 1,400 1,200 Millions of dollars 1,000 800 600 400 226 193 200 152 71 62 58 37 33 31 26 5 5 3 2 0 um To y o as r p n ts ts s um s l a e oi ed na ga Te t aiz rim to cc uc nu ici G Su ut in ni na se ot ba M tr od sh um ta w on Ba C ec e he pr Ti m en oc Al El as sa d oz C oo C Se Fr W CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 73 Figure 3.14 Mozambique’s Export Market Penetration, by Country, 2006 90 80.5 80 70 60 Percent 50 40 30 20.4 20 9.3 10 0.4 0.5 0.6 0.7 1.7 2.2 0 sia ia n d a na i on ca w di tio lan ys ri hi ala ne ni In ala Af ra C er U do M de itz M h an In ut Fe Sw pe So ian ro Eu ss Ru Source: UN COMTRADE (various years); Gillson 2008. data indicate that the national poverty headcount fell expected promises. The model—which focuses on main- rapidly, from 69 percent in 1996 to 54 percent in 2003 taining macroeconomic stability, making public invest- (MPD 2004). Reduction in rural poverty was even more ments in infrastructure and increasing access to public pronounced, declining from 71 percent to 55 percent during services (education, health, water, and electricity), and the same period. attracting capital-intensive mega-projects (which are capital The pace of poverty reduction appears to be slowing, intensive)—is still not bringing the increase in jobs and pro- however, now that the post-war reconstruction effort and ductivity required to set the economy on a more diversified the agriculture catch-up has been exhausted, and the exist- growth pattern and eventually to reduce poverty by more ing pattern of economic growth is not generating a suffi- than under the current strategy. In addition, lack of progress cient number of jobs. The preliminary results of the 2009 on the good governance agenda is putting strain on nationally representative household budget survey are Mozambique’s relationship with some of its development mixed. Although there was significant progress on many partners, and raising further questions about the sustain- nonmonetary poverty indicators (such as access to educa- ability of Mozambique’s development strategy (box 3.1).11 tion and health services, increase in asset ownership by A more inclusive, labor-intensive economic growth strat- households, and improvements in housing quality), the egy is needed to reduce poverty and improve living stan- poverty headcount suggests that poverty may have stag- dards (World Bank 2010a). Diversifying the economy into nated at about 55 percent during 2003 and 2009 (MPD labor-intensive sectors (including agriculture, agroprocess- 2010). Urban poverty continued to decline, although at a ing, manufacturing, and tourism) and increasing the com- much slower rate, reaching 50 percent, but rural poverty petitiveness of domestic production to replace imports and increased to 57 percent.10 Although the food and fuel crisis diversify exports requires a new strategy focused on elimi- played a role in this outcome, the stagnation of poverty nating barriers to private investment (both domestic and reduction while the economy continued to grow at high foreign). Such a strategy would attempt to reduce excessive rates suggests that growth has become less inclusive in the regulation; simplify the trade and tax regime; reduce the past few years. costs of hiring and firing workers; improve labor force skills; These results are important because they seem to cor- eliminate the rigidity in the land tenure system (freeing up roborate what many academics in Mozambique (see, among access to land use rights and allowing their tradability); others, Castel-Branco 2002; Castel-Branco and Ossemane strengthen transport logistics and facilitate better services/ 2009) have been arguing for a number of years: that even articulation for exports; address the lack of standards; and though GDP growth rates have been very high in Mozam- maintain sound fiscal, monetary, and exchange rate policies. bique, the current development strategy is lagging behind A number of constraints have to be addressed through a 74 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 Box 3.1 Signs of Political Maturity: Is Mozambique’s “Aid-Supported Growth Engine” Beginning to Shift? Aid flows needed to finance public investments and private sector-led investments in these areas but expand service delivery are becoming increasingly also in the accompanying infrastructure for export uncertain in Mozambique but also for other Low Income corridors. Countries. This is due, in part, from the budgetary con- Inevitably these prospects brought more attention sequences of the 2008 global financial crisis in ODA allo- to issues of economic governance such as the relation- cation in many OECD countries. ship between an emerging class of businesses and OECD donors have been considering Mozambique entrepreneurs with the dominant political elite. Fol- an emblematic success story since the 1990s, especially lowing several years of unhappiness about mixed since the end of the Civil War (1992). Official develop- progress in the good governance agenda and fears that ment assistance flows were significant and relatively increasing growth of non-aid revenue would allow stable both for projects and for direct budget support, Mozambique to become less aid-dependent. Some given the country’s good performance and “donor- OECD donors in late 2009 threatened to “cut budget agencies” high ratings. Throughout the early 2000s support” if very specific actions on political reforms changes occurred in Mozambican politics: from a rela- were not urgently undertaken. While the 2009–10 tively strong showing in the first, post–Civil War elec- standoff has since been largely resolved, some of the tions in 1994, RENAMO’s electoral results declined reasons and the political dynamics that led to the subsequently and posed a lesser threat to FRELIMO- impasse appear to signal a structural change in dominated politics. Changes also affected the FRELIMO Mozambique’s relationship with the budget support leadership, rotating toward a more business-oriented donors. Many donors, however, consider such interfer- framework. With more investment opportunities ence in the country's political affairs as unwarranted brought by macroeconomic stabilization, positive and refrain from entering into this type of conditional- donor sentiment, and increased aid inflows, among ity. In fact, the government has also started to consider other things, a shift toward a market-oriented frame- different possible scenarios to reduce its political vul- work took place, accentuated by the discovery in the nerability to aid-dependence, ranging from the possibil- mid-2000s of significant mineral resources that trig- ity of revising some contracts with some mega-projects gered private sector interest and investment. In recent that benefit from generous tax incentives, or turning to years, the discovery of such sizable reserves of miner- some of BRIC countries for support, in addition to the als began to engender business interest in the country need to expand the tax collection base and diversify by global players, triggering expectations of large, revenues, among other strategies. mix of simultaneous policy reforms, investments, and applying for them. The standard statutory tax regime strengthening of institutions. They can be categorized as (excluding incentives) continues to place a substantial bur- business environment, factor markets, trade logistics, and den on investors, notably small domestic entrepreneurs. The export-supporting institutions. combination of a statutory tax regime with relatively high Aggressive reforms are required to improve the business tax rates with a generous system of tax incentives influences environment and entice potential investors. First, the regula- investment decisions and the allocation of resources in a tory environment is very unfriendly to small and medium way that is not immediately visible but distortionary in the enterprises, and licensing, inspections, and red tape are a long run. One widely recognized distortion is that the past heavy burden. This is arguably the most important priority tax regime encouraged industries that used more capital needing attention to unleash private sector potential, starting rather than labor. Mozambique’s fiscal climate could be with domestic investors. Second, the Mozambican tax and significantly improved by simultaneously reducing fiscal tariff system is characterized by many different exemptions incentives and the number and level of tax rates without whose cumulative effects are unclear. Additionaly, small compromising government revenues. Third, despite entrepreneurs normally do not benefit from current exemp- progress in recent years, tax administration remains tions, because they cannot afford the fixed cost involved in another impediment to Mozambique’s competitiveness CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 75 and productivity. A large number of the firms interviewed Legislation for the financial sector has been evolving in the recent Investment Climate Assessment (World Bank continuously, and all banks were privatized in the 1990s. 2010b) reported that tax administration is a major or severe However, capital markets remain very shallow. The banking obstacle to doing business. Reducing the number of taxes on system is now sound and highly profitable. Exceptionally the books and simplifying compliance with tax legislation high ratios of capital adequacy (17 percent on Basle I) and would reduce transaction costs of businesses. liquidity (55.4 percent) of commercial banks raise a legiti- Factor markets—land, labor, and capital—need to allocate mate question as to whether there is sufficient financial resources to their most productive uses. The 1997 Land Law intermediation for a fast-growing economy. Only a quarter is generally considered to be exemplary in Africa. However, of all districts have any kind of banking service, housing 10 years after its approval, its implementation is weak, and the finance faces big constraints, and trade finance is still reality on the ground is one of extensive extralegal land mar- underdeveloped. kets, multiple claims to the same pieces of land, and a lack of Trade logistics remain a serious constraint. Mozam- guidance on how to protect or compensate customary and bique has the potential to be a major outlet for southern good-faith occupation rights. This situation has led to exten- Africa based on its strategic location and congestion in sive land speculation and corruption, potentially leading to Durban port, provided that it becomes more logistics land conflicts and landlessness in the future. In the short run, friendly. Maputo port is the closest route for the largest the land situation acts as a deterrent to investors, particularly mining and manufacturing region of South Africa, while in tourism and agribusiness. Access to land for productive Beira and Nacala are the closest ports for the other neigh- purposes needs to be improved. The new labor law approved boring countries. Yet corridor traffic potential remains in 2007 has substantially improved labor regulations for new mainly untapped. Traffic going to and from Gauteng and investments. However, the labor market could and should be Mpumalanga provinces equals at least 700 million tons. If allowed to work even better to attract more domestic and for- Maputo could capture 1 percent of this traffic, its total eign investment and sustain high growth rates. throughput would be doubled and port revenues would be In addition, despite significant progress, education levels increased by several tens of millions of dollars. Despite a in Mozambique are still very low and constitute a structural rather successful concession process, transit traffic at constraint for job creation in the formal sector. Quality of Maputo port remains relatively low, and the port operates education is key for inclusive growth. Professional services at less than 30 percent of its capacity. Beira port has also (such as engineering, auditing, legal, and medical services) operated at less than 40 percent capacity for the past five are expensive and in short supply in Mozambique; at the years. Mozambique’s main deficiencies for trade mostly same time, many highly skilled Mozambicans are emigrat- derive from logistics problems and low transport reliability. ing because they can earn more money abroad. This Low ratings in the logistic performance index (LPI) (2.29) paradox of not having enough supply of skills (making pro- and in the timeliness index (2.83) put Mozambican ports at fessional services very expensive) while at the same time not a disadvantage compared with the port of Durban (with an having enough demand for them (leading to emigration) LPI of 3.53 and a timeliness index of 3.78). The World needs to be analyzed carefully to come up with an appropri- Bank’s 2008 Doing Business report highlights a web of pro- ate package of policy reforms. Inappropriate standards often cedures taking 27 days for exports and 38 days for imports. stifle demand for services in areas such as engineering and Some large South African shippers are still reluctant to accounting. Professional associations can play a key role in shift their transport routes from Durban to Maputo because creating, with government support, a framework for regula- they still perceive the business climate there as unpre- tion and appropriate standards, and also monitor quality dictable, including a higher incidence of bribe payments for and skills development. Immigration rules for qualified spe- the port of Maputo than for Durban (World Bank 2010a). cialists who can provide managerial and technical expertise Also, the poor integration of trucking services with the rest to Mozambican enterprises are still very restrictive, particu- of the subregion explains why South African shippers are larly for companies providing professional services. In the still reluctant to use Mozambican ports. High transport short run, a more liberal policy of granting work visas to unpredictability is closely linked to low traffic volumes. As skilled foreigners would help to develop the market for pro- a result, shipping lines do not call directly in Maputo and fessional services in Mozambique and would enable cargo has to be transshipped in Durban or Mombasa. Mozambican companies to compete better with foreign Physical investments and reforms of customs procedures companies providing such services. should be carried out simultaneously to reach a critical mass 76 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 in port transport volume. At that point, international ship- mineral resources has produced positive results. The country ping lines will make direct calls, which will greatly improve has achieved high growth rates, is getting significant predictability. Constructing the “one-stop border post” in amounts of export revenue and has begun to reduce its aid- Ressano Garcia for trade with South Africa is of highest pri- dependency, especially vis-a-vis traditional Western donors. ority and should be quickly emulated in other large border However, the trickle-down effect of such a bet has not (yet) posts. At the same time, all the customs clearance proce- produced the expected impact on local businesses and labor dures including at the dry ports have to be revised. Long markets. Moreover, uneven income distribution in urban truck immobilization creates unnecessary delays and costs areas and rising food prices during the last global commod- for shippers and transporters. Capital dredging of the Beira ity price cycle triggered protests. port is also very important, as is the reconstruction of the Shifting growth dynamics toward a more inclusive road between Nampula City and the border of Malawi. pattern based on economic diversification and increased Strengthening specialized export-supporting institu- participation of small and medium enterprises will require tions that can assist small- and medium-size exporters is policy action. Businesses’ interest in expanding and initiat- important for product and market development. Needed ing operations in Mozambique is palpable. This ongoing are a national system for standards and quality controls; an diversification of productive activities is bound to spur ICT-based proactive trade information system; product ID transformations in the economic landscape. Economic cards, and CITEs (centers supporting knowledge, innova- diversification and increasing engagement of small and tion, and technology transfer and services otherwise not medium enterprises leveraged by sustained efforts to partic- available). In addition, more “articulators” are needed to ipate in world markets represent a strategic orientation in connect Mozambican businesses and products with the Mozambican growth model that is required to acceler- regional and global markets. For a country where produc- ate growth and reduce poverty. Harnessing such transfor- tion is highly fragmented, such articulation initiatives are mations for economic development will require strategic critical to secure scale benefits and knowledge transfer. policy steps. By promoting competition and trade integra- There are already a number of ongoing articulation efforts tion measures, the government should induce further in Mozambique through nongovernmental organizations, private sector demand for Mozambique’s factors of produc- producer associations, private sector firms, and government tion, diversifying the economy and achieving shared and agencies. But more needs to be done to get the benefits of sustained economic growth. economies of scale. A more balanced growth path, that would rely on local business initiatives, is fundamentally about creating the right incentives and necessary underlying conditions to CONCLUDING REMARKS foster entrepreneurial drive. These conditions can be The government of Mozambique has consistently pursued achieved through policies to improve the business environ- sound fiscal, monetary, and external policies, demonstrating ment, the functioning of factor markets and land alloca- a clear understanding of the benefits of a solid underlying tion, and the trade infrastructure, aiming to strengthen or macroeconomic framework for fostering further private create the institutions necessary for more absorption of sector-led growth. Mozambique’s stable macroeconomic labor, in particular helping small and medium enterprises environment has allowed the economy to undergo signifi- to succeed. Such a shift requires political commitment and cant transformation over the past two decades, in part sup- attention to new institutional bottlenecks and implementa- ported by a stable source of ODA funding. Sectoral output tion capacity. composition has witnessed significant changes, with the The bet that Mozambique has taken is showing some participation of agriculture in output diminishing as more positive results. It was perhaps, given its chronic aid- productive sectors grow. Although employment is still over- dependency during most of its post-independence history, whelmingly concentrated in agriculture, job growth in the the only possible way to depart from it. This bet is not private sector (mostly informal) has led to increased partici- exempt from risks, and there are several indications that the pation of private services in output. Nevertheless, Mozam- political leadership is fully aware of many of them. Losing bique’s current growth model is not creating enough jobs or too soon significant portions of ODA would certainly be widening the productive base in order to create an economy harmful to both macroeconomic stability and the quality of in which wealth is broadly shared. However, the strategy policies, although it should be acknowledged that the implemented since the discovery and exploitation of large country has fully owned and implemented the standard CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 77 agenda aimed at stability. The benefits from the export such as Sulename (2001); Benito-Spinetto and Moll windfall could be captured by vested interests and not re- (2005); IMF (2005); Tahari, Akitoby, and Aka (2004); and invested into the rest of the economy. There could be a sub- Ndulu and O’Connell (2003). Vitek (2010) provides a stitution of aid-dependency by an excessive reliance on the growth accounting exercise in Mozambique between 1990 business interests of large private groups. and 2008. Nevertheless, it is overall quite a success story, despite 4. The shrinking of the labor force in agriculture in favor these risks and perhaps because of these risks. To the extent of other more productive sectors is a general pattern in the development process of many other advancing economies that the country understands these challenges and works to and should not be viewed as a negative phenomenon. address them, especially in the efficient allocation of wind- 5. Job creation in the nonagricultural informal sector falls to development purposes, chances are that Mozambique between 1997 and 2003 was also an important determinant could be a bright spot in terms of a successful strategy that of poverty reduction (World Bank 2008b). It provided eco- overcame a civil war, instabilities of all sorts, lack of initial nomic opportunities to the poor and led to economywide resources, lack of human capital, poverty and excessive productivity gains. dependency on external assistance. 6. The enterprise survey of 600 firms was carried out in 2008 by the World Bank for the Investment Climate NOTES Assessment (World Bank 2010b). Each firm in the sample was visited to collect employment information on 2003 1. In primary education (grades 1–7), the number of and 2006. children more than tripled, from about 1.3 million in 1992 to 4.2 million in 2008. Net enrollment rates in pri- 7. The index essentially compares each product Mozam- mary education doubled, from 45 percent in 1998 to bique exports with all importing countries of the product more than 96 percent in 2008, with rates for girls rising from all world sources. By tracing all bilateral flows for from 40 percent to 93 percent. The number of primary each exported good, Gillson (2008) proposes a measure that school teachers increased from 30,000 in 1992 to 73,900 captures the extent to which markets have been penetrated in 2008. The gross enrollment rate in lower secondary by Mozambican products. For instance, if country A exports education (ES1) increased from 4.8 percent in 1998 to 28.0 good X only to country B, but good X is also imported by percent in 2008. The gross enrollment rate for upper sec- countries C and D only from country E, then country A is ondary school (ES2) increased from 1.3 percent in 2008 to exploiting one of three (or one-third of) bilateral flows for 8.0 percent in 2008. Under-five mortality rates decreased good X. from 212 per 1,000 live births in 1996 to 178 in 2003 and 8. Market penetration measures disaggregated by largest 138 in 2008. Infant mortality decreased from 145 per 1,000 export partners indicate that Mozambique is already live births in 1996 to 122 in 2003 and 93 in 2008. Maternal exploiting 80 percent of its potential bilateral flows to mortality fell from an estimated 1,000 per 100,000 live South Africa. births in the early 1990s to 408 in 2003 and 340 in 2007. 9. This section draws heavily on the recent Country Eco- The share of the population with access to an improved nomic Memorandum on “Reshaping Growth and Creating water source increased from 39 percent in 1995 to 48 per- Jobs through Trade and Regional Integration” (World Bank cent in 2008. Although the rate of HIV prevalence 2010a). remains very high, the capacity of the health system was 10. These results should be treated with caution, however. expanded to start providing free antiretroviral treatment Substantial underreporting of food consumption has been for HIV infection. identified in the 2008/09 household survey data, which 2. The methodology applied to obtain the decomposition could affect both the poverty levels and their trend over was based on a Cobb-Douglas production function with time. The data problem also affects the 1996/97 and 2002/03 capital and quality-adjusted labor as inputs. Shares of the household surveys, albeit to a different extent. Additional inputs were considered fixed at 0.4 and 0.6. Data from the analytical work is required to test the robustness of the Instituto National de Estatistica (INE 1998, 2004) were used poverty analysis in Mozambique by correcting for the data to adjust for labor quality. problems. 3. Jones (2006) provides a thorough discussion of growth 11. The considerable amounts of aid and the presence of a accounting in Mozambique between 1980 and 2004. He great number of aid agencies led to the establishment of considers alternative and more flexible functional forms donor coordination mechanisms that were considered best for the production function; discusses perils of the growth international practices until recently. A number of donor- accounting exercises with specific Mozambican data con- government working groups were established during a straints in mind; and reviews results from other authors, “golden era” for coordination that brought benefits to both 78 CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 donors and the government—in terms of transaction ———. 2004. Relatório final do inquérito aos agregados famil- costs—and culminated with the adoption of the Paris and iares sobre orçamento familiar (IAF), 2002/3. Maputo. Accra Declarations. In parallel, a distinction began ———. 2006. National Labor Force Survey 2004/5. Maputo. between donors providing direct budget support (such as ———. 2010. Relatório final do inquérito sobre orçamento a regrouping into a group of nineteen OECD countries or familiar (IOF), 2008/9. Maputo. the G-19) and those involved in more traditional project Jones, Sam. 2006. Growth Accounting for Mozambique, activities. While important donors (such as the United 1980–2004. Discussion Paper 22E, Ministry of Planning States) concentrated efforts in specific sectors (like health and Development, National Directorate of Studies and and clean water), the G-19 elaborated a memorandum of Policy Analysis, Maputo. understanding to define quasi-contractual rules for dis- bursing budget support, considered as a sign of a more MPD (Ministry of Planning and Development), Govern- mature relationship with recipient governments. In fact, ment of Mozambique. 2004. Second National Poverty while Mozambique’s success led the country to benefit from Assessment. Maputo. larger shares of aid under direct budget support (by the ———. 2010. Third National Poverty Assessment. Maputo. G-19), this theoretically more “advanced” form of aid, ade- Ndulu, B., and S. O’Connell. 2003. Revised Collins/ quate for countries with a solid track record, also makes the Bosworth Growth Accounting Decompositions. AERC country more dependent on the political and budgetary Growth Project, African Economic Research Consor- cycles of OECD donors (see box 3.1). tium, Nairobi. Sonne-Schmidt, C., C. Arndt, and M. Magaua. 2008. Contri- bution of Mega-Projects on GDP. Ministry of Planning REFERENCES and Development, Direcção Nacional de Estudos e AfDB (African Development Bank). 2009. African Develop- Análise de Políticas (DNEAP), Maputo. ment Indicators 2009. Tunis. Sulename, Jose. 2001. “Economic Decline: A Study Benito-Spinetto, M., and P. Moll. 2005. “Macroeconomic with Reference to Mozambique.” Ph.D. diss., Department Developments, Economic Growth and Consequences for of Economics, University of Notre Dame, South Bend, IN. Poverty.” Background Paper for the 2005 Country Eco- Tahari A., B. Akitoby, and E. Aka. 2004. “Sources of Growth nomic Memorandum. World Bank and International in Sub-Saharan Africa.” IMF Working Paper 04/176, Monetary Fund, Washington, DC. International Monetary Fund, Washington, DC. Castel-Branco, Carlos Nuno. 2002. “Mega projectos e UN COMTRADE (various years). United Nations Com- estratégia de desenvolvimento.” http://www.iese.ac.mz/ modity Trade Statistics Database. http://comtrade lib/cncb/Mega_projectos_Moz_texto.pdf. .un.org/db/ Castel-Branco, Carlos Nuno, and Rogéio Ossemane. 2009. Vitek, Francis. 2010. “Economic Growth in Mozambique: “Crises cíclicas e dasfios de transformação do padrão de Experience and Policy Challenges.” International Mone- crescimento económico em Moçambique.” In Desafios tary Fund, Africa Department, Washington, DC. para Moçambique 2010, ed. L. Brito, C. Castel-Branco, World Bank. 2008a. Expanding the Possible in Sub-Saharan S. Chichava, and A. Francisco, pp. 141–82. Maputo: Insti- Africa: How Tertiary Institutions Can Increase Growth and tuto de Estudos Sociais e Económicos (IESE). Competitiveness. Poverty Reduction and Economic Clement, Jean A. P., and Shanaka J. Peiris, eds. 2008. Post- Management (PREM), Washington, DC. Stabilization Economics in Sub-Saharan Africa. Washing- ———. 2008b. Mozambique, Beating the Odds: Sustaining ton, DC: International Monetary Fund. Inclusion in a Growing Economy. A Mozambique Poverty, Gillson, Ian. 2008. “Mozambique’s Export Market Penetration.” Gender, and Social Assessment. Poverty Reduction and World Bank, Poverty Reduction and Economic Manage- Economic Management (PREM), Washington, DC. ment (PREM), Africa Region, Washington, DC. ———. 2010a. Mozambique Country Economic Memoran- IMF (International Monetary Fund). 2005. Republic of dum: Reshaping Growth and Creating Jobs through Trade Mozambique: Selected Issues and Statistical Appendix. and Regional Integration. Washington, DC. IMF Country Report 05/311, Washington, DC. ———. 2010b. Mozambique Investment Climate Assessment INE (Instituto National de Estatistica). 1998. Relatório final 2010. Washington, DC. do inquérito sobre orçamento familiar (IOF), 1996/7. ———. Various years. World Development Indicators. Maputo. Washington, DC. CHAPTER 3: RAPID GROWTH AND ECONOMIC TRANSFORMATION IN MOZAMBIQUE, 1993–2009 79 CHAPTER 4 Botswana’s Success: Good Governance, Good Policies, and Good Luck Michael Lewin ver the past 60 years, Botswana’s economy has O Annual growth in per capita income averaged 7.0 percent been one of the most successful in the world. The between 1966 and 1999 (table 4.1 and figure 4.1). The coun- country’s achievement is remarkable, because at try’s performance is particularly impressive compared with independence, in 1966, its prospects were not encouraging. that of other African economies (figure 4.2). Critics have argued that the social gains from this growth have been somewhat limited. In fact, in addition to ECONOMIC AND SOCIAL INDICATORS the gains in health and life expectancy noted above, there Botswana is a sparsely populated, arid, landlocked country; have been gains in poverty reduction and education. The at independence it was also one of the poorest countries in proportion of poor people fell from about 50 percent in the world, with per capita income of just $70 a year. In the 1985 to 33 percent in 1994, and the proportion of people first few years of independence, about 60 percent of current completing at least primary school rose from less than government expenditure consisted of international devel- 2 percent at independence to about 35 percent in 1994 opment assistance. There were only 12 kilometers of paved (Leith 2005). roads, and agriculture (mostly cattle farming for beef pro- Not all indicators are as positive: income distribution in duction) accounted for 40 percent of gross domestic Botswana remains very unequal (the Gini coefficient was product (GDP). about 0.55 in 1994).2 Unemployment remains high, reflect- By 2007 Botswana had 7,000 kilometers of paved roads, ing to a large extent rural to urban migration, although it and per capita income had risen to about $6,100 ($12,000 too has fallen, dropping from about 21 percent in the 1990s at purchasing power parity), making Botswana an upper- to about 17 percent in 2008. middle-income country comparable to Chile or Argentina. Its success is also evident in other measures of human DIAMONDS AND DEVELOPMENT development. At independence, life expectancy at birth was 37 years. By 1990 it was 60, 10 years above the African aver- Botswana’s extraordinary growth was fueled by minerals, age. Under-five mortality had fallen to about 45 per 1,000 particularly diamonds. At independence, beef, the country’s live births in 1990, compared with 180 for Africa as a whole. main export and largest sector, contributed 39 percent of Development assistance has shrunk to less than 3 percent of GDP. From independence until the 1970s, international aid the government budget, and agriculture currently accounts dominated the government budget and was the main source for only about 2.5 percent of GDP. Major strides have also of foreign exchange. At that time the mineral sector, mainly been made in infrastructure and education.1 diamonds, began to take off and soon became the dominant 81 Figure 4.2 Average per Capita Income in Africa and Table 4.1 Annual Growth in per Capita Income in Botswana, 2001–09 Selected Economies, 1966–99 (purchasing power parity at current Economy Average growth rate dollar prices) Botswana 7.0 Chile 2.1 16,000 Hong Kong SAR, China 4.6 Indonesia 3.8 14,000 Ireland 4.1 12,000 Korea, Rep. of 6.1 Singapore 6.2 10,000 Dollars Thailand 4.6 8,000 Source: Adapted from Leith 2005. 6,000 4,000 Figure 4.1 Average Annual Growth in per Capita Income 2,000 in Botswana, 1980–2008 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 8 Africa Botswana 7 Source: OECD n.d. 6 5 Percent 4 “good economics, bad politics.” In almost all developing 3 countries, the government owns the mineral resources and 2 is therefore the main recipient of the revenues from their extraction. This concentration of revenues with the govern- 1 ment as the conduit of benefits to the rest of the economy 0 1980–90 1990–2000 2000–08 can lead to a host of problems, including rent-seeking, cor- ruption, and the efficiency losses that result from them. It is Source: Leith 2005. not surprising that, with easily accessible wealth concen- trated in the hands of the government, malignant dictator- sector. Income growth and the growth of the mining sector ships and predatory regimes often thrive. There also seems accelerated in tandem from about 1974/75 until recently. to be a greater tendency for armed civil conflict in mineral- Luck—the discovery of diamonds as well as other based economies (Collier and Hoeffler 2004). These “bad important minerals—was clearly an element of Botswana’s politics” are extremely costly. success. However, there is considerable disagreement over Botswana has avoided these manmade disasters. Part of whether the discovery of minerals in a developing economy its success may reflect luck: Botswana’s relatively homoge- generally brings good luck or bad (box 4.1).3 Indeed, many neous population has less potential for ethnic polariza- countries in Africa, including Zambia, Nigeria, the Demo- tion, which when combined with mineral rents can be cratic Republic of Congo, Sierra Leone, and others, have particularly combustible. Still, most of the credit must be squandered vast amounts of their natural wealth. One does given to the leadership, which, since independence, has not have to be persuaded that natural resources are neces- designed and fostered the conditions of governance that sarily a curse to conclude that they certainly are not sufficient have ensured stability and social and economic progress. for economic success or even a good predictor of it. In the The government established respect for property rights case of Botswana, minerals turned out to be lucky, but other and the rule of law. It maintained a high degree of trans- key ingredients in the recipe for success were present, includ- parency, which was reinforced by continuing the Tswana ing good governance and good economic management. tribal tradition of consultation. These consultative insti- tutions, known as kgotla, created a degree of trust in the government—the sense that government exists to serve Avoiding the bad governance curse the people and promote development and is not the Mineral-based countries seem to be prone to bad govern- instrument of one group or individuals for the purpose of ment, a phenomenon Acemoglu and others have termed getting hold of the wealth. Tswana tradition also 82 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK Box 4.1 The Resource Curse The counterintuitive notion that the endowment of nat- can be financed by the revenue from the mineral ural resources is a curse rather than a blessing has exports. become received wisdom. Sachs and Warner (1995) Because the entire process is initiated by an increase report regression results showing that being a natural in income, expenditure, and consumption, it is not clear resource or mineral exporter reduces a country’s devel- why it should be thought of as a disease. One reason opment prospects. Many other researchers find evidence given is the presence of positive externalities in the to support this claim (see for example, Auty 2004 and declining sector. The declining traded goods sector is the references therein). The fact that most of the East thought to have spillover effects that contribute to long- Asian tigers (the Republic of Korea; Taiwan, China; run growth and industrialization. Hence, the real Hong Kong SAR, China; and Singapore) plus Japan are appreciation caused by mineral exports is said to cause resource poor has added to the perception that endowed “deindustrialization,” contributing to poor long-run wealth is an obstacle to growth and development. performance. Of course, in Africa and Botswana in There is certainly evidence that a disproportionate particular, the declining traded goods sector was agri- number of the poorest countries are dependent on culture, not industry (there was no significant industry mineral exports. The direction of causality, however, is to deindustrialize). Some might argue that mineral far from clear: does mineral dependence retard devel- exports prevent the real depreciation needed to indus- opment, or does retarded development create mineral trialize. Additionally, it is argued, the mineral sector dependence? fails to stimulate the nonmineral sector through link- Accounting for the endogeneity of mineral depen- ages to the rest of the economy. dence overturns the Sachs-Warner conclusions. Coun- Against this effect, one must weigh the advantages tries that fail to develop remain resource dependent, but that minerals bring in terms of increased saving and resource abundance may be correlated with higher hence investment, which should offset any loss of com- growth (see, for example, Brunnschweiler and Bulte petitiveness caused by the real exchange rate. Moreover, 2008; Alexeev and Conrad 2009). Many rich countries— because government is usually the chief beneficiary of including Canada, Australia, Norway, the United States, the resource revenues, the increase usually eases con- and Latin America’s top performer, Chile—were or are straints on public goods, such as infrastructure, educa- mineral dependent. In Africa the two richest countries, tion, and health, which should contribute to growth South Africa and Botswana, owe much of their wealth and industrialization. If private investment or public to gold and diamonds. goods expenditure fails to materialize, the problem lies In short, resources are not necessarily a curse, but not in the real exchange rate but rather in the failure of many countries have squandered their resources, and in financial markets to allocate the saving and investment some cases, the presence of mineral resources may have productively or of government to invest wisely. contributed to economic stagnation or decline. What is The Volatility Curse it about mineral endowments that can make things go wrong? Mineral prices tend to be volatile, which, it is argued, destabilizes the economy (Hausmann and Rigobon Dutch Disease 2003). This volatility affects the economy through the Dutch disease refers to the deleterious effects that pur- exchange rate, which appreciates in the boom and portedly result from the real appreciation of the cur- depreciates in the bust, and through fiscal policy, which rency caused by a booming resource export sector tends to be procyclical, leading to inefficient “stop-go” (Corden and Neary 1982). The discovery of minerals provision of government services and infrastructure or an increase in their international price leads to an projects.a It is argued that the perceived riskiness that increase in income and expenditure. As long as some of this volatility engenders leads to lower investment, par- the increased expenditure goes to domestically pro- ticularly foreign direct investment, which seeks out duced goods, there will be a real appreciation of the safer markets. In addition, governments tend not only currency, which causes productive resources to move to increase spending when current revenues rise but often from nonmineral traded goods, the output of which also to believe that revenues will continue to rise; they declines, to nontraded goods, the output of which therefore borrow against future revenues. Instead of increases. The resulting excess demand for tradables being able to use accumulated assets to help in the bust, (continued next page) CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK 83 Box 4.1 (continued) economies find themselves contending with debt Exhaustibility overhang. Minerals are a windfall, even in the long run, in the It is possible to insulate the domestic market from sense that they are finite and that when they are the volatility of the commodity markets. Because the depleted an economic slump will occur. That a blessing government is the conduit of the cycle to the economy, does not last forever hardly makes it a curse, however. If it is also the solution. The key is to delink expenditure at least some of the windfall is saved, in the form of and revenue, essentially by saving all or part of the rev- physical and human capital or financial assets, future enues, thereby replacing a volatile income stream with a generations will benefit from it. Moreover, the deple- more stable one. Expenditure can then be based on this tion usually does not occur suddenly. With some fore- more stable, permanent income. In this way, consump- sight, the adjustment and transition can be cushioned tion can be smoothed over time, in some cases over by accumulated saving.b generations. Many countries, including Botswana, Chile, Kuwait, Norway, Qatar, and Saudi Arabia, have How Has Botswana Fared? successfully applied such an expenditure-smoothing To a large extent, Botswana overcame the threat of the policy. resource curse: The Governance Curse 1. Dutch disease was avoided, most significantly by In most developing countries the government claims government investment in public goods and infra- ownership of all minerals. If government institutions structure. The government also took measures to are weak to begin with, there is a strong likelihood that help boost productivity, by limiting parastatals and the resources will be squandered. In addition, the pres- avoiding import substitution policies. ence of the resources may weaken and corrupt govern- 2. The volatility curse was overcome by unlinking pub- ment. The relative ease with which mineral revenues are lic expenditure from revenue. The government collected can lead to a lack of transparency and established savings funds and avoided typical pro- accountability. The resulting interdependence of gov- cyclical behavior and real exchange rate volatility. ernment and the mining industry reinforces this ten- 3. The governance curse was avoided, through a series dency. Government itself is seen as the means to of actions described in the chapter. Botswana’s suc- acquiring wealth, which typically leads to rent seeking cess in this regard is exceptional. and corruption. There is also evidence that in many 4. The question of whether the exhaustibility chal- cases mineral wealth provokes or fuels internal conflicts lenge has been met remains unclear. If current pre- (Collier and Hoeffler 2004). dictions are correct, diamond revenues will begin In an ironic metaphor, underground minerals are declining in 2016 and could be exhausted by 2029. often referred to as one of the “commanding heights” of Botswana has adopted many of the appropriate the economy. Postindependence governments claimed policies to prepare for the depletion of its mineral the need to control these commanding heights for the base, by accumulating funds for the future, building benefit of all the governed. One of the unforeseen con- infrastructure, and investing in health and educa- sequences of this policy was the perception that control tion. However, the growth of the nonmineral traded of government is the path to wealth, which in turn, goods sector, which has remained flat at about unsurprisingly, has often led to regional and ethnic con- 5 percent of GDP, will need to accelerate to avoid flict for control of these resource rents (Collier and balance of payments problems as diamond revenues Hoeffler 2004). begin to decline. a. Rather than being procyclical, in most developing countries, the government initiates the cycle because it is the recipient of the rents. The boom therefore begins with government expansion and the bust with its contraction. Hence, real exchange rate vari- ability is also the result of volatile government expenditure. b. A full discussion of mineral resource management in the context of intertemporal and intergenerational allocations goes beyond the objectives of this chapter. Essentially, the optimal management of resource rents requires expenditure smoothing not only to avoid cyclical volatility but also to spread the benefits of the windfall over time. Like a rational household, an opti- mizing government will attempt to base its long-run expenditure on some calculation or estimate of its “permanent” or stable revenue rather than on its current or transitory revenue. For some calculations of permanent income in the case of Botswana, see Basdevant (2008). 84 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK respected private property; the fact that many of the tribal economies seem to spawn the predatory type of dictatorship leaders who helped usher in modern government were rather than the relatively benevolent ones of East Asia’s past. also large cattle owners may have reinforced this respect. Although this notion is largely speculative, it seems that a Acemoglu and Robinson (1999) emphasize that property democratic government dependent on mineral wealth is rights and the rule of law are the key factors explaining more likely than a dictatorial one to be responsive to devel- development success. They also emphasize the importance opment needs, to settle disputes peacefully, and to respect of the preindependence colonial regime in determining the the rule of law. Very few, if any, mineral-rich countries in adoption of these institutions after independence. More Africa have been ruled as peacefully and productively as heavily settled colonies tended to establish institutions to Botswana; it is hard to escape the conclusion that the dem- maintain the rule of law and enforce property rights; in ocratic institutions established at independence are an sparsely settled colonies, the colonial regimes tended to be important part of the explanation. more “extractive” or “exploitative,” unconstrained by law Social science cannot rigorously assess the relative and respect for property. Although Botswana’s colonial past importance or contribution of leadership in the evolution is not typical of the kind of regime that usually led to a of successful institutions. However, in the case of Botswana, postindependence government constrained by law and leadership, particularly that of its first president, Seretse property rights, its unique colonial history had the same Khama, may have been crucial, especially in the areas of effect.4 Before independence, Botswana was a British pro- mineral exploitation and the rights of the state versus those tectorate, the colonial power having been “invited” in. of the tribes. The discovery of minerals can easily lead to Because Botswana was not colonized for economic or civil war and the dissolution of the state. To prevent this strategic advantage, the colonial rulers, it is argued, did not from happening, even before independence, Khama’s party, impose the extractive-type regime often found in other the Botswana Democratic Party (BDP), wrote into its plat- sparsely populated areas. Thus the regime that evolved after form its intention to assert the state’s rights to all mineral independence was one that respected the law and property resources. After independence, the government reached and was dedicated to development. agreement on ownership of mineral resources with the More controversial is the role of democracy. Botswana tribal authorities. Although the largest diamond deposits has maintained a parliamentary democracy since indepen- were discovered in Khama’s own district of Bamangwato, he dence. To be sure, its democracy is not perfect: the country chose the country over his tribal land, thus helping limit the has had one-party rule since independence, women’s repre- possibility of conflict. sentation is limited, and there has been some criticism that minorities (particularly the San people) are not treated Implementing good policies equally. Nevertheless, the government functions in a demo- cratic manner, elections are “free and fair,” and the govern- All mineral-based economies face the issue of an appreciat- ment is responsive to the electorate and transparent in its ing real exchange rate (box 4.2). The well-known possible dealings. harmful consequences of this are often referred to as Dutch Is democracy a positive factor for economic growth and disease (see box 4.1). Prevention of Dutch disease in development? None of the miracle East Asian economies or Botswana consisted of three key components: fiscal saving, a Latin America’s high performer (Chile) was democratic surplus on the current account of the balance of payments, during the first years of rapid growth and development, and heavy government investment in infrastructure and leading many commentators to conclude that property human capital. Together these policies limited the erosion of rights and the rule of law are more important than democ- domestic productivity and competitiveness that can result racy (Acemoglu, Johnson, and Robinson 2003). For some from the appreciation of the real exchange rate. High fiscal observers, democracy is actually a hindrance to economic saving limits current consumption, reducing pressure on development. Although this conclusion is probably unwar- domestic price inflation, a typical problem in natural ranted, the fact that many democracies (such as India) have resource booms. Some of the government saving took the not fared well economically undoubtedly means that form of offshore investments, which directly limited real democracy is neither sufficient nor necessary for growth exchange rate depreciation and diversified the sources of and development.5 future foreign exchange revenues. The accumulation of In the case of mineral-rich countries like Botswana, reserves is a form of self-insurance against short-run democracy may be an important catalyst. Mineral-dependent declines in mineral revenues as well as long-run reductions CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK 85 Box 4.2 Windfall Economics 101 The mechanism by which mineral resource flows affect money into the economy, however, resulting in domestic the real exchange rate depends on the nominal price inflation, which will eventually erode local compet- exchange rate regime maintained by the country. itiveness until the balance of payments surplus is elimi- Botswana has a fixed exchange rate system. Initially, it nated. This is the effect of real appreciation when the pegged its currency to the U.S. dollar. When the South nominal rate is fixed. To prevent this from happening, African rand depreciated relative to the dollar, the government has to sterilize the effects of the resource Botswana suffered from the appreciation of the pula rents on the domestic money supply. In the medium to (Botswana’s currency) relative to the rand (the currency long run, the fiscal authority can do so by absorbing the of its main trading partner).a The pula is therefore now excess domestic money through fiscal budget surpluses. pegged to the rand. This policy of twin surpluses—a balance of payments In a fixed-rate regime, the first effect of a resource surplus in tandem with a fiscal budget surplus—is essen- windfall is the creation of an excess supply of foreign tially the policy that Botswana has followed. (Some ana- exchange. To maintain the peg, the monetary authority lysts have argued that China has followed a similar policy must accumulate reserves. Doing so will inject domestic to maintain an undervalued currency.) a. In contrast, Argentina maintained a dollar peg after the devaluation of the Brazilian real. Many analysts consider this a key ele- ment leading to the Argentine crisis. in these revenues as a result of mineral depletion. The instru- seems to have served Botswana well. As noted above, the ments of public saving took the form of the Public Service number of paved roads increased from around 20 kilome- Debt Management Fund and the Revenue Stabilization ters in 1970 to 2,300 in 1990; 90 percent of the population Fund. These funds have proved to be successful mechanisms had access to safe water (compared with 29 percent in for managing fiscal saving; income from the Revenue 1970); and the number of telephone connections rose from Stabilization Fund is now a stable source of government around 5,000 in 1970 to 136,000 in 2001.8 Similarly, in edu- revenue. cational achievement Botswana leads Africa. For example, Another part of the fiscal saving was channeled to adult literacy, male and female, is around 80 percent domestic assets, combating the effect of the loss of compet- compared with 69 percent and 50 percent, respectively, for itiveness by raising productivity. When this investment is the rest of sub-Saharan Africa.9 focused on public goods (for example, infrastructure, As a result of its prudent fiscal policy, Botswana has health, and human capital), it will contribute to growth enjoyed relative macroeconomic stability and avoided the without crowding out private sector investment and devel- boom-slump cycles that characterize many mineral-based opment. Public sector saving was positive in every year economies. Monetary policy was also restrained: for most of between 1975 and 1996, fluctuating between 10 and 40 per- the postindependence period, inflation was moderate, aver- cent of GDP. Public sector investment was fairly constant at aging about 10 percent (Maipose n.d.). The periodic slow- about 10 percent of GDP.6 However, if one counts expendi- downs in the diamond industry have thus by and large not ture on health and human capital as investment, government been passed on to the rest of the economy. By withholding investment has consistently remained about 20 percent of some of the benefits from the economy during the booms, GDP. The capital budget focused on basic infrastructure the government has, to some extent, been able to insulate it with about 30 percent devoted to water, electricity, roads, from the busts.10 communication, and transportation. Twenty percent of the capital budget, on average, went to education and around Shunning import substitution and parastatals 30 percent of recurrent expenditures was devoted to educa- tion and health.7 While quantifying the effectiveness of Two things Botswana did not do are also significant. Unlike public expenditure on growth and development is notori- many African countries, it did not adopt a policy of import ously difficult, particularly in resource-based economies substitution, and it did not expand the extent of state- where soft budget constraints may lead to overinvestment, owned productive entities, which employ only about 5 per- the emphasis on infrastructure, health, and education cent of the workforce in Botswana (figure 4.3). 86 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK Figure 4.3 Percentage of GDP Accounted for by Botswana’s combination of policy and governance evidently State-Owned Enterprises in Selected helped it avoid the worst effects of the resource curse. African Countries, Early 1990s 30 PLANNING FOR A FUTURE WITHOUT MINERALS 25 Botswana is still a mineral-based economy: it has not suc- Percentage of GDP 20 ceeded in significantly diversifying its economy away from diamonds. Government remains the largest employer 15 (employing 30 percent of the active workforce); other than 10 minerals (which employ a relatively small share of workers), most production is in nontraded goods. This pattern is typ- 5 ical of mineral-dependent economies. Real GDP and mining income have moved in tandem since 2000 (figure 4.4), and 0 mining consistently accounts for about 40 percent of GDP, a n na a ia go a e an ny bi bw oo an completely dwarfing the contributions of manufacturing ha To m w Ke nz er ba G Za ts am Ta m Bo and agriculture (figure 4.5).11 Minerals, mostly diamonds, Zi C Source: Leith 2005. Figure 4.4 Mining Revenues and Real GDP in Botswana, 2000–09 Being part of the South African Customs Union (which 30,000 required cooperation with the then odious regime in South Pula (millions) at constant 25,000 Africa) meant maintaining a fairly low tariff regime and 1993/94 prices 20,000 provided a steady stream of government revenue. Avoiding an activist import-substituting policy and maintaining lim- 15,000 ited government involvement in production seems to have 10,000 paid off for Botswana by allowing it to avoid many of the 5,000 inefficiencies and structural deficits that so often arise from 0 such policies. 00 01 02 03 04 05 06 07 08 09 Trade policy is also an area where good governance and 20 20 20 20 20 20 20 20 20 20 good policies reinforce one another. A government rich with Mining Real GDP mineral revenues is an inviting target for rent seekers and worse; restricting the avenues for rent seeking and corrup- Source: Bank of Botswana 2010. tion thus helps preserve the efficiency and integrity of the government. Even if the theoretical merits of import substi- Figure 4.5 Contributions of Mining, Manufacturing, and tution or the existence of state-owned enterprises seem per- Agriculture to Botswana’s GDP, 2000–09 suasive, in practice both often result in inefficiency and 45 drain fiscal resources. 40 Good fiscal policies by themselves may not be sufficient 35 for success. Many mineral-based economies with high rates 30 of investment have not enjoyed the positive results that Percent 25 Botswana has. The quality of investment is evidently as 20 important as the quantity. Moreover, in the hands of the 15 venal and corrupt, government savings funds can easily turn 10 into slush funds for the favored elites. There may be a lesson 5 in this from Botswana: good governance will aid the effec- 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 tiveness of good policy, and good policy encourages better Agriculture Mining Manufacturing government. Policy formulators should therefore not ignore the political economy consequences of economic policy. Source: Bank of Botswana 2010. CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK 87 comprise about 85 percent of exports, and mineral revenues ships this may cause, will be necessary to create the compet- account for about 50 percent of government revenues (Bank itiveness needed to replace falling mineral exports. of Botswana 2008).12 In 2009 manufacturing contributed less than 5 percent The diamond industry is expected to decline in the near of GDP (see figure 4.5). This share will have to rise if future, with revenues projected to begin falling in 2016 and Botswana is to avoid decline. The fact that manufacturing to be depleted by 2029 (Basdevant 2008). If these projections has been flat as a proportion of income conceals its growth, are correct, adjustment is inevitable and unlikely to be pain- because just keeping pace with Botswana’s growing econ- less, although Botswana’s prudent policies—particularly omy has required rapid growth.13 However, manufacturing regarding saving and investment—have left it in a relatively growth will have to accelerate if Botswana is to raise or even strong position to facilitate a soft landing. For most years, maintain its standard of living in the long run. domestic saving has been above 40 percent and investment Key to such growth will be Botswana’s ability to attract about 35 percent of GDP. (The difference is the surplus on foreign direct investment (FDI). Although FDI fell between the current account of the balance of payments.) This 2002 and 2007 (Bank of Botswana 2010), Botswana’s repu- implies that from the perspective of the economy as a whole, tation as a stable country and successful economy should all revenue from minerals is being saved rather than con- help it attract FDI in the future.14 Policy makers should sumed, because total saving from national income is roughly examine the country’s overall investment environment with equal to mineral income. When mineral revenues begin to a view to attracting desirable investment.15 decline, some of the adjustment could thus be absorbed by Clouding Botswana’s smooth transition to a more diver- saving rather than consumption. Thus, even if growth slows, sified economy is HIV/AIDS, which hit Botswana harder consumption growth can be maintained as the proportion of than any country in Africa (with the possible exception of consumption to income rises. South Africa). The epidemic has reversed many of the The government has also been a big saver. Central bank impressive gains in health indicators achieved over the past profits and other income from accumulated assets now 100 years and reduced productivity. The number of deaths make up about 30 percent of government revenue, most of from AIDS began declining in about 2003 (World Bank it income from the accumulated saving of mineral revenues. 2009), but the costs and consequences of the epidemic will At times, the overall fiscal surplus rose to more than 30 per- persist for many years. cent of GDP (Maipose n.d.), and public sector investment (as noted) has been consistently about 10 percent of GDP. If CONCLUSION investment is defined more broadly to include human cap- ital and health, public investment rises to about 20 percent Landlocked Botswana seems to have defied the odds by of GDP, which meets the IMF benchmark for sustainability creating a successful economy. Poverty has been reduced, (Basdevant 2008). The government has used this accumu- education has become more widespread, and health indi- lated savings to smooth expenditure over the business cators had improved before the HIV/AIDS epidemic undid cycle, thus providing the economy with a short-run shock some of that progress. absorber. However, as argued in box 4.1, the accumulated The country’s vast natural resources played a key role in saving is also a long-run shock absorber, because it has this accomplishment, but the mere endowment of resources built up human and physical capital and financial assets is clearly not the whole story. In much of Africa—and in that should serve the economy well during the adjustment other parts of the world—natural resources have not always period. In addition, the accumulated foreign assets man- been conducive to growth and development; in many cases aged by the Bank of Botswana now amount to about they seem to have brought out the worst in countries, in the 40 months of imports. form of conflict and predatory governments. These advantages notwithstanding, Botswana faces seri- Studying the effect of mineral wealth on economic out- ous challenges. Dependence on mineral exports is the key comes is timely, because the increase in the prices of natu- weakness of the current economic outlook and diversifica- ral resources as a result of the rise of China and India is tion of the traded goods sector is the most important policy likely to result in windfalls for many African countries. objective. Accumulated foreign assets and the income from How can countries turn these windfalls into long-run them can ease the welfare cost of adjustment, but it cannot growth and development? In Botswana’s case, the key to eliminate it. Barring a miraculous jump in productivity, successfully harnessing natural resources lay in good gover- some real depreciation of the pula and the consequent hard- nance and good policies. Governance has not been perfect 88 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK in Botswana, but it has been good. Botswana has been was defense expenditure, which average about 11 percent of largely free of kleptocracy and civil conflict; it has main- total expenditure. While this is high relative to most coun- tained a transparent, law-abiding government; and it has tries Botswana arguably had good reason to do this. implemented good policies, including a hyper-prudent fis- 8. World Bank cited in Lange and Wright, 2002, 32. cal policy, which has done much to diversify foreign 9. World Bank 2009. Botswana, however, lags other upper exchange earnings and prevent the volatility that typifies middle income countries in these categories. many resource-based economies. Investments in human 10. Botswana has not been able to escape the effects of the and physical capital and vast improvements in infrastruc- current global crisis: GDP declined 4 percent in 2009, and ture have also raised Botswana’s productivity, which, for the first time in many years, Botswana had fiscal and together with its substantial financial reserve in the form of balance of payments deficits. However, the economy foreign assets, should help ease the transition to a more appears to have weathered the worst, with the Bank of Botswana projecting real growth of more than 3 percent for diversified economy. 2010 and 2011. 11. Mining’s share fell in 2009 because of the slump in NOTES world trade, not because of the declining importance of the 1. For discussions of Botswana’s success, see Maipose sector. (n.d.), Acemoglu and Johnson (2003), and Leith (2005). 12. Between 2005 and 2008, before the effects of the global 2. Leith (2005) notes that when measured accurately, tak- slump set in, exports grew by about 40 percent. The share ing into account social, health, and educational services pro- of diamonds fell from about 75 percent in 2005 to about vided by the government to the poor, this measure falls to 65 percent in 2008. Total mining remained more or less about 0.53. No measure of inequality is without serious ana- constant, however, at about 85 percent of total exports. The lytical problems; the Gini coefficient is probably the best share of diamonds fell because of rising copper and nickel available and most widely used index. The coefficient ranges exports (Botswana Central Statistical Office 2009). from 0 (perfect equality) to 1 (perfect inequality). The 13. Other areas of interest are services, including finan- higher the index, the more unequal the society. An index cial services; downstream diamond processing and trad- above 0.5 is thought to denote an unequal distribution ing; and tourism, which represented about 10 percent of (Leith 2005). GDP in 2008 and is a potential growth sector for export 3. See, for example, Auty (2004) and Sachs and Warner earnings. (1995). For a contrary view, see Stevens (2003) and 14. FDI rose again in 2008. It is too early to tell whether this Brunnschweiler and Bulte (2008). See also Lederman and is a trend. Maloney (2007). 15. Botswana had the best ranking in Africa on Trans- 4. The analysis is based on the idea that some colonial parency International’s corruption perception index in 2011 regimes were mostly “extractive”—that is, the regime and ranked 33rd worldwide. (See http://www.transparency existed to reap the maximum out of the colony’s economy. .org/policy_research/surveys_indices/cpi/2010.) On the Other regimes, usually ones in which there were more colo- World Bank’s 2010 Ease of Doing Business index, Botswana nial settlers, were less extractive. They often established ranked 3rd regionally but only 52nd worldwide, indicating institutions that put more constraints on the extractive and some room for improvement (see http://www.doingbusiness arbitrary powers of government. These constraining institu- .org/rankings). tions often carried over into independence. 5. All of the high-performing developing countries men- REFERENCES tioned have established strong democratic institutions. The question therefore arises whether the causality runs the Acemoglu, D., S. Johnson, and J. A. Robinson. 2003. “An other way, from growth to democracy. It can also be argued African Success: Botswana.” In In Search of Prosperity: that democracy is necessary to sustain a high-income econ- Analytical Narratives on Economic Growth, ed. D. Rodrik. omy (see Barro 1997). Princeton, NJ: Princeton University Press. 6. Isham and Kaufmann (1999) estimate this proportion Acemoglu, D., and J. A. Robinson. 1999. “On the Political (10 percent of GDP) to be the likely ceiling for public invest- Economy of Institutions and Development.” American ment to remain productive. Beyond this they find public Economic Review 91 (4): 938–63. sector investment is likely to be detrimental to growth. Alexeev, Michael, and Robert Conrad. 2009. “The Elusive 7. Lange and Wright (2002) point out that the only single Curse of Oil.” Review of Economics and Statistics 91 (3): “nonproductive” item of comparable weight in the budget 586–98. CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK 89 Auty, R. (ed.) 2001. “Resource Abundance and Economic Implications.” NBER Working Paper 9424, National Development.” Wider Studies in Development. Oxford Bureau of Economic Research, Cambridge, MA. University Press, Oxford, UK. Isham, J., and D. Kaufmann. 1999. “The Forgotten Rationale Bank of Botswana. 2008. Annual Report. Gabarone. of Policy Reform: The Productivity of Investment Proj- ——. 2010. Botswana Financial Statistics. Gaborone. http:// ects.” Quarterly Journal of Economics 116 (1). www.bankofbotswana.bw/index.php/content/2009 Lange, G., and M. Wright. 2002. “Sustainable Development 110615035-botswana-financial-statistics. in Mineral Economies: The Example of Botswana.” Barro, R. 1997. Determinants of Economic Growth: A Cross- Ceepa Discussion Paper Series, University of Pretoria, Country Empirical Study. Cambridge, MA: MIT Press. Pretoria, South Africa. Basdevant, O. 2008. “Are Diamonds Forever? Using the Per- Lederman, D., and W. F. Maloney, eds. 2007. Natural manent Income Hypothesis to Analyze Botswana’s Resources: Neither Curse nor Destiny. Washington, DC: Reliance on Diamond Revenue.” IMF Working Paper, World Bank. International Monetary Fund, Washington, DC. Leith, J. C. 2005. Why Botswana Prospered. Montreal: Botswana Central Statistical Office. 2009. External Monthly McGill-Queen’s University Press. Trade Digest January. Gaborone. http://www.cso.gov.bw/. Maipose, G. n.d. Policy and Institutional Dynamics in Sus- Brunnschweiler, C. N. and E. H. Bulte. 2008. “The Resource tained Growth in Botswana. University of Botswana, Curse Revisited and Revised: A Tale of Paradoxes and Gaborone. Red Herrings.” Journal of Environmental Economics and OECD (Organisation for Economic Co-operation and Devel- Management 55: 248–64. opment). n.d. Africa Economic Outlook. http:// www Collier, P., and A. Hoeffler. 2004. “Greed and Grievance in .africaneconomicoutlook.org/en/countries/southern- Civil War.” Oxford Economic Papers 56, Oxford Univer- africa/botswana/. sity, Oxford, UK. Sachs, Jeffrey D., and Andrew M. Warner. 1995. “Natural ———. 2005. “Democracy and Resource Rents.” Oxford Resource Abundance and Economic Growth.” NBER University, Oxford, UK. Working Paper W5392, National Bureau of Economic Corden, W. M., and P. J. Neary. 1982. “Booming Sector and Research, Cambridge, MA. Deindustrialization in a Small Open Economy.” Eco- Stevens, P. 2003. “Resource Impact: Curse or Blessing? A Lit- nomic Journal 92: 825–48. erature Survey.” Journal of Energy Literature 11 (1): 3–42. Hausmann, R., and R. Rigobon. 2003. “An Alternative Inter- World Bank. 2009. Country Partnership Strategy for the pretation of the ‘Resource Curse’: Theory and Policy Republic of Botswana. Washington, DC. 90 CHAPTER 4: BOTSWANA’S SUCCESS: GOOD GOVERNANCE, GOOD POLICIES, AND GOOD LUCK CHAPTER 5 Mauritius: An Economic Success Story Ali Zafar n 1961 James Meade, a Nobel Prize recipient in econom- $7,000. Imports and exports have boomed; together, they I ics, famously predicted a dismal future for Mauritius because of its vulnerabilities to weather and price shocks and lack of job opportunities outside the sugar sector. The reached more than 100 percent of GDP during the late 1990s and early 2000s. At the same time, efforts at eco- nomic diversification have been successful, allowing the small island nation in the Indian Ocean of approximately country to move from sugar to textiles to a broader service 1.3 million people has defied those predictions, trans- economy. Mauritius’s reliance on trade-led development forming itself from a poor sugar economy into a country has helped the country achieve respectable levels of export with one of the highest per capita incomes among African performance. Along the way, measures of human develop- countries.1 Mauritius’s combination of political stability, ment have improved substantially. strong institutional framework, low level of corruption, Despite being a small island economy vulnerable to and favorable regulatory environment has helped lay the exogenous shocks, Mauritius has been able to craft a strong foundation for economic growth, while its open trade growth-oriented developmental path. Natural disasters and policies have been key in sustaining growth. The govern- terms-of-trade shocks have never prevented the economy ment functions as a parliamentary democracy, and the from having strong and regular growth. Constrained at country has an efficient administration that is both tech- inception by a monocrop sugar economy, low amounts of nically competent and adaptive to changing global arable land, and a high rate of population growth, Mauritius economic circumstances. Mauritius’s financial sector is has emerged as a regional entrepôt and tourism destination sufficiently well developed that it has begun to position as well as the top-ranked African country in the World itself as a platform for investment linking East Africa with Bank’s Doing Business survey (in 2010, it ranked 20th of India and China. 183 countries). Headline figures related to Mauritius’s economic perfor- Although a variety of explanations have been advanced mance are impressive. Growth of real gross domestic prod- to explain Mauritius’s growth, the country’s focus on inter- uct (GDP) has averaged more than 5 percent a year since national trade has been, without doubt, a critical element of 1970, and annual growth in real per capita income has like- that performance.2 Mauritius’s preferential access to trading wise been strong. GDP per capita increased about sevenfold partners, particularly the European Union (EU), in the between 1976 and 2008, from less than $1,000 to roughly sugar, textile, and clothing sectors resulted in subsidies for Ali Zafar is a macroeconomist in the World Bank Group. 91 the export sector and provided important foreign exchange to adapt to shifting economic circumstances. Although for the economy. Preferential trading deals accounted for Mauritius benefited from the sugar protocols, the strong growth in Mauritius’s total exports between the government also recognized early on the advantages of 1970s and the 1990s. Although imports tariffs in Mauritius diversification. As a result, it relied heavily on EPZs but were kept high, they have never been high enough to inter- ensured that there was no anti-export bias. Mauritius has fere with the overall trade regime. For example, the average also proven very adept at embracing new sectors, particularly tariff for manufacturing was 86.2 percent in 1980, but it light manufacturing, offshore banking and financial services, fell to 30.1 percent in 1994 (Wignaraja and O’Neil 1999). and service-related information and communication tech- Simultaneously, Mauritius pursued a very liberal invest- nology (ICT). It has adapted and transformed its ethnic ment regime and used incentives to attract foreign direct pluralism into a tangible economic force.3 It has forged a investment (FDI). Mauritius also offset the burdens on its vibrant democracy with competing parties, a strong electoral exporters with tariff-free access for productive inputs, with system, and an open media.4 In sum, Mauritius’s impressive tax incentive subsidies and relaxed labor market regula- economic performance has not been an accident, but rather tions in the export sector. Furthermore, it has used export the result of careful planning and policies. During and in the processing zones (EPZs) to export key manufacturing aftermath of the biggest exogenous shocks to its economy in goods, mostly apparel and textiles. Mauritius’s overall trade recent times—the phasing out of the Multifibre Arrange- and investment policy has been based on a managed embrace ment governing textiles, significant reductions in EU sugar of globalization and cultivation of market access. protocol prices, the 2008 food and fuel crisis, and the Aside from its trade and investment policy, Mauritius has 2008–09 global financial crisis—Mauritius’s economy has benefited from prudent fiscal, exchange rate, and monetary displayed strong resilience. policy, the latter of which has also been beneficial to export performance. To compensate for the myriad disadvantages THE GROWTH STORY of limited scale, Mauritius has developed a plethora of strong institutions and good governance. Additionally, the Since the 1970s Mauritius has recorded very high growth public and private sectors maintain a vibrant partnership rates and sustained increases in human development that manifests itself in a range of areas. indicators through a combination of good macroeco- Finally, Mauritius has always displayed receptivity to nomic policies and strong institutions (table 5.1). With new ideas and adaptability. At various points in its history, the advent of the sugar preferences and the EPZs in the Mauritius has used intervention, subsidies, and targeting 1970s and 1980s, Mauritian authorities have succeeded in Table 5.1 Major Economic Indicators for Mauritius 2006–09 Indicator 2006 2007 2008 2009 National income (%) Real GDP growth 3.9 5.4 4.2 1.5 Consumer prices 11.8 8.7 6.8 4.0 Unemployment 9.1 8.5 7.2 8.3 Fiscal (% of GDP) Revenue 19.3 22.0 22.5 21.1 Expenditure 23.6 25.7 26.1 25.6 Fiscal balance –4.6 –4.0 –3.4 –4.5 Total external debt 6.1 4.6 6.1 8.0 Total domestic debt 46.6 45.6 43.8 42.0 External (%) Import 16.4 6.0 20.6 –24.5 Export 8.9 –4.7 7.3 –23.5 Current account (% of GDP) –9.4 –5.6 –10.4 –8.1 Terms of trade –5.8 –1.0 –1.1 1.6 Domestic investment (% of GDP) Public 6.9 6.4 5 6.1 Private 17.3 18.7 19.6 20.2 Nominal GDP ($ billions) 6,507 7,521 9,321 8,527 Source: IMF (2010); Government of Mauritius. 92 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY transforming the economy and laying the foundation for Mauritius experienced steady growth, low inflation, and stable growth in the future. Sugar and textile revenues increased employment. GDP per capita, meanwhile, have been used to facilitate service-sector development increased approximately sevenfold between 1976 and 2008, and contribute to socioeconomic progress and higher liv- from less than $1,000 to nearly $7,000 (figure 5.2). At the ing standards. same time, consumer price inflation in Mauritius has remained in the low single digits through the 1990s and early 2000s, and the country’s debt stock has been manage- A well-managed economic regime able.5 Windfalls from sugar and textile preferences have Proper macroeconomic management—fiscal discipline been used wisely to help promote diversification and boost during boom times, monetary management that kept infla- growth. The structural transition away from agriculture and tion in the single digits and that produced interest rates into manufacturing and services shows the success of the that encouraged domestic savings, and an exchange-rate government’s efforts at economic diversification. policy that maintained flexibility and competitiveness for exporters—have all been key to Mauritius’s economic suc- Favorable international comparisons cess. At the same time, flexible public policy, especially in the form of experiments creating EPZs in the 1980s and Both a cursory examination of economic indicators and embracing the ICT industry in the 2000s, was also an detailed diagnostics show that Mauritius has been one of important feature. Trade reforms, the development of a the best-performing and most stable economies in Sub- social welfare system, and policies that favored human cap- Saharan Africa in recent decades, even in the presence of ital development also played a part. The result has been terms-of-trade and oil shocks. Using purchasing power manageable fiscal and current account deficits, high rates of parity (PPP) data for 44 Sub-Saharan African countries, private investment, and respectable and stable growth rates. Arbache and Page (2008) examine country-level dynamics of long-run growth between 1975 and 2005 and conclude that Mauritius was one of the best performers, both in per A very successful economic trajectory capita growth performance and in low volatility of growth, Between 1977 and 2009, and despite the volatility caused by alongside Botswana, Cape Verde, Gabon, Namibia, the exogenous shocks during those years, real GDP in Mauritius Seychelles, and South Africa. grew on average by 5.1 percent annually, compared with Decomposing the standard deviation of GDP per capita 3.2 percent for Sub-Saharan Africa (figure 5.1). The growth and economic growth into within-country and between- rate accelerated in the 1980s as a result of the macroeco- country components, one finds that Mauritius’s growth nomic reforms instituted in response to protracted balance stability was much higher than that of many oil economies, of payments and fiscal troubles. Following the reforms, such as Angola and Nigeria, and of many resource-poor Figure 5.1 Real GDP Growth in Mauritius, 1981–2009 12 10 8 Percent 6 4 2 0 81 83 85 87 89 91 93 95 97 99 01 03 05 07 09 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 Source: Government of Mauritius. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 93 Figure 5.2 GDP per Capita in Mauritius, 1976–2008 than in the past and to a greater extent by productivity growth, with TFP growth during this period averaging 8,000 more than 1 percent a year.6 Their estimates suggest that in 7,000 the EPZ sector, TFP grew by 5.4 percent from 1991 to 1999. 6,000 In another study, Rojid and Seetanah (2009) provide evidence that TFP gains in Mauritius have reached a U.S. dollars 5,000 4,000 plateau. TFP growth averaged 1.4 percent in the 1980s, 3,000 1.0 percent in the 1990s, and 0.7 percent in the 2000s, but the TFP contribution to overall growth varies depending 2,000 on the different growth rates (table 5.2). For the entire 1,000 period 1980–2007, TFP contribution to growth in Mauritius 0 averaged 1.0 percent annually, considerably higher than the 76 80 84 88 92 96 00 04 08 0.3 percent average in Common Market for Eastern and 19 19 19 19 19 19 20 20 20 Source: Government of Mauritius. Southern Africa (COMESA) countries over the same years. The TFP change in Mauritius resulted from economic reforms and human capital improvements. economies, such as Burundi and Central African Republic, in Sub-Saharan Africa. While Mauritius was growing, Structural transformation much of Sub-Saharan Africa was not. In 1975–94 growth decelerations were twice as frequent as accelerations: Over time the sectoral composition of the Mauritian econ- 29 percent versus 14 percent of all country-year observa- omy has changed profoundly. Between 1976 and 2010 the tions in the Arbache-Page dataset. share of primary-sector production declined from 23 per- cent of the overall economy to 6 percent, while the second- ary sector (including manufacturing, electricity, water, and The role of total factor productivity some construction) increased from about 23 to 28 percent Empirical work decomposing total factor productivity (with manufacturing making an even bigger jump). The ter- (TFP) in Mauritius during different time periods shows the tiary sector, which includes tourism and financial services, strong role of factor accumulation. As in many countries in grew from just over 50 percent to nearly 70 percent of GDP East Asia, the principal drivers of growth in Mauritius have (figure 5.3). Projections by the Mauritian government sug- been capital and labor accumulation, with TFP growth gest further expansion of the tertiary sector as a share of the making a significant but varying contribution. economy in the future. In general, the share of manufactur- In a paper using a growth accounting framework analy- ing output increased in the 1980s because of the presence of sis for Mauritius and highlighting the performance of EPZs but stagnated as the sector faced adjustment in the the 1980s and 1990s, Subramanian and Roy (2001) uncover 1990s and 2000s. Table 5.3 shows patterns for specific indus- diverse performance in the two periods. For the period try groups and subgroups between 1990 and 2010. Sugar, 1982–99, productivity growth in the EPZ sector was which represented more than 20 percent of Mauritius’s 3.5 percent, substantially more than double the 1.4 percent GDP in 1976, accounted for approximately 4 percent of for the economy as a whole. In the first period (1982–90), GDP in 2009. At a disaggregated sectoral level (whether at economic growth was rapid and driven predominantly by current prices or constant prices), there has been a strong the growth of inputs—capital and labor—which together structural change, with the decline in sugar, a rise of finan- accounted for more than 80 percent of the annual average cial services and real estate, and a mixed performance of tex- rate of GDP growth of 6.2 percent. During this period, tiles as the service sector has strengthened (figure 5.4) By employment growth averaged 5.2 percent a year, reflecting the 2000s the service sector had become the dominant fea- a sharp decline in the unemployment rate from 20 percent ture of the economy in terms of contribution to output. of the total labor force in the early 1980s to 3 percent in the This noteworthy structural transformation has helped the late 1980s. TFP growth was respectable at more than 1 per- country deal with decreasing returns to scale of capital cent, but capital accumulation was the main driver of accumulation at the sectoral level. The various economic growth. The authors find that economic growth from 1991 pillars in Mauritius have thus contributed to growth and to 1999, however, was driven less by capital accumulation mitigated output volatility. 94 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY Table 5.2 Total Factor Productivity Decomposition and Contribution to Growth Contribution to the Growth Rate Growth rate Capital Labor TFP Economy/ Years (percent) (percentage points) (percentage points) (percentage points) Mauritius 1980–1990 6.3 3.0 2.0 1.3 1991–2000 5.6 2.5 2.1 1.0 2001–2007 4.1 2.5 0.9 0.7 1980–2007 5.3 2.6 1.7 1.0 COMESA 1980–1990 3.1 1.5 1.7 –0.2 1991–2000 2.4 0.7 1.6 0.1 2001–2007 4.5 2.1 1.5 0.9 1980–2007 3.3 1.5 1.6 0.3 Source: Rojid and Seetanah 2009. Note: Percentages are rounded to nearest tenth. Figure 5.3 Sectoral Composition of GDP in Mauritius, Poverty levels in Mauritius have also fallen significantly. 1976–2010 In 1975, 40 percent of Mauritian households were below the presumed poverty line, according to the government’s 70 Central Statistics Office.8 By 1991/92 the proportion had 60 fallen to 11 percent, and by 2010 absolute poverty was less 50 than 2 percent. The poverty decline was achieved with no Percent of GDP 40 exacerbation of inequality. Significant improvements in gender equality have resulted from the massive inflow of 30 women into the labor market starting in the 1980s. The 20 country recently ranked 11th of 102 countries in the OECD 10 Social Institutions and Gender Index. Mauritius’s housing 0 stock has also improved, both in quality and quantity, as a 1976 1980 1985 1990 1995 2000 2005 2010 result of government investment. Finally, Mauritius has Primary Secondary Tertiary developed a sophisticated pension system covering retire- ment benefits and general social security. Source: Mauritius authorities; World Bank 2009. MACROECONOMIC MANAGEMENT Improvement in human development indicators Prudent, proactive fiscal policy In contrast to many Sub-Saharan African and comparator economies, rapid economic growth in Mauritius has A hallmark of economic management in Mauritius in recent occurred in parallel with substantial improvement in decades has been prudent fiscal policy, which has helped human development indicators and a decrease in income maintain macroeconomic stability and contributed to inequality. Life expectancy at birth, for example, increased growth. Fiscal policy has focused on ensuring that spending from 62 years in 1970 to 73 years in 2008, while infant remains linked to resource availability. While there have been mortality dropped from 64 per 1,000 live births in 1970 to fiscal imbalances, there is no history of the government bor- 15 in 2008 (figure 5.5). The Gini coefficient, a measure of rowing from the central bank or from aid agencies. The income inequality in which 0 represents perfect equality strong growth in the 1980s led to a decrease in recourse to among households and 1 represents perfect inequality, foreign financing. The budget deficit, which was at 3 per- declined from 0.5 in 1962 to 0.37 in 1986–87 and was stable cent in 1983, turned into a budget surplus by 1987, as at 0.38 in 2008. Following heavy government investment in current expenditures shrank from 26 percent of GDP to 21 education improvements in the 1980s and 1990s, primary percent in 1987 while government revenues remained flat, school enrollment rates reached very high levels, averaging around 24 percent over the same period. On the expenditure more than 90 percent in the 1990s and 2000s, although chal- side, the government withdrew from subsidies on food items lenges remain.7 and kept the wage bill under control during that period. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 95 Table 5.3 GDP by Industry Group at Current Basic Prices, 1990–2010 (percent) Industry group 1990 1995 2000 2005 2010a Agriculture, hunting, forestry, and fishing 12.9 10.4 7.0 6.0 4.3 Sugarcane 8.0 5.7 3.6 3.2 1.6 Other 4.8 4.6 3.4 2.8 2.7 Mining and quarrying 0.2 0.2 0.2 0.1 0.0 Manufacturing 24.4 23.0 23.5 19.8 19.1 Sugar 3.4 1.6 0.8 1.0 0.5 Food, excluding sugar 0.0 — 4.1 5.1 6.6 Textiles 0.0 — 12.0 6.7 5.2 Other 0.0 — 6.6 7.0 6.8 Electricity, gas, and water supply 1.5 2.4 1.7 2.1 2.5 Construction 6.7 6.4 5.6 5.6 7.1 Wholesale and retail trade; repair of motor vehicles, motorcycles, and personal and household goods 13.0 12.8 12.2 12.1 11.9 Wholesale and retail trade 12.6 12.3 11.7 11.4 11.1 Other 0.4 0.5 0.5 0.7 0.8 Hotels and restaurants 3.9 5.1 6.5 7.7 7.5 Transport, storage, and communications 10.4 11.4 13.0 12.6 10.8 Financial intermediation 4.9 6.5 9.7 10.3 12.3 Insurance 1.5 2.1 2.3 2.9 2.9 Banksb 0.0 4.4 6.6 6.2 8.0 Other 0.0 — 0.8 1.2 1.4 Real estate, renting, and business activities 8.9 8.5 8.9 10.2 11.9 Owner-occupied dwellings 6.4 5.3 4.5 5.0 5.0 Other 2.5 3.2 4.4 5.2 6.9 Public administration and defense; compulsory social security 6.4 6.7 6.7 7.1 6.5 Education 4.1 4.4 4.5 4.8 4.6 Health and social work 2.5 2.8 3.0 3.4 3.8 Other community, social, and personal service activities and private households with employed persons 2.0 2.8 3.3 3.7 4.7 Financial intermediation services indirectly measured (FISIM) –1.8 –3.3 –5.7 –5.5 –7.0 GDP at basic prices 100.0 100.0 100.0 100.0 100.0 Manufacturing industries previously operating with an EPZ certificate 11.9 11.4 11.9 7.4 6.7 Source: Central Statistics Office, Government of Mauritius. Note: Figures may not add up to the totals due to rounding. — = not available. Data are at current prices. a. Forecast figures. b. For years 1991 to 1996, figures for other financial intermediation are included in banks. Figure 5.4 Major Economic Sectors in Mauritius, Similarly, in the 2000s, the government built up reserves 1976–2009 that allowed it the freedom to expand fiscal policy in the 100 aftermath of the 2008–09 global financial crisis. Moreover, the stock of international and domestic debt has remained 80 well below an unsustainable threshold (total public debt Percent of GDP 60 was projected to be 60.4 percent of GDP at the end of 2010). Excluding a period in the early 1980s, current 40 expenditures have never much exceeded 20 percent of GDP 20 and have been used mostly for wage policy, with a small amount devoted to food subsidies. Taken together, the 0 1976 1980 1985 1990 1995 2000 2005 2009 composition of current expenditure has been mostly Sugar Textile Construction oriented in recent decades to the outlays for wages and sub- sidies, but there has been little expansion of the federal Hotels and restaurants Financial intermediation Real estate apparatus as a share of GDP because fiscal profligacy has Source: Mauritius authorities; World Bank 2009. been an anathema to the Mauritian policy makers. In the 96 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY first four years of the 1980s, the ratio of current expendi- trade taxes. The diversified stream of revenue helped the ture to GDP was close to 25 percent but it steadily dropped fiscal system absorb shocks and provide stable revenue to around 20 percent by the mid 1980s, and that ratio has flows to the government. stayed relatively constant, in contrast to that of many devel- oping countries. Capital expenditures, which have averaged Monetary policy as an anchor less than 5 percent of GDP since independence, have been for economic growth used productively to invest in infrastructure—especially In tandem with fiscal policy, monetary policy in Mauritius roads—and to provide a necessary operating environment has helped anchor economic growth and ensure competi- for EPZs.9 Mauritian policy makers have been remarkably tiveness. Since its creation in 1967, the Bank of Mauritius fiscally prudent to avoid any macroeconomic instability. In has been concerned with ensuring the competitiveness of sum, fiscal policy has helped lay the foundation for manage- the country’s export sectors and, secondarily, with price ment of volatility and robust growth. stability. In important ways, a series of exchange rate deci- In terms of both revenue management and expendi- sions early on had ripple effects on economic activity. As tures, fiscal policy in Mauritius has been proactive. Inter- part of the liberalization program, the Mauritian rupee national trade taxes anchored the system’s revenue (MUR) was devalued by 30 percent in 1979 and read- system, accounting for close to 50 percent of GDP during justed by 20 percent in 1981, when the rupee was officially the 1970s, 1980s, and 1990s. High import tariffs and delinked from the IMF’s special drawing right and pegged export levies on sugar helped give the government to a trade-weighted basket of the currencies of its major resources during the early years, although the introduction trading partners. By the mid-1980s the Bank of Mauritius of a value added tax (VAT, which replaced the sales tax) in was intervening to smooth currency fluctuations, and the 1998 has played an important role in improving tax buoy- country has maintained a managed float since the mid- ancy at the level of direct and indirect taxes and has 1990s. Figure 5.6 shows the path of the real effective allowed the fiscal system to evolve.10 As a result, the tax exchange rate (REER) of the rupee (base year is 1970 = 100). system has not been affected by import tariff liberaliza- The results of the managed float exchange rate regime have tion in recent years. In addition, in 2007 the implementa- been favorable to the economy, with the trade-weighted tion of a flat tax of 15 percent on corporate and personal REER depreciating on average by more than 5 percent a incomes streamlined tax administration. By 2008 tax rev- year from 1981 to 2007.11 Although inflation has been enue amounted to 19 percent of GDP, of which 20 percent higher in other Sub-Saharan African countries than in came from income taxes, more than 35 percent from the Mauritius, the central bank allows the accommodation of VAT, and slightly less than 5 percent from international these inflation differentials by letting the nominal exchange rate slide in order to keep the REER competitive. The pol- icy has worked well in terms of being flexible and being Figure 5.5 Social Indicators in Mauritius, 1970 and 2008 sensitive to the country’s export sector. 80 70 Figure 5.6 MUR Real Effective Exchange Rate, 60 1981–2009 50 Number Index (base year 1970 = 100) 130 40 120 30 110 20 100 10 90 0 80 Life expectancy (years) Infant mortality (per 1,000) 70 81 83 85 87 89 91 93 95 97 99 01 03 05 07 09 1970 2008 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 Source: World Bank 2010, UNICEF. Source: IMF. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 97 As inflation fell in other parts of the developing world Figure 5.7 Trade and Tourism Growth, 2006–09 in the 1990s, Mauritius adopted an informal inflation tar- geting approach. Over time, monetary policy in Mauritius 40 has evolved from a strong reliance on direct monetary 30 instruments, such as credit ceilings, to a gradual intro- 20 duction of market-based instruments such as weekly Percent auctions of Treasury and Bank of Mauritius bills. In 10 practice, monetary management has been oriented to 0 ensure a positive interest rate differential relative to major –10 currencies while reacting to domestic inflation when it is above a certain level. Interest rate policy has been used –20 successfully to provide savers with positive real rates of –30 return in order to mobilize domestic capital and to help 2006 2007 2008 2009 sterilize excess liquidity. As a result, national savings has Exports of goods Tourism receipts Imports of goods consistently exceeded investment expenditure since the mid-1980s.12 Source: IMF 2010. Effective response to economic shocks about 5 percent of GDP over 2009–10. The plan focused on A tribute to Mauritius’s successful macroeconomic manage- infrastructure spending, providing financial relief to the ment can be seen in the response to the global financial firms hit hardest by the global crisis, and social and job crisis and the various shocks that preceded it. The combi- protection measures. At the same time, the government nation of the phasing out of the Multifibre Arrangement introduced offsetting measures that are expected to bring starting in 2004, the reduction in sugar price guarantees the primary budget into surplus by the end of 2011, and a from the European Union starting in 2006, and dramatic public debt management act was passed limiting public sec- increases in world commodity prices (especially for food tor debt to 60 percent of GDP, with a goal of reducing it to and fuel) had already started to act as a brake on the coun- 50 percent by 2013. The government has also been using try’s growth trajectory and current account and fiscal special “rainy day” funds to help finance the stimulus. These positions when the financial crisis began. The financial cri- funds had been prudently put aside in previous financial sis hit the small, open economy of Mauritius hard. The years (to the tune of 3 percent of GDP), reducing current country experienced a sharp decline in demand for tourism debt financing needs. Monetary policy in Mauritius was also and textile exports (figure 5.7). Labor-intensive sectors, loosened, the key discount rate at the central bank was cut especially construction and textiles, contracted as the crisis by 250 basis points, and reserve requirements were reduced. made its way to Mauritius. In addition to easing fiscal and monetary policy, As a result of the crisis, growth slowed to 4.2 percent in Mauritius introduced measures to assist the private sector 2008 and to less than 2 percent in 2009, while unemploy- in the wake of the crisis. Firms facing liquidity difficulties ment increased from 7.2 percent in 2008 to 7.7 percent in were given temporary financial relief (conditional on credi- 2009. Through prudent macroeconomic management, ble restructuring plans), with costs shared by banks, the however, international reserves continued to expand and government, and the firm’s shareholders.13 A tax suspension the fiscal deficit was kept below 5 percent of GDP. Moris- program was also introduced for the labor-intensive and set, Bastos, and Rojid (2010) find that the country’s vulnerable tourism, construction, and real estate sectors. resilience to the shocks derived from a combination of four factors: reforms to sustain long-term growth, which A TALE OF THREE SECTORS: SUGAR, EPZS, accelerated in 2006; a timely, targeted, and temporary short- AND TOURISM term response to the crisis; institutional arrangements to face the crisis that promoted private sector collaboration; At various times in Mauritius’s history, sugar, EPZs, and and strong relationships with development partners. tourism each have been a mainstay of the economy, and the In response to the global financial crisis, Mauritius interaction among the three sectors has been essential for passed a fiscal stimulus and monetary easing package of the country’s economic take-off. As a small, open economy 98 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY with a high ratio of trade of goods and services to GDP form of import licensing, which covered nearly 60 percent (averaging more than 50 percent from 1970 to 2010), of imports, and an extensive system of exemptions and con- Mauritius has long been well positioned to embrace cessions. Protection was especially high in the clothing, globalization. Over the years Mauritius has used trade footwear, furniture, and rubber sectors, all of which had policy as a means both to protect domestic industry and tariffs above 50 percent, while tariffs for electronics and to launch export growth. plastics averaged more than 40 percent. Corporate taxes were also very high, and bureaucracy was quite heavy. Pro- tectionism faded through the course of the 1980s, however, The sugar sector and by the early 1990s import licensing was eliminated on In its heyday in the 1970s the sugar sector in Mauritius all but a limited list of items subject to health, sanitary, or accounted for close to one-third of employment, one-third strategic controls, while export licensing was abolished for of export earnings, and one-quarter of GDP. Through smart most products. negotiations and building on a preexisting relationship with the United Kingdom, Mauritius succeeded in obtaining preferential treatment from the European Economic Com- The rise of EPZs in the 1970s and 1980s munity (EEC) through the Sugar Protocol of the Lomé Having studied the success of export processing zones in Convention in 1975, under which it received more or less East Asia, a group of visionary policy makers in Mauritius free access for its sugar exports to the EEC.14 Mauritius sold put forth the idea that the country’s small economic size its sugar to the EEC at a premium—three times the interna- and distance from large developed markets presented a tional market price, on average. For years Mauritius’s export potential opportunity to develop an export-oriented textile quota was fixed at more than 500,000 metric tons annually, industry. In 1970 Mauritius passed the Export Processing the largest quota share among African, Caribbean, and Zone Act, which provided powerful incentives to manufac- Pacific (ACP) countries. Even with these international trade turers that catered to foreign markets. The EPZ was not agreements in place, however, the collapse in international restricted to one physical location but was envisaged as a fis- sugar prices in the mid-1970s hit Mauritius’s sugar sector cal regime that encompassed the entire island, and the idea hard, leading to balance of payments difficulties and was to develop a fast-track approval for all administrative recourse to external assistance. Nevertheless, by 2005, in the procedures in the EPZ. Ambitious and well-crafted legisla- aftermath of the World Trade Organization’s ruling that the tion provided the underpinnings for the new regime. Key above-market prices paid to sugar producers constituted components of the new legislation included protective unfair trade, the European Union ended the preferential import duties and quotas for infant industries, suspension deals by slashing sugar prices. of import duties on materials and equipment for industrial use that were not locally available, rebates of import duties Import substitution industrialization (ISI) on other raw materials and components for specified indus- and restrictive trade policies following tries, duty drawback schemes, and favorable long-term the colonial period loans. The granting of duty-free inputs for manufactured Following the establishment of a development certificate exports was key in expanding Mauritius’s export competi- scheme by colonial authorities in 1964 to promote import tiveness on world markets, while tax incentives helped sub- substitution industrialization and provide incentives for sidize exports. Firms within EPZs also benefited from the local manufacturers through concessions and tariffs, Mauri- availability of relatively cheap labor, drawn from unem- tius invoked a series of protectionist measures in an effort to ployed workers and women who were outside the labor develop local industry. Besides providing government with force at the time. According to interviews with textile exec- some revenue and helping to incubate entrepreneurial tal- utives located in the EPZs, 80 percent of workers in the EPZs ent, these policies, which were concentrated in the EPZs, in the 1980s were women. The rate has decreased somewhat had a marginal impact on the local economy. Subramanian in the 1990s and 2000s, but more than 60 percent of the and Roy (2001) find that Mauritius’s trade policy was highly workers in the zones are women. The lower wages that were protective during the 1970s and 1980s. In 1980 average paid to the workers in the EPZs in the early years allowed effective protection exceeded 100 percent, although this fell the firms to accumulate capital and reinvest the earnings into to 65 percent by the end of the 1980s. During the 1970s and the firms’ expansion.15 However, over time the wages in the 1980s, there were extensive quantitative restrictions in the EPZ became higher than those in the non-EPZ economy. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 99 Also important, Mauritius did not restrict EPZs to one geo- reunification with China. Third, there was a strong political graphical location, and the government invested heavily in stability, proper governance, and a clear legal and institu- the infrastructure needed to set up EPZs. Finally, it is tional demarcation between the ISI and the EPZ regimes. important to note that the government provided strong Fourth, quotas on Asian textile exports into Western institutional support for marketing EPZ products. markets led investors to look to alternate production coun- By the 1980s EPZs had exceeded the expectations of tries. Those investors brought capital, marketing networks, even visionary policy makers in Mauritius. The zones and technological know-how to Mauritius’s nascent textile accounted for more than 60 percent of Mauritius’s gross sector. Greenaway and Milner 1989, however, find that the export earnings and employed one-third of the Mauritian decision to grant Mauritius trade preferences in garments labor force. Goods produced in EPZs more than tripled as through the Multifibre Arrangement was more important in a share of GDP between 1980 and 1988, from 4 percent to giving the country privileged access to developed markets more than 14 percent (figure 5.8). More people worked in relative to established Asian producers. Fifth, EPZs bene- EPZs than in the agricultural sector by the end of the 1980s. fited from the inflow of local capital and of indigenous The growth rate of the EPZs’ value added was close to 30 managerial capacity, which had partly been incubated percent annually between 1983 and 1988. Most of the under import substitution policies. Bheenick and Schapiro goods produced in EPZs were exported to Europe under a (1989) find that local participation in EPZ equity in Mauri- preferential regime. Notably, there was significant interac- tius was roughly half, a much higher ratio than in EPZs in tion between the sugar sector and EPZs. Much of the other developing countries. And finally, Mauritius benefited start-up capital for EPZ firms, as well as technical and from a strong entrepreneurial class and a trainable labor managerial expertise, came from the well-established sugar force, that together with the political class, had a strong companies in the aftermath of the sugar boom in the 1970s. ownership in the success of the EPZs. In turn, the job Together with the sugar sector, the textile sector provided opportunities offered by EPZs played a significant part in the capital accumulation that allowed Mauritius to unemployment falling from 20 percent to less than 5 per- decrease reliance on foreign capital and start down the path cent between the mid-1970s and 1990. Over time the EPZs to becoming a middle-income economy. have helped the country’s exports match the imports, Several additional factors contributed to Mauritius’s although the proportion of service exports has increased in success with EPZs. First, the country took advantage of the recent years as the economy has changed (see figure 5.8). depth of EU and U.S. demand for textiles and apparel, which provided a solid base for expansion. Second, Mauritius’s timing was good, as its initial entry into the Post-EPZ economic drivers: Tourism, business U.S. market got a boost from investors based in Hong Kong process outsourcing, and financial services SAR, China, who were seeking to move capital and factories Although EPZs brought dramatic economic improvements out of Hong Kong SAR, China, in anticipation of the 1997 to Mauritius, the textile and apparel sectors have met chal- lenges in more recent years. Not least of these was the phas- ing out of the Multifibre Arrangement, which started in 2004 Figure 5.8 Value of Mauritian Exports and Imports, and led to a contraction of 30 percent in value added of the 1980–2009 products produced in EPZs. The number of Mauritians 80 employed in EPZs fell by about 25,000 people between 2005 70 and 2010.16 In parallel, the European Union’s reduction of 60 sugar prices by more than 50 percent starting in 2005 and 50 continuing onto 2010 was a significant blow to the Maurit- 40 ian economy, given that from 1975 to 2005, about 90 percent 30 of its sugar production was exported to the European mar- 20 10 ket. The European Union’s decision to end price guarantees 0 on raw sugar for all countries has thus been a shock for the 1980 1990 2000 2009 $10 billion Mauritian economy. In short, both the sugar and Exports/GDP Exports of goods/GDP textile and apparel sectors are in the process of adjustment in Exports of services/GDP Imports of goods and services/GDP order to remain globally competitive, with the textile sector Source: Mauritian authorities. making inroads into fully integrated activities (for example, 100 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY spinning and weaving) and higher-end manufacturing and policy framework led to a strong private sector response. the sugar sector increasing its focus on refined sugars. And in all cases, the government acted as a facilitator of pri- As the sugar and EPZ sectors in Mauritius have struggled vate sector expansion. in recent years, the tourism sector has expanded rapidly, backed by a master plan reflecting the government’s dislike DYNAMIC INSTITUTIONS AND ADAPTABILITY for mass tourism and high-rise buildings. Public sector efforts TO CHANGE to expand tourism have been complemented by the promo- tional activities of the hotels and by Air Mauritius. Indeed, Aside from its success in macroeconomic management, one Mauritius’s combination of beautiful beaches, a multieth- of the keys to Mauritius’s economic success has been its rich nic society, and excellent hotels dotting the coastline has web of trade links, effective institutions, and history of been quite effective in attracting tourists. According to the public collaboration. In surveys of institutional quality, Mauritius Chamber of Commerce, tourist arrivals reached Mauritius repeatedly ranks high relative to comparator 240,000 in 1988, 400,000 in 2000, and 900,000 in 2008. An countries, particularly in measures of governance, rule of estimated 1 million tourists were expected to visit the law, and control of corruption. The combination of political country in 2010. Tourism is also among the strongest for- stability, democratic legacy, rule of law, and quality of judi- eign exchange–earning sectors of the Mauritian economy. cial institutions sets Mauritius apart from many Sub- In addition to tourism, the government of Mauritius has Saharan African countries. Moreover, a set of informal and also encouraged diversification of the economy into business formal mechanisms guide the interaction between the pub- process outsourcing (BPO), financial services, and informa- lic and the private sector, with the result that the private sec- tion technology. According to government figures, the BPO tor plays a seminal role in the policy formulation process industry has been growing 70 percent a year and is currently (all Mauritian delegations to international organizations, worth $1.6 billion, employing more than 100,000 people. for example, have a private sector member). Indeed, coop- Offshore banking was introduced in 1988 as a first step eration between the public and private sectors has a long toward developing Mauritius into an international financial history in Mauritius. center, and the offshore sector is now emerging as a growth For decades, Mauritius served as a trade hub for Chinese vehicle for the economy.17 Development of the information and Indian traders and an entrepôt for shipping across the technology sector, meanwhile, is intended to transform Indian Ocean, and its inhabitants were known for their Mauritius into “a cyber island” by creating a high-tech, entrepreneurship. Over time those trade links coalesced into multi-story tower that provides a home to companies from formal trade associations and entities, some of which came all over the world to set up BPO operations, including data to be represented at the political level. At the same time, management, e-commerce, and call centers. The government Mauritius has been quite effective at promoting its trade has also encouraged use of Mauritius as a transshipment links on the international level. center and a re-export base and, more recently, as an inter- national medical hub and regional knowledge center. As a result of these efforts, the Ministry of Finance finds that the The importance of forging consensus services sector showed consistent growth over 2006–09, with The search for consensus is one of the remarkable features financial intermediation growing by 10.1 percent in 2008. of the Mauritian political economy. Essentially, the Maurit- ian state is modeled on the British system of government, with a cabinet headed by a prime minister and a legislative Lessons of the three sectors assembly serving as the law-making body. As in the United The story of the three sectors in Mauritius offers a number Kingdom, the system of government in Mauritius is based of lessons. Most important, the constant search for new on the principle of separation of power between the legisla- drivers of economic growth reflects a desire by policy mak- ture, the executive, and the judiciary. Gulhati and Nallari ers to adapt to the future rather than wait for and respond (1990) argue that since no single political party has ever to shocks. The central lesson here is that the public sector secured a majority in the assembly, which would allow it to can effectively formulate and implement sectoral policies to form a government on its own, there has always been a need stimulate the private sector. In the case of sugar, the to work together across party lines, putting a distinctive protocols were signed and the private sector rushed into the stamp on economic policy process. Despite being based on activity. In the case of EPZs and tourism, a well-articulated loose agglomerations of ethnic and economic interests, CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 101 political parties in Mauritius have not been vehicles for eth- and a large number of double taxation avoidance agree- nic separation. To the contrary, political parties have long ments; together, these attributes sometimes make Mauri- recognized that building consensus is necessary to avoid tius more attractive than larger financial sectors for busi- adverse economic effects in a small economy. nesses. International rankings consistently give Mauritius high marks for business and investment climate. The main lesson from Mauritius in this regard is that geography is The nexus between the public and private sectors not destiny and that policies to improve investment cli- A paramount role in state-business relations in Mauritius is mates can have large positive multiplier effects. played by the Joint Economic Council (JEC), which occu- pies a central place in the country’s institutional landscape Business climate improvement in recent years and represents an umbrella association of a number of sector-specific groupings. As such, it carries a certain The World Bank’s Doing Business 2010 ranks Mauritius as amount of institutional weight, meeting with the prime the best country in which to do business in Africa. Overall, minister on a regular basis and providing input on major Mauritius is ranked 20th of 183 countries included in the policy decisions. Being funded entirely by the private sector, 2010 survey, up from 24th of 183 in 2009. Currently, it also has a degree of financial autonomy. The overarching Mauritius is among the top-performing developing coun- goal of the JEC is to ensure private sector representation in tries in starting a business, paying taxes, and protecting all key government economic decisions. It also ensures that investors, and it has been a consistent reformer since it its members’ ideas are conveyed to political leaders. began being included in Doing Business in 2005. More Another example of public-private sector cooperation is broadly, Mauritius has taken steps to improve business the establishment of the EPZs, which would not have been facilitation since the late 1990s. In earlier years, while the possible without support from several key public sector private sector was thriving in certain sectors such as sugar institutions. Central among these were the Mauritius Export and textiles, the climate for domestic firms in the non-EPZ Development and Investment Authority (MEDIA), the sector was unfavorable. A study by Lall and Wignaraja Export Processing Zone Authority (EPZDA), and the Devel- (1998) found several major obstacles for enterprises oper- opment Bank of Mauritius (DBM). MEDIA was formed in ating in Mauritius at the time: high interest rates; heavy 1985 as a public trade and investment promotion agency bureaucratic procedures resulting in delays in obtaining (with some private sector membership) and was a pivotal foreign investment approvals; difficulty getting loan institution behind the country’s drive for export growth and approvals from the Development Bank of Mauritius; delays industrial upgrading. Providing overseas marketing support in receiving refunds on import duties; difficulty obtaining for exports and arranging buyer-seller meetings, it helped work permits for foreign technical staff; lack of access to explore niches for Mauritian garments in European and finance for small enterprises; and high sea-freight costs. American markets. Formed in the early 1970s, the EPZDA Government reforms have helped alleviate all of these helped represent the interests of firms in the EPZs, while the obstacles in the years since. Development Bank of Mauritius provided much of the One major piece of legislation, the Business Facilitation credit and start-up capital for the economy as it was taking Act 2006, provided a new, streamlined legal framework for off from its narrow monocrop base. More recently, the Board business operations in Mauritius. The legislation facilitates of Investment has played a role in helping to promote Mau- doing business and acquisition of properties by foreigners ritius as an international investment, business, and service and, among other things, enables small enterprises to start center, providing counseling on investment opportunities in business activities within three working days. As a result, the Mauritius and helping in setting up businesses. private sector and foreign investors can more easily venture into new sectors, such as real estate and financial services, in which growth rates are eclipsing traditional sectors. BUSINESS CLIMATE AND INVESTMENT Between 2006 and 2010, the share of private sector invest- Alongside successful trade policy and adaptability, another ment, mainly in infrastructure projects such as commercial major reason for Mauritius’s economic success has been its and office buildings, hotels, and resorts, grew to account for business climate and incentives for foreign companies to more than 80 percent of gross domestic capital formation. locate there. Mauritius has no capital controls, a relatively While infrastructure deficiencies and difficulties obtaining stable currency, a low flat corporate tax rate of 15 percent, credit remain for small firms, along with a lack of skilled 102 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY labor for larger firms, the thrust of the reforms has been to Mauritius through private equity, hedge, and mutual funds. alleviate these constraints. Under the current double taxation treaty between India and Mauritius, capital gains on Indian shares that are held by a Mauritian company are not subject to Indian tax laws The role of foreign direct investment and rates, an issue that has been vexing to Indian regulators FDI inflows to Mauritius have increased rapidly in the past and policy makers. Simply put, Mauritian companies are several years, attracted by reforms such as the removal of taxed according to Mauritius tax laws, which are extremely the tax on capital account transactions and the waiving of favorable, compared to Indian laws. There is some chance the requirement that foreign investors need approval of the that India will try to amend the treaty in 2011 to try to Bank of Mauritius to carry out activity. In addition while capture more of the revenue from this activity. Neverthe- the corporate tax rate is a low 15 percent, foreign firms less, FDI from Mauritius to India, which has been mostly receive a subsidy of close to 10 percent, leading to an effec- in the electrical equipment, cement, telecommunications, tive tax rate of 5 percent. The country attracted more FDI and financial sectors, has helped Mauritius establish the during 2004–07 than the cumulative stock of FDI during attributes needed to compete globally in high-value ser- the previous 25 years (figure 5.9). Importantly, FDI inflows vice sectors. are accompanied by new business ideas, technologies, and managerial skills. Most FDI inflows have gone to the hospi- THE ECONOMIC FUTURE OF MAURITIUS tality and tourism, property and real estate, banking and finance, information technology, health, and education sec- The long-term challenge for the government of Mauritius is tors. The main FDI source countries are France, South to maintain its unique combination of resilience in the face Africa, and the United Kingdom, although total FDI inflows of changing economic circumstances and adaptability to are equally divided between developed and emerging coun- new paths for growth. A number of key lessons can be tries. Interestingly, Mauritius has also been a beneficiary learned from Mauritius’s experience. First, the forging of of a high inflow of FDI into India. Because of special tax consensus between the Franco-Mauritian business elite and treatment given to investments that come through Mauritius the Indian political elite has provided a sound foundation to India,18 Mauritius has become a quasi-tax haven for for- for economic growth. Consensus building was mapped into eign funds invested in India, and currently, about 80–90 per- the management of ethnic interests while the country cent of foreign direct investment into India flows through opportunistically moved forward with growth strategies. Figure 5.9 FDI Inflows to Mauritius, 2002–08 12,000 10,000 Millions in local currency 8,000 6,000 4,000 2,000 0 02 03 04 05 06 07 08 20 20 20 20 20 20 20 Manufacturing Wholesale and retail trade; repair of motor vehicles Hotels and restaurants Financial intermediation Real estate, renting Other Source: World Bank 2009. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 103 Even an occasionally contentious political environment has that on balance, the disadvantages outweigh the advan- not jeopardized the continuity of state policies and admin- tages: initial conditions have slowed growth by about 1 istrative stability. The combination of adaptable institutions percentage point a year relative to the average African and a rich interface between the public sector and private country and by nearly 2 percentage points relative to the sectors has ensured effective economic policy. The impor- fast-growing developing economies of East Asia. tant role the private sector plays in the formulation of 2. See Sachs and Warner 1995, 1997; Meade 1961; Rodrik economic policy, especially through the Joint Economic 1999; Romer 1993; and Subramanian and Roy 2001, among Council, is relatively unparalleled in Africa. others. Second, Mauritius benefited from pragmatic macroeco- 3. Mauritius has three ethnic groups: Mauritian Creoles, nomic management that was supportive of long-term growth Indo-Mauritians, and Franco-Mauritians. The Creoles (African) were brought to Mauritius as slaves to work for aspirations. The rents generated within the system during the owners of sugar cane fields. Indo-Mauritians came to 1970s and 1980s were used to finance capital accumulation Mauritius as indentured laborers after slavery was abolished rather than consumption. The real exchange rate was kept in 1835 and eventually became the country’s political elite. competitive, fiscal discipline was maintained, and debt bur- And Franco-Mauritians were the French who remained in dens were kept at respectable levels, while there was a willing- Mauritius after the British took over in 1810 to look after ness to correct external and internal imbalances when needed. their large sugar estates and other businesses, including Third, Mauritius recognized the benefits of economic trading and banking. openness at an early stage, implementing effective sectoral 4. Mauritius’s legislative system is based on a classic parlia- polices and building a good investment climate. The policy mentary Westminster system. Legislative power is vested in framework was used to facilitate private and foreign invest- the National Assembly, which is composed of 62 elected and ment, particularly in the textile and tourism industries. The up to 8 designated representatives. The four main current interplay between the relatively closed import substitution political parties in Mauritius are the Labour Party (PTR), industrialization on one hand and export-driven initiatives the Movement Mauricien Militant (MMM), the Mauritius on the other (particularly the EPZs) provides a fascinating Socialist Militant (MSM), and the Parti Mauricien Xavier Duval (PMXD). tale. In general, though, Mauritius’s openness, which allowed it to successfully penetrate developed markets 5. Although Mauritius’s public debt is relatively high, at about 60 percent of GDP, most of it is held domestically by through exports, has been unrivaled by countries in Africa the National Pension Fund and commercial banks. and the Middle East. Through smart tax policy, the country 6. Note that TFP calculations are very sensitive to the start has become a source of a large portion of FDI flows into and end year. India. Throughout, Mauritius has demonstrated a capacity 7. One challenge Mauritius faces is the large proportion of to capitalize on good international relationships. young people who are unable to access secondary education While its future success is not guaranteed, Mauritius has because of the very competitive system for moving from proven that it has the right instruments to weather a range primary to secondary schools and who thus cannot con- of economic shocks. Its history is one of adaptability, inno- tribute to the skilled labor force required by the economy. vation, and anticipating global changes. Its combination of Almost 35 percent of students fail to pass the completion- good leadership, consensus building, sound macroeco- of-primary-education examination and drop out of the nomic management, and positive policies for the private school system at the age of 12. sector will serve it well in the future. 8. Mauritius does not have a national poverty line; poverty figures are derived by the Central Statistics Office using census and survey data. NOTES 9. It should be noted that capital expenditures are under- 1. Careful empirical work by Subramanian (2009) shows estimated in the Mauritian budget, because many public how initial conditions in Mauritius—especially the investments take place through parastatals, while expendi- income level, geography, and commodity dependence— ture classifications are detailed for the central government. have hurt long-term growth, while favorable demography As a result, some capital expenditures may be classified as and high levels of human capital have been mitigating current expenditures. factors. For example, Mauritius is disadvantaged by being 10. The VAT was introduced in Mauritius in September at least 25–30 percent more distant from world markets 1998, close to six months ahead of schedule, and its perfor- than the typical African country. Statistical analysis shows mance has exceeded expectations. 104 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 11. The IMF (2008) finds that using the macroeconomic Bheenick, R., and M. O. Schapiro. 1989. “Mauritius: A balance approach and the single-equation equilibrium Case Study of the Export Processing Zone.” EDI exchange rate approach, the real exchange rate at the end of Development Policy Case Studies 1, World Bank, 2007 was broadly in line with its equilibrium value (as Washington, DC. determined by economic fundamentals). Greenaway, D., and C. Milner. 1989. “Nominal and Effective 12. Since the mid 1980s private sector investment has also Tariffs in a Small Industrializing Economy: The Case of exceeded private sector savings. Mauritius.” Applied Economics 21 (8): 995–1009. 13. Technically, the Mechanism for Transitional Support for Gulhati, R., and R. Nallari. 1990. “Successful Stabilization the Private Sector (MTSP) under the Additional Stimulus and Recovery in Mauritius.” EDI Development Policy Package (ASP) was a combination of equity support, Case Studies. World Bank, Washington, DC. liquidity/working capital (including guarantees for bank IMF (International Monetary Fund). 2008. “Mauritius: support), and asset purchases and swaps. Between mid- Selected Issues.” Washington, DC. December 2008, when the mechanism began functioning, ———. 2010. “Mauritius Article IV Consultation Staff and September 2010, 11 companies had received assistance Report.” Washington, DC. under the MTSP, to which the government contributed Lall, S., and G. Wignaraja. 1998. “Mauritius: Dynamizing MUR 140 million (36 percent of the total MTSP support for Export Competitiveness.” Commonwealth Secretariat, these companies; the remainder was provided by banks and London. shareholder equity) in the form of debentures at 5 percent interest. Meade, J. E., et al. 1961. The Economics and Social Structure of Mauritius: Report to the Government of Mauritius. 14. The Sugar Protocol was negotiated as a condition of the London: Methuen. United Kingdom’s membership in the EEC to protect the developing countries from which it had traditionally Morisset, J., F. Bastos, and S. Rojid. 2010. “Facing Off a Man- imported sugar under the Commonwealth Sugar Agreement. Made Disaster: Global Financial Crisis and Policy Response in the Small Tropical Island of Mauritius.” 15. Firms within EPZs were subject to general labor laws World Bank Note, Africa Region, Washington, DC. (including minimum wages) but were free to fire workers, to demand compulsory overtime work, and to penalize work- Rodrik, D. 1999. The New Global Economy and Developing ers heavily for absenteeism. Countries: Making Openness Work, London: Overseas Development Council. 16. Despite the challenges, the government has supported the restructuring of EPZ textile firms in order to avoid a Rojid, S., and B. Seetanah. 2009. “Using Growth Accounting possible collapse of the sector. These efforts include estab- to Explain Sources of Growth: The Case of COMESA.” lishing the Textile Emergency Support Team initiative in International Journal of Business Research. July 2003 and encouraging the National Productivity and Rojid, S., B. Seetanah, and R. Shalini. 2009. “Are State Busi- Competitiveness Council to carry out a diagnostic study of ness Relations Important to Economic Growth: Evidence textile firms to assess their cost structure and point to areas from Mauritius.” DFID/University of Mauritius working in need of improvement. paper. 17. At the end of October 2002, the number of companies Romer, P. 1993. “Two Strategies for Economic Development: registered in the offshore sector reached 20,111. The Mauri- Using Ideas and Producing Ideas.” In Proceedings of the tius Freeport, the duty-free zone in the port and airport, World Bank Annual Conference on Development Econom- aims at transforming Mauritius into a major regional distri- ics. Washington, DC: World Bank. bution, transshipment, and marketing center. Sachs, J., and A. Warner. 1995. “Economic Reform and the 18. India has double taxation avoidance agreements with Process of Global Integration: Comments.” Brookings approximately 65 countries, including France, Germany, Papers on Economic Activity, 1: 1–118. Japan, the United Kingdom, and the United States, although ———. 1997. “Sources of Slow Growth in African Mauritius is the most preferred route for FDI inflows. Economies.” Journal of African Economies 6 (3): 335–76. Subramanian, A. 2009. “The Mauritian Success Story and its Lessons.” Research Paper 2009/36, United Nations BIBLIOGRAPHY University–World Institute for Development Economics Arbache, J., and J. Page. 2008. “Hunting for Leopards: Long- Research, Helsinki. Run Country Income Dynamics in Africa.” Policy Subramanian, A., and D. Roy. 2001. “Who Can Explain the Research Working Paper 4715, World Bank, Washington, Mauritian Miracle: Meade, Romer, Sachs, or Rodrik?” DC. IMF working paper, WP/01/116, Washington, DC. CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY 105 Wignaraja, G., and S. O’Neil. 1999. SME Exports and Public ———. 2009. Mauritius Investment Climate Assessment. Policies in Mauritius. Commonwealth Trade and Enter- Washington, DC. prise Paper. London: Commonwealth Secretariat. ———. 2010. World Development Indicators Database. World Bank. 1989. Mauritius Country Economic Memo- Washington, DC. randum, Washington, DC. 106 CHAPTER 5: MAURITIUS: AN ECONOMIC SUCCESS STORY CHAPTER 6 Cotton Dependence in Burkina Faso: Constraints and Opportunities for Balanced Growth Jonathan Kaminski urkina Faso experienced more than a decade of B 85 percent of Burkina Faso’s active labor force, has been the robust growth in gross domestic product (GDP) main driving force of growth. Within agriculture, the cotton following the devaluation of its currency in 1994. sector has shown particular dynamism. The cotton boom Average growth of 6 percent a year—nearly double the during the 2000s lifted cotton’s share of GDP from around growth of the previous decade—was led largely by the cot- 4 percent in the 1980s to 12 percent in 2008, contributing to ton sector, which accounted for about 60 percent of exports. higher GDP growth (figure 6.1). GDP per capita in Burkina Faso rose from $214 in 1997 to As cotton production in Burkina Faso posted unprece- $260 in 2007 ($430 in real terms). Reforms in the cotton dented growth in the 2000s, the share of cotton earnings in sector contributed to rapid growth of this sector, but export revenues shot up from less than 40 percent in the 1990s spillover in terms of structural transformation of the econ- to 85 percent in 2007. At the same time, increased export omy has been slow, especially because the increase in cotton dependency on cotton has exacerbated vulnerability to exoge- production was not productivity-based. While reforms that nous shocks over the past decade, which was characterized by led to management improvement and institutional upgrad- a pattern of falling world cotton prices and rising input prices; ing in the cotton sector were necessary, they were not suffi- a decline in local profitability and farm productivity; and cient to translate productivity increases in the cotton sector poorly performing cotton firms that lack the ability, informa- into economic consolidation, structural transformation, tion, and resources to adjust to evolving international markets. and diversification. Instead, the economy’s dependence on Although cotton represents a large proportion of Burkina cotton has exacerbated growth volatility and vulnerability to Faso’s exports, its contribution of export earnings to GDP is exogenous shocks. Despite this, Burkina Faso’s economic small (10 percent) and trade openness is limited. Most of the prospects remain positive due to record cotton prices in the country’s growth is from domestic demand.1 Limited export world market, a growing level of gold exports, effective gov- earnings highlight the unsustainable growth path that Burkina ernment-led growth strategies, and continuing improve- Faso is currently on. Although the country’s terms of trade ment of the business environment. have recovered since the early 2000s, structural deficiencies have led to persistent trade deficits (Savadogo 2009). ECONOMIC GROWTH IN BURKINA FASO EFFECT OF GROWTH ON DEVELOPMENT Burkina Faso’s average annual GDP growth rate accelerated from an average 3.8 percent in the 1980s to 6 percent during Sustained growth has had only a limited effect on poverty 1994–2008. The agricultural sector, which accounts for reduction in Burkina Faso (World Bank 2009a). Despite 107 Figure 6.1 Annual Changes in Cotton Production and GDP in Burkina Faso, 1980–2008 12 100 10 80 60 8 40 6 Percent Percent 20 4 0 2 –20 0 –40 –2 –60 –4 –80 81 83 85 87 89 91 93 95 97 99 01 03 05 07 19 19 19 19 19 19 19 19 19 19 20 20 20 20 Percent change in GDP (left axis) Percent change in cotton production (right axis) Source: World Bank 2010; FAOSTAT, SOFITEX. sustained annual per capita GDP growth of about 2–3 percent some amount of poverty convergence across the West between 1994 and 2003, the national poverty headcount African region as a result of the migration of less productive index remained about 45 percent at the end of that period. and poorer farmers in cotton regions to more marginal Granted, estimation of poverty measures has been subject to lands (Gray and Kevane 2001). In addition, the increase in controversies (including methodological), but no significant cotton production reduced poverty largely among house- effect of growth on poverty reduction has been found at the holds that were able to increase their per capita agricultural aggregate level (Lachaud 2005, 2006; Grimm and Gunther incomes through more productive assets and factors. There 2004). Grimm and Gunther (2004) show that only cotton- was no significant poverty reduction effect as a result of the producing households increased their expenditures signifi- decline in cereal prices or spillovers into the local economy cantly between 1995 and 2003, a finding that attests to the led by higher domestic demand (Kaminski, Headey, and pro-poor growth effect of smallholder cotton production. Bernard 2009). Several studies find that the impact of Burkina Faso’s GDP As global food and fuel price increases took hold in 2007, growth was mostly neutralized by its population growth, the purchasing power of households in Burkina Faso which limited scaled-up investments in infrastructure, declined and the incidence of poverty began to rise again. education, and health. Cotton households did not escape these negative impacts, Several pieces of research show evidence of poverty and in subsequent years, negative shocks such as flooding reduction in the mid- to late 2000s: the incidence of poverty have hurt cotton-producing households even more. At the fell from 46 percent in 2003 to 38.5 percent in 2007; the inci- same time, the deceleration in economic growth had a dra- dence of poverty in rural areas fell from 52 percent to matic effect on poverty, alleviation of which requires higher 44 percent over the same years. There is no evidence, how- rates of economic growth, mainly in the rural sector (cotton ever, that cotton-producing regions had lower incidence of and livestock). Measures of the depth and severity of poverty than other rural areas (Grimm and Gunther 2004). poverty in Burkina Faso have also increased since 2007. The apparent contradiction between pro-poor cotton Although poverty is much higher in rural areas than in growth on one hand and the absence of rapid poverty urban areas, unemployment in Burkina Faso is largely an reduction in cotton-producing areas on the other reflects urban problem that disproportionately affects women, youth, both the effect of political and economic turmoil in Côte and highly skilled workers.2 The urban unemployment rate d’Ivoire (which reduced remittances to Burkina Faso) and fell to 8.6 percent in 2007, down from 13.8 percent in 2003 108 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH (16 percent for the labor force ages 15–25, down from In the health sector, the prevalence of HIV/AIDS 30 percent in 2003). According to the World Bank (2009a), has fallen, and progress has been made in containing the quality of jobs has not improved, however, and a large epidemic and endemic diseases. Both morbidity and proportion of the workforce remains employed in low- mortality rates decreased between the mid-1990s and paying jobs and jobs that do not provide social benefits. mid-2000s. Absorbing young graduates in professional jobs remains a major challenge for the government in the coming years. A REFORM OF THE COTTON SECTOR new law that promotes more flexibility in the labor market went into effect in 2008, but creation of new jobs has been Since the early 1990s Burkina Faso’s dependence on cotton hampered by the economic slowdown. has grown as a result of the implementation of institutional Other indicators of social welfare and development reforms, which have brought new land and producers to have exhibited encouraging signs over the past decade. cotton production (Kaminski and Thomas 2009). Table 6.1 Education-related infrastructure has been scaled up, with details the chronology of the reforms. the number of classrooms increasing by 9 percent annually The cotton sector provides income for 15–20 percent since 2006. Primary school enrollment rates have also of the active labor force of Burkina Faso, supporting 1.5 risen, reaching 78 percent in 2009. Although regional dis- million–2 million people. It is composed mostly of small parities are significant, gender disparities in enrollment farms and smallholders, with a small number of large rates have narrowed. These results have been achieved farms led by rural elite. In 2005 Burkina Faso became the thanks to long-term policy frameworks, substantial leading West African producer of cotton, ahead of Mali, increases in public investment in education in the 2000s, producing 500,000 to 800,000 tons of seed cotton between and partnerships with the private sector (which has con- 2005 and 2010. In 2006 and 2007, Burkina Faso was the tributed about 15 percent of investment in education leading cotton producer and exporter among all African infrastructure). countries. Table 6.1 Chronology of Cotton Reforms in Burkina Faso, 1992–2008 Year(s) Development 1992–93 Formal commitment made by Societé Burkinabé des Fibres et des Textiles (SOFITEX), the national cotton parastatal company, to let producers’ representatives participate in reform debate. Contrat-Plan Etat SOFITEX, in which the state committed not to interfere with management of SOFITEX, established a plan to streamline accumulated debts of producers and the parastatal. 1994 Laws pertaining to establishment of farmer groups amended. 1996–99 Free membership introduced in formation of groups of local cotton farmers; the groupements villageois were replaced by market-oriented organizations (groupements de producteurs de coton) at the subvillage level, with implementation of new local governance rules. 1996–2001 National cotton union (Union Nationale des Producteurs du Coton du Burkina, or UNPCB) is established with support of l’Agence Française de Développement, the Burkinabe government, and SOFITEX, based on membership of local groups and their integration into regional unions. 1998 Accord Interprofessionnel signed by SOFITEX, the state, UNPCB, donors, and financial consortium (Caisse Nationale du Crédit Agricole; Banque Internationale pour le Commerce, l’Industrie, et l’Agriculture; Banque Internationale du Burkina), replacing the Contrat-plan and defining the reallocation of responsibilities. 1999 State partially withdraws from the sector through partial privatization of SOFITEX; half of government’s share in SOFITEX is transferred to UNPCB. 2000–06 Economic activities—including provision of cereal input credit; management assistance of cotton groups; and participation in quality grading, financial management, and price bargaining—progressively delegated from SOFITEX and the government to UNPCB. The state downsizes support of research and extension services. 2002–06 New players—including private input providers, new regional private cotton monopsonies (SOCOMA, FASOCOTON), and private transport companies—begin operating in the sector. 2004–06 Interprofessional association (AICB) is established with cooperation of cotton farmers, banks, private stakeholders, the government, and research institutes. Association of cotton firms (APROCOB) established to interact with UNPCB. 2006–08 Price-setting mechanism changed to better reflect world price levels; new smoothing fund managed by an independent organization becomes operational in 2008. Source: Kaminski, Headey, and Bernard 2009. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 109 A partial equilibrium analysis based on estimation of coun- Figure 6.2 Seed Cotton Production in Selected terfactuals suggests that cotton reforms explain two-thirds Sub-Saharan African Countries, 1980–2010 of the threefold increase in production from 1996 to 2006, 700 Average annual production of cotton which enabled farmers to cultivate more land, raise their seed (thousands of metric tons) incomes, and improve food security (Kaminski, Headey, 600 and Bernard 2009). Figure 6.2 compares average production 500 of seed cotton in Burkina Faso and neighboring West 400 African countries with that in the Southern African coun- tries of Zambia and Zimbabwe, in which conventional pri- 300 vatization and reforms were pursued (table 6.2). As shown 200 in figure 6.2, the volume of cotton production growth in Burkina Faso has been substantially higher than in com- 100 parator countries since 2000. 0 Importantly, increased cotton production in Burkina 4 9 4 9 4 6 8 0 98 98 99 99 00 00 00 01 –1 –1 –1 –1 –2 –2 –2 –2 Faso in recent years was not based on significant productiv- 80 85 90 95 00 05 07 09 19 19 19 19 20 20 20 20 ity increases but rather on accumulation of factors such as Benin Burkina Faso Côte d'Ivoire land, labor, and inputs. Rising cotton yields also hide the negative dynamic effect resulting from the entry of less able Mali Zambia Zimbabwe farmers and land in cotton production. Hence, stagnant Source: Kaminski, Headey, and Bernard 2009. (even slightly increasing) yields mean that individual per- formance increased over the last decade. Overall competi- tiveness of cotton farmers and of the cotton industry adopting reforms using a cautious, piecemeal approach increased much more rapidly than in neighboring countries (Kaminski, Headey, and Bernard 2009). The approach dif- and has been hampered only by low world market prices fers substantially from the conventional approaches of (which include the depreciating effect of U.S. cotton subsi- other countries, which ignored the specificities of the dies) and the appreciation of the euro. institutional set-up (Jayne and others 1997). Reinforcing Nevertheless, cotton-led growth faces several challenges. the institutional framework has ensured better market Kaminski, Headey, and Bernard (2009) point out the lack of coordination along the value chain and higher levels of sustainability of cotton reforms as well as the limitations of contract self-enforcement, the main bottleneck to better the factor-driven Burkinabe economy, especially with regard performance of cotton industries in the region.3 The to the underlying environmental constraints to extensive reforms in Burkina Faso also countermanded several gov- agricultural growth and the need for the intensification of ernment interventions and policies that had directly con- sustainable agricultural. The cotton sector also faces envi- tributed to inefficiency, while reallocation of activities ronmental challenges, including the threat of soil impover- among the new institutions and tightened vertical coordi- ishment and emerging conflicts regarding access to land nation improved efficiency in the cotton sector. (Gray and Kevane 2001). These problems notwithstanding, Burkina Faso’s success in creating an efficient cotton the reform has led to upgraded institutional arrangements value chain stems from the institutional capacity for between producers and other stakeholders and helped farm- improved contractual coordination and collective action— ers organize in more professional-oriented associations to something that was achieved through the creation of pro- carry out a growing number of responsibilities (World Bank fessional (under free membership principles) cooperatives 2004). All these factors improved the overall management of of cotton growers, which substantially improved cotton the sector. The reform has resulted in higher profit-sharing marketing and input credit repayment, yielding significant for farmers, better extension services, fairer quality grading, operational cost savings. The improved institutional frame- and better access to agricultural inputs with improved credit work has allowed cotton firms to provide better-quality repayment mechanisms (Kaminski 2007). technical extension services and research in the course of The Burkinabe reform model is unique in Sub-Saharan sectoral privatization and ensured coordination of the Africa in that it addresses government failures and local delivery of public goods, quality control, picking and gin- realities within the current institutional framework, ning, and marketing activities. At the same time, access to 110 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH Table 6.2 Reforms and Outcomes in Benin and Mali, 1992–2010 Benin Mali Period Development Outcome Development Outcome 1992–99 Supervised liberalization of input Production increased Traditional integrated commodity Production increased distribution; progressive by more than chain experienced financial by more than establishment of professional and 50 percent troubles and embezzlement 50 percent interprofessional associations 2000–05 Quasi-competition and Slight increase in Producer boycott, opening of reform Stagnation, with periodic dismantlement of parastatal, production; dialogue; establishment of cotton declines as result of emergence of several farmers’ problems in input cooperatives (but limited adoption farmers’ collective networks, establishment of financial recovery de facto) action clearinghouse, disappearance of governance former professional associations; institutional failures 2006–10 Move toward a private monopolistic Yields and production Reform process under way, with Production plummeted organization with new public decreased, privatization and convergence as a result of delays in regulation and new institutional stabilizing only toward zoning approach (splitting and uncertainties framework for public-private recently of parastatal into four private surrounding reform; partnership territorial entities linked through farm profitability newly established interprofessional declined body) Source: Kaminski, Headey, and Bernard 2009. finance and inputs has been eased for smallholder farmers, strategies, most of the income vulnerability of cotton could and transportation services are now provided by the private be removed. sector. Figures 6.3 and 6.4 present the main organizational Although the Burkinabe cotton story is largely one of changes. removing institutional constraints, further reforms and new In the long term, mitigating vulnerability requires that policy frameworks are needed. The growth in cotton areas the main stakeholders improve their risk-management in the early 2000s reflects the lack of alternative (crop) solu- strategies and that the government increase the quality of tions, combined with a strong comparative advantage for agricultural services provided to farmers (research, exten- cotton growing in a very constrained market environment.4 sion, and quality grading). There is also a need to invest in Farmers need a secure access channel to output markets, human capital in a number of areas. To deal with world mar- because marketing is typically too costly for them to do ket volatility, the new price-determination mechanism— on their own—or they lack the human capital to do so. which has suffered from lack of adjustment—has been Because of market failures (incomplete or missing input supplemented by an independently managed smoothing and rural credit markets), Burkinabe farmers must estab- fund. Price and weather risk-management instruments are lish contractual arrangements with downstream stakehold- currently in development. Another risk-mitigation strategy ers (processors and traders) to access inputs and credit.5 is to moderate U.S. dollar–euro exchange rate volatility Effective contractual arrangements involve a high degree of (using, for example, the holistic approach proposed by the market coordination and trust-building relationships. At World Bank 2009a). Input price risk is another substantial the same time, Burkina Faso’s challenging business environ- risk faced by cotton sector participants. In response, stake- ment and poor overall institutional framework discourages holders are exploring the possibility of establishing an input entrepreneurs from investing in agrifood or large-scale fund that could lead to as much as 20 percent savings on trade because of unmanageable risks and hidden costs. This input procurement costs. Last, stakeholders are working to lack of investment limits the scope for developing other provide a more adequate set of microfinancial and micro- high-value commodity chains, even where the potential insurance instruments to farmers to stimulate small-scale exists. The unsatisfactory outcome is that most tradable agri- farm investment and protect against production risks, cultural output markets are tied to the cotton sector for including those related to weather. With the sustainable accessing inputs and credit, increasing the vulnerability of management of smoothing schemes and risk-mitigation rural income to changes in the cotton market. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 111 Figure 6.3 Prereform Organizational Model of the Cotton Sector in Burkina Faso GV Cotton producers Input Cotton credit seed Other farmers SOFITEX Cotton fibers Cotton seed purchases, ginning, oil mills, purchases and distribution of agricultural inputs, Farmers unions and faitières Imported inputs input credit International cooperation (WB, AFD) Banks DAGRIS Government Source: Kaminski 2008. Note: GV are village groups; AFD is Agence Française de Développement; DAGRIS is Développement des Agro-Industries du Sud (DAGRIS), a French parastatal involved in the agroindustrial sectors of many French-speaking countries. Figure 6.4 Postreform Organizational Structure of the Cotton Sector in Burkina Faso in 2007 Interprofessional agreement 30% of SOFITEX Government Cotton companies UNPCB Banks 35% SOFITEX SOCOMA FASOCOTON Cotton Credit committees unions 34% 100% 100% DAGRIS Reinhardt GPC Research Technical agents Cereal inputs Management assistance for GPCs and cotton unions Cotton inputs Technical assistance Cotton producers Source: Kaminski 2008. Note: SOCOMA and FASOCOTON are the regional cotton firms that were granted regional monopolies in the more marginal center and eastern cotton- producing regions. UNPCB is Union Nationale des Producteurs de Coton du Burkina Faso; GPC is Groupement de Producteurs de Coton . DAGRIS is no longer operating in Burkina Faso after its withdrawal from SOFITEX shareholding in 2008. The sector then went through financial restructuring with state recapital- ization, which resulted in a substantial reduction of producers’ shares in cotton companies’ capital (from 30 percent to around 10 percent for SOFITEX). 112 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH DRIVERS OF COTTON GROWTH; MARKET within the commodity chain while the state withdrew from AND INSTITUTIONAL ADVANTAGES most of its activities. Although most of the other key sectors of the economy—the agricultural input supply chain, energy Structural characteristics of rural markets partly explain and other public utilities, and a portion of the financial why the Burkinabe economy has remained so dependent on sector—remained at least partially state controlled until the its cotton sector. Historically, the cotton sector has been early 2000s, they have been progressively privatized and lib- better supported by institutions, stakeholders, and infra- eralized in the years since. All banks are now privately con- structure than cereal sectors, which face enormous logistical trolled, as are some public utilities and input markets. This constraints and high transport costs. The cotton sector in completed the privatization and liberalization process that Burkina Faso has benefited from better social organizations occurred in the trade, transport, and cereals sectors. The evo- of farmers, better public and private institutions, and better lution of organizational structures in the cotton sector also infrastructure, ensuring better-functioning markets (Bassett has created better incentives for farmers and manufacturers 2001). Infrastructure is funded by cotton revenues through and allowed them to benefit from profit-sharing. Apart from private (local public goods) and public investments. Since these improvements, stakeholders in other agricultural sec- the 1980s investments in the cotton sector have also fostered tors had to incur more economic risks in the course of liber- technology adoption, thanks to high-quality research, coop- alization (in terms of the profitability of different marketing eration between French and local researchers, farmer organ- channels, processing, and retailing activities), although they izations, and top-down implementation policies. stand to reap potentially higher financial margins. Technological progress in agricultural sectors is more Cotton producer organizations have benefited rural civil likely to diffuse when land becomes increasingly scarce society and revitalized rural communities in Burkina Faso (Boserup 1965). Constraints in the farmer environment and neighboring countries (SWAC Secretariat and OECD (liquidity and market failures) hamper the adoption of 2007). Such organizations help farmers establish profes- technology, notably for the poorest households that seek sional structures and participate as political actors. Gover- to meet their food-security objective and minimize pro- nance structure and management capacities are critical to duction risks (Abdoulaye and Lowenberg-DeBoer 2000). this success, in addition to local social conservatism and Fertilizer use is constrained by lack of experience, access to other social norms at the village level. As Bernard and others inputs and capital markets, and learning (Abdoulaye and (2008) show, community-oriented organizations are often Sanders 2005).6 These problems suggest the need for more captured by traditional authorities and local elites; in con- extension services, quality control, and research to foster trast, when elaborate administrative rules are established, fertilizer use on higher-value crops and efforts to reduce market-oriented organizations are much more democra- the risk associated with fertilizer use through training and tized and efficient. In the cotton sector, market-based quality standards for lower-value crops. As a result of cooperatives of cotton farmers have been established and demographic pressure in and migration into cotton areas, producers have become significant players in policy making as well as the relatively high value of cotton compared to and the public debate. In other rural sectors, by compari- other crops, cotton has attracted increasing levels of tech- son, institutional constraints remain more or less unchanged nology, capital, and inputs. at the farmer level. The advantages of growing cotton (rather than other Importantly, cotton provides better risk-management crops) include better access to rural finance and agricul- arrangements to producers than do other sectors, especially tural inputs, thanks to interlinked transactions with with regard to price, outlets, and marketing. Cotton pro- powerful agribusinesses (contract farming and other con- ducer prices are administratively determined at the begin- tractual arrangements) and the ability to overcome inher- ning of the cropping season and the pricing formula allows ent market failures in the realms of input and rural credit for ristournes (after-harvest bonus payments) to farmers. In markets. The strong position of these agribusinesses addition, cotton farmers have better access to market infor- ensures contract self-enforcement (in that it prevents sig- mation than farmers in other sectors, thanks to their effec- nificant room for side-selling). This is not the case in the tive organizations and national union, expansion of local cereal sector, which is characterized by small traders who infrastructure, and increased use of mobile phones in lack physical capital. cotton-producing areas. Better information sharing can also In Burkina Faso, most reforms in the cotton sector help agribusiness entrepreneurs provide insurance schemes addressed government failures and institutional weaknesses by to farmers. building the capacity of producers and upgrading institutions CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 113 Despite several shortcomings (including the weakness of price fell in the late 2000s. As a result, smallholders’ prof- the farmers’ union, corruption, and adjustment lags in the itability is threatened and trust in the newly established price determination mechanism), the revamped cotton arrangements and partnerships declined from 2007 to policy framework in Burkina Faso has contributed to an 2009. Indeed, due to a misapplication of the pricing for- increased concentration of cotton exports and dependency mula in 2006–07, the smoothing fund was depleted and on cotton production at both the national and farm levels. farm gate prices were adjusted downward in 2007–08 Consensus building was forged by top-down establishment once the formula started to be rigorously applied. In addi- of the cotton union, state authoritarianism, trust-building tion to growing deficits, SOFITEX had trouble paying with foreign partners, and accumulation of social capital at farmers and delivering inputs to them on time. Some the village level (Kaminski and Bambio 2009). Consensus observers point to management problems in SOFITEX building followed by institutional innovations built part- once DAGRIS left the shareholding and the company was nerships along the supply chain with the professionalizing recapitalized by the state. But the recent spike in cotton of farmers’ groups and reinforced their bargaining power. prices, even in the face of higher marketing costs attribut- As a result of upgraded institutional and policy frame- able to the political crisis in Côte d’Ivoire, means that works, profit-sharing has evolved in favor of producers, financial prospects are optimistic in the mid-run. The while the management, input procurement and cost recov- expectation of good financial prospects is all the more ery, and overall market coordination of the sector has true since the price-determination mechanism has been become more efficient, even in the face of the decline in the in effect (and rigorously applied) over the past five produc- world price of cotton and the appreciation of the euro tion cycles. Burkinabe farmers’ profitability is tightly linked against the U.S. dollar. A larger profit share for farmers to world prices—with predefined pricing at the beginning reinforced production incentives once coordination was of the crop season, they receive roughly 60 percent of the improved, contributing to the impressive cotton growth world price of cotton for their crops. pattern observed in the 2000s (figure 6.5). In sum, while growth strategies aimed at economic diver- Aside from these positive outcomes, the state bailout of sification should be pursued within an adequate policy SOFITEX and the lack of political sustainability of the framework, governance, management, and institutional cotton policy framework have reduced sectoral efficiency improvements are still required in the cotton sector if it is to and led to failures in market coordination once the world remain a sustainable pillar of the Burkinabe economy. Figure 6.5 Farm Gate and World Price Indexes of Seed Cotton, 1990–2009 14 12 10 8 6 4 2 0 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 Burkina Faso farm gate price index, 10 = 2000 Liverpool price index, 10 = 2000 Source: Author, based on monthly data on international prices and exchange rates from data construction from IMF 2010 and Burkina Faso national sources. Note: The Liverpool cotton price index is the index of world cotton prices. All series were calculated from averages of monthly data covering the period February–January, the production year for Burkina Faso’s cotton sector. 114 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH COMMODITY DEPENDENCE AND ECONOMIC tend to vanish, and ensuring a minimum amount of agri- VULNERABILITY cultural inputs to the noncotton agricultural sectors. The cotton-led growth of Burkina Faso has been nega- Over the past decade, GDP growth in West African coun- tively affected by volatile world cotton prices (competing tries has been derived largely from agriculture (export crops synthetic fibers, cotton subsidies in the United States and in coastal countries and livestock in landlocked countries) Europe) and the appreciation of the euro (World Bank and mining. As many economists, including Hausman, 2009a).7 Moreover, increases in production have been driven Hwang, and Rodrik (2007), note, export sectors are key to mainly by increases in factors of production (labor, land, economic growth, as they typically have a cascade of effects inputs) rather than through increases in total factor produc- on other sectors. However, export earnings cannot compen- tivity. Economic vulnerability from commodity dependence sate for weak internal conditions for sustained growth. has thus been exacerbated by the nonintensive pattern of The problematic issue here is commodity dependence. growth, which has had a limited effect on overall economic Because of few absolute advantages, poor countries such as transformation and limited spillovers through domestic Burkina Faso often experience unbalanced growth pat- demand-led growth. Due to environmental and demo- terns driven by a restricted number of export commodities graphic limitations, this growth path looks to be unsustain- traded on very volatile world markets. Furthermore, these able in the long run. In general, the Burkinabe economy is commodity markets are characterized by distortive poli- still highly agriculture-oriented with a growing importance cies and barriers to entry or participation for farmers in of cotton during the late 2000s and increasing importance of developed countries. In addition, production is subject the mining sector. Figure 6.6 compares the economic growth to climatic variability (droughts and floods) and prob- of each economic sector over the past decade. lems of soil fertility. Relying on a few commodities with Economic diversification is one of Burkina Faso’s uncertain profitability leads to economic vulnerability national priorities, as stressed in recent official strategic and calls for an in-depth examination of viable economic documents. Of particular importance is diversification in alternatives. the rural sector toward higher-value agricultural exports, Globally, the outlook for cotton profitability is uncertain such as fruits and vegetables. In the medium term, diversifi- because of enduring institutional and governance limita- cation to manufactured products does not appear necessary tions (despite recent improvements in both areas) and for growth. In contrast, increases in the production of volatile world markets. In addition, cotton sectors some- higher-value natural-resource-based products are required. times still serve political ends—for example, offering good Such a shift may raise income levels by developing capacities prices to producers before elections, even if these offers Figure 6.6 Real GDP Average Growth, by Economic Sector, 1994–2008 16 14 12 10 Percent 8 6 4 2 0 re g g w and n om d e es s ck ice in rin ad io lec n vic tu to te rt a in er ct Tr rv tu ul , s M as er ru at ve se ac ric o ,g ls st sp Li uf Ag er on ty cia an an th ici C an Tr M O tr Fin ec El Source: World Bank, CEM Burkina Faso, and IMF 2009. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 115 for manufactured goods together with the capacities of diversification by establishing facilitating conditions. This marketing and exporting a diversified basket of primary process involves general policies aimed at building insti- products. The application of technology to increase cotton tutions, providing necessary public goods, and loosening profitability should also be encouraged. economic constraints in areas such as public administra- Although growth of the cotton sector has been accom- tion, market imperfections and transaction costs, fiscal and panied by some amount of economic diversification, that trade policies, and weak governance. Such policies should be diversification has not been sufficient to stimulate struc- combined with sectoral policies according to absolute and tural transformation through macroeconomic spillovers. comparative identified advantages, emerging champions Although cotton provides skills to the other agricultural (World Bank 2009a), and the structural role of key eco- sectors, the growth of the cotton sector has not brought nomic sectors (Hirschman 1958). Hence, diversification and improvement in labor and land productivity and has not specialization may exist side by side because of segmenta- enabled sufficient redistribution of public goods such as tion of the economy. This is the case for most West African education and health. In addition, spillovers of cotton countries, where there is a lack of economic integration activities to other economic sectors in Burkina Faso are between the primary sector (agriculture and mining) and restricted because of the limited industrial nature of cotton the industrial sector. growing and its limited size in the rural sector (not more The most important challenge for spurring sustainable than 20 percent). A natural way to encourage spillovers growth in Burkina Faso lies in increasing total factor pro- would be to scale up textile activities. Finally, it is impor- ductivity, which depends to a large extent on the investment tant to recognize that the increase in demand for agricul- climate. Policies that make it easier to do business, reduce tural inputs from cotton growers themselves has not been the burden borne by entrepreneurs, and upgrade the insti- met by corresponding increases in local supply. Most agri- tutional framework are worth considering. From an invest- cultural inputs and nonfood goods are produced abroad, ment climate perspective, Burkina Faso has improved limiting transmission of demand increases in the rest of the significantly over the past decade, placing it in the economy. midrange of African countries in the World Bank’s Doing Business rankings (World Bank 2009b) and other indexes. Yet important bottlenecks remain, such as lack of access to Constraints to economic diversification finance and electricity, high tax rates, unfair (on fiscal A number of factors constrain the diversification of Burkina grounds) practices in the informal sector, corruption, Faso’s economy.8 Inefficient and incomplete policy reforms, inadequate transportation, lack of access to land, ineffec- institutions, and investments (in infrastructure, capacity- tive bureaucracy, lack of technology, inadequate health building, and public goods, for example) are all hindrances services, and lack of an educated labor force. These con- to diversification. Yet diversification is clearly a necessity for straints raise labor costs, limiting producers’ competitive- more sustainable and balanced growth. Macroeconometric ness. Large firms in Burkina Faso, however, are roughly as studies (Berthelemy 2005) show that diversification indica- productive as smaller ones (World Bank 2009a), implying tors positively influence GDP growth rates, notably that differences in labor productivity are not necessarily a through increases in total factor productivity. matter of scale efficiency but rather rely on higher capital In general, there is a U-shaped relationship between eco- intensity for larger firms. Indirect and “invisible” costs nomic concentration in exports (the Herfindahl index) and account for an estimated 8.2 percent of firms’ sales in per capita income (World Bank 2009a). In the low part of Burkina Faso (table 6.3). the curve (where the relationship is negative) stand low- income countries such as Burkina Faso, characterized by Constraints to technology adoption and diffusion export earnings derived mainly from agriculture (Imbs and Wacziarg 2003). The key role of diversification is to smooth Promising technologies exist in the rural sector, but low export earnings derived from volatile (agricultural and levels of technology adoption and diffusion hamper the resources) world markets in developing countries (Massel process of diversification.9 Several underlying constraints 1970), which fosters economic structural changes and con- are at play. First, diffusion of a new technique or technol- solidation. A more sustainable and balanced growth path ogy requires complementary techniques and inputs, for can be attained by economic diversification and a set of which extension services and sustainable arrangements to policies and institutions that accompanies the process of ensure the provision of inputs are needed. Given the market 116 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH Table 6.3 Indirect and Invisible Costs as Share of Firms’ Sales, Selected Sub-Saharan African Countries (in percent) Invisible costs Cost of Indirect costs Cost of corruption Total Losses manager’s (informal indirect due to time to payments to Cost of and power deal with get things security invisible Country/year Transport Telecommunications Customs outages regulations done) measures costs Burkina Faso 2006 2.23 0.70 1.63 1.20 0.05 1.48 0.86 8.20 Benin 2004 — — 0.48 1.16 0.06 4.27 0.57 6.55 Mali 2007 1.96 0.86 0.26 1.39 0.05 0.24 0.12 4.60 Uganda 2006 1.08 0.50 0.52 28.16 0.06 2.18 0.43 32.93 Zambia 2007 0.61 0.71 0.68 1.79 0.04 0.05 0.93 4.81 Source: World Bank Enterprise Surveys, 2004–07. Notes: Estimates are based on at least 15 observations. — = not available. failures affecting rural credit and agricultural inputs in of cereal profit margins is significant, meaning variability at Burkina Faso, it is understandable that technology diffu- producers’ gate prices when combined with low market sion may be constrained. integration. Public investments in market infrastructures Apart from institutional and market constraints, technol- have been low, and only some rural cotton producing areas ogy diffusion and adoption need a clear and consistent pol- have been unlocked thanks to cotton benefits and the inter- icy framework for new marketing strategies, risk-mitigation ests of agribusinesses. Not surprisingly, the most integrated strategies, or export promotion set-ups to be sustainable cereal markets are the ones located close to the most pro- and provide the necessary incentives for long-term invest- ductive cotton markets, because they benefit from better ment. In the rural sector, Abdoulaye and Sanders (2006) infrastructures and institutions. have shown that new marketing strategies (such as the After the liberalization of cereal markets in Burkina Faso, Intsormil project10) do not ensure sufficient investment many small-scale traders coexisting with larger private incentives if food security policies are inconsistent and do traders have replaced government marketing. However, not allow a greater number of farmers to derive more prof- because of information asymmetries with farmers and the its during adverse years (removing export quotas, for lack of access to reliable information services, traders have a instance). Technology adoption also relies on comprehen- comparative advantage and better information regarding sive exchange rate and trade reforms. In the manufacturing local and central markets, retailers, and processing facilities. and services sectors, policy frameworks are also critical. The Even if this has created competitive emulation, there has needed reforms combined with long-term commitment of been an associated lack of investment in human capital and the public sector in export promotion and investment innovation, few long-term contractual relationships with strategies should, in turn, facilitate technology diffusion and agribusinesses, and no significant change in access to finan- diversification in higher-value production and exports. cial markets. Due to several market failures and institutional weaknesses (contract enforcement), informal traders rely on social networks, rendering competition imperfect (Barrett Market constraints: Missing and incomplete markets 1997). In addition, Burkina Faso’s network economy— in the rural sector retention of profits among insiders—tends to limit business Agricultural producers face missing and incomplete mar- expansion, as noted by Badiane and others (1997). kets for their goods. This is especially true for cereal pro- Not only do producers face constraints in output mar- ducers. Cereal markets are still characterized by high kets, which slows down diversification, but they also face transaction costs, ineffective institutions, and discrimina- constraints with respect to input markets. Hence, the poten- tory policies, maybe because of their low value to weight tial for income-enhancing diversification of the rural sector ratio and logistical disadvantages (information access, is limited. Input market imperfections result from struc- inland freight cost). This notably contrasts with the dynam- tural deficiencies such as high transaction costs, liquidity ics of cotton markets. Hence, the transport cost component constraints of producers, and asymmetric information. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 117 Lack of well-maintained and high-quality infrastructure such as agricultural or industrial innovation and quality also impedes market integration in Burkina Faso, translat- grading. ing into higher market risks borne by producers and traders, In Burkina Faso organizational structures rely on formal together with capacity constraints (risk-management and and informal institutions to coordinate the decisions of storage).11 Cost-effective ways to provide infrastructure major players and to regulate the industry. Formal structures may use existing infrastructures with the involvement of include state rules and formal regulations or collective organ- user communities (mainly for maintenance with own- izations such as producers’ associations, partnerships, and managed funds). Productivity benefits could be achieved by farmer unions. The evolution of the regulatory framework in improving off-road transport and intermediate means of Burkina Faso can be qualified as uneven. Overregulation in transport with capital-savings techniques for road construc- several sectors has been quite counterproductive and has tion or using labor-based techniques to overcome usual notably contributed to restrictive environments, such as in problems related to equipment use and availability in the the trade sector. Despite several improvements, there is a per- region. Soft infrastructure, such as information, is also in sistent lack of confidence in formal institutions, including the short supply. Diversification from cotton to other agricul- legal system. There are no effective institutions dedicated to tural commodities needs to rely on better market informa- investment and export promotion, indicating the absence of tion services and farmer literacy to foster adoption of more a long-term strategic vision for economic development and profitable crops and techniques. diversification. In contrast, standardization and quality pro- motion systems are being developed, and business associa- tions are beginning to adopt quality control systems within Institutional constraints their marketing systems. Continued improvements in these A country’s institutional environment is composed of areas would benefit the whole set of private investors and market and nonmarket relationships between different business expansion. stakeholders. The role of institutions is all the more cru- cial in Burkina Faso given that markets are subject to The institutional environment. Because formal insti- structural deficiencies and constraints for participants. tutions in Burkina Faso face a lack of credibility, many Nonmarket institutions define social rules and norms producers and private stakeholders rely on informal insti- that, in turn, overcome the other market and institutional tutions, social networks, and personal relationships. How- imperfections. But informal institutions, such as social ever, this poses a problem, since the prevalence of informal networks and personal relationships, have been shown to institutions weakens the credence of the formal ones. be less efficient than formal ones, and they severely Informal village norms—such as those for credit, risk, or restrict the degree of competition, information, and busi- access to land—also enable local actors to overcome several ness expansion. market constraints. In matters of land access, however, informal norms often entail exclusion of some ethnic Supply chains and vertical relationships (regula- groups or difficult access to the fertile or accessible land for tory, informal institutions, contract enforcement, some societal groups. Political and ethnic biases in the allo- information asymmetries). In the formal realm, the cation of land imply allocation inefficiencies. Informal and institutional structure of many supply chains was deeply traditional risk-sharing arrangements provide only imper- reconfigured in the course of structural adjustment plans fect assurance of low effectiveness. and sectoral reforms. Vertical relationships now entail The decentralization of rural development policies in more specific arrangements between stakeholders and Burkina Faso, together with more participatory approaches, more coordinated provision of public goods. Stakehold- has contributed to a democratization process whereby ers face more options as marketing channels may be responsibilities have been gradually transferred from potentially beneficial for them if stakeholders can national institutions to newly established local institutions. improve their capacities (information, management, bar- However, capacity constraints are still very strong. gaining, storage, infrastructures). However, competition To improve the business environment, national stake- in several sectors can be detrimental because of the per- holders in Burkina Faso have begun exploring various sistence of other market failures (Lipsey and Lancaster options: public-private partnerships, business associations 1956) and the fact that it can lead to coordination prob- and interprofessional bodies, information services, and lems regarding the provision of certain public goods, consensus-building institutions. While all of these options 118 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH have the potential to improve vertical and contractual rela- cesses can be attributed to a consistent policy framework tionships within supply and value chains, the experience of in which the state gained credibility with its economic building the interprofessional association in the cotton sec- partners, farmers’ groups were strengthened, and private tor was especially encouraging in that it provided a credible arrangements (such as contract farming) were not hin- institutional framework to the main stakeholders and dered by second-generation controls. Policy inconsistency ensured market coordination while being limited by the remains in the cereal sector, however, in the form of food predominant position of SOFITEX, which can sometimes security policies (lowering agricultural profits of the most impose some decisions on collective management issues efficient farmers), protectionist trade policies (lowering (quality grading, input distribution system or regulation, incentives for diversified exports), and restrictions on and so on). input markets. Fortunately, these short-lived interven- tions that were common in the past are increasingly less Risk management. Along with being a market constraint, evident. If needed, they should be decided upon harmo- risk is an institutional constraint. Inefficiencies in the cur- niously at the regional level, one of the purposes of the rent institutional arrangements in Burkina Faso require Comprehensive Africa Agriculture Development Pro- mitigation of business and production risks as well as eco- gramme (CAADP), an initiative of the New Partnership nomic ones. For example, minimizing production risks for for Africa’s Development (NEPAD) that calls for more all agricultural commodities and livestock requires a better harmonized agricultural and trade policies at the regional institutional framework. The development of new level in West Africa. microinsurance schemes may help farmers facing external Sufficient agricultural extension services remain a chal- shocks to production. Indeed, improved transmission of lenge in Burkina Faso. Currently, there is a lack of means, information about weather conditions would help in the training, and funding for extension services, along with lack development of weather index microinsurance schemes of involvement by the state. (as already achieved in Southern Africa), providing farmers with producers’ incentives. Another innovative insurance Political economy, social contracts, and the urban scheme that could be applicable is parametric insurance. bias. The political economy of agricultural policies Although the development of markets in innovative insur- explains the persistence of contradictory policies and ance schemes requires a strong financial framework to be incomplete market reforms. Compared to neighboring established, enhancing information access through new countries, Burkina Faso stands between the situation of technologies for information and communication could Mali—steady adherence and commitment to market help spread these schemes. reforms—and that of Benin—de jure reforms and de facto state control with regard to noncotton sectoral policies (the reverse with regard to cotton policies). Private incentives in Political constraints Burkina Faso have been jeopardized by entrepreneurs’ fear Constraints to economic diversification also include the of government intervention and by the somewhat volatile ones that come from policy making and that contribute political environment. Fewer private sector opportunities to market and institutional failures. Most of them can create scope for a larger government involvement in mar- be classified as government failures: public governance kets, which, in turn, fuels this pattern. Therefore, political problems, transparency, democratization, nonbenevolence, economy conditions have contributed to constraining agri- corruption, and lack of foresight. Despite substantial cultural diversification while favoring the cotton sector. improvements in the political environment in Burkina Faso These political conditions also apply to the services and over the past decade, enduring political practices hamper industrial sectors in Burkina Faso. the effectiveness of upgraded policies and the questions of democratization and authoritarianism remain. The political environment. Despite several improve- ments, the political environment in Burkina Faso contin- Commodity reforms and the political economy of ues to present constraints in the form of enduring cotton-cereal reforms. Although commodity reforms political practices resulting in weak or poor governance, in many Sub-Saharan African countries have been the collusion of interests between officials and executives, and subject of criticisms (Akiyama and others 2001), Burkina lack of political and democratic accountability of political Faso has been viewed as a positive experience. These suc- leaders. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 119 CONCLUSION constraints described in this study in Burkina Faso are still of crucial importance. This calls for a policy-led economic From an overarching perspective, cotton continues to transformation induced by structural changes, a consistent account for a significant share of exports in Burkina Faso’s policy framework, and institutional innovations. Innova- very concentrated economy. Future global demand for cot- tion needs to take place in a range of sectors: agriculture and ton, however, does not appear consistent with an economic livestock, mining, trade, tourism, and the food industry. strategy in Burkina Faso that relies heavily on cotton In the mining sector, the Burkinabe government has exports as a driver for growth, not least because world cot- recently established a mining law to improve incentives for ton demand is expected to remain at volatile levels because cost-effectiveness, export promotion, and for informal of imperfect substitution among textile fibers and rising oil orpailleurs. In the agricultural sector, income-enhancing prices (oil is used in the production of synthetic fibers). But options exist and are achievable but would not lead to sig- in the short run, there is a need to maximize cotton rev- nificant changes in the current environment of smallholders enues and benefit from the historic price swings on the and traders, partly because of enduring market imperfec- world markets. tions, structural deficiencies, and macroeconomic instabili- A useful approach for Burkina Faso would be to exam- ties. Higher-value production is also of potential interest ine other sectors with significant export growth potential, and has already driven growth and diversification in India such as gold and livestock. Ramping up exports in these and in several countries in Eastern Africa (for example, in sectors may reduce dependence on cotton incomes. horticulture, fruits, and nuts). A strategic set of sequenced Expanded production and export of other agricultural actions could be led by officials and local leaders.12 products (vegetables, fruits, poultry, and rice) and nonpri- Another area in which Burkina Faso has space for mary commodities may also serve to reduce the economic reform is trade. Currently, the trade environment in the concentration on cotton. country is characterized by high tariffs and low market Even with the potential expansion of other sectors, cot- integration. Burkina Faso would also benefit from diversi- ton is likely to continue providing livelihoods for a signif- fying its primary products, the production of which is con- icant share of Burkina Faso’s rural population and to strained by the scope of development of manufacturing remain important in the national economy. The most industries (for example, the food industry). Reasons typi- urgent concern is to tackle the issue of “sustainable inten- cally invoked for the limited scope of manufacturing in the sification” at the farm gate, together with increasing the country include its abundance of land, scarcity of skilled cost-effectiveness of cotton firms. Overall, there is a need labor, insufficient or poor-quality infrastructure, lack of to consolidate the cotton sector in the course of ongoing policy reform, high transaction costs, and mismanaged reforms and institution building, and to initiate a signifi- or uninsured risks (Collier 1998, 2002; Habiyaremye and cant process of economic diversification through the for- Ziesemer 2006; Collier and Gunning 1999). A consistent mulation and implementation of growth-enhancing policy framework toward structural transformation should policies aiming to overcome the most relevant con- deal with these constraints. straints. Nonmarket factors, such as the quality and effec- Several of the issues raised in this study are already being tiveness of institutions, are also at play, meaning that addressed by officials and foreign donors in policy formula- improvements in the business environment and invest- tion and project implementation. Stylized facts, in fact, ment climate would not automatically lead to a scale-up of show that policy makers have been successful in bringing nontraditional export industries as a prerequisite for about change in several areas: structural transformation of the economy. While policies are moving in the right direction— generalization of GM Bt cotton (genetically modified cot- ■ Established priorities: food security, poverty reduction, ton with Bacillus thuringiensis inserted into the seeds) and private investments, and public-private partnerships in diffusion of improved agricultural productivity practices infrastructure and human capital. (such as irrigation) and soil conservation, improved busi- ■ Policies for emerging and developing economic sectors ness environment, implementation of more adequate risk- ■ On-going reforms management instruments and the new accelerated growth ■ Strategic action plans with donors: special economic strategy, explicit support to growth poles as a means to zones and planning for employment, competitiveness, create faster growth and diversification—removing the and diversification of economic growth poles 120 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH ■ Capacity to improve the institutional framework (for weights used to construct the urban CPI (Grimm and example, the cotton reform) even in the face of external Gunther 2004). shocks (declining world cotton prices and economic and 3. Market coordination is essential to ensure the delivery political crisis in neighboring Côte d’Ivoire) of crucial public goods (inputs and credit) in an incomplete ■ Accumulated experience marker environment. More competition broadens the scope ■ Government capacity to react to shocks and resistance to for strategic defaulting on input credit and side-selling economic turmoil while maintaining the pace of reforms. (selling outside the contractual arrangement to another company) and reduces the incentives to provide public goods, such as research or extension services. The outlook for Burkina Faso—albeit promising—could 4. This phenomenon explains the rise in the Herfindahl be made more secure with scaled-up investments in crucial index from 0.2 to 0.6. sectors, and deepening market and political reforms. As 5. Cotton inputs are provided by ginning companies, such, the road is half-traveled. Burkina Faso is still at the which also directly purchase and gin seed cotton; inputs stage of establishing policy and institutional frameworks. have been subsidized lately by the government. Governance has mixed records, despite progress in areas 6. The use of fertilizer increases with roads and rainfall. such as public investments. Demographic growth is a key 7. Cotton production still ensures most cash earnings for challenge but the way government is responding to it, rather farmers as well as most of their inputs and credit for other than hiding behind cultural values and norms, may be crop productions. Cotton earnings also provide funds for viewed as a success. Creating and using fiscal space in increasing capacities of smallholders’ unions, agricultural the near future, coping with demographic growth, and pur- extension services and research. The strong linkages suing reforms will all be crucial. Key investments in infra- between cotton and other crop productions must be borne structure (first priority) and in human development (second in mind while not forgetting agronomic and economic priority) will be the cornerstone of subsequent development. complementarities that are harnessed in a very constrained While progress thus far has been in the right direction, it environment for farmers. must be complemented by sector-specific policies in line 8. Economic diversification is defined here as the process of with structural economic change and better promotion of progressive enlargement of the number of goods produced private investment and exports. Several further improve- in an economy without a decline in productivity. ments in the public administration and the political systems 9. Promising technologies include integrated livestock- are also expected. Indeed, internal political sustainability of crop farming systems; soil conservation; regreening pro- the policymaking and decision processes is a necessary con- grams; small-irrigation schemes; intercropping and alley cropping; pest management; early cultivars for rain-fed rice, dition for successful policy-led development and reforms millet, and sorghum; and new generations of maize varieties (Rodrik 1996). (Kaminski 2008). Finally, in the course of market reforms in Burkina Faso, 10. They involve the widespread use of inventory credit, policy makers should bear in mind that removing market agroprocessing of traditional cereals such as millet, producers- imperfections (with investments in infrastructure to reduce processors contracts in the poultry industry. and so on. transaction costs) needs to be done very carefully. As Lipsey 11. As noted by Platteau (1996), infrastructural constraints and Lancaster (1956) note, the persistence of other market are a major cause of the low long-run supply response of failures may have an adverse effect even after imperfections farmers to price incentives, notably for transport and com- are removed. Timing of market and institutional reforms munications. Although tax reforms and infrastructure and infrastructure investment should proceed accordingly. investment both compete for public funds, their effects are complementary and differently impact producers. NOTES 12. The mitigation of agricultural incomes’ vulnerability to weather and external shocks is a key strategy through 1. Other significant sources of export earnings are live- income-enhancing diversification under intensified farming stock and gold. systems for selected products (comparative advantages and 2. The higher incidence of poverty in rural areas is dis- market/trade potential), an improved business environment putable, because the official poverty line established by the to encourage private sector investment (decreasing trans- Institut National de la Statistique et de la Démographie action, financial, labor, and energy costs), and an improved (INSD) in 2003 places too much weight on the prices of trade environment (infrastructures, export promotion, necessary goods (such as staples and water) relative to the market integration, and quality standards). CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 121 BIBLIOGRAPHY Conley, T. G., and C. R. Udry. 2000. “Learning about a New Technology: Pineapple in Ghana.” Yale Economic Abdoulaye, T., and J. Lowenberg-DeBoer. 2000. “Intensifica- Growth Center Working Paper 817, New Haven, CT. tion of Sahelian Farming Systems: Evidence from Niger.” Agricultural Systems 64 (2): 67–81. Coulter, J., A. Goodland, A. Tallontire, and R. Stringfellow. 1999. Marrying Farmer Cooperation and Contract Farm- Abdoulaye, T., and J. H. Sanders, 2005. “Stages and Determi- ing for Agricultural Service Provision in a Liberalizing nants of Fertilizer Use in Semiarid African Agriculture: The SSA. ODI Natural Resources Perspectives, 48, Overseas Niger Experience.” Agricultural Economics 32: 167–79. Development Institute, London. ———. 2006. “New Technologies, Marketing Strategies and Fafchamps, M. 2004. Market Institutions in Sub-Saharan Public Food Policy for Traditional Food Crops: Millet in Africa: Theory and Evidence. Cambridge, MA: MIT Press. Niger.” Agricultural Systems 90: 279–92. Foster, A. D., and M. R. Rosenzweig. 1995. “Learning by African Economic Outlook. n.d. “Burkina Faso.” http://www Doing and Learning from Others: Human Capital and .africaneconomicoutlook.org/en/countries/west-africa/ Technical Change in Agriculture.” Journal of Political burkina-faso/. Economy 103 (6): 1176–1209. Akiyama, T., J. Baffes, D. Larson, and P. Varangis. 2001. Gray, Leslie C., and Michael Kevane. 2001. “Evolving Tenure Commodity Markets Reforms: Lessons of Two Decades. Rights and Agricultural Intensification in Southwestern Washington, DC: World Bank. Burkina Faso.” World Development 29 (4): 573–87. Badiane, O., F. Goletti, M. Kherallah, P. Berry, K. Govindan, Grimm, M., and I. Gunther. 2004. “How to Achieve Pro- P. Gruhn, and M. Mendoza. 1997. Agricultural Input and poor Growth in a Poor Economy? The Case of Burkina Output Marketing Reforms in African Countries: Final Faso.” Working paper, University of Göttingen, Germany. Report. Washington, DC: International Food Policy Habiyaremye, A., and T. Ziesemer, 2006. “Absorptive Capac- Research Institute. ity and Export Diversification in Sub-Saharan African Barrett, C. B. 1997. “Food Marketing Liberalization and Countries.” UNU-MERIT Working Paper Series. United Trader Entry: Evidence from Madagascar.” World Devel- Nations University, Maastricht Economic and Social opment 25 (5): 763–77. Research and Training Centre on Innovation and Tech- Bassett, T. 2001. The Peasant Cotton Revolution in West nology, Maastricht, the Netherlands. Africa, Côte d’Ivoire, 1880–1995. Cambridge: Cambridge Hausmann, R., J. Hwang, and D. Rodrik. 2007. “What University Press. You Export Matters.” Journal of Economic Growth 12 (1): Bernard, T., M-H. Collion, A. de Janvry, P. Rondot, and 1–25. E. Sadoulet. 2008. “Do Village Organizations Make a Hausmann, R., and B. Klinger. 2006. “Structural Transforma- Difference in African Rural Development? A Study for tion and Patterns of Comparative Advantage in the Prod- Senegal and Burkina Faso.” World Development 36 (11): uct Space.” CID Working Paper, Center for International 2188–204. Development, Harvard University, Cambridge, MA. Berthelemy J-C. 2005. “Commerce international et diversifi- Hirschman, A. 1958. The Strategy of Economic Development. cation économique. “ Revue d’Economie Politique 115 (5): New Haven, CT: Yale University Press. 591–611. Imbs, J., and R. Wacziarg. 2003. “Stages of Diversification.” Boserup, E. 1965. The Conditions of Agricultural Growth: The American Economic Review 93 (1): 63–86. Economics of Agrarian Change under Population Pressure. IMF (International Monetary Fund). 2007. Burkina Faso. Chicago: Aldine. Propositions de réformes fiscales: Simplification, équité et Boughton, D., and T. Reardon. 1997. “Will Promotion of efficacité. Washington, DC. Coarse Grain Processing Turn the Tide for Traditional ———. 2008. Burkina Faso: Third Review Under the Three- Cereals in the Sahel? Recent Empirical Evidence from Year Arrangement under the Poverty Reduction and Mali.” Food Policy 22: 307–16. Growth Facility. Washington, DC. Collier, P. 1998. “Globalization: Implications for Africa.” In ———. 2009. Burkina Faso: “Third Review under the Trade Reform and Regional Integration in Africa, ed. Three-Year Arrangement under the Poverty Reduction Z. Iqbal and M. S. Khan. IMF Institute, International and Growth Facility. Staff Report.” Press Release on the Monetary Fund, Washington, DC. Executive Board Discussion. IMF Country Report No. ———. 2002. Primary Commodity Dependence and Africa’s 09/38, Washington, DC (January). Future. Washington, DC: World Bank. Jayne, T., and S. Jones. 1997. “Food Marketing and Pricing Collier, P., and J. W. Gunning. 1999. “Why Has Africa Grown Policy in Eastern and Southern Africa: A Survey.” World Slowly?” Journal of Economic Perspectives 13 (3): 3–22. Development 25: 1505–27. 122 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH Jayne, T., J. D. Shaffer, J. M. Staatz, and T. Reardon. 1997. Moseley, W. G., and L. Gray, eds. 2008. Hanging by a Thread: “Improving the Impact of Market Reform on Agricul- Cotton, Globalization, and Poverty in Africa. Athens, OH: tural Productivity in Africa: How Institutional Design Ohio University Press. Makes a Difference.” MSU International Development Newbery, J. 1989. “The Theory of Food Price Stabilization.” Working Paper 66, Michigan State University, East Economic Journal 99: 1065–82. Lansing, MI. Pinkney, T. C. 1993. “Is Market Liberalization Compatible Jones, S. 1998. Liberalised Food Marketing In Developing Coun- with Food Security? Storage, Trade, Price Policies for Maize tries: Key Policy Problems. Oxford Policy Management. in Southern Africa.” In Agricultural Policy Reforms and Kaminski, Jonathan. 2007. “Interlinked Agreements and the Regional Market Integration in Malawi, Zambia and Zim- Institutional Reform in the Cotton Sector of Burkina babwe, ed. A. Valdés and K. Muir-Leresche. Washington, Faso.” Working paper, Atelier de Recherche Quantitative DC: International Food Policy Research Institute. Appliquée au Développement Économique, Toulouse Platteau, J-P. 1996. “Physical Infrastructure as a Constraint School of Economics, Toulouse, France. on Agricultural Growth: The Case of Sub-Saharan ———. 2008. “Cotton-Cereal Systems in West and Central Africa.” Oxford Development Studies 24 (3): 189–219. Africa: Opportunities and Constraints for Revenue- ———. 2007. “Constraints on African Economic Growth: Raising Diversification and Marketing Strategies.” FAO The Institutional Legacy.” Paper presented at the first All-ACP Background Paper, European Union, Rome. IERC conference, “The Economic Performance of Civ- Kaminski, J., and Y. Bambio. 2009. “The Cotton Puzzle in ilizations: Roles of Culture, Religion, and the Law,” Burkina Faso: Local Realities versus Official Statements.” University of Southern California, Los Angeles. Paper presented at the annual meeting of the African Reardon, T. 1993. “Cereals Demand in the Sahel and Poten- Studies Association, New Orleans, LA. tial Impacts of Regional Cereals Protection.” World Kaminski, J., D. Headey, and T. Bernard. 2009. “Navigating Development 21 (1): 17–35. through Reforms: Cotton in Burkina Faso.” In Millions Reij, C., G. Tappan, and M. Smale, 2009. “Re-Greening the Fed: Proven Successes in Agricultural Development. Sahel: Farmer-Led Innovation in Burkina Faso and Washington, DC: International Food Policy Research Niger.” In Millions Fed: Proven Successes in Agricultural Institute. Development. Washington, DC: International Food Policy Kaminski, J., and A. Thomas. 2009. “Land Use, Production Research Institute. Growth, and the Institutional Environment of Smallhold- Rodrik, D. 1996. “Understanding Economic Policy Reform.” ers: Evidence from Burkinabe Farmers.” CAER-LERNA Journal of Economic Literature 34 (1): 9–41. Working Paper, Hebrew University, Jerusalem, and Insti- Savadogo, K. 2009. Le contexte macroéconomique. Back- tut National de la Recherche Agronomique, Toulouse ground paper for the Burkina Faso Country Economic School of Economics, Toulouse France. Memorandum. World Bank, Washington, DC. Labaste, Patrick. 2009. Background paper for the Country Stringfellow, R., J. Coulter, A. Hussain, T. Lucey, and Economic Memorandum. World Bank, Washington, DC. C. McKone. 1997. “Improving the Access of Smallholders Lachaud, J-P. 2005. “A la recherche de l’insaisissable to Agricultural Services in Sub-Saharan Africa.” Small dynamique de pauvreté au Burkina Faso. Une nouvelle Enterprise Development 8 (3): 35–41. évidence empirique.” Document de travail 117, Centre Stringfellow, R., J. Coulter, T. Lucey, C. McKone, and d’Economie du Développement de l’Université Mon- A. Hussain. 1996. The Provision of Agricultural Services tesquieu Bordeaux IV, Bordeaux, France. through Self-Help in Sub-Saharan Africa. Research report ———. 2006. “La croissance pro-pauvres au Burkina Faso. R6117CA, Natural Resources Institute, Chatham, United L’éviction partielle de l’axiome d’anonymat en présence Kingdom. de données transversales.” Document de travail du SWAC Secretariat, and OECD (Organisation for Economic département de l’école d’économie du développement, Co-operation and Development). 2007. “Economic and l’Université Montesquieu Bordeaux IV, Bordeaux, France. Social Importance of Cotton Production and Trade in Lipsey, R. G., and K. Lancaster. 1956. “The General Theory of West Africa: Role of Cotton in Regional Development, Second Best.” Review of Economic Studies 24 (1): 11–32. Trade and Livelihoods.” Working Paper, Paris. Massel, B. F. 1970. “Export Instability and Economic Struc- Teravaninthorn, S., and G. Raballand. 2008. Transport Prices ture.” American Economic Review 60 (4): 618–30. and Costs in Africa. A Review of the International Corri- McMahon, G., and N. J. Ouédraogo. 2009. “Ouagadougou, dors. Washington, DC: World Bank. Burkina Faso.” Draft mining chapter for Country Eco- Tschirley, D., C. Poulton, and P. Labaste, eds. 2009. Organi- nomic Memorandum. World Bank, Washington, DC. zation and Performance of Cotton Sectors in Africa. CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH 123 Learning from Reform Experience. Washington, DC: ———. 2009a. Burkina Faso: Country Economic Memoran- World Bank. dum. Promoting Growth, Competitiveness and Diversifica- World Bank. 2004. “Cotton Cultivation in Burkina Faso: A tion. Washington, DC. 30-Year Success Story.” Paper presented at the conference ———. 2009b. Doing Business 2009. Washington, DC: “Scaling up Poverty Reduction, a Global Learning World Bank. Process,” Shanghai. ———. 2010. World Development Indicators. Washington, ———. 2007a. Burkina Faso: The Challenge of Export Diver- DC: World Bank. sification for a Landlocked Country. Diagnostic Trade Inte- ———. Various years. Enterprise Surveys. Washington, gration Study for the Integrated Framework Program. DC. Washington, DC. Yanggen, D., K. Kelly, T. Reardon, and A. Naseem. 1998. ———. 2007b. World Development Indicators. Washington, “Incentives for Fertilizer Use in Sub-Saharan Africa: A DC: World Bank. Review of Empirical Evidence on Fertilizer Response and ———. 2008. World Development Indicators. Washington, Profitability.” Michigan State University Working Paper DC: World Bank. 70, East Lansing, MI. 124 CHAPTER 6: COTTON DEPENDENCE IN BURKINA FASO: CONSTRAINTS AND OPPORTUNITIES FOR BALANCED GROWTH PA R T I I Post Conflict Situations – Building Institutions and Governance CHAPTER 7 Postconflict Economic Governance Reform: The Experience of Liberia Vishal Gujadhur hen Ellen Johnson Sirleaf began her tenure as W that poor economic governance during the 1970s and president of Liberia in January 2006, she took 1980s was one of the causes of the conflict. Corruption charge of a country facing enormous chal- and mismanagement of the economy worsened dramati- lenges. It had been more than two years since the signing of cally during the war and prolonged the conflict, as cor- the Accra Comprehensive Peace Agreement that brought an ruption grew out of control and there was little political uneasy end to a conflict that killed more than 250,000 will for a well-run government. Sound economic policies Liberians and left more than 500,000 others displaced. The were therefore, crucial for Liberia’s long-term prospects scale and intensity of the violence had destabilized the for peace. region politically and economically, making Liberia a global Initial steps toward reform of economic governance in exemplar of a failed state. Liberia were taken during the transitional period between Chief among the president’s many challenges was restor- the end of the civil war and start of the new administration. ing public trust in economic governance. As the country Among these was the launch of the Governance and Eco- collapsed during the conflict, so did government finances nomic Management Assistance Program (GEMAP) several and management of those finances. By 2005 revenue was months before the January 2006 elections. A partnership less than $95 million a year, with public spending only between the government of Liberia and a group of interna- about $25 a year on each citizen, one of the lowest levels in tional donors, GEMAP was intended to improve the the world. Corruption was rampant. Liberia’s external debt accountability and transparency of fiscal management in ballooned to $4.7 billion, or roughly 800 percent of GDP Liberia. President Sirleaf continued the implementation of and 3,000 percent of export in 2005. Domestic debt and GEMAP after her election, and in addition she committed arrears exceeded $600 million by 2006, largely due to weak to making further economic governance reforms. Five years government spending controls that allowed ministries and on, Liberia’s recovery in terms of public finances and agencies to spend without central authorization. Economic responsible public policy is striking. Annual government mismanagement and the uncertainty surrounding public revenue has nearly tripled since the start of the Sirleaf finances was reflected in the continuation of UN sanctions administration and substantial inroads have been made in on the export of Liberian diamonds and timber. the budgeting and expenditure processes. Civil servants are The low level of trust in the government’s economic now paid on time, the government has accumulated no new management capabilities was even more significant given domestic arrears, and it has reached the completion point in Vishal Gujadhur is a former Center for Global Development Scott Fellow at the Ministry of Finance in Liberia. 127 the enhanced Heavily Indebted Poor Countries (HIPC) astonishing 90 percent, and by the time elections took process, allowing more than $4 billion of external debt to be place in 2005, average income in Liberia was one-fourth of written off. Liberia’s international reserves also increased, what it had been in 1987 and one-sixth of what it had been reaching $108 million in December 2008 and $312 million in 1979. in December 2009.1 At the same time, Liberia has climbed Finally, after 14 years of instability covering two civil the ranks of Transparency International’s Corruption Per- wars, the Accra Comprehensive Peace Agreement signed in ceptions Index, moving from 137 of 158 countries in 2005 August 2003 provided for a National Transitional Govern- to 97 of 180 countries in 2009. The nation has also adopted ment of Liberia (NTGL) to guide Liberia toward elections the Kimberley Process Certification scheme for sourcing in late 2005. The country faced enormous reconstruction diamond exports, established new forestry regulations, and needs, and international donors stepped in to help set is now fully compliant with the Extractive Industries Trans- Liberia on the path to recovery. The transitional govern- parency Initiative (EITI). Thanks to the improvements in ment took several initial steps toward addressing economic economic governance, the UN-imposed ban on diamond governance issues. The chairman of the NTGL signed Exec- and timber exports from Liberia has been removed. utive Order No. 2, which consolidated all government Liberia’s success at gaining control over public finances accounts at the Central Bank of Liberia (CBL). The in a postconflict environment has prompted a closer look at government requested audits of practices under the past the factors underpinning the turnaround. Establishing a regime. The Results-Focused Transitional Framework simple causal story is difficult because two major policy (RFTF), meanwhile, set out a comprehensive reconstruction changes happened nearly simultaneously: the start of Presi- framework. The transitional government also partnered dent Sirleaf ’s administration and the implementation of with the World Bank, the International Monetary Fund GEMAP. The story is further complicated by a general lack (IMF), and the European Community (EC) to work on of data resulting from the realities of a postconflict country issues related to governance, economic revitalization, and and the difficulty of measuring outcomes related to financial management. economic governance. Many observers have been quick to credit GEMAP for the economic governance successes that have taken place in Persistent fiscal problems and the Liberia in recent years. While GEMAP played a useful role, genesis of GEMAP it also had shortcomings, and an understanding of the con- Following a hopeful start, doubts surfaced at the end of text in Liberia and the efforts of the Sirleaf administration is 2004 about the NTGL’s commitment to improving eco- necessary to establish a more complete story of the recovery nomic governance and fighting corruption, while Liberia’s and the reform effort in Liberia. overall economic recovery remained sluggish. A group of international partners known as the International Contact Group (ICG) expressed concern that the lack of economic THE CONTEXT BEHIND ECONOMIC progress would threaten peace. Audits commissioned by GOVERNANCE REFORM the EC in early 2005 revealed widespread corruption, not Several major events in Liberia’s recent history lie behind just during the time of the Taylor regime but also during the country’s need for economic governance reform, the NTGL. At a review of RFTF progress in May 2005, many namely, the long period of conflict; a difficult economic international partners focused on the mismanagement of recovery following the conflict; and the start of a new, dem- public resources as a primary reason for the sluggish recov- ocratically elected administration. ery, kick-starting negotiations with the NTGL for a more robust international engagement on public financial man- agement in Liberia.2 Descent into war and emergence from conflict The NTGL was reluctant to submit to international Liberia’s economic problems began before the first out- partners’ recommendations. For the NTGL, the two main break of conflict in 1989. Even then, the country was points of contention were the submission of GEMAP for marked by deep social and economic divisions, and Samuel UN Security Council endorsement and the international Doe’s coup d’état in 1980 began a decade of gross misman- recruitment of the chief administrator of the CBL. Only agement until his assassination in 1989. Between 1987 and after repeated threats of halting all international aid did the 1995, Liberia’s gross domestic product (GDP) fell by an NTGL representatives agree to GEMAP. In the end, explicit 128 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA linkage to the UN was removed, although the exit from month after the signing of GEMAP. There was little progress GEMAP was now linked to Liberia reaching the Enhanced on GEMAP implementation at that point, because Liberia HIPC Completion Point. This had the benefit of both tying was focused on the elections. The IMF, which had cut off the completion of GEMAP to a significant incentive (clear- relations with Liberia during the war, began discussions ing Liberia’s enormous debt burden) and providing about providing assistance to the country and possibly insti- detailed triggers that could be determined at a later date, tuting a staff-monitored program. because the triggers in a HIPC program generally relate to At her inauguration in January 2006, Ellen Johnson public financial management and goals that are in concert Sirleaf acknowledged the scale of the challenge of rebuilding with GEMAP. the war-torn country. Toward this end, the administration Broadly, GEMAP targeted transparency in public finan- announced its “150-Day Action Plan” shortly thereafter. The cial management and revenue collection, expenditure con- plan detailed the objectives and deliverables that both it and trols, and government procurement and concession the donor community would assume upon taking office. practices. Thus conceived, GEMAP ambitiously sought to The 150-Day Action Plan organized the administration’s accomplish six objectives: securing Liberia’s revenue base, work into four pillars: expanding peace and security, revi- improving budgeting and expenditure management, talizing economic activity, strengthening governance and improving procurement practices and granting of conces- the rule of law, and rebuilding infrastructure and providing sions, establishing effective processes to control corruption, basic services. supporting key institutions, and capacity building. Revitalizing economic activity and strengthening gover- A key innovation undertaken as part of GEMAP was the nance, in particular, are both closely related to the areas cov- system of internationally recruited financial controllers ered by GEMAP. While the capacity for overlap between the posted alongside Liberians in key agencies. This system was administration’s 150-Day Plan and GEMAP was certainly meant to enable the controllers to help establish transparent possible—as was the possibility, at the time of GEMAP’s financial procedures, train and build local capacity from creation, that the new administration would not want to within the agencies, and report on revenue and spending. work within the GEMAP framework—the scenario played The centerpiece of GEMAP design, for which it is known out otherwise. A passage from President Sirleaf ’s inaugura- internationally, is the cosignatory authority these experts tion speech addresses the issue directly: wielded, which ensured that no major financial transactions could take place without being scrutinized by both a Liber- “If we are to achieve our development and anti- ian manager and an international adviser. The basic corruption goals, we must welcome and embrace the GEMAP goal was to “ensure that all revenues due to the Governance and Economic Management Program government are collected, and [that] those revenues are (GEMAP) which the National Transitional Govern- spent according to a budget,” all in an environment of ment of Liberia, working with our international part- increasing transparency.3 ners, has formulated to deal with the serious Although GEMAP was intended to be an interim plan for economic and financial management deficiencies in the body of international partners involved in postconflict our country. reconstruction in Liberia, in practice it provided an embed- We accept and will enforce the terms of GEMAP, ded control system to maintain a transitional economic recognizing the important assistance which it is governance framework. The intention was to ensure that expected to provide during the early years of our gov- government and donor resources were secured and chan- ernment. More importantly, we will ensure compe- neled through the budget, but the overall setup was one of tence and integrity in the management of our own increased oversight. In the end, it would be up to the gov- resources and insist on an integrated capacity build- ernment to put in place the internal systems and use the ing initiative so as to render GEMAP nonapplicable in space that this oversight afforded in order to set up long- a reasonable period of time.” (Sirleaf 2006) term systems. In short, the new administration recognized that with scarce government resources and limited capacity, the Elections and the Sirleaf administration international community would play a prominent role in Liberia’s Economic Governance Steering Committee Liberia’s reconstruction and economic governance (EGSC) held its inaugural meeting in October 2005, one reform. CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 129 KEY COMPONENTS OF THE APPROACH TO external agencies to place some checks on corruption in the ECONOMIC GOVERNANCE REFORM absence of domestic impetus. That said, GEMAP also left ample room for domestic authorities to implement it in The simultaneous timing of the GEMAP implementation either more extensive or more limited ways, partly through and the change in administration in Liberia is an important the vagueness of such areas as capacity building, but also in one, because the government of Liberia is both the driver of the possibility that GEMAP would enable more resources to interventions and the target of change. Clearly, this duality be channeled through the budget. complicates assessments of causality. In addition, GEMAP’s The Sirleaf administration approached its governing broad objectives and multiple actors mean that it is “difficult plans thoughtfully and outlined its key goals and tactics to trace the boundaries of GEMAP assistance,” as noted in from the start. These plans can be traced to the 150-Day a midterm GEMAP evaluation (Morsiani et al. 2008, 13). Action Plan, a framework similar to Poverty Reduction Finally, GEMAP and the change in administration were nec- Strategy Papers prepared by IMF and World Bank member essarily complementary from the start, with GEMAP provid- countries. That framework, which was announced at the ing a set guideline the government could rely on but would start of the Sirleaf administration, would continue to form also be constrained by. Indeed, GEMAP was meant to pro- the overall strategic governance platform as the administra- vide a transitional economic governance system and enable tion embarked on an iterative process that moved from the long-term systems to be put in place. The key question to 150-Action Plan to an Interim Poverty Reduction Strategy focus on then is the appropriateness of the GEMAP inter- Paper in 2007 to a full Poverty Reduction Strategy Paper in ventions and their effect on government policy initiatives. 2008. The administration also used the poverty reduction While GEMAP was an extensive initiative, its essential strategy program as an opportunity to set out annual plans thrust was clear. As stated in an early World Bank/UN joint for economic governance reforms. In fact, it could be argued review, “GEMAP is notable for three key features: the scope that the government’s iterative planning process, including and intrusiveness of its key features, the consistency of its efforts such as the IMF’s Staff Monitored Programs and the features throughout the negotiating process, and a lack of Poverty Reduction and Growth Facility, are part of the detail on funding and implementation” (Dwan and Bailey GEMAP process insofar as they are linked to the Enhanced 2006, 19). The “intrusive” key feature referred to the cosig- HIPC Completion Point. natory authority for state-owned enterprises (SOEs), the From the start, the Sirleaf administration exerted bold Ministry of Finance, the Bureau of the Budget, and the CBL; political leadership and made concerted efforts to work with the position of chief administrator was created in each of donors collaboratively and ensure coordination. In particu- these agencies to be the cosignatory with the international lar, the administration supported a series of locally sourced financial expert. Additionally, linking the exit from GEMAP and targeted interventions in key ministries and agencies, to Liberia achieving HIPC completion shows the imposition while making use of donor resources and coordination. Also of forces outside Liberia on the program, despite GEMAP’s key in the administration’s plans was its effort to make effec- mentions of the “full respect for the sovereignty of Liberia.” tive use of donor resources. Given the ravaged state of the (Government of Liberia 2005). country at the end of the war, the administration recognized Other GEMAP features, such as support for Liberia’s from the start that the reconstruction of Liberia would General Auditing Commission and the Public Procurement require significant external assistance. The challenge was to and Concession Commission, are also important. Those channel the resources effectively, no easy task given the mul- features, however, do not differ substantively from interna- titude of partners, projects, and methods of funding. Flow tional postconflict reconstruction programs in other coun- of donor information was poor, and only in the 2009/10 tries, and much of the work process related to them could be national budget were preliminary estimates of total aid to completed outside the general framework of GEMAP. Liberia collected by the government. Indeed, the other interventions contain much less detail The Sirleaf administration introduced a new mechanism about how they are to be implemented, underlining the cen- to coordinate government and donor activities: the Liberia trality of the cosignatory aspect of GEMAP. Reconstruction and Development Committee (LRDC). The Establishing space for reform was also a key component LRDC’s Steering Committee was initially kept small, to of GEMAP. At the time of its formulation, an optimistic allow decisions to be made more quickly and easily, and outcome for GEMAP would have been its full utilization by resembled a subcabinet group chaired by the president. an incoming administration. At the least, though, it could Members of the Steering Committee included four have been a fail-safe within the government and a way for 130 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA ministers, each with responsibility for one of the recon- revenues from customs duties, import levies and taxes” struction pillars, and representation by the four largest (Government of Liberia 2005). For the administration, donors (the United States, United Nations, World Bank, and procedures to improve tax collection were included in the the European Commission). Representatives of the Eco- 150-Day Plan under the banner of “economic revitaliza- nomic Community of West African States (ECOWAS) and tion,” while increased revenues were central to its ability to of the IMF joined meetings, as did other donors and minis- begin to fulfill its campaign promises. ters when appropriate. The Steering Committee met every The scale of the revenue collection problem was enor- two weeks at the start of the Sirleaf administration, reflect- mous. In fiscal year 2004/05, the government received only ing the large amount of work to be done, before moving to $75 million in taxes, although it had spent more than monthly meetings. $85 million a year between 2000 and 2005. Corruption The LRDC resembled, in some ways, the Economic Gov- affected public spending at all levels of government. Spend- ernance Steering Committee envisioned by GEMAP. By ing per capita of only about $25 was already among the low- adopting an administrative structure that mirrored the sub- est in the world, and for the government to begin showing stantive pillars of the government’s own comprehensive peace dividends and deliver urgently needed services to its action plan, as well as including key donors, a small group citizens, revenue had to be raised quickly. The international of decision makers was able to make policy decisions and community noted in the GEMAP that progress on increasing ensure coordination across efforts. The LRDC also allowed revenue collection was not fast enough under the NTGL, say- problems to be addressed as they emerged, permitting ing that “large leakages were taking place due to difficulties donors to see and participate in the decision-making in fiscal administration, absence of verifiable mechanisms processes of the government and understand the realities for financial control, and malfeasance” (Government of the politicians faced. The strength of this arrangement was Liberia 2005). GEMAP designers also focused on the need to ultimately reflected in the decision in 2009 to merge the enforce the NTGL Executive Order mandating that all rev- EGSC gatherings into the LRDC. enue due to the government be consolidated by the CBL. At their core, both GEMAP and the incoming Liberian While financial management was weak across govern- administration’s plans attempted to accomplish the same ment, GEMAP noted that “SOEs present particularly large goals. GEMAP provided an external oversight and control risk given their weak management and operational struc- function, while the government was responsible for carrying tures and systems” (Government of Liberia 2005). This con- out the actual reforms. After several years, it is now possible cern was reflected in GEMAP’s primary instrument for to investigate the underlying efforts and examine their enforcing financial management and accountability. Per- gains. As discussed earlier, GEMAP’s objectives were broad haps influenced by the audits of the SOEs that had been and ambitious, which has made it difficult to take stock of conducted by the EC during the NTGL, GEMAP led to the the program’s effectiveness. Further fuzziness around the imposition of management contracts and international program’s edges is the result of programs undertaken by controllers for several SOEs (the National Port Authority, GEMAP’s international partners on behalf of the govern- Roberts International Airport, the Liberia Petroleum Refin- ment of Liberia but not explicitly outlined within GEMAP. ery Corporation, the Forestry Development Agency, and the Given the broad scope of public financial management and Bureau of Maritime Affairs) and backed an adviser in the establishing fiscal accountability and transparency, the focus Bureau of Customs and Excise and a new post of chief in the future will need to be on two priority areas: securing administrator in the CBL. Liberia’s revenue base and establishing an effective expendi- In hindsight, the focus of GEMAP on placing experts at ture process. SOEs may have been misdirected, because the SOEs were somewhat peripheral to the central government’s opera- tions. The Forestry Development Authority, which would Improving revenue collection have been the focal point for forestry revenues, was not as Securing Liberia’s revenue base was a primary initial goal of important after the implementation of UN sanctions on both the Sirleaf administration and GEMAP. Indeed, first timber exports. That was also true for SOEs as a whole, on the list of GEMAP components is financial management because in the absence of profits that can be turned over as and accountability. As described in the GEMAP agreement, dividends, there was little financial connection between the “It is critical to protect the revenue streams of key revenue central government and SOEs. In fact, the only transfer to generating agencies and institutions, as well as to secure the the government from SOEs during the first few years of the CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 131 administration was $1.7 million from the LPRC in 2007. As electronic computerized receipt system during 2005 in a GEMAP evaluation report noted, total revenue collection collaboration with USAID, the Ministry of Finance, and by the Ministry of Finance in fiscal year 2006/07 was $139 the CBL. However, the system was not fully used and rec- million—approximately four times the combined gross onciliation of the CBL/Ministry of Finance account did revenues of the four SOEs with GEMAP controllers (Dod not begin until the new administration took office. By the and Nelson 2008). end of January 2006, before any GEMAP advisers had been As for the Bureau of Customs and Excise at the Ministry deployed across government, the newly installed Liberian of Finance, which collects over half of government revenue, authorities accepted only computer-produced receipts at the improvement in revenue performance has not been a the CBL. Liberian authorities also began to ensure more result of GEMAP. An evaluation of the first short-term stringent preshipment inspections for goods moving expert working in those two agencies found him to be of through the port, and began to eliminate noncash tax substandard quality, and three years after the signing of payments—previously, the government had accepted “in GEMAP there had been no systematic donor support under kind” payments. This sort of insistence on adherence to GEMAP auspices (Morsiani et al. 2008). In fact, government existing rules and use of systems already in place allowed action and IMF support were much more salient in improv- the government to produce the dramatic 35 percent ing revenue collection, as evaluations have concluded. A increase in revenue collection, and in the process demon- U.S. Agency for International Development (USAID) strated newfound political will. report, however, has noted that while there have been efforts The government’s success in increasing revenues was not under GEMAP to support systemic improvements in rev- limited to collection of customs duties; in fact, it was real enue collection in Liberia, particularly in the customs arena, and sustained across several areas of the government. Cus- the assistance was “not being fully utilized for the benefit of toms duties nearly doubled in the first year of the Sirleaf Liberia” and has produced little follow-up for functions that administration, from $35.3 million in fiscal year 2005/06 to had been installed (Dod and Nelson 2008). $69.9 million in 2006/07. This is all the more impressive How, then, was the Liberian government able to have a considering that the percentage increase in imports during nearly instantaneous effect on revenue collection? The this time was much smaller (imports rose from $294 million 150-Day Plan announced a target of increasing revenues in 2005 to $418 million in 2007). The government also by at least 15 percent year-on-year, and indeed, the overall began to reorganize its methods for collecting income taxes. revenue from international trade increased by 35 percent Efforts in this area paid off immediately, with actual tax rev- in the first five months of the Sirleaf administration (Dod enues increasing 19 percent between February and June and Nelson 2008). This was accomplished largely by better 2006 (IMF 2006). GEMAP experts did not arrive in Liberia use of solutions already proposed. As the new government until mid-April 2006. came to office, customs payments were collected in an Table 7.1 shows that the increase in Liberian revenues was “unwieldy and complex manual system,” according to a not limited to customs duties (shown in the table as interna- GEMAP assessment (Dod and Nelson 2008). To address tional trade taxes). The deputy minister of revenue of the this, a technical expert from the U.S. Treasury designed an Ministry of Finance, in tandem with the IMF, introduced a Table 7.1 Revenue Collection in Liberia (millions of dollars) Revenue 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Total revenue 56.0 79.3 84.6 146.8 200.8 211.3 Total tax revenue 55.3 75.7 81.0 140.0 168.8 190 International trade tax 24.3 30.2 35.3 69.9 79.1 87.9 Income tax 9.4 28.4 25.1 42.5 52.6 65.8 Goods and services tax 21.3 16.9 20.3 26.1 34.9 33.7 Other 0.3 0.3 0.3 1.4 2.2 2.6 Nontax revenue 0.8 3.6 3.6 6.9 32.0 21.4 From SOEs 0 0 0 1.7 0 0 Grants 3.0 1.0 1.0 1.5 0 23.6 Source: IMF Liberia Program Reviews. 132 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA series of actions aimed at both short- and medium-term evaluation reports, which found that GEMAP performed revenue collection. A partial list of actions includes com- “unremarkably” in regard to the improvement of central puterization of the tax payment system; strengthening cus- banking and securing revenue at the Bureau of Customs toms administration and steadily reducing exemptions; (Morsiani et al. 2008). increasing penalties for not undergoing preshipment inspection, reorganizing the tax collection system into large, Establishing an effective expenditure process medium, and small taxpayer units, with an emphasis on the large taxpayer units; and increasing the excise tax on beer Just as important as improving revenue collection in Liberia and cigarettes. The turn away from exemptions and short- was increasing the transparency and effectiveness of public term fixes that had been common during the periods of spending. One of the stated aims of the government was to instability in favor of modernization of tax collection and “restore the government’s credibility in using the country’s rules-based collection is reflected in an increase in the scarce resources efficiently and effectively,” while a primary income tax revenue take of more than 50 percent during the GEMAP objective was to improve the controls in public first year of the Sirleaf administration. expenditures. Gains in effective expenditure, however, are Another central platform in the GEMAP effort to more difficult to isolate than gains in revenue collection, increase revenue collection in Liberia was the institution of because metrics of the quality and efficiency of spending are the chief administrator position at the Central Bank of generally not objective or are unavailable in Liberia. Liberia. The original idea was for the administrator to serve During the NTGL period, the meager amounts of rev- under the guidance of the governor of the CBL, to possess enue the government collected were not spent effectively, binding cosignatory authority on operational and financial and reports were widespread in the Liberian media of cor- matters, to advise on banking operations and to emphasize ruption and similar accusations among various factions in internal controls and audits. These plans never came to the NTGL. Because the Bureau of Budget was initially oper- fruition, however, because the governor of the CBL consid- ationally independent from the Ministry of Finance, the ered the cosignatory arrangement unacceptable and the structure required cooperation and collaboration in budget first GEMAP adviser found it difficult to work within the formulation and execution. Past governments, meanwhile, system and left the position after 15 months. After this, a had kept the national budget secret and allowed unlimited GEMAP adviser was placed at the CBL in a revamped posi- transfers between line items at the executive branch’s whim. tion, focused more on operational matters. Procurement did not follow standardized procedures. In The second GEMAP adviser was able to contribute sig- short, the structure the Sirleaf administration inherited did nificantly to improvements in the CBL system. The CBL was not fit its aims. able to register surplus in its financial position, introduce a As an interim measure, foreign policy advice and techni- travel policy harmonized with the rest of the government, cal assistance during the NTGL established a cash-based and increase its dollar reserves from $6.4 million at the start balanced budget with budget execution handled through of 2006 to $35.1 million in 2007. The midterm GEMAP the Cash Management Committee (CMCo), which consisted evaluation, though, recommended “a careful scrutiny in the of the Bureau of Budget and Ministry of Finance officials, effectiveness of GEMAP assistance to CBL given the fact who approved all government expenditures. To introduce that the conflicts it creates are not conducive to the achieve- better procurement practices, international partners helped ment of results” (Dod and Nelson 2008). draft a new Public Procurement and Concessions Act. By Despite their successes, GEMAP efforts were a relatively the time of the 2005 IMF Article IV consultation, however, small part of the overall task of improving revenue collec- IMF staff found that “significant expenditures were still tion in Liberia, primarily because of GEMAP’s targeting of made without the committee’s prior authorization” and that SOEs. During the war and the NTGL period, resources con- the CMCo was “not operating as intended” (IMF 2005). tinued to flow to SOEs and agencies, while the central gov- Reviews of the Ministry of Finance found that it “operated ernment remained weak. Centralization of government in constant crisis mode, with the CMCo under constant and accounts significantly reduced opportunities for leakage. enormous pressure to clear payments,” while “poor docu- Although GEMAP’s support to SOEs was useful in improv- mentation, thirty stages of voucher review, lack of informa- ing those institutions’ performance, it was peripheral to the tion and opaqueness regarding prioritization exacerbated revenue recovery led by government efforts and non- [problems]” (Morsiani et al. 2008). With no reliable esti- GEMAP-related donor support. This finding is echoed in mates of budgetary receipts and balances, it was hard for CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 133 line ministries to learn of their budgetary allotments, and both short- and medium-term targets: the monthly allot- the lack of commitment controls led to a buildup of gov- ments and current budget had to be executed appropri- ernment liabilities. By the end of 2004 the government had ately, while the following year’s budget had to be prepared approximately $12 million in domestic arrears, roughly simultaneously. The arrival of the GEMAP adviser brought 15 percent of the total budget. increased focus on both immediate controls in terms of GEMAP interventions aimed at improving budgeting cosigning of allotment requests and capacity building and expenditure management consisted of the establish- through training of staff at the Bureau of the Budget. The ment of an Integrated Financial Management Information importance of the adviser was also discernible in reports System and the placement of advisers with cosignatory noting instances of abuse when the adviser was temporar- authority in the Bureau of Budget and Ministry of Finance. ily absent (Dod and Nelson 2008). The incoming government also took action in this area, In the lead-up to the merger of the Bureau of the Budget indicating in its 150-Day Action Plan that the CMCo should and the Ministry of Finance, reactivation of Liberia’s inter- be effectively operational upon the start of the new admin- agency National Budget Committee led to improved coordi- istration in January 2006. In this case, the work done by nation and oversight of both budget policy formulation and GEMAP and the government were mutually reinforcing. expenditure. In addition, the budget preparation process Expenditure controls first took root through the CMCo, became more open and participatory, particularly during which quickly began to function effectively and arguably the fiscal year 2007/08 budget cycle, when both the GEMAP became one of the successes of GEMAP and the govern- adviser and local staff began to focus on long-term planning ment. Under GEMAP’s commitment control system, all and processes. This sequenced intervention—first providing purchase orders and expenditure vouchers from line min- for the immediate stabilization of the procurement system istries were submitted to the CMCo, which verified that and then moving toward long-term planning—was an each proposed expenditure was consistent with the spend- example of effective implementation of the GEMAP. ing agency’s monthly expenditure allocation and compliant A key enabling factor for Liberia’s success in establishing with the procurement law. The CMCo also ensured that an effective expenditure program was senior management there were sufficient funds in the government’s accounts at at both the Ministry of Finance and Bureau of the Budget. the CBL. The system in which payments required CMCo Given the political nature of a budget in a democracy, this authorization prevented expenditure arrears and unautho- also involved strong links with Liberian civil society, and rized government purchases. Although similar systems had the lengths to which the administration went to work with been in place beforehand, they were cumbersome, had little civil society were rewarded with a relatively high Public political support, and consequently were rarely enforced. Expenditure and Financial Accountability (PEFA) score of In a September 2006 review, IMF staff found that “B” by 2007.4 Liberia’s commitment control system had been well imple- While Liberia should be lauded for its success in manag- mented. From the start of the administration, no new ing its budget and expenditures, it should also be noted that arrears accumulated and all expenditure authority was con- the Integrated Financial Management Information System solidated with the CMCo. Through public notices, the gov- did little to improve Liberia’s budget and expenditure ernment made it clear that commitments not accompanied reform process. As of 2010 the system still had not been by CMCo vouchers were not legal claims on the govern- implemented. On the other hand, the Resource Manage- ment. Although the foreign adviser at CMCo likely played a ment Unit, the pool of donor money originally intended to significant role, CMCo also had a foreign adviser during the fund technical assistance and support for the system, was NTGL period. What changed with the Sirleaf administra- able to be used to attract talent to and finance a manage- tion was not the presence of a foreign adviser but the polit- ment training school and an aid management unit within ical environment and support for that adviser. the Ministry of Finance. In fact, the RMU became an Within the Bureau of the Budget, complementary adaptable method of ensuring that the Ministry of Finance improvements were implemented. Even before the received technical assistance for its changing needs and GEMAP-recruited adviser began to work with the Bureau could fill immediate capacity gaps where advisers had of the Budget in April 2006, there was increased communi- direct operational responsibilities. With this flexible setup, cation with ministries to ensure that commitments did not Liberian leaders were able to implement their own pro- exceed available revenues. The task of the Bureau of the grams and initiate the day-to-day operational work needed Budget, however, was complicated by the importance of to get Liberian institutions running effectively again. 134 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA Overall, evaluation reports are favorable for GEMAP on perhaps complicated by the government of Liberia being securing and stabilizing revenue collection and improving both an agent and target of change. Still, the effect of the expenditure management, despite the fact that the processes change in political leadership in the sequence of reform is were not fully in line with what was intended at program undeniable. Figure 7.1 provides an overview of the timeline conception. That said, the GEMAP-appointed advisers at of economic governance reforms in Liberia. the Ministry of Finance and Bureau of the Budget had a Without a doubt, the shift between the NTGL period positive impact, and reports suggest cosignatory authority and the Sirleaf administration in terms of economic man- was useful in these cases, although the advisers were also agement was marked. President Sirleaf was democratically significantly enabled by the leadership in various min- elected, selected members of her government, and had a istries. In both expenditure management and budget coor- comprehensive plan to rebuild Liberia. She signaled a clear dination, GEMAP was able to play a role in stabilizing the break with the policies of the past, and one of the first system and creating capacity for more significant reforms to agenda items in the new administration’s 150-Day Action take place. How this capacity would be used, however, is Plan was to dismiss all NTGL political appointees from the another question. Several additional accomplishments Ministry of Finance. She also endorsed GEMAP in her made under GEMAP are illustrated in box 7.1. inaugural speech and included GEMAP implementation benchmarks in the 150-Day Action Plan. Political leader- ship, then, was clearly important to both the functioning of THE LINK BETWEEN THE GOVERNMENT GEMAP and the change in economic governance. GEMAP OF LIBERIA AND GEMAP evaluators recognized this. Indeed, without this change in While it is difficult to assess how the Liberian government leadership, GEMAP’s operations would necessarily have would have performed in the absence of GEMAP, it is clear been vastly different. that without government buy-in, GEMAP would have oper- All of this said, the openness of Liberian politicians to ated very differently. In the case of revenue collection, the receiving technical assistance and accepting foreign advisers Liberian government acted almost independently of working alongside them is also striking. In part, it reflects GEMAP, driving the reform process, whereas in establishing the scale of work that needed to be accomplished. It likely expenditure and budgeting controls GEMAP had a larger also reflects a country in which both the president and the role in helping the government stabilize the system. The minister of finance were international technocrats with sig- commitment and actions of the government in Liberia to nificant policy-making experience. Either way, it made for a the changes in public accountability are striking, yet many collaborative process that helped to drive economic gover- stock-taking reviews have focused only on the role of nance reform forward. GEMAP. Any overall report card for the administration’s Government buy-in was particularly important in performance, on economic governance or otherwise, is Liberia’s case because at its core, GEMAP established only a Box 7.1 Other Accomplishments under GEMAP GEMAP consisted of six pillars. In addition to the pillars Liberian) and the refurbishment of the General focused on improving revenue collection and establish- Auditing Commission, and provided general advisers. ing effective expenditure processes, the GEMAP pillars The result has been an independent General Auditing focused on supporting key institutions and establishing Commission that provides, in a very public manner, effective processes to control corruption. an ex post check on fraud and support for accounta- These two pillars had contrasting outcomes. In the bility. first case, the evaluation teams appointed by GEMAP By contrast, the Liberia Anti-Corruption Commis- to support the General Auditing Commission were sion has yet to take off. Although donors explicitly out- successful, playing a strong role in rebuilding institu- lined an Anti-Corruption Commission in the GEMAP tional and capacity building. The European Commu- document, the legislation to establish such a commis- nity provided long-term technical support, financed sion did not pass until 2008, and then donor funding an auditor general position (currently held by a support proved to be lacking. CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 135 Figure 7.1 Economic Governance Reform Timeline 250 August 2003: Civil September 2005: NTGL November 2005: war ended by reluctantly signs GEMAP April 2006: Ellen Johnson Accra after international First GEMAP Sirleaf wins run-off comprehensive concern over corruption controllers arrive 200 election peace agreement leads to pressure 150 100 January 2009: First EITI report 50 May 2007: Budget published; Liberia January 2006: June 2006: submitted to legislature March 2008: soon becomes first President Sirleaf Customs receipts as per agreed time, Liberia reaches African country assumes office, up by 35% yoy in followed by HIPC decision designated initiates 150-day plan first 150 days participatory hearings point EITI-compliant 0 2003 2004 2005 2006 2007 2008 2009 GoL revenue collection Note: Revenue data is exclusive of grants and is from the IMF. Liberia’s fiscal year runs from July-to-June, all figures reported are for the fiscal year ending in June (i.e., 2003 is for the fiscal year 2002/2003) transitional economic governance framework that merely long-term building of systems to the government. So, while created the space for the government to carry out the more the interim controls carried out under GEMAP were suc- operational aspects of reform. While a GEMAP-style pro- cessful in helping to prevent new arrears and rooting out gram might work in the case of a hostile recipient govern- opportunistic corruption, their effectiveness resulted in a ment, such a situation is unlikely. One program, the top-heavy structure that led to a difficulty in allowing a USAID-sponsored “Liberia Economic Stabilization Support more successful rollout of spending. Program,” which strongly resembled GEMAP, can serve to By 2008 it was apparent to both the government and its inform the feasibility of GEMAP under a different gover- international partners that a lack of capacity at the line min- nance regime. Implemented in 1989, that program sought istry level was hindering the effective implementation of the to improve revenue collection and expenditure control by government’s agenda. Difficulties implementing the pro- giving 17 experts financial control over government curement law led to a slowdown in critical government accounts (Dwan and Bailey 2006). Its failure in less than a spending that GEMAP’s stop-gap, concentrated measures year underscores the importance of dynamic leadership and were not equipped to address. The capacity building that political will. GEMAP was supposed to deliver could possibly have GEMAP certainly complemented the government’s removed this bottleneck but did not materialize in a signif- efforts and put in place a stop-gap system, but questions still icant way. Instead, the focus on the immediate controls surround the mix of its activities. Although GEMAP’s inten- reduced the long-term relevance and effectiveness of tion, in part, was to lay the foundations for reform, the GEMAP, leaving the government to drive the capacity- activities carried out under it concentrated on keeping the building agenda with donors. system moving forward as constructed. The implementa- In addition to enabling GEMAP to be a success, the polit- tion of the CMCo and financial controllers, for example, ical leadership in Liberia also took independent actions to while successful, left most of the heavy lifting in terms of improve financial management and transparency from the 136 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA start. In her inaugural speech, President Sirleaf promised ity and institutional building” (Morsiani et al. 2008) and that there would be actions taken “so as to render GEMAP recommended a transition from a focus on immediate con- unnecessary,” implying that she intended for the new trols to longer-term system building. administration to target not just GEMAP actions but also The Sirleaf administration made its views clear early on. the larger economic governance areas that GEMAP covered. In a paper prepared for the Liberian Partners Forum in Feb- In the end, the Liberian government’s drive for reform out- ruary 2007, the administration gently noted that donors grew the outlines of GEMAP assistance once the system continued to focus on the short term: “Previous break- had stabilized. downs in governance led to considerable focus being placed on the ‘cosignatory’ aspects of GEMAP. Considering the progress and positive changes that have taken place in the FURTHER CAPACITY FOR ECONOMIC leadership of the country, [the government of Liberia] has GOVERNANCE REFORM encouraged partners to focus on the development of atti- Despite significant progress since 2005, Liberia is still a frag- tudes, systems and procedures that can promote and sustain ile state, and much work remains to be done. That said, crit- integrity, rather than on a transient policing role” (Govern- icism of the government or of donor interventions should ment of Liberia 2007). not downplay the enormity of the tasks before it just five While this focus on stabilization was appropriate and years ago, and many of the concerns today were not of top useful to economic governance reform in an immediate priority during the initial launch; when stabilizing the econ- postconflict situation, a second phase was needed to deepen omy was the government’s unique focus. upstream and downstream controls on spending, including Liberia’s most problematic economic governance short- building up internal and external audits and increasing the fall continues to be capacity building, the most vaguely public procurement capacity in spending agencies, a task defined objective in the GEMAP program document, which that GEMAP frequently bypassed. The controls introduced noted the other components would be accompanied by “a under GEMAP were at best transitory, and it is possible that plan and resources to enable major progress for capacity they were actually a hindrance to establishing deeper, more building” (Government of Liberia 2005). In general, capac- strategic reforms. Indeed, despite GEMAP efforts to ity building has never been a priority under the GEMAP strengthen systems, most donors have yet to trust Liberia program. A two-year capacity-building training program at with budget support. Like GEMAP’s transitional systems, the Ministry of Finance, though useful, is the only explicit donors continue to intervene in Liberia in an ad hoc fash- capacity-building initiative conducted under GEMAP ion, from providing project support to multidonor trust auspices. More informal programs took place in other funds to contributing to a variety of parallel implementa- ministries. tion units designed to work around the Liberian system One GEMAP evaluation reported that even though line rather than build it up. ministries and agencies were committed to economic gover- As further confirmation of the need for sustained nance improvements, their lack of capacity severely limited capacity-building efforts in Liberia, the national capacity their performances. This shortcoming was exacerbated by a development plan currently being formulated under the new public procurement law that, while necessary, decentral- auspices of GEMAP (the same report discusses progress on ized procurement without first building capacity to facilitate the implementation of Liberia’s poverty reduction strategy) its implementation. Overall, however, the will to retrain finds that capacity building is the primary constraint to Liberian government workers on new procedures and regu- faster progress on economic governance reform in Liberia. lations has been lacking, and the role of GEMAP in capacity- Without the capacity to implement, the long-term systems building initiatives has not always been constructive. upon which sustained success depends will remain elusive. In hindsight, GEMAP should have paid closer attention to capacity building from the start. Rather than allowing its KEY LESSONS LEARNED FROM THE LIBERIAN focus to pivot as needed based on the situation in Liberia, it ECONOMIC GOVERNANCE EXPERIENCE appears that GEMAP efforts focused on form, rather than function, with too much focus on SOEs, despite GEMAP Despite the difficulties in attributing causality to gover- reviews suggesting a change in focus. One evaluation noted nance policy interventions given the two policy changes that that, “When best practices are in place, cosignature arrange- took place in early 2006, it is possible to draw lessons from ments are not a function of control but complement capac- Liberia’s public financial management gains. CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 137 ■ Placing international advisers in postconflict environ- governance reform program should, for example, be able ments can be useful in some cases. The GEMAP pro- to provide for rapid-progress scenarios. Although gram was primarily aimed at top-down controls, and this GEMAP met with success in several of its interventions, target was appropriate and valuable in the Ministry of it focused on the control process and did not go much Finance and SOEs in the initial transitional NTGL period beyond it, which lessened its impact in terms of building given the situation at the time. Building up systems, how- institutions that go beyond the initial stabilization of ever, requires more than top-down checks on abuse. For the system. the most part, stationing advisers in the ministries both In terms of revenue collection, for example, the initial helped Liberia fill immediate capacity gaps and estab- GEMAP interventions targeted SOEs, which were impor- lished controls and set examples for those willing to fol- tant but not especially relevant for long-term fiscal gains in low them. This was dependent, however, on local author- Liberia. More generally, as pointed out in one evaluation, ities being open to the process. The cosignatory process “assistance provided by GEMAP partners outside GEMAP was useful in some cases, but GEMAP tended to be a arrangement were . . . more relevant to the specific needs focus on the process itself, rather than the function the of the various institutions” (Morsiani et al. 2008). reforms were to intended to carry out. ■ Supporting local leaders can improve economic gover- ■ Finding the appropriate role for international partners nance outcomes. Ultimately, success in economic gover- is key. In 2005 a key question in the process of economic nance reform depends on local efforts, and leadership in governance reform appeared to be whether discipline particular was a key ingredient in Liberia’s success. Addi- could be imposed from the outside. While it is simplis- tionally, local leaders took charge of programs through tic to view governance interventions on a continuum constant dialogue with donors and took ownership of between international trusteeship and full country own- the reform process. Although it was impossible to know ership, most people familiar with the situation in Liberia at the time GEMAP was signed whether the new admin- considered that a significant amount of external assis- istration would fully support the program, the new tance was needed. In the case of Liberia, however, the administration found GEMAP to be, in the words of government had changed by the time GEMAP was for- President Sirleaf, “a necessary intrusion.” Indeed, the mulated, complicating assessment of how effectively reform progress under GEMAP was only a start for the GEMAP fulfilled its intentions. Indeed, GEMAP helped government, and the impetus provided by the new gov- improve economic governance in Liberia, but only after ernment in addressing economic governance, imple- the new administration changed the parameters of its menting strong expenditure controls, and securing and operation. expanding revenue cannot be underestimated. In the ■ Partners must strike a balance between capacity build- end, strong leadership may be both the most relevant ing and delivery. Under GEMAP, international con- point and the hardest to replicate. trollers were to serve two roles in Liberia. First, they were to impose temporary controls and give the government NOTES the necessary breathing space to put long-term systems into place. Second, they were to help fill in for the lack of 1. Special SDR allocation included. capacity and help to build up the long-term systems 2. For more detailed discussions on the genesis of GEMAP, from the inside. While those two roles are interrelated, see Dod and Nelson (2008). the distinction between the two continues to matter. 3. Retrieved from http://www.gemapliberia.orsg/pages/ GEMAP was ostensibly designed to cover both, but the accomplishments. focus ended up being on delivery. Once Liberia’s top- 4. www.pefa.org. level governance problems were resolved, capacity became the binding constraint to fully rebuilding gov- ernment institutions, and by some accounts, GEMAP did REFERENCES not do enough to address that. Dod, D., and E. Nelson. 2008. “USAID Activities under ■ Provisions must be made for flexibility. The timing of GEMAP in Liberia: Impact Assessment Report.” U.S. GEMAP, which was signed during the transitional gov- Agency for International Development, Washington, DC. ernment in Liberia, required that donors be flexible in Dwan, R., and L. Bailey. 2006. “Liberia’s Governance and their support of the incoming government. An economic Economic Management Assistance Program (GEMAP): A 138 CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA Joint Review by the Department of Peacekeeping Opera- ———. 2006. “First Review of Performance under the Staff- tions’ Peacekeeping Best Practices Group and the World Monitored Program.” Country Report 06/412. Washington, Bank’s Fragile State’s Group.” United Nations, Department DC. of Peacekeeping Operations, New York. Morsiani, G., et al. 2008. “Mid-term Evaluation of the Government of Liberia. 2005. “Government and Economic Governance and Economic Management Assistance Management Assistance Program.” Monrovia. Program.” GEMAP, Monrovia. ———. 2007. “Progress under GEMAP.” Monrovia. Sirleaf, Ellen Johnson. 2006. Inauguration Speech Transcript. IMF (International Monetary Fund). 2005. “Liberia: Article www.emansion.gov.lr/doc/inaugural_add_1.pdf). IV Consultation: Staff Report.” Country Report 05/166. Washington, DC. CHAPTER 7: POSTCONFLICT ECONOMIC GOVERNANCE REFORM: THE EXPERIENCE OF LIBERIA 139 CHAPTER 8 Decentralization in Postconflict Sierra Leone: The Genie Is Out of the Bottle Vivek Srivastava and Marco Larizza hen Sierra Leone emerged from more than a W finances, and local councils are now fully staffed. Each coun- decade of conflict in 2002, it was one of the cil has a core staff of development planners, internal audi- poorest countries in the world. It faced huge tors, monitoring and evaluation officers, and procurement development challenges, with much of its infrastructure officers with requisite capacity for managing their service having been destroyed during the war. delivery functions. Both the legislation and its implementa- The areas outside Freetown had traditionally been tion leave open the possibility, however, of the center dom- excluded and marginalized. In fact, the overcentralized sys- inating and manipulating the subnational governments by tem of rule, which excluded the majority of the population, playing off the traditional authorities (chieftaincies) against was one of the key causes of the conflict. There were internal the local authorities through a “divide and rule” strategy and external pressures on the government to be more inclu- and by minimizing the autonomy of the local councils over sive to establish its legitimacy and reverse the conditions the control of financial and human resources. that led the country to conflict (Truth and Reconciliation This chapter examines the devolution of power in Sierra Commission 2004; Hanlon 2005; Kieh 2005). Leone since the end of armed conflict. The first section The reestablishment of local governments, through the briefly reviews the history of decentralization in Sierra Local Government Act of 2004, was an important initiative Leone and discusses the incentives and motivations that in this direction undertaken by the Sierra Leone People’s may have influenced the government’s decision to decen- Party (SLPP) government of Abdul Tejan Kabbah. The leg- tralize in 2004. The second section highlights the key fea- islative framework provided by the act and the associated tures of fiscal, administrative, and political decentraliza- regulations for political, fiscal, and administrative decen- tion by comparing the legal (de jure) provisions of the Local tralization (with some exceptions) provided a robust foun- Government Act with the actual (de facto) implementation dation for the establishment of decentralization through experience during the period 2004–10. The third section devolution of key functions from the central government to summarizes the major achievements of decentralization to local councils. They also provided a simple and easily date, focusing on the impact on service delivery and local understood system for intergovernmental transfers. governance. Addressing the politics of decentralization, the With financial support from donors and through the fourth section identifies potential threats and emerging evi- efforts of a set of donor-supported agencies established dence that suggests that the national government may be within government, a functioning system is now in place. trying to regain control and manipulate local politics in a Although the process has been somewhat slower than way that would be optimal for the center. The last section desired, there has been a steady devolution of functions and summarizes the main arguments and suggests politically 141 feasible options available for development partners to move staffed, with employment provided as a reward for political the decentralization agenda forward and prevent a reversal. support, and accountability for service delivery was under- mined. This dramatic retraction of local government, which coincided with the period of the All People’s Congress THE LEAD UP TO THE ESTABLISHMENT (APC) single-party rule, meant that traditional authorities OF LOCAL GOVERNMENTS represented the only form of governance in the provinces. Under colonial rule, the British authorities established a Although chiefs became increasingly active in the collection strong political and administrative divide between the of revenues on behalf of the central state, they had no ser- Colony (Freetown and the Western Areas) and the Protec- vice delivery or development functions. torate (the rest of the country, which was divided into three This reliance on chiefs, who were often repressive, and provinces) (figure 8.1). The Colony was ruled by an elected a lack of formal state structures in most of Sierra Leone local government and a British governor representing the led governments to start considering decentralization as a monarch. The Protectorate was administered through a viable option to mitigate popular discontent. Following system of “indirect rule” in which traditional authorities multiparty elections in 1996, the new government prepared (the chieftaincy) were appointed by the state for the collec- a national document entitled the “Good Governance and tion of revenue, the maintenance of law and order, and the Public Sector Reform Strategy.” The document focused on resolution of local disputes. decentralization as a major instrument for reform, espe- After independence, outside of Freetown the local cially in deprived rural areas. It suggested that decentral- administration retained many of the features of the British ization could help rebuild service delivery and improve colonial indirect rule system, but elected local councils also quality; improve resource allocation, by moving resources functioned until 1972, when President Siaka Stevens abol- to the service delivery level; ensure greater citizen engage- ished them, moving their responsibilities to the central gov- ment in the processes of government; and involve the ernment (Fanthorpe et al. 2006). Management committees community in its own development by enhancing citizens’ superseded local town councils, but they focused largely on participation in the planning and implementation of the collection of market dues. They became grossly over- development strategies and the setting of priorities. Figure 8.1 Territorial Organization of Sierra Leone a . Pr o v in c e s an d t h e W e s ter n A r e a b . D i s tri cts a nd chiefdoms IBRD 38553 MAY 2011 10° 13° 12° 11° 10° 10° 13° 12° 11° 10° GU INE A GUINEA GUINEA GUINEA Kabala Kabala KOINADUGU BIA BOMBALI Kambia NORTHERN K AM Kambia 9° 9° 9° 9° Makeni Makeni Port Loko Port Loko KONO Magburaka Magburaka Sefadu PORT FREETOWN FREETOWN LOKO TONKOLILI Sefadu URBAN WESTERN RURAL Kailahun WESTERN Kailahun Moyamba EASTERN AREA Moyamba BO UN LAH MOYAMBA KAI 8° Bo 8° Kenema 8° Bo SOUTHERN Kenema KENEMA Bonthe ATLANTIC Bonthe ATLANTIC BONTHE Pujehun Pujehun LIBERIA PUJEHUN LIBERIA OCEAN OCEAN 7° 13° 12° 11° 7° 7° 13° 12° 11° 7° Source: Zhou 2009. Note: The Republic of Sierra Leone is composed of three provinces (the Northern, Southern, and Eastern Provinces) and a region known as the Western Area, which is governed by a rural council and a city council for Freetown, the nation’s capital. The provinces are divided into 14 districts, which are divided into chiefdoms. 142 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE In 2004 the government embarked on a nationwide DECENTRALIZATION IN PRINCIPLE decentralization reform program, with the official goal of AND IN PRACTICE addressing some of the root causes of civil war and The Local Government Act and its statutory instruments improving the delivery of basic services. The decentraliza- provided the legislative framework governing decentraliza- tion program was officially launched in 2004, when the tion in Sierra Leone. With some exceptions, the act laid Parliament passed the Local Government Act and its down a robust framework for political, fiscal, and adminis- related statutory instruments. The process of decentraliza- trative decentralization in Sierra Leone. The framers of the tion was designed around two major components: the re- document took a pragmatic approach, putting in place a creation of the local councils, which had been abolished in framework that included “transitional” provisions that were 1972; and an attempt to re-create and re-legitimize the acceptable in the absence of a policy on decentralization. institutions of the chieftaincy, which had suffered greatly Policies on decentralization and chiefdom governance were during the period of one-party rule and the civil war. The to be prepared subsequently. The act recognized the need government and its development partners viewed the for change and identified the period up to 2008 as the first revival of subnational political institutions—which had phase of the arrangements for fiscal decentralization. existed before the period of one-party rule—as a primary An important area on which the act did not bring closure strategy with which to build popular legitimacy, sustain is the relationship between the local councils and the chief- political stability, and reverse the massive urban/central taincy. Although by law the chiefdoms are subordinate to bias that was a feature of Sierra Leone’s postcolonial poli- the local councils, chiefs have not accepted this hierarchy, tics and led the country into war.1 and ambiguity on the part of the national government per- Although the efficiency gains normally associated with sists. In particular, the issue of the revenue domain was not decentralization may have played a role in determining satisfactorily resolved with respect to the local tax. Under the government’s decision to decentralize, their influence existing law, the tax rate and the share of the local tax (“pre- was probably of secondary importance.2 The dominant cept”) to be paid to the council are to be determined by the factors in the choice of this strategy were the political and councils. The Chiefdom Councils are required to collect economic interests of national politicians in the SLPP this tax, presumably on behalf of the local councils, and to government, who supported decentralization for two hand over the precept to the local councils.5 Moreover, the main reasons.3 identification of the chiefdoms as a lower unit of adminis- First, the SLPP was the prime victim of prewar political tration and of local councils as the highest political author- economy: the dismantling of local councils went hand in ity at the local level is not acceptable to the chiefs, creating hand with the concentration of power in Freetown and the continuing tensions between the two institutions. The vari- consolidation of APC power in Sierra Leone (Reno 1995). ous laws governing the chieftaincy have not been repealed. The SLPP decision to decentralize in 2004 may have been The delay in addressing this issue leaves open the possibility part of a more general attempt to reconfigure political insti- of manipulation by the national government to pursue a tutions in a way that reduced the urban/central bias and the “divide and rule” strategy (Robinson 2010; Acemoglou, potential for a return to the prewar political economy, Robinson, and Verdier 2004), thereby keeping the local which tended to benefit the APC more than the SLPP. councils weaker than they could be. Second, the SLPP may have perceived decentralization Several other laws are inconsistent with the Local Gov- as a good opportunity to meet popular expectations, ernment Act 2004 or duplicate its provisions. Four laws in enhance the government’s legitimacy, and increase politi- particular place more control in the hands of the ministries cal support for the ruling party—all likely outcomes given of education, health, and energy and power than envisaged that a key source of popular discontent before the war was in the Local Government Act 2004.6 the “local despotism” (Richards 1996) of the paramount chiefs. The SLPP addressed that discontent by reestablish- Political decentralization ing the institution of local councils, which reduced the authority of the paramount chiefs. The SLPP president, The Local Government Act of 2004 identifies local councils Ahmad Tejan Kabbah, had been a district commissioner as the highest political authority in their jurisdiction. The himself and often made public his view of local councils as legislation sets out a detailed political framework covering cornerstones of democratic life and citizens’ participation the election and composition of councils; the qualifications in local politics.4 of councilors; procedures for the election of mayors (urban) CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 143 and chairpersons (rural); powers to make and execute the grant for each devolved function was to be sufficient to bylaws; the role and responsibilities of ward committees; provide the service at the standard at which it was provided and provision for citizen participation, transparency, and before devolution. The act did not provide a rule or for- accountability. The act also recognizes the laws and regula- mula for determining the vertical pool of resources to be tions governing the chieftaincy and chiefdom administra- devolved; in practice, each sector allocation was deter- tion, which were not repealed. Chiefdoms are identified mined through negotiations with the line ministry, inter- as the lowest unit of administration. The Local Government mediated by the Local Government Finance Department Act provides paramount chiefs representation in councils (LGFD).10 Allocation across councils was to be determined and membership in the ward committees. on the principal of “equity.” Equity was not defined in the Local elections in 2004 and 2008 were successfully law but, in practice, transparent formulas based on popula- completed, and transitions were peaceful. Elections were tion and existing infrastructure were devised and are being fairly competitive. In 2004, 1,112 total candidates regis- used for the horizontal allocation of grants for devolved tered with the National Electoral Commission for the 394 functions. Additional administrative grants were provided constituencies. In 84 constituencies (21 percent), coun- based on expenditure needs and fiscal capacity and indexed cilors were elected unopposed. Elections in urban areas to inflation.11 were more competitive than in rural areas, and the elections For the period after 2008, the law indicates that the vol- were more competitive in 2008 than in 2004, with the num- ume of grants should allow councils to provide devolved ber of uncontested wards dropping from 84 to 38 (less than services “at an appropriate standard” and that the annual 10 percent of all constituencies). There was a significant changes must grow at least as fast as the total budgetary turnover of councilors between the two elections, creating a appropriation made to government ministries. For all new class of local politicians, perceived by the population as grants for devolved functions, parliament retains the “the young generation of leaders on the political scene” authority to specify the functions on which these tied grants (Zhou 2009, 105). These trends suggest that channels of must be spent. Horizontal distribution is expected to political accountability are taking root at the local level, with depend on the expenditure needs and revenue-raising citizens willing to reward or punish politicians based on capacity of local councils (the specific relationships are not their performance. articulated). Transfers are significantly tied not only to sectors but to specific activities and programs, making for a large number Fiscal decentralization of separate grants. Although the formulas for the horizon- The Local Government Act provides a framework for fiscal tal allocation of grants are transparent, there is scope for decentralization for a first phase (2004–08).7 With the simplification and improvement. The design of a “second- exception of the point regarding local taxes noted earlier, it generation” grant system has been delayed and is currently clearly lays down the revenue domain of the local councils. ongoing. Three sources of financing exist for local governments in Budgeted transfers to local councils increased from Le Sierra Leone: central government transfers for devolved 19 billion in 2005 to Le 34 billion in 2009, an increase of functions and administrative expenses; local councils’ 44 percent (figure 8.2). With the exception of 2007 own revenues from taxes, fees, licenses royalties, mining (a national election year), actual transfers increased during revenues, and other sources; and loans and grants from the same period by 150 percent, reflecting the fact that other sources.8 The act recognizes that in the short term, the ratio of actual to budgeted transfers rose from about own revenues are unlikely to finance the functions 70 percent in 2005 to about 98 percent in 2009. The provi- devolved to the subnational level. It therefore provides for sional figures for 2010 indicate that the budgeted amount a “first-generation” system of intergovernmental transfers for 2010 was more than 50 percent higher than for 2009 to fill the gap.9 and that the actual transfers were higher than the budgeted The act provides for tied grants to the local councils to amounts. This is a remarkable achievement for a country carry out the functions devolved to them and to meet their with a history of centralization. Transfers as a share of cen- administrative costs. The law separates the grant arrange- tral government’s nonsalary, noninterest expenditures also ments for a first phase (2004–08) from those for the period increased, from 4.9 percent in 2005 to 11.9 percent in 2009 after 2008. The description of the goals and bases for grants On average, however, actual transfers remain low, and local is different for the two periods. For the period up to 2008, councils are still underfunded relative to the functions that 144 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE Figure 8.2 Budgeted and Actual Transfers to Local Figure 8.3 Local Council Own Revenues as Share of Councils, 2005–2010 Total Revenues, 2005–2010 55 54.86 35 33.3 32.9 50 51.42 30 Percentage 45 26.4 Billion leones 40 35.0 33.2 33.7 35 25 24.5 24.5 30 25.4 33.1 20 25 19.0 25.2 18.8 20 25.6 15 15 2005 2006 2007 2008 2009 2010 10 13.2 7.5 5 Source: IRCBP 2010b. 2005 2006 2007 2008 2009 2010 Note: Figures for 2010 are provisional. Budgeted transfers Actual transfers Source: IRCBP 2010b. Note: Figures for 2010 are provisional. waste management, youth and sport activities, and some fire and social welfare functions. Central ministries and agencies retain responsibility for strategic planning, setting have been devolved to them. Given these circumstances, the of standards, quality control, and monitoring, as well as move to a more rational “second generation” of grants is procurement of certain priority commodities, such as text- long overdue. books and drugs. Central government also retains adminis- Not surprisingly, revenue generation by the local coun- trative control over staff responsible for performing cils continues to be weak, with local councils on average devolved functions. able to finance only about 25–30 percent of their expendi- The devolution of functions has been slower than tures from their own revenues (Searle 2009). Local taxes expected, with varying degrees of responsiveness. Although and property taxes are the sources of revenues for the local the act envisioned completion of the transfer of authority councils. The main sources of nontax revenues are market by 2008, by mid-2010 only 46 of 80 functions had been for- dues, business registrations, license fees, and mining royal- mally devolved to local councils (Decentralization Secre- ties (in selected councils). tariat 2010). Important areas such as the devolution of Local council revenues as a percentage of total revenues feeder roads remain politically contentious. indicate that limited and uneven improvements have taken Each local council has a political head (mayor and deputy) place over time (figure 8.3). This trend is slightly more pos- and administrative head (chief administrator) (figure 8.4). itive if one considers revenues in absolute terms: in 2008 the The chief administrator is responsible for management and local councils collected Le 8.3 billion, a substantial increase all administrative and technical matters. In addition, each of over the Le 4.7 billion collected in 2005. Revenues in 2009 the local councils now has its complement of core technical were Le 16.2 billion, an increase of 95 percent over 2008.12 staff under the supervision of the chief administrator. How- There was a further increase of about 10 percent between ever, staff for devolved functions continue to remain under 2009 and 2010 although the share of own revenues in total the administrative control of the central ministries, depart- revenues fell due to the significant increase in the total vol- ments, and agencies. Under the new decentralization policy ume of transfers. (Government of Sierra Leone 2010a), staff for devolved functions are to be fully devolved to the local councils by 2016. In the interim, it is proposed that administrative con- Administrative decentralization trol of local councils over such staff will be enhanced The statutory instrument accompanying the Local Govern- through “letters of deployment.” ment Act 2004 provided a detailed framework and timetable for devolving functions housed in 17 ministries, depart- MAJOR ACHIEVEMENTS OF ments, and agencies to the local councils. The functions that DECENTRALIZATION were to be devolved between 2005 and 2008 included pri- mary and mid-secondary education, primary and secondary Sierra Leone’s success in reestablishing local government in health facilities, feeder roads, agriculture, rural water, solid a fragile postconflict environment is notable. Despite some CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 145 Figure 8.4 Intergovernmental Relationships under Decentralization CENTRAL GOVERNMENT Ministry of Finance Ministry of Local Government and Line Local Government Rural Development ministries Financial Department Decentralization Secretariat Capacity support 3 provinces (deconcentrated units) Grants 19 local councils Staff for District devolved functions devolved officers (LGA 2004) functions Share of 149 chiefdoms local taxes paramount chiefs • Local taxes • Land Source: Authors. Note: The shaded boxes identify central government institutions and deconcentrated units (provinces, district officers). The dashed box identifies the key actors of decentralized administration. The solid arrows indicate a clear relationship of authority across levels of government. The dotted arrows denote the lack of clarity in defining relationships between government institutions, including the recently re-introduced district officers and local councils; local councils and chiefdoms; and local councils and local staff, who are formally accountable to the local councils but appointed by the central administration. The hollow downward-slanting arrow indicates the flows of resources from the central government to local councils (resources come from capacity sup- port from the Ministry of Local Governance and grants from the Ministry of Finance). The hollow upward-slanting arrow indicates the flows of resources (share of local taxes) from chiefdoms to local councils. Historically, there were 149 chiefdoms established under the Chiefdoms Councils Ordinance 1938 (and subsequent amendments). A local council thus has more than one chiefdom within its jurisdictions. of the shortcomings outlined above, decentralization is local governments are able to work with centrally managed now well established, arguably more strongly on the politi- frontline staff to manage service delivery in the areas cal dimension than on the others. Two council elections devolved to them. The pace at which local councils assumed have been completed; all local councils have the core staff full identity as democratic, effective, and legitimate political to carry out planning, budgeting, accounting, and procure- actors and institutionally oriented themselves to discharg- ment functions; a system of intergovernmental transfers is ing devolved functions has been a notable feature of the in place; and, although there is scope for improvement, decentralization process. 146 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE Service delivery 90.6 percent (see table 8.1). Between 2006 and 2008 the largest gains were in the number of staff, particularly sen- With all councils decentralizing at the same time, it is dif- ior staff. Although progress is being made in filling vacant ficult to rigorously establish the extent to which improve- positions, the percentage of clinics open fell between 2006 ment in services is a result of decentralization.13 That and 2008, from 88 percent in 2006 to 82 percent. Also, the said, on average the quality of services did not decline, percentage of clinics receiving supervision by Ministry of and in some cases there have been significant improve- Health and local council members declined after 2005. ments, according to data from a series of national public Despite decentralization, local councilors in particular do service and sector (health and education) surveys carried not appear to be taking a strong supervisory role, with only out by the evaluation unit of the Institutional Reform and one in four clinics receiving a visit from a councilor in the Capacity Building Project Evaluation Unit (IRCBP) year preceding the survey. These trends suggest that more together with the Abdul Latif Jameel Poverty Action Lab effective supervision by the local councils and ministry (table 8.1).14 staff, rather than more hiring, should be the main focus Communities far from Freetown but close to a district moving forward. capital saw the biggest improvements in services, even holding constant the remoteness of these communities. A reduction in “distance from power”—a direct byproduct of Education. Education has seen less devolution than other decentralization—thus appears to have had a positive sectors. Because only one detailed survey of school quality impact on service delivery. has been undertaken by the IRCBP, it is harder to assess These improvements are remarkable if one considers the gains in education. Household surveys show that more short time since the launch of decentralization and the fact rural Sierra Leoneans are within reach of a primary school that local government performance continues to be con- than they were in 2005. The percentage of households with strained by several factors, including the incomplete devo- access to a school within 30 minutes’ walking distance lution of functions and line staff, the relatively small size increased from 68.3 percent in 2005 to 74.3 percent in 2008 and tied nature of the transfers, and the unresolved ten- (see table 8.1). Overall public satisfaction with primary sions between local councils and the traditional authorities. schools has improved. Informal school fees remain high, however, a key reason why children are not in school. Health. Health is the sector that has progressed furthest Although only 3 percent of schools were reported to be on devolution, with about $3 million, just less than one closed, teacher absenteeism was estimated at 22 percent in quarter of the national health budget, budgeted to grants to 2005, highlighting the lack of adequate supervision. local councils as early as 2006. Access to and quality of health services have improved dramatically since 2005, Agriculture. Access to drying floor and storage space with most of the gains taking place between 2005 and 2006. improved sharply between 2005 and 2007, although most Clinic infrastructures, availability of drugs, and numbers of households still lack access. Only 18 percent of households staff have all improved, with the result that public satisfac- had contact with an extension worker in 2007, down from tion with health services improved from 81.0 percent to 23 percent in 2005. Nearly half of farmers sell their produce Table 8.1 Quality of Service Delivery Reported in Household Surveys, Selected Years (percent of respondents) Service indicator 2005 2007 2008 Access to school within 30-minute walking distance 68.3 73.9 74.3 Satisfaction with primary schools 87.7 94.4 90.3 Satisfaction with health clinic 81.0 90.9 90.6 Spoke to an extension worker in past year 23.0 17.8 9.0 Access to sufficient storage space (farming households only) 8.4 11.8 14.3 Drivable road within 30-minute walking distance 67.1 73.2 77.5 Market area within 60 minutes 31.9 45.8 50.9 Water source within 15 minutes 61.0 73.4 80.9 Source: IRCBP 2010. CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 147 to traders who come to the village, a third sell at the market in rural communities and areas without transport or themselves, and the rest sell to a trader at market. telecommunications (IRCBP 2010). Other sectors. Although the Local Government Act was Accountability and participation. Provisions of the supposed to devolve responsibility for the rehabilitation Local Government Act stipulate that councilors must and maintenance of roads to local councils, devolution has declare their assets and councils must maintain an inven- not yet taken place, and no money has been transferred tory of assets, print receipts, and maintain other accounting from the central government to the local councils for this documents. Notice boards are also to be maintained in function. Resistance by the powerful Sierra Leone Roads wards and at the council displaying financial information Authority is the main reason for this delay.15 and strategic documents, such as development plans and No central government agency is responsible for mar- procurement contracts, to enhance accountability and kets, another area in which local councils do not yet receive information sharing with the community. To ensure greater transfers. Local councils have, however, spent a significant accountability and support a participatory approach to the share of their discretionary funding on roads and markets, decision-making process, ward committee members partic- which are perceived as key means to increasing the effi- ipate in council meetings and review the council’s develop- ciency of the agriculture sector. ment plan. It is the responsibility of the ward committee to Access to markets in rural areas remains weak, with prioritize the community’s development needs, which are half of the rural population having to travel more than an then finalized and passed at the council meeting. hour to reach a market. As indicated in table 8.1, however, Evidence suggests that local communities are becoming there have been significant and steady improvements: in increasingly active in demanding services and holding local 2005 only 32 percent of rural respondents had a market authorities accountable, looking at local elections as a means less than 60 minutes away. In 2008 this figure increased to to achieve the promise of development.16 Participation in 51 percent. Road access also improved, but the percentage local elections remains moderately high—albeit lower than of respondents having access to regular public transport in national elections—with civic activism tending to be declined between 2005 and 2008. Access to water and higher in remote areas than in urban areas.17 water sources saw significant improvements: in 2008 Marginalized groups, such as women and ethnic about 81 percent of respondents had access to a water minorities, have been the largest beneficiaries of the new source within 15 minutes, a sharp increase from the 2005 space for political participation. In 2004 women occupied baseline (61 percent). about 13 percent of council seats in council elections; in 2008 this share increased to 18 percent, more than three times the 5 percent share of seats that women have in the Local governance national Parliament (IRCBP 2010). In 2004 and 2008 repre- A central argument in support of decentralization is that it sentatives from minority ethnic groups such as the Kono, brings government closer to the people. By expanding the Loko, and Sherbro, were elected to local councils (Zhou political space, decentralization is expected to allow greater 2009)—a remarkable achievement given Sierra Leone’s his- government accountability and citizen participation to pub- tory. Decentralization has thus made significant contribu- lic affairs, hence strengthening state legitimacy (Bardhan tions to promoting prospects of political stability by allowing and Mookherjee 2006). greater participation and power-sharing dynamics. In Sierra Leone national elections tend to become “win- ner takes all” games focused on region-based ethnic identi- Performance of local councils. In 2006 the Decen- ties. The role of ethnicity is likely to be less relevant at the tralization Secretariat, supported by the IRCBP, designed local level, where voters are more likely to share a common a Comprehensive Local Government Performance Assess- ethnic background. Decentralization may provide citizens ment System (CLoGPAS) tool to serve as a sustainable with greater opportunities to monitor the performance of local council management accountability mechanism for local authorities on service delivery and hold them account- the local councils.18 The first assessment was carried out able through elections. The experience of Sierra Leone sug- in June 2006. A follow-up assessment was conducted in gests that decentralization is indeed contributing to better early 2008. The results show that performance of local local governance by providing greater scope for citizens’ par- councils is improving in several development and man- ticipation and engagement with local authorities, especially agement functions. 148 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE Figure 8.5 Number of Local Councils Meeting Minimum councils making the largest gains among political institu- Conditions, 2006 and 2008 tions between 2007 and 2008 (table 8.2). Together with chief- dom officials, local (political) authorities are more trusted 18 by citizens than national government officials. These 16 14 trends are encouraging, although some of the findings may 12 partly reflect a temporary boost from the publicity sur- Number 10 rounding the July 2008 local council elections. This line of 8 argument seems to be (indirectly) confirmed by the findings 6 4 of the 2008 National Public Services survey that trust in local 2 councils is higher among citizens who are more actively 0 engaged with local politics. t ng g t cy io , ity nc luat ion en en tin en ni ac na n em em un t lan ev nta ar ap co ag ur sp lc tp nd me an oc ac an THREATS AND CHALLENGES en a g a ple lm Pr Tr nd tio pm rin im cia ga elo Fu an in ito ct Six years after the decentralization initiative was launched, it ev et on je Fin m Pro dg D has arguably taken firm root, leading several observers to Bu suggest that it would be very hard to reverse the process and 2006 2008 concentrate power in Freetown again. Thus, from a Source: Decentralization Secretariat 2007, 2009. Freetown-centric perspective, decentralization is a “genie” that is out of the bottle. Recent developments, however, suggest that pressures might be building to weaken decen- Figure 8.5 shows the number of councils that fully met tralization, casting some doubt on the central government’s the seven “minimum conditions” of the CLoGPAS.19 It intentions to unequivocally move the agenda forward. indicates that the number of local councils that met the Ironically, this apparent weakening may well be a result of minimum conditions rose between 2006 and 2008 in all but the successes of the initiative. Rather than empowering two areas (transparency and, to a lesser extent, functional local councils further, national elites seem to be following a capacity). Local councils hold regular meetings and pro- strategy of “divide and rule,” preventing local councils from duce minutes, citizens participate in meetings, participa- becoming strong enough to seriously challenge the political tory development planning has taken root (although there hegemony of the center. is scope for improving its quality), accounts are completed on time, and regularly audited and financial information is Stance of the national government disclosed. These promising trends have been confirmed by the latest (draft) report on public expenditure and financial Tensions between the local councils and the chieftaincy accountability (Government of Sierra Leone 2010b), which result largely from a lack of clarity about their respective assesses the performance of central and subnational gov- domains (Fanthorpe 2005; Sawyer 2008). It is difficult to ernment authorities across various dimensions of public judge whether this situation is intentional or reflects simple financial management.20 According to the report, in 2010 local councils received the highest scores on key dimen- Table 8.2 Citizens’ Trust in Public Institutions, sions of budget execution (namely, competition, value for 2007 and 2008 money, and controls in procurement) and accounting prac- (percent of respondents) tices (namely, timeliness and regularity of accounts recon- Type of official 2007 2008 ciliation), scoring higher on these dimensions than the Justice sector national government. This performance is remarkable in Police 36 38 view of the fact that local councils were established as Local court 41 45 Magistrate court 34 37 recently as 2004. Political authorities Central government 42 44 Citizens’ trust in public institutions. Results from Chiefdom 47 51 National Public Services surveys indicate that public confi- Local council 33 49 dence in local councilors has increased over time, with local Source: IRCBP 2010. CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 149 delays in decision making, although some evidence suggests senior government functionaries have remarked that the that it is intentional. A national decentralization policy has local councils—whose key function is development—were recently been drafted, but a critical companion piece—the becoming political. Although these actions by the central policy on chiefdom governance—has not been prepared. government may be a response to real political change The Chieftaincy Act of 2009 codifies and adds to customary engendered by decentralization, they represent a potential law on the election and removal of chiefs. With the enact- threat to the autonomy of the local councils and, more ment of this law, any reforms aimed at the democratization generally, to prospects for further strengthening decen- of the chieftaincy are unlikely, at least in the medium term. tralization in Sierra Leone. The incentives of members of Under the law, the central executive has effective leverage parliament (MPs) are aligned with those of the central over the chieftaincy, because it plays an important role in executive. The national election of 2002 was based on the election and removal of paramount chiefs. The ministry proportional representation and the MPs did not have responsible for local government has traditionally dealt with geographical constituencies. With the change in 2007, the chiefs and even today is much more comfortable with MPs represent single-member constituencies and now see this function than with its new role with respect to the local the local councils as competitors. councils.21 These links are important for the national elec- tions because, by some estimates, the chiefs are able to influ- Champions of change ence 10–20 percent of voters in their jurisdictions. In 2009, contrary to the provisions of the Local Govern- The Decentralization Secretariat (Dec Sec) and the Local ment Act 2004, the minister for local government advised the Government Finance Department (LGFD) have played crit- chiefs not to share local tax revenues with the local councils. ical roles in implementation and have been strong advocates Since then, also contrary to the provisions of the law, the for decentralization. Both are extragovernmental agencies ministry has played a role in determining the amount of the created under the Institutional Reform and Capacity local tax precept, establishing a range of 0–20 percent for Building Project financed by the International Development different classes of chiefdoms. Until 2008 most chiefs were Association (IDA) and a multidonor trust fund financed by paying a precept of 60 percent. This ministry action will the European Commission and the U.K. Department for undermine the fiscal autonomy of the local councils. International Development. Although, in principle the Dec The National Decentralization Policy approved by the Sec is an arm of the Ministry of Local Government and Rural cabinet in September 2010 provides strong support to the Development and the LGFD is a unit in the Ministry of hypothesis of a containment strategy by the center to limit Finance and the secretariat to the Local Government Finance political power of local governments. According to the pol- Committee (LGFC), all of the staff and expenses associated icy, “Local councils shall continue to exist as the highest with these agencies are currently financed by the IRCBP.22 development and service delivery authority” (Government With the project due to close soon, it is not entirely clear of Sierra Leone 2010a). This policy is inconsistent with the how and how well their functions will be integrated into Local Government Act (2004), which defines local councils government. Without the mainstreaming of these functions, as the highest political authority at the local level. The pol- the future of decentralization in Sierra Leone will be in jeop- icy also reintroduces the position of district officers. ardy. If progress on this issue of mainstreaming continues to In June 2010 the APC government decided to reintro- be slow, donors will need to make a judgment fairly soon duce the post of district officers, who traditionally repre- about whether the government is genuinely handicapped by sented the national government in the districts and financial and capacity constraints or whether this lack of provided the links with the chieftaincy. The official explana- progress reflects apathy toward the decentralization agenda. tion is that there is a need to establish a stronger channel of communication between the national government and the Dependence on national government chiefs. Representatives from the SLPP opposition and civil society groups, however, have articulated the concern that The local councils are highly dependent on grants from the the district officers—acting as representatives of the central national government. Although the volume of these grants government at the local level—will try to influence decision has increased over time and a larger proportion of budgeted making by the local councils, which may further reduce the amounts are now being transferred, the lack of financial councils’ financial autonomy and increase the control of the autonomy leaves the local councils very vulnerable to the ruling party as the 2012 elections draw closer. Several whims of the national government. The lack of control over 150 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE frontline staff weakens the ability of the local councils to and help local authorities in their ongoing efforts to meet influence the quality of services. Both of these risks are exac- citizens’ demands and perform the functions devolved to erbated by the fact that the national government appears to them. Donors have played an important role so far. The want to keep the local councils weak. main instrument has been the IRCBP, which, through the Dec Sec and LGFD, has supported the establishment of a simple but robust intergovernmental grant system; made LESSONS LEARNED AND CONCLUDING resources available for small capital works; and, probably REMARKS most important, supported the development of capacity in Decentralization has had a significant positive impact on the the local councils to enable them to perform their core political landscape in Sierra Leone. The sharing of political functions and become established as credible governments. power beyond Freetown is unprecedented; on this dimen- On some dimensions, such as procurement, capacity is sion alone, the initiative can be viewed as a major success. rated to be better at the subnational level than at the There is scope for improving the capability of local coun- national level. Continuing support for strengthening ser- cils to deliver on development outcomes. The volume of vice provision at the local level and enhancing the resources resources available to the local councils remains small, and and autonomy of the local councils is being provided the relatively low level of own revenues and dependence on through the $20 million IDA–financed Decentralized Ser- central grants undermines accountability, limits auton- vices Delivery Project (DSDP). This second-generation omy, and makes the local councils vulnerable to manipu- project—which tilts the balance toward significantly aug- lation by the center. menting the grants provided by the central government Functioning and effective local governments provide a while providing support for a “maintenance” level of capac- viable and promising alternative to the persistent patrimo- ity development—is expected to attract additional donor nial nature of Sierra Leone’s national politics for a shift contributions. An increasing allocation of donor resources toward public goods and better service delivery to the to the subnational level through well-designed projects will poor.23 In particular, the local councils are in a position to strengthen both the autonomy and the capacity of local contribute to the improvement of services and the provi- councils and help mitigate countervailing pressures from sion of public goods in the social sectors and in those the national government. In the authors’ view, there is a aspects of infrastructure (water and sanitation, feeder continuing need for building technical capacity beyond roads, off-grid power) that can efficiently be provided at what was originally envisaged in the DSDP and, more the subnational level. important, for dialogue and advocacy. Although it is unlikely that any government will openly These donor initiatives will need to be complemented seek to recentralize power in Freetown, as the APC led by with additional reform efforts to improve the prospects of Siaka Stevens did in 1972, there are risks that the center will decentralization and create the conditions for a greater try to undermine the emerging political power at the sub- impact on service delivery: national level. The reintroduction of the district officers and the downgrading of the political status of the local ■ A second-generation system of grants needs to be councils provide some evidence of a new containment strat- designed and implemented as soon as possible to egy designed to protect the interests of the national elites. enhance predictability and autonomy and provide incen- These efforts by the national government to undermine the tives for efficiency. The divisible grant pool needs to be autonomy and strength of local councils may be the ulti- determined based on a set of well-understood forward- mate outcome of an effective decentralization process—the looking criteria, the allocation across councils should be “paradox of success”—implicitly suggesting that local gov- based on an agreed philosophy for equalization, and ernments are indeed emerging as important players in the grants need to be gradually untied to provide greater dis- political and development landscape of Sierra Leone. The cretion to the local councils. risks associated with the containment strategy of the center ■ Local councils need to be gradually provided with greater suggests that donors and civil society groups will need to administrative control over frontline staff associated play an important role, however, if decentralization is to suc- with service delivery. As noted, the provisions of the new ceed and central government efforts to undermine it are to national decentralization policy identify a vision for be neutralized. Continuing engagement by donors and key 2016, when comprehensive devolution of staff to the national stakeholders is crucial to support decentralization local councils should have taken place and interim CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 151 arrangements made during which authority is gradually instability, by raising the stakes of politics, making the cen- handed over to the local councils. While this is, in prin- ter attractive to capture in a zero-sum game. ciple, a sensible approach the transfer of responsibility to 2. The standard arguments about the benefits of decen- local councils will have to be carefully balanced with tralization focus on the efficiency of service delivery. The local council capacity to undertake this responsibility. basic idea is that by bringing representatives closer to the ■ Issues concerning the functions and jurisdictions of sub- people, decentralization leads to socially more desirable national institutions need to be resolved to ensure that service provision, because local politicians are more local councils and paramount chiefs work better accountable and because they have better information about people’s preferences and what needs to be done. together.24 This strategy would enable subnational polit- Arguments that link decentralization to better service deliv- ical actors to present a more united front, which would ery may be particularly compelling in situations in which be effective in strengthening the decentralization process there are important ethnoregional divisions in a society, while mitigating the risk of manipulation from the cen- making it difficult to reach agreement about which policies ter. The national government needs to take the lead in to adopt or public goods to supply (Robinson 2010; see also clarifying functions and jurisdictions. Bardhan 2002 and Eaton, Kaiser, and Smoke 2010 for a ■ Across all devolved functions, improved supervision and review of this argument). monitoring of service delivery by local councils and cen- 3. An additional factor that may have driven the SLPP to tral ministries is needed to ensure that implementation decentralize can be found in the use of decentralization as meets required standards. an electoral strategy to ensure political survival at the ■ As more resources become available at the subnational local level in the event of a loss of power at the national level, the risk of rent-seeking behavior and corruption elections. Although theoretically plausible and historically will increase. Close watch will need to be kept to contain relevant in other contexts (O’Neill 2003), this kind of opportunities for corruption and elite capture; additional electoral incentive was hardly a dominant factor in Sierra Leone. Historical evidence suggests that at the time decen- resources and autonomy will have to be complemented tralization was launched (2004), the SLPP was confident with robust accountability arrangements involving the of remaining in power and that it remained confident of ward committees and citizens’ participation. doing so until early 2007, when the sudden realization of possible electoral loss prompted uncoordinated and inef- Key country stakeholders and donors took advantage of fective actions (see Kandeh 2008 for an excellent analysis of the window of opportunity that became available after 2002 the 2007 elections). by supporting the enactment and implementation of the 4. At the launch of the Local Government Reform and Local Government Act. They did so knowing that it was not Decentralization Programme, on February 20, 2004, Presi- a perfect document and that several loose ends needed to be dent Kabbah remarked, “People have the right to elect the resolved. Their actions helped to establish a system that is leaders, men or women, who are supposed to serve them at reasonably robust and to create a new class of subnational the national level, in Parliament. They also have the right to actors that is gaining central prominence in shaping new choose those who should serve them at the district and political and economic dynamics. The genie is now out of other local levels. This, in my view, is what the principle of the bottle. The question is whether it will be able to work its democratic decentralization is all about. It should and must remain a cornerstone of the process of nation-building in magic or whether it will be “contained” by the center. the country.” Donors, together with civil society and other stakeholders, 5. The Local Taxes Act of 1975 identified the “local” tax as will have to play a critical role in maintaining momentum a tax to be collected by the “local authority” and paid into along the trajectory that has been established, including the Consolidate Revenue Fund/Accountant General. Under advocacy initiatives and continuous policy dialogue with this act the rate of the tax was to be determined by the min- the government. Any letup in effort at this stage could jeop- ister for internal affairs. Under this formulation, the local ardize the initiative and compromise the remarkable tax is collected by the “local authority” on behalf of the progress achieved since 2004. national government. Section 45 (4) of the Local Govern- ment Act 2004 made the “precept” payable to the local coun- cils but did not make any reference to the Local Taxes Act NOTES 1975, which is still on the statute books. 1. This massive bias not only created large socially unde- 6. The laws are the Education Act, the Hospitals Board Act, sirable biases in resource allocation, it also led to political the SALWACO Act, and the Local Tax Act. 152 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 7. See Fox 2009 and Searle 2009 for detailed discussions of and development second; for local elections the promise of fiscal decentralization in theory and practice. development was the priority (Zhou 2009). 8. Under section 65 of the Local Government Act 2004 and 17. According to the latest National Public Services survey section 17 of the Public Debt Management Bill 2010 data, respondents were significantly more likely to report (gazetted but not enacted at the time of writing), local gov- voting in national (87 percent) than local council (77 per- ernments can borrow domestically up to limits agreed with cent) elections. Official electoral data show substantially the Minister of Finance. This borrowing need not be guar- lower participation rates in local elections, indicating a anteed by the central government. So far, local governments decrease from 55 percent in 2004 to 39 percent in 2008. have not used this option as a source of financing and local 18. The design and implementation of CLoGPAS involves government debt is not currently a problem. the setting up of a multidisciplinary task team comprising 9. We refer to the grant arrangements envisaged under the technical staff of IRCBP and the Ministry of Internal Affairs, Local Government Act 2004 for the period 2004–2008 as the Local Government and Rural Development. “first generation” system of intergovernmental transfers. 19. The minimum conditions deal with aspects of local The design of a “second generation” system is ongoing and council management accountability and examine func- has not been completed. tional capacities of the local councils in terms of their 10. The LGFD is a unit in the Ministry of Finance that per- preparedness to take over devolved functions and deliver forms the secretariat functions for the Local Government services at acceptable standards. They also assess/review the Finance Committee (LGFC), the entity that determines the compliance of local councils with existing laws and regula- volume and distribution of grants under the law. tions that guide the decentralization process. 11. In practice the administrative grants are in two parts. 20. More precisely, the public expenditure and financial The first part covers the sitting fee and transport allowances accountability report “first examines the credibility of the of councilors and is based on the number of councilors. The Budget as a tool for implementing government policy, and second part finances their general administrative expenses; then looks at two key crosscutting issues relating to Public it is positively related to expenditure needs for undertaking Financial Management (PFM), the comprehensiveness and revenue collection and administering devolved functions transparency of PFM systems. It then rates performance and negatively related to fiscal capacity. through the four key stages in the budget cycle: budget for- 12. Several factors (beyond decentralization) may have mulation, budget execution, accounting and reporting and driven the positive results in 2009; considered alone, this finally external scrutiny and audit. Under each dimension, a sharp improvement is not sufficient to establish a trend. set of performance indicators is identified, and scoring cri- 13. For example, the fact that satisfaction with health teria is set out” (Government of Sierra Leone 2010b). improved more sharply than satisfaction with education 21. In 2004 the ministry was known as the Ministry of does not necessarily reflect the fact that health decentralized Local Government and Community Development. In 2007 earlier than education. It may be that health decentralized it became the Ministry of Internal Affairs, Local Govern- more rapidly than education because there was greater ment and Rural Development. In November 2010, Internal capacity in the Ministry of Health and that satisfaction with Affairs was hived off, so the ministry is now the Ministry health improved more because of that capacity. of Local Government and Rural Development. 14. The National Public Services survey was administered 22. The IRCB Project was launched in mid-2004, with a three times between 2005 and 2008. The quality of services World Bank credit of $25.1 million. Financing was at health clinics was surveyed three times, and a baseline for extended in 2006 with an additional $25 million provided education quality was created in 2005. The surveys include by the Department for International Development and the large, nationally representative samples of households, clin- European Union through a multidonor trust fund. ics, and primary schools; the data collected can be com- 23. See Robinson (2008) for a compelling analysis of the pared over time. sources of patrimonialism in Sierra Leone. Robinson 15. The Sierra Leone Roads Authority is a powerful inde- (2010) further elaborates on the political economy of pendent statutory body established in 1993 responsible for decentralization, analyzing the reasons why local govern- policy formulation and implementation in the roads sector. ments are expected to be less subject to the capture of pat- 16. A household survey by GoBifo (Sierra Leone’s rimonial politics. community-driven development project) and the IRCBP in 24. Fanthorpe and Sesay (2009) make a number of well- 2009 in the Bonthe and Bombali districts asked respondents informed and constructive recommendations for how to their main reasons for voting. It found that in national elec- reform the chieftaincy and make it work better with local tions, voters ranked political party as the most important councils. CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE 153 REFERENCES Hanlon, Joseph. 2005. “Is the International Community Helping to Recreate the Preconditions for War in Sierra Acemoglu, Daron, James A. Robinson, and Thierry Verdier. Leone?” The Round Table: The Commonwealth Journal of 2004. “Kleptocracy and Divide-and-Rule: A Model of International Affairs 94 (381): 459–72. Personal Rule.” Journal of the European Economic Associ- ation 2 (2–3): 162–92. IRCBP (Institutional Reform and Capacity Building Project Evaluation Unit). 2010. Report on the IRCBP 2008 Bardhan, Pranab K. 2002. “Decentralization of Governance National Public Services Survey: Public Services, Gover- and Development.” Journal of Economic Perspectives 16 nance, and Social Dynamics. Freetown. (4): 185–205. Bardhan, Pranab, and Dilip Mookherjee, eds. 2006. Decen- IRCBP (Institutional Reform and Capacity Building Project tralization and Local Governance in Developing Countries: Evaluation Unit), and J-PAL (Abdul Latif Jameel Poverty A Comparative Perspective. Cambridge, MA: MIT Press. Action Lab). 2009. Basic Services and Decentralization in Sierra Leone: Trends and Lessons. Abdul Latif Jameel Decentralization Secretariat. 2007. Comprehensive Local Poverty Action Lab, Massachusetts Institute of Technol- Government Performance Assessment System (CLoGPAS): ogy, Cambridge, MA. 2006 Summary Report. Freetown. Kandeh, Jimmy. 2008. “Rogue Incumbents, Donor Assis- ———. 2009. Comprehensive Local Government Perfor- tance and Sierra Leone’s Second Post-Conflict Elections mance Assessment System (CLoGPAS): 2008 Summary of 2007.” Journal of Modern African Studies 46: 603–35. Report. Freetown. Kieh, George. 2005. “State-Building in Post-Civil War Sierra ———. 2010. Status Report on the Devolution of Functions Leone.”African and Asian Studies 4 (1–2): 1. to the Inter-Ministerial Committee. Freetown. O’Neill, Kathleen. 2003. “Decentralization as an Electoral Eaton, Kent, Kai Kaiser, and Paul Smoke. 2010. The Political Strategy.” Comparative Political Studies 36 (9): 1068–91. Economy of Decentralization Reforms in Developing Countries: A Development Partner Perspective. World Reno, William. 1995. Corruption and State Politics in Sierra Bank, PRMPS, Washington, DC. Leone. New York: Cambridge University Press. Fanthorpe, Richard. 2005. “On the Limits of the Liberal Richards, Paul. 1996. Fighting for the Rainforest: War, Youth Peace: Chiefs and Democratic Decentralization in Post- and Resources in Sierra Leone. Oxford: James Currey. War Sierra Leone.”African Affairs 105 (418): 27–49. Robinson, James. 2008. Governance and Political Economy Fanthorpe, Richard, and Mohamed Gibril Sesay. 2009. Constraints to World Bank CAS Priorities in Sierra Leone. Reform Is Not against Tradition: Making Chieftaincy Rev- World Bank, AFTPR, Washington DC. elation in 21st Century Sierra Leone. Campaign for Good ———. 2010. The Political Economy of Decentralization in Governance, Freetown. Sierra Leone. World Bank, AFTPR, Washington, DC. Fanthorpe, Richard, Taylor Brown, Janet Gardener, Lansana Sawyer, Edward. 2008. “Remove or Reform? A Case for. Gberie, and M. Gibril Sesay. 2006. Sierra Leone: Drivers of Restructuring. Chiefdom Governance in Post-Conflict.” Change. IDLgroup, London. African Affairs 107: 387–403. Fox, William F. 2009. “Devolution and Grant Structures for Searle, Bob. 2009. Sierra Leone: Issues in Fiscal Decentraliza- Local Councils in Sierra Leone.” World Bank, AFTPR, tion. World Bank, AFTPR, Washington, DC. Washington, DC. Truth and Reconciliation Commission. 2004. Final Report of Government of Sierra Leone. 2010a. National Decentraliza- the Truth and Reconciliation Commission of Sierra Leone. tion Policy. September. Freetown. Freetown. _____ . 2010b. Public Expenditure and Financial Accounta- Zhou, Yongmei, ed. 2009. Decentralization, Democracy and bility (PEFA) Sub-National Government Summary Development: Recent Experiences from Sierra Leone. Report. Freetown. World Bank, AFTPR, Washington, DC. 154 CHAPTER 8: DECENTRALIZATION IN POSTCONFLICT SIERRA LEONE: THE GENIE IS OUT OF THE BOTTLE CHAPTER 9 Transport Infrastructure and the Road to Statehood in Somaliland Jean-Paul Azam ow can economic success emerge in the midst of H traditionally fairly porous, as might be expected in a land political chaos and civil war? This is the question mostly devoted to nomadic pastoral activity. Refugees have raised by the experience of Berbera, a successful crossed this border back and forth over the past decades, port located on the southern shore of the Gulf of Aden, in depending on the changing intensity of fighting during the the former British protectorate of Somaliland. Berbera’s Ethiopian civil war, which ended in 1991, and the ongoing traffic has been rising steadily over the past decade. For one in Somalia that began in 1991. The best grazing land, example its traffic of import containers has nearly doubled called the Haud, straddles this border, which nomadic between 2003 and 2007 (MNPC 2010). This port’s promis- herdsmen also routinely cross (Doornbos 1993, map 6.1, ing position is confirmed by the fact that Bolloré Africa 101). While ethnically homogenous, these Somali-speaking Logistics, the biggest port operator in Africa, has recently people of Djibouti, Ethiopia, Kenya, Somalia, and Soma- announced its plan to invest massively in Berbera’s port liland are traditionally affiliated with different clans that and its transport corridor with the Ethiopian capital city, straddle these countries’ borders and that at times enter Addis Ababa.1 into violent conflict with one another. The chaos that pre- Somaliland’s small population, estimated at about vails in most of Somalia, which has been engulfed in war- 3.5 million people, benefits from two main kinds of physical lordism and banditry ever since the collapse of the state in assets: its pastoral assets of grazing land and livestock, on the 1991, testifies to the threat of violence that looms over these one hand, and its transport infrastructure, on the other people. Siyad Barre’s military government had launched an hand. The latter includes mainly the port of Berbera and, to unsuccessful invasion of Ethiopia’s Ogaden region in 1978 a lesser extent, the airport of Hargeysa, its capital city, as well and eventually signed a peace agreement with the Mengistu as their connecting roads. The nomadic herdsmen need the regime in 1988, which was perceived as a disgrace by many port of Berbera for exporting their livestock to the Arabian Somalians. Many military officers from the north, mainly Peninsula and other Middle Eastern countries, which pro- from the Isaaq clan, which is the largest one in Somaliland, vide their main outlet. Moreover, Berbera is ideally located accused Siyad Barre of having mishandled the operation. to give a convenient access to the Indian Ocean to neigh- They felt that the conquest of a more limited area in the boring landlocked Ethiopia, and there is a main road Ogaden, including mainly the grazing land of the Isaaq between the two countries.2 clan, could have been a success. The peace agreement with Ethiopia’s Ogaden province across the border is also Ethiopia triggered a full-blown civil war in Somalia, peopled by Somali-speaking herdsmen, and that border is because many clans wanted to hang on to the pan-Somali 155 project, and a coup d’état toppled the military regime in that was widely used in the days of the British Empire and 1991. Chaos ensued, and Somalia collapsed into a stateless that has been used in many developing countries since their entity. Nevertheless, in the nearly two decades since independence. In particular, Boone (2003) shows how this Somaliland seceded from Somalia, it has managed to system was applied quite successfully in several parts of develop the port of Berbera and to enforce an acceptable postcolonial Francophone West Africa, where the central enough level of security on the paved road linking it to government delegated the task of controlling some areas to Ethiopia for the traders to adopt this route. A large share some local traditional authorities, in return for some trans- of Ethiopia’s international trade is now shipped through fers. I argue here that this system cuts through a vexing this port. This achievement stands in sharp contrast to the “bootstrap” problem that faces all new states: a state needs to chaotic remainder of Somalia. have fiscal resources to extend its control to various parts of This chapter presents a framework for analyzing this its territory, but the state needs to have a fairly serious level unexpected success and draws some lessons for under- of control to be able to raise those fiscal resources in the first standing state formation in general. A very simple model place. This problem explains to a large extent why the con- is presented that shows how the political equilibrium that trol many African states exert is in fact extremely limited, emerged in Somaliland is fundamentally rooted in the leaving de facto large parts of their country without any need to provide security to the traders who provide most effective state presence, as emphasized by Herbst (2000). of Berbera’s activity. This is a sine qua non for the traffic I argue that Somaliland was put on the fast track to solve through Berbera to be active and flourishing. This model this problem thanks to two of its preexisting assets. First, sheds some light on the fairly unusual political institutions this country inherited a valuable transport infrastructure, that emerged from less than two decades of self-rule by which only required establishing an efficient political Somaliland. Although some limited fighting occurred for a regime to become competitive in the Horn of Africa. I sug- while in the early 1990s, this former British colony quickly gest that “ports are the taxman’s best friends,” because they engaged in a political process that led to the creation of a provide a “choke point” where taxable resources are con- fairly successful democracy in about a decade. centrated and make revenue collection relatively cheap. Sec- What seems very important when looking at the experi- ond, the traditional institutions of this nomadic pastoral ence of Somaliland is that this gradual buildup of a function- society had not been destroyed by either British colonial rule, ing state started from the grassroots, with very little outside or the subsequent “modernizer” national government of interference. Eubank (2010) emphasizes that the Somaliland post-colonial Somalia, despite the brutal attacks by the Republic has not been recognized internationally, which Mogadishu government in the final years of Siyad Barre’s makes it ineligible for foreign aid. In his view, this is an asset rule in the late 1980s (Lewis 2008). That indiscriminate vio- rather than a liability, because it forced the Somalilanders to lence against civilians and soldiers alike probably helped the develop accountable political institutions and to engage in Somalilanders to achieve a consensus on the project to state formation in a non-Eurocentric fashion. The aim of this secede from Somalia and build a state of their own. The clans chapter is to go one step further in analyzing the type of gov- with which Somali nomadic herdsmen are affiliated are ernment that resulted and in particular to explain clearly why themselves subdivided into kinship groups, which are sub- the “fiscal decentralization” that Eubank is looking for does ject quite informally but firmly to the leadership of the eld- not exist in Somaliland. The model argues instead that the ers. There is no real “chiefdom” among the Somali, unlike in redistribution of fiscal resources from the government to the many other African societies (Lewis 2008), but the elders different regions lies at the heart of these peaceful and demo- exert a significant level of authority. cratic institutions. The key role of redistribution in peaceful This power was successfully harnessed to the emerging state-building in Africa has been emphasized in particular by Somaliland Republic by creating a House of the Elders, Azam (2006). In Somaliland this redistribution of fiscal called Gurti in Somali, in addition to a more standard resources has mainly funded a significant expansion in the elected House of Representatives. This upper house is play- education sector in all regions of the country. Between 1997 ing a part in Somaliland’s bicameral system close to that and 2006 the number of primary schools rose from 165 to played by the House of Lords in Britain’s Westminster sys- 516, while the number of universities increased from 1 to 5 tem, allowing the traditional authorities to be involved (MNPC 2010). directly in running the country’s affairs. I argue that this is The model presented later argues that the Somaliland one of the fundamental pillars of this country, because it is Republic shares many features of the “indirect rule” system the key to establishing—at a low up-front cost—the required 156 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND level of security for making the port of Berbera an economic externality that has the potential to enlarge the opportunity success and hence a reliable source of fiscal revenues. The set of the players, provided a fair compensation is paid to emerging Somalilander government is in fact delegating to the investors. This is why redistribution plays such a key the elders the task of controlling violence and banditry, with role in African state-building, as mentioned above. The rea- a view to protecting the traders who then pay taxes in return son for this result is that the basic negative externality that for the transport services of Somaliland and its port. Then, plagues African states is the threat of civil war or, more gen- the redistribution of fiscal revenues pointed out by Eubank erally, the threat of violence. If I become armed, then your (2010), as mentioned above, is the natural compensation expected welfare goes down, because there is a non-zero for the investment made by the elders in providing the key probability that I will use these weapons to attack you. This public good that makes this lucrative trade possible. Hence, is what the “social contract” aims at preventing, by provid- my approach to Somaliland’s state-building shares some ing a fair and credible compensation for giving up one’s features of the so-called “property rights” approach to the weapons (Azam and Mesnard 2003). In Somaliland the theory of the firm (Hart 1995). The returns to the transport threat was more directed at the economy, because any inse- infrastructure inherited by Somaliland thus depend cru- curity felt by the traders would have brought the port of cially on the “relationship-specific investments” (Hart 1995) Berbera to a halt. Nevertheless, this implicit threat was made by the elders in controlling violence and banditry, overcome through a gradual bottom-up process leading to which is in turn rewarded by some redistribution of fiscal the emergence of the democratic regime. revenues. Similarly, Hart’s theory rests on “incomplete con- tracts theory,” which assumes that only a fraction of the A MODEL OF TRADERS UNDER THREAT observable information can be used as part of an enforce- able contract, while the rest is not verifiable by a court, After the breakup of Somalia, law and order collapsed, and although it is observable by the parties to the contract. My the country became prey to roving bandits and warlords, model pushes this to the extreme, because there is no third making trade highly risky. However, Somaliland itself man- party that can be called upon to enforce any agreement aged to isolate a relative safe haven for traders. The following between the government and the clans’ elders. model aims at bringing out the two levels of political organi- The solution offered here to this fundamental commit- zation that made this possible. I first analyze how the tradi- ment problem brings out the key theoretical contribution tional system of social control was mobilized for reining in made by the current model relative to Alesina and Spo- uncontrolled violence. The next section shows how a higher laore’s theory of the size of nations (Alesina and Spolaore level of political cooperation was needed to create the 2003). These authors define the government as a country’s required level of security for making Berbera a success. monopoly producer of a public good that affects its people differentially. They then raise serious doubts about the pos- The traders sibility of compensating people for these differential bene- fits by transfers, because of the lack of commitment of the A very simple model was chosen for capturing the key part democratic government that they assume. In the model played by security in determining the level of trade going presented here, the government is unable to produce the through Somaliland and Berbera. Let V be the value of the public good alone and must rely on the traditional author- goods transported through the country by the traders. The ities that are in a position to control violence and banditry, traders potentially incur three types of costs while moving provided their participation is satisfied through a transfer. across the country. First, there is a resource cost involved Then, the promise of this transfer can be made credible in in trucking the shipments, including fuel and labor. Sec- a repeated-game framework because the recipients can ond, there is a possibility that bandits might rob the trader punish any deviation by the government by reducing dras- along the way, leaving him without anything to sell at the tically its payoff in case of cheating. This implicit threat is port; the probability that this happens is denoted as p. credible because the recipient of the transfer is incurring a Last, I assume that the lucky traders who have not been positive opportunity cost in delivering its part of the deal. raided have to pay a tax on the goods leaving the country Hence, what looks like a transfer is in fact the price paid for at the fixed rate t. On the export side, I thus assume that a productive service sold to the government. the tax is paid at the port, and on the import side, at the More generally, the state-formation theory sketched end of the trip, that is, mainly at the border for the transit below views the state as a means to internalize some key trade to Ethiopia. We assume that the government controls CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND 157 corruption well enough so that the tax rate is fixed before The potential bandits and the government other agents make their decisions and does not respond to How would the potential bandits behave in a hypothetical the observed trade flow coming through its control points. society where the traditional clan authorities would fail All the parameters of the model are assumed to be common somehow to organize their activity? I assume that there are knowledge, so that all the players can correctly anticipate N bandits who sequentially choose whether or not to raid a the decisions made later on by the other players. The shipment. I make the simplest assumption regarding the traders are assumed to play last, that is, to make their deci- cost of raiding the traders, namely, that a fraction 0 < g < 1 sion to start the trip or not while taking t and p as given. of the shipment is lost in each raiding. This might capture It is realistic to assume that there is no free entry in the the collateral damage of any fighting between the bandits trading business, mainly because of the limited warehouse and the traders, or any other form of cost that the bandits capacity and restricted credit that is typical in poor coun- incur in raiding. This assumption is the simplest, but most tries. In Africa long-distance traders usually belong to some results below are robust to several extensions. Given this long-established family networks (Grégoire and Labazée cost function, a given bandit i ∈{1, . . . ,N} takes the decision 1993). Assuming for the sake of simplicity a quadratic cost xi ∈{0,1} to raid a trader to maximize his expected profit function, then the representative trader chooses Berbera simply defined as: rather than any other port of the Red Sea or the Gulf of Aden area if: Bi = maxxi xip(1–g )vi, (9.5) v2 where p is the probability a raid is successful once under- E = max (1 − τ ) (1 − π ) v − ≥r (9.1) taken and vi is the value of the traffic faced by bandit i, v 2v which takes value 0 if the trader has been successfully raided where r is the trader’s reservation profit, which he could earlier, and v otherwise. – expect to earn while using an alternative trade route and v This expression implies that the bandits incur no cost at is the maximum carrying capacity of this route. From the all if xivi = 0 either because there is no traffic to attack, or first-order condition, the amount of trade going through because they have chosen not to undertake any raiding. the country if the weak inequality in (9.1) holds, is: Otherwise, they loose a constant fraction of the catch when they raid a trader. When the bandits are not organized, they v = v (1 − τ ) (1 − π ) . (9.2) do not take into account the externality that they inflict on Substituting back into (9.1), I find that a positive level of other bandits when performing an attack, namely, that this trade will go through the country if: will in turn reduce the level of traffic that can be raided. 1/2 Thus, any bandit has an incentive to raid a trader, which (1 − τ ) (1 − π ) ≥ ⎛ ⎞ . 2r (9.3) implies that the probability of a successful raid eventually ⎜ ⎟ ⎝v ⎠ converges to 1 as the number of bandits grows large: – Assuming that v > 2r, (9.2) and (9.3) jointly imply that pN = 1–(1–p)N and lim π N = 1. N →∞ when this is profitable, the level of trade will be such that: Thus, the resulting expected value of the trader’s ship- ment converges to zero: ( 2rv ) 1/2 ≤ v ≤ θ. (9.4) lim v N = 0. (9.6) N →∞ Otherwise, it will fall to zero. This simple setting thus captures the idea that the competitiveness of the port of Against this background, I can now analyze the first con- Berbera depends first on two parameters, namely, r and v , – tribution made by the assembly of elders in the Somali clan which determine respectively the profitability of doing busi- society in reducing banditry. In this nomadic society, where ness with the competitors and the physical transport cost the young men are spread all over the land in search of fod- inside the country, and then on two choice variables, p and t, der for their flocks, collective identity is not defined so much which are determined by two key players, respectively, the by reference to a territory as by genealogy. Moreover, in this bandits and the tax authority. In the following, I assume that potentially violent society, where herdsmen could fight over – v is much larger than 2r, in order to capture the geographi- grass or water at any time without any witnesses, escalation cal advantage of Berbera over its competitors. Then, Berbera is prevented by the widespread use of “blood compensa- should win, provided the two key players manage to coordi- tion,” or diya in the Somali language. This is a clever system nate their decisions efficiently. for creating joint liability within the group of origin of the 158 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND perpetrator of a violent crime, because the whole group is authorities to coordinate their action makes a positive responsible for paying the compensation required by the contribution toward efficiency by reducing raiding. This victim’s group for settling the issue. According to Bradbury institution helps the clans’ leaders to internalize the nega- (2008), the diya for the murder of one man is 100 camels tive externalities that they would inflict on one another by (half of this for one woman), which is a very high cost for raiding the traders without control, inducing them to the group. If the diya is not paid, then the aggrieved group reduce their raiding activity. is committed to launch a war against the criminal’s group. Now, absent any political arrangement that could help This provides a strong incentive for those involved to exert the government and the potential bandits to coordinate some control over their fellow clan members, so that any their action, the government would simply maximize its man found guilty of a crime against a member of another expected fiscal revenues, taking into account the traders’ clan would potentially be punished by his own clan, thereby best-response function (9.2) and the bandits’ best choice: avoiding interclan violence as far as possible. Hence, an important service delivered by the clan is to control the vio- ( ) G = max τ 1 − π s v , s.t. (2) holds. t ,x (9.9) lence that could be perpetrated by its own members against both the clan’s members and the members of the other Proposition 1 below describes the resulting Nash equilib- clans. This gives the elders a key role in the Somali society’s rium that prevails in this model when the two players do not control of violence. This power was gradually aggregated in coordinate their actions through some political arrangement. Somaliland during the political buildup toward democracy by organizing first a large series of local meetings of elders, PROPOSITION 1: The uncoordinated Nash equilibrium choice of which then developed in a kind of pyramidal fashion, cul- p and t by the government and the syndicated bandits, minating with the creation of the national assembly of the respectively, is: elders at the Gurti. This process is well documented by t N = p N = 1/2, (9.10) Bradbury (2008). The most striking point about this process is that it was mainly organized and funded by the diaspora entailing a level of traffic: of Somalilanders who had fled the repression under the v N = v 4, (9.11) Siyad Barre regime. The diaspora played a key role in many other parts of Somaliland’s political development and the and the following payoffs for the bandits and the government, emergence of democracy there. respectively, are: What can such a consolidation process deliver within my G N = v 16 and B N = (1 − γ ) v 8 . (9.12) model society of traders and bandits? Quite obviously, such a consolidation process would end up creating a kind of The next section shows how a more inclusive political syndicated banditry, which could internalize the negative arrangement can harness this social control mechanism externality involved because an increase in the raiding activ- provided by the traditional clan authorities to improve effi- ity against traders would reduce the size of the trade flow ciency still further and reduce raiding to zero. itself. Now, instead of problem (9.5), this coalition would choose the number of raids or, equivalently given equation REDISTRIBUTION IN THE EFFICIENT (9.6), the probability of a successful raid so to maximize: POLITICAL EQUILIBRIUM B S = max π (1 − γ ) v , (9.7) In a clan society, genealogy is the essence of social identity, π as mentioned above. It is then natural to assume that the such that (9.2) holds. elders have a strong interest in the continuation of the clan It can be readily checked that this “syndicated banditry” and thus care for the welfare of the next generation. This equilibrium yields the following levels of raiding and traffic: can be captured by using a dynastic family model à la Barro π S = 1 2 and v S = v (1 − τ ) 2 . (9.8) (1974). In this kind of model, each generation is affected by intergenerational altruism, such that the next genera- This implies that the mere fact of forming a coalition of tion’s welfare is an argument in the current generation’s util- clans is not sufficient in general to explain why raiding ity function. Choi and Bowles (2007) have coined the would stop. Nevertheless, the model predicts that the cre- expression “parochial altruism” to describe such an inter- ation of an institution that helps the traditional clan generational externality and have shown how these links CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND 159 across generations are an important asset for the survival of knowing that it will redistribute part of the resulting revenue, human groups in a violent society within an evolutionary as explained below, to compensate the potential bandits for framework. Hence, the dynastic family assumption seems their restraint. Moreover, the chosen tax rate is the same as in especially appropriate for describing the behavior of a tradi- the Nash equilibrium of the previous section, at the level that tional clan society like the Somali one. In this case, it is nat- maximizes the Laffer curve t (1–t ). Then, the traffic level is ural to assume that the players have an infinite horizon, twice as large in this equilibrium as in the uncoordinated because their concern for the next generation creates a chain Nash equilibrium of the previous section. of intergenerational links up to infinity. I thus discuss the We can now prove the following: political setting in which banditry and warlordism can be eradicated by embedding the simple model of the previous PROPOSITION 2: There exists at least one efficient political equi- section within an infinite-horizon repeated game frame- librium with p ∗ = 0, t ∗ = 1/2 and: work. I then show how Somaliland’s political institutions 4 − 3δ 3v δ cater for the key mechanisms brought out by this model. (1 − γ )v ⎛ ⎜ ⎞ ⎟≤ g∗ ≤ , (9.15) To capture their common ethnic heritage in Somaliland, ⎝ 8 ⎠ 16 assume that the potential bandits and the government have the same discount factor 0 < d < 1. Then, assume that the if: government can offer at each period the following contract (1 − γ ) 8 to the potential syndicated bandits: “I will give you g > 0 if δ ≥δ = . (9.16) 9 − 6γ you refrain from raiding the traders and enforce p = 0”. This contract clearly entails that the potential syndicated bandits The proof is rejected in the appendix, while figure 9.1 play first and the government second, after having observed helps the reader understand proposition 2 intuitively. The whether the raiding was avoided or not. Moreover, no third downward sloping line represents the left-hand part of con- party is available to enforce the promise made by the gov- dition (9.15). All the points located above this line are ernment to deliver the transfer once the potential bandits acceptable for inducing the potential syndicated bandits to have refrained from raiding the traders. I define an efficient cooperate. The required transfer is lower, the more patient political equilibrium as an efficient outcome that can be the potential bandits are. The upward sloping line repre- sustained ad infinitum in the repeated game between the sents the right-hand part of this condition; all the points government and the potential bandits. located below it are acceptable for inducing the government to cooperate. The government is willing to pay more, the DEFINITION 1: An efficient political equilibrium is a triplet more patient it is. Then, the figure makes it clear that the tri- {t, p, g} that lies on the Pareto frontier in the game angle on the right labeled “acceptable triangle” is the set of between the government and the potential syndicated bandits and that can be sustained by a standard trigger- Figure 9.1 The Efficient Political Equilibrium Set strategy equilibrium. In this simple setting involving a transfer, the Pareto- g efficient {t, p} pair of actions by the two players can be derived by solving the following problem: (1 –g )n max ⎡τ (1 − π ) + π (1 − γ ) ⎤ v , s.t. (2) holds. ⎣ ⎦ (9.13) 2 t ,π Using standard maximization techniques (Kuhn and Tucker theorem), one finds easily that the Pareto-efficient 3n outcome is a corner solution that implies the following: 16 Acceptable π ∗ = 0 , τ ∗ = 1 2 and v ∗ = v 2 . (9.14) triangle (1 –g )q The intuition for this result is pretty straightforward. 8 Because of the unit cost g > 0, raiding is an inefficient way of d d 1 collecting revenues from the traders. The government thus will perform all the tax collection in the efficient equilibrium, Source: Author’s calculations. 160 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND all the points that are acceptable to induce cooperation by leaning in the opposite direction, suggesting that they were both players. Notice that d , whose value is given at (9.16), – just regarding the “New Puntland State” as a mere building is strictly lower than 1. This means that the set of acceptable block to reconstruct Somalia. Second, they did not build any {d, g∗} is not empty, ensuring existence of at least one pos- institutional representation of the elders, because they cre- sible efficient political equilibrium point. Moreover, d is a – ated a unicameral parliament, which did not have many decreasing function of g, so that cooperation is easier to resources in any event. In terms of the model described achieve, the higher is the unit cost of raiding. Hence, (9.16) above, this lack of resources might be blamed on the lack of defines a credibility frontier in the {d, g } space, which is significant transport infrastructures and thus on the lower used in the next section. The intuition for this result is need for internal security. This is illustrated by the fact that that the stronger the comparative advantage of the govern- piracy is quite active off the coast of Puntland, imposing ment at collecting revenues, the easier it is for it to buy off some negative externality on all the ports of the Red Sea credibly the potential bandits with a transfer for producing and the Gulf of Aden. However, the failure to give the eld- security for the traders. The next section exploits this intu- ers a large enough role in the New State of Puntland is also ition to diagnose why a solution similar to Somaliland’s probably due to the towering figure of Colonel Abdillahi did not emerge elsewhere in Somalia. Yusuf, who assumed first the presidency in Puntland, up to 2001, and then won the federal presidency in Mogadishu, which is in fact an empty shell. Abdillahi obviously did WHY NOT THE REST OF SOMALIA? not draw the lessons of the failure of Siyad Barre’s military The foregoing modeling exercise begs the question of why regime to create top-down a viable state in Somalia, because such an efficient political equilibrium did not emerge in the he tried also to impose his authoritarian rule on the New rest of Somalia. One easy answer is given by Eubank (2010), State of Puntland. However, imposing the “rule of fear” also who claims that foreign aid played a detrimental role there, requires resources, which are dramatically lacking in by relaxing the need to build accountable institutions in that Puntland, while they are somewhat higher in Mogadishu, part of the country. Nevertheless, it is worth going into if only because of foreign aid as mentioned above. more detail to look at the missed opportunities in eastern Southern Somalia once had the resources to support and southern Somalia in order to bring out the kind of diag- the strong authoritarian government led by Siyad Barre, nosis that the model above is pointing out. It emphasizes who for a while was aligned with the Soviet Union, which some more structural characteristics. was playing a complicated game in the Horn of Africa. The first point to notice is that without declaring seces- Then alliances switched, and the United States became sion formally, the northeastern part of Somalia also built a involved. Although southern Somalia inherited a valuable bottom-up institutional solution known as the “New infrastructural asset, in addition to the sovereignty rent due Puntland State of Somalia,” which was founded at a con- to international recognition and foreign aid, it was facing a ference in Garowe in 1998. This promising solution started more complicated political problem than Somaliland. among some Darod clans, in particular the Mijerteyn While the latter is very homogenous, with most of its pop- (Lewis 2008), but it was missing two of the key ingredients ulation involved in nomadic pastoral activity, the former of Somaliland’s success. First, that part of the country did has a sizable agricultural area, between the Shabelle and not inherit an infrastructure asset of the same caliber as Jubba rivers. The Somali clans living there have a distinct the port of Berbera. In fact, the Puntland ruling elite never sedentary culture, in which territory matters at least as lost sight of the nearby formal capital city of Mogadishu in much as genealogy, so the elders have a weakened role in the south. That is where the infrastructural assets are nat- social control. Moreover, there are some Bantu farmers in urally located, despite the massive destruction brought the midst of this Somali population, loosening further the about by the war. Mogadishu has two ports, one new and ethnic ties in that part of Somalia. Hence, the traditional the other old, and an international airport. Moreover a authorities are too weak in that part of Somalia to deliver road to Addis Ababa, called the “Strada Imperiale,” could the kind of social-control services available in Somaliland, be restored. Hence, the Puntlanders never severed their on the one hand, while the government was too authori- links to the rest of Somalia, realizing probably that they tarian to make credible promises of redistribution, for lack would never be in a position to levy the fiscal resources of checks and balances, on the other hand. required to cement a Puntlander social contract similar to Figure 9.2 summarizes the foregoing discussion within the one prevailing in Somaliland. Their strategy was clearly the analytical framework presented earlier. It represents CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND 161 Figure 9.2 Why Not the Rest of Somalia redistribution mechanisms for compensating the latter for giving up their highly lucrative violent activities. d 1 CONCLUSION 8/9 Somaliland This case study of the emerging Somaliland Republic pro- vides a natural experiment that sheds some useful light on the theory of state-building. It shows that the Hobbesian Other Leviathan is not the only path available for controlling vio- Somalia lence and building up a peaceful state. It suggests that a Lockean “horizontal social contract” model may be a viable g solution in some circumstances, for “breaking up” a state of 1 anarchy, using the expression coined by Hirshleifer (1995). Source: Author’s depiction. In Somaliland, one observes a separation of the power to control violence, which belongs to the clans’ elders, on the condition (9.16) as the downward sloping frontier repre- one hand, and the power to tax and to produce some of the sented in the {d, g } space. This is the credibility frontier public goods that a modern state is expected to provide, on defined above. The parameter g measures the cost advantage the other hand. Among other things, this study thus enjoyed by the Somalilander government over the potential shows the benefit that political economists can gain from bandits in collecting revenue, thanks to the preserved stock using the work of the social anthropologists for under- of infrastructure inherited from colonial days, and enhanced standing the political economy of developing countries. In over the recent past by some investment. In the case of return, the modeling exercise demonstrates the key part Puntland, no such stock is available, as explained above, played by an inherited infrastructural asset, namely the suggesting that g is small there, because the government is port of Berbera and the road that links it to Ethiopia. not in a position to collect revenue much more efficiently This model sheds some light on the political institutions than the potential bandits. In Southern Somalia, g is poten- that have been put in place in Somaliland. The key problem tially much larger, because of the infrastructure available in to be solved was for the business-oriented elite to delegate to Mogadishu, although the latter needs massive investments to the traditional authorities the task of controlling violence recover its potential efficiency. However, the mixed ethnic and banditry effectively so that Berbera became an attrac- composition of the population living between the Jubba and tive outlet for the traders from Ethiopia as well as for the Shabelle rivers, as well as the sedentary culture of the exporting the output of the livestock sector. The first step Somali agriculturists living there, suggest that d is much was to help the clans’ elders to cooperate by organizing sev- lower there than in Somaliland. Hence, this analysis empha- eral local conferences. But the model shows that this is not sizes that Somaliland exploited the two-dimensional edge enough to provide the incentives for reducing banditry to that it had over the rest of Somalia for creating its efficient zero. The second step was aimed at making credible the political equilibrium. This is captured in figure 9.2 by noting promise of redistributing the enhanced fiscal resources that Somaliland’s efficient political equilibrium prevails for a resulting from the increased trade flow to the clans. A {d, g } pair lying above the frontier, while the other parts of bicameral system was put in place to ensure that the elders Somalia are found below that frontier. had a key role to play in the law-making process, by giving This finding also suggests that simply trying to export them a direct access to the required information, as well as the solution that worked in Somaliland to the rest of Soma- some veto power in the implicit bargaining problem. As lia would probably not bring about the same benefits. Some noted, the redistribution of the fiscal resources was focused additional imagination is surely needed to devise an appro- on the development of the education sector in all the priate system for the rest of Somalia with a view to create a regions, as one would expect in a country where genealogy lasting peace there. Nevertheless, the analysis presented is the key social identifier that determines each person’s here points to the two dimensions where economic and affiliation to a clan. The dynastic family model thus seems political innovation is required: enhancing the relative effi- to apply perfectly in this case and explains why people felt ciency of the government at collecting revenue, relative to compensated for their efforts by seeing their children going warlords and bandits; and making credible the necessary to school and to university. 162 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND This model may thus be viewed as an extension of Hart’s for decades. The qualification brought out by this case study “Property Rights Approach” to the theory of the firm (Hart is that a correct political setting is required, aimed at mak- 1995). The clan authorities can invest in providing security, ing the redistribution of the benefits from cooperation an asset that enters the production function for transport among the different actors credible. This redistribution is services as a complement to infrastructure. This insight is the compensation due to the potential bandits for refraining fundamental for understanding why a bicameral demo- from raiding the traders and thus participating in the effi- cratic institution lies at the heart of Somaliland’s political cient political equilibrium. This suggests that the inability to institutions, for providing a balanced representation of both set up a correct political system, partly because of external the traditional authorities and the business-oriented mod- interference, is what makes the infrastructural assets of ern actors. Hence, Somaliland’s experience provides a fruit- southern Somalia, like Mogadishu’s old and new ports, its ful line of arguments in favor of a qualified support for the airport, and the “Strada Imperiale” road that links them to traditional “project aid” doctrine, with its emphasis on Addis Ababa, as well as the rent to sovereignty provided by funding infrastructural projects, which has inspired the access to foreign aid that comes with international recogni- action of the World Bank and other development agencies tion, largely useless for producing peace. CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND 163 APPENDIX PROOF OF PROPOSITION 2 2. Markakis (1992) provides a regional perspective on the ports in the Horn of Africa. Assume that both players adopt the standard trigger strategy (see, for example, Gibbons 1992). If they choose to cooper- ate and refrain from raiding the traders with a view to REFERENCES receiving the transfer g from the government, the potential Alesina, Alberto, and Enrico Spolaore. 2003. The Size of syndicated bandits receive the following present value: Nations. Cambridge, MA: MIT Press. g VC = . (A9.1) Azam, Jean-Paul. 2006. “The Paradox of Power Reconsid- 1−δ ered: A Theory of Political Regimes in Africa.” Journal of If they choose instead to deviate and attack the traders, African Economies 15 (1): 26–58. the syndicated bandits will be punished first within the Azam, Jean-Paul, and Alice Mesnard. 2003. “Civil War and same period, as the government withholds the transfer g, the Social Contract.” Public Choice 115 (3): 455–75. and the static Nash equilibrium outcome will then prevail Barro, R. J. 1994. “Are Government Bonds Net Wealth?” ever after. Assuming that this deviation is not expected by Journal of Political Economy 81: 1095–1117. the traders, this yields the following present value: Boone, Catherine. 2003. Political Topographies of the African (1 − γ ) v δ (1 − γ ) v State: Territorial Authority and Institutional Choice. Cam- VD = + . (A9.2) bridge, U.K.: Cambridge University Press. 2 8 (1 − δ ) Bradbury, Mark. 2008. Becoming Somaliland. London: The potential syndicated bandits will thus choose to Progressio. refrain from raiding the traders if VC ≥ VD, that is, if: Choi, Jung-Kyoo, and Samuel Bowles. Samuel. 2007. “The ⎛ 4 − 3δ ⎞ Coevolution of Parochial Altruism and War.” Science 318 g ≥ (1 − γ ) v ⎜ ⎟. (A9.3) no. 5850: 636–40. ⎝ 8 ⎠ Doornbos, Martin. 1993. “Pasture and Polis: The Roots of On its part, if the government chooses to cooperate, that Political Marginalization of Somali Pastoralism.” In is, to deliver the agreed amount g when observing p = 0, it Conflict and the Decline of Pastoralism in the Horn of receives the following present value: Africa, ed. John Markakis, 100–121. Basingstoke, U.K.: v − 4g Macmillan. WC = . (A9.4) 4 (1 − δ ) Eubank, Nicholas. 2010. “Peace-Building without Exter- nal Assistance: Lessons from Somaliland.” Working If it chooses instead to deviate and to withdraw its trans- Paper 198, Center for Global Development, Washing- fer despite the potential bandits’ compliance with the ton, DC. promised p = 0, then the government will first keep the Gibbons, Robert. 1992. Game Theory for Applied Economists. whole fiscal revenue in the current period, and then get its Princeton NJ: Princeton University Press. Nash-equilibrium payoff GN from then on. This yields the Grégoire, Emmanuel, and Pascal Labazée, eds. 1993. Grands following present value: Commerçants d’Afrique de l’Ouest. Paris: Karthala-Orstom. v δv Hart, Oliver. 1995. Firms, Contracts, and Financial Structure. WD = + . (A9.5) 4 16 (1 − δ ) Oxford, U.K.: Oxford University Press. Herbst, Jeffrey. 2000. States and Power in Africa. Compara- Then, the government will choose to cooperate if tive Lessons in Authority and Control. Princeton, NJ: WC ≥ WD, that is, if: Princeton University Press. 3θδ g≤ . (A9.6) Hirshleifer, Jack. 1995. “Anarchy and its Breakdown.” Journal 16 of Political Economy 103 (1): 26–52. Then (A9.3) and (A9.6) together make (9.16). Last, d is _ Lewis, Ioan M. 2008. Understanding Somalia and Somaliland computed as the value of d such that the range of values of New York: Columbia University Press. g∗ defined at (9.15) is just empty, as illustrated in figure 9.1. Markakis, John. 1992. “The Regional Significance of Sea Ports in the Horn of Africa.” In Beyond Conflict in the Horn, eds. Martin Doornbos, Abdel Gaffar M. Ahmed, NOTES and John Markakis, 130–31. London: James Currey. 1. “Bolloré proposes Berbera-Addis Ababa Transport Cor- MNPC (Ministry of National Planning and Coordination). ridor.” WorldCargo News online. December 10, 2009. 2010. Somaliland in Figures 2009. Hargeysa. 164 CHAPTER 9: TRANSPORT INFRASTRUCTURE AND THE ROAD TO STATEHOOD IN SOMALILAND PA R T I I I Leveraging Sectoral Advantages to Expand Exports CHAPTER 10 Growing Mali’s Mango Exports: Linking Farmers to Market through Innovations in the Value Chain Yéyandé Sangho, Patrick Labaste, and Christophe Ravry lthough less than 4 percent of land in Mali, a A major leaps toward developing a competitive horticulture landlocked country in West Africa, is arable, agri- export sector. culture accounts for 45 percent of the country’s economy and employs 80 percent of its workforce. Indus- MALI’S EXPORT DIVERSIFICATION STRATEGY try represents 17 percent of the country’s gross domestic product (GDP), with food processing, construction, and A key objective of Mali’s poverty reduction strategy over the phosphate and gold mining as the principal industrial past two decades has been—and still is—to increase rural activities. Mali’s main agricultural export is cotton, followed incomes and employment opportunities by promoting agri- by livestock. The fact that Mali is landlocked, however, has cultural diversification and developing exports of high- always made it dependent on the transport infrastructure value commodities. In the early 1990s the government of and other logistical arrangements of its neighbors for trade Mali recognized a need to design policies to diversify and exports. exports and foreign exchange earnings, which had been Thanks to excellent geographical and weather conditions heavily dependent on only three export products: gold, cot- prevailing in the southern part of the country, mangoes ton, and livestock. These products, however, are quite sus- have always been abundant in Mali, particularly in the ceptible to fluctuations. For example, cotton exports Bamako and Sikasso regions. The fruit was traditionally sold dropped dramatically in 2008 to less than half of their pre- in the domestic market. In the 1970s Mali was the first vious level; also, because of several years of financial crisis, country in West Africa to explore opportunities to export the contribution of the sector to fiscal revenues has been fresh mangoes. These exports were shipped exclusively by negative throughout the whole decade. air freight to a niche market—high-end retail shops in France selling tropical fruits—and reached a volume of Striving for diversification between 1,000 and 1,500 metric tons a year Starting in the early 1990s Mali undertook several In the face of these fluctuations, the government of Mali transformations in its mango subsector that have allowed began, in the 1990s, to focus on high-value and nontradi- the country to overcome logistical dependencies and con- tional agricultural products as a means to generate income straints, expand exports of fresh mangoes, and make and achieve greater diversification of exports, based on the Yéyandé Sangho is Senior Operations Officer at the World Bank. Patrick Labaste is Lead Agriculture Economist at the World Bank. Christophe Ravry is Senior Agribusiness Specialist at the World Bank. This paper builds on an early draft and research work done by Malick Antoine at the World Bank in 2006. 167 country’s comparative advantages. Besides offering small- In 1996 the government of Mali used $6 million in World holder farmers the opportunity to diversify the source of Bank financing for an agricultural trading and processing their livelihoods, high-value and nontraditional agricultural pilot project, PAVCOPA (Project d’appui à la valorisation et products also offer countries the opportunity to diversify à la commercialisation des produits agricoles). The project away from low-value bulk commodities. sought to promote agribusiness and exports. Specifically, it Several horticulture crops were considered as possible aimed to “improve the enabling environment needed to targets for these diversification efforts, including cashews, enhance private sector business opportunities and encour- tomatoes, shallots, and mangoes. Mangoes were a prime age [state] disengagement from commercial activities, candidate both because of the excellent agroclimatic condi- improve technical support to producers through effective tions for growing them in the southern regions of Bougouni research and extension aimed at enhancing and diversifying and Sikasso and because of the fast-growing demand for production and improving the international competitive- mangoes in European markets (figure 10.1). Furthermore, ness of Malian exports, and boost private investment in because mangoes were already being produced by small- agricultural trading and processing” (World Bank 1996). holder farmers throughout the country, the subsector had In addition, the PAVCOPA project aimed at supporting the potential to contribute to rural livelihood improve- producers, processors, and traders in the Sikasso, Segou, ments. However, despite the high quality of Mali’s fresh fruit and Koulikoro regions and the Bamako district, providing and vegetables, the high cost of air freight was severely lim- technical and promotional assistance (including price iting marketing and exportation. In fact, significant volumes information) and organizing commercial forums, sup- of Mali’s mangoes were purchased and processed for export porting professional associations, and carrying out studies by operators based in Côte d’Ivoire, thus leaving little on the markets for high-value crops. potential for value addition in Mali. The Agricultural Value Chain Promotion Agency (APROFA, by its French acronym), established by the gov- ernment in 1993, was designated as the project executing Early efforts toward agricultural diversification agency for PAVCOPA. APROFA’s goal was to create sustain- In 1992 Mali’s Ministry of Agriculture prepared a national able growth in the agrifood sector by increasing exports to rural development strategy, the Schéma Directeur du the European and African regional markets and establishing Développement Rural (SDDR), emphasizing commercial import substitution activities. Further, the agency was to agriculture, export promotion, and value addition, and support the improvement of the technical, managerial, the government began directing resources toward those organizational, and professional capacities of public and ends using financing from international donors such as private actors engaged in agribusiness. the World Bank and the U.S. Agency for International Development (USAID). The global market for fresh mangoes Figure 10.1 Imports of Mangoes into the European Demand for mangoes in the European Union (EU) has Union, 2004–08 grown significantly in recent years, increasing by approxi- 250,000 mately 55 percent between 2001 and 2008, from 136,000 tons to more than 230,000 tons respectively (table 10.1). The leading exporter countries, in order, are Mexico, Brazil, 200,000 Peru, India, Pakistan, and the Philippines; together, they represent around 75 percent of the market. Although Metric tons 150,000 African mango exports grew by 69 percent during the same period, export volumes remain far lower than those of the 100,000 leading producers. In 2007, for example, Mexico was exporting a total of 236,000 metric tons of mangoes, while 50,000 the entire African continent recorded 46,300 metrics tons of exports only (FAOSTAT). 0 Within Africa, leading producers, in order, are South 2004 2005 2006 2007 2008 Africa, Côte d’Ivoire, Sudan, Kenya, the Arab Republic of Source: Eurostat. Egypt, and Mali. Historically, Côte d’Ivoire has exported 168 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN Table 10.1 Mango Imports to the European Union, by Country, 2004–09 (metric tons) Growth 2004–08 (percent) Country 2004 2005 2006 2007 2008 2009a Annual Total Brazil 69,319 82,293 84,858 82,993 96,870 69,590 9 40 Peru 19,817 26,394 41,027 36,854 50,756 25,062 27 156 Pakistan 10,938 12,306 10,120 13,224 12,941 12,913 4 18 Israel 8,059 12,548 11,181 14,808 12,261 12,606 11 52 Côte d’Ivoire 11,426 9,856 14,428 14,706 11,249 11,659 0 –2 United States 7,612 6,894 5,971 7,404 7,516 5,536 0 –1 Senegal 2,810 3,011 6,194 4,702 6,034 6,219 21 115 Costa Rica 3,983 6,271 7,545 4,664 5,360 5,685 8 35 Mali 2,096 2,560 3,477 4,317 4,902 3,480 24 134 Dominican Rep. 1,228 1,591 1,618 2,767 4,307 4,179 37 251 India 915 1,720 2,472 2,425 2,577 2,470 30 182 Burkina Faso 928 1,164 2,152 3,191 2,406 1,957 27 159 Other 23,516 20,035 19,786 19,003 13,209 12,893 –13 –44 Total 162,646 186,643 210,829 211,057 230,388 174,248 9 42 Source: Eurostat. a. January to November. significantly higher volumes of mangoes than its closest transport issues prevented an initial substantive scaling up competitors on the continent, apart from South Africa. In of exports. the early 2000s, before political disturbances in the country, Côte d’Ivoire exported around 11,000 metric tons; the next Overcoming transport and logistics constraints closest African rival was Burkina Faso with 3,500 metric tons exported. That same year Mali reported 900 metric High transportation costs made Malian mangoes uncom- tons of exports (FAOSTAT). petitive in the global market. Whereas competitors in Latin America could take advantage of more economic sea freight, Mali’s producers were limited to the more expensive air THE CHALLENGES freight option, which reduced their exports and relative Identifying market opportunities position as a mango exporter in West Africa (table 10.2). Historically, Mali relied on Côte d’Ivoire as a link to port The large amount of mango exports from Côte d’Ivoire can facilities. Cotton lint and other products, for example, were be partly explained by the fact that the country was working exported through the port of Abidjan. In addition to relying with buying agents from Mali in the 1980s and 1990s. on its neighbor’s ports, Mali also relied on its neighbor’s Mango exporters in Côte d’Ivoire operated pack houses in infrastructure to move its export products to the ports. the northern part of the country, sourcing mangoes from Until the 1990s the only rail line in Mali with international growers across the border in Mali. The arrangement led to a links was run inefficiently, leading to significant high prices surge of mango exports, sold under an Ivorian label. and delays. This protracted transport process was especially Malian mango producers did not anticipate the increas- detrimental to those Malian agricultural products that have ing demand for mangoes in the European market, which high spoilage and shrinkage rates. A research report was moving from a luxury, niche market to a volume (CARANA Corporation 2004) also points out that produc- market. Mangoes, like some other tropical fruits, such as ers’ logistical costs were negatively affected by the cost of the bananas and pineapples some decades earlier, became a consolidation of goods and poor business practices. fruit that was in high demand all year round by European customers. Other challenges In an attempt to increase its mango exports, Mali sought alternatives to air freight that would increase its competi- Although infrastructure was the most serious constraint to tiveness and market share. Mali’s landlocked status posed a Malian exports of fresh produce, inadequate access to serious challenge to that effort, however, and logistics and finance and land; a poor business climate; inadequate CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 169 addition, limited foreign direct investment in the country and Table 10.2 Mango Exports to the European Union, 1970–95 poor enforcement of fair business practices further reduced (metric tons) competitiveness. Mali Burkina Faso Côte d’Ivoire Poor Land Titling and a Nonexistent Land Market. 1970 35 122 — 1980 1,172 2,116 281 Bureaucratic inefficiencies and land tenure rules kept 1990 1,300 2,700 1,000 Malian producers from owning large tracts of land, thereby 1995 850 714 7,107 creating another barrier to large-scale horticulture. Until Source: FAOSTAT. the Agriculture Orientation Law of 2006, there were no pro- Note: — = not available. visions for the establishment of commercial-scale irrigated plots, nor could producers purchase irrigated land that management, harvesting, and handling techniques; and could be used as collateral against which to secure loans. little investment in mass production further weakened producer export capacity. Inadequate Working Capital and Lack of Supply Chain Financing. With limited working capital, Malian Lack of Market Information and Organization. production and subsequent exports were constrained because Malian producers and exporters did not have a good grasp exporters were required to finance the cost of shipping. This of the requirements of the intensely competitive European problem was of particular urgency to exporters, who did market. In addition, poor organization and coordination not always receive payments from previous shipments before between producers, government inspection officials, and they needed to send additional shipments. exporters led to inefficiencies affecting the industry’s ability In short, not only did Mali need to make changes that would to meet international orders on time. enable it to improve its main constraint—infrastructure—it also needed to undertake a series of transformations improv- Lack of Proper Harvesting Practices and Posthar- ing technical capacity and business regulatory practices in vest Handling Techniques. Buying agents typically vis- the mango sector (box 10.1). ited orchards once a season and harvested all of the fruit at once, regardless of its ripeness. In addition, many growers INTERVENTIONS did not manage their orchards, did not prune undergrowth, and did not clean the ground under the trees, creating a Supporting agricultural diversification good habitat for fruit flies. The transformation of Mali’s mango export sector began in the mid-1990s with PAVCOPA, the pilot project funded by Little Investment at the Production Level. Invest- ment in mango production—including the establishment of Box 10.1 The Mango Export Value Chain before commercial orchards—is important because it provides Reforms producers and exporters a degree of organization and effi- ciency. Traditionally, Malian farmers saw mango production Before reforms in the sector, the value chain included as mainly a subsistence activity, a business opportunity. producers, intermediaries, and exporters. Even though orchards were smaller than five hectares, as much as 50 percent of total mango production was wasted ■ Producers sold produce to buying agents, called each year (Club du Sahel/OECD 1998), and little investment “pisteurs,” independent traders who select, harvest, was made in increasing production. Additionally, because wash, sort, package, and transport the fruit from Malian mangoes were exported through Ivoirian pack orchard to pack house. houses, Malian producers received little return from their ■ Agents worked exclusively for a particular exporter. crop and thus had even fewer incentives to invest in com- ■ Quality control was undertaken by exporters, who mercializing crop production. selected fruit that were of export quality and returned the second-grade fruit to the pisteur for sale on the domestic market. An Uninviting Investment Climate. Foreign investors, unfamiliar with Mali’s policies and procedures for conducting Source: Authors. business, perceived investment in the country as high risk. In 170 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN the government of Mali and the World Bank. Initially, overcome poor access to international ports while develop- Malian buying agents used existing facilities at the Bamako ing alternate and effective transport routes. airport, to export about 1,500 metric tons of mangoes to the European market each year. In the early 1990s, however, as demand in Europe increased, competition intensified A key innovation: Introduction of multimodal among producers as South American exporters, notably transport in the export value chain Brazil, began exporting sea-freighted mangoes to Europe on The design and implementation of a multimodal trans- a large scale. Even though air freighting allows mangoes to portation system in 1995 was a key innovation for the Malian be picked at a more advanced stage of maturity (resulting in mango industry that eased transportation logistics con- sweeter fruit), it also comes at a significantly higher cost. straints and costs. Through this new system, mangoes are In addition to the government and World Bank initia- loaded into temperature-controlled containers and driven tives, USAID contributed to jump-starting initial work in into Ferkessedougou, Côte d’Ivoire. There, the containers are Mali’s mango subsector through support to the Office de la transported by rail to Abidjan and shipped to Europe. The Haute Vallée du Niger (OHVN) project, that also focused on containers are maintained at 5 degrees Celsius. Transit times high-value agriculture and food products (HVAFs). This per shipment between the mango-producing regions of program served as a pilot project while the preparation of Sikasso and northern Europe have declined from 25–30 days PAVCOPA was ongoing. Aid organizations continued to to approximately 12–15 days (Danielou, Labaste, and Voisard be involved in Mali’s agriculture sector for over 15 years 2003). As figure 10.2 illustrates, multimodal transit also following the initial project in 1993. allowed large volumes of mangoes to be exported from Mali at an affordable cost. The existence of this uninterrupted cold chain (refrigerated containers) greatly decreases the rate Rethinking and redesigning the of spoiling of mangoes during transit. mango export value chain The first step that triggered the agricultural export reform Designing the innovation and testing the system process in Mali occurred when a small but critical mass of stakeholders were brought together to thoroughly analyze To hedge risks involved with the new system, APROFA nego- and assess opportunities for diversification into higher- tiated and signed a partnership agreement with a privately value crops. Initial implementation of PAVCOPA was not owned Ivorian company, SN Tropical Expressions (SNTE), unsuccessful; however, the preparation of its midterm that split the costs, profit margins, and risks associated with review in January 2000 provided an opportunity to restruc- the first shipment of 200 tons. SNTE was responsible for ture the project and create a new business plan. both the logistics and packaging involved in shipping the Another critical point was reached when APROFA and produce to their ultimate port of destination in Europe. A its technical team of advisors began analyzing what had profit margin was built into the fixed price to ensure that made Côte d’Ivoire successful in developing sea-freighted both parties respect their mutual costs objectives. If the aver- mango exports. Although Mali had been exporting mangoes age selling price is greater than the fixed price, APROFA and for years, the industry was not broadly based but rather lim- SNTE share the difference. This partnership resolved the ited to a narrow market targeting an exclusive club of problem of lack of finance for exporters because SNTE had exporters on the periphery of Bamako. the necessary working capital to fund the operation. Observing and understanding the success of Côte Under the same agreement, APROFA was able to obtain d’Ivoire entailed a thorough analysis of the global market a guarantee from a local bank enabling SNTE to lease an demand and its trends, and an assessment of the supply existing pack house in Sikasso with the necessary equip- chain in Côte d’Ivoire. This evaluation suggested that three ment for precooling fruit pallets. In addition, SNTE sec- critical steps were necessary to change the mango industry onded two experienced managers to operate the Sikasso in Mali. The industry had to penetrate and compete in the facility. Precooling the mangoes at the pack house stage is growing market for fresh mangoes (that is, the sea-freighted critical for the continuous cold chain. The establishment of trade to major European ports and hubs such as Rotterdam, continuous cold chains in the export of horticulture and Antwerp, Algesiras, and London); establish the drivers of other high-value crops remains one the most serious chal- change and a value chain within the system to improve lenges to the expansion of exports from other Sub-Saharan competitiveness; and identify and implement solutions to African countries. CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 171 Figure 10.2 Supply Chains for Fresh Mango Exports from Mali a. Air b. Truck/sea Air Truck Farm Smallholder mango producers Farm Smallholder mango producers M M a a Small traders Small traders l Traders l Traders (pisteurs) (pisteurs) i i Pack houses / Packers/exporters Pack houses / Packers/exporters exporters exporters D E C i u ô v r o Exporters t i o e r Importers Exporters p e e E 2 days u High cost, r Importers 25 days low volume o p Lowest cost, e Wholesalers Retailers low volume c. Multimodal Multi-Modal Smallholder mango producers Farm M Small traders a Traders (pisteurs) l i Pack houses / Packers/exporters exporters D Ferke C i train station ô v t o Exporters i e r e Exporters E u r Importers 12 days o Low cost, p high volume Wholesalers Retailers e Source: Geomar International. Early success quality: none of the 63,000 cartons of mangoes shipped to the Netherlands was rejected. To the contrary, there were The test carried out by APROFA and SNTE was a great demands for steady shipments of Malian mangoes. The success from many points of view, particularly in terms of 172 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN validation of the feasibility and profitability of the new management, monitoring, and evaluation of project value chain for mango exports represented a critical break- implementation. through and opened completely new horizons for the Using a series of analytical tools and data, the project mango business in Mali. team identified value chains for export markets, providing Despite sociopolitical turmoil in Côte d’Ivoire, Mali’s a basis for prioritized interventions that would ultimately cold chain transportation system remained uninterrupted. create value and improve livelihoods of Malian small farm- The pilot project helped Malian producers and exporters ers. The analysis used five modules to assess the competi- address the three major constraints of Mali’s export supply tiveness of a range of nontraditional agricultural value chain: finance, management, and infrastructure. Producers chains (box 10.2). Each module built on the previous one were able to receive a higher price for their mangoes at the and progressed from a comprehensive list of sectors to farm gate—125 CFA francs versus 50 before, a 150 percent those with true marketability, competitive advantage, and increase. Similarly, exporters increased the volume of fruits comparative advantage. This process also took into account they were able to ship from Mali, which translated into the demand in existing end markets and identified new increased revenues. The impact of the pilot project has been potential end markets, regional climate and growing factors, far reaching and has encouraged entrepreneurs to engage in production capacity, access to finance, and infrastructure, similar ventures within and outside the mango sector. among other determinants. Although the full export potential of Malian mangoes is yet to be determined, APROFA conducted a number of Improving quality by investing in studies to produce estimates of such figures. In the Sikasso infrastructure and technical assistance region, for instance, where production in 2001 was esti- mated at 48,000 metric tons, it is expected that the region Mali’s revised surface transport system vastly increased the could export 14,400 metric tons (under the assumption that country’s export capacity for mangoes. In fact, once the new only 30 percent of the mangoes would be of export quality). transportation system was established and proven economi- APROFA estimates that with total production of approxi- cally efficient and reliable, the government pursued additional mately 200,000 metric tons, Mali would be able to export 50,000 metric tons (25 percent) of its mangoes. Box 10.2 The Strategic Profile Approach Consolidating success and transforming The approach that was followed to identify and prior- agricultural exports in Mali itize value chain interventions in Mali entailed the fol- lowing sequence of modules, each one comprising In 2005 the government of Mali launched the Agricultural several steps: Competitiveness and Diversification Project, a six-year investment project, funded by a $46.4 million World Bank ■ Module 1: Defining Mali’s Broad Portfolio of Agri- credit. The project’s goal is to increase revenues and com- cultural Sectors petitiveness for a range of agricultural products with grow- ■ Module 2: Analyzing Market Demand and Market ing (yet mostly untapped) markets and strong demand, thus Entry Conditions diversifying the economy’s foreign exchange earnings. In ■ Module 3: Analyzing the Competitiveness of addition, its development objective is to improve the per- Potential Malian Offerings formance of supply chains across a range of nontraditional ■ Module 4: Defining Priority Sectors ■ Module 5: Competitiveness Planning: Putting the agricultural, livestock, fisheries, and produce in which Mali Analysis into Action enjoys a strong competitive advantage, including mangoes, cashews, shallots, potatoes, dairy products, beans, papayas, Through this process, a series of operational tools sesame, and shea nuts. was developed for each of the selected value chains— The project includes five components: demonstration the Strategic Development Plan, the Competitiveness and dissemination of irrigation, postharvest, and value- Plan, and the Priority Action Plan. PCDA is now in adding technologies; improvement of the performance of the process of implementing these action plans with existing and developing supply chains; facilitation of the respective value chain stakeholders. access to finance for producers and operators; investment Source: Authors. in key collective, market-oriented infrastructure; and CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 173 interventions to increase the quality and quantity of man- continuum of projects and initiatives that supported the goes exported. A number of successive agricultural diversi- development of the mango industry in Mali between 1993 fication projects were thus started during the 2000s: and 2009. USAID’s Centre Agro-Entreprise (CAE), Trade-Mali, the To strengthen human and physical capacity and to Initiatives Intégrées pour la Croissance Economique du improve the competitiveness of mango exports, a pack Mali (IICEM) program, and the World Bank’s Programme house and logistics facility known as the PLAZA (Périmètre Compétitivité et Diversification Agricoles (PCDA). Inter- Logistique Aménagé en Zone Aéroportuaire), with capacity ventions covered by these programs included infrastructure to handle 2,000 tons of fresh produce a year, was built in cold-chain and conditioning improvements, phytosanitary 2007 (box 10.3). The PLAZA has been integral in helping improvement programs (especially the control of fruit fly mango exporters (mainly sea-freight exporters) prepare infestation), certification programs, traceability programs, their products for export with precooling and storage training in orchard management practices, and postharvest rooms, meet international standards for quality and safety, handling training programs. Figure 10.3 summarizes the and train staff. The PLAZA is the only modern pack house Figure 10.3 The Stages of Development of the Mango Export Industry in Mali Mali introduces multi- Mango exports to EU increase modal transport by Partnership b/w APROFA 600% from 1996–2006 shipping mangoes by sea & Ivorian firm to test the model API-Mali replaces CNPI New PLAZA begins APROFA established operations New investment code adopted 10 8 6 Tons 4 exports in MT Malian mango 2 0 93 96 98 00 01 03 05 06 07 09 19 19 19 20 20 20 20 20 20 20 PAVCOPA PCDA World Bank World Bank DHV Trade MALI USAID USAID CAE IICEM USAID USAID Source: J. E. Austin Associates. Note: DHV, API and CNPI are all project acronyms. Box 10.3 The PLAZA The World Bank project, PCDA, with funding from the away from the growing region and the border with Côte Dutch Embassy, built a modern pack house in Bamako, d’Ivoire. However, the PLAZA currently operates only near the airport, to help exporters improve their capaci- during the mango season. PCDA is looking at other per- ties in handling and shipping mangoes. The PLAZA has ishable export products, such as papaya and other fresh proven to be an effective pack house despite its location, produce, that could make use of the pack house. Source: Authors. 174 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN for horticulture in Mali. A precooling and shipping facility handling techniques. Since the fruit fly problem is a in Sikasso is also being considered. regional one, several regional initiatives have been estab- With assistance from the donor community, the govern- lished to eliminate the flies. Over the past three years, the ment has provided technical assistance and training in phy- West African Fruit Fly Initiative (WAFFI), jointly financed tosanitary issues. Continuous tightening of the EU food by the World Bank, the European Union, and the World safety regulations, coupled with stricter demands in terms Trade Organization (WTO), has been piloting fruit fly of traceability, are making it necessary to devise and deliver surveillance and mitigation protocols in West African substantial interventions in agricultural export countries. orchards, including those in Mali. The knowledge and Figure 10.4 illustrates the general stages of horticulture experience gained thus far has led to the design of a West export quality requirements. African Regional Action Plan to control fruit flies. Once Phytosanitary requirements are still a challenge for the substantial funding required to implement this Malian produce and several other West African countries. program (€25 million) is gathered, the large-scale inter- In the past, many shipments of mangoes were rejected ventions planned by the program should translate once they reached Europe because of fruit fly infestations into much-reduced prevalence of fruit flies in Malian or other phytosanitary concerns. These problems arise orchards, thus improving the quality of the marketed partly from a lack of proper harvest and postharvest fruit. Figure 10.4 Continuum of Commercial and Regulatory Horticulture Export Requirements + specifications for more + specifications advanced/quite for more specific process advanced and standards, yet + specifications implemented in often quite for selected, the context of specific process basic standards, highly standards with basic GAP, good integrated associated + internal hygiene and supply chains greater detail in Upgrading steps quality approaches to and where the record-keeping characteristics safe pesticicde supplier has a of products use/storage and relatively associated sophisticated + quality + basic record-keeping management grades and requirements systems structure for varietals on pesticide use quality control preferences and risk + visual + consistent management characteristics quality and quantities 1st and 2nd + 2nd and 3rd party + 2nd and 3rd party conformity + 2nd and 3rd inspections/ party conformity assessment party conformity + visual + visual assessment inspections inspection testing assessment Level 1 Level 2 Level 3 Level 4 Level 5 Level 6 Stringency of official and buyer’s requirements Level of sophistication of the conformity assessment systems Informal sector - Domestic high end horticulture Horticulture exports to high-end Int’l outlets domestic - Ethnic produce exports Low risk products High risk - Exports to wholesale markets products/sophisticate (fresh and dried products) postharvest operations Source: Jaffee 2009. CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 175 Phytosanitary issues are closely linked to the certification harvest practices, developing tools and linking value process. Recognizing that many of its mangoes are sold in chain actors to financing. International partners have supermarket chains in Europe, Mali has pursued several cer- also provided assistance in small business development tifications, including GLOBAL G.A.P. certifications. Stream- techniques. lining the export process at the airport has contributed to As part of PCDA, for example, the Dutch firm Bakker the success of Mali’s certification programs. Barendrecht teamed with five exporters operating at the PLAZA to teach the exporters how to reach European mar- Improving orchard management and kets with their products. Bakker invests in knowledge in postharvest handling interventions mango production and pays a minimum on the fruit if the exporters abide by a code of practice detailed in the contract. Orchard management and postharvest handling interven- Varieties, ripeness, size, and other details are agreed upon tions have been very important in Mali as well. To have with exporters for the entire season, with the possibility of high-quality fruit to export, quality improvement must payment of premiums if the mangoes sell well. begin at the orchard level. Training in such improvements has been provided by several institutions in recent years, including the government, international partners, and RESULTS AND IMPACTS even exporters themselves. Training has been comprehen- Mali’s development agenda places a high priority on agri- sive, with topics ranging from harvesting best practices, cultural growth and diversification. The achievements transport handling aimed at reducing damage to the realized through the PAVCOPA and PCDA projects have product, and grafting techniques to improve the varieties made a direct and tangible contribution to this agenda. A and productivity of trees. Over time, training has whole range of stakeholders participating in the value increased the capacity of mango producers, allowing for chain—small farmers, traders, agroprocessors, exporters, continued success in the subsector. service providers (technicians, financiers, and accounting specialists), and input and equipment providers—have Supply chain financing been involved since project launch and are benefiting from the expansion and improvements brought about by Through the PCDA project, local banks and other financial these projects. institutions have regained trust and interest in Malian hor- ticulture. In 2008 CFAF 150 million were lent to operators in the subsector, a record level compared with the past 10 Quantitative results years. This loan was disbursed in the form of credit for the import of shipping material (CFAF 56 million) and seasonal Results achieved after 15 years of reform in Mali’s mango credit (CFAF 94 million) for three exporters. Between 2007 subsector are summarized in table 10.3. In many cases, the and 2008, PCDA provided a guarantee to a commercial improvements are significant. The volume of mangoes bank to fund the import of transport boxes from Côte exported, for example, reached 11,995 tons in 2008, an d’Ivoire; the bank held them in bond through a third all-time record with a growth rate of 24 percent a year. party, releasing them on credit terms when the exporters (figure 10.5). Mango exports generated revenue of CFAF needed to prepare an export shipment. In 2009 the same 9.7 billion ($25 million) in the same year, a significant banks agreed to pre-finance the import of boxes directly to proportion of the earnings generated by Mali’s traditional the exporters. PCDA has established its credibility in this exports, such as cotton lint. A range of stakeholders— area and continues to play an important facilitative role farmers, harvesters, processors, and exporters—are between farmers, exporters, and professional operators on benefiting from expansion in mango trade and improve- the one hand, and the banks and financial institutions on ments along the value chain. This means increased market the other hand. share, value creation, and improved prices at all stages of mango production. The price producers receive for their mangoes at the farm gate increased by 150 percent Business training between 1993 and 2008. That price increase has allowed Capacity building has been critical for the transition for additional capacity for other on-farm activities as from subsistence mango farming to commercial farming. farmers became aware of the margins to be gained from Producers received training in crop husbandry and best horticulture production. 176 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN Table 10.3 Key Quantitative Improvements in Mali’s Mango Subsector, 1993–2008 1993 2008 Impact 1,050 metric tons exported 11,995 tons exported 1,042% Marginal sea-freighted mango exports; exports not recorded as Sea-freighted exports total 4,600 tons 460% originating in Mali $460,000 of revenue generated by mango exports Value of mango exports reached $3.4 million in 2007 639% 25-day transit time from Sikasso to northern Europe 12-day transit time from Sikasso to northern Europe –13 days Farm gate price of CFAF 50 Farm gate price of CFAF 125 150% 237 tons of European imports European imports increased to 4,560 tons 1,824% Source: Author research; revenue contribution figures from FAOSTAT. Figure 10.5 Mali Mango Exports, 1993–2008 9 8 7 6 5 4 3 2 1 0 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Mango export quantity (million metric tons) Mango export value (million US$) Source: FAOSTAT and J. E. Austin. The average quality of fruit exported from Mali has Box 10.4 Fair Trade Market Opportunities improved. The number of sea container rejections due to fruit flies, for example, dropped from 14 containers in Recent data from the Fairtrade Labelling Organiza- 2007 to 5 containers in 2008. Backward linkages at the tions show that about 8,000 tons of fair-trade-certified production level also have improved tremendously, as fruit were sold in Europe in 2005, including pineap- have relationships between exporters and farmers. Mali ples, mangoes, avocados, and citrus and deciduous has also been able to set a foothold on the Fair Trade fruits (excluding bananas). By volume, the United niche market (box 10.4). Exporters provide support ser- Kingdom accounted for more sales than any other vices to farmers such as helping to manage their planta- country, with 4,700 tons of fruit. On a per capita basis, tions, working to reduce fruit flies, and implementing however, spending was highest in Switzerland. Sales of certification or traceability programs on the plantation. fair-trade products have grown strongly since 2005. In return, exporters purchase farmers’ final product— As a brand, fair trade is making its way into the often without a contract. The trust established by these mainstream and is being taken up by supermarkets. interactions allows exporters to obtain a higher-quality The sector already has the same performance require- product, because farmers are more willing to respect phy- ments as conventional products: year-round supply and quality and price guarantees are to be expected. tosanitary controls when provided assistance. In fact, trust among all actors in the mango value chain has Source: Authors. increased over time. CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 177 In addition to improvements within Mali, regional Improving market positioning and cooperation also has improved. For example, mango diversifying market outlets exporters in Mali have organized to obtain multicountry While Mali has succeeded in increasing sea-freighted exports support from donors to fight against fruit flies, and the since the 1990s, there is still room for better market posi- multimodal transport initiative involved coordination tioning and diversification of market outlets. First, Mali, among several stakeholders across borders. These efforts whose mangoes represent only 2 percent of the total mango have led to an improved value chain not only in Mali but in imports to the European Union, is still a marginal player in Côte d’Ivoire, Senegal, and other countries. the global market for mangoes (figure 10.6). Second, the European Union and other importing markets are dynamic, Qualitative changes meaning that there is constant need for adjustment and improvement to meet market demand. Mali is currently The initial, and now sustained, take-off in the growth of supplying the EU market based on the seasonality of its pro- mango exports has led to a complete transformation of the duction, which is concentrated between mid-May and early subsector, not only in quantitative terms but also qualita- July, after Burkina Faso but before Senegal (figure 10.7). tively. The expansion of the subsector has brought about a Going forward, it will be important for Mali to extend progressive and likely irreversible change in business prac- its harvest season and develop markets other than Europe. tices in the sense of increased professionalism and attention To address this, PCDA is testing the irrigation of mango to product quality, better compliance with trade standards, orchards, which may offer opportunities to induce and increasing interest in private investment. flowering and fructification at a different period of the For one, major players in the mango business are now year. Local exporters also have begun seeking market present in Mali and have an interest in expanding their opportunities in the Middle East and North Africa, operations and making long-term investments in the coun- among other places. Mali has begun to export mangoes to try. AHOLD, a major Dutch supermarket retail chain, has Morocco. There are also substantial opportunities in operated in Mali since the PLAZA and has been providing other Sub-Saharan country markets. technical assistance to Malian exporters and PCDA. In addition, substantial work has been undertaken on the Continuing to invest in quality and upstream/production level of the production chain. Mango product differentiation production is now considered a legitimate agricultural In recent years Mali has focused on obtaining certifications activity, not just fruit collection. (PCDA, for example, has that will allow mangoes to enter the European market, such invested in a study that aims to map tree crop plantations using satellite imaging.) Financial institutions have also shown renewed interest as demonstrated by the increasing volume of credit to the subsector, the low default rates, and Figure 10.6 Volume of Mango Imports to the European the emergence of innovative financing instruments, such as Union, by Country of Origin, 2008 input prefinancing. Finally, Mali’s Mango Task Force has Other Costa Rica Mali been active for some years to improve coordination in the 2% 10% 2% subsector, develop an agenda for collective action between Senegal 3% professionals, and provide a platform to address issues of common interest with the public sector. USA 3% Cote Brazil THE WAY FORWARD 42% d'Ivoire 5% While the mango industry in Mali has grown in recent years, it will continue to develop in the years ahead. Mali has Israel given priority to a number of areas, including market posi- 5% Peru tioning, improving quality and quantity, attracting capital, 22% adding value to mango exports, improving transportation Pakistan 6% infrastructure, and potentially establishing more links at the regional level. Source: Eurostat. 178 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN Figure 10.7 Seasonality of Mango Imports to the European Union, by Country of Origin, 2008 (metric tons) 25,000 20,000 15,000 10,000 5,000 0 1 2 3 4 5 6 7 8 9 10 11 12 Months Others Mali Senegal Costa Rica USA Israel Pakistan Cote d'lvoire Peru Brazil Source: Eurostat. as GLOBAL G.A.P. the certification required by supermar- of mangoes a year with improved management of existing kets in the European Union. Collaboration along the value mango plantations. Effective management includes intro- chain led exporters to assist producers to achieve certifica- duction of new varieties, improved grafting and replanting tion through training, with the support of PCDA. Some of new trees for increased yields and quality, additional producers have even obtained certification of their man- investments in pack houses and cold storage, renewed goes as organic. The market for organic fruit and vegeta- efforts in the control of fruit flies, and procurement of tech- bles is a growing one, particularly in countries of northern nical advisory services. However, in 2009, against an initial Europe and Switzerland. export program of 104 containers, the PLAZA could finally ship only 42 containers because of the lack of mangoes of Ensure compliance with standards and exportable quality. phytosanitary requirements Attracting capital The constant evolution of EU food safety regulations toward more stringent phytosanitary requirements, coupled with Developing commercial value chains is a private sector busi- demands for traceability, make it necessary to devise and ness. It requires capital and know-how. If an economic sec- deliver specific interventions in the produce-exporting coun- tor does not manage to attract investors and capital, both tries to meet market requirements. Recently PCDA financed national and foreign, sustaining growth will be difficult. In an intervention by the Comité de liaison EU-ACP pour les general, governments should encourage commercial farm- fruits et légumes (COLEACP) specific to the mango value ing by facilitating access to land concessions. In Mali’s chain aimed at building capacities in health and phytosani- mango subsector, the government should make financing tary risks control. A series of workshops, which brought more readily accessible and provide incentives to banks to together mango stakeholders of Mali and Burkina Faso, were lend to stakeholders. These actions would stimulate private held to train participants on risk assessment and mitigation. sector investment into value-addition activities such as dry- This training led to the creation of a guidebook for the mango ing, canning, and juice production. Burkina Faso has had value chain that is now followed in both countries. some success in producing dried mangoes, and Mali has the potential to achieve similar success. Continuing to focus on improvements in quantity and quality Diversify to better utilize the PLAZA terminal Although estimates vary greatly, some experts estimate that To maximize the use of the PLAZA and ensure sustainabil- Mali could export between 20,000 and 50,000 metric tons ity, commercial operations must be able to cover PLAZA CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 179 operating costs by the end of 2012, when the public funding and higher risks as Malian exporters operate through Côte stops. Several exporters are looking at opportunities to d’Ivoire. Improvements in alternative transport routes export green beans and other vegetables that could be would certainly benefit the mango subsector. Mali could packed at the terminal. To ensure profitability and com- also improve the road-to-rail corridor in order to reduce petitive pricing for its members, there are plans for a transport times and increase quality. strong professional organization to eventually take over To develop a vibrant and competitive export horticulture management of the PLAZA. sector, Mali also needs to encourage the development of related industries and services, such as certification, packag- Further invest in marketing infrastructure ing material, and inputs. These related industries are key to improving the competitiveness of the value chain(s) by A mechanical grading line is being installed to improve reducing the cost of inputs, technology, and services, and turnover at the PLAZA, enhancing the ability to quickly also by providing an important source of revenue and meet demand quantity and quality in the European employment in Mali. market. The equipment was ordered by PCDA, and the calibrating machine should be installed for the 2010 sea- son. Assuming an adequate fruit harvest, the PLAZA Joining forces at the regional level should now be able to handle over 100 containers in the Finally, when examining trade figures, it is evident that the 12- to 14-week mango season. competition for West African mango sectors does not come from neighboring countries, but from Central and South Adding value to consolidate the achievements America. The concentration of mango exporters within the West African region, however, suggests advantages from Strengthening and diversifying the value chain, in addition coordinating the production and export of mangoes to to valorizing the large mango surplus that is not exportable, Europe at a subregional level. Mali could reap significant is critical for the future growth of the mango subsector. advantages from coordinating mango production and While fresh mango exports were an entry point for Mali to export with Côte d’Ivoire, Burkina Faso, and Senegal. Such reach the EU market (and indeed, demand for fresh man- coordination could result in an increase in the scale of goes is expected to continue in the foreseeable future), fresh export and greater efficiencies in shipping and standardized mangoes may not be extremely profitable in the long run, quality control, better handling of traceability, and brand- and Malian professionals need to seek different ways to add ing. All these factors would contribute to increased volumes value in the mango subsector. PCDA has been working on of exported mangoes and, ultimately, to increased revenues several clear opportunities: processing, increasing the range for producers. One possibility would be to create a West of products offered, expanding other value chains (papaya), Africa brand for mangoes. and establishing new value chains. A recent study (Royal Tropical Institute 2010) shows the potential for exports of dried mango from Mali to European LESSONS AND REPLICABILITY markets, provided the product meets the expected quality standards. For dried mangoes, Mali has an advantage over The driving forces behind Mali’s experience in developing Burkina Faso in the sense that it is relatively unencumbered its mango sector, as well as the lessons learned during the by an obsolete technology and enjoys a large surplus of development process, are useful to point out for other coun- mango varieties that are not exportable as fresh. Recently, tries in the course of such a transition. PCDA agreed to take up one of the study’s recommenda- tions and finance a pilot project to produce dried mangoes Drivers of success based on technology established in South Africa, which cur- rently meets the quality standards in EU markets. Innovation Is Critical in Triggering and Driving Change. In the case of mango exports from Mali, the key initial innovation was in transport and logistics. Together, Further improving transport and these changes created greater opportunity to access the large fostering cluster development EU market. The innovations also created the dynamics of The inefficiencies the Bamako-Dakar railway and at the change and initiated a learning process. There is no univer- Dakar port are resulting in higher costs, longer wait times, sal formula for achieving innovation, however; innovation 180 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN derives from a combination of factors that depend on the tools in identifying opportunities and articulating opera- specific country and sector context. In the case of the mango tional strategies to create greater value in agriculture and subsector of Mali, two key resources, expertise and entre- agribusiness. It is also essential to find a key entry point that preneurship, were brought together to improve the export responds to market demand and helps increase scale. For capacity of the industry. Additionally, creativity in funding Malian mangoes, this meant finding a more effective way to mechanisms is another important lesson learned in the case get the product to the market, while innovations in trans- of Mali. port and logistics system allowed farmers and exporters to achieve the economies of scale they needed. A structured, Time Is of the Essence. Building capacity—human, phys- holistic approach is required. ical, or otherwise—in a new or small industry requires sus- tained efforts over time, especially when starting from Sustaining Development Efforts over Time Is Neces- scratch or from a very low base, as was the case in Mali. sary. Mali could not have successfully transformed its mango sector if activities had not been pursued year after High-Quality Technical Work Is Key. The importance year. Building an industry takes years, not months. The les- of market research, value chain cost analysis, benchmarking, son here for governments and development partners is that and assessment of industry constraints cannot be over- if they not prepared to dedicate time and resources to a stated. This work is critical in identifying and designing reform over a long period of time, it is probably better not action plans, programs, and business solutions. In the case to start at all. That said, early success in the reform process, of Mali, it was perhaps the most critical factor contributing such as what Mali achieved through the positive outcome of to launching the reforms. the multimodal export pilot launched by APROFA in 2000, can serve as a foundation on which to build momentum for Private Sector Leadership Is Also Important. Even further reforms. though private sector involvement in the mango sector in Mali was weak at the start of the transition, it was neces- Aid Funds Can Play a Catalytic Role in Change sary to work with the existing private operators and even- Processes. Project aid has the capacity to provide both tually bring in new ones, such as the Ivoirian company financial resources and know-how to share risks in order to that conducted the pilot export test. In the case of Mali, facilitate innovation. In Mali, technical assistance programs the partnership with SN Tropical Expressions was unique targeting postharvest handling, supply chain finance, and provided a good model of what public-private part- export, and other areas have helped improve the perfor- nerships can achieve. mance of the mango value chain and are now doing the same for other agricultural and nonagricultural sectors. Knowledge and Funding Must Be Packaged and Technical assistance by aid organizations can also help Delivered Properly. Building capacities in emerging address market failures and can be combined in a proactive subsectors and industries takes not only time and perse- fashion with efforts and interventions by the private sector. verance but also the ability to deliver investments in “hardware” (infrastructure and other means of produc- Collaboration and Partnerships Are Part of the tion) and “software” (training and knowledge transfer) in Solution. Public-private partnerships (between donor a flexible manner. In the case of Mali, capital investments agencies and government, or between the private sector and such as PLAZA were imperative to the success of the government), such as the arrangement to build and manage mango subsector. Training alone could not bring about the PLAZA, are key in developing infrastructure that will, in such a transition. Improvements to cold chain, trans- turn, help a country expand a specific sector. In Mali the portation, and conditioning facilities within the horticul- emergence of the Mango Task Force helped provide a space ture sector significantly improved Mali’s capacity to for public-private cooperation, as well as an opportunity for export quality mangoes. exporters to coordinate their respective shipments through a common buyer. In addition, the use of “Mali mango” logo on boxes created closer cooperation among shippers. Lessons learned A Targeted, Organized Approach Is Necessary to A Favorable Investment Climate and Investment Sector Reform. Sector and value-chain analysis are essential Policies Remain Important. Thus far the government of CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 181 Mali’s commitment to reform in its mango sector has been farmer groups, receptive businesses, and a facilitating public carried out through diversification, donor financing, and sector. Actors in all categories must be willing and able to improved technical expertise. The industry, however, is now play their role, and even then establishing sustainable mar- experiencing a new generation of issues that will require sig- ket inclusion needs a public sector body to facilitate the nificant reforms in the general business environment by process. This is the role that APROFA and its successive promoting private sector investment, and improving access projects have played for the mango subsector in Mali and to land. Recent difficulties in implementing a Foreign the role that such projects are still playing for a number of Investment Advisory Service (FIAS) technical assistance other emerging value chains in the country (papaya, cashew, program on agribusiness do not bode well in terms of the sesame, potato, and onion). determination of the government to move to the next stage of development of agribusiness value chains. CONCLUSION Since the early 1990s Mali has achieved a spectacular Scalability and transferability increase in its exports of fresh mangoes, seizing opportuni- Two general lessons from Mali’s experience with mangoes ties offered by an increasing market demand in Europe and can be applied to other countries. First, designing market- finding solutions to overcome significant physical hurdles. led strategies and investing in applied value-chain analysis is The initiatives taken in the mango sector in Mali clearly necessary if the objectives of creating higher value in agri- demonstrate that even a physical constraint as immutable as culture, raising incomes, and lifting the rural population out being landlocked can be overcome to some extent through of poverty are to be achieved. While general, cross-cutting innovative solutions. In Mali the entry point was the identi- agricultural support programs are necessary, they are in fication and economic validation of a new transport and themselves not sufficient to make a difference. Specific logistics arrangement that allowed mangoes to be exported information and analysis must be generated on the issues to in large volumes and in good condition. be addressed throughout the value chain(s). Mali currently is focusing on consolidating and expand- Second, it should be kept in mind that sustainable mar- ing its initial success by strengthening stakeholder involve- ket inclusion requires multiple interventions. As demon- ment and private sector partnerships to ensure growth and strated by a recently completed research program funded by sustainability. The Mango Task Force has emerged as the the U.K. Department for International Development called sector’s professional organization, with potential to become “Re-governing Markets,” inclusion of small farmers in mod- a permanent trade/business association capable of handling ern agricultural value chains does not happen by itself. In the new issues facing the horticulture industry in Mali. fact, inclusion requires interventions by multiple actors, as summarized in figure 10.8. The study indicates that three BIBLIOGRAPHY major categories of actors need to be involved: farmers and Antoine, M. 2005. “Growing Malian Mango Exports: How New Business Models Translate to Increased Opportuni- ties for West African Producers.” World Bank. Figure 10.8 Multiple Actors Necessary for Sustainable Markets APROFA (Agricultural Value Chain Promotion Agency). 2002a. “Mango Exports to Europe out of Sikasso-Mali: Multiple interventions Multimodal Shipping and Cross-border Partnerships.” Bamako. ———. 2002b. “Note sur l’expérience de Sikasso concer- Receptive business Trained/ nant l’amélioration des revenus des petits planteurs de Partnership sector facilitation empowered mangues. ” Bamako. farmers ———. 2002c. “Projet de diversification des filières agri- coles et de l’horticulture : Livrable 1.1: Diagnostic des principales filières agricoles (mangue, haricot vert, pomme de terre et oignon).” Bamako. Facilitating public Birthal, P., A. Jha, and H. Singh. 2007. “Linking Farmers to sector Markets for High-Value Agricultural Commodities.” Source: UK Department for International Development. Agriculture Economics Research Review 20: 425–39. 182 CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN CARANA Corporation. 2004. “Impact of Transport and Logis- Debate.” Policy Research Working Paper 3348, World tics on Mali’s Trade Competitiveness.” Washington, DC. Bank, Washington, DC. Chemonics International. 2003. “Mali SEG Projet Centre Jaffee, S., P. Labaste, R. Kopicki, and I. Christie. 2003. “Mod- Agro-Entreprise: Final Report Deliverable 4.” Washington, ernizing Africa’s Agro-Food Systems: Analytical Frame- DC. work and Implications for Operations.” Africa Region Club du Sahel/OECD (Organisation for Economic Co- Working Paper 44, World Bank, Washington, DC. operation and Development). 1998. “L’économie locale Jones, C., and G. Livingston. 2009. “Supporting Agribusi- de Sikasso.” Paris. ness Investment in Mali.” J. E. Austin Associates, Danielou, M., P. Labaste, and J.-M. Voisard. 2003. “Good Arlington, VA. Practice Note: Linking Farmers to Markets: Exporting PCDA (Programme Compétitivité Diversification Agri- Malian Mangoes to Europe.” Africa Region Working coles). 2007. “Mango Value Chain Competitiveness Paper 60, World Bank, Washington, DC. Plan.” Bamako. IF (Integrated Framework for Trade-Related Technical Rey, J.-Y., et al. 2004. “La mangue en Afrique de L’Ouest Assistance to Least-Developed Countries). 2004. Francophone: variétés et composition variétale des verg- “Expanding and Diversifying Trade for Growth and ers.” Fruits 59 (3): 191–208. Poverty Reduction: A Diagnostic Trade Integration Royal Tropical Institute. 2010. “L’amélioration des perform- Study.” http://www.integratedframework.org/files/mali ances de la filière des produits séchés de la mangue au _dtis_17nov04.pdf. Burkina Faso et au Mali.” Amsterdam. Ezzine, A. 2009. “Etude des corridors pour l’export des USAID (U.S. Agency for International Development). 2002. mangues du Mali et du Burkina Faso vers l’Union “Mali Trade Capacity Needs Assessment—Sub Sector Européenne.” World Bank, Washington, DC. Analysis.” Technical Report, Washington, DC. Fairbanks, M., and S. Linsay. 1997. “Plowing the Sea: Nur- World Bank. 1995. “Staff Appraisal Report, Agricultural turing the Hidden Sources of Growth in the Developing Trading and Processing Promotion Pilot Project.” World.” Harvard Business School Press, Cambridge, MA. Washington, DC. Geomar International and PCDA. 2006. “Strategic Profile of ———. 1996. “Mali Agricultural Trading and Processing Mali’s Agricultural Exports.” Geomar International, Promotion Pilot Project.” Project Appraisal Document Montreal. 25884, World Bank, Washington, DC. Gulati, A., N. Minot, C. Delgado, and S. Bora. 2007. “Growth ———. 2003a. “Implementation Completion Report in High-Value Agriculture in Asia and the Emergence of Republic of Mali Agricultural Trading and Processing Vertical Links with Farmers.” In Global Supply Chains, Promotion Pilot Project.” Washington, DC. Standards and the Poor, ed. J. Swinnen, pp 91–108. ———. 2003b. “Mali: Exporting Mangoes to Europe.” Wallingford: CABI Publishing. Washington, DC. Hallam, D., et al. 2004. “The Market for Non-Traditional ———. 2004. “Project Appraisal Document on a Credit and Agricultural Exports.” Food and Agriculture Organiza- Proposed Grant to the Government of Mali for the tion, Rome. Transport Corridors Improvement Project.” Washington, Jaffee, S. 2009. “African Smallholders, Higher Value Agri- DC. Food Markets and Measures to Achieve Compliance with ———. 2005. “Project Appraisal Document on a Proposed Emerging Standards.” World Bank, Washington, DC. Credit to the Republic of Mali for an Agricultural Com- Jaffee, S., and S. Henson. 2004. “Standards and Agro-Food petitiveness and Diversification Project.” Washington, Exports from Developing Countries: Rebalancing the DC. CHAPTER 10: GROWING MALI’S MANGO EXPORTS: LINKING FARMERS TO MARKET THROUGH INNOVATIONS IN THE VALUE CHAIN 183 CHAPTER 11 Economic Liberalization in Rwanda’s Coffee Sector: A Better Brew for Success Karol C. Boudreaux n 2008 Solberg and Hansen, a Norwegian importer of I income or profits were limited. Starting in the late 1990s, the high-quality specialty coffee, bid just under $40 a kilo- government liberalized the sector, removing a variety of gram (more than 21,000 Rwandan francs) for a load of barriers to trade, creating new incentives for groups and coffee from Rwanda.1 While this bid was exceptional— individuals to invest in coffee production, and facilitating specialty coffee normally sells for between $3 and $4 a entrepreneurship in the coffee industry. kilogram—it also represents a real and positive transfor- Working with the private sector and international mation within Rwanda’s coffee sector. In the not-too- donors, the government of Rwanda has reshaped the coffee distant past, the country was known as a producer of industry: the regulatory framework for production has been mediocre coffee that attracted little attention from dis- modified; more than 100 coffee-washing stations have been criminating importers or consumers.2 Today, Rwandan built; donors have supported the development of market coffee is increasingly recognized as a high-quality prod- links between producers and foreign buyers; cooperatives uct, one for which importers such as Solberg and Hansen have formed; and smallholder farmers are working together and, in turn, consumers are willing to pay a premium. in an effort to improve the quality, marketing, and branding Although the Rwandan economy is diversifying, agricul- of their coffee. ture continues to be the primary source of livelihood for These changes have had important effects on the ground the vast majority of the population. The overwhelming in Rwanda. Coffee continues to generate important export majority of agricultural workers are subsistence farmers. revenue for the country: just over $47 million in 2008, com- Just over 10 percent of these farmers planted coffee in pared to $35 million in 2007. Higher incomes benefit farm- 2008, and the crop remains a major source of export rev- ers, their families, and their communities in a variety of enue for Rwanda, generating more than 36 percent of ways: farmers can improve a home, pay medical expenses total export revenue in 2009. or school fees, or better ensure food security. And when cooperatives earn a profit they are able to hire workers, purchase capital, and support community projects such as The evolution of coffee production in Rwanda improved schools. The transformation of Rwanda’s coffee sector happened rel- The transformation of the Rwandan coffee sector may be atively quickly. In 2000 Rwandan farmers were producing creating social benefits in addition to economic benefits. semi-processed coffee for sale on world markets. Farm gate Working together at coffee-washing stations—typically prices paid to farmers were low (60 Rwandan francs a kilo- located in the rural, hilly, and relatively inaccessible areas gram), and prospects for farmers and exporters to increase where coffee grows and where little other commercial 185 infrastructure exists—farmers have new opportunities to marketing and sales problems resolved, incentives interact with other Rwandans. These repeated interactions strengthened to produce higher-quality beans, and harm- may reduce the sense of ethnic distance among members ful government interference avoided, then the positive of Rwandan society. As farmers and other workers at gains of the past several years should continue, and coffee-washing stations experience the increased eco- Rwanda’s smallholder farmers can look forward to earning nomic satisfaction that comes with higher income earned more income from coffee production. This income should, from coffee, they may also feel greater levels of trust in turn, filter through local economies to spread benefits toward people with whom they interact and thus develop to other Rwandans. more positive attitudes toward reconciliation. And because coffee in Rwanda is grown by poor, smallholder farmers who make up 90 percent of the population, these positive REFORMING A VITAL SECTOR changes have the potential to benefit a broad swath of Government control and limited markets Rwandan society. In many developing countries, governments are heavily involved in the agriculture sector, and that certainly holds Challenges and concerns true in Rwanda, where coffee has been a major export for Despite this good progress, the coffee sector faces a number decades. The Belgian government as well as the two inde- of serious challenges. To start, Rwanda’s landlocked geo- pendent, pre-genocide governments controlled important graphical status and poor infrastructure mean that coffee aspects of the coffee trade for their political and financial producers face high transport costs. Moving coffee cherries gain. Through compulsory production, export taxes, and a quickly over Rwandan roads is difficult, as is moving monopsony export control agency, these regimes captured processed beans out of the country in a timely and cost- the profits of mostly poor coffee farmers, and used the effective manner. Other concerns are related to the costs in funds to help maintain political power (Bates 1981). Pro- the industry. Production costs remain high. Many of the ducers had little incentive to invest in the production of newly built coffee-washing stations are operating at much high-quality coffee, and so for decades Rwandans produced less than full capacity, and labor costs are higher than in a small volume of low-quality coffee. neighboring countries. Rwandan farmers are also less pro- Significant government involvement in Rwanda’s coffee ductive than farmers in neighboring countries. Although sector began in the 1930s, when the Belgian colonial gov- some have received support and training from nongovern- ernment launched a series of “coffee campaigns.” Govern- mental organizations (NGOs), regular visits from extension ment authorities built nurseries and supplied seeds, but agents are limited. they also required Rwandan farmers to plant coffee trees In addition, a variety of management concerns have (Dorsey 1983). The government also introduced price plagued the cooperatives that many coffee farmers join. restrictions, imposed mandatory quality guidelines, and One specialist of Rwanda’s cooperatives observed, “After issued special licenses that allowed only some firms to pur- five years of extensive cooperative capacity building, chase coffee. Export taxes were imposed on coffee sales, and Rwanda’s coffee cooperatives remain surprisingly fragile, individual income taxes were imposed on producers, who unorganized, and dysfunctional” (SPREAD 2007). Some at the time were mostly Hutu farmers. Tutsi chiefs collected cooperatives have mishandled loans, while others have failed these taxes, which helped support them and the colonial to fulfill contracts in a timely manner or have encountered government. trouble marketing their products. Some of these problems Following Rwanda’s independence, the Kayibanda gov- are the result of a lack of training or financial management ernment (1962–73) retained most of these policies because skills. it had limited alternatives for raising revenue. A government Other challenges involve the broader institutional envi- marketing board (OCIR, subsequently OCIR-Café), ronment. As Rwanda implements a new land law, some together with a monopsony export company, Rwandex, smallholder farmers may face uncertainty about the rights purchased, and then sold on world markets, the vast majority to the land that they work. Women, in particular, may be of coffee grown in Rwanda. The farm gate price was set by especially vulnerable. The government, NGOs, and other the government.3 Middlemen bought beans from farmers stakeholders are taking measures to deal with these various and sold them to Rwandex, which in turn sold them to challenges, however. If capacity issues can be addressed, foreign buyers. The locations where smallholders brought 186 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS their beans for purchase acted as “the economic arm of the To rebuild its popularity, the regime diverted attention Gitarama (that is, Kayibanda) regime” (Verwimp 2003). from its own economic policies to the Tutsi/RPF threat Heavy government involvement in the coffee sector con- and increased levels of repression within Rwandan soci- tinued under the Habyarimana regime (1973–94). During ety. The government demonized the invaders, arguing the 1970s and 1980s, as world coffee prices rose, coffee that allowing Tutsis into the country would lead to Hutus exports provided between 60 and 80 percent of Rwanda’s having less access to already scarce land, and used the export revenue (Berlage, Capéau, and Verwimp 2004). media to foment ethnic hatred. Most repression was President Habyarimana ensured control of these important directed at the Tutsi minority, although some spilled over rents by appointing relatives and political supporters to to Hutus. The ultimate results were, of course, disastrous. In positions of authority at OCIR-Café.4 the three months between April and June of 1994, approxi- mately 800,000 people were murdered in Rwanda, mostly Tutsis but also moderate Hutus and Hutu opponents of the Crisis and response Habyarimana government. During the early and mid-1980s, rising coffee prices allowed the government to modestly increase the price it paid to Liberalization to open markets and farmers for their beans. Verwimp (2003) indicates that increase opportunity high market prices for coffee had another effect—they “allowed the [Habyarimana] regime’s elite to increase both Today, the Rwandan government is less directly involved in its personal consumption and its power over the popula- the coffee sector. Farmers have more choice about what to tion” (p. 172). The government then used the additional grow, to whom to sell their beans, and how to market their revenue to buy support in rural areas (through higher farm product. Private sector investment in the sector is rising. This gate prices for coffee and subsidized agricultural inputs) increased openness is part of a larger government effort to and to spend more on monitoring the population. improve economic growth in the country. Rwanda’s Vision Tumbling coffee prices in the late 1980s meant reduced 2020 is a strategic plan for economic change. This plan has, government revenue. For a few years, the government since 2003, provided a guideline for sectoral policy setting attempted to keep payments to farmers stable, but this policy with Rwanda’s ministries (Rwanda Ministry of Finance and proved unsustainable, especially after 1990, when the govern- Economic Planning 2000; IFAD 2006). ment needed resources to fight the invading Rwandan Patri- The goals created by Vision 2020, together with Rwanda’s otic Front (RPF) forces. By the early 1990s the government Poverty Reduction Strategy Paper (PRSP) and the subse- was forced to lower prices paid for coffee cherries to small- quent Economic Development and Poverty Reduction holder farmers. Price supports ended completely in 1992. Strategy (EDPRS), include improving the institutional envi- The fall in coffee prices, coupled with growing military ronment to allow for increased private sector development expenditures, meant that the Habyarimana government and infrastructure improvements, focusing on good gover- faced severe fiscal constraints. Seeking alternate sources of nance (including democratization, national reconciliation, revenue, it began confiscating property and raising taxes to political stability, and security), improving agricultural supplement the budget, and there was some amount of productivity, improving human capital through invest- reduction in consumption by elites. Foreign aid also became ments in health and education, creating a service-based an increasingly important part of Rwanda’s budget. How- economy with a focus on ICT (information and communi- ever, the reliance upon, and competition for, foreign aid cre- cation technology), reducing external support, relying more ated serious problems among governing elites. As one on exports, and promoting regional integration. observer puts it, “the various gentlemen’s agreements which Some progress has been made toward achieving the Vision had existed between the competing political clans since the 2020 goals. Real GDP growth has been strong for more than end of the Kayibanda regime started to melt down as the a decade: 10.8 percent over 1996–2000 and 6.4 percent over resources shrank and internal power struggles intensified” 2000–06 (Rwanda Ministry of Finance and Economic Plan- (Prunier 1995, 47). With its growing dependence on for- ning 2007), reaching a high of 11.2 percent in 2009 (although eign aid, and in a bid to remain in power, the government it was estimated to fall to under 5 percent for 2010) (Index agreed to an International Monetary Fund (IMF) struc- Mundi 2010) With an emphasis on private-sector-led tural adjustment program in 1990 that imposed further growth and improvements in the environment for doing hardships on farmers. business, the economy is diversifying, merchandise trade CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 187 levels are rising, and the service sector is expanding. Partic- More substantial reform efforts began in 2000, when the ularly for rural Rwandans, reform in the coffee sector is government, working with consultants (the OTF Group has playing an important part in helping thousands of farmers been the leader in these efforts) and donors, studied the increase their incomes, by creating jobs and providing potential to add value to Rwandan coffee through the pro- opportunities for new skills training. The reform measures duction of higher-quality, washed, and fermented specialty are also strengthening human and social capital and, in the coffee. Part of the catalyst for this redoubled effort came as process, may also be generating valuable social benefits. prices for coffee of ordinary quality fell to a historic low in Most Rwandans—especially rural Rwandans—are still 1997 and farmers became increasingly unwilling to invest in very poor, however. More than one-third of the population is the crop. The quantity of coffee produced declined as the unable to meet its minimum food requirements and rou- tree stock aged and was not properly tended, soil fertility tinely goes hungry (Rwanda Ministry of Finance and declined, and farmers battled with insects and fungal dis- Economic Planning 2007). Although the U.K. overseas devel- eases. As a result, the quality of Rwanda’s coffee crop tum- opment agency, the Department for International Develop- bled: by 2000, 90 percent of Rwanda’s crop was classified as ment (DFID), reports that the poverty rate dropped from low-quality, “ordinary” coffee (Rwanda Ministry of Agricul- 70 percent of the population in 1994 to less than 57 percent ture and Animal Husbandry and Ministry of Trade and in 2006 (DFID 2008), the World Bank (2010) estimates that Industry 2008). Rwanda’s gross national income per capita was only $460 in 2009. Thus Rwanda has a long way to go to meet the Vision Rwanda’s national coffee strategy 2020 goal of transforming itself from a low-income to a middle-income society. In 2002 the government issued a National Coffee Strategy that outlined a plan for capturing a larger share of the spe- cialty coffee sector. A production target of 44,000 tons of Modest gains and encouraging signs coffee was set for 2010, 63 percent of which would be fully Changes in the coffee sector began shortly after the geno- washed (figure 11.2).6 These production targets have, how- cide, when the government opened the market for coffee ever, never been met. export to increased competition and began to focus on The shift away from low-quality coffee to high-quality improving the value chain for coffee (figure 11.1). The specialty coffee was designed to break a perceived “low- method for pricing beans also changed. Rather than dictat- quality, low-quantity trap.” As noted, Rwandan farmers ing a single price for the entire season, the Rwanda Coffee were producing low-quality coffee that sold for a low price. Development Authority (the former OCIR-Café) sets a Low sales prices meant that farmers lacked revenue to invest minimum weekly reference price, in consultation with in improvements. Without sufficient income, farmers could stakeholders, a basis from which a sales price per kilogram not invest in capital to improve the quality of their beans— may be negotiated (Mutandwa et al. 2009).5 hence, the trap. Figure 11.1 Value Chain for Washed Coffee Buyers Farmers private coffee- Millers/ Auction Importers Retailers Consumers washing stations exporters (Mombasa) or co-op Source: Author. 188 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS Figure 11.2 Targets Set Out in Rwanda’s Fully Washed Coffee Strategy 2002 50,000 45,000 40,000 35,000 30,000 63% Tons 25,000 20,000 15,000 10,000 25% 5,000 12% 0 02 03 04 05 06 07 08 09 10 20 20 20 20 20 20 20 20 20 Fully washed Standard Ordinary Source: Rwanda Ministry of Agriculture and Animal Husbandry and Ministry of Trade and Industry (2008). The 2002 National Coffee Strategy was designed to break established cafés that feature local coffee. In other words, this trap, increase income and revenue, and improve price along a value chain, Rwandans are benefiting from the stability, because specialty coffee prices fluctuate less dra- opportunity to produce and sell specialty coffee. matically than does the price for semiwashed coffee. In an At the production level, thousands of Rwanda’s small- attempt to meet these targets, coffee-sector stakeholders holder farmers are benefiting from higher coffee prices for focused on strengthening and supporting producer cooper- fully washed specialty coffee. NGOs such as PEARL and atives, identifying sites for and supporting the building of SPREAD have helped farmers establish cooperatives and coffee-washing stations, replanting aging tree stock, have trained co-op members in quality control, processing, improving quality control throughout the industry, and and marketing efforts. To date, more than 100 washing sta- strengthening the Rwandan brand globally. And as one of tions have been built around the country with the support four focal sectors identified in its 2008–12 Economic Devel- of the government, donors, NGOs, and the private sector opment and Poverty Reduction Strategy, the coffee sector (table 11.1). As a result, Rwanda is producing more high- remains a high priority for the government. quality coffee and demand for the country’s specialty coffee is increasing. DIRECT ECONOMIC BENEFITS OF LIBERALIZATION Incomes are rising for farmers and cooperatives The liberalization of the coffee sector has had a number of Perhaps the most important effect of the liberalization of positive effects. First, farmers now have an incentive— Rwanda’s coffee industry is that more of the farmers increased income—to shift some production from semi- (approximately 500,000) who grow coffee have an oppor- washed to fully washed coffee. Entrepreneurs are investing tunity to sell their beans for higher prices. The price that in building coffee-washing stations, where cherries are cooperatives and private sector coffee-washing stations are processed for sale. Rural communities are forming coopera- paying farmers for cherries rose from between 60 and 80 tives, some of which also build washing stations and process Rwandan francs in 2004 to between 160 and 180 Rwandan cherries. Exporters are competing for opportunities to sell francs in 2008 (Rwanda Ministry of Agriculture and Animal Rwandan coffee to foreign buyers. Other entrepreneurs have Husbandry and Ministry of Trade and Industry 2008).7 One CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 189 Table 11.1 Growth in the Specialty Coffee Sector 2002 2003 2004 2005 2006 2007 2008 2009 Number of washing stations 1 10 25 45 76 112 112 112 Tons of green specialty coffee exported 30 300 800 1,200 3,000 2,300 2,455 3,045 Number of specialty coffee buyers 2 8 16 25 30 30 — — Total value of specialty coffee exported ($1000s) 90 720 1,850 3,168 8,000 7,800 8,060 11,600 Source: USAID (2009). Note: — Not available. study finds that farmers who sell coffee cherries to wash- months between harvests more easily than before (personal ing stations increase their annual expenditures by 17 per- interviews of the COOPAC cooperative 2006). cent compared with farmers who sell lower-quality parch- Once coffee cherries are washed, fermented, and dried, ment coffee (Murekezi and Loveridge 2009). The same they can be sold to buyers for an even higher price. In 2004 study indicates that since reform, coffee farmers have the Maraba cooperative sold washed coffee for $3.26 a kilo- increased their food consumption and their overall house- gram; in 2007 the cooperative was able to charge $4.08 for hold expenditures, leading to improved food security and the same amount. The COOPAC cooperative was selling its to generally improved economic conditions for coffee cherries for $4.00 a kilogram in 2007, and the Rusenyi coop- farmers. erative was selling in a range between $4.40 and $5.50 In a 2008 survey of 239 farmers and coffee-washing sta- (Swanson and Bagaza 2008). On average, in 2007 coffee- tion workers, Tobias and Boudreaux (2011) asked farmers washing stations sold their fully washed coffee for $3.60 a to identify benefits they received as a result of being a kilogram. In a notable achievement, importers paid a record member of a coffee cooperative.8 Farmers listed a number high of $55 a kilogram (approximately $25 per pound) for of direct financial benefits, such as increased prices the best Rwandan coffee in September 2007, a price compa- received for their cherries, employment opportunities, and rable to the world’s most expensive coffees (Rwanda Devel- better and easier access to loans, particularly access to opment Gateway 2007). credit to purchase inputs such as fertilizer. Farmers also Cooperatives use the income they generate to pay indi- noted that their families are now better fed, that they are vidual farmers for cherries, to repay loans taken build wash- able to hire laborers, that they have help with marketing ing stations (or for other equipment), to pay salaries for and sales, and that they receive some medicines for free. washing-station staff, and sometimes to provide other bene- Some farmers also report that they benefit from access to fits to members—short-term microloans and improvements coffee bicycles to transport coffee cherries. Less directly, to local schools, for example. farmers stated that they benefit from socializing with and learning from others. Some farmers felt their work was Government revenue is increasing now easier (they no longer process cherries at home) and that they could spend time on things other than coffee The Rwandan government reports that coffee receipts production. increased by an average of 30 percent a year between 2002 A 2006 analysis for the U.S. Agency for International and 2006 (figure 11.3). In February 2010 the head of Development (USAID) reports that “approximately 50,000 Rwanda’s Coffee Development Agency said that he expected households have seen their incomes from coffee production 26,000 tons of coffee to be produced during the year (up double, and some 2,000 jobs have been created at coffee from 16,000 tons in 2009) and that the value of coffee washing stations” (USAID/Chemonics 2006). An NGO exports would increase significantly, to $69 million, from involved in the USAID project (ACDI/VOCA n.d.) $38 million in 2009.9 reports that “incomes (in the specialty coffee sector) have Despite the increases in production and export value, the doubled or tripled, and business skills, labour conditions amount of fully washed coffee being produced in Rwanda is and community spirit have been enhanced.” With addi- still below targets. It now accounts for 20 percent of the tional income, farmers can repair their homes, buy clothes, annual crop, compared with 1 percent in 2002. A greater pay school fees for their children, and get through the long concern is that washing stations and cooperatives need to 190 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS Figure 11.3 Average Farmer and Export Prices, 2003–08 from Europe, China, and Japan also routinely visit Rwanda, bringing income to farmers and expertise that helps $3.1 improve the local industry. The benefits from specialty cof- $3.0 fee extend beyond the cooperative. Another observer notes $2.4 that “[a]s income levels of the cooperative members have $/kg green coffee $2.1 $2.0 $1.9 increased so has the flow of money in the community . . . . $1.6 $1.8 The positive feelings among community members are a reflection of increased incomes in the area (of the coopera- $1.0 tives)” (Goff 2006, 70). While global growth in ordinary-grade coffee consump- $0.0 tion remains modest, the consumption of high-quality spe- 2003 2004 2005 2006 2007 2008 cialty coffee (currently, 7 percent of the coffee volume in the Average export price Cherry price premium international market) has been rising by 15 percent annu- Cherry price OCIR cafe ally in recent years. Even with increased competition and Source: Rwanda Ministry of Agriculture and Animal Husbandry and Min- somewhat lower prices, Rwanda’s specialty coffees should istry of Trade and Industry (2008). continue to command a good price (Economist Intelligence Unit 2007).11 Increased entrepreneurship is leading do more improve technical capacities and operate prof- to more jobs in the coffee sector itably, creating incentives for more farmers to choose to sell cherries for washing rather than processing cherries them- As of 2006, 4,000 jobs had been created at coffee-washing selves at home. stations (USAID/Chemonics 2006). Though many of these Identifying precisely how much coffee is being produced are part-time jobs during the harvest season, others are in Rwanda is difficult. Government figures from different full-time positions managing stations and cooperatives. sources do not necessarily coincide, and figures from the Co-op and washing-station employees are learning valuable U.S. Department of Agriculture and the Food and Agricul- business skills: accounting, marketing, and negotiating. ture Organization (FAO) do not match figures produced by An estimated 100 Rwandans have learned to cup coffee.12 the Rwandan government. The government’s trade statistics At milling operations and at exporters, other Rwandans sort put the volume of coffee traded in 2008 at 18,185 tons, beans, operate milling equipment, and prepare beans for although a more recent news story puts that figure at 22,000 shipment. And as the new coffee culture grows in Rwanda, tons.10 The news story indicates that output in 2009 was jobs are starting to be created in retail outlets such as the approximately 24,000 tons and that the Rwandan govern- popular Bourbon Café in Kigali. ment projects that output would rise 13 percent in 2010, Although Rwanda has made good progress in the spe- to 27,000 tons. Estimates for fully washed coffee are more cialty coffee sector, stakeholders recognize that more must solid: in 2002 Rwanda produced 48 tons of fully washed be done to consolidate these gains and to direct attention coffee, a figure that grew exponentially in by 2006, to to the most pressing problems in the sector. To refocus 3,000 tons (Rwanda Ministry of Finance and Economic attention, the Rwandan government issued a revised Planning 2007). National Coffee Strategy (NCSR) in 2009. The NCSR cre- Rwanda’s former Minister of Agriculture Anastase ated a modified strategy for the coffee sector, one that Murekezi has indicated that the specialty coffee industry’s builds on lessons learned implementing the 2002 strategy. most successful trading partners to date, in terms of creat- It sets a new, lower production target of 33,000 tons of ing wealth for Rwandans, were American importers. Amer- coffee by 2012, 19,000 tons of which is to be fully washed. ican companies such as Starbucks and Green Mountain The government anticipates revenues of $115 million if Coffee buy from Rwanda’s producers, bringing much- this much coffee is produced (Rwanda Ministry of Agri- needed income to smallholder farmers. Other less-well-known culture and Animal Husbandry and Ministry of Trade and but highly discriminating American importers such as Industry 2008). Intelligentsia, Thanksgiving, and Counter Culture Coffee The new policy identifies five target projects: improving also routinely purchase Rwandan coffee. Coffee buyers farming techniques, providing support to make washing CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 191 stations more profitable, helping private exporters improve reduce levels of prejudice and hostility. Therefore, positive marketing and sales, conducting a census of all coffee-growing contact is considered one of the most effective strategies for areas, and partnering to do toll roasting in China and the reducing intergroup conflict (Pettigrew 1998; Dovidio, Middle East.13 In an effort to add more value locally, the Gaertner, and Kawakami 2003). When contact between Rwanda Coffee Development Authority (RCDA) has part- groups in postconflict societies is intense and deep, meaning nered with the Hunter Foundation to build a factory to roast when it involves interactions at social events, helping others, and package coffee locally. The hope is that this effort will or celebrating together, it can promote reconciliation and produce more than 100 full-time jobs and another 2,000 help prevent a renewal of violence (Gibson 2004; Straub indirect jobs.14 While coffee production cannot, by itself, 2006). If formerly antagonistic groups find ways to cooper- solve the many problems faced by the poor in Rwanda, ate, they may develop a new shared identity that helps changes in the industry are helping coffee farmers and other reduce prejudice and creates a sense that a more collabora- workers in the industry better cope with poverty, and these tive future is possible (Gaertner et al. 1990). changes also seem to be generating positive social benefits. As noted, farmers come to coffee-washing stations either to sell beans or to do seasonal work. Seasonal employees, working side by side, help at drying tables. During the har- INDIRECT AND SOCIAL BENEFITS vest season, farmers may bring cherries to a washing station OF LIBERALIZATION several times a week. This means that farmers and seasonal The liberalization of Rwanda’s coffee sector is distinguish- workers are at the washing stations repeatedly from March able from other liberalization efforts, such as privatizations or April through June or July. This organizational structure in Eastern Europe and land titling reforms in Kenya, that is relatively new in Rwanda (in 2000, there were only two have tended to benefit elites. Because coffee sector liberal- coffee-washing stations in all of Rwanda, and as of 2009 ization in Rwanda has helped to raise income for the rural there were more than 100), and farmers may well be experi- poor—rather than imposing costs on them—it may be less encing a new type of contact. likely to promote conflict than have liberalizations in which costs are spread widely (such as when subsidies are Perceptions of economic satisfaction removed) and benefits are narrowly concentrated (Collier et al. 2003).15 The Tobias and Boudreaux survey explored the possible connection between perceptions of economic improve- ments in the lives of farmers and coffee-washing station Coffee facilities are a locus of cooperation workers and of improved interpersonal relations, compar- Cooperatives and private washing stations may also serve as ing current levels of economic satisfaction with levels of sat- vehicles for increasing cooperation among members. Build- isfaction in the past. Economic satisfaction was measured on ing on reports in the media of informal reconciliation at a four-item scale, with low scores indicating high levels of sat- coffee cooperatives, Tobias and Boudreaux (2011) con- isfaction. A high “economic satisfaction change” score indi- ducted an exploratory survey of 10 coffee-washing stations cated an improvement in economic satisfaction in recent in 2008 to investigate possible social benefits associated with years. Only 3 percent of participants indicated that they the liberalization of the sector.16 A total of 239 completed were very satisfied with their economic situation five years surveys were obtained from a subsection of rural Rwandans earlier, whereas 40 percent reported that they are very satis- associated with these stations—some were farmers with sea- fied with their current economic situation. Forty-five sonal jobs at the washing stations and others were people percent of all participants reported a one-point improve- who were at the stations to sell cherries. Statistical analysis ment (on a four-item scale) in economic satisfaction in of the surveys showed significant correlation between eco- recent years, 22 percent reported a two-point increase, nomic satisfaction or general perceptions of life satisfaction 10 percent reported a three-point increase in economic and more positive attitudes to reconciliation. This finding satisfaction, 15 percent experienced no change in economic was particularly true the older a washing station was and the satisfaction, and as noted, 3 percent of survey participants longer it had been a part of the local community. were more satisfied economically in the past. These respon- The survey drew on insights from inter-group contact dents indicated a decrease of one or two points. theory (Allport 1954). Extensive evidence indicates that Life satisfaction ratings today and in the past were also positive interactions between antagonistic groups can measured. Eighty percent of participants reported a positive 192 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS life satisfaction change; for 10 percent, life satisfaction with the creation of more than 100 washing stations since remained unchanged compared with recent years. Only 2000 is substantial. 7 percent indicated less life satisfaction today than in the Taken together, the study’s findings suggest that the past. These figures, like the figures for economic satisfac- enhanced entrepreneurial activities in this particular sector tion, indicate that the overwhelming majority of the sample of Rwanda’s economy not only produce positive economic experienced positive life satisfaction gains in recent years. change among individuals touched by this institutional To explore the impact of increased happiness among change, but that they may also be triggering a chain of farmers and washing-station workers on other people, the mediating effects linked to positive social change among survey explored possible links between perceptions of eco- people working at or with coffee-washing stations. The nomic and life satisfaction and willingness to engage in con- observed effects were not dependent on ethnicity or on the tact with people of other ethnic groups. The findings indi- particular ethnic mix of participants in a given location. cate that people who are less willing to engage socially with This finding suggests that forgiveness and increased levels of members of another ethnic group have a greater degree of trust may be experienced broadly in this environment. Par- “ethnic distance” than do people who engage with others ticipants felt less ethnic distance than in the past and are willingly and more frequently. Measures of ethnic distance now engaging in deeper social contact. It is possible that the were determined in the following manner: an “ethnic dis- collaboration initiated by the liberalization of the coffee sec- tance today” score was obtained by counting each of five tor, while difficult to quantify, is one of the most important possible interaction types from a classic social distance scale benefits of the government’s coffee sector policy reform. (high scores indicate low ethnic distance), and the “ethnic distance change” score was obtained by calculating the dif- CHALLENGES AND CONCERNS ference between a positive answer today versus a statement IN THE COFFEE SECTOR that in the past the participant would not have engaged in these interactions. A high numeric score for “ethnic distance Over the past decade, the Rwandan government has change” was interpreted as a signal of less ethnic distance worked together with donor organizations, NGOs, the pri- today than previously. In general, participants reported high vate sector (foreign and domestic), and local producers in a degrees of reduction in ethnic distance as well as very fre- sustained and concerted manner to help this important quent social and work-related contact. sector of the economy become more competitive and, in Survey results show that meaningful contacts with mem- turn, grow and expand. Political champions for these bers of another ethnic group are significantly correlated efforts existed at the highest levels of government. Donors with low distrust and conditional forgiveness. Participants not only supported financing and credit programs, they who expressed satisfaction with their economic and overall also facilitated programs and projects that brought individ- life situation had significantly correlated responses in terms uals with a variety of technical expertise to Rwanda to share of positive attitudes toward reconciliation. In particular, experiences and knowledge in areas of agronomy, the participants with greater economic security also reported development of market links, coffee production, and cup- low ethnic distance, low distrust toward the other group, and ping. Coffee cooperatives received support to improve their a tendency toward conditional forgiveness. Life satisfaction production efforts and to become competitive producers. significantly correlated with economic security variables, Other donor projects supported the entry of Rwandan and participants reporting greater satisfaction with life also entrepreneurs into the market. expected a more positive, peaceful future in Rwanda. These joint efforts required commitments of time, Results of the survey also show that responses of partici- resources, and talent to address weaknesses in this supply pants at coffee-washing stations that had been in operation chain. This iterative process lasted for years. The manner in for longer periods of time are significantly correlated with a which the Rwandan government was able to create an effec- reduction in ethnic distance over time. It is reasonable to tive alliance of partners to transform Rwanda’s coffee sector, assume that positive social change in the coffee sector takes through these iterative efforts, from a low-volume, low- time, and the survey data support this reasoning. All of the quality market to a high-quality market provides a useful washing stations in the study had been in operation for less example of effective policy leadership. than seven years, and most of them were created fewer than Although the specialty coffee industry is having positive five years ago. If the observed pattern were to continue, effects in Rwanda, the smallholder farmers and the entre- however, the potential for positive social change associated preneurs who work in the sector face several challenges. CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 193 One of these is implementing strategies that create price identified, suggesting that additional research is needed to incentives for farmers to improve both quality and washing- clarify this important issue. station management, thereby increasing production effec- The National Coffee Strategy Rwanda report also notes tiveness. At the cooperative level, management needs to be that too few farmers are familiar with and able to imple- improved so that costs are better controlled and members ment good farming practices. Farmers need good-quality continue to benefit. The government also needs to avoid dis- inputs, such as seedlings and fertilizers, to increase their locations to poor Rwandans – particularly women – that crop yields and quality, but they also need technical knowl- may result from the 2005 Land Law. Finally, transportation edge about how to handle seedlings, how to deal with pests costs remain high in Rwanda. and disease, and how to apply fertilizers. Although RCDA/OCIR-Café supplies farmers with fertilizer and employs agronomists, there are only a small number of Production capacity and pricing problems these professionals, and they have a difficult time reaching One continuing concern in the coffee sector is that farmers all the farmers who need assistance.17 The National Coffee are not producing a sufficiently large volume of high- Strategy Rwanda also cites a lack of coordination among quality coffee to meet continuing demand. A number of agencies that further limits the effectiveness of these factors may be contributing to this problem. First, prices extension services (Ministry of Agriculture and Animal paid to farmers do not provide sufficient incentives for Husbandry and Ministry of Trade and Industry 2008). them to focus on quality. Farmers continue to produce a Management and profitability in Rwanda’s coffee- much greater amount of ordinary-quality coffee than fully washing stations is a growing concern. Stations are operat- washed coffee. Second, farmers may lack knowledge of how ing at partial capacity, producing smaller amounts of fully best to manage their trees. Third, despite the growing num- washed coffee than anticipated. The government reports ber of washing stations, many farmers still have limited that operating costs associated with the stations (labor, access to them and so process cherries at home. transportation, and electricity and water) are high. In As noted above, OCIR sets weekly minimum prices for addition to these challenges, washing-station personnel the purchase of cherries at washing stations. In some areas, often lack technical skills, particularly in finance, account- washing stations pay farmers amounts that fluctuate ing, and management. These capacity problems can trans- modestly to reflect changes in supply. Washing stations do late into difficulties securing financing and working with not, however, seem to pay farmers a premium based on the lenders, which in turn can translate into difficulties paying quality of cherries they deliver. Farmers thus concentrate on farmers in a timely fashion. The government recognizes quantity produced versus quality. the potential benefits of increasing coffee yields at washing If there is continuing unmet demand for Rwanda’s stations, particularly for fully washed coffee. Farmers who fully washed specialty coffee, washing-station operators sell coffee as a cash crop are projected to see their incomes should respond by demanding, and paying a premium for, grow 4 percent a year more than that of other farmers higher-quality cherries. One possibility why this is not (Ministry of Agriculture and Animal Husbandry and happening may be that washing-station personnel are Ministry of Trade and Industry 2008). unable to identify higher-quality beans. If washing sta- tions cannot distinguish a low-quality from a high-quality Concerns with cooperative management cherry, they will choose to pay a low rather than a high price. Additionally, high labor and input costs also make Smallholder farmers who voluntarily join cooperatives it difficult for washing stations to operate at full capacity face the difficult task of creating a culture of entrepre- and as a result they generate limited income. Therefore, neurship within the cooperatives so that they become washing stations may lack sufficient working capital to more “business minded.” A key problem identified by pay farmers for higher-quality beans. Competition among local NGO SPREAD is the need to attract and retain more washing stations could resolve this problem. At the professional managers in cooperatives and, at the same moment, however, it is unclear how best to shift this pat- time, to reduce the influence of volunteer boards of direc- tern if there are no other legal impediments to paying tors. An assessment of a group of Rwandan cooperatives higher prices for higher-quality beans. Although the states that “[a] professional, entrepreneurial [g]eneral problem of pricing is identified in the National Coffee [m]anager is the most important individual to a cooper- Strategy Rwanda report, the causes of resistance are not ative’s ultimate success” (Swanson and Bagaza 2008). 194 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS However, the report’s authors find that no cooperatives in state of the cooperative, and about likely prices for cherries the group under investigation had such a manager. The rea- and benefits of membership, the possibility of corruption son seems to be that the boards of directors are reluctant to and conflict over resources rises. These are especially impor- pay high enough salaries to attract a professional manager. tant issues to resolve because cooperatives face increasing Further, boards often prefer to have to have a local repre- competition from other coffee entrepreneurs. sentative, rather than an “outsider,” fill the general manager Cooperatives have been an important asset for role. Local representatives, however, are less likely to have Rwanda’s smallholder farmers, allowing them to earn the skill set needed to manage the cooperative effectively. more money from coffee, develop additional skills, and Cooperatives are capable of producing very high-quality work cooperatively with others in ways that may promote coffee, but they are experiencing real difficulties creating reconciliation. But they must address serious shortcom- effective management structures. ings in management practices and capabilities if they hope Rwanda’s coffee producers, with support from donors, to continue playing this role in the future. Because coop- NGOs, and the government, have done an impressive job of eratives seem to provide a space for interaction and even generating interest in their products. Rwandan coffee is reg- informal reconciliation, further support efforts to help ularly available at retailers such as Starbucks and Whole accomplish the goal of creating transparent and account- Foods across the United States and Marks & Spencer in the able management may well be justified. United Kingdom. But this good progress may stall if prob- lems related to marketing and processing are not adequately Concerns with Rwanda’s land law addressed. Buyers have expressed concerns that contract terms for quality, quantity, and timely delivery are not being A separate challenge in Rwanda involves use and control of met (Rwanda Ministry of Agriculture and Animal Hus- land. Land is an extremely scarce and highly contested bandry and Ministry of Trade and Industry 2008). Some resource. More than three-fifths of families working as buyers have reported long delays receiving shipments, and farmers cultivate less than 0.7 hectares, and more than a many have complained that the shipped product did not quarter cultivate less than 0.25 hectares (Rwanda Ministry match samples. These kinds of concerns, if not addressed, of Finance and Economic Planning 2007). According to one will lead to loss of business for Rwanda. study, “land was a factor behind social tensions before every A related problem is that boards of directors in Rwanda major open conflict. Even today, more than 80 percent of all often interfere inappropriately in the daily management disputes in Rwanda are related to land” (Food and Agriculture of the cooperative. The assessment notes that “[board Organization and Rwanda Ministry of Lands, Environment members], particularly [p]residents, do not want to relin- Forestry, Water and Mines 2006, 7; see also Van Hoyweghen quish their authority to a strong [g]eneral [m]anager” 1999). For much of its history, Rwanda’s rulers have owned (Rwanda Ministry of Agriculture and Animal Husbandry most of the country’s land. With control of land in the and Ministry of Trade and Industry 2008, 13). While it is hands of government, formal land markets did not develop. essential that boards take seriously their fiduciary duties to Rather, transfers often took place informally, and confusion create general policies and oversee management activities, and insecurity were common. Local officials had great dis- the assessment recommends that they give managers cretion over land allocation and favored politically power- increased decision-making authority and discretion. ful individuals over marginalized people who may have held Managers of cooperatives and boards of directors need to traditional land use rights.18 communicate more effectively with members so that mem- In 2003 the Rwandan parliament approved a land reform bers understand the ownership structure of the cooperative decree that provides for individualized rights to property. as well as their rights within the structure. To date, this has This policy was followed, in 2005, by passage of the Organic not been done effectively. Members report that they are Land Law. The law is implemented by a series of decrees, unclear who “owns” the cooperative (Rwanda Ministry of many of which are just now going into effect. The Land Law Agriculture and Animal Husbandry and Ministry of Trade abolishes all customary forms of tenure, and in their place, and Industry 2008). In addition, cooperatives need to establishes government-issued titles for 99-year leases. Rural develop and communicate effective business plans and to land will be registered locally, and urban, commercial prop- improve financial record keeping and documentation. erty will be registered in a national cadastre in Kigali. The There is a clear need for increased capacity in these areas. government maintains a role in the resettling of people and When members lack clear information about the financial in devising land use and land planning policy. CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 195 The government “sees increased security of tenure or Security and clarity of tenure rights, whether customary rights of address to land, and more effective land manage- or leasehold, are essential both to avoid future conflicts and ment, as important factors for the improvement of the agri- to encourage increased investment in agriculture. However, cultural sector and the economy as a whole, helping to create the land law raises serious concerns, especially for women the resources needed to reduce poverty and to consolidate and for uneducated farmers who might be dispossessed of peace and social cohesion” (Pottier 2005, 511). Although their land. Surely, these risks are undesirable in a nation the government says that it wants to increase tenure secu- with such high levels of poverty and such strong depen- rity, the new law is likely to create a host of problems. dence on agriculture as a livelihood. The government hopes the land law will promote con- solidation of land parcels into larger units. By allowing sales Concerns with transportation infrastructure of property and increasing freedom within the land market, small parcels could be sold to commercial farmers who will Finally, as a landlocked country with limited paved roads in consolidate the land and create viable agribusinesses. How- rural areas where most coffee is grown, transport costs in ever, one observer notes: Rwanda are high. Diop, Brenton, and Asarkaya (2005) argue that Rwanda’s smallholder subsistence farmers are discon- Land fragmentation in Rwanda serv[es] as a coping nected from markets as a result of “extremely high” transport mechanism in smallholder agriculture, the typical costs. The authors estimate that transport costs from farm Rwandan household farms an average of five plots. gates to the export port of Mombasa were 80 percent of the Some are in the valleys, others are upland and some producer price.20 Transport within Rwanda itself was esti- near the household. In some parts of southern mated at 40 percent of the producer price. If transport costs Rwanda, a household may have up to 14 crops grow- were reduced, through the development of better rural infra- ing in different fragments at different seasons. . . . the structure and, in particular, more effective rural transport costs of consolidation in Rwanda may not exceed the routes, access to markets would improve and poverty levels benefits of using land fragmented over the years in would likely be reduced. The authors find that a 50 percent adopting to land scarcity (Musahara 2006, 11). reduction in the transport costs in rural areas would lead to a 20 percent increase in producer prices for coffee, which in turn While a more open land market is desirable, the law would reduce poverty levels among coffee farmers by more (Organic Law No. 08.2005) contains provisions that allow than 6 percent. Given the continued emphasis on coffee pro- the government to interfere in the market. For example, it duction as a strategy to alleviate rural poverty, improving the allows the government to bar people who own less than one rural transport system will be an important way to connect hectare from registering their property. More troubling is farmers to markets and to increase their household income. the provision that “subsistence farmers can have their land confiscated should they fail to exploit it diligently and effi- CONCLUSION ciently” (Pottier 2005, 521, [Articles 62–65]). Though the government is supposed to provide compensation for such Despite good economic growth and real benefits of liberal- confiscations, it has not established clear standards for such izing the coffee sector, more remains to be done to move payments. Rwanda toward its Vision 2020 goal of becoming a stable, The new law may pose a special problem for women middle-income country. Most Rwandans remain very poor smallholders and their children. Under the law, the gov- and the rural areas of the country provide few off-farm ernment is supposed to register all parcels of land in the employment opportunities. Creating more employment country. Only legally married women and their children, opportunities for millions of Rwandan smallholder farmers however, (not women married under customary norms and remains a pressing challenge. their children, or poor women who do not formally marry Rwanda’s specialty coffee industry is helping to address because of the associated costs) can register and inherit some of these concerns. As a result of improvements in the land. Further, there is uncertainty in the law regarding sector, smallholder coffee farmers now keep more of the inheritance—specifically, whether women inherit through value of the product they grow. In turn, rising incomes for the inheritance law or through the land law.19 Also of con- tens of thousands of farmers make them better able to feed cern is the fact that local custom continues to bar women themselves and their families, to send their children to from exercising their legal rights under the land law. school, to buy insurance, and to repair or improve their 196 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS homes. By freeing the coffee sector from the heavy-handed 5. Mutandwa et al. also note that “due to the low level of involvement of the government, the Rwandan government production, the milling factories operate under capacity and has shifted incentives in the coffee sector and created greater exporters tend to lower the reference price in order to cover scope for citizens to pursue entrepreneurial opportunities. their relatively high milling costs” (p. 212). Though additional research needs to be done, evidence 6. A conflicting statement, however, can be found in the suggests that joint efforts in commercial activities related to Economic Development and Poverty Reduction Strategy the specialty coffee sector are providing an alternative path 2008–2012, which states that the “the proportion of fully- washed coffee production will increase from 10 percent to to postwar reconciliation in Rwanda. Initial exploratory 100 percent . . .” (p. 38). survey data provide some supports for these claims (Tobias 7. See also Swanson and Bagaza (2008). and Boudreaux, 2011). By working together in cooperatives and at washing stations, Rwandans are experiencing both 8. Survey results are available from the author of this paper. social and economic benefits from liberalization. They feel 9. Reuters, “Rwanda Sees 20010 Coffee Crop Rising 63 Pct,” greater economic satisfaction, which is correlated to lower February 11, 2010. levels of distrust, higher levels of forgiveness, and reduced 10. “Coffee Output May Climb 13 Percent in 2010,” New ethnic distance. Times, January 8, 2010. Whether this kind of reconciliation is effective in the 11. The report notes, “Stagnation in many of the tradi- medium to long term in Rwanda remains to be seen. The tional coffee-drinking markets of North America and West- ern Europe will restrict growth in demand, although positive experience of workers in the coffee sector, however, demand for high-quality specialty coffees, including strongly suggests that in postconflict environments, govern- Rwanda’s finest fully-washed Arabica, will remain more ments should follow Rwanda’s lead and promote broad-based buoyant” ( p. 10). trade liberalization that encourages commercial interaction 12. To “cup” coffee refers to a tasting technique used to cre- between rival groups. ate a flavor profile for coffee. Cuppers evaluate coffee based Although real concerns surround the economic viability on fragrance, aroma, and taste, a process that helps identify of some coffee-washing stations and cooperatives, liberal- defects in coffee. Fewer than half of the 100 cuppers, how- ization of the Rwandan coffee sector has created a wider and ever, are trained and licensed to international standards. deeper space for positive entrepreneurship: a space being “Twenty-One Coffee Cuppers Trained on Quality,” New filled by thousands of Rwandans, from smallholder farmers Times, August 29, 2009. to local exporters. At the same time, adding value to the cof- 13. Toll roasting involves a contractual arrangement fee supply chain is producing both direct economic benefits whereby an experienced coffee roaster accepts green coffee and important indirect social benefits for individuals and from a client and roasts it to the client’s specifications. communities in Rwanda. 14. “Two Million Invested in Coffee Value Addition,” New Times, August 7, 2009. NOTES 15. Collier et al. (2003, 53) argue that “the key root cause of conflict is the failure of economic development. Countries 1. “Coffee ‘Cup of Excellence’ Slated for Next Year,” New with low, stagnant, and unequally distributed per capita Times, September 23, 2009. incomes that have remained dependent on primary com- 2. Traditionally, Rwandan farmers removed the fruit of modities for their exports face dangerously high risks of their coffee cherries using either a hand-pulper or rocks. prolonged conflict.” Beans would then be dried and fermented in buckets, for 16. Surveys were conducted at five cooperatives and five varying lengths of time, in water of varying quality. As a privately owned coffee-washing stations. Surveys were result, coffee was of lower, industrial quality. This home- administered in Kinyarwanda by students from the National processed coffee still makes up the majority of coffee being University of Rwanda over a two-week period. Surveys were sold from Rwanda. previously field tested in February 2008. 3. Prices were set by OCIR until 1994 and remained con- 17. Farmers receive fertilizer from RCDA, and the agency, in stant for the entire harvest season; price differences based turn, imposes a fertilizer fee on exporters. Exporters then on quality were not permitted. See Mutandwa et al. 2009 choose either to pass these costs along to cooperatives and and deLucco 2006 for more information. washing stations in the form of lower prices paid per kilo or 4. Under the Habyarimana regime, the powerful OCIR- they absorb the costs themselves. Café agency was run by relatives of the dictator’s wife, mem- 18. Sales were restricted according the size of the buyer and bers of the clan de Madame (Verwimp 2003). seller’s total land holdings; see Musahara (2006). CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 197 19. Dyer (2007) notes that although Rwanda’s 1999 Inheri- Economist Intelligence Unit. 2007. “Country Report: tance Law gives women equal rights to men this has “yet to Rwanda” (February). make any difference in practice.” FAO (Food and Agriculture Organization) and Rwanda 20. Note these prices should now be lower as a result of Ministry of Lands, Environment Forestry, Water and Rwanda’s accession to the East African Community. Mines. 2006. “A Case Study on the Implications of the Ongoing Land Reform on Sustainable Rural Develop- ment and Poverty Reduction in Rwanda and the Out- come Report of the Thematic Dialogue Held on the 20th REFERENCES January 2006.” Presented at the International Conference ACDI/VOCA. n.d. “Specialty Coffee: Increased Quality & on Agrarian Reform and Rural Development (ICAARD), Profits for Smallholders.” http://www.acdivoca.org/ Porto Alegre, Brazil, March 7–10. http://www.icarrd.org/ 852571DC00681414/Lookup/coffeebroweblayout/$file/ en/icard_doc_down/case_Rwanda.pdf. coffeebroweblayout.pdf. Gaertner, Samuel L., Jeffrey A. Mann, John L. Dovidio, Allport, Gordon W. 1954. The Nature of Prejudice. Oxford, Audrey J. Murrell, and Marina Pomare. 1990. “How Does UK: Addison Wesley. Cooperation Reduce Intergroup Bias?” Journal of Person- Bates, Robert H. 1981. Markets and States in Tropical Africa: ality and Social Psychology 59 (4): 692–704. The Political Basis of Agricultural Policies. Berkeley, CA: Gibson, James L. 2004. “Does Truth Lead to Reconciliation? University of California Press. Testing the Causal Assumptions of the South African Berlage, Lode, Bart Capéau, and Philip Verwimp. 2004. Truth and Reconciliation Commission.” American Jour- “Dictatorship in a Single Export Crop Economy.” House- nal of Political Science 48 (2): 201–17. holds in Conflict Network. http://www.hicn.org/papers/ Goff, Samuel Neal. 2006. “A Case Study of the Management dicta.pdf. of Cooperatives in Rwanda.” Unpublished master’s the- Coffeeresearch.org. “Coffee Cupping.” http://www.coffeere- sis, Texas A&M University. search.org/coffee/cupping.htm. IFAD (International Fund for Agricultural Development). Collier, Paul, V. L. Elliott, Håvard Hegre, Anke Hoeffler, 2006. “Enabling the Rural Poor to Overcome Poverty Marta Reynal-Querol, and Nicholas Sambanis. 2003. in Rwanda.” http://www.ifad.org/operations/projects/ Breaking the Conflict Trap: Civil War and Development regions/Pf/factsheets/rwanda_e.pdf. Policy. Oxford, UK: Oxford University Press and World Index Mundi. 2010. “Rwanda GDP – Real Growth Rate.” Bank. http://www.indexmundi.com/rwanda/gdp_real_growth_ deLucco, Paul. 2006. “Raising the Bar: Producing Quality rate.html. Coffee in Rwanda.” World Report Spring 2006: Specialty Murekezi, Abdoul, and Scott Loveridge. 2009. “Have Coffee Coffee; Improved Market Linkages & Increased Profits. Reforms and Coffee Supply Chains Affected Farmers’ http://www.acdivoca.org/852571DC00681414/Lookup/ Income? The Case of Coffee Growers in Rwanda.” Paper WRSpring06-Page13-14-RaisingtheBar/$file/WRS presented at the Agricultural & Applied Economics pring06-Page13-14-RaisingtheBar.pdf. Association’s 2009 Joint Annual Meeting, Milwaukee, DFID (UK Department for International Development). WI, July 26–28. 2008. “Good Taxes Reduce Poverty in Rwanda.” January 10. Musahara, Herman. 2006. “Improving Tenure Security for Diop, Ndiame, Paul Brenton, and Yakup Asarkaya. 2005. the Rural Poor: Rwanda—Country Case Study.” “Trade Costs, Export Development, and Poverty in FAO/LEP Working Paper 7. Rwanda.” Policy Research Working Paper 3784, World Mutandwa, Edward, Nathan Taremwa Kanuma, Emmanuel Bank, Washington, DC. Rusatira, Theopile Kwiringirimana, Patrice Mugenzi, Dorsey, Learthen. 1983. The Rwandan Colonial Economy: Ignatius Govere, and Richard Foti. 2009. “Analysis of 1916–1941. Unpublished PhD dissertation, Michigan Coffee Export Marketing in Rwanda: Application of the State University. Boston Consulting Group Matrix.” African Journal of Dovidio, John F., Samuel L. Gaertner, and K. Kawakami. Business Management 3 (5): 210–19. 2003. “Intergroup Contact: The Past, the Present and Personal interview with Anastase Murekezi, Rwandan the Future.” Group Processes and Intergroup Relations Minister of Agriculture. 2006 (March). Kigali, Rwanda. 6: 5–21. Personal interviews with members of the COOPAC cooper- Dyer, Rodney. 2007. “Land Issues in Rwanda: Drawn from ative. 2006 (March). Gisenyi, Rwanda. DFID Technical Assistance for Land Tenure Reform, Pettigrew, T. F. 1998. “Intergroup Contact: Theory, Research 2005–2008.” http://www.oxfam.org.uk/search?q=dyer& and New Perspectives.” Annual Review of Psychology 49: x=17&y=19. 65–85. 198 CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS Pottier, Johan. 2005. “Land Reform for Peace? Rwanda’s Straub, Edwin. 2006. “Reconciliation after Genocide, Mass 2005 Land Law in Context.” Journal of Agrarian Change Killing or Intractable Conflict: Understanding the Roots 6(4): 509–37. of Violence, Psychological Recovery and Steps toward a Prunier, Gerard. 1995. The Rwanda Crisis: History of a General Theory.” Political Psychology 27 (6): 867–95. Genocide. New York: Columbia University Press. Swanson, Richard, and Tom Bagaza. 2008. “SPREAD-Growers Republic of Rwanda. Organic Law No. 08.2005: Determin- First Coffee Cooperative Assessment and 2008 Coopera- ing the Use and Management of Land in Rwanda, Article tive Development Work Plan.” Report prepared for the 20. Kigali. July 14. Texas A&M University System, Borlaug Institute of Inter- Rwanda Development Gateway. 2007. “Coffee Sells at national Agriculture and Growers First. Record Prices.” http://www.rwandagateway.org/article Tobias, Jutta, and Karol Boudreaux. 2011. “Entrepreneur- .php3?id_article=6848. ship and Conflict Reducation in the Post-genocide Rwanda Ministry of Agriculture and Animal Husbandry Rwanda Coffee Industry.” Journal of Small Business and and Ministry of Trade and Industry. 2008. “National Cof- Entrepreneurship 24 (2): 217–242. fee Strategy Rwanda 2009–2012.”http://amis.minagri USAID (U.S. Agency for International Development). 2009. .gov.rw/sites/default/files/user/National_Coffee_ “Rwanda.” http://www.usaid.gov/rw/our_work/programs/ Strategy_Rwanda_2009-2012.pdf. docs/factsheets/coffee.pdf. Rwanda Ministry of Finance and Economic Planning. 2000. USAID/Chemonics. 2006. “Assessing USAID’s Investments “Rwanda Vision 2020.” http://www.gesci.org/assets/files/ in Rwanda’s Coffee Sector: Best Practices and Lesson Rwanda_Vision_2020.pdf. Learned to Consolidate Results and Expand Impact.” ———. 2007. “Economic Development and Poverty Van Hoyweghen, Saskia. 1999. “The Urgency of Land and Reduction Strategy 2008-2012.” http://planipolis Agrarian Reform in Rwanda.” African Affairs 98: 353–72. .iiep.unesco.org/upload/Rwanda/Rwanda_EDPRS_ Verwimp, Philip. 2003. “The Political Economy of Coffee.” 2008-2012.pdf. European Journal of Political Economy 19 (2): 161–81. SPREAD. 2007. Annual Report. http://www.spread.org.rw/ World Bank. 2010. “Country Data: Rwanda.” http://data spread_project.php. .worldbank.org/country/rwanda. CHAPTER 11: ECONOMIC LIBERALIZATION IN RWANDA’S COFFEE SECTOR: A BETTER BREW FOR SUCCESS 199 CHAPTER 12 Cocoa in Ghana: Shaping the Success of an Economy Shashi Kolavalli and Marcella Vigneri o other country comes to mind more than Ghana N depending on global prices, marketing costs, explicit taxes when one speaks of cocoa. Likewise, one cannot on the sector, and macroeconomic conditions such as infla- think of Ghana without thinking of its cocoa tion and overvaluation of exchange rates and inelasticity of sector, which offers livelihoods for over 700,000 farmers in cocoa supplies. Regardless of the level of extraction, the the southern tropical belt of the country. Long one of need for sound macroeconomic management, of inflation Ghana’s main exports, cocoa has been central to the coun- and exchange rates in particular, becomes evident for con- try’s debates on development, reforms, and poverty allevia- tinuing to offer incentives for production. The other is the tion strategies since independence in 1957. The cocoa sector need for Ghana’s cocoa pricing policy to arrive at a market- in Ghana has not been an unmitigated success, however. ing arrangement that does not kill the goose that lays the After emerging as one of the world’s leading producers of golden eggs. Ghana appears to have achieved such as cocoa, Ghana experienced a major decline in production in arrangement without fully liberalizing the sector as other the 1960s and 1970s, and the sector nearly collapsed in the producers in West Africa have. early 1980s. Production steadily recovered in the mid-1980s after the introduction of economywide reforms, and the OBSERVABLE ACHIEVEMENTS 1990s marked the beginning of a revival, with production IN THE COCOA SECTOR nearly doubling between 2001 and 2003. These ups and downs offer interesting lessons. Since the introduction of cocoa in Ghana in the late 19th Various administrations in Ghana, including the colonial century, the crop has undergone a series of major expansions one, have used cocoa as a source of public revenue, and in and contractions. Ruf and Siswoputranto (1995) suggest that so doing the Ghanaian experience offers a recurrent exam- cycles are intrinsic to cocoa production because cocoa is ple of a policy practice followed by many other African influenced by environmental factors such as availability of countries: taxing the country’s major export sector to forest land; ecological factors such as deforestation, out- finance public expenditure (Herbst 1993). Revenue extrac- breaks of disease, and geographic shifts in production; and tion by the state has had varying effects on production economic and social factors such as migration. Shashi Kolavalli is senior research fellow and leader, Ghana Strategy Support Program, International Food Policy Research Institute. Marcella Vigneri is a visiting scholar, Department of Economics, University of Oxford. 201 Emergence as a leading producer buying companies on the coast of West Africa that were pre- pared to trade the new crop (Hill 1963; Amanor 2010; and Four distinct phases can be identified in regard to cocoa Gunnarsson 1978). production in Ghana: introduction and exponential growth Three social classes: land-owning farmers, peasants, and (1888–1937); stagnation followed by a brief but rapid laborers emerged among cocoa producers as a second wave of growth following the country’s independence (1938–64); migrants from Akyem moved to the region. Without sufficient near collapse (1965–82); and recovery and expansion, start- money with which to buy land, these migrants sharecropped ing with the introduction of the Economic Recovery Pro- with earlier settlers under a system called abusa, in which gram (ERP) (1983 to present). Figure 12.1 shows long-term laborers were paid one-third of the sales price of the harvested trends in levels of production. cocoa. Simultaneously, there was a large influx of migrants from relatively distant Upper Volta (now Burkina Faso), Niger, Exponential growth (1888–1937). Cocoa was introduced and Mali, who were attracted by the generous remuneration in the southern region of the Gold Coast in the mid-19th that cocoa production offered in southern Ghana. century by commercial farmers from the Eastern region dis- The growing population of cocoa farmers reinvested its tricts of Akuapem and Krobo, who had moved west toward profits in cocoa production in the western end of Ghana’s the adjacent district of Akyem to purchase mostly unoccu- Forest Zone, rapidly shifting the production frontier into pied forest land from the local chiefs for cocoa cultivation the Ashanti and Brong Ahafo regions, and consolidating (Hill 1963). Ghana as the leading world producer between 1910 and The conditions that encouraged these farmers to migrate 1914. Facilitated by the rapid expansion of the road and rail and buy land for cocoa are well documented: a fall in the network which began in 1920 and the organization of cocoa world price of palm oil after 1885, which pushed farmers to marketing by Ghanaian middlemen, cocoa earnings search for alternative export crops; a boom in rubber accounted for 84 percent of the country’s total exports by exports in 1890, which provided the capital for the purchase 1927. By the mid-1930s, production reached 300,000 tons. of new land; increasing population pressure in the Akuapem area, which encouraged commercial farmers to go further afield in search of alternative export agriculture Stagnation and growth postindependence (1938–early opportunities; and the establishment of European produce- 1964). The interwar period marked a slowdown in cocoa Figure 12.1 Ghana’s Cocoa Production, 1900–2008 800 700 600 Production, thousands of tons 500 400 300 200 100 0 19 0 19 3 06 19 9 12 19 5 19 8 19 1 24 19 7 30 19 3 19 6 39 19 2 19 5 48 19 1 54 19 7 19 0 63 19 6 69 19 2 19 5 78 19 1 84 19 7 90 19 3 96 20 9 02 20 5 08 0 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 0 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 Source: Gill & Duffus Group, various issues; Ghana Cocoa Marketing Board, various issues. 202 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY production, caused by decreasing demand and growing dif- the Second Development Plan at a time when the govern- ficulties in transport (Gunnarsson 1978). Outbreaks of ment was also receiving soft loans from the CMB. These pests and diseases (swollen shoot virus in particular) events made it obvious that by then the CMB had been reduced production in the Eastern region in the early 1940s, transformed into an instrument of public finance. The cap- pushing cocoa cultivation further into the western Brong turing of windfall profits from high cocoa prices had impor- Ahafo frontier (Amanor 2010). Production picked up again tant fiscal implications. Government expenditures grew during the second half of the 1940s but was now concen- dramatically over the 1950s: in real terms total consolidated trated in the Western region. In 1947, the colonial govern- public expenditures increased almost sixfold during this ment established the Cocoa Marketing Board (CMB) and period. The share of government expenditure in GDP grew gave it a monopoly over the purchase of beans. Until 1951 from 7 percent to 18 percent over the decade, and the share the bulk of profit made by the CMB went into its reserves, of extraordinary and development expenditure grew from which were then used for public investment (Brooks, Crop- 27 percent to 36 percent. In 1961, a cooperative society was penstedt, and Aggrey-Fynn 2007). In 1961 a cooperative given the monopoly right to purchase cocoa. From 1957 to society was given the monopoly right to purchase cocoa 1964 exports grew steadily, and production reached an replacing the network of private agents, brokers, traders, unprecedented level of 430,000 tons despite the significant and middlemen who until then had controlled internal decline in world prices between 1960 and 1962. marketing. In the early 1960s, when world prices plummeted, farm- As Beckam (1976) noted, the Convention People’s Party ers were required to save 10 percent of their earnings in (CPP), founded by Kwame Nkrumah, benefited from National Development Bonds, redeemable after 10 years. In extremely favorable postwar market conditions and accu- 1963 this scheme was replaced by a farmers’ income tax mulated cocoa income on a massive scale: following the charged at a flat rate equal to previous saving deductions. sharp increase in market prices in the 1950s, farmers were The government started to rely heavily on the CMB’s paid two to three times more than they received before the reserves, and the producer price was reduced from 224 to war, and between 1947 and 1965 the government collected 187 new cedi per ton between 1961 and 1964. With foreign almost one-third of the total value of cocoa export as export exchange reserves declining and the budget deficit rising duties. In 1950/51 the government increased export duties sharply, the government introduced a number of strong and began to take a much larger share of cocoa revenue by restrictive measures, an increase in taxes, foreign exchange means of a graduated ad valorem tax that increased with the controls, and comprehensive import licensing. The austerity increase of the average selling price per ton of cocoa. To of these measures lost Nkrumah much of his political con- extend its influence to the rural sector, in 1953 the Nkrumah sent, especially from cocoa farmers who had been aggra- regime also created the United Ghana Farmers’ Council vated by declining producer prices and by the conversion of (UGFCC), which was mainly concentrated in the cocoa- the compulsory saving scheme into an explicit export tax. growing regions despite its remit to cover the interest of In the second half of 1964 the world cocoa price collapsed farmers all over the country. The UGFCC was made the with a bumper crop in West Africa—Ghana alone reaching monopoly buyer of cocoa to create a platform for organizing an unprecedented production record of 538,000 tons. After the farmers behind the government and its administration. the purchasing and marketing costs of the CMB and UGFCC Following the second elections in 1954, the cocoa export were covered, virtually nothing was left for the government, tax was further increased while the producer price remained and the CMB’s liquidity resources were nearly exhausted. To at the same level for four years. This generated unrest and meet its expenses, the government started printing money, political agitation among cocoa farmers, ultimately forcing which ignited a 35 percent rise in inflation between October the government to increase the producer prices and to sta- 1964 and July 1965. In the face of such pressure, cocoa pro- bilize them during 1956–57 despite declining world cocoa ducer prices were reduced to their lowest levels in years. The prices. As a result, the share of government revenue in cocoa introduction of such highly restrictive measures represented sales dropped from 60 percent to 13 percent between a turning point in the fortunes of the Nkrumah government, 1954/55 and 1956/57. After its third political victory of which was overthrown in February 1966 and replaced by the 1957, the government increased its share of cocoa revenues National Liberation Council (NLC). by reducing producer prices to the 1954 levels. It also obtained a “voluntary contribution,” announced by the The downturn (1964–82). The collapse of world cocoa UGFCC on behalf of cocoa farmers, to share the burden of prices in 1965 triggered another downturn (Stryker 1990). CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 203 Real producer prices dropped consistently through the was introduced, and interest rates and cocoa producer 1960s because of inflation fueled by the government’s print- prices were raised. Cocoa production sunk to its lowest level ing of money to compensate for loss of revenue from cocoa ever in 1980–81; the world price at the official exchange rate and the introduction of an exchange rate policy that led to was lower than the producer price plus marketing costs. the heavy overvaluation of the cedi, the local currency. By The domestic conditions that led to the downturn in 1983, market exchange rates were nearly 44 times the official Ghana’s cocoa sector took place against an international rate. Between 1970s and early 1980s, it is estimated that as backdrop of increasing supply of cocoa from new producers much as 20 percent of Ghana’s cocoa harvest was smuggled such as Indonesia and Malaysia and expanded production into Côte d’Ivoire (Bulír 2002). Meanwhile, an aging tree in Côte d’Ivoire and Brazil. By the early 1970s Ghana had stock and the continued spread of disease made investment also lost much of its cheap labor supply from Burkina Faso in cocoa unattractive. Farmers in old cocoa production and Côte d’Ivoire, as migrant farmers, reluctant to work in areas, who found that sales prices barely covered their costs, the old cocoa-producing areas that had become less pro- increasingly turned from cocoa to food production ductive, were attracted to the neighbouring Ivorian (Amanor 2005). Ghana’s cocoa production dipped to a low regions, where policies granted migrants access to land at of 159,000 tons in 1982/83, a mere 17 percent of the total favorable terms. world volume, down from the 36 percent in 1964/65. The National Liberation Council dissolved the UGFCC The recovery and second expansion phase (1983–2008). and established the Producing Buying Company as a sub- The turnaround in Ghana’s cocoa sector began with the sidiary of the CMB. Producer prices were raised and farm- implementation of the ERP in 1983, which included a spe- ers were paid a bonus for top grade cocoa beans to upgrade cial program to revive the sector (the Cocoa Rehabilitation the quality of cocoa being exported. Shortly before the Project). Policy changes included increasing the farm gate Busia government came to power the cedi was devalued by prices paid to Ghanaian farmers relative to those paid in 43 percent and cocoa prices were raised by 30 percent. neighboring countries, thus minimizing the incentive to Cocoa production stagnated in the face of unchanged real smuggle, and devaluing the cedi, thus reducing the level of producer prices that remained at their 1950s levels. The implicit taxation of farmers. Busia administration took advantage of windfall profits As part of the Cocoa Rehabilitation Project, farmers from high cocoa prices in 1970 to enable a rapid expansion were also compensated for removing trees infected with of public expenditure. swollen shoot virus and planting new ones. This effort led In 1971 the Busia regime was replaced by the to substantial rehabilitation, with a large number of farms Acheampong-led National Redemption Council. Because of planting higher-yielding cocoa tree varieties developed by high world cocoa prices, this administration was initially the Cocoa Research Institute of Ghana. Production able to offer higher prices to farmers without cutting public rebounded to 400,000 tons by 1995/96 and productivity revenues, creating positive incentives to production. But a increased from 210 to 404 kilograms per hectare. Another progressively worsening balance of payments situation important reform took place in 1992, when Cocobod (as fueled inflation and undermined subsequent increases in CMB was renamed in 1984) shifted responsibility for real wages, producer prices, and other real incentives. domestic cocoa procurement to six privately licensed com- With the fall in world cocoa prices in the mid 70’s, the panies (commonly known as licensed buying companies general macroeconomic picture began to worsen: the gov- or LBCs) and reduced its staff by 90 percent between 1992 ernment budget deficit rose to 127 percent of total govern- and 1995. ment revenue and inflation accelerated to 116 percent. The Growth in cocoa production became more pronounced strong overvaluation of the cedi implied that little was left starting in 2001, possibly driven by a combination of of export revenues to divide between the government and record-high world prices, increased share being passed onto the farmers. Cocoa revenue went from 46 percent in 1974 farmers, and a set of interventions rolled out by the Coco- to 23 percent in 1979 and into negative figures between bod to improve farming practices: mass spraying programs 1980 and 1981 because of the exchange rate misalignment. and high-tech subsidy packages to promote the adoption of The rising costs of the CMB further reduced government higher and more frequent applications of fertilizer (Vigneri revenues. and Santos 2008). Some of the growth during this period In July 1978 the government underwent another regime may also have been due to the influx of cocoa smuggled change, and the cedi was devalued again, an austerity budget from Côte d’Ivoire. One study estimated this amount at 204 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY between 120,000 and 150,000 thousand tons in 2003/4 Table 12.1 Fertilizer Use in Cocoa-Producing (Brooks, Croppenstedt, and Aggrey-Fynn 2007). Regions: 1991/92–2003/04 Brong Technical change in the cocoa sector Crop year Ashanti Ahafo Western Total Number of farmers Since 2001 a significant share of Ghana’s agricultural produc- 1991/92 112 71 137 320 tivity gains have been generated by export crops, with cocoa 1997/98 132 54 227 413 accounting for 10 percent of total crop and livestock produc- 2001/02 108 94 226 428 2003/04 108 94 226 428 tion values (World Bank 2007a) and contributing to 28 per- Quantity of fertilizer used (50-kilogram bags) cent of agricultural growth in 2006, up from 19 percent in 1990/91 0.28 0.13 0.03 0.14 2001. At the same time, economic growth has been solid, Adoption rate (%) (13) (8) (6) (9) 1997/98 0.10 0.06 0.10 0.09 averaging more than 5 percent since 2001 and reaching 6 per- Adoption rate (%) (10) (13) (19) (15) cent in 2005–06. Coupled with the effects of greater access to 2001/02 0.35 0.17 0.74 0.52 education, health services, and land ownership (World Bank Adoption rate (%) (5) (7) (12) (9) 2003/04 4.17 4.39 6.10 5.24 2008), this rate of growth has contributed to the near halving Adoption rate (%) (57) (52) (41) (47) of the national poverty rate since the beginning of the 1990s, from 51.7 percent in 1991/92 to 28.5 percent in 2005/06 Source: Authors’ calculations from GLSS3, GLSS4, and Ghana Cocoa Farmers Survey, 2002 and 2004 rounds. (Breisinger et al. 2008). Over time, cocoa farmers have changed the way they access land and labor in response to the changing produc- markets in 1996/97, which eliminated subsidies but tion conditions of a constantly moving cocoa frontier. Until improved private distribution (Vigneri and Teal 2004). the early 1940s, when both land and labor were abundant, large farms were able to attract rural workers to establish Adoption of improved varieties. Hybrid cocoa varieties new farms by selling them small plots of land, an arrange- were introduced in 1984 through the government’s Cocoa ment that often also drew the workers’ family members to Rehabilitation Project (CRP). Hybrid varieties outperform establish and maintain new farms. By the second half of the the older “Amazons” and “Amelonado” varieties in two 1960s, when land became scarce, sharecropping arrange- ways—by producing trees that bear fruit in three years com- ments increasingly replaced land sales. During times when pared with at least five years for the older varieties, and by the cost of hiring waged workers became too high, alterna- producing more pods per tree.1 But hybrid cocoa trees tive forms of labor were used—mostly, either sharecrop- underperform older varieties in that they require optimal ping arrangements or informal labor groups known as weather conditions and complementary farming practices nnoboa (Berry 1993; Blowfield 1993; Vigneri, Teal, and such as the application of chemical inputs, adoption of new Maamah 2004; and Amanor 2010). Since 1990 noticeable planting procedures, pruning, and spraying. Hybrids vari- changes have taken place in the technology of cocoa pro- eties also require that farmers make more harvest rounds at duction, in particular increased use of fertilizers; the adop- the beginning and the end of the season, something they are tion of hybrid cocoa varieties, and better control of pests reluctant to do when it conflicts with other farming or trad- and diseased trees (Boahene, Snijders, and Folmer 1999; ing activities (Boahene, Snijders, and Folmer 1999; Bloom- Edwin and Masters 2003; Gockowski and Sonwa 2007; Teal, field and Lass 1992). Zeitlin, and Maamah 2006; Vigneri, Teal, and Maamah Despite the increased labor input for hybrid cocoa trees, 2004; and Vigneri 2008). farmers have increasingly adopted them. In the late 1980s only 10 percent of cocoa grown in Ghana was of the high- Increased use of fertilizer. Fertilizer use in Ghana has yielding type (Nyanteng 1993). By 2002, 57 percent of farm- increased significantly since the 1990s. Surveys of cocoa ers in the three main cocoa-producing areas were growing farmers in the three main cocoa-producing regions of hybrid trees (Vigneri 2005). Traditional varieties may have Ghana show that fertilizer application rates increased from disappeared entirely from all fields planted after 1995 9 percent in 1991 to 47 percent in 2003 (table 12.1). (Edwin and Masters 2003). Although the quantity of fertilizer used decreased between 1991/92 and 1997/98, the proportion of farmers applying Better disease and pest control. Control of disease fertilizer increased, possibly from liberalization of input and pests, swollen shoot virus and capsid in particular, CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 205 has improved significantly in recent years. After Cocobod their living conditions compared with food crop farmers initiated a free mass spraying program in 2001, 93 percent (McKay and Coulombe 2003). Poverty reduction among of cocoa farmers who participated in a survey conducted cocoa farmers is clear. Household surveys indicate that in 2002 linked their yield improvements to the effects of poverty among cocoa-producing households dropped to the program (Steedman 2003). Similarly the cocoa farm- 23.9 percent in 2005, down from 60.1 percent at the begin- ers’ panel referred to above for the crop years 2002 and ning of the 1990s (World Bank 2007b). 2004 suggests that nearly all farms were sprayed in 2003/04, when producers reported an average of more Reputation for high-quality cocoa than four spraying applications during the crop year, of which 46 percent were carried out by the government Cocoa, like many other commodities, is often differentiated (Vigneri 2005). by country of origin, and this in turn is associated with a The effect of all these improved practices has been an reputation based on average quality. The reputation, a increase in productivity of about 30 percent, which brought national public good, enables the country to earn a pre- productivity to the levels achieved in the 1980s (figure 12.2). mium in the global market for the crop it is producing. Productivity was stagnant until the late 1980s, with produc- Generally, Ghana receives a price premium for its cocoa in tion largely related to area harvested. The first big jump in world markets because of the slightly higher-than-average productivity occurred in the 1980s, corresponding to the fat content; low levels of debris, which results in higher year of the Cocoa Rehabilitation Program rolled out under cocoa butter yields than beans containing high levels of the ERP, and the second more recently, with improved prac- debris; and low levels of bean defects, which generate a tices. The correlation between production and area har- cocoa liquor flavor preferred by some end users. In addi- vested remains strong. tion to these attributes, the reputation of the Cocoa Mar- keting Company (the government division in charge of all exports) in ensuring the consistency and reliability of Cocoa’s contribution to economic growth and cocoa-related shipments and documents has played a cen- poverty reduction tral role in establishing the country’s reputation for high- In the Southern Forest Belt, where cocoa is produced, aggre- quality beans (Agrisystems Ltd. 1997). Using trade NYSE gate figures suggest that through the 1990s, cocoa-farming Liffe cocoa market information, Gilbert (2009) suggests households, along with those engaged in mining or timber that Ghanaian cocoa draws a premium of 3 to 5 percent rel- (the other predominantly export-oriented activities) and ative to Côte d’Ivoire, currently the world’s largest producer other commercial activities, experienced improvements in of cocoa (table 12.2). Figure 12.2 Cocoa Production, Area Cultivated, and Yields, 1961–2008 2.5 800 700 Millions of hectares harvested 2.0 600 Yields (kgs/ha) 1.5 500 400 1.0 300 200 0.5 100 0 0 61 19 3 65 67 19 9 71 19 3 75 77 19 9 19 1 19 3 85 87 89 91 93 19 5 19 7 99 01 20 3 05 07 6 6 7 7 8 8 9 9 0 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 Area harvested, left hand side Production (000 tons), right hand side Yields (Kgs/ha), right hand side Source: FAOSTAT. 206 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY Table 12.2 Cocoa Unit Values and Terminal Market Differentials Percent Cameroon Ghana Nigeria Period Unit value Differential Unit value Differential Unit value Differential 1988–1991 2.7 — 3.7 — –0.4 — 1992–2002 –3.0 0.20 1.1 4.8 –2.1 –0.5 2003–2008 –7.8 — 5.2 4.9 –0.7 –0.9 1988–2008 –3.3 — 2.8 4.9 –1.4 –0.7 Source: Adapted from Gilbert (2009). Figures reported are relative to those of Côte d’Ivoire, the reference country. Note: — = not available. Characteristics that determine the quality of cocoa Figure 12.3 Share of Processed Cocoa Products in Total include content and quality of fat, consistency in the size of Cocoa Exports in West African Countries, 1990–2007 the beans, and their moisture content. These characteristics 30 determine the quality of cocoa butter and cocoa liquor produced from the beans, the two ingredients that control 25 texture, aroma, color, and flavor of chocolate. The fermen- 20 tation, drying, storage, and evacuation of wet beans can Percent alter the quality of cocoa beans dramatically, particularly in 15 the development of the flavor of cocoa liquor. The classic 10 “West African” cocoa flavor is obtained by fermenting beans in a heap under banana leaves for about six days with fre- 5 quent manual turning and thorough drying in the sun. 0 Drying beans slowly on raised platforms is very important 90 92 94 96 98 00 02 04 06 19 19 19 19 19 20 20 20 20 for the quality of flavor because it quickly decreases the Cameroon Côte d'Ivoire Ghana Nigeria acidity level of the beans. Quality is also maintained by quickly collecting properly fermented and dried beans from Source: FAOSTAT. smallholder farmers and promptly shipping them to avoid Note: Cocoa processed products include cocoa butter and cocoa paste. the buildup of moisture, mold, and free-fatty acids that can rapidly deteriorate the quality of the bean. Partly because of its reputation for high-quality cocoa, used for local processing, which has resulted in consider- Ghana is able to sell most of its annual production able underutilization of existing capacity in the country. through forward contracts, which fix the price farmers are given for their cocoa for the entire crop year. The value Increased share of free on board prices that international firms place on Ghana’s cocoa is also going to farmers reflected by the amount of investment they have made in processing facilities in the country. Ghana’s export earn- Agricultural exports continue to be the most important ings from processed cocoa products more than tripled source of foreign exchange for the majority of Sub-Saharan between 1991 and 2004, from $32 million to $105 million African countries (Gilbert 2009). In virtually every country (figure 12.3).2 However, because of the limited conditions in Africa with a major export crop, including Ghana, the under which semiprocessed cocoa can be transported effec- government has intervened through state-owned marketing tively (Fold 2002), it is not clear whether local value-adding boards, or caisses de stabilization, to coordinate the produc- efforts will be sufficiently profitable for international com- tion and marketing of the crop, offering farmers stable farm panies to expand their operations in Ghana. Thus far, gate price that shield them from price volatility. Many schol- informal discussions with the private sector participants ars (Bates 2005; McMillan 1998; Akiyama et al. 2001) hold indicate that the net benefits from processing locally may that marketing boards in Africa have long operated as cor- not be significant, particularly because the government rupt institutions taxing farmers through the power to set allows only a limited quantity of low-quality beans to be prices and indirectly by maintaining overvalued exchange CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 207 rates. That said, the role of governments in the agricultural the farmers, in response to pressure from multilateral sector has changed substantially since independence. organizations to streamline its operations (Brooks, Crop- Despite granting Cocobod the monopoly over market- penstedt, and Aggrey-Fynn 2007). Figure 12.4 shows the ing, Ghana has managed to develop a marketing system that share of f.o.b. prices paid to producers, the share retained by passes on an increasingly larger share of export prices to Cocobod (shown as direct taxation), and the share of indi- farmers. Prices received by Ghanaian producers have been a rect taxation imposed by the exchange rate.4 Exchange rate function of government interest in using the sector as a distortions can further erode the share producers receive. source of revenue and a balance against global prices, These distortions were high in the mid-1980s but have com- exchange rate distortions, and inflation. Price policies were pletely disappeared. also made ineffective by macroeconomic policies. In the early 1980s, for example, the Provisional National Defense REASONS FOR SUCCESS OF THE Council (PNDC) had to choose between supporting cocoa COCOA SECTOR farmers and continuing to maintain highly overvalued exchange rates (Stryker 1990). A number of factors have contributed to the success of In Ghana, the price producers are paid for cocoa is cur- Ghana’s cocoa sector: a favorable price regime, both in rently set at the beginning of the harvest season for the entire terms of the f.o.b. share passed on to producers and the real crop year by the Producer Price Committee.3 The price is price received by farmers; improved marketing through par- based on the price Cocobod expects to receive, having tial liberalization; and Cocobod’s interventions to raise already sold nearly 70 percent of the crop. To this price, cocoa productivity. Cocobod adds the costs of its operations and the export tax to arrive at what it calls “net free on board (f.o.b.) price.” Favorable prices The share of the net f.o.b. price received by cocoa farm- ers in Ghana has increased to nearly 80 percent after having With the exception of 1998–2000 and 2003–06, world cocoa fallen below 20 percent before the economic reforms of the prices have steadily increased since 1990. This, combined 1980s, and as low as almost 5 percent between 1975 and with a higher share of the price being passed on to farm- 1981. By 1987/88, real producer prices in Ghana had ers, has offered farmers increasing real producer prices increased threefold compared with 1983/84, largely as a (figure 12.5). A variety of models estimating the sensitiv- result of Cocobod’s revised policy of paying higher prices to ity of production supply to farm gate prices find that Figure 12.4 Farm Gate Prices, Direct Taxation, and Exchange Rate Taxation for Ghanaian Cocoa, 1966–2008 Percent 120 100 80 60 40 20 0 –20 19 67 19 69 19 /71 19 73 19 /75 19 /77 19 79 19 /81 19 83 19 85 19 87 19 /89 19 /91 19 /93 19 /95 19 /97 20 /99 20 /01 20 03 20 /05 20 /07 9 /0 / / / / / / / / 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 Farm gate price Direct taxation Exchange rate taxation Source: Vigneri 2005. 208 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY Figure 12.5 Ghana Cocoa Production and Real Producer Price, 1990–2008 180,000 800 160,000 700 140,000 Price per ton (2005 cedis) 600 120,000 500 Metric tons 100,000 400 80,000 300 60,000 200 40,000 20,000 100 0 0 19 0 19 1 19 2 93 19 4 19 5 19 6 97 19 8 20 9 20 0 20 1 02 20 3 20 4 05 20 6 07 08 9 9 9 9 9 9 9 9 0 0 0 0 0 19 19 19 20 20 20 Real producer price Production Source: Cocobod and ICCO. small-scale cocoa producers in Ghana have responded posi- reduced from 100,000 in the early 1980s to 10,400 in 1999 to r tively to these price incentives (Bulíˇ 2002; Hattink, Heerink, just over 5,100 in 2003, bringing down costs considerably. In and Thijssen 1998; and Vigneri 2005, among others). the same year, Cocobod ended its control over all domestic Although strictly comparable data are not available, purchases by allowing a number of private licensed compa- informed inference on the returns on cocoa farms using the nies to compete with its former purchasing agency, the results from two rural surveys, one conducted in 1996 Producing Buying Company (PBC), to buy and transport (Agrisystems Ltd. 1997) and one in 2006 (Barrientos and the cocoa crop from farms; the board, however, specifies a Asenso-Okyere 2008), show that cocoa production has not minimum price. This partial liberalization appears to have become more profitable for farmers. In fact, calculations benefited producers. The internal marketing of cocoa has show that cocoa, which usually is the largest source of earn- also become more competitive in recent years, with nearly 20 ings in cocoa-producing households, accounting for more licensed buyers, along with PBC, procuring cocoa through than 67 percent of revenues, has actually declined over time: nearly 3,000 buying stations manned by purchasing clerks or net cocoa profits for cocoa-producing households were individuals from cocoa communities who purchase the crop 7 percent lower in 2005 than in 1996. While the real price of on the buyers’ behalf. Although the total number of licensed cocoa increased by 47 percent between these two years, the buyers is relatively large, five dominate the market: the Pro- cost of inputs increased more. These estimates, however, do duce Buying Company, Kuapa Kokoo, Olam, Armajaro, and not suggest a trend because they are based on observations Global Haulage, a former transport company comprising for two specific years. three Ghanaian buyers (Federal Commodities, Transroyal, and Adwumapa) Additionally, Cocobod extends funds to producers at rates slightly below the market rate to finance Liberalization of domestic cocoa marketing their operations. It also monitors producers’ operations, par- Following pressures from multilateral organizations in the ticularly with regard to quality of beans. Though licensed early 1980s, wide-ranging changes were introduced to buyers are free to export, none of them has thus far because improve Cocobod’s efficiency: transport was shifted to the none is large enough to acquire the minimum amount private sector, feeder road development was transferred to needed to be eligible to export. the Ministry of Roads and Highways, and in 1988–89 input Zeitlin (2006) finds a positive correlation between the subsidies were phased out (Brooks, Croppenstedt, and concentration of licensed buying companies at the village Aggrey-Fynn 2007). Following the 1992 elections more dras- level and production. But the direction of causality is not tic measures were undertaken: Cocobod staff levels were clear, because buyers are also likely to locate themselves CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 209 where large quantities of cocoa are available for purchase. fertilizer on credit. The program collapsed after one year, The PBC continues to operate as a buyer of last resort. however, because of poor repayment rates. Following this While Cocobod sets a minimum price that must be paid to pilot, a private agri-input company, Wienco, tested a package producers, the buying companies are free to pay higher of agricultural inputs and farm practices known as the prices. Even in the absence of price competition among “Abrabopa package.” In 2003, its first year of testing, the pack- licensed buyers, farmers have benefited. Payments to farm- age raised yields from 510 to 1,081 kilograms per hectare and ers have become more reliable, and corruption, which char- to 2,317 kilograms per hectare after the third year. acterized the contractual negotiations when the PBC was In 2006 the Cocoa Abrabopa Association (CAA) was the only buyer, has diminished. While licensed buyers may established, under which groups of farmers with mature not compete on prices, they do offer occasional price trees on at least one hectare of land were given the Abrabopa bonuses, subsidized inputs, or credit extensions for produc- package on credit and offered technical and business train- ers (Laven 2007). Because the licensed buyers buy cocoa ing. The number of farmers participating in this program throughout the year, however, the new buying system puts a reached 11,000 in 2008. An evaluation of the program in steadier stream of money into the hands of producers 2008 (Opoku et al. 2009) suggests that the principle of (Vigneri and Santos 2008), giving farmers working capital group liability employed in this program ensured, to some to buy labor and other inputs when they need them. extent, the effective use of the fertilizer and other inputs Although the efforts at liberalization are likely to have provided by the CAA package. That said, a large proportion made procurement and transport more efficient than of farmers, nearly 40 percent, dropped out of the program, before, it is unclear whether Cocobod’s costs have been so the benefits of the CAA package reached only a small reduced by “outsourcing” procurement and transport and share of cocoa growers. to what extent liberalization may have helped Cocobod pass on a higher share of f.o.b. prices to farmers. But regardless, Cocobod’s role in maintaining quality retaining control over exports and other aspects of market- ing has enabled Cocobod to support producers in ways that In terms of quality practices by government marketing would not have been feasible had it devolved these respon- boards among West African cocoa-producing countries, sibilities to other organizations. Ghana is an exception, because maintenance of quality continues to be Cocobod’s mandate even after its restruc- turing. In other countries, dismantling and restructuring of Cocobod’s impact on productivity marketing boards in the 1980s radically reduced quality Importantly, Cocobod’s continued involvement in the cocoa control systems (Fold 2001; Gilbert 2009). One rationale for sector in Ghana has allowed surpluses generated in good a government role in maintaining quality is that cocoa is years to be used to finance deficits during years when prices transported in bulk, and poor-quality cocoa beans can were low. Similarly, Cocobod has invested in research, dis- diminish the quality of other beans in the same shipment, ease control, and credit programs that are of general benefit thereby affecting the price of all beans in the shipment. to the cocoa industry (Stryker 1990). In 2001 the Cocoa Maintaining a government role is also important because it National Disease and Pest Control Committee was estab- allows the government to control the national reputation of lished to develop strategies to control capsid and black pod Ghana’s cocoa and keep its premium in the world market through a nationally coordinated spraying program under (Fold and Ponte 2008). This quality maintenance comes at a which Cocobod, through a network of regional offices, cost, however, including the cost of ensuring that lower- undertakes spraying of all cocoa fields at no cost to the quality beans are not mixed into those prepared for export producers.5 By Cocobod’s estimates, the scheme has had a and the costs of administration. positive impact on national cocoa production, particularly during the 2003/04 and 2005/06 seasons. Cocobod also SUSTAINABILITY OF THE COCOA SECTOR reports that the protection of the cocoa plants that the pro- gram offers has encouraged farmers to undertake additional Ghana’s cocoa sector faces a number of challenges. For one, spraying applications. productivity levels are lower than they are in other countries. In 2002/03, Cocobod rolled out the “Cocoa High-Tech” Ghana also faces the possibility that its quality advantage program designed to encourage farmers to apply a minimum may disappear in the coming years. In addition, Ghana must of 5 bags of fertilizer per hectare of planted cocoa, supplying determine how to keep its cocoa sector competitive as 210 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY cocoa-producing households change. Finally, the environ- (Gockowski and Sonwa 2007). One evaluation (Opoku et al. mental impact of current farming practices may soon con- 2009) suggests that the high dropout rate from the CAA strain cocoa production expansion. On the other hand, program may result from high variability in the expected however, Ghana has been quite successful in taking advan- returns from fertilizer applications. tage of niche cocoa markets. The low level of tree replanting is an additional threat to the sustainability of Ghana’s cocoa production. Often, farm- ers find it more economical to expand their farms rather Productivity and competitiveness than to replace old and diseased trees (Vigneri 2005; Ruf Notwithstanding the technical changes that have occurred in and Burger 2001), because it takes twice as long to clear an cocoa production, Ghana still needs to close a large produc- old farm as it does to clear new forest land (Masdar Ltd. tivity gap to remain competitive. The gap between observed 1998). Additionally, farmers regard the expansion of land on and achievable yields is 50–80 percent (Gockowski 2007), which cocoa is planted as both an investment and a means depending on the production practices adopted by farmers to establish land ownership. Given that migrants and share- (for example, thin shading and the amount of fertilizer croppers represent an increasing share of the cocoa-farming applied). A survey conducted in the 1980s, however, indi- population, this dual view means that many farmers seek to cated that Ghana was the lowest-cost producer in the world acquire permanent land rights by expanding into unculti- (Bloomfield and Lass 1992). Ghana’s yields are low com- vated land, where land ownership is established by clearing pared to those of its leading competitors, Côte d’Ivoire and land and planting new trees (Amanor 2010; Berry 2009; Indonesia (figure 12.6). Additionally, it is not clear which Takane 2002).6 Further opportunities to increase produc- technologies intended to increase productivity are attractive tion by land expansion may be limited, though, by the to farmers. For example, farmers may not have much incen- decreasing availability of virgin forest land. tive to apply fertilizers to hybrid trees, because the returns from doing so may not be higher than those achieved on tra- Longevity of the quality advantage ditional varieties (Edwin and Masters 2003). On experimental farms, application of fertilizers to Although Ghanaian cocoa draws a premium price for its young trees has increased yields as much as threefold reliable quality, this advantage may be eroded in the future Figure 12.6 Cocoa Yields, by Country, 1990–2008 1,200 1,100 1,000 900 kg per hectare 800 700 600 500 400 300 200 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 Cameroon Côte d'Ivoire Ghana Indonesia Nigeria Source: FAOSTAT. CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 211 because of technological advances in processing most densely populated regions of the cocoa belt. The study (Agrisystems Ltd. 1997; Fold 2001). On the other hand, estimates that this could result in the replanting of up to current quality control processes in Ghana guarantee mini- 24,000 hectares of land, and that integrating this interven- mum parameters that are important to large industry tion with the expansion of fertilizer use would achieve pro- players like Cadbury, which is known to use Ghanaian cocoa ductivity gains in excess of 50 percent. For less efficient beans exclusively in all its U.K.-retailed chocolate products. cocoa producers, the STCP recommends implementing a A second, potential threat is when other cocoa-producing different set of policies that would either allow these produc- countries improve on their quality. Currently, this is not ers to exit the sector or support their transition to alternative much of a threat to Ghana, because smallholder farmers in production systems. One option for these less efficient farm- countries such as Malaysia and Indonesia lack the institu- ers would be the conversion from a no-shade cocoa system to tions to support quality. In Côte d’Ivoire, the mixing of a partial-shade system with cocoa and non-cocoa trees inter- good cocoa beans with the bad ones in shipments for export cropped, allowing producers to augment their incomes from results in variability in quality (Bloomfield and Lass 1992). the sale of forest products, and possibly from the additional payments for higher carbon sequestration associated with shaded tree systems. Competitiveness of cocoa on farms Cocoa is a mixed crop system in which other crops may be Environmental impact of current consumed or sold. Intercropping with plantain and farming practices cocoyam, for example, provides early returns when cocoa trees are still young. Studies conducted in the 1970s and in An issue closely related to the competitiveness of cocoa on the 1990s (Rourke 1974; Masdar 1998) report that almost all farms is the environmental impact of existing farming prac- cocoa farmers grew alternative crops for subsistence and tices. Since its introduction in West Africa, cocoa has been sale, mostly roots and tubers but also a variety of cereals and the major cause of land use change in the high forest zones vegetables. Both studies also suggest that many farmers of the regions in which it is grown, where it has replaced shifted to crops other than cocoa (mixed plantain and agricultural activity that incorporated fallowing to maintain cocoyam, mixed maize and cassava, and oil palm inter- land fertility (Gockowski and Sonwa 2007). Although the cropped with maize and cassava) on a scale greater than that initial expansion of cocoa production did not entail a com- needed to satisfy subsistence needs. This shift occurred for plete removal of the forest shade because the traditional several reasons: the crops offered farmers greater income shade-dependent and tolerant tetteh quarshie variety of continuity throughout the year, and returns were perceived cocoa did not require forest clearing, trees have been cut to be higher relative to cocoa, especially in the presence of down en masse in recent years to accommodate the open- significant problems with the rehabilitation of the existing field hybrid variety, which grows in full sun conditions. In cocoa tree stock. nearly three-quarters of Ghana’s production area, there is More recent research has questioned the viability of little to no shade (table 12.3). cocoa on small farms. A 2001 survey conducted by the Sus- Farmers in Ghana have a strong preference for full-sun tainable Tree Crop Programme (STCP) in four cocoa- crops because their much shorter growing cycle is linked to producing countries in West Africa shows that the top 25 higher short-term profits (Obiri et al. 2007). The damage to percent of households (ranked by the amount of cocoa cocoa trees from capsid attacks tends to be higher for cocoa produced) have average costs of production four times trees growing in full sun than for those in shaded systems, lower and yields nearly four times greater than the bottom 25 percent, and that a significant share of small cocoa farms incur losses (Gockowski 2007). The study recom- Table 12.3 Shade Levels in the Cocoa Belt of Ghana (percent) mends the urgent adoption of policies that vary for larger, more efficient producers and poorer marginal ones as a Region None to light Medium to heavy necessary step in keeping Ghana’s cocoa sector competitive Ashanti 52 47 Brong Ahafo 52 47 and efficient. Eastern 50 49 For the larger producers, the STCP study recommends Western 77 21 implementing innovations through the strategic distribu- Ghana 72 29 tion of improved planting material (hybrid pods) in the Source: Adapted from Gockowski and Sonwa (2007). 212 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY however, and the carbon sequestration potential of full-sun obtain certification for organic production, with more than cocoa systems is significantly less than that of traditional 500 members meeting the required standards. shaded cocoa systems (Norris 2008). The best possible environmental alternative to the cur- LESSONS FROM GHANA’S EXPERIENCE rent cocoa-growing practices in Ghana would be a mixed WITH COCOA agroforestry system, where the forest is selectively thinned and fruit trees with economic value—such as oil palm, avo- Cocoa was developed in Ghana, largely by commercial cado, and citrus—are grown next to cocoa trees, providing farmers, many of whom were smallholders and laborers both shade for the cocoa trees and food and income for the drawing on their own savings and labor, in response to mar- farming household (Gockowski and Sonwa 2007). This ket opportunities and the development of infrastructure. practice, which is used in southern Cameroon, could offer Policies and institutions have played an important role. The farmers up to 23 percent of total revenues from their non- importance of macroeconomic management, the avoidance cocoa holdings, but it is rarely practiced in Ghana. One rea- of distortions in the exchange rate in particular, is clearly son is that in the biodiversity hotspots in remote areas of the evident from the effect of its absence on farm gate prices in Western region, the profitable marketing of agroforestry the mid-1980s. products would not be easy. Additionally, past logging prac- Ghana appears to have emerged with an appropriate tices, in which concessionaires harvested in a way that institutional mix in which competition has been introduced destroyed cocoa farms with no compensation for producers, in internal marketing to benefit from efficiencies in pro- have discouraged the use of fruit and timber-producing curement and transport, while the government marketing trees in cocoa fields (Obiri et al. 2007). board retains control over setting minimum prices for the year, maintaining quality, and managing exports. The con- trol it has retained over exports enables it to stabilize prices Ghana’s role in a changing global and use the surpluses to offer some services such as plant market for cocoa protection, research, and extension that may not be forth- Ghana is well positioned to expand its position in high- coming from the private sector, as suggested by the experi- value markets, with Cocobod proving to be responsive to ence of the fully liberalized producing countries in the trends in international markets. The chocolate industry also region. Public support to farmers to rehabilitate the dis- has expanded into secondary markets, such as fair trade in eased tree stock, public research that produced new hybrids, the late 1980s. Although these markets offer strategic oppor- and the continued state intervention to promote fertilizer tunities for countries to build competitiveness, estimated in use have all been instrumental in reviving the sector. More 2000 at 2.6 percent of world cocoa bean trade (Abbott recently public spraying and dissemination of technical 2002), they largely remain niche markets because of their packages have spurred private action. limited capacity for expansion. Would the cocoa sector have been better if it were fully Ghana’s considerable progress in the fair-trade cocoa liberalized? Examining the experience of liberalization of market began with the establishment, in 1993, of Kuapa cocoa sectors in four West African countries, Gilbert (2009) Kokoo, a farmers’ cooperative that operates as a private, suggests four criteria to address this question: (1) the level licensed buying company. Its share in the domestic market of competition achieved on both the export and import side is now estimated to be around 10 percent of total purchases, of producing countries, (2) the ability to sustain quality and a panel survey of farmers spanning 2002 to 2006 shows standards, (3) the share of the f.o.b. price passed on to the the cooperative to be farmers’ second preferred outlet for farmers as an indication of the degree of state taxation, and selling beans (Vigneri and Santos 2008). Within Cocobod, a (4) the extent of producer price stabilization achieved. The special channel exists for fair-trade cocoa sourced and evidence suggests that it may not be so. exported from Kuapa Kokoo, although the system traces Liberalization has not resulted in competition in the such cocoa back to the cooperative rather than to the indi- value chain, particularly in exports. Local companies vidual farmer. The social premium earned on fair-trade engaged in exports without access to global financing have exports, which in 2000 was reported to be $150 per ton withdrawn over the years, leaving exports largely in the (Abbott 2002), goes into a trust fund that sponsors develop- hands of multinationals, either converters or their agents ment projects in cocoa-producing communities. Recently, (Gilbert 2009). But, there has been greater competition in the CAA became the first cocoa cooperative in Ghana to internal trade. As for the share of the f.o.b. price passed on CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 213 to the farmers, the proportion is higher in countries such as twice as much cocoa per hectare as similar-aged fields Cameroon and Nigeria, but Ghana’s government has made planted with traditional trees. concrete efforts in the recent past to raise the share similarly. 2. Ghana maintains a state-owned processing plant, the Finally, in relation to the price stabilization objective, Ghana Cocoa Processing Company (CMC). Historically, CMC has has clearly been successful in reducing farmers’ exposure to operated at low capacity. A five-year rehabilitation and price variability during the crop year through its practice of expansion program, however, allowed it to double its annual forward sales. This, combined with the more stable inflation processing capacity between 2004 and 2009. rate of the past decade, has de facto acted as an insurance 3. The committee includes a variety of representatives mechanism against the variability in the world price of the from the cocoa sector: Cocobod, government officials, and commodity. Global businesses like the Ghana model representatives of cocoa buyers, the national cocoa farmers’ association, and haulers and transporters. because it delivers consistently high quality. Local businesses are also content because they can continue to participate in 4. Indirect taxation is measured as the difference between world prices converted using the official exchange rate and the sector (Gilbert 2009). world prices converted using the market exchange rate. Direct The interesting question is whether it is possible to arrive taxation includes Cocobod’s marketing costs and export at this mix of public and private institutions and also be cer- duties imposed by the government (export duties have been tain that a parastatal organization such as the Cocobod close to 25 percent in recent years). The share of f.o.b. prices would operate reasonably efficiently. Ghana’s experience received by farmers does not correspond with global prices. suggests that external pressures as a part of the ERP to For example, between 1971 and 1983, the farmer share reform the sector were instrumental in making Cocobod lib- declined sharply while global prices were rising. This period, eralize some of its operations and streamline its own work- however, coincided with acute domestic currency overvalua- ing to reduce costs. Ghana appears to have done enough to tion in Ghana, which further eroded farmers’ real producer fend off pressures for further liberalization of the sector. To prices. Similarly, in the mid-1990s, producers’ share of world what extent it will strive to continue to pass on a higher share prices increased while global prices were falling. of prices to farmers without external pressures and whether 5. How much of the program is funded by cocoa revenues there is a recognition of the benefits from appropriate man- and whether any of the program is subsidized by the gov- agement that survives political changes are not clear. The ernment is not clear, however. affairs of the Cocobod are not as transparent as they should 6. In Ghana, the distinction between land ownership and be, and the line between cocoa revenues and government usufruct rights over what grows on land has traditionally shaped smallholders’ investment choices. finances remains fuzzy. Whether the Cocobod will be able to stabilize prices if the world market were to become more volatile than it has been in recent years is not clear. The pressure on the government and on its marketing BIBLIOGRAPHY institution to improve their efficiency (as measured by the Abbott, P. 2002. “Towards more Socially Responsible Cocoa share of the world price going to producers) rather than to Trade.” Working Paper 03-3, presented at the Annual seek full liberalization appears to have worked well in Meeting of the International Agricultural Trade Research Ghana. Given the preponderance of smallholders in the sec- Consortium (IATRC), Monterey, CA, December 15–17. tor and the risks associated with the total withdrawal of the Agrisystems Ltd. 1997. “Study of the Cocoa Sector to government’s services, the partial liberalization experience Define Interventions on Behalf of Ghana’s Smallholder of Ghana’s cocoa sector has so far offered a unique example with Particular Reference to the Framework of Mutual of how it is possible to learn from past reforms and to con- Obligations.” Prepared for the Stablex 1992 and 1993 tinue to seek further reform to sustain the sector. However, Allocation. the scope for future improvements and for further learning Akiyama, T., J. Baffes, D. Larson, and P. Varangis. 2001. opportunities will require appropriate pressures from both Commodity Market Reforms: Lessons of Two Decades. World Bank Regional and Sectoral Studies. Washington: local political processes and from external sources. World Bank. Allman, J. M. 1993. The Quills of the Porcupine. Madison, WI: University of Wisconsin Press. NOTES Amanor, K. 2005. “Agricultural Markets in West Africa: 1. Using survey data collected in 2002, Edwin and Masters Frontiers, Agribusiness and Social Differentiation.” IDS (2003) show that the new tree varieties yield approximately Bulletin 36 (2). 214 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY ______. 2010. “Family Values, Land Sales and Agricultural Distortions Working Paper 47, World Bank, Washington, Commodification in Rural Ghana.” Africa 80 (1): DC. 104–125. r Bulíˇ, A. 2002. “Can Price Incentive to Smuggle Explain the Austin, D. 1978. Politics in Africa. Manchester, U.K.: Man- Contraction of the Cocoa Supply in Ghana?” Journal of chester University Press. African Economies 11 (3): 413–39. Awanyo, L. 1998. “Culture, Markets, and Agricultural Pro- Coulombe, H., and Q. Wodon. 2007. “Poverty, Livelihoods, duction: A Comparative Study of the Investment Pat- and Access to Basic Services in Ghana.” In “Ghana Coun- terns of Migrant and Citizen Cocoa Farmers in the West- try Economic Memorandum: Meeting the Challenge ern Region of Ghana.” Professional Geographer 50: of Accelerated and Shared Growth.” World Bank, 516–30. Washington, DC. Barrientos, S. W., and L. Asenso-Okyere. 2008. “Mapping Edwin, J., and W. A. Masters. 2003. “Genetic Improvement Sustainable Production in Ghanaian Cocoa: Report to and Cocoa Yields in Ghana.” Working Paper, Purdue Cadbury.” Institute of Development Studies, University University, West Lafayette, IN. of Sussex, and University of Ghana. Fold, N. 2000. “A Matter of Good Taste? Quality and the Bateman, M. 1965. “Aggregate and Regional Supply Func- Construction of Standards for Chocolate Products in the tions for Ghanaian Cocoa: 1946–1962.” Journal of Farm European Union.” Cahiers d’économie et sociologie Economics 47 (2): 384–401. rurales, 55–56. Bates, R. H. 2005. Markets and States in Tropical Africa. The ______. 2001. “Restructuring of the European Chocolate Political Basis of Agricultural Policies. Berkeley, CA: Uni- Industry and Its Impact on Cocoa Production in West versity of California Press. Africa.” Journal of Economic Geography 1: 405–20. Beckman, B. 1976. Organising the Farmers: Cocoa Politics ______. 2002. “Lead Firms and Competition in ‘Bi-polar’ and National Development in Ghana. Uppsala: Scandina- Commodity Chains: Grinders and Processors in the vian Institute of African Studies. Global Cocoa-Chocolate Industry.” Journal of Agrarian Berry, S. 1993. No Condition Is Permanent: The Social Change 2 (2): 228–47. Dynamics of Agrarian Change in Sub-Saharan Africa. Fold, N., and S. Ponte. 2008. “Are (Market) Stimulants Inju- Madison, WI: University of Wisconsin Press. rious to Quality? Liberalization, Quality Changes and the ______. 2009 “Building for the Future? Investment, Land Reputation of African Coffee and Cocoa Exports.” In Reform and the Contingencies of Ownership in Con- Globalization and Restructuring of African Commodity temporary Ghana.” World Development 37 (8): 1370–78. Flows, ed. N. Fold and M. N. Larsen. Uppsala, Sweden: Bloomfield, E. M., and R. A. Lass. 1992. “Impact of Struc- Nordic Africa Institute. tural Adjustment and Adoption of Technology on Com- Ghana Cocoa Marketing Board. Various years. “Cocoa Mar- petitiveness of Major Cocoa Producing Countries.” keting Board Newsletter.” Working Paper 69. Paris: Organisation for Economic Co- Ghana Ministry of Manpower, Youth and Employment. operation and Development. 2006. “Labour Practices in Cocoa Production in Ghana Blowfield, M. 1993. “The Allocation of Labour to Perennial (Pilot Survey).” Accra. Crops: Decision-making by African Smallholders.” NRI- Gibbon, P., Y. Lin, and S. Jones. 2009. “Revenue Effects of Socioeconomic Series 3, Natural Resources Institute, Participation in Smallholder Organic Cocoa Production University of Greenwich, Kent, U.K. in Tropical Africa: A Case Study.” DIIS Working Paper Boahene, K., T. A. B. Snijders, and H. Folmer. 1999. “An Inte- 2009:06, Danish Institute for International Studies, grated Socio Economic Analysis of Innovation Adoption: Copenhagen. The Case of Hybrid Cocoa in Ghana.” Journal of Policy Gilbert, C. L. 2009. “Cocoa Market Liberalization in Retro- Modelling 21 (2): 167–84. spect.” Review of Business and Economics 54: 294–312. Breisinger, C., X. Diao, S. Kolavalli, and J. Thurlow. 2008. Gill & Duffus Group. Various years. “Cocoa Statistics.” “The Role of Cocoa in Ghana’s Future Development.” Gockowski, J. 2007. “The Analysis of Policies, Productivity IFPRI Background Paper 11, IFPRI, Ghana. and Agricultural Transformation in the Cocoa-Produc- Breisinger, C., X. Diao, J. Thurlow, and R. M. Al Hassan. ing Rural Economies of West Africa.” STCP Technical 2009. “Potential Impacts of a Green Revolution in Africa Report Executive Summary. – The Case of Ghana.” Paper presented at the 27th IAAE Gockowski, J., and D. Sonwa. 2007. “Biodiversity Conser- Conference, Beijing, August 16–22. vation and Smallholder Cocoa Production Systems in Brooks, J., A. Croppenstedt, and E. Aggrey-Fynn. 2007. “Dis- West Africa with Particular Reference to the Western tortions to Agricultural Incentives in Ghana.” Agricultural Region of Ghana and the Bas Sassandra region of Côte CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 215 d’Ivoire.” Draft paper, Institute of Tropical Agriculture, Okali, C. 2010 “The Organisation of Cocoa Production on Ibadan, Nigeria. http://www.odi.org.uk/events/2007/ the Farm and its Evolution over Time.” Unpublished 11/19/434-paper-discussion-biodiversity-conservation- paper. smallholder-cocoa-production-systems-west-africa.pdf. Opoku, E., R. Dzene, S. Caria, F. Tea, and A. Zeitlin. 2009. Government of Ghana. 2003. “Ghana Poverty Reduction “Improving Productivity through Group Lending: Strategy, 2003-2005: An Agenda for Growth and Pros- Report on the Impact Evaluation of the Cocoa Abrabopa perity.”Government of Ghana, Accra. Initiative.” Centre for the Study of African Economies, Gunnarsson, C. 1978. “The Gold Coast Cocoa Industry University of Oxford, Oxford, United Kingdom. http:// 1900–1939. Production, Prices and Structural Change.” www.csae.ox.ac.uk/output/reports/pdfs/rep2008-01.pdf. PhD thesis, Department of Economic History, Lund Uni- Piasentin, F. and L. Klare-Repnik. 2004. “Quality Matters.” versity, Lund, Sweden. Global Research on Cocoa and CABI, Conservation Gyimah-Brempong, K. 1987. “Scale Elasticities in Ghanaian International. Cocoa Production.” Applied Economics 19: 1383–90. Quisumbing, A. R., E. Payongayong, J. B. Aidoo, and K. Hattink, W., N. Heerink, and G. Thijssen. 1998. “Supply Otsuka. 2001. “Women’s Land Rights in the Transition to Response of Cocoa in Ghana: A Farm-Level Profit Individualized Ownership: Implications for the Manage- Function Analysis.” Journal of African Economies 7 (3): ment of Tree Resources in Western Ghana.” Economic 424–44. Development and Cultural Change 50 (1): 157–81. Herbst, J. 1993. The Politics of Reform in Ghana, 1982–1991. Rourke, B. E. 1974. “Profitability of Cocoa and Alternative Berkeley, CA: University of California Press. Crops in Eastern Region, Ghana.” In Economics of Cocoa Hill, P. 1963. The Migrant Cocoa Farmers of Southern Ghana. Production and Marketing, ed. A. Kotey, C. Okali, and A Study in Rural Capitalism. Cambridge, UK: University B. E. Rourke. Proceedings of Cocoa Economics Research Press. Conference, University of Ghana, Legon, 1973. International Cocoa Organization. Various years. “Quarterly ______. 2007a. “The Cocoa Sector: Expansion, or Green Bulletin.” London. and Double Green Revolutions?” Overseas Development Konings, P. J. J. 1986. The State and Rural Class Formation in Institute Background Note, London. Ghana: A Comparative Analysis. Leiden, Netherlands: ______. 2007b. “International Perspectives on the Cocoa African Studies Centre. Sector: Expansion or Green and Double Green Revolu- Laven, A. 2007. “Marketing Reforms in Ghana’s Cocoa Sector: tions?” http://www.odi.org.uk/events/2007/11/19/434- Partial Reforms, Partial Benefits?” Background Note, presentation-session-1-international-perspectives-cocoa- Overseas Development Institute, London. http://www.odi. sector-francois-ruf.pdf. org.uk/resources/download/420.pdf. ______. 2009. “Libéralisation, cycles politiques et cycles du Masdar Ltd. 1998. Socio-Economic Study of the Cocoa Farm- cacao: le décalage historique Côte-d’Ivoire-Ghana.” ing Community. Wokingham, United Kingdom. Cahiers Agricultures 18 (4). McKay, A., and H. Coulombe, 2003. “Selective Poverty Ruf, F., and K. Burger. 2001. “Planting and Replanting Tree Reduction in a Slow Growth Environment: Ghana in the Crops. Smallholders’ Investment Decision.” Unpublished 1990s.” Human Development Network, World Bank, paper, Centre de coopération internationale en recherche Washington, DC. agronomique pour le développement (CIRAD), Paris. McMillan, M. 1998. A Dynamic Theory of Primary Export Ruf, F., and H. Zadi. 1998. “Cocoa: From Deforestation to Taxation. Discussion Paper 98-12. Department of Eco- Reforestation.” Prepared for the First International nomics, Tufts University. Workshop on Sustainable Cocoa Growing, Smithsonian Norris, K. 2008. “Carbon, Biodiversity and Cocoa Farming Institute, Panama. in Ghana.” Climate Change Unit, Forestry Commission ______ and P. S. Siswoputranto. 1995. Cocoa Cycles: The of Ghana, Accra. Economics of Cocoa Supply. Cambridge, U.K.: Woodhead Nyanteng, V. K. 1993. “The Prospect of the Ghanaian Cocoa Publishing Ltd. Industry in the 21st Century.” Paper presented at the Shepherd, A. W., and S. Farolfi. 1999. “Export Crop Liberal- International Conference on Cocoa Economy, Bali, ization in Africa: A Review.” Agricultural Services Bul- Indonesia, October 19–22. letin 135, FAO, Rome. Obiri, B. D., G. A. Bright, M. A. McDonald, L. C. N. Anglaaere, Sjaastad, E. and D. Bromley. 1997. “Indigenous Land Rights and J. Cobbina. 2007. “Financial Analysis of Shaded Cocoa in Sub-Saharan Africa: Appropriation, Security and in Ghana.” Agroforestry Systems 71: 139–49. Investment Demand.’” World Development 25: 549–62. 216 CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY Steedman, C. 2003. “Agriculture in Ghana: Some Issues.” Vigneri M., and P. Santos. 2008. “What Does Liberalization Unpublished paper, World Bank, Washington, DC. without Price Competition Achieve? The Case of Cocoa Stryker, J. D. 1990. “Trade, Exchange Rate, and Agricultural Marketing in Rural Ghana.” IFPRI-GSSP Background Policies in Ghana.” World Bank Comparative Studies, Paper 14. International Food Policy Research Institute, World Bank, Washington, DC. Washington, DC. Takane, T. 2002. The Cocoa Farmers of Southern Ghana: Vigneri M., F. Teal, and H. Maamah. 2004. “Coping with Incentives, Institutions, and Change in Rural West Africa. Market Reforms: Winners and Losers among Ghanaian Chiba, Japan: Institute of Developing Economies and Cocoa Farmers.” Report to the Ghana Cocoa Board, Japan External Trade Organization. Accra. Teal, F., A. Zeitlin, and H. Maamah. 2006. “Ghana Cocoa World Bank. 2007a. World Development Report: Agriculture Farmers Survey 2004: Report to Ghana Cocoa Board.” for Development. Washington, DC: World Bank. Centre for the Study of African Economies, University of ______. 2007b. “Ghana: Meeting the Challenge of Acceler- Oxford, Oxford, United Kingdom. ated and Shared Growth.” Country Economic Memoran- Vigneri M. 2005. “Trade Liberalisation and Agricultural Per- dum, World Bank, Washington, DC. formance: Micro and Macro Evidence on Cash Crop Pro- ______. 2008. “Country Brief: Ghana.” World Bank, Wash- duction in Sub-Saharan Africa.” DPhil thesis, University ington DC. of Oxford, Oxford, United Kingdom. Zeitlin, A. 2006. “Market Structure and Productivity Growth ______. 2008. “Drivers of Change in Ghana’s Cocoa Sector.” in Ghanaian Cocoa Production.” Unpublished paper, IFPRI-GSSP Background Paper 13, International Food Centre for the Study of African Economies, University of Policy Research Institute, Washington, DC. Oxford, Oxford, U.K. CHAPTER 12: COCOA IN GHANA: SHAPING THE SUCCESS OF AN ECONOMY 217 CHAPTER 13 Apparel Exports in Lesotho: The State’s Role in Building Critical Mass for Competitiveness Mallika Shakya esotho, a country of 1.9 million people, has been L global apparel trade. Over time, Lesotho has started to climb quite successful in capitalizing on the African the apparel value chain and overcome issues of corporate Growth and Opportunity Act (AGOA), the eco- social responsibility. nomic partnership agreement between the United States To better understand the policy measures behind the and Africa launched in 2000 to increase African exports to success factors involving Lesotho’s apparel trade, this chap- the U.S. market. Lesotho’s experience with its apparel ter asks three main questions. First, how are spatial and exports demonstrates that competitiveness is not only economic factors balanced in pursuing a coherent compet- about firm-level costing but about developing a broader itiveness strategy in Lesotho’s apparel sector? Second, what coordination mechanism through which the government were some of the industry-specific strategies employed by delivers public goods necessary for overcoming market fail- the government of Lesotho in overcoming the government ures in such a way that private sector investment is trig- and market failures? How are they complemented with gered and sustained. economywide measures and socioeconomic measures? And A notable aspect of Lesotho’s experience is the govern- third, what public-private collaborative efforts have led to ment’s decision to combine an apparel industry competi- reforms that maximized private sector investments? How tiveness initiative with a series of early-stage incentives for have bottlenecks related to transportation, logistics, innova- investors—a tactic that has yielded positive results for tion, and quality been overcome to produce the success of Lesotho thus far. The sequencing and balancing of such pol- the apparel industry in Lesotho? icy measures has differentiated Lesotho’s apparel export In exploring these questions, this chapter takes as a basic path from those of other African countries in recent years. premise that international trade is no longer a random phe- Its proximity to South Africa and its membership in the nomenon where anonymous producers and buyers engage Southern African Customs Union offers Lesotho unique in exchange of what they produce and consume. Rather, geographic and fiscal advantages that other countries may international trade and investment decisions increasingly not have, but the industrial policy measures and institu- depend on a successful ecosystem of producers, suppliers, tional reforms undertaken by the government of Lesotho and providers of specialized services and infrastructure, are relevant for several African countries. Lesotho’s success along with cooperation from the public sector on essential can especially be attributed to its direct engagement with issues of tax, finance, innovation, and quality control. global and regional apparel buyers and an overall policy There is a growing tendency for global apparel investors, acknowledgement of the rapidly changing structure of the buyers, and producers to rely on concerted public support Mallika Shakya is Wolfson Research Fellow at the School of Interdisciplinary Area Studies, University of Oxford. 219 that harnesses new production methods, facilities, and and even middle-income countries in Sub-Saharan Africa logistical models. Beyond macroeconomic soundness, such as Mauritius and South Africa are rapidly losing their countries should also give attention to micro- and meso- comparative and competitive advantages in the apparel level issues related to delivery of necessary public goods for sector. Although both these countries exported amounts private investment. All of these factors add up to a better comparable with that of Lesotho in 2001, the situation business-enabling environment necessary for sustained changed dramatically in 2008, when Mauritius exported competitiveness. only about $100 million worth of apparel and South Africa Not only is the apparel industry an entry point for a exported a negligible amount to the United States. While broad range of light manufacturing industries but, if man- South Africa exports some of its apparel to the European aged well, it has the capacity to unleash the growth and Union (EU) and has a significant domestic market, it is competitiveness potential of an economy. Growth of the quite striking that Lesotho, a country with a very small apparel industry makes immediate contributions to domestic market, exports three times as much apparel than employment and creates tremendous potential for back- Mauritius and is decisively ahead of larger African countries ward and forward linkages that are important to developing Madagascar and Kenya. Even though the global financial countries such as Lesotho. crisis negatively affected Lesotho’s apparel exports in 2008 Lesotho exported just under $350 million worth and 2009, it is difficult to challenge the country’s success of apparel to the United States in 2008, approximately between 2001 and 2007. 29 percent of apparel exports (Edwards and Lawrence Figure 13.1 shows that Lesotho’s apparel exports to the 2009) from all of Sub-Saharan Africa to the United States American market grew exponentially in the late 1990s and for that year (figure 13.1). Although this marks a dip from early 2000s, before more or less stabilizing in recent years. A the amount exported in 2004, Lesotho still exports sub- brief slowdown due to the uncertainty in international trade stantially more apparel than its closest rivals in the sector, agreements resulted in the closure of six firms and a mod- Madagascar and Kenya. The value of Lesotho’s apparel erate compromise on exports in 2004 and 2005, but four exports to the United States is about three times higher new firms entered the market at the same time, and there than those of Swaziland, a country that has similar geopo- was a significant increase in total employment in the apparel litical conditions and that follows Lesotho’s growth path to industry. Despite the impact of the 2008/2009 global finan- a certain extent. Lesotho is also far ahead of neighboring cial crisis recovery has been relatively quick. Mozambique, which exported less than $400,000 worth of The magnitude of Lesotho’s apparel industry success apparel to the United States in 2008, approximately 0.1 per- becomes clearer when its export figures are weighed against cent of the amount Lesotho exported. the size of its population. While high population density Other African countries, such as Cameroon, Ghana, and availability of cheap labor explain the success of the and Tanzania, lag far behind Lesotho in apparel exports, apparel industry in Bangladesh, China, and India, the situa- tion is not that straightforward in Africa. As measured by apparel exports per capita, smaller countries with lower population density are outperforming larger countries with Figure 13.1 U.S. Imports of Clothing and Textiles from large labor forces (figure 13.2). Lesotho, in fact, outranks AGOA Countries other apparel-producing countries in Sub-Saharan Africa, with apparel exports per capita of $177 as of 2009. Swazi- 500 450 land and Mauritius have the second- and third-largest per 400 capita apparel exports, respectively, among Sub-Saharan 350 $ millions 300 African countries. 250 200 The apparel industry represents a significant proportion 150 of Lesotho’s gross domestic product (GDP), trade, and 100 50 employment. Clothing, in fact, has represented the majority 0 of Lesotho’s total exports since the late 1990s (figure 13.3). 19 9 90 91 19 2 19 3 19 4 95 19 6 97 19 8 99 20 0 20 1 02 20 3 20 4 05 20 6 07 08 8 9 9 9 9 9 0 0 0 0 0 19 19 19 19 19 20 20 20 20 In addition to Lesotho’s successful entry into the U.S. Lesotho Madagascar Kenya apparel market, investors are producing higher-value- Mauritius Swailand South Africa added items such as woven shirts and men’s trousers and Source: U.S. Trade Development Agency. are now targeting the regional (South African) and EU 220 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS Figure 13.2 Per Capita Apparel Exports to the Figure 13.4 Employment Trends in Lesotho United States 60,000 200 50,000 Number of workers 177 40,000 150 30,000 20,000 95 100 $ 80 10,000 0 50 00 01 02 03 04 05 06 07 20 20 20 20 20 20 20 20 15 0.1 0.2 0.7 5 7 All other Textiles and clothing Total 0 ia a i a a r s nd o w iu ca ric an ny Source: Lesotho Bureau of Statistics. th op ala ila rit as w Ke so Af hi az M au ts ag Le Et h Sw Bo ad M ut M So Source: Apparel data, Edwards and Lawrence 2009; population data, United GLOBAL AND REGIONAL TRADE AGREEMENTS Nations database. SET THE STAGE FOR APPAREL TRADE AGOA, which went into effect in October 2000, designated 34 “least developed” countries in Sub-Saharan Africa as eli- Figure 13.3 Composition of Lesotho’s Exports gible for specialized trade benefits in the United States.1 Swaziland was designated as the 35th AGOA-eligible coun- 100 try in January 2001. Under AGOA, Lesotho’s apparel exports 80 enjoy duty-free access to the American markets until 2012, making them up to 15 percent cheaper in the U.S. market 60 Percent than exports from non-Sub-Saharan African countries. And 40 because Lesotho is still categorized as a least developed country under AGOA, it is exempt from the rules of origin 20 clause, meaning that it is free to import fabrics and other 0 inputs from anywhere in the world. Because it was among the first countries to put necessary 98 99 00 01 02 03 04 05 06 07 19 19 20 20 20 20 20 20 20 20 systems into place, Lesotho was among the first African Clothing Diamonds Television sets countries to reap the benefits of AGOA. It was not until Footwear All other goods February 2002, for example, that neighboring Mozambique was technically qualified for the U.S. apparel provision. To Source: Lesotho Bureau of Statistics. qualify for the AGOA, each country must ensure the United States Trade Representative’s office that it has the adminis- trative capacity to effectively monitor against possible abuse markets. Lesotho has also emerged as one of the few African of the privileges granted under AGOA. countries where regionally grown cotton is processed into The period from 2000 to 2004 was a unique time for denim and exported as part of final apparel products to the African countries seeking to seize opportunities in the United States. global apparel market. During these years African countries Beyond economic considerations, direct employment not only received preferential treatment in the U.S. market benefits make Lesotho’s apparel industry an important part under AGOA, but the preference came at a time when of the national economy. In 2010 just under 40,000 of a total quota-based restrictions were being imposed on exports population of 1.9 million people were employed in the from large Asian countries to the United States under the apparel industry. Figure 13.4 shows the extent to which Multifibre Agreement (MFA).2 For some large Asian coun- labor conditions in the apparel sector dictate the overall tries, the MFA-imposed quota meant that they could only employment conditions in Lesotho. export less than half of what they had historically exported. CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 221 The MFA quota, however, could be borrowed from the Figure 13.5 Breakdown of Nonmaterial Costs of Apparel forthcoming year and any unused quota from one year Production in Lesotho could be carried over to the next. More importantly, quotas could be transferred between countries (that is, countries wishing to export more could purchase quotas from coun- Overhead/ tries not using them). At times the value of quota reached Labor Profit Misc utilities (49%) (10%) (20%) up to 30 percent of the original value of apparel being (20%) traded. Over time, this quota system encouraged large Asian manufacturers to relocate their operations to less- 0 20 40 60 80 100 developed countries, such as Lesotho, that could participate Percent in preferential trade arrangements with the United States. In Source: World Bank 2006. addition to the tariff-based preferences it offered, Lesotho’s historical ties to East Asian countries made it attractive as a location for apparel manufacturing. and utilities are 20 percent of such costs. Both of these costs are dependent on the presence of public goods that are nec- essary for factories’ efforts to reduce cost and enhance com- KEY FACTORS BEHIND LESOTHO’S petitiveness. COMPETITIVENESS IN APPAREL There is no set formula to explain why some countries are Factory market more competitive than others in a given industry. In general, competitiveness stems from sophisticated firms engaging A typical apparel factory in Southern Africa spends approx- in productive ventures, supported by a policy and institu- imately 20 percent of total production costs in overhead tional environment in which firms are able and encouraged and utilities, including land, electricity, and transportation. to forgo rent-seeking behavior and government and market Land ownership in Southern Africa was historically under failures are overcome. In this regard, competitiveness is a a communal system, and later a public system, hence there kind of continued filtration process through which well- is no commercial land market. Availability of land is a performing firms and markets distinguish themselves from major problem for any foreign investor seeking to start a weaker firms. business venture in Southern Africa. In Lesotho, even if Although a sound macroeconomic environment is nec- land becomes available, the cost of renting it is sometimes essary for increased competitiveness of a sector, that alone as high as in developed countries, and obtaining a permit will not create competitiveness when the basic building to construct a factory or other commercial building on the blocks of industrialization, such as having a critical mass of land takes an average of 16 months. According to the Doing firms that can develop a sustainable economic ecosystem, Business report, it takes 601 days to obtain a construction are missing, as is often the case in Sub-Saharan Africa. permit in Lesotho, compared with the Sub-Saharan African Development of a critical mass of firms, however, is subject average of 260; and 157 days in developed countries. The to public support for necessary institutional and infra- cost of obtaining a construction permit in Lesotho is 670 structural mechanisms. In the case of Lesotho’s apparel percent of per capita gross national product, compared industry, the institutional framework was created through with 56 percent in developed countries, while the unsubsi- a coordinated public-private partnership. dized rate of renting a factory shell in Lesotho is more than To answer the question of why Lesotho succeeded in twice the rate for a comparable facility in South Africa and apparel while most other African countries failed, despite other developed countries. having equally preferential market access through AGOA, a While national land ownership systems are not likely to closer examination of the costs of apparel manufacturing is be reformed overnight, it seems logical for governments to needed. Beyond having the appropriate macroeconomic fun- offer transitional arrangements for commercial land use in damentals in place, success in apparel manufacturing is the form of hassle-free leasing contracts. This point was related to labor and overhead costs, which together represent highlighted in a study commissioned by the government of the bulk of the cost of apparel manufacturing. Figure 13.5 Lesotho in 2002 to develop a comprehensive apparel indus- shows that while labor makes up as much as 49 percent of try development plan. Among the 11 principal constraints the nonmaterial costs of apparel manufacturing, overhead to the growth of the apparel industry in Lesotho identified 222 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS in the report, lack of industrial land and factory shells were Because the land tenure system in Lesotho does not pro- the top two constraints. vide for land being bought or sold, investors were offered To compensate for high land costs, the government of long-term leases within these zones and were given the Lesotho constructed industrial zones and serviced factory option to sublease if necessary. Rental procedures and reg- shells available for rent by foreign investors, allowing ulatory processes were simplified and centralized so that investors to avoid cumbersome bureaucratic processes. Two the export promotion agency, Lesotho National Develop- large industrial zones—Maputsoe on 29 hectares of land ment Corporation (LNDC), could undertake all necessary and Ha on 10 hectares—were built between 1991 and 1995. administrative procedures and cater to the business needs Additional zones were built subsequently on the of potential foreign investors. In fact, LNDC’s success as Lesotho–South Africa border areas of Mohale’s Hoek, an export promotion agency draws enormously on its suc- Mfeteng, and Butha Buthe. Currently these fully functional cess as a property developer and service provider, which industrial zones in Lesotho together house about 60 facto- catalyzed development of a critical mass of first-generation ries and employ just under 40,000 workers (figure 13.6). apparel manufacturing companies in Lesotho. After man- The industrial zones provide a range of real estate ser- aging these services in house for several years, LNDC vices to assist manufacturing and commercial enterprises, recently began subcontracting day-to-day property opera- including fully serviced industrial plots, customized factory tion and maintenance tasks to JHI Real Estate, a prominent buildings, and development of commercial properties for estate management company active in several other African leasing. The buildings are fully serviced with electricity, tele- countries. phone, water, and sewerage. All six zones can be reached by Timing was crucial to the success of industrial zone tarred road, are within reach of the public transport system, construction in Lesotho, a point that is often omitted from and have access to rail links connecting them to apparel dis- policy discussions. After AGOA became effective in 2000, tribution hubs in Johannesburg. and foreign investors operating in East and South Asia Figure 13.6 Industrial Zones in Lesotho IBRD 38607 MAY 2011 28° 29° SOUTH AFRICA Butha-Buthe INDUSTRIAL PRODUCTION ZONES BUTHA- DISTRICT CAPITALS Leribe BUTHE 29° LERIBE 29° NATIONAL CAPITAL Teyateyaneng BEREA MOKHOTLONG DISTRICT BOUNDARIES Mokhotlong MASERU INTERNATIONAL BOUNDARIES MASERU Thaba-Tseka THABA-TSEKA MAFETENG Mafeteng 30° MOHALE’S HOEK QACHA’S NEK 30° LESOTHO Qacha’s Nek Mohale’s Hoek SOUTH AFRICA QUTHING Quthing 0 50 Km. 0 30 Mi. 29° 29° Source: Government of Lesotho. CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 223 began to operate factories in Africa, and Lesotho was regulation and facilitation on various aspects of production among the few countries that already had industrial zones and sales, and management of government incentive with well-equipped and serviced factory shells. As a specific schemes and tax procedures (box 13.1). Through its Invest- response to AGOA, Lesotho offered the apparel investors ment Promotion Center (IPC), the LNDC markets Lesotho subsidized rent for at least the first five years of operation. as a destination of choice to potential manufacturers. To While the sustainability and macroeconomic rationale of achieve this, it has developed an information dissemination this policy decision may be questioned, it virtually pre- system that reaches out to potential investors within empted the efforts of other countries in the region to Lesotho, Southern Africa, and beyond. The LNDC has sent attract foreign direct investment (FDI). trade missions to South and East Asia to mobilize networks of firms that are already invested in Lesotho. One factor behind the success of Lesotho’s export and investment pro- Export and investment promotion motion is that the IPC has adopted tailor-made strategies to Observers often claim that export and investment promo- attract and sustain investments from very different tion activities have not yielded results and that they have investors, including those from South Africa; China; Tai- instead become rent-seeking activities. This is true in sev- wan, China; and elsewhere. In recent years, the IPC has also eral cases both in Africa and elsewhere in which export sought to attract yarn, trim, zipper, button, bag, hanger, and promotion agencies are captive to poor leadership due to carton manufacturers to Lesotho. To some extent, these spe- heavy government involvement, inadequate funding, cialized manufacturers can assist Lesotho in taking full cumbersome bureaucracy, and lack of client orientation. advantage of AGOA and allow a structured approach to At times, countries suffer from strong antitrade bias and inward investment. anti-FDI sentiment. By many measures, Lesotho’s success in export and investment promotion defies the general Labor coordination and skill development trend in developing countries. The LNDC is the main wing of the government charged with the implementation At 49 percent, labor is the largest nonmaterial cost in of export and investment promotion programs. Founded Lesotho. It goes without saying that labor costs directly in 1967 by an act of parliament, LNDC was amended in influence investors’ decisions about where to locate their 1990 and again in October 2000. LNDC is primarily operations and buyers’ decisions about where to choose owned by the government of Lesotho (90 percent); DEG, suppliers. In the case of Lesotho, however, labor costs are a German finance company, owns the remaining 10 per- not the primary point of consideration for investors and cent. The LNDC is part of the Ministry of Trade and buyers. In other words, investors have chosen Lesotho Industry, Cooperatives and Marketing, which is responsi- despite its relatively higher labor costs compared with ble for overall policy direction on industrialization in other Sub-Saharan African countries. Nonetheless, because Lesotho. production and sourcing dynamics are constantly chang- The LNDC works with medium and large companies in ing in the global apparel market, it is important that the manufacturing and services sectors to strengthen Lesotho takes measures to reduce labor costs while Lesotho’s exports and generate foreign investment in improving labor productivity. Lesotho. Rather than relying on a broad strategy to pro- mote all exports, the LNDC focuses on nontraditional Labor costs are high in Lesotho compared to other exports. Because of Lesotho’s extreme concentration in Sub-Saharan African countries. Hourly wages range apparel, the LNDC also has functioned as a de facto gar- from approximately $0.14 to $1.80 in apparel-producing ment promotion agency in the past several decades, devel- African countries as well as in several non-African coun- oping with it several sector-specific skills and insights. Such tries with significant apparel sectors such as Bangladesh, insights have proved instrumental in ensuring that incen- China, and Sri Lanka. Wages in Lesotho average $0.46 an tive packages offered to investors are not poorly utilized or hour, lower than those in Mauritius and South Africa but captured by rent-seekers. significantly higher than those in Ethiopia, Kenya, Mada- In addition to the LNDC’s role as real estate developer, gascar, Mozambique, and Swaziland. Even when productiv- facilitator, and service provider, it acts as a one-stop shop ity is taken into account (labor costs per shirt), Lesotho still that covers most aspects of company operations, ranging is an expensive manufacturer compared with Ethiopia, from business inquiries, site selection, firm registration, Kenya, Madagascar, and Mozambique (figure 13.7). 224 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS Box 13.1 LNDC as a One-Stop Shop Company support ■ Short- and long-term loans ■ Import value added tax (VAT) credit facility for local ■ Business registration procedures purchase of raw materials and capital goods ■ Acquisition of permits and manufacturing licenses ■ License and residency paperwork for foreign work- Tax management oversight ers, managers, and owners ■ Arrangement of site visits and assistance in selection ■ No tax on income generated from exports outside of suitable sites the Southern African Customs Union ■ Key focal point for contact with relevant ministries ■ Permanent maximum manufacturing tax rate of 10 on business regulations percent on profits ■ Facilitation of contact with business companies sup- ■ No tax on dividends to local or foreign shareholders plying services ■ Free repatriation of profits ■ Facilitation of skill and technology development ■ Double taxation agreements with Germany, Mauri- programs tius, South Africa, and the United Kingdom ■ Industrial relations if disputes arise with workers Investment attraction Management of incentive schemes ■ Information dissemination to targeted investor pools ■ Unimpeded access to foreign exchange ■ Trade missions to targeted markets and investor ■ Export finance facility pools Source: Consultations with LNDC officials. Labor unions in Lesotho have undergone a major hand, workers were upset that job stability was no longer transition. The first apparel factories were set up in guaranteed; on the other, factory owners were frustrated Lesotho by South African companies in the 1980s, and were, about low labor productivity. Fueling the antagonism were for the most part, subsidiaries of bigger industrial conglom- several cases of exploitation of workers by foreign factory erates that had specialized for decades in the South African owners and foreign supervisors. market. Primarily located in Maputsoe, these companies The Lesotho Clothing and Allied Workers Union initially established higher labor standards for Lesotho. (LECAWU) is the predominant union representing employ- They negotiated recognition agreements with the trade ees of the apparel industry. It was formed under the terms union of employees at the LNDC and the Department of of the national labor code of Lesotho in1992 and is recog- Labor. While those negotiations were frequently difficult nized by the government of Lesotho as the representative of and protracted, they inevitably resulted in pay increases for labor. Historically, anti-union sentiment has made it diffi- workers. Over the years, negotiations with trade unions cult for LECAWU to organize workers; the unwillingness of became commonplace among apparel companies in management, particularly in Maseru and Thetsane, to Lesotho, even among those where the majority of employ- engage in meaningful negotiation also impeded the union. ees were not union members. Nevertheless, LECAWU membership has grown in recent The new generation of apparel manufacturers and buy- years as a result of AGOA provisions, changes in apparel ers that moved to Lesotho in response to the MFA and companies’ codes of conduct, and increasing interest in AGOA provisions created a demand for a very different set labor relations among company owners and managers. The of labor skills and expectations. In short, the industry labor movement in Lesotho encouraged the government to became less regulated by local labor codes and more influ- create the legislative framework that set up the Industrial enced by the global momentum around labor-related cor- Relations Council, the Directorate for Dispute Prevention porate social responsibility. The apparel industry in Lesotho and Resolution, and the Labor Appeal Court. This impor- underwent a major labor crisis in the 2000s as a result, one tant piece of legislation provides for the settling of labor dis- that involved policy deadlocks and street riots. On one putes through mediation, conciliation, and arbitration, CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 225 Figure 13.7 a. Factory-Floor Productivity and Labor Costs in Apparel Assembly; b. Hourly Wages for Apparel Assembly (US cents) a b Labor Ethiopia 14 Daily cost per Ghana 18 Country productivity shirt ($) Bangladesh 21 Mozambiquea 25 Mozambiquea 10–11 .16 Pakistan 26 Ghana 12 .12 Madagascar 26 Ethiopia 10–12 .12 India 27 Kenya 12–15 .18 Kenya 30 Madagascar 14–15 .16 Nepal 31 Lesotho 18 .19 Swaziland 33 South Africa 15 .65 Sri Lanka 35 India 16 .17 Lesotho 46 Chinab 18–22 .29 EPZ China 72 Mauritius 125 South Africa 179.4 Source: Labor productivity figures and costs are from Eifert, Gelb, and Ramachandran (2005) except for those for Ethiopia, which were calculated using Global Development Solutions (2005). Hourly wages were collected from government sources. Note: Daily productivity is the number of men’s casual shirts made by a machine operator in one day. a. Mozambique wage indicates the minimum industry wage specified under the national labor law. b. Data for China is for export processing zones only. effectively creating an independent authority that is repre- social accountability much like the ISO9000 has done for sentative of the government, employees, and employers. quality. Meanwhile, several multinational apparel manufac- The global labor movement has benefited the local labor turing companies have formed the Worldwide Responsible movement in Lesotho. As manufacturing has shifted from Accredited Production (WRAP) principles to accredit sup- developed to developing countries, there has been growing pliers. In addition, all major multinational apparel manu- concern among developed-country consumers, as well as facturing companies have their own codes of conduct. from activists and nongovernmental organizations (NGOs), The need to comply with the various codes of conduct regarding the wages and work conditions under which prod- has created a major change in most apparel factories in ucts are manufactured. As a result, most major multinational Lesotho. And though the inspections surrounding such companies have adopted codes of conduct they require their codes of conduct can measure a degree of compliance, suppliers to meet. In addition, a number of NGOs have industrial relations will be smooth and sustainable in the devised codes of conduct in partnership with the multina- long term only if they contribute to productivity. tionals: these include the U.K.-based Ethical Trading Initia- tive (sponsored, in part, by the Department for International Labor productivity must increase. A typical apparel Development, or DFID), the Institute of Social and Ethical factory now works with its workers, regulators, and service Accountability, the Global Reporting Institute, the Nether- providers as a group to drive down costs and improve lands-based Clean Clothes Campaign, the U.S.-based Fair standards. By working with providers of training and Labor Association, the U.S.-based Worker Rights Consor- skills, factories are able to shift to new production meth- tium, and the EU’s European Code of Conduct. A major ods that reduce costs, improve product quality, and private initiative called Social Accountability 8000 enhance factories’ commitment to their workers. Nike’s (SA8000), along with the nonprofit organization Account- shift from a normal assembly line production system to Ability’s AA1000, are attempting to create a benchmark for “lean manufacturing,” for example, has been marked with 226 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS significantly higher productivity and better labor condi- of machinists, its primary aim of launching entrepreneurs tions. The initiative has been followed by the incorporation into the formal garment subsector was not successful, and of more sophisticated production planning and merchan- it was terminated in 2003. Lesotho’s experience with the dising systems, such as “Just in Time,” that aim to eliminate LGC is similar to that of several other developing coun- inefficiencies in space, time, and activity and also manage tries, where the public sector has developed channels to the inventory stock. support the private sector that turn out to be ineffective. Most buyers of apparel assess their producers on four key Coordination and incentive structures have been major criteria: cost, quality, timeliness, and corporate responsibility problems in making public sector efforts useful for the pri- (that is, labor and environmental standards) (figure 13.8). vate sector. After an initial baseline is developed, buyers regularly mea- No other training institutes have been created to fill the sure factories’ performance to track delivery, quality, and void left by the termination of the LGC. Several other train- prices over time. Indicators are typically expected to improve ing establishments, such as Lerotholi Polytechnic and the each quarter. Such proactive measures toward labor produc- Institute of Development Management, offer a range of tivity are not yet under way in Lesotho, although a number generic management and general studies, although none are of initial steps have been taken. The fact that Lesotho is at the tailored to the needs of the formal garment sector. Lerotholi lower end of productivity in apparel manufacturing means Polytechnic offers a two-year series of courses related to the that it needs to develop systematic analyses to diagnose bot- apparel industry, for example, but the focus is on design, tlenecks and then design institution measures to tackle those pattern construction, and bespoke tailoring. The Commer- bottlenecks accordingly. cial Training Institute offers similar courses that lead to the award of a certificate in tailoring and dressmaking. The Successful skill development initiatives. In 1999 Basotho Enterprise Development Corporation (BEDCO) the Lesotho Garment Center (LGC) was opened to provide also runs a series of short courses, including marketing skills necessary training for the apparel workforce. Developed and, actively seeks to identify and focus on the future needs with funding from the DFID, the LGC had the objective of the garment industry. Saint Luke’s Mission in Maputsoe of training workers and entrepreneurs in the manage- offers a three-year course providing students with a range of ment of industrial apparel production in anticipation of apparel industry skills, including pattern construction, their participation in the opportunities presented by machine knitting, small business studies, leadership, man- AGOA. Although the LGC succeeded in training hundreds agement skills, and industrial garment production on the appropriate machinery. Upon completion of these courses, many students are Figure 13.8 Buyer’s Scorecard for Assessment of employed in the local craft industry or their own enter- Producer Competitiveness prises, but very few graduates from these institutes end up working on the apparel shop floors. The training institutes are seldom approached by apparel manufacturers to train Gross margin Quality n Product defective rate managers, workers, or technicians, underscoring the need n Labor productivity against target 0.9% for sector-specific contents and modules in training insti- n Scale n Vendor test submit rate tutes for them to be truly useful to employers. This scenario against target 10.0% for use n Locational advantage of restricted substance in highlights a larger problem in the apparel industry in n Connectivity production, finishing and Lesotho: while existing training institutes do not cater to the packaging apparel industry, apparel factories report that their in-house Corporate training facilities are not sufficient. For the most part, Delivery responsibility trainees in factories are taught single machine skills rather GAC: Goods at consolidator, ‘C’ rated factories: lack of than being cross-trained. the date and the product basic terms of employment leaves the factory and legal non-compliance Most apparel factories in Lesotho also acknowledge the OTP: On time receipt to the ‘D’ rated factories: unwillingness need for trained supervisors and indicate that they are final port after dispatching, to comply with CR standards and denial of access to prepared to pay for it. In the factory survey, 69 percent of loading and shipping inspectors interviewees confirmed that they would support training programs that addressed their particular sector needs if such Source: World Bank 2006. programs were delivered in a way that minimized employees’ CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 227 time away from the workplace. Approximately 11 percent Lesotho are owned and operated by Spoornet, the South said they might support such a facility, while 20 percent said African rail service. they would not. Only one of the companies surveyed had Customs processes in Lesotho have significantly improved actually developed a supervisory training course, and is since AGOA took effect. The government has put in place a using its own staff to deliver the training. The course covers rigorous training and regulation system among the customs a range of topics, from organization and planning to indus- staff to meet the requirements of exports under the AGOA trial relations and health and safety. The most common rea- visa scheme and now offers duty rebates on fabrics imported sons companies gave for developing their own courses were in the production of apparel exports. Further, a duty credit that no such courses were available locally and that they had scheme was put in place for exporters of manufactured goods concerns about quality of delivery and training materials that use raw materials sourced from Southern African Cus- provided by third parties. toms Union countries. For certain factories, customs requires The fact that services offered by existing training insti- closer inspection of waste and offcuts disposal before rebates tutes are not used by apparel factories while the same on imported materials are issued, though such inspections factories acknowledge need for more effective training are expedited after factories establish a good track record or indicates a coordination gap. While an initiative is under after in-factory inspections are streamlined with other regu- way in Lesotho to develop broader public-private coordi- latory measures. Such multiple inspection requirements nation on skill development, it is too early to comment on necessitate an efficient, well-trained customs team to ensure its success. Among the commendable initiatives in this smooth implementation. direction has been the development of an interactive data- base that would collate information on leading global and LESSONS FROM LESOTHO’S APPAREL regional experts in specific areas. Depending on the inter- INDUSTRY EXPERIENCE est from the private sector, selected experts would be invited to develop training modules that would feed into Integration into global supply chains has been an important this interactive database. Eventually, private sector partici- avenue for growth and competitiveness for Lesotho’s apparel pants would be expected to use those modules to commis- industry. Geographical advantages and disadvantages—of sion training at their own expense. The success of such an weather, location, and connectivity—cannot be ignored, but initiative will depend on the degree of collaboration among discovery and pursuit of new opportunities through correct various stakeholders. A breakthrough in this area will be formulation of policies and development of effective public- necessary for Lesotho to climb up the global apparel qual- private partnerships has been the key channel through ity value chain. which Lesotho has achieved apparel industry success. Sev- eral key features of this success story can be identified. First, the government of Lesotho was successful in Transport, logistics, and customs attracting foreign investors through its policy of construct- The time and cost of transport is often an important barrier ing industrial zones and serviced factory shells. Throughout to competitiveness in the apparel industry. A smooth trans- this process, the government has remained cognizant of the port corridor supported by competent logistical service geographical implications of such facilities, ensuring that providers and punctuated by efficient border crossing mech- they are not far from the public goods necessary for indus- anisms is a prerequisite for any exporting economy. Thanks trialization. Construction and effective management of the to the critical mass of apparel producers that regularly use factory shells compensated for inherent inefficiencies in the the corridor between Maseru and Johannesburg, Lesotho has land market in the form of prohibitions on private owner- made enormous strides in developing systems for transport, ship and sales-based transactions of land. customs, and logistics. Second, when AGOA-related opportunities became The road infrastructure in Lesotho is substantially devel- available, Lesotho made stronger, clearer efforts to attract oped and is well connected to the extensive road networks FDI than other countries in Southern Africa. Lesotho not of South Africa. Trucking services are provided by private only moved quickly to administer AGOA in country but operators that cater to a wide range of apparel producers. also offered generous tax and financial incentives to early Although there is only one railhead in Lesotho, it is located investors to help develop a critical mass of apparel manu- in Maseru, just across the river border with South Africa. facturers that then made way for service providers and reg- Railway lines, station buildings, and handling facilities in ulatory bodies to join in. 228 CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS Third, Lesotho’s export and investment promotion Eifert, B., A. Gelb, and V. Ramachandran. 2005. “Business efforts would not have been effective without the LNDC, Environment and Comparative Advantage in Africa: which pursued promotion measures through appropriate Evidence from the Investment Climate Data.” Working information dissemination models and targeted interest Paper, Center for Global Development, Washington, solicitation. Once a critical mass of apparel investors came DC. to Lesotho, the LNDC developed an effective one-stop shop DFID. 2006. “Lesotho Garment Industry Subsector Study” for business development, firm registration, financing, and conducted for the government of Lesotho. state-industry coordination on issues such as tax manage- EIU (Economist Intelligence Unit). 2007. “Country Profile: ment and access to state industrialization incentives. Mozambique.” EIU, London. Fourth, when labor disputes arose around the industrial FIAS (Foreign Investment and Advisory Services). 2006. transition, the government of Lesotho quickly stepped in to “Market Diversification of the Lesotho Garment Indus- mediate among multiple stakeholders who were nurturing try.” Discussion draft, Innernational Finance Corpora- mutual antagonism. tion and World Bank, Washington, DC. And fifth, although it is too early to make informed ver- Global Development Solutions. 2005. “Value Chain Analysis dicts, the government of Lesotho has acknowledged its past for Strategic Sectors in Mozambique.” Reston, VA. mistakes on skill development and is developing a coordi- Horton, C. 1996. “Capital Flows from South Africa into nated skill development initiative in conjunction with the Neighboring Countries: The Case of Textiles and Cloth- private sector. If successful, the initiative will compensate ing.” Naledi papers. National Labour and Economic Development Institute, Johannesburg. Lesotho for its competitive disadvantage in labor costs. Kaplinsky, R., D. McCormick, and M. Morris. 2006. “Dan- gling by a Thread: How Sharp Are the Chinese Scissors?” NOTES Institute of Development Studies, University of Sussex, 1. As per the definition as used in AGOA, a least developed Brighton, United Kingdom. country in Sub-Saharan Africa is a country with gross Minor, P. 2005. “Survey of US Apparel Buyers: Sourching national product (GNP) per capita of less than $1,500 in 1998. from Sub-Saharan Africa in the Post-Quota MFA.” This provision was renewed in 2006 for the period 2007–12. Unpublished paper, Nathan Associates Inc. Arlington, 2. The Multifibre Arrangement governed textile and cloth- VA. ing quotas between 1974 and the end of 1994. Upon forma- ———. 2007. “Developing a Cotton Textile and Apparel tion of the World Trade Organization in 1995, textile Value Chain in Mozambique.” Nathan Associates Inc, exports became governed by the Agreement on Textiles and Arlington, VA. Clothing. Morris, M., and L. Sedowski. 2006. “Report on Government Responses to New Post-MFA Realities in Lesotho.” PRISM Working Paper, Centre for Social Science BIBLIOGRAPHY Research, University of Cape Town. Edwards, L., and R. Lawrence. 2009. “Lesotho’s Export Per- World Bank. 2006. “Lesotho Country Economic Memoran- formance: An African Successful Story?” National Bureau dum: Growth and Employment Options Study.” World for Economic Research, Cambridge, MA. Bank, Washington, DC. CHAPTER 13: APPAREL EXPORTS IN LESOTHO: THE STATE’S ROLE IN BUILDING CRITICAL MASS FOR COMPETITIVENESS 229 CHAPTER 14 The Success of Tourism in Rwanda: Gorillas and More Hannah Nielsen and Anna Spenceley wanda is well known for its mountain gorillas. First R on high-end tourism while maintaining conservation and brought to international attention by the conserva- contributing to poverty reduction through the involvement tion efforts of Dian Fossey in the 1960s and 1970s, of communities. Rwanda’s gorillas have been featured in numerous documen- Besides the VNP, Rwanda has two other national parks taries and have been visited by well-known figures such as Bill that offer a range of wildlife and biodiversity. Furthermore, Gates, Natalie Portman, and Ted Turner, all of whom have the country has been particularly successful in attracting participated in Rwanda’s annual gorilla-naming ceremony. large numbers of business and conference travelers, mainly Rwanda and Uganda are currently the only two coun- from the Democratic Republic of Congo and other neigh- tries in the world where mountain gorillas can be visited boring countries of the East African Community (EAC).1 safely, and the number of tourists visiting the Volcanoes This success is evidenced by the large increase in the number National Park (VNP) has increased dramatically since the of hotel rooms, restaurants, and the planned construction of end of the war. Rwanda also views gorilla tourism as a valu- a convention center. Local and foreign direct investments able conservation tool, and as such enforces strict rules for have been substantial, accounting for 16 and 20 percent of the habituation of, and trekking with, gorilla families. total local and foreign direct investment, respectively, over Tourists are willing to pay high fees for a limited number of the last 10 years. In terms of export revenue, tourism already permits, which are usually sold out. Revenues from gorilla outperforms coffee and tea by a wide margin. tourism provide funds to national parks and facilitate con- Several key characteristics have contributed to the suc- servation activities. cessful revival of the tourism sector in Rwanda. First and Although Rwanda is known for its violent past, interna- foremost, the government has shown a clear commitment tional perception of the country is shifting. As of 2010 to the development of tourism and has established Rwanda Rwanda is considered one of the safest destinations in East as a safe destination in the region. The early development Africa. This rebranding goes hand in hand with the market- of a strategy and policy demonstrated this commitment. ing of the country and, in particular, the mountain gorillas. Furthermore, the government involved the private sector The revival of gorilla tourism demonstrates that with the from the start and has implemented policies that enhanced right strategy, a postconflict country can successfully focus the business environment and promoted private sector Hannah Nielsen is an economist in the Poverty Reduction and Economic Management Network of the Africa Region of the World Bank. Anna Spenceley, PhD, is an independent consultant based in South Africa, and formerly a senior tourism advisor at SNV in Rwanda. 231 investment in tourism, thereby marketing Rwanda as a des- the first attempts at habituation for this purpose occur- tination. The business environment has improved ring as early as 1966 (Murnyak 1981). These early tourism markedly in recent years, promoting private sector involve- programs displayed an almost complete lack of structure ment in tourism. In addition, Rwanda has always seen and control. The focus was on revenue rather than con- tourism as an instrument to reduce poverty, for example by servation, and there are many anecdotal reports of large directly involving local communities. groups of tourists visiting groups of nonhabituated or semihabituated gorillas (Fawcett, Hodgkinson, and THE SUCCESS OF GORILLA TOURISM Mehlman 2004). In 1979 the Virunga region’s first official mountain Background: How did Rwanda start to develop gorilla tourism program was launched by Bill Webber and gorilla tourism? Amy Vedder with funding from the African Wildlife Foun- The Virunga mountain gorilla (Gorilla beringei beringei) is dation, World Wide Fund for Nature, and Fauna and Flora a highly endangered African ape subspecies, with a total International (Bush 2009). It was one part of the three- estimated population of 380, that exists only in the part approach of the Mountain Gorilla Project, which also Virunga Conservation Area encompassing Rwanda, the included antipoaching and education programs. The Democratic Republic of Congo, and Uganda(figure 14.1). gorilla tourism program had a dual purpose: providing the The distribution of the Virunga mountain gorillas is lim- Rwandan government and park authorities an incentive to ited to an approximate area of 447 square kilometers, conserve the VNP and the animals within it from the which encompasses the Mgahinga Gorilla National Park in threat of proposed conversion of 5,000 hectares of the Uganda, the VNP in Rwanda, and the Mikeno sector of the VNP for agricultural purposes; and generating local Parc National des Virunga of the Democratic Republic of employment and tourism-related revenue (Weber 1982, Congo (Gray et al. 2005). The VNP consists of about 1985; Vedder and Weber 1990). The program subsequently 160 square kilometers of montane forest. Until Rwanda’s evolved into what is now the International Gorilla Conser- independence in 1962, the VNP was part of Africa’s first vation Program, still organized as a coalition of the three national park, the Parc National Albert, which was created agencies (Bush 2009). in 1925 with the intention of protecting the great apes Two wild groups of gorillas were initially habituated for (ORTPN 2004). tourism visitation purposes, with strictly enforced limits on Tours to view wild mountain gorilla groups have been the number of visitors and length of visits (box 14.1). The organized since 1955 (Butynaski and Kalina 1997), with combination of quality control and international interest in Figure 14.1 Area of Distribution of the Mountain Gorillas IBRD 38552 MAY 2011 DEM. REP. OF CONGO PARC NATIONAL UGANDA DES VOLCANS Gahinga, 3474 m. Muhavura, ORTPN H.Q. 4127 m. Visoke, 3711 m. Karisimbi, Kinigi 4507 m. PARC NATIONAL DES VOLCANS Ruhengeri R WA N D A 0 5 10 Kilometers Source: Bush and Fawcett 2007. 232 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE Box 14.1 Current Rules Controlling Gorilla Tourism in Rwanda A number of rules designed to protect both gorillas and ■ Tourists must remain together in a tight group tourists have been established, as follows: ■ No loud noises or pointing ■ Eating, drinking, and smoking are not permitted ■ Maintenance of a distance of 7 meters between the within 200 meters of the gorillas tourists and the gorillas ■ Tourists must turn away and cover their mouths ■ A maximum of eight tourists per visit when coughing and sneezing ■ A limit of one tourist group per day to each gorilla ■ Human feces must be buried in a hole of a minimum group depth of 30 centimeters ■ A limit of one hour per visit ■ No trash may be deposited in the park ■ Tourists who are visibly unwell or declare themselves ■ Tourists are not allowed to clear away vegetation to to be ill may not visit get a better view As of 1989 the number of tourists allowed in a single visit was increased to eight people (six people for These rules (adapted from Litchfield 1997) were smaller gorilla groups). In 1999 the required separation designed and set to minimize behavioral disturbance distance between tourists and gorillas (to reduce the and disease transmission to the gorillas from tourists. risk of disease transmission) was increased from 5 to Although the welfare of the gorillas has always been the 7 meters. Other rules have been added over time: primary concern, the majority of these regulations were created based on expert opinions rather than specific ■ Minimum age of 15 years for tourists research findings. ■ No flash photography Source: Fawcett, Hodgkinson, and Mehlman 2004; Homesy 1999. Dian Fossey’s highly publicized gorilla studies resulted in be visited safely is Uganda. With a broad client base2 and a steadily increased visitation throughout the 1980s, peaking limited number of permits (around 17,000 per year), around 6,900 in 1989 (ORTPN 2008b). By the mid-1980s demand is higher than availability of permits. The accessi- local attitudes toward and political support for conservation bility of the gorillas is another advantage. Because of increased significantly as a direct result of this program Rwanda’s small size, tourists can reach the gorillas in two (Weber 1987). Stimulated by the attraction of gorilla hours from Kigali; by comparison, it takes six hours to reach tourism, Rwanda received almost 22,000 visits to its three the gorillas from Kampala in Uganda. In addition, the con- national parks in 1990 (Bush, Hanley, and Colombo 2008), dition of infrastructure in Rwanda, especially roads, is rela- before conflict brought tourism to a halt. tively good compared with that of its peers. Since the VNP reopened in 1999, visitation has Besides viable tourism assets and relatively good infra- rebounded from 417 visits that year to nearly 20,000 visits structure, Rwanda has shown a strong commitment to pro- in 2008 (of which 17,000 were to see the mountain gorillas) moting the tourism sector. It has developed a clear tourism (ORTPN 2008b). Gorilla visitation shows some amount of strategy, marketed the destination Rwanda successfully, seasonality, with a peak in the number of permits sold involved the private sector in the policy dialogue, and gen- between June and September (figure 14.2). erally improved the country’s business environment. Overall strategy and vision. Between 1994 and 2001 Factors contributing to the success of the Rwandan government worked to establish a tourism- gorilla tourism friendly environment. The first meetings with the private A number of factors have contributed to the success of sector on the development of the tourism sector were held gorilla tourism in Rwanda. A prerequisite is the relative ease in 1999. From 2000 onward Rwanda participated in major of habituating mountain gorillas to the presence of humans, tourism fairs, and in late 2001 the Tourism Working Group, facilitated by the temperate climate and benign habitat. The which included the public and the private sectors, was only other country where mountain gorillas can currently established. The Rwanda Tourism Strategy was developed CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 233 Figure 14.2 Gorilla Trekking in Volcanoes National Park a. Park activities in VNP, by product, 2008 b. Gorilla trekking permits by month, 2008 Golden 2,000 monkey 1,800 7% 1,600 1,400 Gorilla 1,200 85% Mountain climbing 1,000 5% 800 600 Other 400 3% 200 0 Fe ry ry ch ril ay ne ly st r ov er r r be be be Ju gu ua ua Ap M ob ar Ju em em em Au Jan br M ct pt ec O Se D N Source: ORTPN 2008b. and approved by the Cabinet in 2002. A revised Rwanda incentives, conferences, and exhibitions) tourism and for Tourism Strategy (“Sustaining the Momentum”) was elabo- birding activities (OTF Group 2008a, 2008b). rated in 2007. A National Tourism Policy was put in place in Although the different strategies have not yet been 2006, a revision of which is currently under way. With sup- implemented completely, the Rwandan government has port from the United Nations World Tourism Organization, consistently demonstrated strong commitment to the exe- the government of Rwanda has also prepared a 10-year Sus- cution of tourism-related reforms and to the overall tainable Tourism Development Master Plan (Republic of improvement of the performance of the sector. Despite the Rwanda 2002, 2007b, 2009a, and 2009b). limited number of staff in the Rwanda Office of Tourism Rwanda’s overall strategic vision is to focus on high-end and National Parks (ORPTN), its leadership has effectively ecotourism rather than mass tourism. In the first Rwanda led the advancement of reforms. Further, Rwanda has Tourism Strategy, three core market segments were identi- learned from the experience of other countries. Study tours fied: ecotravelers, explorers, and business travelers. The tar- have been undertaken to Kenya and Mauritius to learn from gets set in that document were soon surpassed, however, the tourism development strategies of those countries. mainly through the success of the gorilla product. The revised strategy in 2007 identified the primates as Rwanda’s Marketing. Following the passing of formal tourism plans unique selling proposition but recognized the need to diver- in Rwanda, a national campaign was launched to improve sify the tourism sector and identified international confer- the image of tourism in the country. (The word for tourism ences and birding as two additional core segments. Tourism in Kinyarwanda, the local language, means “wandering receipts are already higher than the targets set for 2012 in around aimlessly” and has therefore a negative connotation.) the revised strategy. In the latest tourism policy, objectives A media campaign was launched to sensitize the population are set within the framework of other national strategic and convey that the country can benefit from tourists and documents, such as the Vision 2020 and the Economic should therefore welcome foreigners. Simultaneously, Development and Poverty Reduction Strategy. Rwanda has worked to improve its image on an interna- The Sustainable Tourism Development Master Plan for tional basis. In the late 1990s international perceptions of Rwanda consolidates previous strategies and policies, gives Rwanda were primarily associated with the genocide. Even clear and detailed recommendations, and sets ambitious still in 2002 market research conducted in neighboring targets. Tourist arrivals to Rwanda are projected to increase countries showed that more than half of international visi- from about 980,000 in 2008 to more than 2 million in 2020, tors believed that Rwanda was an unsafe destination with an expected increase in foreign exchange earnings (Grosspietsch 2006). Despite concerns about the safety situ- from about $200 million to more than $600 million. Sepa- ation, however, surveys carried out in 2003 showed that rate strategies are being developed for MICE (meetings, the satisfaction level of visitors in terms of safety and 234 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE stability was very high after they had visited the country tourism in Rwanda. As a result, there is now a strong public- (Grosspietsch 2006). private dialogue surrounding tourism in Rwanda, and a Specific public relations and marketing efforts pursued tourism working group composed of private and public by Rwanda include contracting international public rela- stakeholders in the tourism industry is now in place. The tions and marketing agencies in the United Kingdom and private sector is consulted in the development of new United States and launching a new Web site in 2003. In policies and strategies, such as the Sustainable Tourism addition, Rwanda has been featured extensively in docu- Development Master Plan. In addition, the private sector mentaries on international television channels and has federation, of which the tourism chamber is a member, is received positive coverage in more than 350 credible inter- consulted before new strategies and laws are adopted. national press publications, as well as in major travel Rwanda has implemented a number of market-based guides. Rwanda has also represented itself well at major reforms to strengthen the role of the private sector in tourism fairs since 2000, earning first prize for the best tourism. Several important laws and codes have been African stand at the International Tourism Bourse in Berlin revised, including the investment code, company law, secure for three consecutive years, 2007–09, and at the World transactions law, labor law, and insolvency law. The new Travel Market in London in 2009. ORTPN’s financing of insolvency law facilitates the access to finance, allowing mov- travel fees for several Rwandan tour operators to these ables, such as livestock, to be used as guarantee. Customs pro- trade fairs has contributed to cooperation between the gov- cedures are also being simplified. A pilot, one-stop-window ernment and the private sector on issues related to tourism. was successfully launched at one border crossing, and plans To foster the interest of tour operators and travel agents in are in the works to replicate it at other border posts. Rwanda, the government has also organized familiarization Rwanda’s business environment has also improved sub- tours for international investors and tour operators, during stantially in recent years. A one-stop-window has been which the Minister of Commerce and ORTPN have introduced to register a business, and the administrative received the delegates to demonstrate the importance of costs of registering a business have been lowered. It is now their visit. possible to register a business within one day for a flat fee of Rwanda’s annual gorilla naming ceremony (Kwita Izina), RF 25,000 ($43). Rwanda’s success in this area has been doc- launched in 2005, during which mountain gorillas born in umented by a substantial improvement in the World Bank’s the previous 12 months are named, has attracted a number Doing Business indicators: Rwanda was named the top per- of international celebrities. The baby gorillas have been former in 2009. Rwanda outperforms all other countries in named, among others, by the president of Rwanda and his the EAC in the rankings and has shown a strong commit- wife, ambassadors, Hollywood stars, international conserva- ment to further improving private sector conditions, partic- tionists, and performing artists. The ceremonies provide a ularly in the tourism sector. good platform to promote Rwanda as a destination and the A number of tourism sector–related incentives are need for efforts to protect gorillas and conserve their habi- offered to investors. According to the investment code, tax tat. The ceremony is now accompanied by several other exemptions are granted to investors who invest $100,000 or events, including a cross-country cycling tour and a conser- more in a facility. Airplanes imported to transport tourists vation conference. Thanks partly to the awareness of the are tax exempt, and specialized vehicles such as hotel shut- need to protect the gorillas that the gorilla naming cere- tles are exempt from import and excise duties. An investor mony and gorilla tourism in general have brought about, in the tourism and hotel industry is also exempt from pay- poaching has been significantly reduced and the number of ment of import duties on equipment such as bedroom fit- gorillas has increased steadily. tings, swimming pools, and outdoor leisure equipment. Improved business environment and involvement of BENEFITS OF GORILLA TOURISM the private sector. Initially, promotion of Rwanda’s Implications for communities tourism sector was almost entirely driven and implemented by the government. The private sector lacked the capacity Bush, Hanley, and Colombo (2008) note that the Virunga and funding and was not well organized. The government, mountain gorilla represents an isolated island population in however, made efforts to involve the private sector from the an upland area surrounded by a sea of humanity at some of start with the long-term objective that the private sector the highest densities found on the African continent (in some would take over as the driving force for encouraging areas of Rwanda, population density reaches 820 people per CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 235 square kilometer), and much of the land surrounding areas overseen a revenue-sharing scheme whereby 5 percent where gorillas live is inhabited by extremely poor people who of tourism revenues from VNP fees are injected into derive their incomes from agricultural activities. Gorillas are local community projects around the national park to severely threatened by anthropogenic disturbances such as ensure that the local people feel some ownership of the agricultural conversion and illegal extraction of resources (for parks (box 14.2). Although it is not known what pro- example, snare setting for smaller mammals that entrap portion of the budget of local councils is represented by young gorillas). While gorillas are no longer hunted for their the shared revenue, it is clear that local governments meat in the Virunga range, they are the focus of illegal animal must be actively involved in selecting local projects to trafficking. Members of gorilla groups are killed and finance. wounded in an effort to trap infants for the black market, ■ Employment opportunities are offered through sometimes leading to the disintegration of groups. This hunt- national parks: guides, trackers, and antipoaching agents, ing pressure currently represents the greatest threat to the for example. Some of the private tour operators also offer survival of the mountain gorillas and the integrity of their community-based tourism activities, such as stays with habitat. Illegal hunting is mainly motivated by meeting sub- local families, village walks, banana beer production, and sistence needs for the poorest people around the VNP even volunteer opportunities in local communities. (Plumptre et al. 2004). To address local welfare needs to mitigate some of these Disbursement of funds to communities. Since 2005 poverty-related conservation threats, a key focus of contem- nearly $428,248 has been directly invested in community porary conservation strategies is on local communities projects and used to empower communities. The total (Hulme and Murphee 2001). Combining conservation with amount, however, equates to an investment of only $1.45 local development through integrated conservation and devel- per person since the program’s inception, or an average of opment projects is now a standard approach in many devel- $0.36 per person per year. Projects for which funds have oping countries (Barrett and Arcese 1995). Rwandan commu- been used include education, environmental protection nities are involved in gorilla tourism in the following ways: (tree planting, soil erosion control, and fencing in protected areas to limit access by poachers), food security, basic infra- ■ Creation of a department for community conservation structure, and water and sanitation (figure 14.3). Specific to work on local education and social infrastructure community projects have included construction of schools, projects (Uwingeli 2009). water tanks, and hospitals; basket weaving; establishment of ■ Revenue sharing: Since 2005 ORTPN (which was culture centers; potato farming; tree planting; bee-keeping; absorbed into the Rwanda Development Board, or RDB, milk cooler construction; goat rearing; and mushroom and in early 2009), with the support of the government, has pepper farming. Education projects have received the most Box 14.2 Process for Disbursing Community Funds Five percent of tourism revenues from the protected Sustainability of projects (gauged through the eco- areas in Rwanda are put into a fund for community nomical, social, and environmental indicators stated projects in administrative sectors that neighbor in the proposal and their likelihood of being achieved) national parks. The Rwanda Development Board and the proportion of community contribution are issues calls for proposals, and a project selection also considered. process is completed at the sector and district levels. Once projects are selected, contracts are signed with Selection criteria include positive impacts on local the district authority and the community. Contracts communities and on conservation of biodiversity in cover lengths of time that depend on the project com- protected areas. Areas that register a large number of plexity and can vary from 1 month to 15 months. The cases of conflict between protected areas and the com- community is often grouped into cooperatives or munity, according to the results of a ranger-based direct-specific target groups when their ownership and monitoring system, have preferential access to funds, level of organization guarantee effective implementa- as do areas that are located close to the protected areas. tion of the project. Source: Tusabe and Habyalimana 2010. 236 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE Figure 14.3 Funds Disbursed to Community Projects around the VNP, by Sector, 2005–08 Water $8,772 Infrastructure $15,789 Types of projects Food security $40,350 Environment protection $42,105 Education $321,232 $0 0 00 00 00 00 00 00 00 ,0 ,0 ,0 ,0 ,0 ,0 0, 00 50 00 50 00 50 $5 $1 $1 $2 $2 $3 $3 Value in US$ Source: Télesphore 2009. Figure 14.4 Funds Disbursed to Community Projects funds because of the high priority education is given within around the VNP, 2005–08 the sectors in the Musanze district, near the VNP. The annual amount disbursed is directly correlated to tourism 140,000 revenues collected in the previous year. The amount of funds disbursed to communities in each year between 2005 120,000 and 2008 is shown in figure 14.4. 100,000 By 2008 seven districts bordering parks in Rwanda with U.S. dollars 80,000 a population of almost 300,000 people had been reached by the community project financing scheme (Bush 2009). 60,000 Although no formal study has yet been carried out to assess 40,000 the impact of the scheme on the livelihoods of people living near the VNP, the RDB and local authorities indicate that 20,000 the scheme has contributed to an increased awareness of 0 tourism benefits to the community and to the need to pro- 2005 2006 2007 2008 tect biodiversity in the VNP (Spenceley et al. 2010). Source: Télesphore 2009. One specific project that has benefited from community fund financing is the high-end Sabyinyo Silverback Lodge. The eight-room lodge located at the periphery of the VNP and operates the lodge, and pays the communities a $50 is a joint venture of the local Kinigi and Nyange communi- occupied bed night fee and also 7.5 percent of net sales ties (represented by the Sabyinyo Community Livelihoods (Makambo 2009). Association, or SACOLA); the private sector (Governors This joint venture operation allows people who live Camps Ltd); international NGOs, in particular, the Interna- close to the VNP to benefit from tourism in four main tional Gorilla Conservation Program (IGCP) and the ways: equity in a tourism business, employment opportu- African Wildlife Foundation (AWF); and the RDB. Planning nities at the lodge, the supply of goods and services, and for the lodge began in 2004, and the first tourists began to dividends from profits. The lodge employs 45 local people, arrive in August 2007 (Makambo 2009). Some initial fund- who receive training and experience in hospitality and ing was obtained from the U.S. Agency for International tourism. Local agricultural produce is purchased for use by Development (Verdugo 2009). The joint venture agreement the lodge, and there are plans to establish traditional danc- includes a 15-year lease agreement between Governors ing, a cultural center, a community walk, and handicraft Camps Ltd and SACOLA. The private sector operator built sales at the lodge. In addition, there are plans under the CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 237 joint venture to use funds from the lodge to finance infra- organizations as crop rangers, conservationists, porters, and structure, such as roads, in the area (Makambo 2009). community awareness representatives. Weaknesses of the joint venture, however, include poor The Iby’Iwacu cultural village has been developed in col- leadership and weak governance of SACOLA and the laboration between a private sector tour operator, Rwanda reliance on the private sector operator to generate revenue Ecotours, and a group of former poachers living near the from tourists (Makambo 2009). VNP. As a result of an academic research project, a partici- patory process began in 2005 to transform the livelihoods of Social benefits from gorilla tourism. Since the poachers toward farming, and then tourism. Part of the tourism revenue-sharing scheme was initiated in 2005, a concept was to benefit conservation, by providing alterna- number of direct and indirect projects with social benefits tive livelihood opportunities to illegal hunting of buffalo of gorilla tourism for local communities living near the and other wildlife in the VNP. Meetings with poachers were VNP have been implemented (Uwingeli 2009): held to gain their trust and insights, and study tours were undertaken to raise understanding of cultural tourism ■ Schools: Ten schools have been constructed, with 56 products elsewhere. classrooms and an average of 65 pupils per classroom per Interviews were undertaken with tourists to establish rotation (morning and afternoon). The school construc- what they would be willing to pay for a cultural experience. tion has reduced the distance traveled by children to the Community members engaged in the design, construction, nearest schools, allowing them to spend more time on and operation of the cultural village. Local architectural their studies after school. techniques were used in the construction along with local ■ Water tanks: Thirty-two water tanks have been con- materials such as thatching grass and wood. Tourists can structed. These provide 20 liters of water per person per stop at the Iby’Iwacu and experience local attire, practice day, and at least 1,250 people are served by each tank. traditional fire-making techniques, archery, drumming, and ■ Income-generating activities: Ten community associa- dancing; visit a traditional healer; and prepare and eat tra- tions have been supported directly through the revenue- ditional food. The village generates around $14,000 per year, sharing scheme, and a number of other projects, such as and community representatives identify projects in the bee-keeping and basket weaving, have been imple- community that they can finance (Sabuhoro 2009). mented. In all of these projects, the focus has been on training for income-generating activities. Challenges. Studies show that tourists to Rwanda do not ■ New partnerships in conservation and community have a particular willingness to pay for community bene- development brought to the construction of the fits. Bush, Hanley, and Colombo (2008), for example, find Sabyinyo community lodge, which is owned by that the percentage of VNP revenues used to enhance local SACOLA, but managed by a specialized ecolodge com- community development does not have a significant effect pany. At least 3,000 households are members of SACOLA on tourism demand. Bush, Hanley, and Colombo also note and benefit from the agreement with the managing com- that these findings do not imply that tourists are unwilling pany to pay SACOLA bed night fees and a percentage of to take part in community-based tourism—rather, that monthly net income. they are not willing to sacrifice other immediate benefits of the trekking experience relative to increases in permit Employment from gorilla tourism. The VNP employs prices that were dedicated to revenue sharing. This repre- at least 180 people, who work as guides, gorilla group track- sents an important departure from common ecotourism ers (for both tourism and research groups), and antipoach- principles about social benefits. Bush (2009) suggests that ing teams deployed in the five protected sectors of the VNP tourists need to be better educated about the human (Uwingeli 2009). In addition, an estimated 800 community dimension of conservation to emphasize the conceptual members around the VNP are involved in day-to-day man- link between the needs of the local populations and biodi- agement activities and benefit from opportunistic and tem- versity conservation. porary employment and the revenue-sharing scheme. VNP Furthermore, Rutagarama and Martin (2006) state that management has helped to form two umbrella associations: there is something of a catch-22 in relation to community one for conservation activities in the VNP (Amizero, or conservation in Rwanda. On one hand, the empowerment hope) and another for community development activities of local partners will be constrained when appropriate (Iby’Iwacu). Several hundred volunteers work with the two powers are not devolved to them. On the other hand, it is 238 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE impossible to impose powers on those who feel neither Box 14.3 Growth in Gorilla Conservation Efforts capable nor inclined to exercise them. Rutagarama and Martin (2006) therefore suggest the need for a flexible Mountain gorillas in the Virunga region of Rwanda, framework that enables capacity and power to coevolve in Uganda, and the Democratic Republic of Congo were locally appropriate ways. Developing the assets that part- censused five times between 1970 and 1989, when the ners need to maximize their opportunities for entering population was estimated at 324 gorillas. War and productive partnerships should be a fundamental part of political unrest in the region since 1990 meant that no plans to widen (and deepen) local participation in tourism. census was conducted over the next decade, and in In some countries this has been done through a hybrid 2000, the observation of 32 groups provided an esti- approach, and by retrofitting community-based tourism mate of between 359 and 395 gorillas. This represents enterprises into joint-venture partnerships with private a 0.9 percent to 1.8 percent annual growth rate over sector operators. Although the transaction costs, such as 10 years and a 1.0 percent to 1.3 percent annual growth rate between 1972 and 2000. the time needed to negotiate deals, can be high, these part- Source: Kalpers et al. 2003. nerships can provide a win-win for the private sector and community members with appropriate agreements (exam- ples are Damaraland in Namibia, Rocktail Bay and Phinda Reserve in South Africa, and Covane Fishing and Safari Lodge in Mozambique). In these situations the capacity of Biological research on mountain gorillas by teams of community members is built over time through partner- researchers in the Karisoke Research Centre in Rwanda pro- ships with businesses that understand the tourism sector, vides important information on trends in sound tourism how to operate a business, and how to establish market management. Of the estimated 380 gorillas in the Virunga linkages. Volcanoes Range, at least 260 are habituated and regularly monitored in Rwanda (Uwingeli 2009): These gorillas are checked on a daily basis; health reports are shared, and Implications for conservation actions taken when necessary. The habitat is patrolled daily Mountain gorilla tourism in Rwanda has long been viewed to detect illegal activities and discourage attempts to set as a valuable conservation tool. An economic incentive to snares. The use of information technology allows the habi- conserve the mountain gorilla is provided by international tat of the mountain gorillas to be mapped—in particular, to tourists paying relatively large sums of money to spend a show where gorilla groups are moving and what illegal short amount of time with the gorillas. Since its conception, activities are occurring and to plan ranger patrol activities organized gorilla tourism has provided funds to VNP accordingly. authorities to assist with conservation activities. Nature- Over the years, the Rwandan government has become based tourism has thus been enthusiastically accepted and more supportive of gorilla conservation, including allocat- supported by governments, conservationists, and tourists ing more land around the VNP for cultivation to reduce alike and, in Rwanda, has been acknowledged as playing a pressure on the park for agriculture and natural resource crucial role in the success of mountain gorilla conservation use (Uwingeli 2009). A consultation exercise is currently in the VNP (Bush, Hanley, and Colombo 2008). under way to assess the feasibility of a VNP expansion pro- The number of mountain gorillas left in the world is esti- gram (Bush 2009). Although the number of snares found in mated to be approximately 700. As of 2003, 380 gorillas the VNP have increased over time (Fawcett 2009), some for- lived in the Virunga volcanoes range, while 320 lived in mer poachers have begun working on conservation efforts, Bwindi Impenetrable National Park in Uganda as of 2006 and there is even an “ex-poachers association” consisting of (Uwingeli 2009). Research indicates that the gorilla popula- about 400 local community members, who patrol with tions in areas frequented by tourists are increasing, with an ORTPN staff and also help with local education, collecting overall growth in the population of 1.1 percent between information, and addressing human-wildlife conflict (for 1989 and 2003 (Fawcett 2009) (box 14.3). Gorilla groups in example, crop raiding) (Uwingeli 2009). the Democratic Republic of Congo are doing less well in Although no systematic method currently exists for regis- terms of population growth, which may be due to the lower tering or training guides in Rwanda, VNP staff have bene- presence of patrols, researchers, and tourist visits there com- fited from capacity-building programs by the RDB (and pared with Rwanda and Uganda (Fawcett 2009). previously, by ORTPN) and their partners in guiding, gorilla CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 239 health monitoring, and general biodiversity conservation. groups3 and prefer the length of their trek to be between Also, with the support of IGCP and the Karisoke Research one and three hours. These two findings could be inter- Centre, additional training has been provided for gorilla preted as showing that tourists support the ecotourism trekking guides on how to work with visitors and how to principle of minimizing ecological impact, since more minimize adverse impacts on the gorillas (Kalpers et al. 2003; people taking longer trips would increase adverse ecolog- Fawcett 2009). ical impacts. Revenues from Rwanda’s national parks are primarily However, price increases in gorilla permits appear to used to fund conservation efforts in the parks and world- affect tourism visitation, at least in the short term. Research wide tourism marketing activities. Salaries for all of the staff by Bush and Fawcett (2007) reveals that the price increase in are paid out of national park fees. VNP park management is June 2007 from $375 to $475 had a marked impact on the cofunded through the research activities of the Karisoke demographics of visitors to the mountain gorillas (figure Research Centre, which provides basic park management 14.5). In particular, the percentage of visits by people in the functions such as monitoring and antipoaching patrols for highest income group increased significantly, while visits by gorillas. Further conservation funds are contributed by all but one of the other income groups decreased. The NGOs such as the IGCP and CARE. These additional funds length of the stay also decreased significantly, from a mean and support have contributed enormously to gorilla conser- number of 4.2 nights to 3.6 nights. vation successes in Rwanda (Bush 2009). The proportion of visitors going to the other national parks in Rwanda was also significantly lower after the price increase, as was the proportion of visitors to the genocide Willingness to pay for conservation memorial and taking the Kigali city tour. The proportion of Research by Bush, Hanley, and Colombo (2008) finds that visitors participating in alternative activities within the VNP gorilla trekking tourists are willing to pay for biodiversity increased, however: hikes on the Karisimbi and Bisoke vol- conservation, for both gorillas and other wildlife. The canoes, viewing golden monkeys, visits to Dian Fossey’s authors also find that tourists prefer to be in small tomb, and nature walks. Figure 14.5 Impact of the 2007 Gorilla Permit Price Increase on Visitation a. Tourist visitation by income group b. Length of trip 50 50 40 40 30 30 Percent Percent 20 20 10 10 0 0 <$25,000 $25,000 to $45,000 to $65,000 to >85,000 1 2 3 4 5 6 8 $45,000 $65,000 $85,000 Nights stayed in Rwanda Respondent annual income group After price rise Before price rise Source: Bush and Fawcett 2007. 240 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE The reduction in visitors’ length of stay and in the fre- ■ Although the number of tourist accommodation facili- quency of other activities caused by the price increase in ties is growing, the facilities are not sufficient at key gorilla tourism has implications for the overall economic tourism sites, including the VNP. Additionally, the qual- impact of international tourism in Rwanda. However, Bush ity of accommodations is not standardized, and prices (2009) notes that it is probably time to repeat the study and are high relative to accommodations of equivalent qual- establish whether the changes in demographics and con- ity in other East African countries. sumption were temporary or lasting. This study would also ■ The quality of customer service is inconsistent and gen- provide guidance for decision making about further price erally poor compared with that in neighboring countries. changes. A study of tourism satisfaction and pricing for Also the focus on high-end consumers implies demand alternative products to ensure value for money is also for higher quality. Bush found that after the increase of needed in order to increase the number of bed nights and the permit price respondents registered significantly overall trip spending (Bush 2009). lower levels of satisfaction with their trekking experience (Bush 2009). ■ Public and private sector collaboration is improving Remaining constraints and lessons learned through the Joint Action Development Forum and Steer- ing Committees, though it is still weak. There are a number of constraints to the further development ■ Although the road distance between Kigali and Musanze of gorilla tourism in Rwanda: is short, the quality of the road could be improved. ■ While regional collaboration is of interest to government ■ Gorilla tourism can present a threat to gorilla conserva- authorities of the Democratic Republic of Congo, tion, affecting, for example, the health and behavior of Rwanda, and Uganda, regional instability has made it gorillas (box 14.4), and needs to be well managed. difficult to harmonize tourism and conservation in Box 14.4 Threats to Gorilla Conservation from Tourism All six great apes—gorillas Gorilla gorilla and Gorilla of behavioral data collected from three gorilla groups beringei, chimpanzees Pan troglodytes and Pan paniscus, during one-hour observation sessions before, during, and orangutans Pongo pygmaeus and Pongo abelii—are and following tourist visits. Results from these data show categorized as endangered on the International Union clearly that the current tourism program is having a sig- for Conservation of Nature’s 2000 Red List. Threats to nificant impact on gorilla behavior. All three gorilla these species include loss of habitat to settlement, log- groups were found to spend significantly less time feed- ging, and agriculture; illegal hunting for bushmeat and ing and more time moving during tourist visits. In addi- traditional medicine; the live ape trade; civil unrest; and tion, the frequency of certain aggressive behaviors, many infectious diseases. The great apes are highly susceptible directed at humans, increased in all three groups during to many human diseases, some of which can be fatal. If tourist visits, and the gorillas increased their proximity to protective measures are not improved, ape populations the silverback members (adult male gorillas usually more that are frequently in close contact with people will than 12 years of age). eventually be affected by the inadvertent transmission Some gorilla behaviors observed during the study of human diseases. Regulations that protect habituated were correlated with the distance maintained between apes from the transmission of disease from people are the gorillas and the tourists and the number of tourists often poorly enforced. Enforcement of existing regula- in the gorilla group. Reducing this impact on gorilla tions governing ape-based tourism, and the risk of dis- behavior may be a simple matter of better training ease transmission between humans and the great apes guides to maintain the 7-meter distance rule between minimized (Woodford, Mutynsky, and Karesh 2002). tourists and gorillas. Many of these changes in gorilla Tourism might also be changing gorilla behavior in behavior during tourist visits are believed to indicate negative ways. In one research study, Fawcett, Hodgkin- higher levels of stress in gorillas (Fawcett, Hodgkinson, son, and Mehlman (2004) assessed more than 10 months and Mehlman 2004). CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 241 accordance with the Virunga Massif Transboundary Plan tourism development. Rwanda also offers business opportu- (Uwingeli 2009; Mehta and Katee 2005). nities, mainly for travelers from the eastern part of the Dem- ■ There is widespread poverty around the VNP and ocratic Republic of Congo and other neighboring countries. increasing pressure for more agricultural land by a grow- Rwanda has also been successful in attracting national, ing rural population. regional, and international conferences. ■ Benefits accruing from the VNP should be in relation to the needs of people living close to the park. Local people Development and structure of the tourism sector desire individual benefits (such as money), and not only collective infrastructure that is used by the whole popu- In the 1970s and early 1980s only a small number of inter- lation, such as water tanks or clinics (Uwingeli 2009). national tourists visited Rwanda. Most tourists visited Akagera National Park, a government-owned, high-end des- Several tourism-related lessons learned at the VNP have tination used mostly for hunting. Only a very limited num- the potential to improve other protected areas in Rwanda: ber of tourists visited the gorillas. Tourism was not a national priority and was not viewed as a tool to reduce ■ Because of the limited number of available gorilla per- poverty. The first hotel, the Mille Collines, was built in 1973 mits, product diversification and promotion is required and the ORTPN was created in 1974. No tour operators to encourage visitors (particularly repeat visitors) to stay existed in the 1970s and 1980s, and the sector was domi- longer, spend more money, and visit other destinations nated by the government, which owned all hotels except the in the country. Mille Collines. ■ Gorilla conservation needs to be balanced with research Tourist arrivals started to increase notably in the 1980s visits and tourism trips to ensure that the health of the (figure 14.6). Most tourists still visited only Akagera gorillas and the integrity of their habitat are maintained. National Park at that point, but the first official mountain ■ Conservation efforts focused on the key species is impor- gorilla tourism program was launched in 1979, leading to tant, but the contribution to the habitat/ecosystem con- continuously increasing visitor numbers in the VNP, peak- servation must also be ensured. Long-term dedication ing at 6,900 in 1989 (ORTPN 2008b). This trend was and partnerships in conservation (research, protection, brought to an abrupt end with the outbreak of the war in and tourism) are essential. mid-1990s, however. Visits to the VNP, which provided ■ Sustainability of gorilla tourism can be achieved only if most of the tourism revenue by the mid-1990s, dropped in regional collaboration is established to conserve trans- 1994 due to the genocide and again between 1997 and 1999, boundary protected areas and cross-border resources. when the VNP had to be closed for some time due to an ■ Standardized and high-quality training for guides is insurgency. Visits increased in the years following, however. needed, for those working both within and outside pro- In 2008 about 17,000 people visited the VNP to see the tected areas. gorillas, a dramatic increase from the late 1980s and an ■ A more diverse range of accommodation and restaurant impressive recovery from only 417 tourists in 1999 after the facilities is required, with higher quality and better value reopening of the VNP following the war. Total visits to all for money. Rwanda’s national parks reached more than 43,000 visitors in 2008. Today, the majority of visitors to the VNP are for- eigners, while Akagera National Park is visited by a more TOURISM IN RWANDA: THE BIGGER PICTURE equal mix of Rwandan and foreign residents. Besides the mountain gorillas in the VNP, Rwanda has other Reliable tourist arrival statistics for Rwanda as a whole are excellent tourism assets that create a wider foundation for available only for 2007 onward, when entry cards were intro- the tourism sector. Rwanda has three national parks that duced. This innovation led to a substantive upward revision cover about 10 percent of the country’s area, one of which is of tourism revenue. The use of statistics on park visits alone the VNP. The Akagera National Park offers a range of wildlife led to a significant underestimation of the total number of including elephants, hippopotamuses, giraffes, and zebras. tourists. Most important, the large number of business and The Nyungwe Forest National Park has a large tract of conference tourists had not been taken into account. mountain forest and is rich in biodiversity. Guided walks and Although data are still not collected electronically and the chimpanzee tracking is offered. Lake Kivu has recreational quality of data is not high, there is little doubt that the recov- facilities as well, but there is still potential for significant ery of the tourism sector in Rwanda has been successful. 242 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE Figure 14.6 Visits to Rwanda’s National Parks, 1974–2008 45,000 40,000 35,000 30,000 Number 25,000 20,000 15,000 10,000 5,000 0 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 Nyungwe Akagera Volcanoes Source: ORTPN. Table 14.1 Tourist Arrivals to Rwanda by Purpose of Visit, 2007–09 Thousands Visiting friends Conference/ Holiday/vacation and relatives business Transit Other Share Share Share Share Share Year Number (%) Number (&) Number (%) Number (%) Number (%) Total 2007 21.5 2.6 332.0 40.2 275.8 33.4 150.1 18.2 47.0 5.7 826.4 2008 59.4 6.1 248.3 25.3 345.9 35.3 307.8 31.4 19.1 1.9 980.6 2009 (Jan - Jun) 21.4 4.9 112.2 25.6 187.9 42.8 83.3 19.0 34.1 7.8 439.0 Source: ORTPN. A total of 980,577 international arrivals were recorded in arrivals came from Africa. The country or region of origin, 2008, up from 826,374 in 2007 (table 14.1). The main char- however, varies considerably by the purpose of the visit. Most acteristics of international tourists entering Rwanda are: of the tourists on vacation came from Europe and the United States (figure 14.7a). The majority of business and conference ■ Most visitors came for business and conferences (35 per- travelers came from the Democratic Republic of Congo cent in 2008). This had already been indicated in a hotel and the other EAC member states (figure 14.7b). Most inter- market study, which found that 75 percent of all tourists national tourists visiting friends and relatives in 2008 came in the country in 2006 were business travelers (IFC 2007). from the EAC and Congo Republic (figure 14.7c). According ■ The share of tourists arriving in Rwanda for vacation is to estimates by ORTPN and the Ministry of Economy and relatively small, but increased from 3 percent in 2007 to Finance, non-African tourists coming to Rwanda for leisure 6 percent in 2008. or conference and business purposes spent the most money ■ The large number of international arrivals includes tran- among the various categories of tourists (table 14.2). sit passengers, thereby reducing the overall number of Most leisure tourists visit the region as part of a multi- arrivals that can be counted as tourists entering Rwanda. country itinerary and do not yet consider Rwanda as a stand- alone destination. A recent survey of tourists, tour operators, The tourist entry cards offer more details about travelers and accommodation providers in Rwanda (SNV and RDB entering Rwanda. Overall, 88 percent of total international 2009) finds that the most common length of stay for CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 243 Figure 14.7 Country/Region of Origin of Foreign Arrivals to Rwanda, by Purpose of Visit, 2008 a. Holiday/vacation b. Business/conference c. Visiting friends and relatives USA and USA and Other Canada Canada 6% Other 2% 3% 3% Europe 4% Other Europe 11% EAC 4% 21% EAC USA and 30% Canada EAC 24% Congo, Dem. Rep. 50% Congo, Dem. 14% Congo, Dem. Rep. Rep. Europe 57% 41% 30% Source: ORTPN. Table 14.2 Average Spending by Tourists in Table 14.3 Growth in Hotel Rooms, Restaurants, Rwanda, by Purpose of Visit, 2009 Tour Operators, and Travel Agencies in Average amount spent per visit Rwanda (US$) Average annual Non-African Service 2003 2009 growth rate (%) Purpose of visit visitors African visitors Hotel rooms 650 4,256 37 Leisure 1,623 1,136 Restaurants 50a 94 17 Conference and Tour operator companies 12 26 14 business 1,623 108 Travel agencies 5 24 30 Visiting friends and relatives 120 84 Source: RDB/ORTPN. Transit/other 119 83 a. Data from 2005. Source: RDB/ORTPN/MINECOFIN, based on visitor expenditure survey from 2006, adjusted for inflation. rates for lower-grade accommodation (2,264 rooms) were only 28.5 percent (ORTPN 2008c, 2008d). domestic tourists was two days, while the most frequently A total of 26 tour operators were active in Rwanda as of cited length of stay by international tourists was four days. 2009, compared to none in the 1980s. The number of The number of hotel rooms and tour operators in restaurants and travel agencies has grown as well. Most Rwanda has increased significantly in recent years, under- Rwandan tour operators started out with little available scoring the successful recovery of the tourism sector. Data finance, which limited their possibilities. They could, for available from RDB and ORPTN show that the number of example, not afford to buy a car, but had instead to rent cars hotel rooms increased from 650 in 2003 to 4,256 in 2009, as they needed them. Most of those tour operators have suc- more than 500 percent overall and 37 percent annually on cessfully expanded their businesses by now, though, and average (table 14.3). Information on occupancy rates, how- some regional operators have also opened offices in Kigali. ever, is very limited. The only available information spans The tourism sector has also experienced a significant January to March 2008 and indicates an average room occu- amount of privatization. The situation is much different pancy rate of 36 percent. There was a large difference in the from the 1980s, when all hotels except one were government occupancy rates by the grade of accommodation. Room owned. Although the government still held a share in two occupancy rates for upper-grade accommodation (a total of hotels in 2010, it was not involved in their management, 453 rooms) were more than 70 percent on average, whereas leaving the tourism sector almost entirely in the hands of 244 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE the private sector. The government still owns and runs the role in the development of the tourism sector in Rwanda. national parks, although in 2009 a concession was awarded Eighty-six percent of all new projects since 1999 that are to the private company African Parks to manage Akagera operational were financed by local investors. Moreover, a National Park. group of private investors has established the Rwanda Tourism income, as recorded by ORPTN, has increased Investment Group (RIG) to pool resources. Several sub- in recent years. The majority of revenue is derived from RIGs have also been created, one of which is currently plan- VNP entrance and gorilla permit fees. Other income is ning the construction of a convention center intended to raised through the gorilla naming ceremony, partners and hold up to 2,000 participants, which would increase donors, and interest from treasury bills and other income. Rwanda’s chance of attracting large conferences.4 Most of the revenue is spent on operating expenditures, Finally, the structure and organization of the tourism sec- whereas capital expenditures paid out of the budget of tor has been reformed to assign clear responsibilities. The ORPTN are limited. Large investments are funded by the Ministry of Commerce, Industry, Investment Promotion, overall central government budget. Revenue shared with Tourism, and Cooperatives (MINICOM) holds overall communities has increased since 2005, representing 8 per- responsibility for tourism. The private sector is represented cent of total operating expenditure in 2008, up from 6 per- by the tourism chamber, which consists of four industry cent in 2006. associations: accommodation, tour operators, transport, and Foreign investment in Rwanda’s tourism sector is sub- private education establishments. The tourism chamber is, stantial. Between 2000 and 2009, foreign direct investment however, still supported by the government because of insuf- of RF 258 billion went into hotels and leisure, accounting ficient resources. It is part of the private sector federation, for 20 percent of total FDI inflows into the country (figure the equivalent of a chamber of commerce and industry. 14.8). Local investment in hotels, restaurants, and other tourism activities amounted to RF 140 billion between 1999 and 2009, representing 16 percent of total local investment Contribution of tourism to the economy over that period. Rwanda has made remarkable progress in terms of eco- Although the amount of foreign investment exceeded the nomic growth since the genocide in 1994 (figure 14.9). amount of local investment, it has been concentrated on a Growth averaged 15.6 percent in the five years after the small number of projects. Local investors play an important genocide in 1994, declined to an average of 6.6 percent between 2000 and 2004, and increased again to an average of 8.4 percent between 2005 and 2008. A significant increase Figure 14.8 FDI to Rwanda, by Sector, 2000–2009 in GDP per capita has been recorded along the way, from $142 in 1994 to $313 in 2008 (both in 2000 prices). While the contribution of tourism to GDP remains small, the sector has become Rwanda’s main source of Hotels and export revenue. The category “restaurants and hotels” has leisure Mining and contributed less than 2 percent to overall GDP and 4 per- 20% energy 33% cent to the services sector on average since 1999, but value added from restaurants and hotels (at constant prices) has recorded a steady increase of 21 percent on average.5 The Other main increase in the services sector came from wholesale 24% and retail trade and other services (education, health, Telecom finance and insurance, and real estate). The measurement of 15% tourism’s contribution to GDP, however, is difficult, because transport services, for example, constitute a large share of tourism revenues but are not included in the “hotel and restaurants” category. Overall, exports of nonfactor services from Rwanda have Construction and real estate outperformed exports of goods (such as coffee and tea) as 24% Rwanda’s main foreign exchange earner (figure 14.10). Source: RDB; included are RDB registered investments. Travel is the largest component of exports of nonfactor CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 245 Figure 14.9 Growth in Real GDP and GDP per Capita, 1993–2008 40 30 300 20 Constant 2000 USD 10 250 Percent 0 –10 200 –20 –30 150 –40 –50 100 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Real GDP (left axis) Per capita GDP (right axis) Source: International Monetary Fund. Figure 14.10 Comparison of Composition of Export of Goods and Nonfactor Services (Averages) a. 1995–98 b. 2005–08 Nonfactor Nonfactor services excl. services excl. Coffee travel travel 11% Tea 20% Coffee 23% 8% 37% Travel Other 20% goods Travel Other Tea 22% 36% goods 11% 12% Source: IMF. services; other categories include other transportation and ducers, and artisans. Tourists visiting the VNP and the freight and insurance. After a decline from $19 million in Musanze area generate around $1 million in income for 1993 to only $6 million in 1995, revenue from tourism poor workers and producers. In addition, the area receives increased tremendously to $202 million in 2008 (figure large amounts of donations and grants7 (SNV and ODI 14.11). A comparison with the preconflict period is diffi- 2008). As much as possible, hotels source their supplies cult, because data on tourism revenue are available only for (particularly food products) from the local market to con- 1992 onward. tribute to the economy. Rwanda has identified tourism in its Economic Devel- Estimates for 2009 indicate that the tourism industry opment and Poverty Reduction Strategy as a national pri- directly employs 33,800 people in Rwanda, whereas indirect ority sector to eradicate poverty (Republic of Rwanda employment accounts for another 40,500 jobs, resulting in 2007). According to rough estimates, each of the three big total travel and tourism–related employment of 74,300 jobs. business hotels in Kigali6 generates about $500,000 per year Tourism thus represents 4.0 percent of total employment in income for semiskilled and unskilled workers, food pro- in Rwanda, only slightly below the Sub-Saharan African 246 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE Figure 14.11 Tourism Revenue, 1992–2008 220 200 180 160 U.S.$ (millions) 140 120 100 80 60 40 20 0 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Tourism revenue (current prices) Tourism revenue (constant 2005 prices) Source: IMF and World Bank, World Development Indicators database. average of 4.6 percent, and well below the amount in Kenya diversified and other attractions promoted, such as bird- (7.1 percent), Tanzania (7.1 percent), and Uganda (6.6 per- ing and primates in Nyungwe, visits to Lake Kivu, and cent). Employment in the tourism industry (direct and conference tourism. Diversity is particularly important indirect) has grown by 2.8 percent on average in Rwanda in establishing Rwanda as a stand-alone destination. over the past 10 years, compared with 3.4 percent in Kenya, ■ Access to finance is still an impediment for the devel- 2.5 percent in Tanzania, and 2.4 percent in Uganda (World opment of the sector. Banks seem reluctant to finance Travel and Tourism Council 2009). tourism projects because they are within a service- oriented sector rather than attached to a sector produc- ing tangible goods. Stakeholders in the tourism sector REMAINING CONSTRAINTS AND EMERGING have proposed the establishment of a guarantee fund by POSSIBILITIES FOR TOURISM the government. Despite the good performance of the tourism sector in ■ Other sectors need to be further promoted through the Rwanda, several challenges remain. The main impediment tourism sector in order to reduce poverty. Pro-poor links cited by almost all actors in the sector is the large skill deficit. that can be exploited further include the food supply This deficit applies to all areas of tourism, including guides, chain to hotels, lodges, and restaurants; assistance to chefs, and hotel service personnel and technicians. Hotels poor households to access training, employment, and and tour operators either train their staff in house or send promotion in hospitality; practical initiatives to help them to neighboring countries to be trained, although businesses enhance their own business models; and part- recently some tourism schools have been opened by the pri- nerships with more domestic and regional tour opera- vate sector. The emphasis of the curriculum is on managerial tors, hotels, and lodges to promote community activities rather than technical skills, meaning that the demands of the such as cultural events and the sale of handicrafts (SNV sector are not taken into account adequately. To accomplish and ODI 2008). the goal of turning Rwanda into a service-oriented economy, ■ Rwanda needs to comply with international standards. skill development is of utmost importance. Currently, hotels are being classified according to the EAC Several other challenges confront the tourism sector: standard. While this is an important step, it poses chal- lenges to a number of hotels. Once Rwanda shifts to an ■ There is an overreliance on gorilla tourism. The number international system of standards and classifies hotels of permits cannot be easily increased, and the existing accordingly, the pricing structure might have to be revised permits are typically sold out. Tourism needs to be to remain competitive with the neighboring countries. CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 247 ■ RDB and MINICOM do not have sufficient staff. In BIBLIOGRAPHY MINICOM, only one part-time person is responsible for Barrett, C., and P. Arcese. 1995. “Are Integrated Conservation- tourism. RDB, as well, has only a limited number of staff. Development Projects Sustainable?” World Development Given the very ambitious agenda for the tourism agenda, 23 (supplement 7): 1073–84. sufficient staffing should be in place. Bush, G. 2009. Personal Communication, Woods Hole ■ Infrastructure needs improvement. Although roads in Research Center, November 30. Rwanda are broadly adequate, the air transport con- Bush, G., and K. Fawcett. 2007. “An Economic Study of nection to the country is still limited. A new airport is Mountain Gorilla Tourism in the Virunga Volcanoes currently being planned, which could attract more Conservation Area.” U.S. Fish and Wildlife Service, Dian international carriers to offer direct flights, especially Fossey Gorilla Fund International, Ruhengeri, Rwanda. from Europe. Bush, G., N. Hanley, and S. Colombo. 2008. “Measuring the Demand for Nature-based Tourism in Africa: A Choice In the face of the challenges, several possibilities are Experiment Using the ‘Cut-off ’ Approach.” Discussion emerging to diversify the tourism sector and increase its Paper 2008-06, Stirling Economics, University of Stirling, contribution to the economy. Regarding leisure tourists, Stirling, U.K. primate tours, and birding in the Nyungwe Forest are the Butynaski, T. M., and J. Kalina. 1997. “Gorilla Tourism: A most promising areas of diversification. New experiences Critical Look.” Africa Biodiversity Conservation Pro- would extend the length of stay of tourists and eventually gram, Nairobi. establish Rwanda as a stand-alone destination. Lessons Fawcett, K. 2009. Personal Communication, Director, learned from gorilla tourism regarding conservation, how- Karisoke Research Centre, September 16. ever, should be taken into account when developing tourism Fawcett, K., C. Hodgkinson, and P. Mehlman. 2004. “An attractions in other national parks. The development con- Assessment of the Impact of Tourism on the Virunga cept of Rwanda’s Destination Management Areas is identi- Mountain Gorillas: Phase I—Analyzing the Behavioral Data from Gorilla Groups Designated for Tourism.” Dian fied in the Sustainable Tourism Development Master Plan Fossey Gorilla Fund International, Atlanta, GA. (Republic of Rwanda 2009b). Conference tourism could Gray, M., et al. 2005. “Virunga Volcanoes Range and Census, also be of great benefit to Rwanda. In this regard, the con- 2003.” Joint organizers report, Uganda Wildlife Author- struction of a conference center that will accommodate up ity/ORTPN/Institut Congolaise pour la Conservation du to 2,000 people is under way. Further, a draft action plan for Nature. the development of MICE (meetings, incentives, confer- Grosspietsch, M. 2006. “Perceived and Projected Images of ences, and exhibitions) tourism has been prepared by the Rwanda: Visitor and International Tour Operator Per- Tourism Working Group. spectives.” Tourism Management 27: 225–34. Homesy, J. 1999. “Ape Tourism and Human Diseases; How NOTES Close Should We Get?” Technical Report, International Gorilla Conservation Program, Nairobi. 1. EAC member countries are Burundi, Kenya, Rwanda, Hulme, D., and M. Murphee. 2001. “Community Conser- Tanzania, and Uganda. vation in Africa.” In African Wildlife and Livelihoods: The 2. The gorillas are visited by independent travelers, over- Promise and Performance of Community Conservation, landers, and high-end tours (ORTPN 2004). ed. D. Hulme and M. Murphee. Oxford, U.K.: James 3. The optimum group size, according to tourists, is six Curry Ltd. people; tourists are not willing to pay more to reduce group IFC (International Finance Corporation). 2007. “Rwanda size beyond this (Fawcett 2009; Weber 1993). Hotel Market Study.” Washington, DC. 4. The currently available conference facilities can host a Kalpers, J., et al. 2003. “Gorillas in the Crossfire: Population maximum of 500 and 1,000 participants. Dynamics of the Virunga Mountain Gorillas over the 5. The share of hotels and restaurants is not available for Past Three Decades.” Oryx 37 (3): 326–37. years before 1999. Litchfield, C. A. 1997. “Treading Lightly: Responsible 6. Serena Hotel Kigali, Hotel de Mille Collines, and Laico Tourism with the African Great Apes,” Travellers’ Medical Umubano Kigali Hotel (former Novotel Kigali). and Vaccination Centre, Adelaide. Cited in Fawcett, 7. Including contributions to community development Hodgkinson, and Mehlman (2004). from ORTPN’s revenue-sharing program and donations by Makambo, W. 2009. “Sabyinyo Silverback Lodge: A Com- tourists. munity Partnership for Conservation.” Presentation 248 CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE from RDB-TC tourism forum, International Gorilla SNV Rwanda and ODI. 2008. “Creating Pro-poor Linkages Conservation Program, Nairobi. around Rwanda Tourism.” Kigali. Mehta, H., and C. Katee. 2005. Viriunga Massif Sustainable SNV Rwanda and RDB. 2009. “Tourism Market Research Tourism Development Plan, DR Congo, Rwanda, for Rwanda 2009: A Survey of Tourists, Tour Operators Uganda. ECDA: Fort Lauderdale FL. and Accommodation—Final Report.” Kigali. Murnyak, D. F. 1981. “Censusing the Gorillas in Kahuzi- Spenceley, A., S. Habyalimana, R. Tusabe, and D. Mariza. Biega National Park.” Biological Conservation 21: 163–76. 2010. “Benefits to the Poor from Gorilla Tourism OTF Group. 2008a. “Building a Winning MICE Tourism in Rwanda.” Development Southern Africa, 27 (5): Product—Target Customer Needs.” Rwanda Tourism 647–62. Workgroup. Télesphore, Ngoga. 2009. Interview. September. RDB, ———. 2008b. “Building a Winning Bird Product—Target Kigali. Customers and Channel Partner Needs.” Rwanda Townsend, C. Forthcoming. “Guidelines for Community- Tourism Workgroup. based Tourism in Rwanda.” ORTPN / UNWTO ST-EP. ORTPN (Office Rwandaise du Tourisme et des Parcs (Sustainable Tourism–Eliminating Poverty). Nationaux). 2004. “Strategic Plan 2004–2008.” Draft ver- Tusabe, R., and S. Habyalimana. 2010. “From Poachers to sion 2, Kigali. Park Wardens: Revenue Sharing Scheme as an Incentive ———. 2008a. “Highlights on National Parks Visitation in for Environment Protection in Rwanda.” Mountain Rwanda—2008.” Vol. l.1, Kigali. Forum Bulletin (10) 1. ———. 2008b. “Highlights of International Visitor Arrivals Uwingeli, P. 2009. Personal Communication, Chief Park in Rwanda—2008.” Kigali. Warden, Volcanoes National Park, November 3. ———. 2008c. “Highlights on Accommodation Statistics in Vedder, A., and B. Weber. 1990. “The Mountain Gorilla Rwanda—January 2008.” Vol. I., Kigali. Project.” In Living With Wildlife: Wildlife Resource Man- ———. 2008d. “Highlights on Accommodation Statistics in agement with Local Participation in Africa, ed. A. Kiss. Rwanda—March 2008.” Vol. I.2, Kigali. Washington, DC: World Bank. Plumptre, A. et al. 2004. “The Socio-economic Status of Verdugo, D. 2009. “Challenges for Developing Community People Living Near Protected Areas in the Central Alber- Understanding of Tourism in an Emerging Destina- tine Rift.” Albertine Rift Technical Reports 4, Albertine tion.” Unpublished paper cited in Spenceley, A. et al. Rift, Program, Wildlife Conservation Society, New York. (2010). Republic of Rwanda. 2002. “A Plan to Unleash Rwanda’s Weber, B. 1982. “People, Gorillas...or Both?” London: Earth- Tourism Potential.” Kigali. scan. ———. 2007a. “Economic Development and Poverty ———. 1985. “Le Parc National des Volcans Biosphere Reduction Strategy, 2008s–2012.” Ministry of Finance Reserve, Rwanda: The Role of Development in Conser- and Economic Planning, Kigali. vation.” Parks 10 (3): 19–21. ———. 2007b. “A Revised Tourism Strategy for Rwanda— ———. 1987. “Socioecologic Factors in the Conservation of Sustaining the Momentum.” Ministry of Commerce, Afromontane Forest Reserves.” In Primate Conservation Industries, Investment Promotion, Tourism and Cooper- in the Tropical Rainforest, ed. C. Marsh and R. Mitter- atives, Kigali. meier, 205–29. New York: Alan R. Liss, Inc. ———. 2009a. “Rwanda Tourism Policy.” Ministry of Trade ———. 1993. “Ecotourism and Primate Conservation in and Industry, Kigali. African Rain Forests.” In The Conservation of Genetic ———. 2009b. “Sustainable Tourism Development Master Resources, ed. C. Potter, 129–50. Washington, DC: AAAS Plan for Rwanda—Final Report.” Project of the Republic Press. of Rwanda and the United Nations World Tourism Woodford, M., T. Mutynsky, and W. Karesh. 2002. “Habitu- Organization. ating the Great Apes: The Disease Risks.” Oryx 36 (2): Rutagarama, E., and A. Martin. 2006. “Partnerships for Pro- 153–60. tected Area Conservation in Rwanda.” Geographical Jour- World Bank. 2008. “Country Assistance Strategy for the nal 172 (4): 291–305. Republic of Rwanda for the Period FY09-FY12.” Report Sabuhoro, E. 2009. “Contribution of Tourism to Poverty No. 44938-RW, Washington, DC. Reduction in Rwanda: A Tour Operator’s Perspective.” World Travel and Tourism Council.s 2009. “Travel and Presentation at the Delphe/EPA Workshop, July 6–7. Tourism Economic Impact—Rwanda 2009.” London. CHAPTER 14: THE SUCCESS OF TOURISM IN RWANDA: GORILLAS AND MORE 249 PA R T I V Boosting Agricultural Efficiency and Output through Targeted Interventions CHAPTER 15 Increasing Rice Productivity and Strengthening Food Security through New Rice for Africa (NERICA) Aliou Diagne, Soul-Kifouly Gnonna Midingoyi, Marco Wopereis, and Inoussa Akintayo espite Africa’s potentially rich land and water D capita rice consumption in West Africa increased from resources, its farmers are among the poorest in the 14 kilograms in the 1970s to 22 kilograms in the 1980s and world. Because the vast majority of people in more than 39 kilograms in 2009. For Africa as whole, annual Africa derive their livelihoods from agriculture, the weak per capita rice consumption increased from 11 kilograms in state of the sector has profound implications for poverty.1 the 1970s to 21 kilograms in 2009 (figure 15.1). Since the Agricultural innovation in Africa needs to internalize the early 1970s rice has been the number one source of caloric region’s biophysical, institutional, and socioeconomic con- intake in West Africa and the third most important source straints and establish efficient value chains to support sus- of calories (after maize and cassava) for the continent as a tainable growth and reduce poverty. whole (figure 15.2). Agricultural research can catalyze agricultural innova- Domestic rice production grew at the rate of 6 percent a tion and the development of the value chain. A prime exam- year between 2001 and 2005.5 But production still falls far ple is the New Rice for Africa (NERICA) varieties developed short of demand. As a result, Africa imports up to 40 per- by the Africa Rice Center (AfricaRice2) and partners, which cent of its rice consumption. The continent as a whole gained popularity among Africa’s rice farmers in a relatively imports 30 percent of world rice imports, a potentially very short period of time.3 NERICA varieties were developed risky and unsustainable situation, as shown during the food from crosses between the African species Oryza glaberrima crisis in 2008. Rice imports cost Africa almost $4 billion in Steud. and the Asian species O. sativa L., using conventional 2009 (Seck et al. 2010)—money that could have been biotechnology to overcome the sterility barrier between the invested in developing the domestic rice sector. two species. This chapter describes the history, develop- Until the early 1990s, rice breeding programs in Africa ment, and impact of NERICA and makes recommendations worked almost exclusively on improving O. sativa lines about its expansion. for Africa’s diverse rice growth environments. Some breeders even voted to accelerate the disappearance of glaberrima (IRAT 1967), even though sativa varieties are THE IMPORTANCE OF RICE AND THE generally much more vulnerable than glaberrima to the DEVELOPMENT OF NERICA VARIETIES numerous biotic stresses (rice diseases, insect attacks, Rice has long been the food staple in many traditional rice- weeds, and so forth) and abiotic stresses (soil acidity, growing communities and in major cities in West Africa.4 It drought, salinity, iron toxicity, and so forth) of the is now the fastest-growing food staple in Africa. Annual per African environment. 253 Figure 15.1 Annual per Capita Rice Consumption in Africa and West Africa, 1961–2009 45 40 35 West Africa Kilograms/person/year 30 25 20 15 Africa 10 5 0 19 1 63 65 67 69 71 73 19 5 77 79 19 1 83 85 87 89 19 1 93 19 5 97 99 01 03 20 5 07 09 6 7 8 9 9 0 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 Source: FAO 2009. Figure 15.2 Share of Calories in African Diet Provided by Various Sources, 1961–2007 18 16 14 Percent of total calorie intake 12 10 8 6 4 2 0 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 07 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 Cassava Maize Millet Rice (milled equivalent) Sorghum Source: FAO 2009. Some visionary breeders thought about combining the ing breeders from the Institut de Recherche Agronomique favorable characteristics of O. sativa with those of O. glaber- Tropicale (IRAT) and the Institute of Savannah (IDESSA) rima. Crossing was difficult, however, because of the steril- Bouaké, in Côte d’Ivoire, worked on the problem during the ity barrier between the two species. A few breeders, includ- 1970s and 1980s (Diagne 2006). Their work did not produce 254 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) good results, with the crosses resulting in either sterile off- table 15.A1 and figure 15.A1 in annex 15.A1, which compare spring or offspring that had some of the undesirable char- the performance of NERICA progenies with that of their acteristics of the glaberrima parent (lodging, shattering, low sativa and glaberrima parents and other sativa checks under yield).6 high and low input conditions and under major stresses in Development of the NERICA varieties began in 1991, upland rice ecologies in AfricaRice’s key research sites in when AfricaRice initiated an interspecific breeding program Cote d’Ivoire.8 for the upland ecosystem. This long-term investment paid Indeed, the agronomic trials data show that the NERICA off, when breeders eventually overcame the obstacles varieties bring to upland farmers the high yield potential encountered earlier, through perseverance and the use of of the sativa varieties under both low and high input con- biotechnology tools such as anther culture and embryo res- ditions (see figure 15.A1), with what is in essence insur- cue techniques. In 1994 the first interspecific line with ance against the risk of significant yield losses in the face promising agronomic performance was obtained (Wopereis of major upland biotic and abiotic stresses. Given that the et al. 2008; Jones et al. 1997a, b).7 Several hundred interspe- sativa and glaberrima parents of these first-generation cific progenies were generated, opening new gene pools and NERICAs are not the best-performing varieties within the increasing the biodiversity of rice. two species, there is reason to expect that the perfor- The main objective of the breeding work that led to the mance of the next generations of NERICAs will be much NERICAs was to combine the high-yielding attribute of higher.9 O. sativa with the resistance of the indigenous O. glaberrima Through participatory varietal selection (box 15.1), to the African environment. Another long-sought attribute farmers all over Africa have evaluated interspecific lines. for a good upland variety is the ability to provide acceptable The most successful lines have been named NERICA vari- yields under the low-input use conditions typical of upland eties. There are currently 18 varieties suited for upland rice farming in Africa. Both objectives have been largely met, growth conditions (NERICA1 to NERICA18) and 60 vari- as evidenced by the experimental trials data summarized in eties suited for lowland growth conditions (NERICA-L1 to Box 15.1 The Participatory Varietal Selection Methodology Participatory varietal selection (PVS) aims to improve African O. glaberrimas, and local varieties as checks. rice production through direct involvement of farmers, Male and female farmers are invited to visit the plot who play an active role in varietal selection, develop- informally as often as possible. Groups of farmers for- ment, and dissemination. PVS is an adjunct to conven- mally evaluate the varieties at three key stages: maxi- tional breeding that markedly reduces the time taken mum tillering (development of the plant leaves), matu- for new varieties to reach farmers. In conventional rity, and postharvest. At the first two stages, farmers breeding schemes, selection and testing procedures compare agronomic traits. Each farmer’s varietal selec- involve a series of multilocation trials over 8 to 12 years tion and the criteria for selection are recorded and later using a diminishing number of varieties. Through PVS, analyzed. a large number of potentially interesting varieties reach In the second year, each farmer receives as many as farmers for evaluation in 5 years, and farmers have a six of the varieties he or she selected in the first year to major input into the selection of varieties to be released. grow. PVS observers, including breeders and techni- By involving farmers in the varietal selection process, cians from nongovernmental organizations and exten- PVS aims to meet a diversity of demands for varieties, sion services, visit participating farmers’ fields to record taking account of site-specificity for agronomic adapta- seed performance and farmer evaluations of the tion and selection criteria. selected varieties. At the end of the year, farmers evalu- AfricaRice’s approach to PVS research involves a ate threshability and palatability to provide an overall three-year program. In the first year, breeders identify view of the strengths and weaknesses of the selected centralized fields near villages and plant a rice garden varieties. In the third year, farmers are asked to pay for trial of up to 60 upland varieties. The varieties range seeds of the varieties they select. Willingness to pay pro- from traditional and popular O. sativas to NERICAs, vides evidence of the value farmers place on the seeds. Source: Gridley and Sié 2008. CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 255 NERICA-L60). Agronomic characteristics of these vari- hectares under NERICA1 in Nigeria in 2007 (ARI 2008). eties vary widely, but they are generally high yielding and The Uganda National Research Institute, citing a report of early maturing (a trait much appreciated by farmers), do the Statistics Office of the Ministry of Agriculture in not lodge or shatter, yield good grain quality, and are Uganda, estimates that 35,000 hectares were under NER- relatively resistant to the biotic and abiotic stresses of the ICA4 in 2007 (ARI 2008). continents harsh growth environment. Weed resistance of Except in Guinea and Benin, where surveys were con- the NERICAs, especially the upland NERICAs, still needs ducted (in 2003 in Guinea and in 2005 in Benin), the esti- improvement. (For details on the development and the char- mated areas under NERICA reported above by national acteristics of different NERICA varieties, see Somado, Guei, agricultural research systems are based on quantities of seed and Keya 2008; Wopereis and others 2008; and Kaneda produced and distributed to farmers. To meet the need to 2007a, b, c, and d.) quantify uptake of NERICA and other improved varieties in a much more reliable manner, AfricaRice and partners are currently conducting surveys in many countries where ADOPTION AND IMPACT OF NERICA VARIETIES NERICA is cultivated. Surveys conducted in 2009 by IN SUB-SAHARAN AFRICA national agricultural research systems and national agricul- The first generation of interspecific progenies was intro- tural statistical services in 21 Sub-Saharan African countries duced through participatory varietal selection trials in Côte provide more recent estimates for some countries d’Ivoire in 1996 and in other member countries of (AfricaRice 2010). According to these data, the area under AfricaRice in 1997. NERICA1 and NERICA2 were officially NERICA was about142,391 hectares in Guinea and 244,293 released in 2001 in Côte d’Ivoire (Diagne 2006; Science hectares in Nigeria in 2009.11 Council 2007). NERICA varieties have now been introduced in more than 30 countries in Sub-Saharan Africa using PVS Adoption of NERICA varieties approaches, mostly thanks to the African Rice Initiative (ARI).10 By 2007, 17 upland NERICA varieties and 11 low- A 2009 study on adoption in Sub-Saharan African countries land NERICA varieties had been adopted. Where needed, estimates actual rates of adoption of NERICA varieties ARI produced foundation seed to stimulate uptake of NER- (table 15.1).12 These rates illustrate a largely unrealized ICA and other improved varieties. Seed production by ARI potential adoption rate if the full rice farming population of has increased steadily, from 2,733 tons in 2005 to 7,238 tons these countries could be exposed to NERICA varieties and in 2007 and 13,108 tons in 2008 (Akintayo et al. 2009). provided access to seed. About 700,000 hectares in Sub-Saharan Africa were esti- Uptake of NERICA varieties was reportedly high in the mated to be under NERICA varieties as of 2009, covering an Kaduna and Ekiti states of Nigeria (Spencer et al. 2006). estimated 5 percent of the continent’s upland rice-growing About 30 percent of farmers in Ekiti cultivated NERICA1 in areas. This level of penetration is impressive 8–10 years after 2005; 42 percent of farmers in PVS villages and 19 percent release of the first varieties. The studies reviewed in this sec- in near–PVS villages did so in Kaduna.13 Adoption of NER- tion describe the areas under NERICA varieties in selected ICA1 appears to have continued during 2004 and 2005, countries, actual and potential rates of adoption among despite the scaling down of PVS activities. About 35 percent farmers, and the impact of NERICA varieties on rice pro- ductivity and the livelihoods of family farmers. Table 15.1 Estimated Actual and Potential Rates of Estimated area under NERICA Adoption of NERICA Varieties in Selected African Countries Surveys conducted by national agricultural research systems (percent) provide estimates of the total area under NERICA varieties Rate of Potential rate in different countries. Diagne et al. (2006) and Adegbola et Country (year) adoption of adoption al. (2006) estimate the total area under NERICA varieties at Benin (2004) 19 47 51,000 hectares in Guinea in 2004 and 5,000 hectares in Côte d’Ivoire (2000) 4 24 Benin in 2003. Based on its estimates of the quantities of Guinea (2001) 20 61 The Gambia (2006) 40 87 seed produced and distributed to farmers, the National Food Reserve Agency (NFRA) of Nigeria reports 186,000 Source: Diagne 2009. 256 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) of farmers in near–PVS villages in Ekiti and Kaduna had not Table 15.2 Yields and per Capita Income in heard about NERICA1, showing the potential for increase in Selected African Countries, by Gender adoption rates (Spencer et al. 2006). In 2005 an estimated 14 percent of farmers in PVS villages in Kaduna and 9 percent Yield of farmers in near–PVS villages had adopted NERICA2. Country (year) (tons/hectare) Income ($) The percentage of households that grew NERICA vari- Benin (2004) eties in Uganda increased from 0.9 percent in 2002 to 2.9 Men 0.26*** 22*** Women 0.97*** 38*** percent in 2003 and 16.5 percent in 2004, according to All 0.71*** 25*** Kijima, Otsuka, and Sserunkuuma (2008). They find that Côte d’Ivoire (2000) membership in a farmers group, formal education of the Men 0.44 — Women –0.45 — head of household, and the number of household members All ^0.44 — significantly increased the probability of adopting NERICA The Gambia (2006) varieties as well as the scale of area planted to NERICA vari- Men 0.18 4.6 Women 0.142*** 10.6** eties. Land size per person had a negative effect on the share All 0.146*** 10.2** of land planted to NERICA varieties, suggesting that land- Source: Authors compilation, based on data from Agboh- poor households tend to allocate a larger proportion of land Noameshie, Kinkingninhoun-Medagbe, and Diagne (2007) for to cultivation of NERICA varieties. This pattern is also Benin, Diagne; Midingoyi, and Kinkingninhoun (2009) for Côte observed among households headed by women, who are d’Ivoire; Dibba et al. (2008a, b) for The Gambia; Dontsop et al. (2010) for Nigeria; and Kijima, Otsuka, and Sserunkuuma (2008) likely to be poorer. Poorer households in Uganda thus tend for Uganda. to allocate a larger proportion of their land to NERICA vari- — = Not available. eties, which may suggest that the adoption of NERICA has *** Significant at the 1 percent level; ** significant at the 5 percent level. the potential to reduce poverty and improve income distri- bution as it increases rice yields. per hectare for men; the potential net income gain from Impact of NERICA varieties on productivity and adopting NERICA is estimated at about $337 per hectare for livelihoods of farmers women and $277 for men. Results in The Gambia indicate NERICA rice varieties were developed with the aims of that an additional yield gain of 0.14 ton per hectare was improving rice productivity, raising income, and reducing achieved by rice farmers, most of them women, adopting food insecurity of poor upland rice farmers (mostly NERICA varieties (Dibba 2010).14 Results from Côte women), who rarely use fertilizer, which they say they can- d’Ivoire show that the impact is heterogeneous, with a siz- not afford. Empirical evidence from impact assessment able and statistically significant impact found for female studies conducted in West Africa points to a heterogeneous farmers (0.7 ton per hectare) and a nonstatistically signifi- impact of NERICA adoption across and within countries cant impact found for male farmers (Diagne 2006; Diagne, (table 15.2). Impacts have been significantly positive in Midingoyi, and Kinkingninhoun 2009). Benin and Gambia. In almost all countries, the impacts have In Uganda NERICA had positive effects on productivity generally been higher for women than for men (Diagne, and allowed farmers to improve their yields (Kijima, Sserunk- Midingoyi, and Kinkingninhoun 2009). uuma, and Otsuka 2006; Kijima, Otsuka, and. Sserunkuuma In Benin farmers adopting NERICA enjoyed an addi- 2008; Bergman-Lodin 2005). According to Kijima, Sserunku- tional yield gain of about 1 metric ton per hectare. The uma, and Otsuka (2006), the average yield of NERICA in impact at the national level was very limited, however, Uganda was 2.2 tons per hectare—twice the average rice yield because of the low diffusion of NERICA varieties in Benin in Sub-Saharan Africa. They find large differences in yields (Adegbola et al. 2006). between experienced (2.46 tons per hectare) and inexperi- Results from another analysis based on data from the enced (1.72 tons per hectare) households, indicating that 2004 season show that the impact of NERICA adoption is experience matters in achieving high yields. Other yield higher for women than for men (Agboh-Noameshie, Kink- determinants were rainfall and cropping patterns. ingninhoun-Medagbe, and Diagne 2007). Female potential A few studies document the impact of adoption of NER- adopters have a surplus of production of 850 kilograms of ICA varieties on household income and poverty. Most find paddy per hectare, compared with 517 kilograms of paddy that adoption increases household income and improves CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 257 income distribution within households. In Uganda, for of average gross income from rice), increasing farmer’s prob- example, adoption of NERICA varieties has the potential to ability of escaping poverty (Dontsop et al. 2010). increase annual per capita income by $20 (12 percent of These results suggest that NERICA can bring hope to actual per capita income) and to reduce the incidence of millions of poor, small-scale farmers in Africa by reducing poverty, measured by the headcount ratio, by 5 percentage poverty and income inequality. Realization of this hope points (from 54.3 percent to 49.1 percent) (Kijima, Otsuka, requires wider dissemination of NERICA, however, which and Sserunkuuma 2008). The poverty gap index and the can take place only if the seed bottlenecks and other pro- squared poverty gap index also declined following the intro- duction constraints examined in the next section are duction of NERICA, suggesting that adoption can increase addressed. incomes among the absolute poor in Uganda. NERICA adoption has had a positive impact on house- LESSONS LEARNED hold rice and total incomes in The Gambia (Glove 2009; Dibba 2010). Dibba (2010) finds that adoption of NERICA The use of the PVS approach allowed farmers to evaluate increases rice farmers’ daily income by about $0.34 on aver- the new varieties in comparison with their own material. It age (about 30 percent of the average daily income of enhanced capacity building and ownership of the NERICA adopters). Sogbossi (2008) identifies NERICA adoption as a varieties among farmers and national research and exten- key determinant in poverty reduction in Benin, where it sion communities. PVS also reduced the time involved in reduced the probability of being poor by 13 percent.15 The the varietal release process in many countries. impact of NERICA adoption was higher for women farmers Different partnerships and approaches were essential for (reduction in probability of being poor was 19 percent) NERICA adoption and seed production (box 15.2). In many than for men (reduction in probability of being poor was countries the active involvement of government authorities 6 percent). Findings from a study conducted in Nigeria indi- at the highest level helped stimulate uptake. Major govern- cate that adoption of NERICA increased total farm house- ment initiatives on rice in Guinea, Mali, Nigeria, and hold gross income from rice by about $555 (about 46 percent Uganda focused on NERICA varieties. In Benin a private Box 15.2 Successfully Disseminating NERICA Varieties in Burkina Faso, The Gambia, Guinea, and Uganda Partnerships with the National Agricultural Research fertilizer and quality seed of high-yielding rice varieties, Extension Systems, nongovernmental organizations, including NERICA varieties. and farmer organizations, with support from donors and development partners, were key ingredients in scal- The Gambia ing up the dissemination of NERICA varieties through- NERICA varieties were introduced into The Gambia out Sub-Saharan Africa.a Governments across the through PVS trials in 1998, initially in three villages and region also supported the dissemination of NERICA later more widely through farmers’ own channels and varieties through major rice development initiatives efforts by the National Agricultural Research Institute well before the 2008 food crisis. and the Department of Agricultural Services. The uptake of NERICA varieties continued through the Burkina Faso African Rice Institute and NERICA dissemination proj- Domestic rice production in Burkina Faso increased by ect funded by the African Development Bank in 2004. an astonishing 241 percent in 2008 (Solidarité et Progres Even the head of state was growing NERICA varieties 2008), partly as a result of the adoption of NERICA vari- on his own land and encouraging farmers to grow these eties. This remarkable achievement reflected the govern- varieties. According to FAO data, from 2001 to 2007, the ment’s response to the world food crisis in 2008. Emer- rice-growing area was relatively constant, at about gency support facilitated farmers’ access to mineral 17,000 hectares. A doubling to 34,000 hectares occurred (continued next page) 258 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) Box 15.2 (continued) in 2008, partly as a result of the enthusiasm generated Bank funded the program through its Special Program by NERICA varieties, the support of the government, for Agricultural Research in Africa (SPAAR). IRAG and and the efforts of researchers and technicians. NERICA SNPRV worked together to select the most appropriate varieties are now cultivated in all six agricultural testing sites and methods of introduction, joining hands regions of the country. Introduction of the varieties has with an international nongovernmental organization, brought hope for increased rice productivity and Sasakawa Global 2000. This cooperation facilitated poverty reduction. access to inputs for farmers involved in the trials. AfricaRice staff provided training to IRAG and SNPRV Guinea scientists and technicians in the conduct of the PVS tri- als and seed production throughout the duration of the The first interspecific lines were introduced in Guinea program (Kaneda 2007d; Diagne 2006). in 1997 by the Institut de Recherche Agronomique de Guinée (IRAG) and the Service National de la Promo- tion Rurale et de la Vulgarisation (SNPRV) through an Uganda intensive program to introduce new upland rice vari- eties. NERICA varieties subsequently experienced a Rice farming is relatively marginal in East Africa, espe- very rapid spread in the regions of Upper Guinea and cially in Uganda. Before 2002 rice farming was practiced Forest Guinea after only two years of experimental test- mostly as a cash crop by only a few farmers on a very ing and PVS (Camara et al. 2002; Diagne 2006). A study small upland area and in very few small irrigated by researchers from AfricaRice, IRAG, and SNPRV in schemes. Research support had been abandoned in the the four regions of Guinea estimates the area under 1970s, and rice yields had gradually declined to about NERICA varieties at 51,000 hectares in 2003, slightly 0.4 ton per hectare in 2002. The introduction of NER- less than 10 percent of the 525,000 hectares used for rice ICA varieties enabled farmers to attain yields of up to production according to Food and Agriculture Organi- 3 tons per hectare on fertile land in Uganda. zation (FAO) statistics (FAOSTAT 2010)—a remarkable By 2002 NERICA varieties were being cultivated in achievement in a relatively short time. In 2007 the area many parts of Uganda. Rice acreage increased sixfold in under NERICA varieties in Guinea was estimated at the six years following the release of the new varieties, 82,930 hectares, 12 percent of the total rice area (ARI from 6,000 hectares in 2002 to 40,000 hectares in 2008, 2008). Estimates from the nationwide survey show that and the number of rice growers rose from 4,000 in 2004 the area under NERICA has grown significantly since to 35,000 in 2007 (Akintayo et al. 2009). Uganda 2007 to 142,391 hectares, representing about 18 percent reduced its rice importation from 60,000 tons in 2005 of the total rice area in Guinea (AfricaRice 2010). to 35,000 tons in 2007—an almost 50 percent reduction Several factors explain the success of NERICA vari- in rice imports, leading to savings of about $30 million eties in Guinea. First, NERICA varieties appeared when (Akintayo and others 2009). Several factors explain the Guinea was experiencing a period of prolonged drought rapid uptake in Uganda. The great interest shown by the that had pushed the Guinean government to seek assis- National Institute of Agricultural Research for the new tance from AfricaRice to put at farmers’ disposal short- varieties at the time of revitalizing its rice program and season varieties that could better cope with drought the dynamism of Sasakawa Global 2000 played very (Diagne 2006). NERICA varieties corresponded well to important roles in the introduction and diffusion of what was needed. Second, Guinean farmers were very NERICA varieties. The interest and dynamism of the receptive to experimenting with new varieties in their private sector was essential in producing and marketing fields. Third, an effective partnership between national seed of NERICA varieties, as was the active involvement and international institutions greatly contributed to the of the highest political authorities of the country, par- rapid adoption of the new rice varieties. The World ticularly Uganda’s vice president (Akintayo et al. 2009). a. Donors included the African Development Bank, the Food and Agricultural Organization, the Government of Japan, the Inter- national Development Research Centre, the International Fund for Agricultural Development, the Rockefeller Foundation, the United Nations Development Programme, the U.S. Agency for International Development, and the World Bank. CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 259 entrepreneur was instrumental in the diffusion of NERICA involved to some extent in rice research in AfricaRice mem- varieties, paying for demonstration plots across the country ber states, according to a 2008 survey—and most of them and promoting seed production (Akintayo et al. 2009). work on many other crops and spend only a fraction of their The ARI–coordinated NERICA dissemination project time on rice. Egypt alone accounts for the lion’s share of this funded by the African Development Bank and numerous research pool, with 50 highly qualified researchers working smaller projects run by the Japan International Coopera- full time on rice, including 12 rice breeders. In comparison, tion Agency, FAO, and many nongovernmental organiza- Nigeria has only 2 rice breeders. tions (NGOs) greatly accelerated the uptake of NERICA Other issues include lack of access to seed, fertilizer, and varieties across Sub-Saharan Africa (map 15.1). Despite credit and inadequate rice production, processing, distribu- these efforts, seed production remains a major bottleneck, tion, and marketing infrastructure. These factors need to be especially in West and Central Africa, where many farm- identified, and relevant actions taken to develop the rice sec- ers lack access to improved rice varieties. In many coun- tor. An integrated approach from seed to plate is needed. tries, national seed regulatory bodies do not exist or are Africa has tremendous potential to expand the area dysfunctional. Functioning bodies are essential to esti- under rice cropping, especially in the rainfed lowland mate and meet the demand for rice seed for different ecosystem. It has been estimated that 200 million hectares of ecosystems and consumer preferences. Incentives should rainfed lowland area could be developed for rice across the be provided to small-scale enterprises to develop viable continent. If just 1 percent of this land were allocated to seed businesses. growing rice and the average yield was just 3 tons per The data on potential adoption rates of NERICA vari- hectare, Africa would produce another 6 million tons of rice eties illustrate its tremendous potential to reach still more a year, making a major dent in the 10 million tons of rice it farmers. The constraints to doing so are numerous and vary imports. The lowlands offer great potential for the sustain- from country to country. One major factor is the desperate able expansion, intensification, and diversification of rice- lack of capacity at all levels in the rice value chain, especially based systems. Given this potential, the lowland NERICA the neglect of Africa’s rice research capacity. Only about varieties and their successors are expected to have an even 250–275 researchers (including only about 15 women) are greater impact than the upland NERICA varieties. Map 15.1 Areas of Sub-Saharan Africa Producing NERICA Varieties, 2005 and 2006 a. 2005 b. 2006 IBRD 38555 MAY 2011 PRODUCTION STATUS (2005): HIGH: >10,000 ha. NERICA DISTRIBUTION MEDIUM: 5,000– IN AFRICA 10,000 ha. LOW: <5,000 ha. 0 1,500 Km. 0 1,500 Km. INTERNATIONAL BOUNDARIES 0 1,000 Mi. INTERNATIONAL BOUNDARIES 0 1,000 Mi. Source: Akintayo et al. 2009. 260 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) CONCLUSIONS including farmers who have never seen a new rice variety, can profit from these new varieties, especially in rainfed As a result of the increase in rice consumption across Sub- ecologies. Saharan Africa, the continent accounted for 30 percent of AfricaRice is already working on new varieties that will world rice imports in 2009, costing African economies an replace the first generation of NERICA varieties, in close estimated $4 billion. Relying on the world market to feed collaboration with national agricultural research institutes, Africa’s rice consumers is a risky and unsustainable strategy. the International Rice Research Institute, and other interna- Encouraging the development of Africa’s rice sector is cru- tional research institutes in Japan, Europe, and the United cial for food security. The sector could also serve as an States. These efforts, together with a better definition of engine for economic growth (IFPRI 2006). farmers’ needs and consumer demands, are expected to lead Development of Africa’s rice sector will require an inte- to the development of varieties that will eventually outper- grated seed to plate value chain approach. Much work is still form the current generation of NERICA varieties. needed to identify the varieties that respond to consumer AfricaRice is an active member of the Coalition of demand and grow well in the various ecosystems in which African Rice Development, an initiative of the Japan Inter- rice is grown in Africa, under both current and future cli- national Cooperation Agency and the Alliance for a Green mate scenarios. Much greater input in varietal development Revolution in Africa, which aims to double rice produc- is needed at the national level, and greater capacity in rice tion in Sub-Saharan Africa between 2008 and 2018. Such breeding is needed across the continent. partnerships and commitment at the national and regional NERICA varieties have shown great potential in both level to hire, train, and retain new staff in rice research and upland and lowland ecosystems in Africa, where they cur- invest in rice production, storage, processing, and distri- rently cover more than 700,000 hectares. Women farmers in bution capacity will be instrumental in boosting Africa’s particular have profited from NERICA varieties. More effort rice sector. is needed to ensure that larger numbers of rice farmers, CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 261 ANNEX 15.A1 Table 15.A1 Performance of NERICA Progenies and Their Parents under Various Conditions Yield under Yield under acid Yield under high- and drought- and and nonacid soil Susceptibility low-input conditionsa nondrought-prone sites sites to gall midge Item Low input High input Drought Nondrought Acid Non-acid attack NERICA progeny line Sample size 19 19 16 16 25 25 100 Mean (across replications and lines) 1.16 3.07 1.62 2.05 1.99 3.19 40.72 Standard deviation (across lines) 0.16 0.50 0.39 0.36 0.27 0.40 8.95 Minimum 0.91 1.70 1.08 1.49 1.57 2.15 8.76 Maximum 1.47 3.86 2.27 2.66 2.63 3.96 58.69 Sativa parent (WAB56-104) Mean value (across replications) 0.90 3.26 1.19 1.96 1.35 3.26 52.77 T-test for equality with mean of NERICA lines p-value 0.00 0.12 0.00 0.33 0.00 0.36 0.00 Glaberrima parent (CG14) Mean value (across replications) 0.64 1.00 1.24 0.86 1.23 1.00 T-test for equality with mean of NERICA lines p-value 0.00 0.00 0.00 0.00 0.00 0.00 Sativa tolerant checkb Mean value (across replications) 1.51 1.96 12.07 T-test for equality with mean of NERICA lines p-value 0.25 0.33 0.00 Sativa susceptible checkc Mean value (across replications) 0.49 1.12 38.36 T-test for equality with mean of NERICA lines p-value 0.00 0.00 0.01 Source: Authors, based on data from experiment in 1999 at AfricaRice headquarters (in M’bé) and at two of its on-farm research sites (in Man and Korhogo). The trials used a randomized complete block design with three replicates at each site: the NERICA progenies, their two sativa and glaber- rima parents (WAB56 104 and CG14), and a select set of sativa checks, which varied across the trials. The set of NERICA progenies also differed across trials. a. Yield data are mean values for two sites (M’bé and Man) from a 1999 randomized complete block design trial with three replicates at each site. b. Moroberekan, a local sativa variety commonly used as check for drought. c. IR20, an improved sativa variety used as check for drought. 262 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) Figure 15.A1 Performance of NERICA Progenies under Low and High Input Conditions and Major Stresses in Upland Rice Ecosystems a. Yield under low and high b. Yield under drought and 5 input conditions nondrought conditions 3.0 4 2.5 2.0 Yield (tones/ha) Yield (tones/ha) 3 1.5 2 38 1.0 1 0.5 0 0 Low input use High input use Nondrought site Drought site NERICA lines Sativa parent Glaberrima parent Sativa tolerant check Sativa suceptible check c. Yield under acid and nonacid soils d. Susceptibility to African gall midge attack 4.5 70 Score for susceptibility to AfRGM attack 60 4.0 197 50 3.5 40 3.0 Yield (tones/ha) 30 2.5 20 2 21 10 2.0 1 0 102 1.5 –10 es nt e ch sa 1.0 ch re lin al le nt pa on ib A ta pt IC iti a sis tiv ce ad ER Sa re su 0.5 tr N a er a tiv tiv th Nonacid soil Acid soil Sa Sa O NERICA lines Sativa parent Glaberrimas parent Source: Authors, based on data from experiment in 1999 at AfricaRice headquarters (in M’bé) and at two of its on-farm research sites (in Man and Korhogo). The trials used a randomized complete block design with three replicates at each site: the NERICA progenies, their two sativa and glaberrima parents (WAB56 104 and CG14), and a select set of sativa checks, which varied across the trials. The set of NERICA progenies also differed across trials. Note: Sample sizes are 19 for panel a, 16 for panel b, 25 for panel c, and 100 for panel d. CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 263 NOTES 5. Seventy percent of the production increase came from land expansion, with just 30 percent attributable to produc- 1. The role of technological innovation in raising agricul- tivity enhancement. tural productivity and fostering overall agricultural devel- opment is documented in the World Bank’s 2008 World 6. Lodging is the process in which rice plants fall over at, or Development Report. just before, maturity; shattering is the process in which plants’ panicles scatter their seeds on the ground at maturity. 2. The Africa Rice Center—AfricaRice for short—is the new official name, adopted in September 2009, for the 7. NERICA rice varieties are interspecific but not geneti- organization previously known as the West Africa Rice Devel- cally modified. opment Association (WARDA). This intergovernmental asso- 8. AfricaRice has five on-farm research sites in Côte ciation of African countries is one of the 15 international d’Ivoire, known as “key sites.” The sites—selected in the agricultural research centers supported by the Consultative early 1990s, when AfricaRice moved its headquarters to Group on International Agricultural Research (CGIAR). Côte d’Ivoire from Liberia—were chosen to cover all nine The center was created in 1971 by 11 African countries. rice ecologies and the main rice-producing regions in the With the continuing expansion of its activities into Central, forest and savanna agroecological zones of Côte d’Ivoire and East, and North Africa, the Council of Ministers passed a to be representative of almost all of the rice- growing ecolo- resolution at their 27th meeting, held in LomÈ, Togo, in gies in West Africa. September 2009, to change its name from WARDA to 9. NERICA varieties have retained only about 5–10 percent AfricaRice. Today its membership comprises 24 countries: of the genome of the African parent, because of backcrosses Benin, Burkina Faso, Cameroon, Central African Republic, to the O. sativa parent. They are also based on crosses Chad, Côte d’Ivoire, Democratic Republic of Congo, Arab between only a handful of O. sativa and O. glaberrima acces- Republic of Egypt, Gabon, The Gambia, Ghana, Guinea, sions. AfricaRice breeders have since developed “bridge Guinea-Bissau, Liberia, Madagascar, Mali, Mauritania, varieties” that show full fertility when crossed with Niger, Nigeria, Republic of Congo, Senegal, Sierra Leone, O. glaberrima, potentially allowing the full exploitation of Togo, and Uganda. all 2,300 accessions of the O. glaberrima species in the gene 3. The importance of the development of NERICA vari- bank of AfricaRice and a much greater retention of the eties has been widely recognized. The Consultative Group on O. glaberrima genome. Molecular assisted breeding is also International Agricultural Research (CGIAR) presented the becoming a routine tool, accelerating varietal development. King Baudouin Award to the Africa Rice Center (then 10. ARI was established in 2002 by AfricaRice and key WARDA) in 2000 for the development of NERICA varieties. donor partners, including the African Development Bank, In 2004 the center’s Dr. Monty Jones became the first African the government of Japan, the Rockefeller Foundation, and to receive the World Food Prize, for his contribution to the the United Nations Development Programme, in order to development of the family of NERICA rice varieties suitable facilitate the dissemination of NERICA varieties and other for upland growth conditions. Dr. Moussa SiÈ, a senior rice improved varieties in Sub-Saharan Africa. This network of breeder at AfricaRice, was awarded the prestigious Interna- rice research and extension organizations and private sector tional Koshihikari Rice Prize of Japan, in recognition of his partners plays an important role in making quality rice seed leadership in developing the family of NERICA rice varieties available. This role is crucial, as seed availability is one of the suitable for lowland growth conditions. weakest links in the rice value chain in Africa. 4. Rice is generally associated with Asia, but it is also an 11. Data from these national surveys, which are still being integral part of the history and culture of Africa, where it has processed, are expected to provide more accurate estimates been grown for more than 3,000 years. Oryza glaberrima was of the total area under NERICA in Sub-Saharan Africa. Pre- grown largely in the central delta of the Niger River and in liminary figures from these surveys and other secondary the Higher Gambia and the Casamance region of what is sources suggest that about 700,000 hectares was under now Senegal (Carney 1998; Poteres 1962, cited in Linares NERICA in Sub-Saharan Africa in 2009. 2002). Oryza glaberrima was the only rice species grown in 12. The potential adoption rate estimated here is the adop- West Africa until the 16th century, when the Portuguese tion rate when the entire population of rice farmers is made introduced the Asian species Oryza sativa in the West African aware of the existence of the NERICA varieties (Diagne and coastal regions (Linares 2002). Over the years, O. sativa grad- Demont 2007). Observed adoption rates are lower than ually replaced O. glaberrima, which has some unfavorable these estimated potential adoption rates because many characteristics. Farmers who still grow O. glaberrima do so farmers simply lack awareness of the NERICA’s existence. mostly for social or religious reasons, for its culinary and 13. PVS villages are villages in which NERICA varieties were nutritional qualities, or because it is better adapted to their tested and evaluated by farmers through participatory vari- agroclimatic environments. etal selection (PVS) trials. The nearby villages are villages 264 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) not hosting PVS trials but within 2–5 kilometers radius of Bentley, J., C. Almekinders, D. Yawovi, and D. Dalohoun. the selected PVS-hosting villages. 2008. Center Commissioned External Review of the 14. With women making up 93 percent of sample farmers African Rice Initiative. AfricaRice, Cotonou. in The Gambia (sample includes 39 men and 532 women), Bergman-Lodin, J. 2005. “The NERICA Conundrum: From it is not surprising that the average for the whole sample and Rice to Riches.” Lund University, Department of Eco- the sample for women are almost the same. The average nomic and Social Geography, Lund, Sweden. gain for men is not significant. Although seemingly small, Camara, B., A. Conde, M. S. Diallo, M. B. Barry, B. Sekou, the 0.14 ton per hectare average gain from NERICA adop- and T. Berhe. 2002. NERICA, the Guinea Experience. tion is statistically significant and represents about 15 per- Report of the Departments of National Research, Exten- cent of the average rice yield obtained by farmers (just sion and Rural Development and Sasakawa Global under 1 ton per hectare). 2000/Guinea Agricultural Project, Conakry. 15. The poor are defined using the income poverty line in Carney, J. A. 1998. “The Role of African Rice and Slaves in Benin, estimated at CFA 51,413 a year (about $100) in rural the History of Rice Cultivation in the Americas.” Human areas and CFA 91,709 a year (about $180) in urban areas Ecology 26 (4): 525–45. (Sogbossi 2008). Dalton, Timothy. J. 2004. “A Household Hedonic Model of Rice Traits: Economic Values from Farmers in West Africa.” Agricultural Economics 31: 149–59. REFERENCES Diagne, A. 2006. “Le NERICA: mythe et réalités?” Grain de Adégbola, P. Y., A. Arouna, A. Diagne, and A. A. Souléïmane. sel, la revue de l’inter-réseaux (March–August): 34–37. 2006. “Evaluation de l’impact économique des nouvelles ———. 2009. “Technological Change in Smallholder Agri- variétés de riz NERICA au Bénin: évidence avec les mod- culture: Bridging the Adoption Gap by Understanding Its èles basés sur l’approche contre factuel.” Paper presented Source.” Paper presented at the conference on Agricul- at the First Africa Rice Congress, Dar es Salaam, July ture for Development sponsored by the African Eco- 31–August 4. nomic Research Consortium, Mombasa, May 28–29. Adékambi, A. S., A. Diagne, F. Kinkingninhoun-Medagbe, Diagne, A., and M. Demont. 2007. “Taking a New Look at and G. Biaou. 2009. “The Impact of Agricultural Tech- Empirical Models of Adoption: Average Treatment Effect nology Adoption on Poverty: A Case Study of NERICA Estimation of Adoption Rates and Their Determinants.” Varieties in Benin.” Paper presented at the meetings of Agricultural Economics 37 (2–3): 201–10. the International Association of Agricultural Econo- Diagne, A., G. S. Midingoyi, and F. Kinkingninhoun. 2009. mists, Beijing, August 16–22. “The Impact of NERICA Adoption on Rice Yield in West AfricaRice. 2010. Rice Data Systems for Sub-Saharan Africa: Africa: Evidence from Four Countries.” Paper presented Contribution to the Japan–AfricaRice Emergency Rice at the meetings of the International Association of Agri- Project. Updated synthesis report submitted to the gov- cultural Economists Conference, Beijing, August 16–22. ernment of Japan, September 30, Cotonou. Diagne A., M. J. Sogbossi, S. Diawara, and Camara, A. 2006. Afriquenligne 2009. “Le riz NERICA pour assurer la sécurité L’étendu du succès de la dissémination des variétés NER- alimentaire.” In Actualités africaine en ligne. http://www ICA en Guinée: estimation des superficies emblavées. West .afriquenligne.fr/actualites/economie/le-riz-nerica- Africa Rice Development Association, Cotonou. pour-assurer-la-securite-alimentaire-200804252475 Dibba, L. 2010. “Estimation of NERICA Adoption Rates and .html. Impact on Productivity and Poverty of Small-Scale Rice Agboh-Noameshie, A. R., F. M. Kinkingninhoun-Medagbe, Farmers in The Gambia.” M.Phil. thesis, Kwame and A. Diagne. 2007. “Gendered Impact of NERICA Nkrumah University, Faculty of Agriculture, Depart- Adoption on Farmers’ Production and Income in Cen- ment of Agricultural Economics, Agri-business and tral Benin.” Paper presented at the meetings of the Extension, Kumassi, Ghana. African Association of Agricultural Economists, August Dibba, L., A. Diagne, and S. Fialor. 2010. “Impact of NER- 20–22, Accra. ICA Adoption on Productivity and Poverty of the Small- Akintayo, I., O. Ajayi, R. A. Agboh-Noameshie, and B. Cissé. Scale Rice Farmers in The Gambia.” Kwami Nkrumah 2009. Report of the Seventh African Rice Initiative Steering University of Science and Technology Kumasi Ghana and Committee/Experts Meeting, May 4–6, Cotonou. Africa Rice Center, Cotonou. ARI (African Rice Institute). 2008: Initiative africaine sur le Dibba, L, A. Diagne, F. Simtowe, M.J. Sogbossi and M. riz: rapport annuel 2008. Centre du Riz pour l’Afrique Mendy. 2008a. “Diffusion and Adoption of NERICA in ADRAO, Cotonou. Gambia.” Africa Rice Center, Cotonou. CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 265 ———. 2008b “The Impact of NERICA Adoption on Yield Rice.” Journal of Japanese Tropical Agriculture 51 (3): in Gambia. Africa Rice Center Cotonou. 79–83. Dontsop, P., A. Diagne, V. O. Okuruwa, and V. Ojehomon. ———. 2007d. “Breeding and Disseminating Efforts of 2010. “Impact of Improved Rice Technology on Income NERICA: Efforts for Dissemination of NERICA in and Poverty among Rice Farmers in Nigeria: The Case of Africa.” Journal of Japanese Tropical Agriculture (51): 4: NERICA Varieties.” University of Ibadan Nigeria and 145–51. Africa Rice Center, Ibadan. Kijima, Y., K. Otsuka, and D. Sserunkuuma. 2008. “Assessing FAO (Food and Agriculture Organization). 2009. “The Spe- the Impact of NERICA on Income and Poverty in Cen- cial Challenges for Sub-Saharan Africa.” High-Level tral and Western Uganda.” Agricultural Economics 38: Experts’ Forum, Rome, October 12–13. 327–337. ———. 2010. Food and Agricultural Organization statis- ———. 2009. “Determinants of Changing Behaviors of tics. http://www.fao.org/corp/statistics/en/. NERICA Adoption: An Analysis of Panel Data from Glove, S. 2009. “Agricultural Technology Dissemination in Uganda.” University of Tsukuba, Graduate School of Sys- West Africa: A Quantitative Analysis of NERICA Adop- tems and Information Engineering, Tsukuba, Ibaraki, tion and Its Impact in Guinea and The Gambia.” MSc. Japan. thesis, University of London, Department of Develop- Kijima, Y., D. Sserunkuuma, and K. Otsuka. 2006. “How ment Economics, the School of Oriental and African Revolutionary Is the ‘NERICA Revolution’? Evidence Studies. from Uganda.” Developing Economies 44 (2): 252–67. Gridley, H., and M. Sié. 2008. “The Role of PVS: Farmer in Linares, O. F. 2002. African Rice Oryza Glaberrima: History the Driving Seat.” In NERICA: New Rice for Africa. A and Future Potential. Smithsonian Tropical Research Compendium, ed. E. Somado, R. Guei, and S. O. Keya. Institute, Balboa-Ancon, Panama. AfricaRice, Cotonou. Ndjiondjop, M-N., K. Semagn, M. Cissoko, M. P. Jones, and IFPRI (International Food Policy Research Institute), in col- S. McCouch. 2008. “Molecular Characterization of NER- laboration with IITA (International Institute of Tropical ICA Lines.” In NERICA: New Rice for Africa. A Com- Agriculture), CORAF/WECARD (Conseil Ouest et Cen- pendium, ed. E. Somado, R. Guei, and S. O. Keya. Coto- tre Africain pour la Recherche et le Développement Agri- nou: AfricaRice. coles/West and Central African Council for Agricultural Porteres, R. 1962. “Berceaux agricoles primaires sur le con- Research and Development), and ECOWAS (Economic tinent africain.” Journal of African History 3 (2): 195–210. Community Of West African States). 2006. Regional Science Council. 2007. Report of the Fifth External Program Strategic Alternatives for Agriculture-Led Growth and and Management (EPMR) of the Africa Rice Center Poverty Reduction in West Africa. Final Draft Report. (WARDA). Consultative Group on International Agri- Washington, DC. cultural Research (CGIAR), Science Council Secretariat, IRAT (Institut de Recherche Agronomique Tropicale). 1967. Rome. “Les variétés de riz au cercle de Banfora. IRAT/Haute Seck P. A., E. Tollens, M. C. S. Wopereis, A. Diagne, and I. Volta.” L’Agronomie Tropical 24: 691–707. Bamba. 2010. “Rising Trends and Variability of Rice Johnson, D. E 1997. Weeds of Rice in West Africa. West Africa Prices: Threats and Opportunities for Sub-Saharan Rice Development Association (WARDA). Bouaké, Côte Africa.” Food Policy 35 (5): 403–11. d’Ivoire. Sogbossi, M-J. 2008. “Riziculture et réduction de la pauvreté Jones, M. P., M. Dingkuhn, G. K. Aluko, and M. Semon. au Bénin: impact de l’adoption des nouvelles variétés riz- 1997. “Interspecific Oryza sativa × O. glaberrima Steud. icoles sur la pauvreté et l’inégalité du revenu dans le Progenies in Upland Rice Improvement.” Euphytica 92: Département des Collines Bénin.” Master’s thesis, Univer- 237–46. sity of Abomey-Calavi, Département d’Economie Socio- Kaneda, C. 2007a. “Breeding and Disseminating Efforts of anthropologie et Communication, Abomey-Calavi, Benin. NERICA: Breeding of Upland Rice.” Journal of Japanese Solidarité et Progrès. 2008. “Mauvais temps pour les Tropical Agriculture 51 (1): 1–4. malthusiens! La production de riz explose en Afrique.” ———. 2007b. “Breeding and Disseminating Efforts of http://www.solidariteetprogres.org/article5783.html. NERICA: Evaluation of Important Characteristics.” Somado, E., R. Guei, and S. O. Keya. 2008. NERICA: New Journal of Japanese Tropical Agriculture 51 (2): 41–45. Rice for Africa. A Compendium. Cotonou: AfricaRice. ———. 2007c. “Breeding and Disseminating Efforts Spencer, D., A. Dorward , G. Abalu, D. Philip, and D. Ogung- of NERICA: Important Characteristics of Lowland bile. 2006. Evaluation of Adoption of NERICA and Other 266 CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) Improved Upland Rice Varieties Following Varietal Promo- Innovation for African Farmers: A Response to Orr et tion Activities in Nigeria. Reported prepared for the al. from the Africa Rice Center.” Outlook on Agriculture Gatsby and Rockefeller Foundations, New York. 37 (3): 169–76. Wopereis, M. C. S., A. Diagne, J. Rodenburg, M. Sié, and World Bank. 2008. World Development Report 2008: Agricul- E. A. Somado. 2008. “Why NERICA Is a Successful ture for Development. Washington, DC: World Bank. CHAPTER 15: INCREASING RICE PRODUCTIVITY AND STRENGTHENING FOOD SECURITY THROUGH NEW RICE FOR AFRICA (NERICA) 267 CHAPTER 16 Fertilizer in Kenya: Factors Driving the Increase in Usage by Smallholder Farmers Joshua Ariga and T. S. Jayne ertilizer use is notably lower in most of Sub-Saharan F food is an important goal, the problem of access to food is Africa than in other developing regions. Too little perhaps the primary cause of food insecurity in Kenya. irrigation and varieties unresponsive to fertilizer may Ensuring access to food requires that the poor are able either explain this low use to some degree. But more often, the to produce or to buy enough food for a healthy diet. causes are lack of credit, long distances between farmers and Because of unpredictable weather and poor infrastruc- the nearest fertilizer retailer, weak market infrastructure, and ture, producers and consumers of agricultural products in lack of government support. Indeed, in many countries, the Kenya face volatile market prices, with periods of surplus withdrawal of state input delivery systems has led to a reduc- production providing a boon for consumers and periods of tion in fertilizer use as commercial distribution systems deficit benefiting a relatively few producers who are net sell- compete with subsidized government programs. ers and hurting the majority of consumers. The justification Kenya, however, stands as a notable departure from this for state participation in the market for agricultural prod- common Sub-Saharan African narrative. In the early 1990s ucts in Kenya has been to maintain prices at levels that both fertilizer markets were liberalized, government price con- provide incentives to raise farm incomes for producers and trols and import licensing quotas were eliminated, and fer- ensure that consumers can access food, the classic “food tilizer donations by external donor agencies were phased price dilemma.” out. Fertilizer use then almost doubled over the 15-year Before 1990 Kenya addressed this dilemma through direct period from 1992 to 2007, with much of the increase attrib- participation in input and output markets for national utable to smallholder farmers. In the productive farming “strategic” crops through either state-run agencies that set areas of western Kenya, rates of fertilizer application on prices at panterritorial levels or ostensible farmer organiza- maize are comparable with rates in Asia and Latin America. tions that were managed by state-connected political agents Kenya’s economy is predominantly agrarian, with more or their surrogates. For example, in the coffee sector, the gov- than 70 percent of its people dependent on agriculture- ernment helped enact laws that created the Coffee Board of related farm and off-farm activities for their livelihoods Kenya and Kenya Planters Cooperative Union; for (Ministry of Agriculture 2004). Food security is a concern pyrethrum (chrysanthemums), it encouraged the creation of for the 60 percent of the population living below the $1-a- the Pyrethrum Board of Kenya; and for milk, tea, and maize, day poverty line. While increasing the available supply of the government helped create the Kenya Cooperative Joshua Ariga is a senior research fellow with the Tegemeo Institute of Agricultural Policy and Development, Egerton University, and a research specialist at Michigan State University. T. S. Jayne is Professor, International Development, Michigan State University. 269 Creameries, the Kenya Tea Development Agency, and the marketing agencies such as the NCPB (which deducted the National Cereals and Produce Board (NCPB). In its heyday, cost of the loans on behalf of the KFA and AFC). To deal the NCPB generally bought maize grain from farmers at with high prices and a weak distribution network for small- higher-than-market prices and sold it to industrial maize holder farmers, fertilizer subsidies were also introduced millers at below-market prices. On the input side, the 1970s through these agencies (Ariga, Jayne, and Nyoro 2006). This and 1980s brought the formation of the state-run Kenya conflict of interest across interlinked agencies generated National Trading Corporation (KNTC) and the Kenya widespread corruption and bureaucratic costs that led to a Grain Growers Cooperative Union (KGGCU), later the policy change in 1972 in favor of introducing another Kenya Farmers Association (KFA), which worked with the agency (KNTC), which was tasked with importing fertilizer, output organizations. while the KFA was to be the distributor. This shift was Of all the crops mentioned in the previous paragraph, intended to increase competition within Kenya’s market for policy makers in Kenya—and indeed, throughout East fertilizer, but it did not succeed in keeping fertilizer prices at Africa—have been most concerned with increasing fertilizer low levels, and the agencies, influenced by the state, fell into use on maize, the main food security crop in the region. For the same patterns of bureaucracy and corruption as before. a number of reasons, however, state efforts in the 1980s to On the output side, the NCPB controlled maize prices at all improve food security through increased production and levels of the market chain (Nyoro, Kiiru, and Jayne 1999). By incomes did not produce desired results. As a result, several setting fixed panterritorial prices and removing arbitrage reform measures were implemented that sought to achieve opportunities for all market participants, these entities sti- food security objectives in a more efficient way, in the lines fled private trade. Requirements that private traders apply of a laissez faire or competitive markets dogma. for permits to transport grain across district boundaries made the situation more difficult. In the 1980s the government started relaxing its monop- Fertilizer and maize market reforms oly, allowing the private sector to compete with state agen- The period before market reforms in Kenya was character- cies, albeit under state rules. Fertilizer traders were to adhere ized by a predictable pattern involving the participation of to official prices, and the state influenced competition state-run agencies or private farmer organizations (with through strict trade licensing requirements and control of heavy state intervention in their management) in input and the allocation of scarce foreign exchange to importers (Arg- output markets for import and export, distribution, and wings-Kodhek 1996). Licensing and allocation of foreign retailing. Although these state agencies continually rein- exchange provided rent-seeking opportunities for public vented themselves under different names, particularly when sector officials (Kimuyu 1994). While the controlled pricing they came under scrutiny for corruption and unsustainable structure was designed to improve farmers’ access to fertil- budgets, their reincarnation followed the same general izer, it had the opposite effect in geographically remote modus operandi, and all eventually failed to achieve their areas, where the controlled prices were too low for fertilizer goal of improving the livelihoods of smallholder farmers. retailers to recoup the costs of transporting fertilizer from To put the reform process into perspective, it is impor- district towns to remote areas. Hence, retailers in remote tant to recognize that agricultural policy in Kenya has locations were less likely to stock fertilizer than those in also gone through a number of key phases characterized more urbanized areas, leaving the average distance traveled by an unpredictable shelf life. In the immediate postinde- by farmers to procure fertilizer relatively high in the 1980s pendence period (late 1960s), agricultural policy was and early 1990s. concerned with supporting a smooth transfer of prime In addition to the market inefficiencies they created, state land from white settlers to indigenous Kenyans with help agencies also imposed a heavy burden on public resources in from state-supported agencies in the production and mar- Kenya, contributing to deficits and inflation in the 1980s. A keting of produce (such as NCPB for maize). Agricultural decline in budgetary support to the agricultural sector by inputs were marketed through the farmers’ union, KFA, late 1980s probably contributed to the subsequent decline in and credit was provided through the Agricultural Finance agricultural growth, as did the mismanagement of agricul- Corporation (AFC). tural institutions, the ad hoc reform agenda, withholding of In the 1960s KFA could, for instance, offer inputs on donor funds over disagreements about democracy and gov- credit (through AFC) to select farmers, who repaid the ernance, and depreciation of the Kenyan shilling, the last of union after harvesting the crop and delivering it to relevant which raised input prices (Argwings-Kodhek 2004). In the 270 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS late 1980s and early 1990s, however, Kenya began easing policy papers emphasized a multisectoral approach to rural trade restrictions in the fertilizer and maize markets. The development in Kenya, including private-public synergies in government started removing some import quota restric- development. By 1996, 12 major importers, 500 wholesalers, tions, for example, in January 1990 and abolished licensing and roughly 5,000 retailers were distributing fertilizer in requirements for fertilizer imports in 1992. Kenya (Allgood and Kilungo 1996), and by 2000 the num- In a major policy change, the government liberalized the ber of retailers was estimated to have risen to between 7,000 fertilizer subsector specifically in 1993 to allow the partici- and 8,000 (IFDC 2001). (These are estimates because there pation of the private sector in importing, local trading, and is no comprehensive business registry or database covering distribution of fertilizer. Coupled with the liberalization of all types of businesses in Kenya.) Even with the easing of the foreign exchange regime in 1992, these changes in the trade restrictions, the high costs of upland transportation policy environment led to the entry of a significant number and logistical problems at the port of Mombasa continued of private sector firms in importing, wholesaling, distribu- to inflate the cost of fertilizer and reduce effective demand tion, and retailing of fertilizer (Wanzala 2001). Government (Wanzala 2001; Ariga and Jayne 2008). Although markups price controls and import licensing quotas were ultimately are less than 11 percent of the farm gate price of fertilizer in eliminated, and fertilizer donations by external donor agen- western Kenya, in a number of farming areas no fertilizer is cies were phased out. Maize trading controls were relaxed in applied because of the risk from markets, poor rainfall, and the early 1990s to allow private traders to transport a few agricultural conditions. But in other areas, application rates bags across districts with permission from government offi- rival those of Asia, and selling fertilizer is clearly profitable. cials, a situation that led to rent-seeking behavior and The overall trend in national consumption of fertilizer has increased costs for businesses (Kimuyu 1994). The NCPB, followed a steady growth path since 1990, with government however, helped stabilize prices for producers and con- imports declining and the role of the private sector increas- sumers by continuing to buy maize (mostly from large pro- ing (figure 16.1). ducers) at above-market prices and, during shortages, sell- Since 2007 a major escalation in the world price of fertil- ing it to consumers at subsidized prices (Jayne, Myers, and izer has led to increased government involvement in fertil- Nyoro 2008). izer marketing. The post-2007 period, which has been With the participation of stakeholders from all facets of marked by uncertain policy regimes, follows a fairly stable society in the 1990s and 2000s, a number of government and transparent period since 1993. In Kenya, in a move to Figure 16.1 Trends in Consumption, Commercial Imports, and Donor Imports of Fertilizer in Kenya, 1990/91–2010/11 600 500 Thousands of metric tons 400 300 200 100 0 9 1 3 5 7 9 1 1 3 5 7 /9 /0 /0 /0 /0 /0 /1 /9 /9 /9 /9 98 00 02 04 06 08 10 90 92 94 96 19 20 20 20 20 20 20 19 19 19 19 Imports Consumption Donor / state imports Projected Source: Estimated from Kenya Ministry of Agriculture data by the authors. Note: In 2004 and 2008, respectively, NCPB imported approximately one-third and 40 percent of national fertilizer needs. The estimate given for 2010/11 is a projection that includes private and government imports. The shaded years cover the time period after 2006/07, when government imports/subsidies were reinstated partly as a reaction to deficits in maize production and postelection disruptions of agricultural activities. CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 271 bolster agricultural production after postelection violence alization of the maize sector, especially the role played by led to disruption of farm activities, NCPB imported fertil- NCPB, the main grain-marketing state agency. izer in 2008 but delivered it to farmers late, which con- Although the increased participation of the Kenyan gov- tributed to the low levels of maize production that year. This ernment in the agricultural market was expected to be situation, in turn, created pressure from some farmer lobby short-lived and not significant enough to disrupt private groups and activists for increased subsidization of inputs sector investments, unforeseen events (such as poor rains in (fertilizer and seed) to raise productivity of maize to recent years) may mean that occasional state subsidies will counter an expected increase in hunger in 2009. In 2009 the be implemented over the upcoming few years in an attempt government of Kenya imported substantial amounts of fer- to meet national food requirements, particularly as political tilizer through NCPB to be distributed through its branches pressure for such measures increases. If recent indications and select private retailers at prices that included a 40 per- are reliable, the incidence of government subsidies will cent subsidy. In early 2010 news reports indicated that the probably decline after a few years as they become unsus- government planned to import 1.5 million bags (75,000 tainable unless international partners shoulder some of the tons) of fertilizer.1 Table 16.1 details key points in the liber- responsibility. Table 16.1 Evolution of Maize and Fertilizer Market Policy Reforms in Kenya, 1988–2010 State marketing agency Maize market policy Fertilizer market policy 1988: NCPB faces deficits and is financially 1988: Cereal Sector Reform Program envisages Pre-1990: KGGCU / KFA and KNTC are restructured. NCPB depots are phased widening of NCPB price margin. In fact, main input agencies. Mismanagement and out. NCPB debts are written off; crop margin narrows. Proportion of grain that deficits are common. Government control purchase fund is established but not millers are obliged to buy from NCPB of the fertilizer industry is heavy. Imports replenished. declines. Limited unlicensed maize trade is are poorly coordinated, leading to allowed. State sets all prices for maize surplus/deficits. Licensing of private trade in grain and flour. the late 1980s is controlled but under panterritorial pricing. State agencies are financially weak. Early 1990s: NCPB narrows its margins. 1991: Local and international pressure for Private sector finds it unprofitable to reforms builds up. Interdistrict trade is reach remote areas. further relaxed. 1992: Kenya moves from a one-party political 1992: Foreign exchange regime is liberalized. system to a multiparty system. Restrictions Fertilizer import restrictions are relaxed. on maize trade across districts are reimposed. 1993: Fertilizer market is liberalized. NCPB is unable to defend ceiling prices. Private traders are allowed to import and 1993: Maize and maize meal prices deregulated. distribute. State and donor imports decline Import tariffs are abolished. No subsidies are dramatically. provided to registered millers. 1994: Custom duty and value added tax (VAT) are removed. 1995: Donor pressure leads to NCPB 1995: Internal maize and maize meal trade is 1996: Number of market entrants is being restricted to the role of buyer and fully liberalized. Maize import tariff of 30 estimated at 12 major importers, 500 seller of last resort. NCPB market share percent is reimposed. 1996: Export ban is wholesalers, and roughly 5,000 retailers declines to 10–20 percent of marketed imposed after poor harvest. (Allgood and Kilungo 1996) maize trade. NCPB operations are 1997: Import tariff is imposed after poor confined mainly to high-potential areas harvest. of western Kenya. 2000 onward: NCPB is provided with funds 1997–2005: External trade and tariff rate levels to purchase a greater volume of maize. change frequently and become difficult to NCPB’s share of total maize trade rises predict. NCPB producer prices are normally to 25–35 percent of total marketed maize. set above import parity levels 2005 onward: The government withdraws the import tariff on maize entering Kenya from East African Community member countries. An official 2.75 percent duty is still assessed, and a variable import duty is still assessed on maize entering Kenya through the Mombasa port. (continued next page) 272 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Table 16.1 (continued) State marketing agency Maize market policy Fertilizer market policy 2008: World food prices are high. NCPB is 2008: Postelection violence erupts. African 2008: High world prices for fertilizer asked to sell subsidized grain to millers, Centre for Open Governance estimates that exacerbate food crisis effects from who in turn could lower prices for 3.5 million bags of maize are destroyed. postelection violence. Prices more than consumers. State has difficulty enforcing NCPB begins importing maize from United double. Petrol and transport costs also and monitoring at millers’ end because of States and South Africa in late 2008. An go up. unknown milling costs. Allegations of estimated 5 million bags arrive, according corruption emerge. to the center. 2009: Briefcase firms and NCPB employees 2009: Imports continue, but domestic maize 2009: NCPB imports state- subsidized take advantage of crisis and subsidy production is greater than expected. Claims fertilizer to aid in recovery from arrangements to favor some firms in of monopoly at Mombasa port involving grain postelection violence. Fertilizer is return for kickbacks. Weaknesses in handling (particularly related to one large grain distributed through private trade networks. disaster preparedness, institutions, and handler, Grain Bulk Handlers Limited) and food policy are revealed. Top management milling arise but are not substantiated. of NCPB and some Ministry of Agriculture officials are fired for corruption during the crisis. 2010: NCPB allocates funds to buy maize 2010: Short rainy season results in large 2010: State imports more than 30,000 tons produced in eastern Kenya during short amount of maize production, but farmers of fertilizer and distributes to vulnerable rainy season. report poor prices from private traders. farmers. Distribution is managed by NGOs. Source: Authors, adapted from Ariga and Jayne (2008). THE EFFECT OF REFORMS IN FERTILIZER within fertilizer and maize markets generated specific AND MAIZE MARKETS responses from the private sector and smallholder farmers. The basic story is one of synergies between liberalization of Data collected in household surveys and found in secondary input and maize markets and public investments in support sources can be used to analyze the effects of policy reforms of smallholder agriculture that led to substantial private sec- on fertilizer and maize markets in Kenya during the 1980s, tor investment in fertilizer retailing and maize marketing, 1990s, and 2000s. The panel data consist of nationwide rural which in turn resulted in an impressive increase in the use household panel survey data covering the 1996/97, of fertilizer by smallholder farmers and in the maize yields 1999/2000, 2003/04, and 2006/07 crop seasons. The panel of smallholder farms over 1997–2007. household survey was designed and implemented under the Tegemeo Agricultural Monitoring and Policy Analysis Proj- ect (TAMPA), implemented by Egerton University/Tegemeo Distance from farm to fertilizer seller Institute with support from Michigan State University. Out of the national sample of 1,260 households, a balanced One indicator of how reforms have contributed to fertilizer panel of 899 households was interviewed in all four periods. availability in Kenya is the distance farmers now have to Other data, such as monthly maize price levels and NCPB travel to buy fertilizer compared with the distance in the maize purchases and sales, were obtained from various prereform period (figure 16.3). This variable is a measure of Kenyan government ministries. increased private sector competition leading to more invest- For analytical convenience, the survey sample has been ment in locating retail services closer to producers, with classified into zones based on agro-ecological characteristics, retailers opening stores in new catchment areas that were districts, and agricultural production potential. Each of hitherto not serviced by the government-run input system these agro-ecological zones has been split into two broader and opening additional stores in already-served areas to categories—high-potential and low-potential regions—based capture more business. on soil quality, rainfall, yield potential, and fertilizer use.2 In general, distances in the low-potential region are longer than those in the high-potential region, which is one reason why the sample is split into two broad groups Processes leading to growth in smallholder in the analysis presented here. The private sector invest- fertilizer use and maize yields ment in fertilizer trade has expanded rapidly after the state Figure 16.2 provides a schematic description of how public allowed competition and removed trade restrictions. investments in market infrastructure and policy reform Although absolute distances are generally higher for the CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 273 Figure 16.2 Synergies Increasing Fertilizer Use and Maize Yields by Smallholder Farmers Public investments: 1. Major investment in rural feeder roads 2. Generation and release of new maize varieties by Kenya Agricultural Research Institute and by private seed firms Policy reforms—fertilizer marketing: Policy reforms—maize marketing: 1. Price controls on fertilizer abolished 1. Barriers to private maize marketing 2. Private fertilizer trade fully legalized eliminated by 1995 3. Fertilizer import quotas eliminated 2. Maize meal price controls eliminated 4. Government auctioning of free donor in 1993 fertilizer phased out; no competing 3. Closing of buying stations in most parts fertilizer subsidy program (1990–2007) of the country; stations remain active in only three or four surplus maize-producing districts Private sector responses: 1. Rapid expansion in private fertilizer wholesaling and retailing, reducing the distance farmers travel to nearest fertilizer retailer 2. Reduction in fertilizer marketing costs observed between off-loading at Mombasa port and farm gate level 3. Reduction in distance traveled by farmers to point of maize sale to private trader 4. Increase over time in maize/fertilizer price ratios Smallholder farmer responses: 1. Rise in the percent of farmers using fertilizer and hybrid maize seed 2. Increase in maize yield and maize production 3. Increase in percent of farmers selling maize Source: Authors. Figure 16.3 Average Distance from Farm to Fertilizer low-potential region, the rate at which distances have Seller declined is generally higher than in the high-potential region. The consumption of fertilizer has not followed the 15.0 same regional pattern, however, implying the presence of Distance to fertilizer seller (kilometers) 15 other constraints. 10.6 10 Increasing proportion of households using fertilizer 6.9 Data from a balanced panel of households in Kenya show 5 4.3 4.1 that the percentage of households using fertilizer on at least 3.0 2.3 2.9 one farm plot rose from 59 percent in 1997 to 72 percent in 2007 for the national sample (table 16.2). However, there 0 are differences in growth across agro-ecological zones and 97 00 04 07 97 00 04 07 the two broad regions of interest. For the high-potential 19 20 20 20 19 20 20 20 High-potential region Low-potential region region, which began with a much higher proportion of Source: Authors, estimated from Tegemeo Institute/Egerton University fertilizer users, use increased from 77 percent to 91 per- household surveys, 1997, 2000, 2004, and 2007. cent of households between 1997 and 2007.Use in the Note: The high-potential region includes the districts of Bomet, Bungoma, low-potential region more than doubled during the same Gishu, Kakamega, Kisii, Meru, Muranga, Narok, Nakuru, Nyeri, Trans period, increasing from 12 percent in 1997 to 26 percent Nzoia, Uasin, and Vihiga. The low-potential region includes the districts of Kilifi, Kisumu, Kitui, Kwale, Machakos, Makueni, Mwingi, Siaya, Taita, in 2007, but the proportion of households using fertilizer and Taveta. remained relatively small. 274 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Table 16.2 Share of Maize-Growing Households Using Fertilizer, by Region and Zone (percent) Region and zone 1997 2000 2004 2007 High-potential region Western transitional 41 65 71 81 High-potential maize zone 84 89 89 92 Western highlands 78 90 91 95 Central highlands 90 91 91 93 Subtotal 77 85 86 91 Low-potential region Coastal lowland 4 4 5 11 Eastern lowland 26 27 47 48 Western lowland 2 5 7 13 Subtotal 12 14 23 26 Grand total 59 65 68 72 Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007. Note: This sample consists of a balanced panel of 899 households interviewed in all four periods. Dynamics of fertilizer application rates Figure 16.4 Average Difference in Mean Household Fertilizer Application Rates from 1997 By examining fertilizer application rates for households Levels, by Region using fertilizer and for the whole sample of households sur- veyed, including nonusers, one can observe a trend in the 20 19.0 intensity of fertilizer use (kilogram per acre of maize) since Differences in means (kilograms per acre) the inception of fertilizer reforms in the 1990s. For compar- 15 ison at the household level, plot-level application rates using 10.7 plot area (acres) as weights are aggregated.3 First, for each 10 region, the differences between the weighted mean house- 7.8 hold application rates (kilograms per acre) for the years 5 2000, 2004, and 2007 are compared with the rates for 1997, 1.5 1.9 revealing whether application rates increased, decreased or 0.1 0 remained the same in subsequent years compared with the 00 04 07 00 04 07 20 20 20 20 20 20 base period of 1997. Figure 16.4 shows the trend in this High-potential region Low-potential region indicator using the base year 1997. Source: Authors, estimated from Tegemeo Institute/Egerton University Clearly, weighted mean household application rates household surveys, 1997, 2000, 2004, and 2007. have increased relative to those in 1997 (and also from one year to the next) for each region. It is also obvious that in absolute terms, the differences are much smaller in includes Trans Nzoia and Uasin Gishu, and the Central the low-potential region, probably as a result of lower Highland zone have some of the highest application rates in application rates and the lower proportion of fertilizer the sample. For example, the HPMZ’s rates for fertilizer users (the latter shown in table 16.2). Table 16.3 supplies users only in 1997 and 2007 are 157.3 and 181.1 kilograms information on fertilizer application rates (kilograms per per hectare, similar to rates in Asia, which benefited from acre) for two sets of households: the whole sample of the green revolution and now has one of the highest fertil- households surveyed (the sum of fertilizer users and izer application rates in the world. nonusers), and a subset consisting of households that used fertilizer (only users). Trend in maize yields for unfertilized plots and As shown in table 16.3, fertilizer application rates have different seed technologies increased from their 1997 levels for all regions and agro- zones, although intensities differ across regions and agro- As shown in figure 16.4 and table 16.3, fertilizer application zones. The high- potential maize zone (HPMZ), which has increased significantly since the reforms of the 1990s, CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 275 Table 16.3 Fertilizer Application Rates for Maize-Growing Households in Kenya, by Region and Zone (kilograms per acre) Sum of fertilizer users and nonusers Fertilizer users only Region/zone 1997 2000 2004 2007 1997 2000 2004 2007 High-potential region Western transitional 23.0 47.1 46.5 57.3 57.5 73.0 63.8 71.8 High-potential maize zone 53.1 58.5 60.9 65.4 63.7 66.5 70.4 73.3 Western highlands 26.9 40.6 49.6 48.4 36.3 45.4 54.0 51.7 Central highlands 62.2 68.4 73.4 67.2 68.8 77.9 84.2 74.1 Subsample 46.3 56.1 59.5 61.6 60.6 66.7 70.0 69.6 Low-potential region Coastal lowland 0.4 0.8 0.1 1.6 10.4 19.6 2.1 13.9 Eastern lowland 3.1 5.7 8.3 9.6 12.1 24.8 19.7 23.9 Western lowland 0.4 0.5 0.9 2.3 21.3 16.4 19.4 18.6 Subsample 1.4 2.8 4.0 5.3 12.7 23.6 19.1 22.0 Grand total 33.3 38.0 43.1 45.0 58.0 63.8 65.2 64.7 Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007. Note: When estimating rates for the whole set of households, “zeros” are used for households not using fertilizer, making these rates lower or equal to rates calculated for the group including fertilizer users only (depending on presence of nonusers). when markets were opened up to competition. But have the in the low- potential region, indicating that the effects of reforms and the rise in consumption of fertilizer, had an fertilizer use vary according to other factors. effect on maize production? Looking at the dynamics of Table 16.4, which expresses yields in metric tons per acre, maize yields for plots that received fertilizer and those that can be used to show why the maize yield estimates reported did not across the survey years and for different seed types by the Ministry of Agriculture are different from the com- provides information on this question. posite yield measure used in this paper. The Ministry of First, as shown in figure 16.5, yields from fertilized and Agriculture estimates are generally lower than the compos- unfertilized plots generally increased between 1997 and ite yield (that is, the sum of the yields for maize and other 2007. Second, irrespective of fertilization, yields for hybrid crops in the same plot converted to an index using price seed plots are higher than those for nonhybrid plots. Third, ratio of each crop to that of maize), and they do not reflect for each seed technology, yields for fertilized fields are the true returns to inputs applied to the plot or field. The higher than those for unfertilized fields. And finally, yields composite yield measure increased from 0.9 ton in 1997 to for plots that receive fertilizer and use hybrid seeds are the 1.4 tons in 2007, whereas the standard measure by the Min- highest for each of the four survey years.4 istry of Agriculture, increased from 0.6 to 1.0 ton per acre Although yields generally increased between each sur- over the same years. vey year and the next, the increase was particularly signif- icant between 1997 and 2000. This may partly be Trends in wholesale price margins explained by the favorable prices for maize following a poor maize crop in the 1998/99 season; as a result, the fer- Price margins can indicate the state of competition and tilizer-to-maize price ratio was lower at the start of the innovations that reduce marketing costs between two planting season for year 2000, an incentive to farmers to points of interest by improving market efficiency. As shown increase their fertilizer use during that season (using in figure 16.7, the margin between the wholesale world naive price expectations based on recent output prices). price of fertilizer, as measured by the cost, insurance, and Figure 16.6 shows the existence of regional differences in freight (CIF) of fertilizer arriving in the port of Mombasa the yield-fertilizer nexus for the period covered by the on the east coast of Kenya, and prices in the hinterland household surveys. town of Nakuru has declined dramatically in the years Figure 16.6 shows clear differences in the potential for between 1990 and 2008. maize production in the two regions. Even without fertiliza- The world price of fertilizer remained fairly constant from tion, yields for unfertilized plots in the high-potential 1990 to 2007, when it rose sharply (see figure 16.7). During region are generally higher than those from fertilized plots this period the wholesale cost of fertilizer in Nakuru steadily 276 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Figure 16.5 Trend in Maize Yields in Fertilized and Unfertilized Plots in Kenya for Different Seed Technologies a. Plots not using fertilizer b. Plots using fertilizer 2,000 1,500 Kilograms per acre 1,000 500 0 97 00 04 07 97 00 04 07 97 00 04 07 97 00 04 07 19 20 20 20 19 20 20 20 19 20 20 20 19 20 20 20 Hybrid seed Nonhybrid Hybrid seed Nonhybrid Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007. Figure 16.6 Trend in Maize Yields for Fertilized and Unfertilized Plots in Kenya, by Region a. High-potential region b. Low-potential region 2,000 1,500 Kilograms per acre 1,000 500 0 97 00 04 07 97 00 04 07 97 00 04 07 97 00 04 07 19 20 20 20 19 20 20 20 19 20 20 20 19 20 20 20 Unfertilized Fertilized Unfertilized Fertilized Source: Authors, estimated from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007. declined, implying that the costs of marketing fertilizer there holders suggest this reduction is a result of increased compe- had declined, leading to lower prices at Nakuru. Studies tition after reforms in the 1990s, economies of scope result- (Kimuyu 1994; Wanzala, Jayne, and Staatz 2002; Allgood and ing from mergers, and access to competitive credit from Kilungo 1996; IFDC 2001), based on interviews with stake- international sources. CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 277 Table 16.4 Trend in Composite and “Standard” Maize Yields in Kenya, by Region and Zone (tons per acre) 1997 2000 2004 2007 High-potential region Western transitional 0.7(0.5) 1.2(0.8) 1.0(1.0) 1.6(1.2) High-potential maize zone 1.3(1.1) 1.3(1.0) 1.6(1.4) 1.6(1.3) Western highlands 0.7(0.5) 1.7(0.9) 0.9(0.8) 1.4(0.9) Central highlands 1.2(0.6) 1.8(1.1) 1.7(1.0) 2.3(1.3) Subtotal 1.1(0.8) 1.4(1.0) 1.4(1.1) 1.7(1.2) Low-potential region Coastal lowland 0.5(0.2) 1.0(0.6) 0.5(0.3) 0.8(0.5) Eastern lowland 0.4(0.2) 1.0(0.5) 0.6(0.4) 0.8(0.5) Western lowland 0.5(0.3) 0.7(0.4) 0.5(0.3) 0.9(0.7) Subtotal 0.5(0.2) 0.9(0.5) 0.5(0.3) 0.9(0.6) Grand total 0.9(0.6) 1.3(0.8) 1.2(0.9) 1.4(1.0) Source: Composite yield estimated by authors from Tegemeo Institute/Egerton University household surveys, 1997, 2000, 2004, and 2007. Maize-only yield from Kenya Ministry of Agriculture. Figure 16.7 Price of Diammonium Phosphate in The analysis conducted for this paper used regression Mombasa and Nakuru methods including random effects (RE), fixed effects (FE), and correlated random effects (CRE) to model fertilizer 4,400 Cost per 50 kilogram bag (constant 2007 K sh) demand (application rate per acre). The RE approach 4,000 assumes strict exogeneity between explanatory variables 3,600 and composite error term, which includes unobserved household-specific heterogeneity. On the other hand, FE 3,200 does not assume strict exogeneity but takes the unob- 2,800 served effects as constant over time and uses a differencing 2,400 approach to remove these effects to generate consistent 2,000 estimates. Unlike FE, the CRE method extends the RE analysis by modeling unobserved heterogeneity using the 1,600 household means of time-varying variables. Therefore, 1,200 with CRE, it is possible to test whether the model captures 800 unobserved effects and to use estimates of these effects to classify households or explain differences between house- 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 holds (Wooldridge 2002). An additional benefit of CRE is Nakuru, wholesale Mombasa, cif that the estimates on the time-varying variables are the Source: Kenya Ministry of Agriculture. same as those in the FE estimation, and unlike in the FE approach, the effect of time-constant factors (such as HOUSEHOLD AND OTHER DETERMINANTS OF gender and location dummies) are estimated as well (not FERTILIZER DEMAND differenced away as in FE). Using variables including education, value of assets, land Using CRE, it is possible to reject the null hypothesis of size, land preparation technology, gender of the household nonexistence of unobserved heterogeneity, implying that head, as well as geographic factors such as distance to fertil- the FE approach is more appropriate than RE, which izer seller, agro-ecological conditions, soil types, and market assumes exogeneity. However, CRE regression offers the conditions, it is possible to conduct a regression analysis of benefit of producing the same estimates as FE regression for fertilizer demand in the high-potential and low-potential time-varying variables, while at the same time providing a regions of Kenya. This analysis will provide a measure of way to model heterogeneity so as to explain differences diversity or heterogeneity in demand across the country and across households based on skills and other factors that can- between different households that face varied surroundings, not be observed or for which data cannot be obtained. For knowledge that is important in setting appropriate policy these reasons, the results of only the CRE method are dis- geared to achieving food security for smallholders. cussed here, using a double-hurdle approach. 278 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Table 16.5, presents the results of the CRE regressions different sign and magnitude in the market participation for the high- and low-potential regions for fertilizer mar- and demand equations, unlike the Tobit model, which ket participation and consumption or use decisions. For assumes the same effect and magnitude in both equa- the double-hurdle model, the same variable can have tions. Table 16.5 Fertilizer Market Participation and Demand Using Correlated Random Effects to Model Household Heterogeneity High-potential region Low-potential region Market Consumption Market Consumption Variables participation (kilograms/acre) participation (kilograms/acre) Dependent variable (units) (0/1) (kilograms/acre) (0/1) (kilograms/acre) Price for nitrogen (K Sh/kilogram) 0.015 –0.831* –0.009 –0.175** (0.005) (0.439) (0.009) (0.081) Price for maize grain (K Sh/kilogram) –0.003 0.313 –0.003 –0.016 (0.003) (0.218) (0.004) (0.026) Age of household head (years) –0.002** –0.001 (0.001) (0.001) Quintiles for value of household assets: 2 0.012 0.853 –0.047 0.359* (0.016) (1.421) (0.029) (0.206) 3 0.001 1.108 0.016 0.571*** (0.019) (1.554) (0.031) (0.219) 4 –0.017 –0.101 –0.001 0.401* (0.022) (1.704) (0.032) (0.241) 5 –0.012 2.454 0.021 0.897*** (0.024) (1.849) (0.036) (0.261) Quintiles for total cropped land: 2 0.021 –3.767*** 0.007 0.130 (0.019) (1.394) (0.026) (0.202) 3 0.038* –4.270*** –0.006 –0.326 (0.021) (1.505) (0.027) (0.216) 4 0.068*** –3.995** –0.007 0.058 (0.022) (1.680) (0.028) (0.235) 5 0.069*** –1.594 0.006 –0.261 (0.025) (1.926) (0.032) (0.264) Categories for education of head of household: 2 1–4 years –0.018 1.475 –0.066 –0.305 (0.022) (2.696) (0.035) (0.338) 3 5–8 years –0.017 0.546 –0.036 –0.155 (0.022) (2.640) (0.034) (0.289) 4 9–12 years 0.001 5.605* 0.043 0.379 (0.024) (2.897) (0.044) (0.357) 5 > 12 years 0.032 6.416* 0.100 0.133 (0.029) (3.698) (0.067) (0.537) Categories for land preparation technology: 2 Oxen 0.101*** 4.330** –0.016 0.011 (0.021) (1.768) (0.027) (0.200) 3 Tractor 0.147*** 5.670*** –0.030 0.056 (0.020) (1.701) (0.040) (0.293) Categories for land tenure: 2 Own land without title 0.005 –0.088 0.027 0.068 (0.014) (1.069) (0.018) (0.142) 3 Renting land 0.054*** –1.047 0.030 –0.299 (0.018) (1.508) (0.036) (0.303) Dummy (1= female head of household) –0.023 –0.647 –0.060*** –0.281* (0.018) (1.669) (0.020) (0.197) Categories of soil types: 2 0.007 –1.595 –0.162** –0.479 (0.040) (4.346) (0.079) (0.872) (continued next page) CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 279 Table 16.5 (continued) High-potential region Low-potential region Market Consumption Market Consumption Variables participation (kilograms/acre) participation (kilograms/acre) 3 0.020 –1.643 (0.028) (2.903) 4 0.009 –1.708 –0.008 –0.143 (0.024) (2.577) (0.075) (0.578) 5 –0.094* –7.913* –0.083 –0.377 (0.054) (4.362) (0.103) (0.907) Agro-zone dummies (central lowland and west transitional dropped): 3 Eastern lowlands –0.036 –0.195 (0.064) (0.710) 4 Western lowlands 0.389*** 2.324** (0.110) (1.133) 6 High-potential maize 0.398*** –0.435 (0.050) (3.891) 7 West highlands 0.263*** –2.239 (0.059) (4.134) 8 Central highlands 0.423*** 19.053*** (0.058) (6.189) Dummy (1= single crop in plot) –0.061*** –0.307 –0.001 –0.042 (0.016) (1.084) (0.018) (0.161) Mundlak–Chamberlain device: Price for nitrogen (K Sh / kilogram) –0.064** –1.246 –0.062 0.525 (0.025) (2.880) (0.061) (0.576) Price for maize grain (K Sh/kilogram) 0.053*** –1.669 0.030*** 0.098 (0.011) (1.075) (0.010) (0.092) Dependency ratio (dependants to productive –0.001 –0.359 0.002 –0.005 members) (0.002) (0.287) (0.004) (0.039) Distance to fertilizer seller –0.023*** 0.641 –0.009*** –0.012 (0.005) (0.629) (0.003) (0.025) Duration as head of household (years) 0.004** 0.001 (0.002) (0.002) Quintiles for value of household assets: 2 0.015 4.074 0.096 –0.495 (0.036) (4.228) (0.065) (0.561) 3 0.088** –3.910 0.046 –0.396 (0.038) (4.261) (0.061) (0.537) 4 0.111*** 0.454 0.058 0.218 (0.040) (4.353) (0.062) (0.567) 5 0.103** –1.483 0.057 –0.064 (0.044) (4.683) (0.064) (0.593) Quintiles for total cropped land: 2 –0.046 –2.737 0.190*** –1.039 (0.037) (4.147) (0.068) (0.548) 3 –0.002 4.387 0.238*** –0.278 (0.038) (4.129) (0.065) (0.529) 4 –0.067* 0.031 0.173*** –0.489 (0.038) (4.117) (0.064) (0.562) 5 –0.007 2.099 0.178*** –0.657 (0.043) (4.483) (0.068) (0.585) Fractions of 20-day periods with < 40 millimeters of rain in season –0.185*** –26.845** 0.100 6.206 (0.124) (12.346) (0.254) (2.479) Observations (plots) 4,051 4,051 1,782 1,782 Source: Authors. Note: Quintiles range from lowest (1) to highest (5) for asset values and land sizes (estimates shown in the table are compared with lowest quintile, which is dropped). Estimates for the remainder of categories (land preparation, zones, tenure, and others) should be interpreted in relation to the omitted category (category 1). Standard errors are in parentheses; *** p < 0.01, ** p < 0.05, * p <0.1. 280 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Effect of fertilizer and maize prices on analysis do not include total land owned for the year 2000, demand for fertilizer land under crops is used as a proxy for this variable). Results of the regression analysis presented in table 16.5 Although the probability of market participation rises with show that when fertilizer prices increase by approxi- an increase in the size of land for households in the high- mately 278 shillings per 50 kilogram bag, household fer- potential region, the amount applied per acre decreases with tilizer application rates (kilograms per acre) decline by land size. Households with more land under crops in the 4.4 kilograms (high-potential region) and 1.1 kilograms low-potential region have a higher probability of market (low-potential region). For the high-potential region, this participation, although the rates per acre are not signifi- translates to a response elasticity of 0.52 (that is, a 10 percent cantly different from those with less land under crops. increase in fertilizer price leads to a 5.2 percent decline in application rates per acre). The elasticity for low-potential Effect of land tenure, gender, land region is 0.32. preparation technology, and mixed From the Mundlak-Chamberlain device (Wooldridge cropping on demand for fertilizer 2002), it is evident that households facing higher fertilizer Households that rent land have a higher probability of fer- prices have lower probability of participation, while those tilizer market participation (by 0.05 points) than those that facing higher maize output prices have higher probability of have title to their land, but differences in application rates participation in fertilizer markets. There is a reduction in across land tenure categories are not significant. Female- probability of market participation of 0.64 points for headed households have a lower probability of market par- households facing a 10 K Sh-per-kilogram higher fertilizer ticipation (0.06) and intensity per acre (0.28) compared price (or K Sh 500 per 50-kilogram bag), while the proba- with male-headed households in the low-potential region. bility of participation by households facing a similar higher Although signs of estimates are the same as those in the maize output price is higher by 0.53 points for the high- high-potential region, gender estimates for the high-poten- potential region (0.30 in low-potential region)—that is, tial region are not significant. The authors find no plausible households facing higher fertilizer prices have lower proba- explanation why the low-potential region has a gender effect bility of market participation, while those facing higher out- while the high-potential region does not. put prices have higher probability. However, the effect of In terms of agro-ecological zones, households in the cen- prices on the decision about how many kilograms of fertil- tral highlands have a 0.4-point higher probability of partic- izer to apply per acre is not significantly different across ipating in fertilizer markets than those in the western tran- households. This finding implies that price signals are sitional zone and apply 5.95 kilograms more fertilizer per important in determining farmers’ decision to participate in acre. On land preparation technology, there is a higher fertilizer markets. probability of market participation for households using animal draught (0.1) or tractors (0.14) compared with Effect of household resource endowments on manual hoeing; fertilizer application per acre for these demand for fertilizer households also increases by 4.4 and 5.6 kilograms, respec- tively, over manual technologies. Households that do inter- The analysis also takes into account the effect of some mea- crop or mix cropping (that is, plant maize with one or more sures of resource endowment (asset values and land size) on other crops in the same plot) have a higher probability of fertilizer demand. Although the probability of market par- fertilizer market involvement than those growing only ticipation for the low-potential region does not change with maize. However, application rates do not significantly differ value of assets, the application rate rises with assets. Raising with intensity of mixed cropping. a household’s assets from the lowest quintile to the third quintile or the fifth quintile, for example, raises the applica- Effect of distance to fertilizer seller, education, and tion rates by 0.571 and 0.897 kilogram per acre, respectively. experience on demand for fertilizer But when contrasting different households in the high- potential region, it is evident that households with assets of Households that are further away from fertilizer sellers have a higher value have a greater probability of participating in lower probability of participating in the market (0.023 points the fertilizer market. per kilometer for the high-potential region and 0.01 points A related measure of resources is the amount of land for low-potential region). In other words, households in the under crops during the season (because data used for this high-potential region that are located 10 kilometers away CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 281 from fertilizer retailers have a 0.23-point lower probability during the 1990s partially eroded the potential response by of participating in fertilizer markets, but application rates per the private sector. Despite the mixed government stance acre do not differ significantly. Households with head who toward maize market liberalization during the 1990s and have more years of schooling have a positive but insignificant early 2000s, evidence of increased private sector investment probability of participating in the market than those who do is tangible. Traders buying maize directly from farmers have not. The intensity of application per acre is positive with years penetrated more deeply into smallholder areas. Increased of education (and significant for high-potential region). Age competition and efficiency in maize milling and retailing is of household head in the high-potential region has a negative also evident in the significant decline in maize marketing relationship with probability of participation but not the margins. There is also strong evidence of increased state intensity of application. The number of years or duration as investment in public goods supportive of private sector the head of household has a positive effect on probability of investment, especially since the creation of the Constituency participation in the market in the high-potential region; Development Fund (CDF) in 2003. The combination of experience as head of household raises chances of market supportive policy changes in the fertilizer, foreign participation. exchange, and maize markets, coupled with improved access to markets and services made possible by public good investments, appears to have stimulated investment LESSONS LEARNED, SUSTAINABILITY, by the private sector in both maize and fertilizer marketing. AND POTENTIAL FOR REPLICABILITY These factors have worked synergistically to bring about When examining the factors driving growth in fertilizer use important gains in maize productivity and benefits to and maize productivity in Kenya from the early 1990s to smallholder farmers and consumers in Kenya. 2007, the basic story is one of synergies between the liberal- Evidence of increased smallholder fertilizer use and maize ization of input and maize markets on one hand and public yields is drawn from nationwide household panel data from investments in support of smallholder agriculture on the four surveys conducted by Egerton University’s Tegemeo other, a situation that has led to tangible private sector Institute between 1997 and 2007. Because the data constitute investment in fertilizer retailing and maize marketing, a balanced nationwide panel of 1,260 households,5 the which in turn has encouraged an impressive rise in fertilizer results provide a fairly reliable indicator of the changes in use and maize yields on smallholder farms over the period fertilizer use patterns over time, although the surveys are not 1997–2007. This narrative is complicated, however, by the strictly nationally representative. The main findings of the many changes that Kenya’s economy and business environ- surveys are as follows: ment has experienced during this period, both positive and negative, which have also undoubtedly affected the incen- ■ The percentage of sampled smallholders using fertilizer tives of farmers, consumers, and private marketing agents. on maize increased from 56 percent in 1996 to 70 percent These factors may not be directly linked to the fertilizer and in 2007. maize markets, but their influence on observed indicators ■ Fertilizer application rates (for all maize fields including cannot be analytically separated from those of the reforms unfertilized fields) rose from 34 kilograms an acre in highlighted in this paper. However, it is reasonable to 1997 to 45 kilograms an acre in 2007, a 32 percent assume that these influences outside the agricultural sector increase. are of second-order magnitude, compared with the more ■ There are wide regional variations in fertilizer use. More direct agricultural policy reforms and investments, in than 90 percent of smallholder farmers use fertilizer on explaining the behavioral responses of farmers and fertilizer maize in three of the broad zones surveyed: the high- and maize marketing agents. potential maize zone, western highlands, and central There are several pathways through which government highlands. Fertilizer use is low and barely rising in most actions in fertilizer and maize markets has positively of the semiarid regions (coastal and western lowlands, affected the agricultural sector and rural and urban living and the marginal rain shadow). However, fertilizer use standards in Kenya in recent years. As shown in figure 16.2, has risen impressively in the medium-potential eastern the government of Kenya implemented a number of policy lowlands and Western Transitional zones, where the reforms affecting the incentives for investment by private percentage of households using fertilizer on maize rose fertilizer distribution firms. The government also legalized from 21 and 39 percent, respectively, in 1997 to 43 and domestic and regional maize trade, although other actions 81 percent in 2007. 282 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS ■ While the total area under maize in Kenya remained general time period. The Tegemeo survey estimates are largely constant over 1997–2007, maize yields increased comparable and in some cases lower than other estimates of by roughly 18 percent during the same period. This yield fertilizer purchases and dose rates. The rise in smallholder improvement is not borne out in official government use of fertilizer in the Tegemeo survey data is also consis- maize production statistics, however, which do not take tent with official Ministry of Agriculture figures (shown in into account the shift over time in the proportion of figure 16.1), which indicate that total fertilizer consump- maize area grown under intercropped cultivation or the tion in Kenya rose 65 percent between 1997 and 2007. shift over time in the proportion of maize area grown in The rise in fertilizer use in Kenya has not been uniform relatively semiarid regions, which has been facilitated by across regions, however. Use rates are much higher in areas the release of improved maize cultivars well suited to where the main-season rainfall is relatively high and stable mid- and low-altitude areas of the country. To assess than they are in the drier areas. Fertilizer use is highly risky changes in maize yield, it is important to account for the in many of the semiarid regions, where environmental fac- gradual shift in the proportion of maize area under tors are likely to limit the role of fertilizer in contributing to monocropped versus intercropped cultivation as well as poverty alleviation and food security unless it is accompa- the expansion of maize production in the more semiarid nied by actions to improve soil organic matter and moisture parts of the country. After stratifying between hybrid and (Marenya and Barrett 2008). Within a given agro-ecological nonhybrid users and between intercropped and zone, it is evident that the decision of households to pur- monocropped maize fields, the household survey data chase fertilizer is only slightly related to farm size and unre- show that maize yields on all types of fields have lated to household wealth. In relatively productive areas, the increased over time, reflecting the influence of many fac- proportion of poorer and wealthier households applying tors in addition to fertilizer use. Fertilizer use and maize fertilizer on maize is similar. In risky environments, only a yields have increased especially rapidly on the inter- small proportion of either poor or wealthy households use cropped fields, less so on monocropped fields. fertilizer on maize. ■ Fertilizer marketing costs declined substantially in con- These gains in smallholder fertilizer use and maize yields stant K Sh between the mid-1990s and 2007. Interviews have been encouraged by Kenya’s decision to liberalize input with key informants in Kenya’s fertilizer sector identified and maize markets in the early 1990s. New entries and four factors responsible for the declining fertilizer mar- investment in fertilizer wholesaling and retailing have been keting costs observed in Kenya: the potential for cheaper massive since the early 1990s. The International Fertilizer backhaul transportation has been exploited by making Development Center (IFDC) estimates that more than more use of trucks transporting cargo from Rwanda and 500 wholesalers and 7,000 retailers are operating in the Democratic Republic of Congo to the port of Mombasa; country. This has led to a denser network of rural retailers private importers are increasingly using international and a major reduction in the distance between farms and connections to obtain credit at lower interest rates and fertilizer sellers, which has contributed to the impressive financing costs than are available in the domestic econ- growth in fertilizer use by Kenyan smallholders from the omy; local and international firms have merged, enabling early 1990s to 2007. The Tegemeo Institute survey data also shared knowledge and economies of scope that save local indicate that the mean distance traveled by farmers to sell distribution costs; and increased competition among their maize to private traders declined over 1997–2007; the local importers and wholesalers has expanded the num- median distance as of 2007 was zero, indicating that assem- ber of firms engaged in fertilizer marketing since the bly traders tend to purchase maize right from farmers’ early 1990s. It is likely that the fourth factor—increased fields. Analysis of wholesale maize grain prices and retail competition—has to some extent stimulated firms to maize meal prices indicate that the miller-retail marketing exploit the other cost-reducing innovations identified in margin has declined significantly over time, conferring order to maintain their market position. benefits mainly to consumers. More than 50 percent of rural farm households are either buyers or net buyers of To assess the robustness of the Tegemeo Institute’s sur- maize, while virtually all urban households purchase vey findings, the proportion of smallholder households maize meal each year (Mukumbu and Jayne 1994; Jayne purchasing fertilizer according to the survey results was and Argwings-Kodhek 1997). compared with estimates based on three other analyses Other signs of improvement in maize markets in Kenya covering a subset of the same districts during the same include farmers’ level of satisfaction with the performance CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 283 of maize markets from their subjective perspective. More relationship that encourages investment in input and out- than 65 percent of farmers surveyed in the nationwide Tege- put marketing services for smallholder farmers. In Kenya’s meo Institute rural surveys indicated that they prefer the case, this relationship was achieved through a combination current liberalized maize marketing system to the former of investments in public goods and institutional reforms controlled marketing system, primarily because grain is eas- supportive of liberalized marketing, even though the ier to sell, farmers are paid in cash at the time of sale, and maize marketing reforms were at times subject to rever- maize is more reliably available for purchase. sals. Considerable additional gains could be made in In 2008, however, the positive trends in Kenya’s maize smallholder and consumer welfare if progress could also and fertilizer markets were reversed by civil disruption, be made in the following areas: drought, and the unprecedented surge in world fertilizer prices. The civil unrest led to the destruction of much phys- Consider changes in government actions in the ical infrastructure in western Kenya, such as petrol stations transport sector that could reduce fertilizer and and grain storage, as well as to the closing of many input grain distribution costs. For example, because of fre- supply stores, in early 2008. Moreover, incentives to use fer- quent delays in off-loading of commodities at the port of tilizer in Kenya have been adversely affected both by Mombasa and because of the erosion of the regional railway drought and world events as maize/fertilizer price ratios system, it is difficult to arrange for upcountry transport of a plunged to their lowest level in at least 18 years. Figure 16.8 full shipload of fertilizer. Because of this coordination prob- plots monthly wholesale maize to wholesale fertilizer price lem, fertilizer importers have invested in storage facilities ratios per ton in Nakuru. The higher the ratio, the more near the port where fertilizer can be temporarily stored profitable the maize, and therefore the greater the incentive until trucks arrive for loading and upcountry distribution. to apply fertilizer on maize. While this ratio has historically These investments make sense if upland transport con- ranged between 0.4 and 0.6 at the time of planting, in 2008 straints and the delays and inefficiency at the Port of it plunged to below 0.25 because of the increase in world Mombasa are taken as given. However, if procedures for fertilizer prices. The price of maize in Kenya has not risen streamlining the efficiency of off-loading at the port could nearly as dramatically as fertilizer. be achieved (for example, by privatizing stevedore services These findings have implications for policy options. and issuing performance contracts or by devolving wider The main general lesson is the need for a public-private management of port operations to professional firms), thus reducing off-loading time and the storage costs incurred at Figure 16.8 Maize/Fertilizer Price Ratios, Nakuru, Kenya, Mombasa for lack of sufficient transport, then fertilizer 1994–2008 importing firms could avoid these extra charges. In a com- petitive marketing environment, these reductions in fertil- 0.8 izer marketing costs would then be passed along in the form tonne maize/tonne DAP fertilizer price ratio of lower farm gate prices.6 0.7 Reduce transaction costs associated with VAT and 0.6 port operations. Currently fertilizer, as well as most other 0.5 farm inputs, is zero-rated with respect to import duties. This means that no duty is charged on fertilizers, although 0.4 at least until 2007, a VAT on related services was still levied. A VAT is charged, for example, on transport and services 0.3 such as bagging at the port of Mombasa. Although the VAT 0.2 is supposed to be refunded, the process is lengthy and is a source of continuing frustration for market participants. In 0.1 addition, port handling charges, Kenya Bureau of Stan- dards charges, and other taxes account for 17 percent of 94 96 98 00 02 04 06 08 19 19 19 20 20 20 20 20 CIF (Gitonga 2004). Port fees, levies, and accessorial charges Source: Kenya Ministry of Agriculture, Market Information Bureau. need to be rationalized and aggregated. In addition, the Note: Price ratio is defined as the wholesale market price per ton of maize in Nakuru divided by the cost of diammonium phosphate (CIF) in Nakuru numerous documentation procedures need to be reduced per ton, in nominal Kenyan shillings. and, if possible, provided through electronic means. 284 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS Interviews with key informants in the fertilizer industry data commonly indicate that the contribution of fertilizer have identified numerous other potential sources of cost to food grain yields varies tremendously across farms savings, many of which require action on the part of gov- even within the same villages. Simply bringing fertilizer ernment to improve efficiency. response rates among the bottom half of the distribution up to the mean would result in substantial improvements Invest in rehabilitating the eroded rail, road, and in household and national food security (Nyoro, Kirimi, port infrastructure, which would reduce distribu- and Jayne 2004). tion costs. The farm gate price of fertilizer in western Kenya is roughly twice as high as the landed cost at Mom- Producer organizations, despite their poor track basa, and transport costs are the major component of this record, will be increasingly important for rural cost difference. High farm gate prices of fertilizer restrict income growth. Assuming that the management prob- demand for its use and depress agricultural productivity. lems and politicization of producer organizations and Hence, efforts to improve the efficiency of port costs and cooperatives could be minimized, they might afford an upland shipping would bring major economywide benefits. important pathway for smallholders to achieve higher levels In particular, rail transport could reduce these costs sub- of input use and to adopt better production and marketing stantially and also save government spending on repairing practices than the current separate and uncoordinated roads damaged by heavy truck traffic. stages in the supply value chains. The role of independent producer groups would be to reduce the transaction costs Tailor fertilizer packages to local demand condi- and risks of private marketing firms dealing with farmers tions. This action would increase demand from smaller and to develop a production base through the transfer of farmers who require and are able to purchase only small credit, inputs, and know-how. The Farm Inputs Promotions packets. Repackaging of fertilizers from 50 kilogram packets and the Kenya Market Development Program/Cereal Grow- into packets of 25, 10, 2, and 1 kilograms is increasingly ers Association farmer training programs are examples of taking place, but the process is sometimes associated with successful attempts by the government, development part- fertilizer adulteration and counterfeit products. (That ners, and NGOs to assist and train groups and to utilize said, adulteration and sales of counterfeit products are farm extension knowledge, supply chain development, and often isolated events rather than a well-organized activi- fertilizer technologies. ties, according to Global Development Solutions.)7 Part of While all of these measures can contribute to increased the wide fluctuation in the nitrogen and phosphorous fertilizer use, none is likely to prove effective on its own. Pol- concentration in fertilizers can be attributed to the icy makers should, therefore, select strategic combinations absence of effective measurement and calibration facilities. of supply- and demand-side measures to allow supply and In this context, the Kenya Plant Health Inspectorate Ser- demand to grow in parallel—strengthening the basis for vice and the Kenya Pesticide Board should become more viable private sector-led commercial fertilizer markets. effective in monitoring and controlling adulteration and The final question is about the role of fertilizer subsidies. counterfeit products, as well as in intensifying farmer and The greatest scope for subsidies to promote fertilizer use is retailer awareness programs to help protect farmers from in areas where fertilizer use is far below its optimal levels substandard products. after taking into account the maize yield response to fertil- izer and the riskiness of applying fertilizer, especially in Raise fertilizer response rates through agronomic semiarid regions where crop failure is not unusual. Recent training of farmers. The profitability of fertilizer use evidence indicates that crop response to fertilizer applica- could be enhanced by improving the aggregate crop yield tion varies widely among smallholder farmers even within response rates to fertilizer application. This requires mak- the same villages because of differences in management ing complementary investments in training for farmers practices, soil quality, timeliness of application, and so on agronomic practices, soil fertility, water management, forth. The evidence also shows substantial scope for raising and efficient use of fertilizer and investing in crop science the efficiency of fertilizer use, at least for farmers who are to generate more fertilizer-responsive seeds.8 Emerging currently getting lower response rates from fertilizer appli- problems of soil acidity in the maize belt of western Kenya cation than their more efficient neighbors (Marenya and indicate that soil pH levels may need to be raised to Barrett 2009; Xu et al. 2009). Moreover, there is little empir- ensure profitable use of fertilizer in these areas. Survey ical evidence to determine how prevailing levels of fertilizer CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 285 application compare with optimal levels when taking these as maize in western Kenya and horticulture throughout factors into account. Fertilizer use rates are clearly low in the the country, most farmers have achieved reasonable levels semiarid areas of Kenya, and fertilizer subsidies in these of fertilizer use without credit. Support for the develop- areas would likely raise fertilizer use, but the contribution to ment of viable credit programs may also help smallholder yields and smallholder incomes may be quite limited farmers maintain their access to fertilizer use despite cur- because of the environmental riskiness and low response rent high prices for households in which liquidity con- rates in such areas. A major question for semiarid areas, straints are the main problem. therefore, is whether poverty reduction and food security The experience of Kenya demonstrates the role of a sup- objectives can be best achieved through fertilizer subsidies portive policy environment that attracts local and foreign or other types of public programs and investments. Given direct investment in improving smallholder farmers’ access that resources are scarce, efforts should be made to identify to input and commodity markets. In Kenya’s case, a stable the types of agricultural expenditures that will generate the input marketing policy environment has fostered a private greatest payoffs. sector response that supports smallholder agricultural pro- In the high-potential areas, a large majority of farmers is ductivity and poverty alleviation. These goals remain elusive already purchasing fertilizer. Although use rates were quite in countries lacking a sustained commitment to the devel- high in 2007, they are likely to have fallen given the ensuing opment of viable commercial input delivery systems. While adverse conditions since then. Fertilizer subsidies are polit- the government’s policy stance toward maize marketing has ically attractive in that they promise increased fertilizer use been prone to vacillation, the operations of the NCPB and and food production, but these outcomes are by no means the elimination of regional trade barriers since the inception assured. In 2009 Kenya faced its lowest maize production of the East Africa Community Custom Union in January level in recent history after having initiated a major fertil- 2005 have both promoted maize price stability (Jayne, izer subsidy program; poor rains in 2009 rendered the fer- Myers, and Nyoro 2008; Chapoto and Jayne 2009). Comple- tilizer subsidy program relatively ineffective, leading the mentary programs to support small farmer productivity, country to import more than 1 million tons of maize in such as the Farm Inputs Promotion program, the agro- 2009. Moreover, providing subsidized fertilizer in areas of dealer training and credit program, and the organization high commercial demand will almost certainly result in a of farmers into groups to facilitate their access to exten- partial crowding out of commercial sales, as shown by the sion and credit services under the Kenya Market Develop- findings of studies conducted in Zambia and Malawi, ment Program have also been important factors in raising where commercial demand for fertilizer is considerably fertilizer use in Kenya. lower than in Kenya (see Xu et al 2009; Dorward et al. Because mean household incomes are higher and infra- 2008). Where purchase of commercial fertilizer is high, structure relatively better in Kenya than in many other then a ton of subsidized fertilizer distributed by govern- African countries, the market-led growth in smallholder ment is unlikely to result in an additional ton of fertilizer fertilizer use in Kenya may not be easily transferable to being applied on farmers’ fields, because the farmers previ- countries where effective demand is highly constrained. ously purchasing fertilizer are no longer likely to buy it if Kenya’s success in increasing fertilizer usage among small- they can acquire the same amount more cheaply from a holder farmers is also tenuous. Sustaining the momentum government program. will depend on continued public investment, good policy In the current high price environment, the availability choices, favorable weather conditions, and avoidance of of seasonal loans for input purchases takes on heightened international events detrimental to Kenya. Governance importance for maintaining farmers’ effective commercial problems and civil disruption are jeopardizing the sustain- demand for fertilizer. Many Kenyan farmers have been ability of the commercially driven input distribution system able to finance fertilizer through the credit offered in the and rural development more generally. Continued access to integrated input-output chains for crops such as tea, input credit for small farmers in many parts of the country sugar, and coffee. These integrated marketing arrange- will require government commitment to limit the potential ments have also provided the means for farmers to obtain for politicization and interference in the management of the fertilizer for their food crops, since the companies can interlinked crop marketing systems for sugarcane, tea, and recoup their loans for other crops as well when the farm- coffee, which have provided a means for farmers to acquire ers sell their cash crop back to the company. But in areas additional fertilizer on credit for use on food crops. Also, where fertilizer use on a particular crop is profitable, such new investment is needed in Kenya’s eroded rail, road, and 286 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS port infrastructure if Kenya is to maintain its competitive- REFERENCES ness. Last, effective systems to improve smallholders’ crop Allgood, J. H., and J. Kilungo. 1996. “An Appraisal of the husbandry and management practices are needed to provide Fertilizer Market in Kenya and Recommendations for incentives for continued expansion of fertilizer use and pro- Improving Fertilizer Use Practices by Smallholder ductivity growth in areas where fertilizer is only marginally Farmers: A Field Report.” International Fertilizer profitable at present. Development Center (IFDC), Muscle Shoals, AL. Ariga, J. M., and T. S. Jayne. 2008. “Maize Trade and Mar- keting Policy Interventions in Kenya.” Paper presented at NOTES the Conference on Food Marketing and Trade Policies in 1. “Kenya to Import 1.5 Million Bags of Cheap Fertilizer,” Eastern and Southern Africa. EST Division, Food and Daily Nation, June 23, 2009, p. 33. Throughout this chapter, Agriculture Organization (FAO), Rome, March 1. “tons” are metric tons. Ariga, J. M., T. S. Jayne, and J. Nyoro. 2006. “Factors Driving 2. The high-potential region has higher productive poten- the Growth in Fertilizer Consumption in Kenya, tial and covers the agro-zones of Western and Central High- 1990–2005: Sustaining the Momentum in Kenya and Les- lands, High Maize Potential, and Western transition areas, sons for Broader Replicability in Sub-Saharan Africa.” which include the districts of Bomet, Bungoma, Gishu, Working Paper 24, Tegemeo Institute, Egerton University, Kakamega, Kisii, Meru, Muranga, Narok, Nakuru, Nyeri, Nairobi. Trans Nzoia, Uasin, and Vihiga. The low-potential region Argwings-Kodhek, G. 1996. “The Impacts of Fertilizer consists of the lowland zones in the coast, east, and west of Market Liberalization in Kenya.” the country (which are generally drier and have poorer- Policy Analysis Matrix Project of the Kenya Marketing quality soil). The low-potential region includes the districts Development Programme. Nairobi. of Kilifi, Kisumu, Kitui, Kwale, Machakos, Makueni, ———. 2004. “Kenya Agriculture Sector Brief.” FAO, Rome. Mwingi, Siaya, Taita, and Taveta. Chapoto, A., and T. S. Jayne. 2009. Open versus Closed Maize 3. The sum of the product of plot fertilizer application Border Policy: A Comparison of Maize Price Instability in rates and ratio of plot size (acres) to total acres is calculated East and Southern Africa. International Development for all plots in the household. This procedure gives more Working Paper. Department of Agricultural, Food, and weight to application rates in bigger plots indetermining Resource Economics, Michigan State University, East aggregate household application rates. Rate = Σ (plot rate * Lansing, MI. plot area / total household area). Dorward, A., E. Chirwa, V. Kelly, T. Jayne, R. Slater, and D. 4. Note that “hybrid” stands for purchased hybrid seed and Boughton. 2008. “Evaluation of the 2006/7 Agricultural open pollinated varieties and “nonhybrid” consists of recy- Input Supply Programme, Malawi.” Final Report of the cled or replanted hybrids and some “traditional” seed types School of Oriental and African Studies, Wadonda Consult, of unknown source. Michigan State University, and Overseas Development 5. In other Tegemeo papers, the balanced panel consists of Institute, undertaken for the Ministry of Agriculture and 1,275 households, but 15 households did not have complete Food Security, Government of Malawi, Lilongwe. information on all variables used in this study, hence the Global Development Solutions. 2005. “From Laboratory to 1,260 sample size. the Dining Table: Tracing the Value Chain of Kenyan 6. Some efficiency improvements in Mombasa port opera- Maize.” Report prepared for the World Bank. Reston, VA. tions have been recently implemented, and more compre- Gitonga. K. T. 2004. “Study on Rationalization and Harmo- hensive reforms are currently under consideration. nization of Policies, Regulations, Procedures, Grades, 7. According to Global Development Solutions (2005), and Standards in the Fertilizer Sub-sector in Eastern nearly 3–5 percent of repackaged fertilizers are sold using Africa.” Kenya Report. Kampala: Eastern and Central counterfeit labels and packages. Specifically, fake brand Africa Program for Agricultural Policy Analysis. name labels are used to sell inferior quality fertilizers. Haggblade, S., G. Tembo, and C. Donovan. 2004. “House- 8. Research indicates that the highest crop yield response is hold Level Financial Incentives to Adoption of Conser- obtained when improved seed, fertilizer, and agronomic vation Agricultural Technologies in Africa.” Working practices to raise soil organic matter are combined Paper 9. Food Security Research Project, Lusaka. (Marenya and Barrett 2008; Kelly 2006). In some areas, IFDC (International Center for Soil Fertility and Agricul- improved management practices may have greater impact tural Development ). 2001. “An Assessment of Fertilizer on yields than use of fertilizer alone (Haggblade, Tembo, Prices in Kenya and Uganda: Domestic Prices vis-à-vis and Donovan 2004). International Market Prices.” IFDC, Muscle Shoals, AL. CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS 287 Jayne, T. S., and G. Argwings-Kodhek. 1997. “Consumer Nyoro, J., M., W. Kiiru, and T. S. Jayne. 1999. “Evolution of Response to Maize Market Liberalization.” Food Policy Kenya’s Maize Marketing Systems in the Post Liberaliza- 22 (5): 447–58. tion Era.” Working Paper 2, Tegemeo Institute, Egerton Jayne, T. S., R. J. Myers, and J. Nyoro. 2008. “The Effects of University, Nairobi. Government Maize Marketing Policies on Maize Prices Nyoro, J., L. Kirimi, and T. Jayne. 2004. “Competitiveness of in Kenya.” Agricultural Economics 38 (3): 313–25. Kenya and Ugandan Maize Production: Challenges for Kelly, V. 2006. “Factors Affecting Demand for Fertilizer in the Future.” Working Paper 10, Tegemeo Institute, Sub-Saharan Africa.” Agriculture and Rural Development Egerton University, Nairobi. Discussion Paper 23, World Bank, Washington, DC. Tegemeo Institute. 2007. “Enhancing Market Access and Kimuyu, P. 1994. “Evaluation of the USAID/Kenya Fertilizer Technology Adoption among Smallholders: A Baseline Pricing and Marketing Reform Program.” U.S. Agency Report for Evaluating Agricultural Interventions Funded for International Development, Nairobi. by the Rockefeller Foundation in Western Kenya.” Report Marenya, P. P., and C. B. Barrett. 2008. Soil Quality and Fer- prepared for the Rockefeller Foundation, Egerton Uni- tilizer Use Rates among Smallholder Farmers in Western versity, Nairobi. Kenya. Ithaca, N.Y: Cornell University. Wanzala, M., T. S. Jayne, and J. Staatz. 2002. “Fertilizer Ministry of Agriculture, Kenya. 2004. “Strategy for Revital- Markets and Agricultural Production Incentives: izing Agriculture: 2004–2014.” Nairobi. Insights from Kenya.” Working Paper 4, Tegemeo Insti- ———. 2008. “Economic Review of Agriculture: 2008.” tute, Egerton University, Nairobi. Central Planning and Project Monitoring Unit, Nairobi. Wooldridge, J. 2002. Econometric Analysis of Cross Section Mukumbu, M., and T. S. Jayne. 1994. “Urban Maize Meal and Panel Data. Cambridge, MA: MIT Press. Consumption Patterns: Strategies for Improving Food Xu, Z., B. Burke, T. S. Jayne, and J. Govereh. 2009. “Do Input Access for Vulnerable Urban Households in Kenya.” Subsidy Programs “Crowd In” or “Crowd Out” Commer- Paper presented at the Symposium on Agricultural cial Market Development? Modeling Fertilizer Demand Policies and Food Security in Eastern Africa, Egerton in a Two-Channel Marketing System.” Agricultural Eco- University, Tegemeo Institute, Nairobi, May 19–20. nomics 40 (1): 79–94. 288 CHAPTER 16: FERTILIZER IN KENYA: FACTORS DRIVING THE INCREASE IN USAGE BY SMALLHOLDER FARMERS CHAPTER 17 Malawi’s Agricultural Input Subsidy Program Experience over 2005–09 Andrew Dorward, Ephraim Chirwa, and T. S. Jayne he implementation in Malawi of a large-scale agri- T land that are predominantly planted to maize (see table 17.1 cultural input subsidy program in the 2005/06 for some key indicators). Continual cultivation of maize on agricultural season and in subsequent years has the same land without addition of organic or inorganic fer- attracted significant international interest. 1 While much of tilizers leads to low yields. Low yields then lead to inability this attention has applauded reported growth in maize pro- to afford the purchase of inputs. Purchase of inputs on duction and food security in the country, there have also credit is also not possible for most farmers because rural been significant criticisms and questions. These have credit markets are underdeveloped and the costs of credit focused on the effectiveness and efficiency of the program administration are too high, as are risks for both borrowers in raising maize productivity, its impacts on the develop- and lenders. Low volumes of input demand, poor infra- ment of sustainable commercial input markets (Ricker- structure, and high transport costs lead to high input costs Gilbert, Jayne, and Chirwa 2011), its high and (from and inhibit the development of input supply systems in less 2005/06 to 2008/09) dramatically rising fiscal and macro- accessible areas, while highly variable maize prices (dis- economic costs, its opportunity costs (in terms of crowding cussed below) add to the risks of input use (whether pur- out of other investments), its overall return on investment, chased with cash or credit). and its sustainability (SOAS et al. 2008; Dorward and Increasing maize productivity is difficult for several rea- Chirwa 2009a, 2009b, 2010; Dorward, Chirwa, and Slater sons. Only 10 percent of Malawian maize producers are net 2010a, 2010b; Kelly, Boughton, and Lenski 2010). sellers of maize, while 60 percent are net buyers of maize The importance of agriculture—specifically, maize—to (SOAS et al. 2008), and hence most (particularly poorer) the Malawian economy and to the livelihoods of most people’s livelihoods and food security are damaged by high Malawian people is the critical backdrop to the agricultural maize prices. Increased maize productivity from the use of input subsidy program (AISP), together with low agricul- purchased inputs requires, however, that the use of the tural and maize productivity and associated high national, inputs is profitable for farmers, and that requires sufficiently individual, and household food insecurity.2 Large numbers high maize prices and yield responses to cover the costs of of very poor people in Malawi work on very small areas of inputs. Unless substantial improvements can be made in Andrew Dorward is Professor of Development Economics at the School of Oriental and African Studies, University of London; Ephraim Chirwa is Professor of Economics at Chancellor College, University of Malawi; and T. S. Jayne is Professor of International Development at Michigan State University. 289 Table 17.1 Key Data on Smallholder Agriculture in Malawi, 2004/05 Indicator North Center South National Rural population (percent of total population) 10 38 40 88 Income and poverty Median expenditure/capita (MK thousands) 17 20.9 16.9 17.5 Poor households (percent of rural population) 56 47 64 52 Nutrition and food security Mean rural daily per capita consumption (kilocalories) for people living below the poverty line 1,738 1,811 1,703 1,746 Incidence of stunting in children 6 months–5 years (percent) 39.6 47.9 40.8 43.7 Incidence of underweight children 6 months–5 years (percent) 16.1 20 17.2 18.3 Share of calories from own production 0.53 0.58 0.47 0.52 Median month after 2004/05 harvest own food crop exhausted (actual)* — — — 4 Suffered large rise in food prices in past 5 years (percent) — — — 79.2 Smallholder agriculture Landholding: Less than 0.5 hectare/household (percent ) 12.1 15.4 25.4 19.9 Less than 1.0 hectare/household (percent ) 31.4 40.6 54.1 46.2 Suffered crop yield loss in past 5 years (percent) — — — 68.8 Maize growers (percent) 93 97 99 97 Access to credit for food crop inputs (percent ) 2.5 4.2 3.0 3.4 Percentage of smallholder farmers purchasing fertilizer (percent) 37 44 39 43 Fertilizer applied on all fields (kilograms)a 32 45 24 34 Fertilizer applied on fertilized maize fields (kilograms/hectare) 139 111 77 101 Source: SOAS et al. (2008) using data from NSO (2005) except * (authors’ calculations from NSO 2006). Note: — = not available. $1 = MK 140 for most of the years covered in this paper. a. Fertilizer rates on tobacco plots are roughly double rates across all plots. yield responses, there is a significant dilemma between the Productivity and investment in productive activities are need for low maize prices for a large numbers of poor maize further constrained by poverty and by vulnerability to a buyers (who are also significant maize producers) and the wide variety of (often related) shocks, particularly low crop need for higher maize prices to allow increased returns from yields, sickness affecting household members, high food input use to reliably cover the purchase costs. Further diffi- prices, and losses of employment or remittance income. culties arise from high maize price variability, which dam- Women, who play a key role in agricultural production and ages both producers and consumers—low prices present rural livelihoods, tend to be particularly vulnerable to these risks to investments in inputs by producers who aim to have shocks. Macroeconomic conditions before 2005—namely, a marketable surplus, while high prices present risks to con- high real interest rates, high inflation, and significant deval- sumers (including the majority of smallholder farmers). uation of the kwacha (MK)—also inhibited growth. How- Poor access to international and domestic markets (caused ever, macroeconomic management has improved dramati- in large part by historically low public investment in trans- cally since 2004. port infrastructure), seasonal scarcities, and poor local Agricultural, rural, and national economic development market development (resulting from low and uncertain vol- in Malawi are therefore constrained by a number of inter- umes, high costs of transport, and uncertain government acting household, local, and national vulnerability, poverty intervention) have led in the past to high intra- and inter- and productivity traps, as illustrated in figure 17.1. These seasonal maize price variation (as well as higher farm gate traps constrain input and maize market development, input prices and lower farm gate produce prices), further investments in maize intensification, diversification out of depressing market development. Risks of high maize prices maize into other agricultural and nonagricultural activities, encourage poor consumers to grow as much of their own the ability of rural people (particularly the poor) to protect staple food as possible, even at very low levels of productiv- themselves from shocks, and wider local and national eco- ity. At the same time, there are limited higher-earning nomic development. The result is a vicious circle of unsta- opportunities within or outside agriculture. The result is a ble maize prices inhibiting net producers’ investment in lock-in to low-productivity maize cultivation. maize production, net consumers’ reliance on the market 290 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 Figure 17.1 Vicious Circle of the Low Maize Productivity Trap Unstable Slow Poor policies private roads sector Unstable development weather Unstable maize prices Low producer investment Consumer Lock-in to low-productivity maize Low maize and agriculture productivity Low demand for nonagriculture goods and services Low and vulnerable real incomes Source: Authors. Note: Light green arrows represent feedback effect. for maize purchases, and poor consumers’ escape from low- adjustment policies, and substantial devaluation (raising productivity maize cultivation. These in turn inhibit the local fertilizer prices). With other policy changes drawing growth of the nonfarm economy. Sustained improvements more productive smallholders away from surplus maize in maize productivity with low and stable prices are production and into tobacco production, there was a wide- required to drive diversification out of low productivity spread perception in Malawi in the mid-1990s that falling maize into a more diversified and productive economy that fertilizer support was leading to diminished maize produc- benefits all Malawians, particularly those who are currently tion and a food and political crisis. From 1998/99 the gov- poor and food insecure.3 ernment, with mixed donor support, reinstated a variety of Input subsidy and maize market intervention policies interventions subsidizing maize fertilizer and seed access, have been a long-standing, major, though often contentious, with intermittent interventions in maize markets. Seed and feature of strategies of both the government of Malawi and fertilizer subsidies shifted from universal price subsidies to varied donors’ to promote agriculture and food security. free provision of small “starter packs” initially provided to From the mid-1970s to the early 1990s, the government all households in 1998/99 and 1999/2000, and then to a financed a universal fertilizer subsidy, subsidized small- more limited (but varying) number of targeted households holder credit, and controlled maize prices. The system in 2001/02 to 2004/05 (see, for example, Harrigan 2003). began to break down in the late 1980s and early 1990s, how- Analysis of smallholder agricultural performance since ever, with cash flow difficulties, rising treasury deficits, par- the late 1990s is complicated by difficulties with data and in tial market liberalization and increasing importance of separating out the effects of poor rainfall and of policy parallel grain markets. The state system of subsidized input changes responding to perceptions of an impending food loans, with loan recovery through farmers’ delivery of grain crisis. In contrast to the widespread perception that maize to the Agricultural Development and Marketing Corpora- production fell during the 1990s and 2000s, official maize tion (ADMARC), collapsed in the mid-1990s as a result of production and overall food production estimates show a the coincidence of widespread harvest failure, multiparty strong rising trend through the 1990s to 2006, together with elections, credit default, the rise of parallel grain markets, a modest rising trend during the same period in per capita partial implementation of liberalization and structural food production (including sometimes disputed estimates CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 291 of increasing cassava production). There were two years of in Malawi, causing major production and welfare prob- very poor rainfall in 1991/92 and 1992/93, two years of good lems in rural areas.5 rainfall with universal distribution of small free fertilizer Improving the profitability of fertilizer use in maize pro- packs in 1998/99 and 1999/2000, poor rainfall and lower duction requires lower fertilizer prices (as a result of greater fertilizer subsidies and production with widespread hunger efficiency in fertilizer supply and reduction in transport in 2000/01, 2001/02, and 2004/05, and good rainfall and a costs for importation or distribution of a subsidy, or both), large fertilizer subsidy in 2005/06. higher maize prices, or greater efficiency in the use of fertil- Fertilizer use also rose impressively through the 1990s, izer (raising the grain output to nitrogen ratio).6 Changes to with an annual average of 6 percent growth in fertilizer use maize prices and improved efficiency of fertilizer use will on all crops and commercial and smallholder farms between not, however, improve the affordability of fertilizer for large 1984/85 and 2004/05 (SOAS et al. 2008). Starting in the numbers of poor rural households in Malawi. Making fer- mid-1990s, private input suppliers took over an increasing tilizer more affordable requires very substantial reductions share of the fertilizer market from ADMARC and the Small- in fertilizer prices, the development of low-cost and accessi- holder Farmers Fertiliser Revolving Fund (SFFRFM), both ble financial services, or both. Development of such finan- parastatals responsible for importing and distributing fertil- cial services, however, requires that maize be profitable, that izers to smallholder farmers. By the end of the 1990s private smallholders have other sources of cash income that can be input suppliers were responsible for more than 70 percent used to repay fertilizer loans when the majority of the maize of national fertilizer imports and for a large proportion of they produce is for home consumption, and that very low- sales to smallholders (SOAS et al. 2008). During the 2002/03 cost systems be used for loan disbursement and recovery. All and 2003/04 crop seasons, 43 percent of smallholders sur- these requirements are difficult to achieve in Malawi. veyed in a nationally representative survey purchased some fertilizer (see table 17.1). Smallholders using fertilizer on REVIEW OF INPUT SUBSIDIES maize applied an average of 101 kilograms per hectare (see table 17.1). Major parastatal involvement in fertilizer To understand the implementation, impacts, strengths, and imports for subsidized fertilizer sales has, however, affected weaknesses of the Malawi Agricultural Input Subsidy private sector sales and confidence in investment in imports Program, it is helpful to examine broader historical and the- and retail systems by varying, and debated, amounts. oretical lessons on input subsidies’ implementation, perfor- Widespread fertilizer use on maize produced by small- mance, and impacts.7 holder farmers is constrained by two problems: prof- itability and affordability. Unsubsidized fertilizer was not Wider experience with agricultural input subsidies generally profitable on maize produced for sale in Malawi from the mid-1990s to the mid-2000s.4 It was, however, Large-scale (so-called universal) agricultural input subsidies more profitable on maize grown for households’ own con- were a common and prominent feature of agricultural sumption, with a higher subjective valuation stemming development policies in poor rural economies from the from farmers’ fears of the effects of a bad year on maize 1960s to the 1980s. They were subsequently criticized, purchase prices. For poorer farmers with this higher sub- however, as a major element in fiscally and economically jective valuation, however, affordability of fertilizer unsustainable policies that were inefficient, ineffective, becomes a major problem. Liquidity presents substantial and expensive in Africa (see, for example, World Bank difficulties for poor farm households, who on the one hand 1981). These policies distorted market incentives, blunted face a “hungry gap” during the cropping period (when competitiveness and farmer incentives, and undermined the farmers need to invest labor, seed, and other inputs in crop growth of private sector agricultural services. While subsi- production while food stocks from the previous season are dized input systems may have seemed attractive to farmers running low, and children are particularly susceptible to (in regard to the services that were supposed to be pro- sickness) and on the other very high borrowing costs and vided), theoretical difficulties with subsidy benefits (see an absence of low-cost financing services for inputs. Hun- below) were compounded by diversion and inefficiency, gry-gap problems at the livelihood level are exacerbated by which often limited actual benefits to farmers. rural economy market effects (which depress wage rates Evaluations of the rate of return to alternative public and asset prices and raise food prices). These problems are investments in Asia tend to rank input subsidies as fourth or widely recognized as very severe for poor rural households fifth, after investments in road infrastructure, agricultural 292 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 research and development, education, and often other types of interest in improving effectiveness and efficiency. Two of public investments (see, for example, Fan, Thorat, and further commonalities are a limited focus on replenishing Rao 2004; Economist Intelligence Unit 2008). There are also soil fertility and a strong prevalence of heavy subsidies arguments, however, that while returns to agricultural input (50 percent to 100 percent subsidy rates) on rationed inputs subsidies were often low, they did yield substantial benefits (Dorward 2009). in some countries at certain times. Such arguments stress the importance of differences between subsidies benefiting Theoretical benefits and costs of agricultural fertilizer suppliers and those benefiting (poorer) farmers, input subsidies falling returns over time where subsidies are effective, and the need for judicious (and changing) decisions on the scale Conventional arguments for subsidies in agricultural devel- of different investments, recognizing trade-offs, comple- opment have focused on the promotion of increased mentarities, differences in the timing of returns, and poten- agricultural productivity through the adoption of new tech- tial diminishing (and sometimes increasing) marginal nologies (Ellis 1992). Reduced costs of subsidized inputs returns across different investments (see Dorward et al. increase the profitability of these technologies and reduce 2004; Djurfeldt et al. 2005; Timmer 1989, for Indonesia; the risks perceived by farmers in adopting them. Together Fan, Gulati, and Thorat 2007, for India; and Dorward 2009). with credit and extension services, input subsidies were sup- Information on the performance of most of the recent posed to help farmers implement, benefit from, and then, input subsidy programs in Africa is limited despite the very with the withdrawal of the subsidy, fully fund efficient input substantial investments of public funds in these programs. purchases and use themselves. However, recent empirical evidence from Malawi and Supply and demand analysis of input subsidies shows Zambia shows that subsidies tend to be targeted dispropor- that because of deadweight losses, a subsidy can generate a tionately to better-off farmers compared with poor and positive net economic return to a country only if it female-headed households, where affordability constraints addresses some market failure (Siamwalla and Valdes 1986). are most severe (Govereh et al. 2006; SOAS et al. 2008); This may occur when: input subsidies have partially displaced commercial fertilizer demand, which has hindered policy objectives to ■ farmers’ private costs of working capital for input pur- promote sustainable development of commercial input dis- chase are greater than the social cost of capital, tribution systems (Xu et al. 2009; Ricker-Gilbert, Jayne, and ■ farmers’ lack of knowledge about the benefits of inputs Chirwa 2011); and the high costs of large-scale input subsi- means that their expectation of the production bene- dies means that there are very substantial opportunity costs fits from input use are less than the benefits they will in terms of forgone public investments, investments that, as gain, shown by the Asian experience discussed earlier, may have ■ there are learning costs with input use, meaning that ini- greater long-term impacts on poverty reduction and agri- tial farmer returns are low but will increase with experi- cultural growth. Moreover, Kenya has achieved impressive ence (Ellis 1992; Crawford, Jayne, and Kelly 2006; Morris growth in fertilizer use on food crops based on strong com- et al. 2007), and mercial demand for inputs after the liberalization of input ■ farmers’ risk assessment and aversion to investing work- marketing and foreign exchange controls, without the use of ing capital in input purchase and use are higher than subsidies (chapter 17; Ariga et al. 2008; Ariga and Jayne society’s risk assessment and aversion. 2009). The potential applicability of the Kenyan model, or parts of it, to other Sub-Saharan African countries needs to The size of the deadweight loss and the distribution of be considered, taking account of particular features of the benefits between consumers and producers also depend on Kenyan situation and experience. the elasticity of supply and demand. Demand or supply Dorward (2009), in a review of a number of input sub- inelasticity tends to be associated with smaller deadweight sidy programs across Africa, also notes apparent tendencies losses. Inelastic demand is associated with larger shares of for these programs to focus on production objectives and consumer surplus benefits, while inelastic supply is associ- producer welfare (largely ignoring potential benefits for ated with larger shares of producer benefits (Dorward consumers and for wider pro-poor economic growth); for 2009). Staple food markets in landlocked countries tend to poor integration of many programs with complementary be associated with more inelastic demand by poor con- investments; and in some programs, for an unfortunate lack sumers (where prices lie between export and import parity CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 293 prices). Demand tends to be more elastic for cash crops, number of difficulties with input subsidy programs: in con- particularly for cash crops that are exported.8 trolling costs, in achieving “exit strategies” after subsidy pro- Subsidy inefficiencies also arise when part of the cost of grams have become entrenched, in effective targeting of the subsidy goes to reducing the cost of production for pro- input subsidies to particular farmer types, in ensuring inputs duce that would be produced anyway; when subsidies bid are not overused, in controlling regressive benefits that favor up demand and prices for land, labor, or inputs, and are larger farmers who can access subsidized inputs, and in pre- passed back to suppliers of these inputs,9 and when venting market distortions where parastatal involvement rationing leads to opportunities for those controlling subsi- crowds out private sector investment in input supply systems dized inputs to demand payments for provision of subsi- and provides opportunities for corruption. dized inputs. Another major concern with input subsides is In recent years, however, some scholars and many gov- the extent of leakage and diversion away from their ernment policy makers have departed from orthodox neo- intended use as a result of diversion between products, classical thinking on input (particularly fertilizer) subsidies. diversion from intended beneficiaries to others within the Factors giving rise to this rethinking in Sub-Saharan Africa country, and cross-border leakage. include perceptions by some that liberalization policies have A final, crucial point is that the technical efficiency of failed to support sustainable intensification of staple food input use in generating additional agricultural production crop production; political demands for fertilizer subsidies; is critical in determining deadweight losses, distribution of tensions among donors facing such demands; concerns benefits between producers and consumers, and wider eco- about declining soil fertility; and interest in using input sub- nomic gains. This efficiency depends upon the quality and sidies as an instrument for social protection policies and as appropriateness of the inputs to the product on which a means of promoting input market development. they are used, the timing of the delivery of inputs to farm- On the basis of this analysis, Dorward (2009) suggests ers, the availability of complementary resources (for example, that the following design and implementation features are seed and fertilizer together), agro-ecological conditions, and important if subsidy programs are to be effective and effi- farmers’ technical skill or competence in using the inputs. cient in stimulating increased productivity and broad-based This analysis suggests that large-scale input subsidies growth: should be focused on: ■ Large unit (or percent) subsidies on rationed supplies ■ producers who are not using inputs because of market targeted to credit constrained farmers to reduce input failure, affordability problems. ■ crops and geographical areas for which increased input ■ Targeting access to subsidized inputs for specific house- use can induce a large supply shift (this may also require hold types where input use is constrained by the market complementary infrastructure and services for input failures that the program effectively addresses; where delivery, extension, and output markets), and these inputs can be used effectively and efficiently;10 and ■ stimulating products with inelastic demand and supply, where substantial political, economic, welfare, equity, particularly inelastic demand, among poor producers and and administrative challenges in effective and low-cost consumers (staple grain production tends to have these targeting can be overcome. characteristics in poor large or landlocked countries). ■ Rationing to control the costs of input subsidies with large per unit or percent subsidies and limited secondary mar- The analysis also suggests the importance of consumer kets in which recipients sell subsidized inputs to others. benefits in addition to (or rather than) producer benefits for ■ Encouragement of private sector input supply systems’ effi- achieving economic and welfare gains from subsidies; sub- ciency and investments by economies of scale and by sidy implementation that reduces deadweight losses and competition in the sale of large volumes of inputs (espe- rents from straight transfers, leakages, and high administra- cially in remote and previously poorer and less productive tive costs; and comparing distributional impacts and multi- areas and producers), with measures to limit uncertainty pliers from expenditure on input subsidies with alternative and the diversion of suppliers’ focus to capturing subsi- (tax, subsidy, or other transfer) instruments for changing dized sales without developing retail systems. income distribution and for stimulating growth. ■ Promotion of dynamic effects on pro-poor growth through The conclusions from this neoclassical supply and higher land and labor productivity in staple food pro- demand analysis influenced conventional wisdom on a duction, lower food prices, and higher producer incomes 294 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 that facilitate wider nonagricultural development, mar- elements of input subsidy programs as discussed earlier. ket thickening, and reduced coordination and transac- Figure 17.2 provides a conceptual framework that identifies tion costs and risks in poor rural economies. key variables and relationships affecting input subsidy pro- ■ Effective and efficient entitlement and distribution sys- gram impacts and guides this discussion of the Malawi tems supporting targeting, rationing, supply system Agricultural Input Subsidy Program. development, control of secondary markets and leakages, and cost control. A combination of paper vouchers (or MALAWI’S 2005/06 TO 2008/09 EXPERIENCE coupons),11 scratch cards, and electronic systems WITH AGRICULTURAL INPUT SUBSIDIES (involving bank cards, electronic “smart” cards, mobile phones, or some combination of the three) may be used Following severe food security difficulties in the early 2000s as evidence of entitlement. Different systems offer differ- and in line with election commitments, the government of ent potential benefits and political, technical, adminis- Malawi decided to implement a large national input subsidy trative, and social challenges within communities and program for the 2005/06 growing season.13 The popular households. Entitlements may be input specific or flexi- program has been repeated and expanded in subsequent ble with regard to inputs allowed and may have a fixed seasons, building on core experience but also incorporating value (with a variable top-up at redemption) or a vari- modifications in components and implementation systems able value (with a fixed top-up). There are also impor- from year to year. Core elements of the program common to tant interactions between entitlement systems, secondary the different years have been its use of vouchers targeting markets, recipient choice (of inputs and suppliers), con- roughly 50 percent of farmers across the country for the trol of fraud and program costs, and gendered access to receipt of fertilizers for maize production, with further and control of subsidized inputs within households. vouchers for improved maize seeds and for fertilizers for ■ Complementary investments, policies, and instruments tobacco. The core objective of the program, which has been critical for subsidy effectiveness and efficiency, with bal- refined over time, has been twofold: to increase resource- anced investments in the subsidy program itself, research poor smallholder farmers’ access to improved agricultural and extension support, transport and communications inputs in order to achieve food self-sufficiency, and to raise infrastructure, and efficient and stable output markets.12 these farmers’ incomes through increased food and cash ■ Matching of political interests with more technical and crop production. The main features of the program across bureaucratic needs for cost control, limited leakages, tar- the four years are summarized in table 17.2, and details of geting, rationing, and private sector development. its design, implementation and various achievements are detailed in the following sections. Given the experience with subsidies in Africa discussed earlier, implementing some of these features is a major Program design and implementation challenge. The 2005/06 program provided the foundation on which subsequent input subsidy programs in Malawi have been Issues to consider in evaluating agricultural built. We therefore describe this program in more detail input subsidies before considering changes made in subsequent programs. The success of an input subsidy program has to be judged The objective of the program was to promote access to and against its objectives, and input subsidy programs can have use of fertilizers in maize and tobacco production in order a wide range of possible objectives: wider (pro-poor) eco- to increase agricultural productivity and food security. Fer- nomic growth, benefits for poor consumers from lower out- tilizer coupons were distributed to districts and within dis- put prices, national (or household) food self-sufficiency or tricts to extension planning areas in two rounds. In the first security, increased input adoption, increased efficiency of round, allocation was broadly in proportion to cropped input use, benefits for poor producers, input supply system maize and tobacco areas. Coupons were distributed to dis- development, soil fertility replenishment, and political ben- tricts and Traditional Authorities (TAs) by the Ministry efits. Most, but not all, of these objectives can be mutually of Agriculture and Food Security (MoAFS). TAs were complementary, depending on how a program is imple- supposed to allocate coupons to Village Development Com- mented. The balance of program objectives, and their context, mittees (VDCs), which were then supposed to identify should then determine the key design and implementation recipients to receive two coupons that they could redeem, at CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 295 Figure 17.2 Conceptual Framework for Investigating Agricultural Input Subsidy Impacts Other macro- 3. Effects on input supply economic system 4. Effects on macro management private sector importers and economy large/ small distributors; fiscal balance Political parastatals foreign exchange balance economy and profits, cash flow, health, education, policy confidence, volumes, prices, infrastructure, other processes investment, innovations, agricultural spending other services Global and 1. Input subsidy implementation regional Scale, cost, prices modalities, timing, targeting, rationing Staple food market policies 2a Effects on 2. Effects on 2b Effects on non- recipients rural households recipients Previous (different hholds and hhold season(s) (different hholds and members) events and hhold members) local and national, input access and use, farm and outcomes input access and use, farm and non-farm market and non- non-farm activities and activities and productivity market relations: productivity, labour hire Weather labour hire in/out (staple food, labour, in/out, crop purchases / crop purchases/sales, cash, land, etc sales, income, food security, Disease income, food security, prices and flows) welfare, vulnerability (HIV/AIDS, welfare, vulnerability malaria, etc) Other rural economic Other social protection and agricutural/rural development measures activities Source: Adapted from SOAS et al. (2008). Note: hholds = households. Table 17.2 Summary of Malawian Agricultural Input Subsidy Programs, 2005/06 to 2008/09 Indicator 2005/06 2006/07 2007/08 2008/09 Fertilizer voucher distribution (mt equivalent) 166,156 200,128 216,000 195,369 Households receiving one or more fertilizer coupons (percent) — 54 59 65 Subsidized fertilizer (mt For maize) 108,986 152,989 192,976 182,309 For tobacco 22,402 21,699 23,578 19,969 Total subsidized Planned 137,006 150,000 170,000 170,000 fertilizer sales Actual 131,388 174,688 216,553 202,278 Redemption price (MK/50-kilogram bag) 950a 950 900 800 Approximate voucher value, (MK/bag) 1,750 2,480 3,299 7,951 Approximate subsidy (%) 64 72 79 91 Subsidized maize seed (mt) — 4,524 5,541 5,365 Hybrid seed (%) 0 61 53 84 Cotton seed (mt) 0 0 390 435 Legume seed (mt) 0 0 24 — Cotton chemical vouchers 0 0 131,848 — Total program Planned 5,100 7,500 11,500 19,480 cost (MK millions) Actual 7,200 12,729 16,346 39,847 Sources: Logistics Unit reports; Nakhumwa 2006; SOAS et al. 2008; MoAFS 2008; Dorward and Chirwa 2009a; MoAFS implemen- tation guidelines; government of Malawi budget statistics; Dorward, Chirwa, and Slater 2010b; key informants. Note: — = not available; mt = metric tons. a. seed or fertilizer coupon (NSO 2008). 296 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 a reduced cash price, for any of four types of fertilizer. There significant numbers of coupons were not used. was considerable variation within areas in the criteria deter- ADMARC/SFFRFM reported total sales of 131,803 metric mining prioritization and selection of beneficiaries, in tons of subsidized fertilizer (representing 2.62 million numbers of people receiving coupons, and in numbers of coupons).14 coupons received per recipient household. A second, sup- Malawi’s 2005/06 agricultural input subsidy program is plementary round of coupon allocation and distribution reported to have cost MK 7.2 billion, against a budget of MK was made later in the season. Under the program, 6,000 5.1 billion (SOAS et al. 2008). The reported program cost metric tons of open pollinated variety (OPV) maize seed excludes overhead costs for ADMARC and SFFRFM and was also offered for sale without coupons, at a subsidized likely allows for only partial deduction of farmer payments price of MK 150 for a 3-kilogram bag, compared with the to ADMARC and SFFRFM for coupon redemption: these market price of MK 500. Although the parastatals payments amounted to a total of MK 2.7 billion. ADMARC and SFFRFM were responsible for distributing Following the popularity of the 2005/06 program and subsidized inputs, 48 percent of subsided fertilizer was sup- the perception of its success, the government decided to plied by private sector importers. implement the program in 2006/07 with a number of mod- Holders of coupons were entitled to redeem coupons ifications (table 17.3). These included an increase in the for fertilizer at the rate of one coupon and MK 950 for one overall amount of maize fertilizers to be subsidized, a stan- 50-kilogram bag of 23:21:0 +4S or urea (“maize fertilizers”), dard redemption price of MK 950 per bag for all fertilizer and at one coupon plus MK 1,450 per bag of Compound D types, improved coupon security (with differentiation by or CAN (“tobacco fertilizers”). These prices represented, on fertilizer type), involvement of the Logistics Unit (a unit average, a two-thirds subsidy to farmers on the market cost largely funded by the U.K. Department for International of inputs. Coupons intended for different types of fertilizer Development, which had played a major role in the logistics were not marked as such, and many coupons allocated for of the nationwide starter pack and targeted input programs “tobacco fertilizer” may have been used to buy “maize fertil- from 1998/99 to 2004/05), involvement of several large izer” instead. Sales continued into January 2005, and in var- input supply companies in retail sales of subsidized fertilizer, ious areas were limited either by a shortage of fertilizer stock and use of maize seed vouchers that could be exchanged at a or a shortage of coupons. In the latter case, supplementary wider range of outlets (including agro-dealers) for different coupons were used in some areas, but unavailability of fer- quantities of OPV or hybrid seeds.15 The seed component, a tilizer in time for it to be agronomically useful meant that portion of the Logistic Unit’s costs, and an independent Table 17.3 Principal Changes in Subsidy Program Design and Implementation, 2005/06 to 2008/09 Year Subsidized inputs Voucher distribution system Voucher redemption systems Other system innovations 2005/06 Maize and tobacco fertilizers, District allocation by maize areas, Only through SFFRFM None maize seed (OPV) distribution through TAs and ADMARC 2006/07 Maize and tobacco District allocation by maize areas, Fertilizers also through major Coupons specific to fertilizer fertilizers, maize seed distribution varied, through local retailers; flexible maize seed type; fertilizer buy-back (hybrid and OPV) government, TAs, VDCs, MoAFS vouchers through wide range system; involvement of of seed retailers Logistics Unit 2007/08 Maize, tobacco, coffee, and District allocation by farm Fertilizers also through major Reduced copies of coupons; tea fertilizers; maize seed households and areas, retailers; flexible maize and remote EPA premium; (hybrid and OPV); legume distribution through legume seed vouchers through fertilizer buy-back system seed (limited); cotton seed MoAFS and VDCs wide range of seed retailers; and chemicals cotton inputs through ADDs 2008/09 Maize and tobacco fertilizers; District allocation by farm Fertilizers only through ADMARC Extra coupon security maize seed (hybrid and households and areas; use of and SFFRFM; flexible maize and features and market OPV); legume seed, cotton farm household register; open seed vouchers through wide monitoring; no remote EPA seed, and chemicals; maize meetings for allocation and range of seed retailers; cotton premium; ADMARC storage chemicals disbursement led by MoAFS inputs through ADDs computers for voucher processing Sources: Logistics Unit; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); key informants; Dorward and Chirwa (2009a); MoAFS Implementation guidelines; NSO (2008). Note: EPA = Extension Planning Area. CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 297 program evaluation were funded by donors, who had not prices, program costs were 29 percent above the budget, directly financed any part of the 2005/06 program (other compared with 18 percent the previous year. However, pri- than through budget support). Donors also funded a buy- vate sector subsidy sales were roughly the same as the previ- back scheme, which reduced the risks to government of ous year (increasing by only 6 percent, from 49,000 metric holding unsold stocks at the end of the year if private sales tons to 52,000 metric tons), whereas parastatal sales led to lower-than-expected sales by ADMARC and increased by approximately 30 percent, from 125,000 to SFFRFM. 165,000 metric tons. Planned and achieved subsidy sales and costs in 2006/07 A number of further changes were made to the 2008/09 (and other years) are shown in table 17.2. Supplementary subsidy program. A farm household register compiled the fertilizer voucher issues and the availability of fertilizer previous year was updated and used to list coupon alloca- for sales by private companies (which sold just under tions to individual beneficiaries in open village meetings led 30 percent of subsidized sales) together led to higher sales by teams involving MoAFS and local government staff. An volumes than budgeted. These, together with higher attempt to print coupons in the government printer was fol- prices than budgeted, led to significant budget overruns. lowed by a significant security breach, and vouchers for the These problems were not faced in seed sales, where no central and northern regions were then printed outside the extra coupons were issued. country with extra security features. The flexible maize and Growing experience with the program led to consolida- legume seed voucher and cotton input systems were contin- tion in 2007/08 of many of the changes made in 2006/07, ued. Grain pesticides were also subsidized, and some subsi- together with further changes to extend the scope of the dized fertilizers were issued to tea and coffee farmers. program. Program objectives and beneficiary targeting cri- Although private retailers were initially involved in the sale teria were amended to give greater emphasis to concerns for of subsidized fertilizers, this practice was discontinued at a vulnerable households. Targeted quantities of subsidized very late stage in the season. The program went massively maize fertilizer and seed were again increased, to roughly over budget, however, largely as a result of soaring interna- equal disbursements to the previous year. Changes were tional fertilizer costs, but there was also an approximate made to coupon allocation systems between districts to pro- 15 percent overrun in quantities of inputs subsidized. vide greater weight to the number of farming households (and less weight to crop areas), leading to an increasing pro- Implementation achievements portion of coupons allocated to the more densely populated southern region, where levels of poverty and poverty inci- Three aspects of Malawi’s achievements in implementing dence are greatest. its agricultural input subsidies are considered: program Following problems in some areas in 2006/07, systems scale, innovation and adaptation, and implementation for allocating and distributing coupons within districts were performance. also modified in 2007/08 to give less power to TAs and more responsibility to MoAFS staff. In addition to the maize seed Program scale. Malawi’s agricultural input subsidy pro- vouchers provided with maize fertilizer coupons, extra gram has grown each year since its creation and involves “flexible vouchers” allowed farmers to choose maize or complex and very significant logistical and organizational legume seeds (although legume seed supplies were very lim- challenges to tight deadlines. The major tasks of the pro- ited). A “remote area premium” was also introduced to pro- gram are shown in figure 17.3. This summary is highly vide incentives to private retailers to extend their networks simplified, however, and in practice a complex set of inter- into areas with low coverage by private retailers; a premium actions between various stakeholders is needed to complete was provided on the subsidy paid to private sector retailers each task. In 2008/09 this process involved the selection of for sales of subsidized fertilizers against identifiable vouch- more than 1.5 million fertilizer coupon beneficiaries from ers issued to beneficiaries in designated “remote areas” with more than 2.5 million farm households, printing and distri- higher transport and distribution costs (with the vouchers bution of 5.9 million coupons, and purchase and distribu- identifiable by their location code). Coupons for cotton seed tion of more than 3.4 million bags of fertilizer—all within and chemicals were distributed through the MoAFS Divi- tight deadlines. Further challenges arise because farmers sional Offices (ADDs). served by the program are widely dispersed across the coun- Subsidized fertilizer volumes were again significantly try (with a significant number being illiterate or semiliterate over budget in 2007/08; with higher-than-budgeted input and living in remote and poorly accessible areas), and 298 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 Figure 17.3 Major Tasks in Implementing Malawi’s Agricultural Input Subsidy Program Planning and Farmer budgeting registration Input Secure coupon Coupon purchase printing allocations Coordination and control Coupon Beneficiary Stakeholders distribution identification Input distribution Farmers (transport and MoAFS: HQ, LU, ADDs, DADOs, Ass, FAs storage) DCs, TAs, VDCs, Police, CSOs Coupon Fertiliser importers, retailers issue Seed producers/importers, retailers ADMARC: HQ, districts, markets SFFRFM: HQ, depots, markets Transporters Donors Coupon Payments redemption and control Source: Authors. because there is a constant threat of fraud or theft of program Following the experience of 2005/06, the government commodities worth a total of approximately $220 million, has, with other stakeholders, implemented further innova- with each fertilizer coupon worth more than 10 percent of tions to improve performance of the program and broaden annual household income for the more than 40 percent of the its impact. These changes emerged from formal and infor- population below the poverty line. mal management and evaluation reviews and lesson learn- The implementation challenges should not be underesti- ing within the government (formal internal evaluations mated, and it is significant that Malawi has been imple- were conducted in 2006/07 and 2007/08); from discussions menting such activities at varying scales annually since the with other stakeholders (donors, private sector fertilizer inception of the starter pack program in 1998 (see below) importers, seed and fertilizer suppliers, a parliamentary and has therefore built up both systems and considerable committee on agriculture, and civil society); from external expertise in these tasks. evaluations (commissioned by CISANet for 2005/06 and by the government, DFID, and USAID for 2006/07 and by the Innovation and adaptation. The 2005/06 subsidy pro- government and DFID for 2008/09); and shifts in policy gram built on Malawi’s innovative experience in imple- within a changing economic and political environment. menting the starter pack and targeted input programs The major program modifications in 2006/07, 2007/08, (TIP). These programs involved large-scale registration and and 2008/09, summarized in table 17.3, concerned: targeting across the country; development of systems using vouchers; coordination across different government, paras- ■ the extent and modalities of private sector involvement in tatal, private sector, donor, and community stakeholders; fertilizer imports, fertilizer sales, and seed sales, with a and substantial logistical challenges. The 2005/06 subsidy buy-back scheme to reduce government stockholding program, however, involved a change in objectives (from an risks, a premium to stimulate private retail network emphasis on social protection and food security for vulner- development in more remote areas in 2007/08, and exclu- able households in the TIP to national food production and sion of the private sector from fertilizer sales in 2008/09; self-sufficiency), an increase in the scale of subsidized ■ recognition of the importance of including vulnerable inputs (from approximately 50,000 metric tons of fertilizer households among targeted beneficiaries, with an in 2004/05 to 130,000 in 2005/06), and the addition of increasing volume of inputs for maize production and tobacco inputs and cash redemption of vouchers. modified district/EPA allocation systems; CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 299 ■ trialing of flexible vouchers for seed inputs and addition which may be affected by the availability of subsidy inputs of cotton inputs and grain storage pesticides; in accessible markets and by any “tip” needed to redeem ■ introduction of beneficiary registration and more open coupons. Household surveys provide the only systematic and more tightly managed beneficiary selection, voucher information available on these areas. Results from focus distribution, and market monitoring systems; group discussions and household surveys examining the ■ coupon design, printing, security, and farmer redemp- 2006/07 and 2008/09 programs (SOAS et al. 2008; Dorward, tion prices; and Chirwa, and Slater 2010a) suggest the following: ■ sharing of the costs of some program components with donors. ■ In 2008/09, 65 percent of farm households received one or more fertilizer coupons, with an average of 1.5 Implementation performance. Effectiveness and effi- coupons per household receiving coupons and of 1.1 ciency of implementation can be assessed in terms of volume coupons per household across all households (figures for of subsidized inputs disbursed, timing of subsidy sales and 2006/07 were 54 percent, 1.7, and 1.0, respectively); supplier payments, targeted beneficiary access to inputs, and ■ Targeting criteria were highly variable across different cost. As shown in table 17.2, both planned and disbursed vol- administrative areas;16 umes of subsidized inputs increased steadily between ■ Overall targeting recommendations were followed to 2005/06 and 2008/09. Although fertilizer disbursement and some extent in that there was a tendency for targeting sales targets were not met in 2005/06, they were exceeded in to reach households that are productive full-time 2006/07, 2007/08, and 2008/09 by 16 percent, 27 percent, and farmers. However, household survey data for the of 19 percent, respectively. Exceeding these targets demonstrates 2006/07 and 2008/09 seasons indicate that coupons considerable success in meeting demand, but also suggests were disproportionately targeted to households with difficulties in controlling disbursement and cost overruns. relatively large amounts of land or other assets, and (in Timing of subsidy sales is determined by the timing of 2006/07 but not in 2008/09) to male-headed house- availability of inputs in markets and the issue of vouchers to holds. Smaller proportions of fertilizer coupons were beneficiaries. Timing is critical for the effective use of seed given to households in the bottom half of the wealth and fertilizer at the start of the agricultural season. For fertil- and income distribution; izer, the timing of input availability depends upon timing of ■ In some areas, particularly the south and center, coupon tendering of input purchases and supplier deliveries to allocations were modified so that in 2008/09 just under depots, the staffing and stocking of input markets (for paras- 40 percent of households in these regions and 36 percent tatal sales), and subsidy redemption contracts with retailers nationally received one fertilizer coupon (rather than and their stocking and staffing of input sales points for private fewer households receiving two coupons); sector sales. The timing of voucher issue depends on the tim- ■ Open meetings for coupon allocation were introduced in ing of beneficiary registration, voucher allocations, voucher 2008/09 and appear to have succeeded to some extent in printing, voucher distribution to districts, and district distri- increasing the proportion of coupons and subsidized fer- bution payments. Information on some of these variables is tilizer going to poorer households (Chirwa, Matita, and given in table 17.4. In general, performance regarding earlier Dorward 2010); award of seed and fertilizer contracts and earlier fertilizer ■ Key informants tended to underestimate the propor- deliveries to depots and uplifts has improved over time. Infor- tion of households receiving subsidized inputs com- mation on the timing of fertilizer sales is incomplete, but it pared with estimates provided in interviews with appears that despite some evidence of improvement between households; 2005/06 and 2006/07 (not shown in table 17.4), there has ■ In 2006/07, 75 percent of ADMARC and private supplier been little improvement since then. It is particularly impor- outlets and 100 percent of SFFRFM outlets were reported tant to increase sales by the end of November. These sales to have suffered from frequent major queues; a similar were highest in 2008/09 but still only 30 percent of the total. figure of 75 percent was reported across ADMARC and Receipt of seed vouchers by the Logistics Unit is determined SFFRFM outlets in 2008/09; by the timing of sales and the speed of voucher processing by ■ In both 2006/07 and 2008/09, household surveys indi- seed suppliers; both were problematic in 2007/08. cated that 5 percent of coupons were reported to be Targeted beneficiary access to inputs is determined by accessed with some payment, with a median price of MK coupon allocation and issue and by the use of coupons, 1,000 in 2006/07 and of MK 2,000 in 2008/09. 300 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 Table 17.4 Implementation Performance Indicators 2006/07–2008/09 Indicator 2006/07 2007/08 2008/09 Fertilizers Tender awards for parastatal supplies Late August Mid-August End-July Depot receipts end October (% of parastatal total sales) 32 58 53 Depot receipts, end November (% of parastatal total sales) 77 76 71 Outstanding payments, end Nov (% and MK millions) 28 1,216 22 1,595 16 3,500 Outstanding payments, end Dec (% and MK millions) 46 4,303 13 1,192 13 3,690 Outstanding payments, end Jan (% and MK millions) 14 1,406 21 2,620 — 7,707 Uplifts dispatched as of end November (% of parastatal total sales) 64 70 75 Total relocation transport costs (MK millions) — 68.4 42.0 Finalization of retail fertilizer contracts Early Nov Mid/late Nov — District voucher allocations Early Sept Oct 9 Sept 12 Voucher printing End Sept End Oct SR early Oct CR/NR early Nov Voucher and list distribution to districts completed Nov 7 Nov 3 Nov18 Sales end Nov (% of total season sales) 8 — 30 Sales end Dec (% of total season sales 74 — 68 SFFRFM/ADMARC voucher returns end Dec (thousands) 0 101 175 SFFRFM/ADMARC voucher returns end Jan (thousands) 111 720 1,057 Finalization of seed supply contracts Mid/late Nov Mid/late Nov Early Nov Seed coupons in LU end Dec (% of season sales) 27 4 6 Seed coupons in LU end Jan (% of season sales) 74 18 22 Sources: Logistics Unit reports; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); key informants. Note: Data are not available for 2005/06. LU = Logistics Unit. — = not available. ■ In 2006/07 a “tip” was paid to retail market staff for Determining the extent of the fraud is rendered difficult redemption of about 20 percent of fertilizer coupons, by the lack of formal and transparent audit systems cover- with a mean price per bag of just over MK 1,000 (com- ing the whole program and by discrepancies between pared with the official price of MK 950) and with no MoAFS and NSO estimates of the total number of farm significant overall differences between parastatal and households in Malawi (MoAFS estimates of farm house- private sector suppliers. In 2008/09, 14 percent of fertil- holds were 33 percent above NSO estimates in 2006/07 and izer coupons were reported to require a “tip” for redemp- 47 percent above NSO estimates in 2008/09). SOAS et al. tion, with a median “tip” of 200 MK, again giving a price (2008) concluded that there is insufficient evidence to sug- of MK 1,000 per bag. gest widespread fraud, and that household survey estimates of subsidized fertilizer access were broadly compatible with There are considerable difficulties in determining the the MoAFS farm household estimates, but there are numer- extent to which fraud affects Malawi’s agricultural input ous anecdotal reports of fraud within the system. Although subsidy program. Fraud can arise in a number of ways— Dorward, Chirwa, and Slater (2010b) suggest that NSO may through allocation of vouchers to nonexistent beneficiaries well underestimate the number of farm households in (and their diversion to government staff, traditional leaders, Malawi, it seems unlikely that underestimate could be off by or politicians), direct allocation of vouchers to people who a third, and there are risks (and anecdotal reports) of do not satisfy beneficiary criteria, printing of extra or coun- increasing numbers of villages and of some “ghost villages,” terfeit vouchers, and payment of “tips.” suggesting significant diversion of subsidized inputs away CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 301 from the intended beneficiaries. The discrepancy between over budget in 2008/09. Program costs rose from just over NSO and MoAFS estimates of the number of farm families 60 percent of the MoAFS budget in 2006/07 and 2007/08, is being jointly examined by the NSO and MoAFS to resolve rising to 74 percent in 2008/09, when the program this issue. accounted for over 15 percent of the total national budget. Regarding the extent to which counterfeit or nonstan- It is, however, important to note that for the 2009/10 pro- dard vouchers (those with serial numbers outside the ranges gram, actual costs were 21 percent below budget and costs recorded by the Logistics Unit) have been accepted by dif- fell back sharply, by 41 percent, because of both a halving of ferent outlets, records for 2007/08 show that these (and sales fertilizer prices and a 20 percent reduction in the amount of without vouchers) accounted for 27 percent of ADMARC/ fertilizer disbursed, which was almost exactly on budget. As SFFRFM sales and 3 percent of private retailer sales (LU a result program costs fell back to 7 percent of the total 2009). Rapid return of vouchers to the Logistics Unit is national budget. important for early identification of markets accepting Data on estimated per unit fertilizer costs and on total counterfeit or nonstandard vouchers. Private retailers gen- program costs, excluding ADMARC overhead costs, are also erally return coupons quickly in order to receive payment, given in table 17.5. As shown in the table, fertilizer prices but ADMARC and SFFRFM have been much slower at this and transport costs rose from 2005/06 to 2008/09. The esti- task although their voucher returns have improved over the mated per unit fertilizer cost increases from 2005/06 to three years for which records are available (see table 17.4). A 2006/07 in Malawi (25 percent) are higher than would be major security breach in the printing of vouchers in 2008/09 expected given that international prices were static over led to reprinting of more secure vouchers for issue in two the same period, but from 2006/07 to 2007/08 the price regions. increase (22 percent) was markedly lower than the increase Overall, the costs of Malawi’s subsidy program were over in international prices, which rose by around 50 percent or budget and increasing from 2005/6 to 2008/9 (table 17.5), more, so the overall cost increase in Malawi over 2005/06 to due to a combination of increasing subsidy volumes and 2007/08 was in line with international price increases. large increases in fertilizer prices. Program costs were just Fertilizer cost increases from 2007/08 to 2008/09 also appear over 40 percent above the budget for 2005/06 and 2006/07, to have been roughly in line with increases in international nearly 50 percent over budget in 2007/08, and 90 percent price increases over the same period (about 125 percent). Table 17.5 Fertilizer and Program Costs in Malawi 2005/06–2009/10 2005/06 2006/07 2007/08 2008/09 2009/10 Planned Actual Planned Actual Planned Actual Planned Actual Actual Fertilizer cost ($/mt): Parastatal: delivered at depots — — — 454 — 555a — 1,204b 575 Parastatal: transport, etc. — — — 36 — 45 — 46 39 Parastatal: total — 393 — 490 — 600 — 1,250 614 Private retailers: total — ... — 490 — 612 — ... ... Average all suppliers — 393 — 490 — 590 — 1,250 614 Program costs: Malawi government 36.4 51.4 51.4 81.4 73.6 109.6 127 227.7 137.6 Donors ($ millions) 0 0 12.5 9.5 5.7 7.1 12.1 37.8 17.5 Total ($ millions) 36.4 51.4 63.9 90.9 79.3 116.8 139.1 265.4 155.1 Net of farmer payments — 32 — 73.9 — 95.4 — 242.3 143.7 Total as % MoAFS budget — — 43 61 51 61 61 74 — Total as % national budget 4.3 5.6 5.4 8.4 6.7 8.9 8.5 16.2 6.7 Total as % of GDP — 2.1 3.1 — 3.4 — 6.6 — Sources: Logistics Unit; Nakhumwa (2006); SOAS et al. (2008); MoAFS (2008); Dorward and Chirwa (2009a); Government of Malawi budget statistics; key informants. Note: The 2005/06 fertilizer costs may also include some seed and coupon production/distribution costs. Parastatal transport etc., costs exclude ADMARC overheads. Program costs exclude buy-back carried forward. — = not available. ... = not applicable. a. Excluding costs of buy-back brought forward. b. Including costs of buy-back brought forward. 302 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 Marked monthly variation in international fuel prices from Fertilizer importers have been responsible for generally mid- to late 2006 and 2007 makes it difficult to calculate increasing proportions and volumes of government subsidy equivalent figures for transport costs. sales, with particularly large volumes in 2008/09 (table 17.6). Aside from the problems of high fertilizer prices in Importers have clearly benefited from their growing share in 2008/09, the 2006/07 to 2008/09 programs faced major imports, although they have faced some difficulties from challenges in controlling the volume of subsidized fertil- exposure to foreign exchange losses caused by delays in pay- izer disbursed. Three alternative (and complementary) ments in local currency. Insofar as currency risks are fac- approaches to limiting the volume may be used: control- tored into tender margins, these raise fertilizer costs for ling the number of coupons issued, controlling the physi- government. Some importers have expressed concerns cal stock of fertilizer available, and controlling sales of about increasing competition in import tenders from new fertilizer by closing further sales once the total budget players without proper qualification criteria, leading to the quantity has been sold. In principle, the first option is the award of some tenders to suppliers who are unable to best approach, although it is undermined by counterfeit deliver. If the award of such tenders leads to late cancellation coupons and by any high-level political pressure that may and reordering at short notice, they may also raise fertilizer demand extra coupon issues. Control of physical stock of costs for the government. fertilizer (the method used in 2009/10) is difficult if the Maize seed suppliers have also benefited from significant private sector is involved in retail subsidy, and may result growth in sales over the life of the program (see table 17.2). in genuine beneficiaries being denied the opportunity to Government and the seed suppliers association negotiate redeem genuine coupons. Closure of the program once prices, which involves a difficult balance between competi- target sales have been achieved suffers from the latter dis- tion and coordination in supply. The subsidy program advantage, and in addition requires timely reporting and affects retail sales through three processes, each of which has monitoring of sales. a different effect on whether retail outlets sell subsidized The rising costs of the subsidy program over the period inputs. These are set out in table 17.7. were met by increasing budgetary allocations to the MoAFS, Displacement of commercial sales occurs when a farmer and did not crowd out other MoAFS activities in terms of chooses not to buy an input received on subsidy when he actually cutting budgetary allocations to them. However, the or she would have bought it commercially if the subsidy opportunity cost of the program is an issue in terms of for- had not been available. This affects private retail outlets gone investments that could have been achieved with those irrespective of their participation as subsidy retailers. Dis- funds. The program also consumes very large amounts of placement is difficult to estimate because even without sub- staff time and other resources, for people must be diverted sidies, farmers’ commercial purchases change from year to from other activities to manage and implement the subsidy year with changes in input prices, output prices, and their program in the critical time before and at the start of the access to seasonal finance. Input suppliers appear to be very cropping season. Similarly, while financial resources allo- concerned about losses of fertilizer sales through displace- cated to the subsidy program have grown dramatically since ment if these are not counteracted by gains in subsidized the start of the program, financial resources allocated to sales and customers from participation in the subsidy other activities have remained largely static or shown only scheme. Displacement rates have been estimated from small increases, posing severe challenges to other essential examination of changes in aggregate sales (SOAS et al. 2008) research and extension activities of the Ministry. Input supply impacts Table 17.6 Private Sector Involvement in Subsidized Fertilizer Sales Malawi’s subsidy program has had major, and mixed, impacts on private sector input suppliers. These effects have Private sector involvement 2005/06 2006/07 2007/08 2008/09 to be considered separately for fertilizer importers, fertilizer Subsidized tender retailers, seed suppliers, and seed retailers. For retailers of deliveries (mt) 70,000 99,386 97,845 162,840 seed and fertilizer, it is important to distinguish between Subsidized tender small independent agro-dealers on one hand and retail out- deliveries (%) 48 72 71 88 Retail sales (%) 0 28 24 0 lets of larger companies involved in importation and both wholesale and retail sales on the other. Source: Logistics Unit; SOAS et al. 2008; Dorward and Chirwa 2009a. CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 303 Table 17.7 Impacts of Subsidy Program on Seed and Fertilizer Private Retail Outlets Processes by which the subsidy program affects retail sales Participating retail outlets Excluded retail outlets Displacement of commercial sales Loss of sales Loss of sales by subsidy sales Sales of subsidized inputs Gain in sales No effect Gain/loss of customers going to outlets Gain in sales Loss of sales to redeem their subsidy vouchers General increases in demand as a result Gain in sales Gain in sales of program-induced growth and income/ cash gains in previous season Private fertilizer retailers Retail chains in 2006/07 and 2007/08 Agro-dealers in all years, retail chains in 2005/06 and 2008/09 Private seeds retailers Agro-dealers and retail chains from Agro-dealers and retail chains in 2005/06 2006/07 to 2008/09 only Source: Authors. and from panel data analysis of farmer purchases (Ricker- 2007/08 for the larger importers with retail outlets with Gilbert, Jayne, and Chirwa 2011). their inclusion in retail subsidy sales, but they again Displacement estimates from examination of changes in reported declines in retail outlet fertilizer sales when they aggregate sales were 20–30 percent in 2005/06 and 30–40 were excluded from the program in 2008/09 (Kelly, percent in 2006/07, with displacement for tobacco fertilizers Boughton, and Lenski 2010). Small agro-dealers had been higher than that for maize fertilizers (SOAS et al. 2008). Dis- excluded from retail sales of fertilizer subsidies during all placement estimated from panel data analysis of farmer four seasons of the program. purchases was 23 percent for 2006/07 (Ricker-Gilbert, Jayne, and Chirwa 2011) and 3 percent for 2008/09 (Ricker-Gilbert Maize market impacts and Jayne 2010), but in all years some further displacement may be expected if some subsidized fertilizers are not The input subsidy program may affect maize markets in a received by smallholders. Estimating displacement from number of ways. We identify four potential impacts: aggregate fertilizer sales for 2007/08 and 2008/09 has not been possible due to lack of data on aggregate commercial ■ Direct impact through increased supply of maize for sale sales. Table 17.8 shows incremental fertilizer use estimates and reduced demand for purchases by net surplus and for 2005/06 and 2006/07 and predictions for 2007/08 and deficit farmers; 2008/09, assuming similar implementation in these years.17 ■ Indirect impacts as a result of policy changes influenced Displacement of maize seed sales appears to be much lower, by the subsidy; with strong growth in commercial seed sales in 2006/07. ■ In the longer term, if the subsidy program leads to rising As table 17.7 shows, however, overall impacts of the sub- incomes, demand for maize should increase as a result of sidy program on input sales depend not only on displace- consumption by both humans and livestock; ment effects but also on the impact of participation or ■ Finally, if the net effect of these impacts is to lower (or exclusion in the program on subsidy sales and on customers raise) maize prices, then a supply response to increase (or visiting the outlet. The last two lines of the table identify the reduce) resources allocated to maize production should status of agro-dealers and retail outlets for larger companies be expected. with regard to subsidized fertilizer and seed sales. Notably, both reported a significant increase in sales in 2006/07 when These impacts arise in the context of wider changes in they were able to participate in subsidized seed sales and in production (as a result of seasonal weather), in policies, in subsidized fertilizer sales (only retail outlets for larger com- regional and national maize markets, and in urban and panies) (Kelly, Boughton, and Lenski 2010). Conversely, the rural incomes (as a result of other processes of livelihood exclusion of the private sector from all retail subsidy sales in change and growth). Although data and analytical limita- 2005/06 led to a substantial drop in reported fertilizer sales tions make it difficult to tease out these different influences, from all retail outlets. Sales recovered in 2006/07 and Malawi does have good information on maize prices. As 304 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 Table 17.8 Estimated Incremental Fertilizer Sales and Maize Production, 2005/06–2008/09 Indicator 2005/06 2006/07 2007/08 2008/09 Incremental fertilizer sales as % of subsidy sales 70–80 60–70 60–70 90 Incremental fertilizer use (mt) 98,541 113,547 140,760 181,800 Incremental seed use (mt) OPV 3,000 1,764 2,604 833 Hybrid 0 2,760 2,937 4,532 Yield response as % of 2008/09 estimate 80 100 70 100 Subsidy program incremental maize Medium estimate 406,348 647,474 566,235 968,900 production estimates (mt) Above 2002/03 and 2003/04 273,609 514,735 433,496 836,161 High estimate: +20 percent 487,618 776,969 679,482 1,162,800 Above 2002/03 and 2003/04 328,332 617,683 520,196 1,003,514 Low estimate: –20 percent 325,078 517,979 452,988 775,200 Above 2002/03 and 2003/04 218,887 411,788 346,797 669,009 National crop production estimates Increment above 2002/03 and 2003/04 (mt) 975,262 1,698,956 1,031,938 2,031,816 Net maize exports in following year (exports minus imports, mt) –78,491 224,972 –101,027 –50,398 Source: Incremental fertilizer and seed sales figures are from Dorward and Chirwa (2009a). Figures for yield response and incremental seed impact for 2008/09 are from Dorward and Chirwa (2010). National crop estimates are from MoAFS. Net maize exports are from Jayne et al. (2010). Note: Production seasons 2002/03 and 2003/04 are considered nondrought years for presubsidy comparisons, although targeted input subsidies of 35,000 and 22,000 metric tons of fertilizer were provided in these years, and a 10 percent displacement is assumed for these years. OPV sales for 2005/06 are estimated to be 50 percent of budgeted sales. noted earlier, prices have varied widely in the past, and, as Exports and purchases of stocks to carry over to the fol- shown in figure 17.4, variability has continued in the period lowing season would reduce maize volumes (this would of subsidy implementation—indeed, maize prices reached not be the case for 2009/10 SGR and other stock pur- historic highs in early 2009. chases subsequently sold later in the 2009/10 season). High maize prices would not be expected given the large The effect of this on figure 17.5 would be to shift the maize production estimates in each subsidy year—rather, 2006/07 and 2008/09 subsidy figures to the left by around low prices would be expected.18 Low prices were observed in 300,000 and 100,000 metric tons, respectively.20 This the 2006/07 marketing season (following the 2005/06 sub- does not, however, bring the three later subsidy years sidy) and initially in 2007/08 (following the 2006/07 sub- anywhere near the pattern of the 1993/94, 1995/96 to sidy). In the latter year, however, prices rose toward the end 2004/05, and 2006/07 seasons. Indeed, there were reports of the season, so that the maximum monthly price within the of maize imports of roughly 50,000 tons from Mozam- year was high, although substantially larger estimated pro- bique and 36,000 tons from South Africa during the duction as compared with the previous year should have led 2008/09 season (FEWSNet 2009; South African Revenue to lower prices. Prices following the 2007/08 subsidy were Service 2009). even higher, with average annual prices exceeding those of ■ Rising real incomes alongside falling poverty rates and the 2001/02 and 2005/06 famine years,19 but these high rising population may lead to rising national demand, prices did not lead to any reports of widespread suffering which would cause the 1995/96 to 2004/05 demand to and distress such as those experienced in previous years with increase over time—but not as suddenly and dramati- equivalent prices. Prices in the first half of the 2009/10 sea- cally as shown in figure 17.5. Such a trend is compatible son (following the 2008/09 subsidy) were considerably lower with the lack of distress in later years despite high prices. than in the previous year, although the estimated production ■ Storage losses may be rising as a result of increasing pro- was still very high by historical standards (figure 17.5). duction of hybrid maize promoted by the 2006/07 and A number of explanations for this pattern of prices and subsequent subsidy programs. However, 2009 household estimated production may be put forward: survey results suggest that storage losses are not particu- larly high, with 50 percent of respondents reporting no ■ In 2007/08, roughly 330,000 metric tons of maize were losses in the 2007/08 and 2008/09 storage years and only exported. In 2009/10, 130,000 tons were purchased by a little over 20 percent reporting high losses (Dorward government for the strategic grain reserve (SGR), and a and Chirwa 2009b). Mangisoni (2010) also reports rela- further 100,000 tons are estimated to have been bought tively low storage losses, in the range of 12 percent over a and held in storage by private traders (FEWSNet 2009). 10-month period. CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 305 Figure 17.4 Mean Annual Maize Price in Malawi, by Marketing Season 0.40 Subsidy years 0.35 Maize price at 1990 US$/kg 0.30 0.25 0.20 0.15 0.10 0.05 0 2 4 6 8 0 2 4 6 8 0 /9 /9 /9 /9 00 /0 /0 /0 /0 /1 91 93 95 97 01 03 05 07 09 /2 19 19 19 19 99 20 20 20 20 20 19 Source: MoAFS. Note: Maize prices are simple averages across all markets obtained from weekly surveys, all deflated to 1990 prices. Figure 17.5 Peak Monthly Maize Prices in Malawi by Estimated Maize Supply per Capita, by Season 0.50 0.45 2000/01 Maximum maize price per kilogram 1994/95 2007/08 2004/05 0.40 1996/97 1997/98 (in 2005 U.S.$) 0.35 1993/94 0.30 2006/07 2001/02 2008/09 0.25 1995/96 1998/99 0.20 2003/04 2002/03 1999/2000 0.15 2005/06 0.10 0.10 0.15 0.20 0.25 0.30 Per capita maize crop estimates plus net imports (metric tons) 1993/94–2004/05 2005/06–2008/09 Source: Calculated from MoAFS annual crop production estimates and weekly maize price data with export-import data from Jayne et al. 2010, Minot 2009, and FAO 2009. Note: Estimated maize supply = crop estimate plus exports – imports. Labels show production season. ■ Higher welfare and real incomes following the 2005/06 ■ Finally, national maize production following the imple- harvest and low maize prices led to greater retention and mentation of the subsidy program could have been over- consumption of the 2006/07 harvest and hence a thinner estimated. Although there are no clear changes in and tighter market. methodology in the last few years, the method appears to ■ Changes in informal cross-border flows could also have rely substantially on field workers’ subjective estimates of occurred (Jayne et al. 2010). crop area and yield, which may be affected by the very 306 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 substantial involvement of field workers in the subsidy kilogram of nitrogen. Results from crop-cutting survey esti- program. Production impacts of the subsidy program are mates in the 2008/09 crop year demonstrate that substantial discussed in more detail below. problems in obtaining precise estimates of crop responses21 make it difficult to obtain precise estimates of incremental None of these explanations (except possibly the last) can production from the subsidy program. They do, however, fully explain the high prices despite the high estimated pro- support the broad response of 12 to 18 kilograms of grain duction following the 2006/07, 2007/08, and 2008/09 sub- per kilogram of fertilizer, with 15 kilograms per hectare a sidy programs. To illustrate this, figure 17.5 plots maximum reasonable “medium expectation.” Using different estimates monthly price against estimated per capita net maize sup- of incremental fertilizer use and of yield responses gives dif- ply. From the 1993/94 to 2004/05 production seasons, there ferent estimates of incremental production as a result of the was a roughly downward sloping relationship, with high input subsidy program. Table 17.8 sets out such estimates by prices following years of low supply and low prices in years year of implementation. following high supply (although the 2002/03 and 2003/04 For each year estimated, incremental fertilizer use is mul- seasons do not fit this relationship). Prices following the tiplied by a grain-to-nitrogen response ratio adjusted to first (2005/06) subsidy fit this pattern. The three subsequent reflect differing conditions in subsidy implementation seasons, however, when prices were high despite high esti- between years (good weather but very little hybrid seed mates of production, do not fit this pattern. in 2005/06 compared with 2008/09; similar conditions in Two clear and important conclusions emerge from this 2006/07 but a bit less hybrid seed; late fertilizer delivery analysis. First, in three out of four years, the subsidy pro- in the southern region and only slightly more hybrid seed in gram did not lead to lower market prices for maize. Second, 2007/08). This is added to a yield gain from subsidized the subsidy program has not led to increases in maize sup- hybrid seed separate from fertilizer to arrive at an estimate ply as large as those suggested by increases in maize crop of total incremental maize production from the subsidy.22 estimates from 2006/07 onward, particularly in 2007/08 and High and low estimates are, respectively, 20 percent above 2008/09, when prices appeared to be very high compared and below the 2008/09 medium estimate (which averaged with estimated supplies. We therefore consider estimates of 15 kilograms of grain per kilogram of fertilizer across the subsidy programs impacts on production. hybrid and local seed plots). While the estimates in table 17.8 are necessarily approxi- mate, indeed indicative, as noted above, they nevertheless Production impacts demonstrate several important points: The major stated objectives of Malawi’s subsidy program have been to achieve food self-sufficiency and to increase ■ Incremental production is very sensitive to yield the incomes of resource-poor households through responses to inputs (hybrid seed and fertilizers) and the increased food and cash crop production. Increased pro- potential is therefore considerable for raising yields and duction is therefore critical to achievement of program yield responses with good subsidy program and crop objectives. This results from incremental use of inputs management—with early subsidy sales, planting and fer- (mainly fertilizers and seeds) leading to increased yields, tilizer application, high plant populations, and greater with yield responses to these inputs dependent upon the use of organic matter, for example. weather and the efficiency of input use and of crop produc- ■ Not explicitly shown in table 17.8 is the importance of tion. Estimated incremental fertilizer sales were discussed hybrid seed in raising yield responses to fertilizer. earlier in terms of the effects of displacement on input sup- Increasing hybrid seed sales (subsidized or unsubsidized) ply markets. Incremental fertilizer sales are also important is therefore another potential way of increasing subsidy for estimating the incremental production effects of the impacts on incremental production.23 program, with responses to fertilizer depending upon rain- ■ Incremental production estimates are considerable, and fall, crop variety, and management (including timing of they grow over the life of the program as a result of the planting, weeding, and timing and methods of fertilizer increasing volume of incremental fertilizer use and application), and soil fertility. increasing supply of hybrid seed. SOAS et al. (2008) and Dorward and Chirwa (2009a) cal- ■ Incremental production estimates are, however, consid- culate estimated incremental production for 2005/06 to erably less than the production increases estimated in the 2007/08 using a range of 12 to 18 kilograms of grain per national crop estimates for maize production since the CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 307 start of the subsidy program, and these differences are Table 17.9 Trends in Macroeconomic Performance too large to be explained by upward revision of the yield Indicators, 2005–09 response to fertilizer.24 (percent) ■ Differences in production between presubsidy and sub- Indicator 2005 2006 2007 2008 2009 sidy years as estimated above are more compatible with Real GDP growth 3.3 6.7 8.6 9.7 6.9 price differences between these years as shown in figure Inflation 15.4 13.9 8.0 8.7 10.1 17.5 (for example the export of 330,000 metric tons fol- Deficit/GDP lowing the 2006/07 harvest would more than cancel out ratio (budget) 2.6 1.5 1.8 1.9 3.7 Deficit/GDP the increased subsidy impact that year, compared with ratio (actual) 0.4 1.4 4.0 6.3 8.0 2005/06). The very high prices in 2007/08 remain a puzzle Source: Reserve Bank of Malawi 2010. but may be explained by the subsidy program’s incre- mental production being insufficient to counteract pro- duction losses from adverse conditions affecting all maize production, and improved macroeconomic management in some parts of the country (late subsidized input delivery and conditions were other important contributors). and local events such as flooding and drought spells). Improved macroeconomic management, together with budgetary support from donors, was also undoubtedly important in enabling the government to finance such a Macroeconomic impacts large program. The large size of Malawi’s subsidy program could be The situation was, changing, however, as increasing vol- expected to have macroeconomic impacts. As a proportion umes and increasing prices in subsequent years led to very of total government expenditure, the subsidy increased high cost overruns. At the same time, Malawi’s economy was from 5.6 percent in 2005/06 to 8.4 percent in 2006/07 to 8.9 facing a number of internal and external pressures that led to percent in 2007/08. With very large increases in fertilizer adverse changes in macroeconomic indicators. The subsidy prices and costs for 2008/09, actual expenditure on the sub- program both contributed to and was affected by these sidy rose to 16.2 percent of total government expenditure macroeconomic changes, this time adversely (other macro- (see table 17.5). As a proportion of gross domestic product economic pressures were very high government expenditure (GDP), subsidy program costs rose from 2.1 percent in and import costs for the subsidy program with high fuel 2005/06 to 3.4 percent in 2007/08 and to 6.6 percent in costs, high maize prices, other government expenditures, and 2008/09 (excluding remittance by ADMARC and SFFRFM lower tobacco prices), although the incremental maize pro- of the farmer’s redemption price to government). (As noted duction from the subsidy program should have exerted a earlier and shown in table 17.5, program costs have subse- downward influence on maize prices. In a fixed exchange quently fallen back markedly.) rate environment, these pressures contributed to a foreign On the positive side, estimates of GDP growth have been exchange crisis in Malawi in November and December 2009. significantly affected by large increases in estimated maize Very high budgetary and foreign exchange allocations to production since the implementation of the subsidy pro- the subsidy program also reduced funding available to activ- gram. Estimates of incremental production attributable to ities such as health, education, and infrastructure develop- the subsidy program are not as high but are nevertheless ment. It is clear that the 2008/9 level of spending on the pro- very large, and incremental maize production and increases gram was not sustainable, and as noted earlier the in land and labor productivity in maize production attrib- government is addressing this: although the very high fertil- utable to the program should have had a significant positive izer prices in the 2008/09 season were a temporary phe- impact on GDP growth. nomenon, the government has committed itself to control- No evidence of negative macroeconomic impacts was ling costs by limiting the volume of subsidized fertilizers in found in the subsidy’s first two years (SOAS et al. 2008). future years. It has also restricted the subsidy to inputs for Important contributors to GDP were sound macroeconomic the production of only maize. management; improving macroeconomic indicators includ- ing growth, inflation, and government deficit (table 17.9); Economic returns and increased growth across the economy at that time; and the subsidy program itself was a contributor to that Economic returns to Malawi’s subsidy program depend growth (good tobacco prices, good weather for agricultural upon the economic price of maize, the price of inputs, and 308 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 production responses to increased input use. Producer ben- dized inputs. These should lead to two main types of direct efit-cost ratios estimated for the 2006/07 program showed benefit for recipients: immediate income transfers (from that the net economic return to the project is very sensitive reselling of coupons or subsidized inputs or from reduced to maize prices and the production response, and, with rea- expenditure on inputs as a result of displacement of unsub- sonable variation in assumptions, these ratios range from sidized purchases by cheaper, subsidized purchases), or 0.81 to 1.30, with a mid estimate of 1.06 (table 17.10). incremental production at harvest if the inputs are used on Adjustments to this analysis using estimated maize and fer- farm. If poorer households sell their coupon(s), then imme- tilizer prices for other program years suggest that both the diate income and welfare gains should also be accompanied 2005/06 and 2007/08 programs should have yielded equiva- by an easing of short-term seasonal cash constraints. This lent or higher returns. However, the very high fertilizer easing may reduce the extent to which they have to hire out prices that prevailed when fertilizers were being purchased their own labor (as casual ganyu labor) to obtain food, thus for the 2008/09 program adversely affected returns in allowing them to work more on their own farms, both 2008/09, despite good weather and yields and high maize increasing end-of-season yields and reducing the supply of prices (although these did offset the effects of high fertilizer labor into the local labor market. At the same time, if less prices to some extent). poor households obtain cheaper inputs directly from the Fiscal efficiency estimates (net economic benefit per program or buy those inputs from subsidy recipients, then unit of fiscal investment) show a similar pattern to eco- this increase in income should increase their demand for nomic returns, but in addition these are (negatively) hired-on farm labor or for local goods and services. The affected by high rates of displacement of unsubsidized sales result should be a tightening of both demand and supply in by subsidized sales (displacement lowers the net benefit of the local labor market, and a consequent rise in wages, to subsidized sales). Key conclusions from the benefit-to-cost the benefit of poorer households. and fiscal efficiency analyses are that economic returns are Further direct and indirect benefits may also accrue in highly sensitive to the yield response to fertilizer (as dis- subsequent seasons. At the end of the season, higher maize cussed earlier under production impacts); fiscal returns are production should result from incremental fertilizer use highly sensitive to displacement rates; and with good pro- and increased crop labor inputs, increasing households’ gram implementation and good (but achievable) yield incomes. Higher production should also depress maize responses to fertilizer, the program can be a very good prices. Higher household maize stocks and lower maize investment. It is therefore critical that the program design prices carried forward into the following season should and implementation deliver low displacement and high again benefit poorer households, reducing their seasonal responses to inputs. cash flow constraints and their need to hire out their labor, so that they can again work more on their own farms. This will again tighten the labor market. Higher incomes from Growth and poverty reduction impacts higher wages should stimulate demand for nonfarm goods While the producer benefit-cost and fiscal efficiency analy- and services, with spin-off benefits and multipliers in the ses can yield valuable information about the efficiency of local economy. the subsidy program, they can be misleading when examin- There is considerable empirical support from Asia and ing the contributions of the program to poverty reduction, from Africa for the importance of some of these processes economic growth, and food security. Understanding the full and of indirect effects of agricultural growth on wider eco- economic benefits of the program requires consideration of nomic growth. Hazell and Rosegrant (2000) show that the the direct effects of the program on subsidy recipients and indirect effects of increased agricultural production in the of the different ways that these effects subsequently work green revolution in Asia were the major process driving pro- through their own and others’ livelihoods and the rural poor growth in the second half of the 20th century. There is economy. Because the program is large, it is very important also a large body of literature on agricultural growth “mul- for the wider market effects of the intervention to be prop- tipliers” in Africa, with estimates that vary from around 1.5 erly recognized. to over 2.0 (a multiplier of 1.5 indicates that $1.00 of extra Figure 17.6 shows three possible uses of the subsidy by income from agricultural production results in further subsidy recipients: reselling of coupons or of subsidized income growth of $0.50) (Hazell and Hojjati 1995; Reardon inputs, incremental use of the inputs in production, or use 1998; Delgado, Hopkins, and Kelly 1998). Studies of countries of the inputs with displacement of purchase of unsubsi- across Africa and Asia also generally show that consumption CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 309 310 Table 17.10 Estimated Economic Impacts of Subsidy Programs in Malawi Economic impact 2005/06 2006/07 2007/08 2008/09 Domestic maize prices ($/mt) May–Oct Nov 2006– May–Oct Nov 2007– May–Oct Nov 2008– May–Oct Nov 2009– 2006 Apr 2007 2007 Apr 2008 2008 Apr 2009 2009 Apr 2010 139.5 142.7 119.3 243.8 336.6 434.6 271.4 307.1 Maize price in benefit-cost and fiscal efficiency analysis ($/mt) 143 154 250 280 Fertilizer price in analysis ($/mt) 393 490 590 1,250 Benefit-cost ratio, high response 1.38 1.30 1.90 1.08 Benefit-cost ratio, moderate response 1.12 1.06 1.54 0.90 Benefit-cost ratio, low response 0.86 0.81 1.18 0.72 Fiscal efficiency, high response 0.76 0.44 1.13 0.09 Fiscal efficiency, moderate response 0.24 0.09 0.68 negative Fiscal efficiency, low response negative negative 0.23 negative 2005 2006 2007 2008 Poverty incidence 50% 45% 40% 40% Meals per day 2.0 2.2 2.3 2.3 Sources: SOAS et al. 2008; Dorward, Chirwa, and Slater 2010b; Food and Nutrition Security Program 2008; NSO 2006, 2009. Note: The benefit-cost ratio and fiscal efficiencies were calculated with high, medium, and low fertilizer responses of 18, 15, and 12 kilograms of grain per kilogram of fertilizer, respectively. The benefit-cost ratio is calculated as the gross incremental benefits divided by the gross incremental cost, valued at social prices. Fiscal efficiency is calculated as the net economic benefit divided by fiscal cost. Figure 17.6 Tracing Direct and Indirect Subsidy Impacts Rural households Farm/non-farm investment Poorer households Resale Y1 Increased Y2 Increased real incomes real incomes Input Incremental subsidy use Y1 Increased Y2 Increased Displacement production production use less poor households Y1 Increased Y2 Reduced Y2 Increased wages maize prices wages Rural Input service demand Farm/nonfarm economy and investment demand and investment Source: SOAS et al 2008. Note: Solid lines indicate positive impacts for poorer, food-deficit producers and sellers of labor; dashed lines indicate negative impacts for the less poor, maize surplus producers, and buyers of labor. linkages are more important than production linkages, SOAS et al. 2008) and for 2008/09 (Dorward 2010). These accounting for between 50 percent (in Senegal) and 98 per- studies provide results consistent with those of the focus cent (in Zambia) of overall multipliers calculated (Delgado, group discussion reports, finding that the subsidy con- Hopkins, and Kelly 1998). These studies do not, however, tributes to higher wages but had lower wage effects in explicitly consider the further effects of transfers and 2008/09 and that subsidy benefits have become more con- growth on seasonal capital constraints, or the knock-on centrated among beneficiaries (because the value of direct effects of relaxing these constraints. production benefits to subsidy recipients is increased with The extent of these effects and of direct and indirect higher maize prices, but the benefits of lower maize prices increases in wages and benefits to poorer households as a and higher wages to poor nonbeneficiaries are absent or result of implementation of the subsidy program is, however, reduced). Fragmented information on wage rates over the an empirical question, as is the extent to which productivity period 2005/06 to 2008/09 suggests that wage rate increases and welfare benefits are carried forward from one year to over the period were higher than maize price increases over the next. Focus group discussions in 2007, reporting on the the same period, representing real increase in wage rates. effects of the 2005/06 subsidy, clearly (and independently) Further empirical work is required to investigate the articulated processes of easing of seasonal cash constraints scale of the indirect benefits through wage effects, because in the hungry gap and tightening of labor markets, with they depend heavily upon the scale of growth multipliers higher wages and low maize prices throughout the season (consumption multipliers of slightly more than 1.5 embed- (SOAS et al. 2008). Similar focus group discussions con- ded in the livelihood and rural economy models are ducted in 2009 reported income benefits from higher wages derived from historical expenditure patterns in Malawi, and increased maize availability, though they were also con- which are augmented by savings multipliers in the model). cerned about the adverse effects of high maize prices—and Without these multipliers, wage impacts would be very the high maize prices experienced after 2006/07 would be low—analysis of 2006/07 and 2008/09 household survey expected to undermine these processes. data suggests that agricultural wage labor constitutes less Livelihood and rural economy models have also been than 20 percent of household income even among the used to investigate and describe this process in 2006/07 (see poorest 20 percent of rural households in Malawi, although CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 311 observed rises in agricultural wages suggest that unskilled, Other African countries considering the introduction of nonagricultural wages have also risen, unless these two labor agricultural input subsidies can learn important lessons markets are completely separated, an unlikely scenario. from Malawi’s experience. The program is a bold, large- The overall increase in real wages over the period is sup- scale initiative that has achieved substantial increases in ported by anecdotal reports of rising rural wages and is con- maize production. The implementation of such a program sistent with the limited political and social impact of and represents a very considerable logistical achievement, and response to high nominal maize prices in early 2009. It is the government is to be commended for this and for its con- also consistent with reports of falling poverty rates and tinuing and often imaginative attempts to improve the decreased wasting among children under the age of five. The program. Nevertheless, higher maize prices and calculation subsidy program is not the only contributor to these of the agronomic yield effects of incremental fertilizer and improvements. Over the same period, there were good rains, seed suggest the production increases resulting from the a marked improvement in macroeconomic management, program are not as large as might appear from the post- and relatively high tobacco prices. These factors will have 2005/06 national maize production increases reported by made direct contributions to economic growth, and also MoAFS. The benefits of the program also have to be facilitated the implementation and impact of the subsidy weighed against its very considerable costs (with an average program (the tobacco prices and macroeconomic manage- of over 9 percent of the national budget going to the subsidy ment contribute to the availability of foreign exchange for program since 2005/06) and the loss of benefits from alter- fertilizer imports, reduce the crowding-out effects of the native investment of these funds. While the food security program, and stimulate other complementary parts of the and growth benefits of the program have been partially economy; good rains promote good yield responses to undermined by high maize prices, there have still been sig- fertilizers). nificant improvements in productivity and welfare from While attribution of these changes to the program is dif- 2005/06 associated with greater maize availability and with ficult, there are reasons to believe that the stimulus to maize increases in real wages. production from the subsidy and the good rains has been a Ongoing implementation challenges that the govern- critical element in rural economic growth and poverty ment is working on include controlling costs, timing of reduction, as compared with the effects of good tobacco input deliveries, effective targeting of subsidized inputs, prices alone. Limited evidence from elsewhere in Africa and reducing diversion and fraud, improving agronomic and the world suggests that even labor-demanding smallholder market returns with complementary investments (for cash crops can drive pro-poor growth in such a broad way. example, in extension, research, organic soil fertility Further empirical work is also needed to establish the extent improvement, and roads), and using the subsidy program to to which productivity and welfare gains in one year are car- extend private sector input delivery systems. Success in ried forward to subsequent years, the conditions under addressing these challenges will lead to new challenges which these gains can be maximized, and the implications because increasing success will lead to changes in the need for subsidy program design and implementation and ques- for such a large-scale program with these objectives—and tions about graduation. continuing political, strategic, economic, technical, and logistical system innovations will be needed to respond to these changes. CONCLUSIONS AND LESSONS Malawi’s experience with its large-scale agricultural This chapter suggests that the Malawi agricultural input sub- input subsidy program offers a number of important les- sidy program has achieved substantial benefits and successes, sons to other countries in Africa considering the introduc- although these are more nuanced than some press reports on tion of agricultural input subsidies. the program suggest. The measured producer benefits of the First, any growth and development strategy that involves program relative to its costs have been relatively modest; agricultural input subsidies must be rooted in the opportu- however, our benefit-cost analysis does not capture all the nities for and constraints to growth and development facing benefits of the program, nor all of its costs. Moreover, there a country and particular groups within it. This chapter has is scope for considerable improvement in the program’s set out in some detail specific difficulties that constrain effectiveness and efficiency, although there are also practical broad-based growth in Malawi, highlighting the reliance on and political difficulties regarding the implementation of low-productivity maize and the difficulties and limited some of these and questions about their effects. options faced by very large numbers of poor, food-deficit 312 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 farmers, and indeed by the Malawian economy as a whole, estimates and the number of farm families in Malawi in breaking out of the low-maize-productivity and poverty also demonstrate the importance of reliable information traps. for issues beyond specific matters related to the imple- This low-productivity trap arises as a result of severe sea- mentation of the subsidy program and assessments of its sonal credit constraints affecting very large numbers of impacts. poor, food-deficit farming families, together with thin and ■ Effective targeting and rationing systems are needed to high-risk, high-margin input and maize markets. Malawi’s limit scale and increase subsidy impacts on productivity, key achievements with its subsidy program have been the but different (geographical or household) approaches ability to raise land and labor productivity and improve face different costs and difficulties (Dorward 2009), and food security for large numbers of poor households by in some situations strict rationing of universal provision relieving both profitability and affordability constraints on may be a practicable alternative. the use of inputs needed to increase staple crop productiv- ■ Entitlement systems are needed for targeting and ity, leading to some combination of increased real wages rationing, and these need to be robust against inevitable and reduced food prices. The Malawian model thus applies counterfeiting and diversion. to other countries only if there are large numbers of people ■ Input supply system development requires close attention facing similar staple-food-productivity constraints along- to the complementary and changing roles and interests side increased input use constrained by thin input markets, of different public sector and commercial stakeholders, poorly developed input supply systems, and widespread but improved farmer access to input services should be a profitability and affordability problems. major objective and outcome of agricultural input sub- Malawi’s experience also shows that, in the right context, sidy programs. large-scale agricultural input subsidy programs have the ■ Complementary policies and investments: if a subsidy pro- potential to yield substantial benefits to people and their gram is seen as part of a broad, long-term strategy for governments with good design and implementation. The poverty reduction and economic development, then chapter has also shown the very substantial costs and investments in complementary activities must be made resources required for such programs, and the difficulties in areas such as extension, research, organic soil fertility and challenges that must be overcome for effective, efficient, improvement, health, education, markets, transport and and sustainable delivery of program benefits. Several issues communication infrastructure, and services. Considera- from Malawi’s experience are relevant to other countries tion of the different roles of these complementary invest- considering similar subsidy programs: ments should also guide decisions on the nature, scale, and implementation of the input subsidy, as well as of ■ Focus: subsidies should be provided for inputs whose use other investments, in order to achieve positive interac- for important staple crops is constrained by affordability tions among investments. difficulties despite high potential responses to input use. ■ Macroeconomic management to promote favorable ■ Consumer gains: strong emphasis should be put on wider growth conditions and provide budgetary resources contributions to economic growth and poverty reduc- needed for such a program is also important. tion through consumer as well as producer gains. ■ Political commitment is required for sustained mobiliza- ■ Scale: the subsidy should affect staple crop prices, labor tion of program resources, but there may also be poten- markets, or both, requiring sufficient local or national tial conflicts between the need for political support on scale to affect markets, but strict limits on scale and the the one hand and targeting, rationing, cost control, and control of costs are needed to limit displacement of exist- performance monitoring needed for efficient and sus- ing purchases, crowding out of critical complementary tainable implementation on the other. investments, and adverse macroeconomic impacts. ■ Sustainability of program implementation should be ■ Logistical systems face major challenges in delivering addressed by attention to cost control, scale, and logistical timely, targeted subsidies to large numbers of widely dis- and performance monitoring and audit systems. There persed farmers, and the establishment of such logistical is also need for investigation of sustainability of impacts, systems requires time and major investments. with examination of the extent to which productivity and ■ Performance monitoring, information, and auditing systems welfare gains carry forward from one year to the next and are needed to develop trust, control fraud, and promote the implications of this analysis for program design efficiency and effectiveness. Debates on crop production and implementation and for questions about graduation. CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 313 An innate dilemma in the design and implementation of months after harvest, and that in 2003/04 one bag of fertil- large-scale subsidy programs is that such programs require izer represented approximately 10 percent of rural house- both stability and flexibility, with innovation. Stability is holds’ median per capita annual expenditure (and more needed to provide stakeholders with confidence and security than 20 percent of median per capita expenditure of the that will justify long-term financial and other investments lowest expenditure quintile). associated with the program’s implementation. Stability can 6. A ratio of 15 was used in the calculations cited above be undermined, however, by the need for flexibility to adjust based on what are believed to represent mean grain-to- nitrogen response rates to fertilizer application on farmers’ to changing conditions (for example in the weather, in fields. Improved management and uptake of hybrid seed international and national markets and economies, or in provide the potential for higher ratios (in the range of 22 to politics), and some of these changes may be anticipated or 28 kilograms of maize per kilogram of fertilizer applied). unanticipated results of the program. Alongside flexibility is Hence, improved farm management practices have the the need for innovation (in technology, systems, and prices) potential to make fertilizer use on maize profitable even to take advantage of learning and change during program without subsidy, although the affordability constraint implementation. Although flexibility and innovation can remains. undermine stability, lack of flexibility and innovation may 7. This section draws heavily on Dorward (2009). also undermine stability if conditions, such as increasing 8. This analysis applies only to subsidies implemented on a incidence of fraud, make the system unsustainable and inef- scale large enough to affect output prices. Small-scale subsi- fective in its initial form. To achieve mutually supportive dies that do not significantly affect product prices are equiv- stability, flexibility, and innovation, trust and stable princi- alent to highly elastic product demand: subsidy benefits are ples must govern both the long-term objectives of and rela- largely captured by suppliers and producers, and dead- tions between different stakeholders on one hand and the weight costs depend upon the elasticity of supply. processes for successful learning, flexibility, and innovation 9. This is not a problem in situations in which the on the other. providers of land and labor are poor. Indeed, subsidies can promote pro-poor growth in such situations. NOTES 10. For this, complementary constraints to effective use of inputs by targeted beneficiaries must be addressed. 1. This section draws heavily on material from SOAS (2008). 11. The terms “coupon” and “voucher” are used inter- changeably in wider discussions of the Malawi Agricultural 2. Reasons for the high dependency on maize as opposed Input Subsidy program and in this paper. to other food crops include dietary preferences, different crops’ relative calorific yields per hectare in different agro- 12. Efficient and stable output markets may require less ecologies, farmers’ familiarity with the crop, and long- government intervention in direct market operations but standing strong government policies aimed at promoting more focused investment in market insurance and in facili- maize production and input and crop marketing subsidies tating infrastructure and institutions. focused on maize. 13. This section draws heavily on SOAS (2008) and on 3. Understanding of the nature, causes, and relative impor- Dorward and Chirwa 2009a. tance of these problems varies (indeed, elements of the 14. No information is available on seed sales. analysis presented here are not universally accepted, nor is 15. A standard maize subsidy pack in 2005/06 consisted of this paper a comprehensive account of the complex issues one voucher for a 50 kilogram bag of 23:21:0+4S, one involved). voucher for a 50 kilogram bag of 23:21:0+4S, and one 4. The postharvest value-to-cost ratio has generally been voucher for improved maize seed. A standard tobacco sub- less than 2, widely considered to be the minimum required sidy pack consisted of one voucher each for a 50 kilogram to make fertilizer use profitable in moderately but not bag of calcium ammonium nitrate (CAN) and a 50 kilo- highly risky situations (Morris et al. 2007). Even for higher gram bag of D Compound. The fertilizer vouchers were preharvest maize prices, the ratio typically has been around redeemable for MK 950. The seed voucher required no top- or below 2 (SOAS 2008), depending on the yield response up and could be used to purchase 2 kilograms of hybrid achieved. With high yield responses, the value-to-cost ratio seed or 3 to 4 kilograms of OPV seed, depending on the has been above 2 in some years (Maize Productivity Task price set by the seed supply company. Force 1997). 16. These variations in targeting stem from vagueness in the 5. Table 17.1 shows, for example, that the median for definition of target beneficiaries in the guidelines and dif- maize stocks running out each year is between four and six ferences in the way communities dealt with problems of 314 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 shortages. Variations meant that those that were targeting and Market Interventions.” Prepared for the COMESA beneficiaries placed different emphasis on different criteria policy seminar “Variation in Staple Food Prices: Causes, and processes. Consequence, and Policy Options.” Maputo, Mozam- 17. Displacement for 2007/08 is assumed to be similar to that bique, 25-26 January 2010 under the Comesa-MSU- for 2006/07 (higher subsidy sales may increase displacement, IFPRI African Agricultural Markets Project. but greater farmer familiarity and higher fertilizer prices Chirwa, E., M. Matita, and A. Dorward. 2010. “Factors would be expected to reduce displacement). A lower displace- Influencing Access to Agricultural Input Subsidy ment is assumed for 2008/09 as a result of much higher fer- Coupons in Malawi.” Future Agricultures Consortium, tilizer prices and earlier (separate) beneficiary registration. Brighton, U.K. 18. Low average prices in some years in the early 1990s were Crawford, E. W., T. S. Jayne, and V. A. Kelly. 2006. “Alterna- brought about by large-scale imports, and at the turn of the tive Approaches for Promoting Fertilizer Use in Africa.” century by the starter pack subsidy program (with good Agriculture and Rural Development Discussion Paper weather). 22, World Bank, Washington, DC. 19. Although the annual average price in 2008/09 was very Delgado, L. C., J. Hopkins, and V. A. Kelly. 1998. “Agricul- high, peak prices were equivalent to those in 2001/02 and tural Growth Linkages in Sub-Saharan Africa.” IFPRI 2005/06 because prices rose much earlier in the season in Research Report. IFPRI, Washington, DC. 2008/09 but then flattened out. Djurfeldt, G., H. Holmen, M. Jirstrom, and R. Larsson, eds. 20. The latter may be higher if there are significant exports 2005. The African Food Crisis: Lessons from the Asian or commercial carryover stocks at the end of the season. Green Revolution. Wallingford, U.K.: CABI Publishing. 21. These include inherent and difficult-to-quantify biases Dorward, A. R. 2006. “Markets and Pro-poor Agricultural in different methods of collecting yield data, multicollinear- Growth: Insights from Livelihood and Informal Rural ity between input use and other management variables, high Economy Models in Malawi.” Agricultural Economics 35 variability in smallholder agriculture, and variation of (2): 157–69. response rates with average rates of subsidized fertilizer use. ———. 2009. “Rethinking Agricultural Input Subsidy Pro- 22. This calculation considers only maize production. grams in Developing Countries.” Non-Distorting Farm Provision of the fertilizer subsidy for tobacco leads to a Support to Enhance Global Good Production, ed. A. much higher displacement than for maize; thus, assuming Elbehri and A. Sarris, pp. 311-74. Rome: Food and Agri- that all incremental fertilizer is used on maize is not culture Organization. Also available at http://www.fao unreasonable. Fertilizer impacts on crops mixed with .org/es/esc/common/ecg/586/en/Dorward_FAO_Subsidy_ maize are ignored here. Paper_FINAL.pdf. 23. It is important to note that increased hybrid seed subsi- ———. 2010. “Informal Rural Economy Modelling of Sub- dies may not be required if farmers are able and willing to sidy Impacts.” Unpublished paper, SOAS (School of Ori- purchase unsubsidized seed—improved access to seed, ental and African Studies), London. increased distributors, and effective extension may be more Dorward, A. R., and E. Chirwa. 2009a. “The Agricultural effective in increasing hybrid seed use. Input Subsidy Program 2005 to 2008: Achievements and 24. For 2007/08, however, low crop output resulting from Challenges.” School of Oriental and African Studies, local droughts and floods affected total production, not just London. incremental production from the subsidy. ———. 2009b. “The Malawi Agricultural Input Subsidy Programme: 2009 Fieldwork Preliminary Report.” School of Oriental and African Studies, London. REFERENCES ———. 2010. “Evaluation of the 2008/09 Agricultural Input Ariga, J., and T. Jayne. 2009. “Private Sector Responses to Subsidy Malawi: Maize Production Impacts.” School of Public Investments and Policy Reforms. The Case of Fer- Oriental and African Studies, London. tilizer and Maize Market Development in Kenya.” Dis- Dorward, A., E. Chirwa, and R. Slater. 2010a. “Evaluation of cussion Paper 00921. International Food Policy Research the 2008/09 Agricultural Input Subsidy Malawi: Prelimi- Institute (IFPRI), Washington, DC. nary Report on Program Impact.” School of Oriental and Ariga, J., T. Jayne, B. Kibaara, and J. K. Nyoro. 2008. “Trends African Studies, London. and Patterns in Fertilizer Use by Smallholder Farmers in ———. 2010b. “Evaluation of the 2008/09 Agricultural Kenya, 1997–2007.” KeWP 28, Tegemeo Institute, Nairobi. Input Subsidy Malawi: Report on Program Implemen- Chapoto, A., and T. S. Jayne. 2010. “Maize Price Instability in tation.” School of Oriental and African Studies, Eastern and Southern Africa: The Impact of Trade Barriers London. CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 315 Dorward, A. R., J. G. Kydd, J. A. Morrison, and I. Urey. 2004. Maize Productivity Task Force, Ministry of Agriculture of “A Policy Agenda for Pro-Poor Agricultural Growth.” Malawi. 1997. “Area-specific Recommendations for World Development 32 (1): 73–89. Hybrid Maize Grown by Malawian Smallholders: A Man- Economist Intelligence Unit. 2008. “Lifting African and ual for Extension Field Assistants.” Lilongwe. Asian Farmers Out of Poverty: Assessing the Investment Mangisoni, J. 2010. “Defects and Storage Loss Analysis of Needs.” Custom research project for the Bill and Melinda Maize Grain in Malawi.” Unpublished paper, Bunda Col- Gates Foundation, New York. lege, University of Malawi. Ellis, F. 1992. Agricultural Policies in Developing Countries. MoAFS (Ministry of Agriculture and Food Security of Cambridge, U.K.: Cambridge University Press. Malawi). 2008. “The 2007/2008 Inputs Subsidy Program Fan, S., A. Gulati, and S. Thorat. 2007. “Investment, Subsi- Review Report: An Internal Review.” Department of dies, and Pro-Poor Growth in Rural India.” Discussion Agricultural Planning Services, Lilongwe. Paper 716, IFPRI, Washington, DC. Minot, N. 2009. “Staple Food Prices in Malawi.” Prepared Fan, S., S. Thorat, and N. Rao. 2004. “Investment, Subsidies, for the COMESA policy seminar “Variation in Staple and Pro-Poor Growth in Rural India: Institutions Food Prices: Causes, Consequences, and Policy Options.” and Economic Policies for Pro-poor Agricultural Maputo, 25–26 January 2010 under the African Agricul- Growth.” IFPRI discussion paper, IFPRI, Washington, tural Marketing Project, IFPRI. DC. Morris, M., V. A. Kelly, R. Kopicki, and D. Byerlee. 2007. FAO (Food and Agriculture Organization). 2009. FAO Agri- “Fertilizer Use in African Agriculture.” World Bank, cultural production data. Rome. Washington, DC. FEWS NET. 2009. “Malawi Food Security Outlook, October Munthali, M. W. 2007. “Integrated Soil Fertility Manage- 2009–March 2010.” Lilongwe. ment Technologies: A Counteract to Existing Milestone Food and Nutrition Security Program. 2008. “Monitoring in Obtaining Achievable Economical Crop Yields in Cul- and Evaluation, 6th Report,” (draft). Lilongwe. tivated Lands of Poor Smallholder Farmers in Malawi.” In Advances in Integrated Soil Fertility Management in Govereh, J., J. Shawa, E. Malawo, and T. S. Jayne. 2006. “Rais- Sub-Saharan Africa, ed. A. Bationo, B. Waswa, J. Kihara, ing the Productivity of Public Investments in Zambia’s and J. Kimetu, pp. 533–536. Dordrecht, The Netherlands: Agricultural Sector.” Working Paper 20, Food Security Springer. Research Project, Lusaka. Nakhumwa, T. O. 2006. “Rapid Evaluation of the 2005 Fer- Harrigan, J. 2003. “U-Turns and Full Circles: Two Decades tilizer Subsidy Program in Malawi.” Policy Paper No 10, of Agricultural Reform in Malawi 1981–2000.” World CISANET, Lilongwe. Development 31 (5): 847–63. NSO (National Statistical Office of Malawi). 2005. “Inte- Hazell, P., and B. Hojjati. 1995. “Farm/Non-Farm Growth grated Household Survey 2004 –2005.” Zomba. Linkages in Zambia.” Journal of African Economies 4 (3): 406–35. ———. 2006. “Welfare Monitoring Survey 2005.” Zomba. Hazell, P., and M. Rosegrant. 2000. Rural Asia: Beyond ———. 2008. “Welfare Monitoring Survey 2008.” Zomba. the Green Revolution. Manila: Asian Development ———. 2009a. “2008 Population and Housing Census: Pre- Bank. liminary Report.” Zomba. Jayne, T. S., J. Mangisoni, N. Sitko, and J. Ricker-Gilbert. ———. 2009b. “Welfare Monitoring Survey 2008.” 2010. “Malawi’s Maize Marketing System.” Report com- Zomba. missioned by the World Bank and Ministry of Agriculture Reardon, T. 1998. Rural Non-farm Income in Developing of the Government of Malawi, Lilongwe. Countries: The State of Food and Agriculture 1998. Rome: Kelly, V. A., D. Boughton, and N. Lenski. 2010. “Malawi Agri- Food and Agriculture Organization. cultural Inputs Subsidy Program: Evaluation of the Reserve Bank of Malawi. 2010. “Financial and Economic 2007/08 and 2008/09 Input Supply Sector Analysis.” Review 2009.” Lilongwe. Michigan State University, East Lansing, MI. Ricker-Gilbert, J., and T. Jayne. 2010. “The Impact of Fertil- Levy, S., ed. 2005. Starter Packs: A Strategy to Fight Hunger izer Subsidies on Displacement and Total Fertilizer Use.” in Developing and Transition Countries? Lessons from the Unpublished presentation, Lilongwe. Malawi Experience, 1998–2003. Wallingford, U.K.: Ricker-Gilbert, J., T. S. Jayne, and E. Chirwa. 2011. “Subsidies CABI. and Crowding Out: A Double-Hurdle Model of Fertilizer Logistics Unit. 2009. “Final Report: Implementation of Agri- Demand in Malawi.” American Journal of Agricultural cultural Inputs Subsidy Program 2008/09.” Lilongwe. Economics 93 (February): 26–42. 316 CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 SOAS (School of Oriental and African Studies), Wadonda Timmer, C. P. 1989. “Indonesia: Transition from Food Consult, Overseas Development Institute, and Michigan Importer to Food Exporter.” In Food Price Policy in Asia, State University. 2008. “Evaluation of the 2006/07 Agri- ed. T. Sicular, pp. 22–64. Ithaca, NY: Cornell University cultural Input Supply Program Malawi: Final Report.” Press. SOAS, London. World Bank. 1981. “Accelerated Development in Sub-Saharan South African Revenue Service. 2009. Monthly maize quan- Africa: An Agenda for Action.” World Bank, Washington, tities exported to Malawi. SARS database. Pretoria. DC. Siamwalla, A., and A. Valdes. 1986. “Should Crop Insur- Xu, Z., W. A. Burke, T. S. Jayne, and J. Govereh. 2009. “Do ance Be Subsidized?” In Crop Insurance for Agricultural Input Subsidy Programs ‘Crowd In’ or ‘Crowd Out’ Com- Development: Issues and Experience, ed. P. Hazell, C. mercial Market Development? Modeling Fertilizer Pomareda, and A. Valdes, pp. 117–25. Baltimore: IFPRI Demand in a Two-Channel Marketing System.” Agricul- and John Hopkins University Press. tural Economics 40 (1). CHAPTER 17: MALAWI’S AGRICULTURAL INPUT SUBSIDY PROGRAM EXPERIENCE OVER 2005–09 317 CHAPTER 18 MoneyMaker Pumps: Creating Wealth in Sub-Saharan Africa I. V. Sijali and M. G. Mwago he first of the Millennium Development Goals T purchases. Agricultural families have remained locked in a (MDGs) is to eradicate extreme poverty and low-input, low-income system, with low and stagnating hunger, with the target of halving the proportion agricultural yields. of people whose income is less than $1 a day between 2000 Agricultural intensification has become the main avenue and 2015. In Sub-Saharan Africa, rural poverty accounts of growth in crop production, with agricultural water man- for 83 percent of total extreme poverty, with 85 percent agement (water harvesting and irrigation) the focus of this of the poor depending partly on agriculture for their intensification in Sub-Saharan Africa. More reliable access livelihoods (World Bank 2000). Considering that agricul- to water for irrigation would stabilize yield, reduce the risk ture can have the greatest impact on poverty and food of harvest failure, and increase productivity and agricul- security if the benefits of its development are reaped by tural growth. Successful investment in agricultural water the poor, priority should be given to the development of management technologies therefore presents an important small-scale irrigation (Kidane, Maetz, and Dardel 2006). opportunity for poverty reduction and economic growth. Agricultural growth is therefore a key to the realization of However, major constraints to using irrigation—including the first MDG. lack of surface storage structures, appropriate irrigation Stimulating agricultural growth is critical to reducing technologies, capacity to exploit underground water, com- poverty in Africa. Commercial agriculture, potentially a munity skills for irrigation, information on appropriate powerful driver of agricultural growth, can develop along a irrigation technologies, and investment in irrigation— number of pathways. But agricultural productivity in Sub- must be addressed. Saharan Africa is the lowest in the world, with per capita In some parts of Sub-Saharan Africa, investment in output only 56 percent of the world average (FAO 2005), irrigation development at the community level is already and agricultural output in Sub-Saharan Africa has not kept bearing fruits. In Tanzania the Participatory Irrigation pace with population increase (Msangi and Rosegrant Development Project has had striking impacts, increasing 2005). More than 80 percent of output growth in Sub-Saha- farm income by 86 percent and enabling irrigator house- ran Africa since 1980 has come from expansion of the holds to enjoy better-quality housing, acquire agricultural cropped area, compared with less than 20 percent for all and household assets, access health services, and finance other regions. Food self-sufficiency among Sub-Saharan children’s education (IFAD 2005). In four representative Africans declined from 97 percent in the mid-1960s to subproject areas, ownership of oxcarts and cattle 82 percent in 1997–99, without a substantial increase in increased considerably, the total number of grinding mills household incomes that would allow people to afford food increased from 2 to 12, and the number of shops 319 increased from 2 to 74. Similarly, in Zimbabwe, irrigator KICKSTART INTERNATIONAL households in the European Union-funded Maunganidze KickStart International is a nonprofit social enterprise Irrigation Scheme reported an increase in income of more organization founded in Kenya in 1991. It now operates in than 200 percent and turned a food deficit into a surplus. Burkina Faso, Kenya, Mali, and Tanzania. KickStart’s mission Investment in new housing and in water and sanitation is to promote economic growth and employment creation in was the most obvious sign of improved livelihoods, with a Africa by developing and promoting technologies that entre- number of modern two- and three-room houses with preneurs can use to establish and run profitable small-scale ventilated pit latrines and in several cases protected water businesses. The entrepreneurs raise small amounts of capital wells. IFAD concluded that these impacts were the result ($100–$1,000) to start a new enterprise, and KickStart helps of investment in irrigation, because there were no other them to identify viable business opportunities and obtain the sources of income in the area. appropriate technologies required. Studies have shown that small-scale irrigation offers KickStart addresses poverty alleviation through devel- opportunities to improve livelihoods in Africa. However, opment of appropriate technologies and innovations to there has been little technology-driven jump in productivity enhance production, value addition, and income genera- for the majority of resource-poor African farmers. IFAD tion. These include, among others, irrigation pumps, an (2005) estimates that of the potentially irrigable 39.4 million oil press, and a brick press. KickStart’s irrigation pumps hectares in Africa, only 7.1 million hectares, or 18 percent enable smallholder farmers to enhance productivity and of the total, have been equipped for irrigation. Expansion improve household incomes and thus contribute to sus- of irrigation has been slow. Over the past forty years, only tainable poverty reduction. KickStart has developed several 4 million hectares have become irrigated in Africa, by far human-powered pumps. These pumps have the distinct the smallest expansion of any world region. The area advantage of being operable without fossil fuels or electric- developed for irrigation as a fraction of the irrigation ity, which is not always available in remote areas. potential for African countries is shown in map 18.1. A majority of African countries have exploited less than 50 percent of their potentially irrigable land, with 10 The business model countries at less than 10 percent, 8 countries at 10 to 25 KickStart’s business model is a fivefold strategy: percent, and 6 countries at 26 to 50 percent. Eleven coun- tries have exploited 50–75 percent, while only 2 countries ■ Market research for business opportunities. KickStart have exploited more than 75 percent of their potential identifies business opportunities through market (FAO 2005). research and feasibility studies that are aimed at assist- The lack of exploitation of irrigation potential in Sub- ing the rural poor. Saharan Africa can be attributed to lack of access to ■ Research and development. KickStart designs tools to end appropriate and affordable irrigation technologies by the poverty that can be used by small-scale entrepreneurs to majority of farmers in these countries. The cost per start their own profitable businesses. hectare of developing land for irrigation in Sub-Saharan ■ Private manufacturing. After developing the tools, private Africa ranges from $2,000 to $4,000 for small-scale irriga- manufacturers are contracted to produce the products in tion and from $9,000 to $15,000 for large-scale irrigation, a way that ensures high-quality standards and optimal making it out of reach for most families and communities pricing for customers. KickStart’s manufacturers are (Kidane, Maetz, and Dardel 2006). In India, the compara- located in China, Kenya, and Tanzania, with the bulk of ble cost for large-scale irrigation ranges from $1,500 to the products originating from China and, to a lesser $2,000 per hectare. extent, from Kenya. As of 2010, the manufacturer in In recent years, however, the development of appropri- Tanzania is capable of supplying only a portion of the ate and affordable irrigation technologies for poor Sub- Tanzania demand. Saharan Africa farmers and communities has been ■ Marketing and distribution. KickStart’s products are remarkable. One of these technologies is the human- introduced to the market through its own marketing and powered irrigation pumps developed by KickStart Inter- promotional activities, and the private sector supply national, which have been used by dynamic farmers and chain is used to distribute the products locally. KickStart entrepreneurs to establish and run profitable small-scale has more than 160 distributors in Kenya, more than agribusinesses. 150 in Tanzania, more than 80 in Mali, and a few in 320 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA Map 18.1 Irrigated Areas as a Percentage of Potentially Irrigable Areas IBRD 38554 MAY 2011 30° 0° 30° 60° Mediterranean TUNISIA Sea MOROCCO 30° Canary Is. 30° (Sp) ALGERIA ARAB LIBYA FORMER REP. OF SPANISH EGYPT SAHARA Red CAPE MAURITANIA Sea VERDE MALI NIGER ERITREA SENEGAL CHAD THE GAMBIA BURKINA SUDAN DJIBOUTI GUINEA-BISSAU GUINEA FASO BENIN NIGERIA ETHIOPIA GHANA TOGO SIERRA LEONE CÔTE CENTRAL LIBERIA D’IVOIRE AFRICAN REPUBLIC CAMEROON EQUATORIAL GUINEA SOMALIA UGANDA SÃO TOMÉ AND PRÍNCIPE KENYA 0° GABON CONGO DEMOCRATIC 0° ATLANTIC REPUBLIC RWANDA BURUNDI INDIAN OF CONGO OCEAN Cabinda TANZANIA OCEAN (ANGOLA) SEYCHELLES COMOROS WATER MANAGED AREAS AS % ANGOLA MALAWI ZAMBIA Mayotte OF IRRIGATION POTENTIALS: (Fr) 76% - 100% ZIMBABWE MOZAMBIQUE MADAGASCAR 51% - 75% NAMIBIA BOTSWANA MAURITIUS 26% - 50% 10% - 25% SWAZILAND < 10% SOUTH LESOTHO 30° 30° AFRICA INTERNATIONAL BOUNDARIES 0 500 1,000 1,500 KILOMETERS Source: Aquastat 2005 0 500 1,000 MILES 30° 0° 30° 60° Source: FAO 2005. Burkina Faso. In countries in which KickStart does not assessments is used to make improvements to the tech- have a marketing program, distributors have been hired. nology and to generate ideas for future innovation. KickStart also liaises with local and international non- governmental organizations (NGOs) and with United Very often, KickStart is asked whether it would be better Nations agencies to serve those organizations’ beneficiar- or more effective to simply give the pumps away. In both ies and programs. cases, the answer cases is no. The KickStart model of selling ■ Impact assessment. KickStart conducts socioeconomic pumps through a profitable supply chain has been shown impact assessments on the incremental benefits resulting to be both profitable and sustainable in many African from its products. Information gleaned from these countries, creating employment opportunities and new CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 321 sources of income. Analysis has shown that the KickStart other than Kenya and Tanzania, the revenue data show an model increases the livelihood of farmers and their families increase of 440 percent for the MoneyMaker pump. more cost effectively than giving the pumps away would. Annex 18.1 provides further details on this topic. Impacts of pumps at the manufacturer and KickStart uses a tipping-point concept to mark the retailer levels beginning of sustainable technology infusion. The tipping point is where the volume of sales of the technology dra- Profits have been generated by the sales and manufactur- matically rises or increases, decreasing the marketing ing of the MoneyMaker range of pumps at the manufac- costs to a minimum, as illustrated in figure 18.1. KickStart turer and retailer levels, as shown in table 18.2. Skilled estimates that the tipping point is achieved when 15 to 20 percent of the market potential in terms of sales is Figure 18.1 The Tipping Point Concept achieved. In Kenya, for example, KickStart aims to surpass the tipping point by 2014. Impacts of MoneyMaker pumps at the farm level Sales The MoneyMaker pump technology has had a significant impact on revenue generation and wages at the farm, dis- Money tributor, and manufacturer levels. Table 18.1 summarizes Production and Sustainable survey data of the income and wage impacts of KickStart’s marketing costs business different types of pumps at the farm level. As shown in the table, household income increased by 191–200 percent in Kenya for all categories of pumps except for the Money- Tipping point Maker Hand Pump, which showed no change. In Tanzania, income doubled for households that used the MoneyMaker Time Hip Pump and tripled for households that used the Super MoneyMaker and MoneyMaker Plus pumps. For countries Source: Authors. Table 18.1 Farm-Level Household Revenues and Wages per Pump, July 1991–November 2009 (dollars) Annual income Annual Percent Wages over without income increase Annual lifespan of Pump pump with pump in income wages pumpa Kenya MoneyMaker 628 1,885 200 126 379 Super MoneyMaker and Super MoneyMaker Plus 1,148 3,443 200 71 214 MoneyMaker Plus 248 744 200 21 63 MoneyMaker Hand 93 279 200 4 12 MoneyMaker Hip 832 2,423 191 60 174 Tanzania Super MoneyMaker and MoneyMaker Plus 493 1,478 200 41 123 MoneyMaker Hip 826 1,651 100 3 6 Other countriesb MoneyMaker pump 183 995 444 18 54 Super MoneyMaker and Super MoneyMaker Plus 345 1,034 200 29 86 MoneyMaker Hand 38 — — 2 6 MoneyMaker Hip 578 1,156 100 3 9 Source: KickStart International. a. Wages over lifespan of pump = annual wages x lifespan of pump. b. Burundi, Democratic Republic of Congo, Malawi, Mali/Burkina Faso, Mozambique, Rwanda, Sudan, Uganda and Zambia. — = Data not available. 322 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA Table 18.2 Manufacturer and Retailer Profits and Manufacturer Skilled Jobs, July 1991 to November 2009 Manufacturing Retailing Technology Pretax profit ($) Skilled man days Pretax profit ($) Kenya MoneyMaker 310,694 1,738 20,918 Super MoneyMaker 61,378 10,391 155,882 Super MoneyMaker Plus 75,604 12,799 215,000 MoneyMaker Plus 33,999 3,837 58,284 MoneyMaker Hand Pump 27,712 1,564 6,455 MoneyMaker Hip Pump 33,197 2,997 37,844 Total 542,584 33,325 494,382 Tanzania MoneyMaker 26,874 219 2,239 Super MoneyMaker 38,248 2,878 35,306 Super MoneyMaker Plus 269,024 20,241 292,034 MoneyMaker Plus 985 132 985 Money maker Hand Pump 925 108 1,013 MoneyMaker Hip Pump a a 17,332 Total 311,870 23,578 348,910 Source: KickStart. a. Not manufactured in the country. manpower jobs have also been created in the manufactur- and manufacturing must ensure that retail prices are ing of the pumps. Between July 1991 and November 2009, affordable, ideally less than $150. retailers’ profits from MoneyMaker pumps totalled ■ Energy efficiency. All pumps are human powered and $494,382 in Kenya and $348,910 in Tanzania. At the man- thus must be extremely efficient at converting human ufacturer level, profits over the same period amounted to power to mechanical power $542,584 in Kenya and $311,870 in Tanzania. The manu- ■ Ergonomics and safety. The pumps must be safe to use for facturing of the pumps involved more than 33,000 and long periods of time without stress or injury. nearly 24,000 skilled man days in Kenya and Tanzania, ■ Portability. Pumps must be small and light enough to respectively. carry home from the point of purchase on foot, by bike, or by minibus. MONEYMAKER PUMPS PRODUCT LINE ■ Ease of installation and use. All pumps must be easy to set up and use, without additional training or tools, not even Criteria for pump creation a hammer or screwdriver. KickStart is guided by a unique set of criteria in designing ■ Strength and durability. The pumps are designed and its products: income generation, return on investment, built to withstand heavy use. The pumps carry a one-year affordability, energy efficiency, portability, ease of use and guarantee upon purchase. installation, strength and durability, design for manufac- ■ Design for manufacturing. For manufacturing to be truly turing, and cultural acceptability. These criteria are effective, pumps must be produced in large quantities, defined as follows: but in the developing world manufacturing capacity is limited. The design of pumps must be such that the ■ Income generation. Every pump model developed and pumps can be manufactured within the capacity of the marketed must have a business model that clearly pre- local manufacturing industry. dicts profitability and can be supplied to entrepreneurs. ■ Cultural acceptability. The pumps must be adapted to the ■ Return on investment. A purchaser of a MoneyMaker culture in the countries where they are sold. pump should be able to fully recoup his or her invest- ment within six months. A team of engineers, designers, and technicians develop ■ Affordability. Because the target purchasers of the and test prototype pumps to ensure their performance, cul- pumps are some of the world’s poorest people, designs tural acceptability, and durability. In addition to design and CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 323 development, KickStart creates marketing awareness and Figure 18.2 Design Diagram and Field Operation of the assembles sales teams to sell the pumps. Importantly, Kick- Super MoneyMaker Pump Start employs innovative marketing techniques that are accessible to potential customers in remote villages without Super-MoneyMaker fossil fuels and electricity, including community pumping pressure irrigation pump competitions during market days and demonstrations at retail shops and individual farms. Operation Manual Types of pumps Original MoneyMaker Pump. KickStart’s debut pump, the original MoneyMaker, was introduced in September Plastic Bearing 1996. This small, treadle-operated pump could pull water Pressure from as deep as 7 meters and be used to furrow irrigate up hose pipe to 0.8 hectares of land. The pumps demonstrate the poten- Two rubber caps tial poverty reduction effects of micro-irrigation: more than Two rubber valve (inside each cyclinder) 4,050 original MoneyMaker pumps were sold between 1996 (inside each cyclinder) Outlet pipe and 1999, generating more than $3.9 million annually Rigid Inlet pipe function among user households in East Africa. User feedback, how- pipe Outlet pipe ever, indicated that farmers needed a pump that could lift Foot valve (below water) Rubber valve and push water through a hosepipe or sprinklers. In (inside) response, KickStart introduced the new Super MoneyMaker Plastic pipe pump in October 1998. The new suction and pressure Source: KickStart International. pump superseded the original MoneyMaker, which was taken off the market in February 1999. Super MoneyMaker Pump. The Super MoneyMaker pump is used to pump water from hand-dug wells, rivers, Figure 18.3 Design Diagram and Field Operation of streams, lakes, and ponds. A twin-cylinder pump, it is ideal MoneyMaker Plus Pump for sprinkler and drip irrigation, for filling overhead water tanks, and for use with nozzles and sprays attached to the end of the delivery hose. The pump can draw water from a depth of 7 meters and pump it 14 meters above the ground. It can be used to irrigate up to 0.8 hectares of land. The design and operation of the Super MoneyMaker pump is shown in figure 18.2. An improved, easier-to-use design of the Super MoneyMaker pump, Super Money- Maker Plus, was subsequently introduced. MoneyMaker plus pump. Responding to demand for a lower-cost pressure irrigation pump, KickStart designed and launched the MoneyMaker Plus pump in July 2001. This small, leg-operated pump has one piston and one cylinder but can still pull water from 7 meters deep and push it more than 14 meters above the ground. It can be used to irrigate up to 0.4 hectares of land. The design and operation of the MoneyMaker Plus pump is shown in figure 18.3. Source: KickStart International. 324 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA MoneyMaker Hip Pump. The MoneyMaker Hip Pump is (six months or more) with all feedback incorporated into a unique pump developed to meet the needs of the very the final design. The first batch of pumps was field tested poor people who cannot afford the twin-cylinder Super through the private sector supply chain that stocked the MoneyMaker pump or the single-cylinder MoneyMaker products and retailed them to the end users. Because Plus. The MoneyMaker Hip Pump debuted in stores in KickStart’s mission includes poverty reduction, an assess- 2006, however, sales and marketing efforts began in 2008. ment of the pumps’ impact is carried out at “zero age,” and The pump weighs only 4.5 kilograms, and can irrigate up to again in the 9th and 18th month after procurement. The zero 0.4 hectares of land. age survey provides the baseline data on which comparison The design of the MoneyMaker Hip Pump includes a is made for the subsequent 9th-month and 18th-month sur- super-efficient valve box and a simple pivot hinge. By veys. Manufacturing begins after the successful testing of attaching a hand pump to a hinged platform, the mechan- the pumps. Manufacturers are trained and provided with ics of the pump were changed to allow operators to use parts produced by KickStart to ensure quality standards for their legs, body weight, and momentum, rather than the all the pumps. The pumps are supplied through distributors few muscles of the upper back and shoulders. This makes and retailers for eventual sale to farmers. the pump more energy efficient, and it is capable of irri- KickStart embarks on intensive promotion and market- gating at least 0.4 hectares. Like the Super MoneyMaker, ing campaigns to create awareness about and generate the MoneyMaker Hip Pump can draw water from 7 meters demand for its products. Marketing and promotion of and pump up to 14 meters above the ground. high-quality products to poor people requires substantial These pumps retail for $35 to $100. Between July 1991 and resources. For KickStart, this is possible through donor November 1999, more than 137,000 MoneyMaker pumps of and government support. Until a critical mass is reached all types were sold, with a vast majority of them purchased by (15 to 20 percent of the market share), the tipping point is residents of African countries. Kenya, Tanzania, and Malawi not realized. However, once the tipping point is achieved, are the countries in which the greatest number of pumps operation costs are substantially reduced and the business have been sold, and the Super MoneyMaker Plus is by far the becomes sustainable and profitable and many more most commonly sold MoneyMaker pump (annex 18.2). Fig- distributors start selling the product. ure 18.4 details the number of pumps that have been sold MoneyMaker pumps are designed to last for three years across the continent. but some have, with good maintenance, operated for more MoneyMaker pumps are developed through design cri- than six years. Critical in this good maintenance is the teria consideration to meet the needs of the target market. replacement of piston cups every 3–12 months (depending Final prototypes are field tested for long periods of time on hours of usage and the quality of water used). Figure 18.4 Number of MoneyMaker Pumps in Various Countries, 1996–June 2009 50,000 Number of pumps sold 40,000 30,000 20,000 10,000 0 a ia i n o a a da i rs w nd ny nd bi da as an he ala an m ru Ke aF Su ga nz ot Rw Za M Bu U Ta in all rk l, Bu ta / To ali M Source: KickStart sales data 1996–2009. CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 325 Figure 18.5 MoneyMaker Pump Supply Chain Sales Product and Production Distribution Marketing & R&D backup promotion service Public sector Private sector Continuous impact monitoring by KickStart Source: KickStart International. maintains business accounts with each distributor or Table 18.3 Costs of Various Steps in the MoneyMaker Pump Supply Chain dealer. Most of KickStart distributors are Agrovets shops (sellers of agrochemicals and veterinary products) fre- Supply Chain Step Cost ($) quently visited by farmers. Initially, KickStart provided Manufacturer 65 KickStart 72 distributors with credit terms that required them to pay Wholesaler 100 KickStart once the consignment was sold. As demand for Retailer 115 KickStart products grew, the credit period was reduced to Source: KickStart International. 30 days after receipt of consignment. KickStart also requires that distributors be willing to demonstrate and MoneyMaker pumps supply chain market the pumps. KickStart identifies different marketing and promotion The MoneyMaker supply chain is a five-phase process strategies to create awareness among potential end users of including research and development (R&D), production, its products. These include above-the-line (those using distribution, marketing and promotion, and sales and mass media) and below-the-line (nonmedia) promotions backup services, as shown in figure 18.5. KickStart carries to close sales. In Kenya, Tanzania, Mali, and Burkina Faso, out R&D and marketing and promotion directly. The sales teams are trained to communicate the benefits of the other phases are executed by the private sector with mon- MoneyMaker pumps and promotion of farming as a busi- itoring by KickStart. Donor participation is very crucial to ness to potential customers. carrying out product R&D and creating product awareness because it enhances demand and consequently creates an attractive business opportunity for the private sector. IMPACT OF MONEYMAKER PUMPS Production, distribution and retailing costs, and profits ON DEVELOPMENT amount to $115 per pump for the Super Moneymaker Impact assessment methodology pump (table 18.3) and are borne by the farmers who pur- chase the pumps. Other costs amounting to $264 a pump, Through its monitoring teams, KickStart assesses key indi- incurred during technology development, promotion, sales, cators of the impacts of the MoneyMaker range of pumps. and impact monitoring, are currently borne by donors. A range of data is collected to carry out this monitoring All the design work behind KickStart’s products is car- process: the number of pumps manufactured and sold is ried out at Nairobi, while products are manufactured in logged, purchasers’ details are recorded in a database, and either Thika, Kenya; Arusha, Tanzania; or in China. Kick- training recipients’ details are recorded. Later, KickStart’s Start recruits and trains a network of distributors and monitoring staff visit a random selection of distributers, dealers (more than 150 dealers in Kenya alone) in cities, retailers, and purchasers to interview, administer question- towns, and small market centers. The pumps are sold, with naires, and gather statistical data for technology impact a markup of 7 percent to KickStart, to distributors in the analysis. These socioeconomic surveys are conducted at private sector through KickStart’s coordination, which three points in time: at “zero age” (before use of the 326 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA pumps) and during the 9th and 18th month of usage. The In Kenya, socioeconomic surveys have shown a substan- parameters monitored include pump ownership; pump tial increase in the area under irrigation at the household management; and impact of the pumps on incomes, food level following the purchase of a MoneyMaker pump. At security, livelihoods, and investment. the zero age survey an average of 0.02 hectares per house- hold was under irrigation (typically bucket irrigation) but Pump ownership and management rose to 0.2 hectares under hip pump irrigation, a nine-fold average increase. KickStart’s impact assessments (KickStart 2008) show that 88 percent of the people who purchase pumps are male. Management of the pumps, however, is not carried out Impact on households strictly by pump owners; rather, it changes with time. At Although rural farmers are risk averse and have very little zero age, 81 percent of the pumps are managed by the cash to spare, demand for MoneyMaker pumps has owners, a figure that fell to 46 percent at 18 months. On increased steadily over the years, a trend that may arise from the basis of gender, the surveys show that at zero age, the change in lifestyles after purchasing the MoneyMaker 23 percent of pumps are managed by females, a figure that pumps, as indicated by the profiles of four typical “farmer- increased to 60 percent at 18 months. This is not surpris- preneurs” shown in table 18.5. ing since, as time passes, women’s interest in the pump According to KickStart survey data, incomes in house- grows because it provides a means of increasing income holds using MoneyMaker pumps have risen from as little as for families’ daily needs. Men, on the other hand, look at $100 to more than $10,000 annually, in addition to extend- other long-term high-capital investments that subse- ing food availability periods from 3 months to 12 months, quently generate higher returns. For example, farmer thus eliminating food insecurity. The change in livelihood of Samuel Ndungu (one of the four profiled below) opted to the irrigating households is manifested in investment of the leave his treadle pump to his wife while expanding his irri- extra income in purchase of assets such as bicycles, motor- gated enterprise with a motorized pump. This change in ized pumps, dairy cows, modern houses, and land. Some management shows how women have been empowered to irrigating households have also expanded their enterprises improve their livelihoods with MoneyMaker pumps. to include dairy, poultry, small-scale maize milling, and transport. With the increased incomes, payment of school Impact on irrigated area fees for their children has eased, and the number of people The area under irrigation using the pumps sold in Sub- employed on such farms has also increased. In addition, the Saharan Africa between 1991 and 2009 is estimated to be MoneyMaker technology has encouraged the introduction more than 31,000 hectares (table 18.4). of high-value crops, with their attendant improved agro- nomic practices. And the technology has been frequently adopted by neighboring farms, whose owners and workers Table 18.4 Area under Irrigation Using the Pumps, have received on-farm training nearby. and Crops Grown in those Areas in Various Sub-Saharan Africa Countries, Job Creation. Investment in agricultural water manage- 1991–2009 ment through small-scale irrigation creates jobs both for Area being the families who own the pumps and the labor they hire, Country irrigateda Crops because irrigation increases farm output and requires Kenya 12,326 Horticultural crops Tanzania 8,785 Horticultural crops more workers, especially where high-value crops are Malawi 4,665 Maize, horticultural crops grown. Statistical analysis by KickStart (1991–2009), for Sudan 1,531 Horticultural crops some Sub-Saharan countries, has shown that, on average, Mali/Burkina Faso 1,433 Horticultural crops Uganda 896 Horticultural crops when families started bucket irrigation, 0.17 jobs were cre- Zambia 679 horticultural crops ated; after the acquisition of a MoneyMaker pump, the Rwanda 540 Horticultural crops number of jobs created rose to 0.55 per household.1 Of Burundi 284 Horticultural crops Total 31,138 that 0.55, family members represented 0.51 and waged labor represented 0.04, while distribution on gender basis Source: Generated with KickStart sales data 1991–2009. a. Area irrigated is generated using a conversion figure of 0.3 was 0.32 and 0.22 for men and women, respectively. The hectares per pump at 80 percent pump utilization. jobs created included pump operation, land preparation, CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 327 328 Table 18.5 Farmerpreneur Profiles before and after Investing in MoneyMaker Pumps Name Samuel Mburu Ndungu Felix Muiruri Catherine Gwambie Mahamoud Guindo Country Kenya Kenya Tanzania Mali Age 45 34 40 48 Gender Male Male Female Male Before After Before After Before After Before After Farm enterprise Bucket irrigation Super MoneyMaker Bucket irrigation MoneyMaker Hip Rainfed and bucket Super MoneyMaker Bucket irrigation Super and petrol pump Pump irrigation MoneyMaker Income $100 $1,000–$2,000 $480 $10,440 — — $400 $700 ($/year) Food availability 3 12 3 12 3 12 3 12 (months) Assets (house, 1 house, 1 cow, 2 houses, 2 cows, — — Chickens 1 modern house, — — bicycle, farm 1 goat 7 goats, 3 bicycles, 2 businesses, animals, etc.) one motorized chickens pump Land size (number 0.6 1.0 (0.4 being 0 0 0.8 1.2 0 0 of hectares) purchased as of time of survey) Land irrigated 0.4 MoneyMaker: 0.2 (rented) 0.8 (rented) 0.8 1.2 150 square meters 300 square (hectares) 0.8 hectares meters (0.4 rented); motorized pump: 2 hectares Family size 8 5 5 6 Number of Workforce has Workforce has 0 2 None None employees grown fivefold doubled on the farm Crops French beans French beans, None French beans, Maize, beans Vegetables Vegetables Vegetables and tomatoes tomatoes, baby fruits corn, green maize Ease of Difficult Easy Difficult Easy Difficult Easy Difficult Easy facilitating children’s education Source: KickStart International. — = Unavailable. weeding, and harvesting. Pump operation represented the countries. Financial resources are particularly limited in Mali largest proportion of the jobs created, with 33 percent of and Burkina Faso, leading to low awareness levels and hence the pumps being lent out to neighbors, generating 0.03 low levels of impact compared with Kenya and Tanzania. jobs at no cost. The Kickstart analysis also shows that Despite the financial constraints, KickStart estimates that MoneyMaker pumps increased the number of jobs created between 1991 and 2009, more than 87,000 small-scale agri- by more than 220 percent. In most countries, where many cultural enterprises were created using the MoneyMaker youths are not currently engaged in productive activities, range of pumps. These businesses generated new profits irrigation provides an opportunity for a net increase in the totalling over $77.2 million annually and employing more jobs created. than 100,000 people. Each pump is used to irrigate 0.3 hectares of land (on average), and generates an average net Income. In addition to job creation, MoneyMaker pumps annual income of $1,100 a year for its owner, representing a have a positive impact on farm incomes generated through 100–450 percent increase in the incomes that poor rural irrigation, as shown in table 18.2. people made from sale of crops before the adopting the MoneyMaker technology. In Kenya by 2009 more than Poverty. Data (1991–2009) from Kenya, Tanzania, Mali, 12,000 hectares of land were estimated to be irrigated using Burkina Faso, and other countries show that 439,839 people KickStart irrigation pumps. Irrigation investment costs less have been moved out of poverty (table 18.6). The poverty than $60 per hectare annually compared with the conven- reduction to pump ratio (number of people moved from tional sprinkler systems, which cost more than $600 per poverty per pump unit) varies but is clearly higher in Kenya, hectare annually. In short, MoneyMaker pumps remove the Tanzania, Mali, and Burkina Faso than in other countries in obstacle of huge initial capital investment for demand- which the pumps are sold. This implies a higher efficiency of driven, small-scale irrigation development. pump utilization in the above four countries and is com- mensurate with the level of technology promotion and Kenya’s experience with MoneyMaker pumps demonstration of on-farm profitability. These last two factors are essential ingredients in high pump utilization Thousands of entrepreneurial farmers in Kenya are now efficiency and need to be replicated in other Sub-Saharan irrigating with KickStart’s range of MoneyMaker pumps, Africa countries to achieve similar success levels in those changing their small subsistence farms into vibrant com- locations. mercial enterprises. The average size of plots held by small- holder farmers is between 0.2 and 1.0 hectare and generates about 25–80 percent of total family income annually IMPACT ON SUB-SAHARAN AFRICA (IPTRID 2005). When they use irrigation, farmers are typi- The impact of a technology depends on its adoption rate, cally able to grow and sell three to four high-value crops which is largely determined by the awareness created. Mar- annually. These “farmerpreneurs” in Kenya over 2007–09, keting of the MoneyMaker technology requires a huge have increased their incomes ten-fold on average, and some financial investment that is not possible for many African have made annual profits of up to $5,400. Table 18.6 Number of People Using Kickstart Pumps Who Have Moved Out of Poverty in Select Sub-Saharan Africa Countries, 1991–2009 Enterprises created People moved Poverty reduction Country Pumps sold or transformed out of poverty to pump ratio Kenya 51,357 40,720 203,600 3.96 Tanzania 36,603 29,495 147,475 4.03 Mali and Burkina Faso 5,970 4,673 23,365 3.91 Other countriesa 43,676 13,079 65,399 1.50 Total 137,606 87,967 439,839 3.20 Source: KickStart 2008. Note: Here, poverty is defined as an income of less than $1 a day. a. Burundi, Democratic Republic of Congo, Malawi, Mozambique, Rwanda, Sudan, Uganda, and Zambia. CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 329 Distribution and Utilization. The distribution of the to reach. Table 18.7 shows the income generated by each MoneyMaker range of pumps in various regions of Kenya is type of MoneyMaker pump. shown in figure 18.6. This distribution reflects the agricul- tural potential in the region, the ability of farmerpreneurs to Income and Employment Creation. In most parts of purchase the pumps, and the availability of surface and rural Kenya, as elsewhere in Sub-Saharan Africa, time is shallow groundwater sources. an available resource that is very often underutilized because opportunities for productive employment activi- Wealth Creation. According to KickStart data, the aver- ties are lacking. KickStart’s ranges of MoneyMaker pumps age net annual income per family using the MoneyMaker are designed to create opportunities for rural people to pumps in Kenya ranges from $800 to $10,440. As of more fully use their available time on productive activi- December 2005 about 33,000 pumps had been purchased ties. In Kenya, the average annual household income rose by farmers and were generating household income from $3,465 (K Sh 242,529) before the pump to $4,700 totalling $31.4 million (K Sh 2.2 billion) annually. Impact (K Sh 328,973) after acquisition of the pump (KickStart monitoring survey data showed that 91 percent of the pur- sales data 1991–2009). Although total household income chased pumps generated total income of $93.8 million (K rose from several sources, including farming, employ- Sh 6.7 billion) during their life span of three years. The ment, pension, remittances, businesses, and collection of other 9 percent are pumps that were not used for various rent, income from irrigation contributed the highest pro- reasons such as water source being too deep for the pump portion of the total, at 47 percent. KickStart’s monitoring Figure 18.6 Regional Distribution of MoneyMaker Pumps Sold in Kenya, 2007–09 2,500 2,000 Pumps sold 1,500 1,000 500 0 al st n i n a y rn ob nz lle er er tr oa te air ya Va en st st C es Ea Ea N N C W ft Ri th or N 2007 2008 2009 Source: KickStart. Table 18.7 Wealth Creation with MoneyMaker Pumps in Kenya Pumps purchased Annual income Cumulative income over Irrigation technology 1996 through 2005 ($ millions) three years ($ millions) MoneyMaker Suction Pump 3,305 2.31 6.93 Super MoneyMaker 13,657 17.44 52.31 Super MoneyMaker plus 7,401 9.45 28.35 MoneyMaker Plus 6,639 2.00 6.00 Hand Pump 2,221 0.23 0.23 Total 33,223 31.27 93.82 Source: IPTRID 2005. 330 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA surveys (2008) also show that Super MoneyMaker pump after they started MoneyMaker pumps. In 2008, the irrigation over an 18-month period on average increased incidence of “improved” houses—those constructed from pumping and watering jobs by 465 percent in Kenyan stone and timber—increased by 38 percent and 10 percent, farms. This is in addition to increased jobs in other crop respectively, demonstrating investment in better housing production activities. Based on the number of pumps and therefore improved livelihoods. sold, the area irrigated and hence the number of jobs created, Kenya has made the largest gains of all the coun- Food Security. Food security exists when people have tries in which MoneyMaker pumps have been sold. physical and economic access to food that is sufficient and nutritious enough to meet their dietary needs and Assets, Housing, and Investment. KickStart surveys preferences for an active and healthy life. Food security show that in Kenyan households using MoneyMaker can be achieved through producing one’s own food (self- pumps, the number of local cattle owned declined by sufficiency), having sufficient income to purchase food, or 31 percent, whereas the number of the more expensive a combination of both. KickStart surveys (2008) show dairy cows increased by 7 percent, indicating availability of that households using the Super MoneyMaker pump surplus resources available to purchase them. Dairy cows increased their food surplus by 35 percent in the produce more milk than the local cattle, thus increasing 18 months following the adoption of the technology. family income. There was also more fodder available to feed the dairy cows, which demand more fodder than local DRIVERS OF SUCCESS cows. Similarly, there was an increase in the number of donkeys and carts in households using MoneyMaker MoneyMaker success stories are numerous. Three such sto- pumps, indicating the demand for extra transport with the ries from Kenya, Tanzania, and Burkina Faso are presented use of MoneyMaker technology. In addition, households in box 18.1. increased their investments in businesses (shops, transport, In developing its irrigation pumps, it is KickStart’s pol- poultry, and dairy) by 32 percent. KickStart surveys also icy to ensure four main outcomes: achieving the highest show improvement in housing among households in Kenya impact in the shortest time possible; achieving a Box 18.1 Money Maker Success Stories Daniel Karanja Njenga and Nancy Gathoni are also helping hungry neighbors and friends struggling in Kenya to get back on their feet after the violence. Daniel has plans to expand his plot to grow cabbages and tomatoes While living in an internally displaced person camp and purchase a dairy cow. after their home and farm were looted during Kenya’s Daniel and Nancy are two of many other farmers in postelection violence in March 2008, Daniel Karanja Kenya who are becoming successful businessmen and Njenga and his wife Nancy Gathoni participated in a women able to feed their families, pay school fees, and promotional demonstration of the MoneyMaker Hip medical expenses. According to KickStart’s research, an Pump. “When I saw the . . . demonstration and heard average farmer can make about $120 a month selling about it on the radio,” Daniel said, “I knew it was the crops produced using the MoneyMaker pump. answer to how we could earn an income quickly and get In Kenya MoneyMaker pumps are also providing back to farming.” hope to people facing a tough combination of challenges Daniel’s first purchase with the relief funds provided in recent years: postelection violence, escalating food to him by the Kenyan government ($130) was the Mon- prices, a difficult economic situation, and continued eyMaker Hip Pump and hoses. The manually operated high levels of poverty. pump is lightweight, costs less than other pumps, and is easy to use. It does not require electricity or fuel. With it, Catherine Gwambie in Tanzania Daniel was able to irrigate his small 1/8th-acre plot and grow crops during the dry season when most farms are Catherine Gwambie and her husband, Hawzi Mwami, bare. Daniel and Nancy now earn a decent living selling are an entrepreneurial couple from Tanzania who had their sukuma wiki (kale), a staple food for Kenyans. They dreams of becoming successful shop owners in Dar es CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 331 Box 18.1 (continued) Salaam. They farmed in their native Kigoma, growing Hawzi now freely admits that his wife was right about and selling maize and beans to save enough to open a the pump, and between their two businesses, they see a shop selling household supplies. bright future for their family. Although the shop was reasonably successful, it did Mahmoud Guindo in Mali not generate as much income as the couple needed to support their family. Hawzi decided to buy land on Mahmoud Guindo, a 48-year-old farmer with a wife and which he could raise chickens and Catherine could four children, had long struggled to make ends meet on start growing vegetables for sale. It was a good busi- credit. He moved from his home in Dogon County to ness, but bucket irrigation on the land took a lot of Bamako, where he was employed as a security guard effort. earning $400 annually. In an effort to boost his income, In early 2007 Catherine heard an advertisement for he began farming a 150-square-meter plot, which was the Super MoneyMaker pump on the radio. She excit- still inadequate to meet his family’s needs. edly told her husband about this new pump that was To increase his annual income, Mahmoud wanted to affordable and made irrigation easier and quicker. build a bigger garden, but he was sceptical about how he Hawzi was not convinced. Catherine, however, insisted would water a larger plot of land. After seeing an adver- that the pump would make her life easier and decided tisement for KickStart’s MoneyMaker pump on television to use her own money to purchase it. Together, the in 2008, Mahmoud wanted to buy it. He did not have Mwamis went to the Kariakoo market in Dar es Salaam enough money, however, so he approached his boss for a to buy a Super MoneyMaker. loan. Both men agreed it was a tangible asset that would The pump increased Catherine’s productivity so provide a profitable and quick return on investment. much that she expanded to another plot and now Since buying the pump in October 2008, Mahmoud employs her daughter and young sister. With the extra has almost doubled his annual income, from $400 to income they are now earning, the Mwamis have plans to $700, by selling fruits and vegetables. The additional cash send three of their young children to good secondary flow is allowing him to pay off some debts while simul- schools and to build a nicer house for their family. taneously providing sufficient food for his family. Source: Authors. cost-effective program; ensuring that the income-generat- Designing New Technologies and Business Packages. ing process is self-sustainable, and ensuring that the KickStart designs and develops the tools, equipment, man- process if scalable and replicable in other areas. uals, and business plans required for establishing small KickStart employs six approaches to achieve its devel- enterprises. It also designs the production protocols and opment goals, as described below. In addition, KickStart’s quality control procedures required for manufacturing success has been driven by the desire of farmer to produce new pumps. on-farm profits and get out of poverty through adoption of appropriate and affordable technology Training Manufacturers to Produce MoneyMaker Pumps. KickStart trains private manufacturers to set up Market Research. KickStart’s technological gap studies assembly lines and quality control mechanisms to mass- identify profitable small enterprises that can be estab- produce its machines and tools. lished by local entrepreneurs with limited capital invest- ment and determine the technological and socioeconomic Promotion of MoneyMaker Pumps. KickStart promotes requirements of those opportunities. The studies evaluate the MoneyMaker and works with the private sector to ensure raw materials, competing products, potential market that pumps are well known and easily available to small-scale demands, and constraints and opportunities for small investors, reducing product risk for entrepreneurs. The risk enterprises. to entrepreneurs is further reduced by the fact that the 332 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA pumps are guaranteed for a period of one year following If each family in Kenya with access to water from purchase. Farmers have the right return their pumps if their underground storage, rivers, streams, ponds, or dams used wells are too deep for the operation of the pump or if the a treadle pump and irrigated 0.3 hectares of land, for pumps break down due any manufacturing fault. example, that country would not be experiencing the food shortages and hunger it has recently. Assuming water On-farm Profitability. The enhanced water utilization accessibility, an irrigation potential of 540,000 hectares, that has come with the MoneyMaker range of pumps has and a 20 percent share of this irrigation potential for the increased farm incomes, a major incentive for practicing MoneyMaker technology, Kenya would need about and potential farmers. 360,000 of the MoneyMaker range of pumps. Based on KickStart’s sales figures, as of 2005, only 9.2 percent of the Farmers’ Need to Get Out of Poverty. Despite being potential number of pumps had been sold, and 14.2 per- hardworking and dedicated, farmers in Sub-Saharan Africa cent by October 2009 (KickStart sales data 1991–2009). have long been seeking a means to escape poverty. The Additionally, KickStart aims to increase access to 20 percent MoneyMaker technology has accorded them an affordable of the potential irrigation market in Kenya, after which the opportunity to do just that. technology will be self-propagating. After the tipping point is reached and assuming 80 percent pump utilization effi- Partnership with Donors. KickStart raises funds for ciency, 17,280 hectares would be under irrigation through research and development, marketing, and awareness pro- the MoneyMaker range of pumps. motion of MoneyMaker technologies and associated busi- In Sub-Saharan Africa, more than 87,967 small agricul- ness plans. KickStart’s donors include a large number of tural enterprises had been created by 2009 using Money- public and private donors such as the Bill and Melinda Maker pumps. These businesses generate $81 million each Gates Foundation, the Rockefeller Foundation, and the year and offer 93,462 new jobs annually. This new wealth David and Lucille Packard Foundation. has helped move 439,839 people out of poverty. A further indication of the benefits of irrigation is the establishment of small-scale businesses within the vicinity of nearby vil- SCALABILITY AND TRANSFERABILITY lage markets, thus creating indirect employment. This has OF THE SUCCESS led to improved access to nutrition, education, health, hous- The scalability and transferability of KickStart depends cru- ing, and welfare services for farming communities. cially on the existence of three things: a sustainable supply KickStart’s systematic, replicable method of measuring chain involving manufacturers, distributors, and retailers; the impacts of its products is an aspect of the organization’s use of appropriate marketing technologies until the tipping success that could be replicated by other organizations. point is reached; thereafter, the technology is self-propagating; Every product comes with a one-year guarantee, and every and policy changes to encourage entrepreneurship and buyer fills out a guarantee form when the product is pur- removal of handouts of technologies. chased. The guarantee reduces the perceived risk of buying Maximization of the use of the technology is essential. For the product, and the forms give KickStart a database of all example, the Super MoneyMaker is currently used to irrigate pump owners. an average 0.3 hectares, despite being designed to irrigate Scaling up the MoneyMaker technology success in Sub- 0.8 hectares. This limitation is attributed to the limited land Saharan Africa should involve technology promotion, available to the farmer. Participatory usage, in which farmers demonstration of on-farm profitability and a sustainable share a single pump, can also increase the use of the pumps supply chain, and high levels of efficient pump use. The and the amount of land area being irrigated. Additionally, still-low pump utilization needs to be addressed through expansion of the pump’s utility from sprinkler irrigation to on-farm pump utilization research that looks at, for exam- include drip irrigation would enhance the water and energy ple, shifting from direct irrigation pumping to indirect use efficiencies and therefore increase the land coverage. In through storage tanks, or changing the application Naro Moru, Kenya, for example, MoneyMaker pumps are method from sprinkler to drip. All these require enhanced used to pump water from shallow wells into raised storage budget levels similar to or greater than the current promo- tanks that feed into drip irrigation systems. This arrange- tional budgets in Kenya and other African countries. In ment has the effect of increasing the area under irrigation. addition to the aforementioned interventions, government CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 333 policies should shift to favor accessible small-scale irriga- Technology infusion tion technologies. The participatory approaches to technology infusion employed in the KickStart model involve marketing through demonstration and competitions. Marketing of LESSONS LEARNED FROM technology is important in creating and enhancing adop- MONEYMAKER PUMPS tion of technology. Creation of organized marketing Prospects groups and structures is important in poverty reduction. In Sub-Saharan Africa, intensification, rather than expan- Cultural and gender sensitivity in technology develop- sion, of cropped areas, is key to agricultural growth and ment and infusion is an important consideration in tech- poverty reduction. One way of achieving this is introduc- nology. Technology evolution changes are driven by users tion of simple innovative technologies such as Money- of the technology. For example, in response to user Maker pumps for agricultural water management. Through demands, the Super MoneyMaker pressure pump, which irrigation, farmers can diversify into high-value horticul- facilitated sprinkler irrigation, was developed to replace tural crops and fodder crops. the original suction-only pump and was quickly adopted The concept of farmer entrepreneurs, in which agricul- by farmers. tural enterprises are run as viable businesses, is now fully inte- grated as policy in Kenya and needs to be introduced in other Challenges African countries. As KickStart’s products have shown, when poor people have access to technology to generate wealth, Despite its efforts and success, KickStart has not been able they use that technology quite effectively to move themselves to design the ideal pump that combines high performance out of poverty. Technology acceptance by farmers and a sus- without maintenance irrespective of the intensity of use, tainable supply chain is assured through the KickStart model. terrain in which it is operated, and water quality. A study (IPTRID 2006) on the experience with treadle The low pump utilization efficiency, irrigating on pumps, introduced in West Africa by an international NGO, 0.3 hectares of the possible 0.8 hectares, could be attributed Enterprise Works, found the need for improved after-sales to direct irrigation pumping, which causes fluctuating service and a coordinating agency with a longer-term pro- sprays. It may be possible, however, to improve pump uti- gram, such as the sort KickStart provides in East Africa. The lization efficiency through indirect irrigation pumping study recommended the adoption of the KickStart model to through storage tanks or using the pumps in combination ensure the sustainability of the West African program. with more efficient water application methods. 334 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA Annex 18.1 Economic Analysis of Pumps Given Away versus Sold Giveaway KickStart Item model selling model Remarks Donor funds $2 million $2 million Start with the same amount of funds. Costs per pump $290 $257 Each pump is $33 cheaper to sell than to give away. Manufacturing 65 65 Why? It costs $65 to manufacture each pump; they are then sold to a wholesaler at Revenue 0 $–72 $72, giving revenue of $7 per unit. This earned income helps support the Distribution $85 0 organization. Promotion and sales 0 $124 Costs of administration, fundraising, technology development, and impact monitoring Admin, technology $140 $140 are the same in either model. The giveaway would not require marketing and development, promotion but would incur costs to distribute the pumps. Staff would be needed impact monitoring, to coordinate, and vehicles and fuel would be needed to transport the pumps. fundraising KickStart would also need to provide hoses or the pumps would be useless. In the KickStart model, distribution costs are handled by the wholesaler. Total units 6,897 7,782 Donor funds divided by cost per pump. The KickStart model puts nearly 900 more distributed pumps in the field. Percent used in 30% 80% KickStart’s goal is not to distribute pumps. Rather, it is to help people create small enterprise enterprises. It is here where the difference in the two approaches becomes Number used in 2,069 6,226 apparent. When a person makes an investment, he or she is committed to making enterprise a better future. Impact monitoring data indicates that 80 percent or more of pumps bought are used to create jobs and income. The same research shows less than 30 percent of pumps given away are used to create a business. The KickStart model creates three times more small businesses than the giveaway model. Average new profits $1,100 $1,100 These businesses will generate an average of $1,100 in new profits and wages each and wages year. Net present worth $9,103,600 $27,394,400 Here is an estimate of the new profits and wages these new businesses will create over 4 years in four years. The KickStart model leverages $2 million in donor funds into more than $27 million in profits and wages for these families—three times more than the giveaway model. This is money that is spent locally, supporting other business and stimulating the local economy. Number of people 10,345 31,130 KickStart estimates that each enterprise supports a family of five. The KickStart moved out of poverty Model moves more than 20,000 more people out of poverty, and it does it at Cost to move one $193 $64 one-third the cost per person of the giveaway model. person out of poverty Source: KickStart International. Annex 18.2 Number and Type of Pumps Sold in Various Countries, July 1991–November 2009 Mali/ Burkina Other Pump type Kenya Tanzania Malawi Sudan Faso Uganda Zambia countries MoneyMaker 3,305 438 0 0 0 212 0 95 Super MoneyMaker 13,683 3,837 0 80 0 539 1 857 Super MoneyMaker Plus 16,985 26,988 18,503 6,138 5,609 2,885 2,368 7,709 MoneyMaker Plus 7,674 264 0 141 0 0 0 30 MoneyMaker Hand Pump 2,217 431 30 0 0 0 0 50 MoneyMaker Hip Pump 7,493 4,645 905 120 361 97 462 2,454 Total 51,357 36,603 19,438 6,379 5,970 3,733 2,831 11,295 Grand total 137,606 Source: KickStart International. CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA 335 NOTE Morris, M., H. Binswanger-Mkhize, and D. Byerlee. 2009. Awakening Africa’s Sleeping Giant: Prospects for Commer- 1. The analysis defines a job as an economic activity that cial Agriculture in the Guinea Savannah Zone and occupies an individual for five hours daily for 150 days Beyond. Washington, DC: World Bank. annually. Msangi, S., and M. Rosegrant. 2005. “World Agriculture in a Dynamically Changing Environment: IFPRI’s Long-Term BIBLIOGRAPHY Outlook for Food and Agriculture under Additional Aquastat. http://www.fao.org/nr/water/aquastat/main/index Demand and Constraints.” Paper written for the FAO and .stm. United Nations Economic and Social Development Arbache, J., and J. Page. 2008. “Hunting for Leopards: Long- Department Expert Meeting on “How to Feed the World Run Country Income Dynamics in Africa.” Policy in 2050.” International Food Policy Research Institute, Research Paper 4715, World Bank, Washington, DC. Washington, DC. Brookings Institution. 2007. “Panel 5: Africa’s Economic Suc- NEPAD (New Partnership for Africa’s Development) and cesses: What’s Worked and What’s Next.” Brookings Blum FAO (Food and Agriculture Organization). 2004. Roundtable 2007. Brookings Institution, Washington, DC. “Government of the Republic of Mozambique: National FAO (Food and Agriculture Organization). 2005. “Irriga- Medium Term Investment Programme.” ftp://ftp.fao.org/ tion in Africa in Figures: Aquastat Survey 2005.” docrep/fao/007/ae415e/ae415e00.pdf. http://www.fao.org/nr/water/aquastat/regions/africa/ Rosegrant, M., and N. Perez. 1997. “Water Resources index.stm. Development in Africa: A Review and Synthesis of IFAD (International Fund for Agricultural Development). Issues, Potentials, and Strategies for the Future.” EPTD 2005. “Agriculture Development in Republic of Discussion Paper 28, International Food Policy Mozambique.” Agriculture Support Programme For- Research Institute, Washington, DC. mulation Report, Working Paper 2, International Fund Tyler, G. 2009. “All-Africa Review of Experiences with for Agricultural Development, Rome. Commercial Agriculture: The African Sugar Industry—A IPTRID (International Programme for Technology and Frustrated Success Story.” Background Paper for the Research in Irrigation and Drainage). 2005. “Kenya Competitive Commercial Agriculture in Sub-Saharan Impact Case Study Report.” IPTRID, Rome. Africa (CCAA) Study. World Bank and FAO, Washington, DC. http://siteresources.worldbank.org/INTAFRICA/ IPTRID. 2006. “Treadle Pump Dissemination and Adoption Resources/257994-1215457178567/Ch6_Sugar.pdf. in West Africa: Performance Problems and Prospects.” GRID Network 24. Rome. UNIDO (United Nations Industrial Development Organi- zation). 2009. Industrial Development Report 2009: KickStart. 2008. “The Super MoneyMaker Pump: The Breaking In and Moving Up: New Industrial Challenges for 18-Months Impact Assessment Report.” the Bottom Billion and the Middle-Income Countries. Kidane W., M. Maetz and P. Dardel. 2006. “Food Security Vienna: UNIDO. and Agricultural Development in Sub-Saharan Africa: World Bank. 2000. Can Africa Claim the 21st Century? Building a Case for More Public Support.” Main Report. Washington, DC: World Bank. FAO Regional Office for Africa, Harare. 336 CHAPTER 18: MONEYMAKER PUMPS: CREATING WEALTH IN SUB-SAHARAN AFRICA PA R T V Engaging the Private Sector to Upgrade Infrastructure CHAPTER 19 ICT in Sub-Saharan Africa: Success Stories Kaoru Kimura, Duncan Wambogo Omole, and Mark Williams s recently as 1998 few people in Sub-Saharan A Voice networks Africa had access to a telephone, and even fewer Voice networks are expanding rapidly, and the number of had access to computers. Since then, this situation people using the networks is growing. Between 1998 and has changed out of all recognition. Africa has experienced a 2008 the number of mobile cellular subscribers in Sub- continent-wide revolution in the growth of the information Saharan Africa leaped from about 4 million to 259 million and communication technology (ICT) sector. Networks (figure 19.1) have expanded, services have been made available on a mass Innovations in the way services are delivered have also market basis, and prices have fallen. ICT is now an everyday made ICT more accessible to customers in Africa. Prepay- service for many Africans, affordable to the majority rather ment has allowed customers to control their expenditures than the privileged few. This growth continues today as and avoid regular monthly subscription payments. African companies innovate, bringing new products and services to operators have also led the way in selling prepay credits in the market. very small units, making calling accessible to large numbers This chapter examines the ICT sector in Africa. The first of people, including the poor. section describes the expansion of voice networks and A combination of new technologies and sound policy broadband Internet. The second section identifies the key choices, including market liberalization, by governments factors that have contributed to the expansion of the sector. has triggered private investment, which has laid the founda- The third section analyzes the impact of ICT on a variety of tions of the connectivity revolution in Africa. The global areas, including mobile banking, telemedicine, and agricul- standardization of mobile telephone network technology ture. The last section examines the role of the World Bank has created international competition in both network Group, to date and in the future. equipment and customer devices, pushing down costs and lowering prices. The liberalization of markets and the estab- lishment of sound regulatory frameworks have allowed pri- EXPANSION OF THE SECTOR vate investors into the market, driving the expansion of The telecommunications sector in Africa has expanded telecommunications networks. Between 1998 and 2008, total rapidly since 1998, with both an increase in network and private sector investment in telecoms in Sub-Saharan Africa services provided. At the same time, prices have fallen every- reached $49 billion (World Bank 2010a), mainly in the where, bringing telecommunications within the economic mobile market (figure 19.2). This investment has resulted in reach of the majority of Africans. a rapid increase in the proportion of the population covered 339 Figure 19.1 Mobile Cellular Subscription in Sub-Saharan Africa, 1998–2008 300 70 258.5 250 60 Subscribers per 100 people 50 200 184.3 40 Million 150 135.3 30 100 90.7 20 54.2 50 36.3 25.1 10 17.0 6.5 1.011.4 1.7 0.6 0 0 98 99 00 01 02 03 04 05 06 07 08 19 19 20 20 20 20 20 20 20 20 20 Mobile cellular subscriptions in Sub-Saharan Africa (left axis) Mobile cellular subscriptions in Sub-Saharan Africa per 100 people (right axis) Mobile cellular subscriptions in world per 100 people (right axis) Source: World Bank 2010b. Figure 19.2 Telecommunications Investment in The success of mobile communication has been very Sub-Saharan Africa, 1998–2008 scalable, with large and small countries experiencing rapid increases in network coverage and access. Both Nigeria, a 12.0 country of about 150 million people, and Rwanda, a coun- 10.0 try of about 10 million people, have experienced steady growth in network coverage (figure 19.4). Billions of dollars 8.0 At the same time as communications networks have 6.0 been expanding across Africa and the number of sub- scribers has risen, the price of telecommunications services 4.0 has been coming down. The cost of mobile communica- 2.0 tions fell by nearly 50 percent between 2000 and 2009. By 2010 the average price of a mobile call in Africa was about 0.0 $0.10 a minute (AICD forthcoming), and prices are con- 98 99 00 01 02 03 4 05 06 07 08 tinuing to fall as competition intensifies. 0 19 19 20 20 20 20 20 20 20 20 20 Source: World Bank 2010a. Broadband internet by mobile networks, which increased from about 10 percent Globally, broadband Internet is becoming an important in 1999 to 60 percent in 2009. More than 90 percent of Sub- part of national economies. Africa lags far behind most Saharan Africa’s urban population now lives within range other parts of the world in access to broadband Internet. of a mobile signal, a remarkable transformation in only Broadband penetration in the region was 6.5 percent in 10 years (figure 19.3). 2008, one-quarter of the world average (ITU 2010). In 2008 340 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES Figure 19.3 Mobile Coverage in Rural and Urban Sub-Saharan Africa, 1999–2009 100 90 80 70 60 Percent 50 40 30 20 10 0 Ju 9 Jan 9 Ju 0 Jan 0 Ju 1 Jan 1 Ju 2 Jan 2 Ju 3 Jan 3 Ju 4 Jan 4 Ju 5 Jan 5 Ju 6 Jan 6 Ju 7 Jan 7 Ju 8 Jan 8 9 -9 l -9 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 l -0 -0 Jan Total Urban Rural Source: Authors based on data from GSM Association. Note: Figures show connection to GSM network. Figure 19.4 Mobile Network Coverage in Nigeria and Rwanda, 2002–08 100 90 80 Percent of population 70 60 50 40 30 20 10 0 02 03 04 05 06 07 08 20 20 20 20 20 20 20 Rwanda Nigeria Source: ITU 2010. fixed broadband Internet cost $100 a month in Sub- KEY SUCCESS FACTORS Saharan Africa, more than three times the world average Many factors have contributed to the success of the ICT sec- of $31.40 (ITU 2010) (figure 19.5). tor in Africa. The most important of these has been the con- This situation is starting to change, as broadband Inter- tinentwide change in sector policy from one based on net becomes more widely available and prices begin to fall. monopoly provision of ICT services to a privately owned, Nigeria, Kenya, and South Africa are leading the way in competitive market. The vast majority of countries in Africa Africa, with rapidly rising numbers of broadband sub- have liberalized their telecommunications markets, and that scribers, mainly through wireless handsets. In most African liberalization has spurred private investment into infrastruc- countries, however, broadband remains beyond the reach of ture and created competition between operators, leading to the majority of the population. expanded availability of services and reduced prices. CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 341 Figure 19.5 Monthly Price of Broadband Internet, by World Region, 2008 120 120 Broadband price as percent of monthly per capita income 100 100 100 101 Monthly broadband internet price (dollars) 80 80 60 60 40 34 40 31 22 23 23 21 20 20 26 9 3 5 8 4 0 0 cif d As d be d Af and ia a ld ric Pa an al an ib an As or ic ia an a Af ric W ia tr e ar a th st h C ic en p As or a ut n C ro e er N eE ra So Eu th Am st ha dl Ea id Sa tin M b- La Su Fixed broadband internet price ($ per month), left hand side Broadband price percent of per capita income (monthly), right hand side Source: ITU 2010. Many governments have also privatized the incumbent The success of the ICT sector in Africa is also attribut- telecommunications operator. Privatization is significant able to innovation within the industry itself. Technologies because it removes one of the main distortionary factors and business models have been adapted to the African in the market and allows the government to carry out market, improving efficiency and tailoring services to the its main role of regulating the market without any conflict specific needs of customers in the region. Finally, the of interest. political economy of the telecommunications sector has These policies have resulted in the rapid growth of the affected government’s willingness to undertake sector mobile telecommunications segment of the market. Broad- policy reforms. band Internet has been much slower to take off in part because of the inadequacy of certain segments of the Creating competitive markets broadband network infrastructure, particularly access to the international communications networks. These bottle- Competition in the telecommunications market was intro- necks have been constraining growth of broadband Inter- duced in Africa in the late 1990s, when mobile networks net, limiting its availability, and making it prohibitively were first established there. All but a handful of countries expensive. Recently, the growth of the submarine fiber- in Sub-Saharan Africa have introduced competition in their optic cable infrastructure has dramatically increased the mobile markets, and more than half have more than two availability of international bandwidth on the continent, mobile operators (ACID forthcoming). This increase in and competition between cable providers has led to rapid competition has been one of the primary drivers of the declines in prices. sector growth and network expansion. 342 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES Competition takes time to develop. Countries that compete effectively with the private sector. The best long- reformed early have experienced better sector performance term solution to these problems is to sell state-owned than late reformers. Performance also improves as competi- enterprises to create a level playing field for all players in tion intensifies. Mobile subscription penetration rates in the market. Africa grew by less than 1 percentage point a year between Realizing this, governments across the region have priva- the establishment of the first mobile network and the entry tized their state-owned telecommunications operators. By of the second operator. By the time a fourth operator enters 2010, 28 African countries had gone through this process the market, the subscriber penetration rate grows by 2–4 (see annex table 19A.1). Mali completed the privatization percentage points a year (AICD forthcoming). of its state-owned telecommunications operator in 2009 The way in which liberalization takes place also has an (box 19.1). impact. In Nigeria, for example, the government awarded the first three GSM (Global System for Mobile communica- Overcoming infrastructure bottlenecks tions) licenses simultaneously, spurring a race to expand coverage and gain market share. Nigeria now has five mobile Most mobile operators have built their own end-to-end net- operators and is considered one of the most competitive works, overcoming infrastructure bottlenecks along the way. markets in the region. With a mobile subscriber penetration As the focus of the ICT industry in Africa shifts toward rate approaching 50 percent, it is also one of the best con- broadband Internet, network infrastructure requirements nected countries in Africa. are changing. Broadband requires networks capable of han- Competition has been the primary driving force behind dling much higher volumes of traffic than that generated by mobile network expansion and, more recently, declines in voice services. As much of the traffic generated by broad- the price of calls. A recent study indicates that competition band crosses international borders, it also requires more could potentially continue to drive this expansion until net- international network infrastructure. This type of network works cover 90 percent of Africa’s population (AICD 2010). infrastructure has traditionally been underdeveloped in Africa. The lack of suitable infrastructure has become a sig- nificant bottleneck, contributing to the slow growth of the Privatizing state-owned enterprises Internet in the region. State ownership of telecommunications operators distorts This situation began to change in 2009 with the develop- markets and adversely affects competition. Governments ment of undersea fiber-optic cables connecting Africa to the often give state-owned operators preferential treatment global communications networks. By 2010 Sub-Saharan by carving out areas of exclusivity for them in the market Africa had 12 operational cables, and another 5 were under or failing to enforce regulatory rules such as interconnec- construction. The operational cables have a combined tion payments. At the same time, these operators often capacity of more than 12 terabytes per second (tbps). A total find it difficult to obtain the necessary capital and skills to of $1.7 billion is being invested in the 5 undersea cables Box 19.1 Privatization and Sector Reform in Mali Mali’s state-owned telecommunications operator, rate rose to 40 percent, and Maroc Telecom is now Société des Télécommunications du Mali (SOTELMA), focusing on revitalizing the fixed-line network and was established in 1989 as the state-owned monopoly rolling out fixed broadband Internet. operator. Despite the continuous growth in the mobile This process was supported throughout by the market in Mali—led by the introduction of competi- World Bank, through a technical assistance project tion, in the form of a second GSM operator, Ikatel in covering the privatization, the reform of the legal and 2003—the fixed-line market remained underserved and regulatory framework, and capacity building for the dominated by SOTELMA. government institutions. A World Bank team contin- In 2009 SOTELMA was successfully privatized, with ues to work with the government of Mali to further a 51 percent stake going to Maroc Telecom. At the same liberalize the telecommunication market and improve time, a new legal framework and universal access strat- connectivity by developing regional links to neighbor- egy were adopted. The telecommunications penetration ing countries. CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 343 currently under construction, which will bring an addi- development of the Internet in Africa. Development of tional 9 tbps of capacity to the region (figure 19.6). these networks has been much slower, although the recent The impact of this growth is already being felt. In East upsurge of interest by African telecommunications compa- and Southern Africa, the SEACOM, TEAMs, and EASSy nies in broadband has prompted new investment in undersea cable projects were all operational by 2010. domestic fiber-optic networks. For the foreseeable future, Wholesale prices for international bandwidth fell, allowing this interest is likely to focus on servicing the profitable operators to provide better-quality service, usually in the intercity routes rather than connecting smaller towns and form of faster download speeds or larger download caps. rural areas. A broadband infrastructure bottleneck is there- The arrival in 2010 of new undersea cables along the west fore likely to remain in these areas for some time to come coast of Africa (Glo-1 and MainOne) together with the (Williams 2010). expected arrival of two more cables (ACE and WACS) in 2011 and 2012 is likely to have a similar effect on markets on Fostering innovation the other side of the continent. The international broad- band infrastructure bottleneck is therefore well on its way to The expansion of communications services creates new being overcome. business opportunities and innovative ways of delivering Domestic fiber-optic networks are needed to carry com- services. The African telecommunications industry itself munications traffic both within and between countries, has also been a source of business innovation. African particularly countries that are landlocked. They are there- operators have had to work hard to find ways of rolling out fore as important as undersea fiber-optic networks for the networks in difficult physical environments and creating Figure 19.6 Undersea Fiber-Optic Cables in Africa, 2010 IBRD 38556 MAY 2011 London, African Undersea Cables England Mediterranean Undersea Cables Marseille, France Monaco Vigo, Spain SEA-ME-WE4, 1280GB, active Sesimbra, Portugal Palermo, Italy I-ME-WE, 3840GB, Q4’09 Chipiona, Spain Asilah, Annaba, Bizerte, Tripoli, EIG, 3840GB, Q2’10 Tunisia Lebanon Morocco Algeria Casablanca, Tripoli, Morocco Alexandria & Libya Cairo, Suez, A.R. of Egypt Fujairah, Altavista, A.R. of United Arab Canary Is. (Sp) Egypt Emirates Karachi, Pakistan Jeddah, Saudi Arabia Oman Mumbai, India Nouakchott, Mauritania Port Sudan, Sudan Dakar, Senegal Massiwa, Eritrea Chennai, Cape The Gambia India Verde Guinea-Bissau Abidjan, Côte d’Ivoire Cochin, Conakry, Djibouti India Guinea Lomé, Togo Freetown, Sierra Leone Accra, Cotonou, Benin Monrovia, Liberia Ghana Lagos, Nigeria Bonny, Nigeria Colombi, Sri Lanka Douala, Cameroon Mogadishu, Bata, Equatorial Guinea Somalia Libreville, Gabon São Tomé and Príncipe Mombasa, Kenya Pointe-Noire, Congo Muanda, D.R. of Congo Cacuaco, Angola Dar es Salaam, Luanda, Angola Tanzania Tamatave, Sub-Saharan Undersea Cables Madagascar SAT3/SAFE, 380 GB, active Toliara, Baie du Jacobet, GLO-1, 640GB, Q4’09 Walvis Bay, Madagascar Mauritius Maputo, St Paul, Namibia Mozambique Réunion (Fr) TEAMs, 1280GB, Q3’09 Seacom, 1280GB, active Mtunzini, Lion, 1300gb, active South Africa EASSy, 1400GB, Q2’10 Melkbosstrand, South Africa ACE, 1920GB, 2011 MaIN OnE, 1920GB, Q4’10 WACS, 5120GB, Q2’11 Source: http://manypossibilities.net. 344 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES commercially viable business models based on a customer Third, the establishment of global technology standards base that was once considered marginal. The introduction in the telecommunications industry and the increasing level of prepaid airtime in very small units is one example of of international competition in the manufacture of how African operators have introduced innovations in telecommunications equipment have put strong downward their commercial practices to suit the African market. pressure on the cost of building and operating telecommu- Zain’s one-network, free regional roaming service is nications networks in Africa. These reduced costs have another innovative business model that originated in allowed many operators to maintain their profitability, even Africa (box 19.2). in the face of increasing competitive pressures. Political economy of the sector IMPACT OF ICT IN AFRICA The rapid pace of policy reform in the telecommunications Several studies examine the macroeconomic impact of ICT. sector in Africa has been influenced by the political econ- Roller and Waverman (2001) analyze 21 advanced countries omy of the sector. Three key factors underlie this influence. over 20 years (1970–90). They find a causal nonlinear First, telecommunications has traditionally not been relationship between telecommunications infrastructure available to anyone other than a small elite section of the investment and economic performance, where impact population able to afford and obtain a fixed line. But mobile increases once countries pass a certain threshold of phones are very popular and there is strong public demand telecommunications access. Qiang and Pitt (2004) find that for them, even in low-income countries. The proven ability ICT has made a significant contribution to economic of private companies to develop and operate mobile phone growth across a wide range of countries. networks has therefore been less politically controversial More recently, research has focused on the economic than in other sectors. impact of broadband Internet. Qiang and Rossotto (2009) Second, the telecommunications business, particularly find that every 10 percentage point increase in broadband mobile telephony, has been very profitable for many Internet penetration in developing countries results in 1.38 investors. As a result private sector interest in the sector has percentage points of additional GDP growth. Koutroumpis grown, and companies have been prepared to pay consider- (2009) finds a significant and positive relationship between able sums for mobile licenses. This has provided an imme- broadband infrastructure and local, regional, and national diate source of revenue for governments that, combined economic growth, including growth in small and medium- with ongoing license-fees and tax payments, has made the size enterprises. These findings are consistent with other telecommunications sector an important source of govern- research that finds that broadband is an important factor in ment revenues, which, in turn, has had an influence on gov- social transformation and improved service provision, par- ernment’s willingness to liberalize the market and encourage ticularly in rural areas, and that mobile broadband services private investment. are particularly well suited for improving the economic Box 19.2 Borderless Roaming In late 2006 Kuwait-based mobile operator Zain (for- Through this service, Zain’s customers (both prepaid merly Celtel) launched the world’s first borderless and postpaid) can make calls at local rates, roaming mobile network, One Network, in the Middle East and without incurring surcharges. They add air time with East Africa. The service began in 4 countries in East locally purchased airtime cards. In 2009 Zain added Africa (Kenya, Sudan, Tanzania, and Uganda) and Internet access, e-mail, Multimedia Messaging Service expanded to 12 other countries (Burkina Faso, Chad, (MMS), BlackBerry services, Short Messaging Service the Republic of Congo, the Democratic Republic of (SMS), international roaming, and mobile portal appli- Congo, Gabon, Ghana, Madagascar, Malawi, Niger, cations to its One Network local rate, further adding Nigeria, Sierra Leone, and Zambia). value to its African customers. Source: Zain web site (http://www.zain.com/muse/obj/lang.default/portal.view/content/Zainpercent20World/Zainpercent20 Connect/Onepercent20Network) and Business Monitor International (BMI) database (http://www.businessmonitor.com/). CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 345 well-being of poor people (Lobo, Novobilski, and Ghosh Telemedicine 2008). Countries that do not develop effective broadband Mobile phone operators, health service providers, health infrastructure may therefore be failing to capitalize on some ministries, and donor organizations are working together to of their economic growth potential (Gaasbeck and Kristin develop innovative ICT-based approaches to healthcare ser- 2008; Tolkoff 2007). vice delivery. One such application is TRACnet, which is used The economic research on the impact of ICT has by the Ministry of Health in Rwanda to improve the quality focused mainly on the impact at the level of the macro- of service in primary health care institutions (box 19.4). economy or firms. ICT is also having a visible impact on the lives of individuals, changing the way people live and work in many different ways. Agriculture Agriculture is an essential part of Africa’s economy, accounting for 13 percent of GDP and employing about 194 Mobile banking million people (World Bank 2010b). ICT is having a positive Banking by cell phone in Africa is one of the most signifi- impact on the sector in many ways (box 19.5). cant developments in the recent history of the continent’s financial sector. The success of some of the early pioneers of IT–enabled services and business process cell phone banking has been replicated in other countries, outsourcing through the launch of other types of financial service prod- ucts delivered by cell phone. In addition to being a platform for delivering services, the Safaricom’s M-PESA in Kenya was one of the first mobile ICT sector has the potential to be a source of economic banking applications to be launched on the continent (box growth and employment itself. The IT-enabled service 19.3 and chapter 20). MAP mobile banking in Uganda fol- (ITES) sector—also known as business process outsourcing lowed suit. WIZZIT, in South Africa, is another mobile- (ITES-BPO)—is becoming established around the world, based “virtual bank,” whose services can be accessed including in some African countries. The sector covers a through any national mobile phone operator. In addition to wide range of industries from applications development providing a valuable service to customers, WIZZIT has had and services through to IT-enabled services such as call cen- a positive economic effect, by employing nearly 2,000 previ- ters and other types of business process outsourcing. ously unemployed “WIZZkids” as its sales force. These com- The ITES-BPO industry is growing and still has consider- panies are revolutionizing the financial services sector in able potential as a source of growth in Africa. Large and small Africa, bringing low-cost financial services to the majority companies around the world are increasingly hiring compa- of the population, that was initially unable to access the tra- nies in Africa to help them deliver efficient, reliable, and cost- ditional banking sector. effective customer support and other key services, such as Box 19.3 Serving the Banking Needs of the Poor in Kenya M-PESA was launched in 2007 to meet the banking money to be safe with M-PESA, 81 percent find it very needs of the financially excluded. By July 2010, M-PESA easy to use, and 84 percent believed the service to be crit- had more than 11 million customers (about 30 percent ical to their socioeconomic well-being (Agrawal 2010). of Kenya’s population) and almost 20,000 agents (up Morawczynski and Pickens (2009) find that incomes from 355 at inception). Person-to-person transactions of rurual recipients increased 5-30 percent since they stood at more than $375 million a month (Safaricom started using M-PESA. Ltd, 2010). M-PESA has succeeded mainly because it has a There is strong demand for M-PESA’s services in broad market positioning, has a built-in accountability Kenya, which have had a positive economic impact. Use structure, is easy and safe to use, provides 24/7 support, of and satisfaction with M-PESA is high. About 40 per- is affordable, is provided by the largest mobile phone cent of households use M-PESA (63 percent of them for network, and has a wide network of agents ensuring regular financial support), 90 percent believe their convenience in sending and receiving cash. 346 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES Box 19.4 Improving the Quality of Service Provided by Primary Health Care Facilities in Rwanda TRACnet is a mobile phone–based platform for moni- The improved information exchange between remote toring HIV treatment in Rwanda. By 2009 it had regis- health facilities and central actors has reinforced tered more than 1,000 service providers, conducted accountability in care and treatment of patients. more than 85,000 annual user sessions, and collected TRACnet has succeeded because it is based on longitudinal data on more than 105,000 patients. simple technology using the widely available platform Thanks to TRACnet, Rwanda has access to robust of the mobile phone. The centralized database is datasets of HIV/AIDS patients located centrally and interoperable with multiple communication channels accessible from any location, allowing faster and better- and it uses open source software giving it flexibility informed intervention. As a result, public monitoring of and scalability. It has the full support of the Rwanda HIV/AIDS transmission patterns has improved. Doc- government—a key to its success—and the program tors and patients also have instantaneous access to more includes a training component that ensures that reliable information. Real-time monitoring of anti- health workers are well prepared to work with the retroviral drug stocks leads to quicker replenishments. system. Source: http://www.un.org/esa/sustdev/publications/africa_casestudies/tracnet.pdf; http://www.kiwanja.net/database/project/ project_voxiva_hivaidsrelief.pdf. Box 19.5 Increasing Access to Market Information in West Africa Esoko leverages mobile phones to enhance productivity Transactions costs for farmers and traders have also gains for African farmers and traders by giving them fallen by $2–$150 per transaction by significantly quicker access to better market information. Starting in reducing the role of middlemen or cutting them out 2007 in Ghana with only $90,000 and operating in the altogether. It has also transformed mobile phones into a red by up to $21,000, Esoko had broken even by the market bulletin increasing their utility beyond voice fourth year and by the fifth year had $1.4 million in rev- and text. enues and about $540,000 in profits. By 2010 Esoko had Esoko has succeeded mainly because it uses open been scaled to seven other countries (Benin, Burkina source software, enabling it to scale up, tailor business Faso, Cameroon, Côte d’Ivoire, Madagascar, Mali, and services to local needs, use affordable mobile telephony, Togo) and had grown to 40 full-time employees and offer free listing of services, allow sending of and receipt about 9,000 users. of text messages in several languages, provide real-time Esoko has led to a 6.4 percent fall in grain price mar- commodity prices, and provide direct access to markets ket dispersion and a 3.5 percent decline in mean prices. worldwide. Source: http://www.esoko.com/; http://www.slideshare.net/slavb/s-bartlett-esoko-cirad-2010; http://www.ifc.org/ifcext/ spiwebsite1.nsf/f451ebbe34a9a8ca85256a550073ff10/0e6f5be010f90329852576ba000e2dad?OpenDocument; http://www.slideshare .net/slavb/davies-esoko-cirad-2010. data entry and document processing. A 2010 study by participation than in the service sector in general McKinsey and Company estimates that the addressable mar- (Sudan et al. 2010). ket for IT and ITES offshoring is $500 billion a year, of which ■ Increase in investment. The IT/ITES sector helps attract only about 20 percent has been realized (Sudan et al. 2010). foreign investment, transform the financial sector, ener- Success in the ITES–BPO sector brings with it a number gize local exports, and nurture ICT skills and innovation of benefits, including the following: in the workforce. ■ Job creation. It is estimated that every job created in the ■ Employment of women. Women now account for a large IT/ITES sector results in the creation of four additional percentage of all professional and technical workers jobs in ancillary sectors, such as transport, training, and in the IT/ITES sector—a much higher rate of female catering. CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 347 Ghana and South Africa have identified ITES as one of development in Central Africa (Cameroon, the Central the key sectors for enhancing economic growth. Both seek African Republic, and Chad) and West Africa. to position themselves as premier BPO destinations in The Bank is also supporting governments in their efforts Africa (box 19.6). to use the ICT infrastructure to improve service delivery. In three stand-alone projects (e-Benin, e-Ghana, and e-Rwanda), it is providing funds to incorporate ICT into the THE ROLE OF THE WORLD BANK GROUP delivery of public services. Implementation of the RCIP The World Bank Group has been involved in the ICT sec- program in Kenya, Mozambique, and Tanzania also has tor in Africa since 1969, when the International Develop- major e-government components. ICT is also integrated ment Association funded an $800,000 telecommunication throughout the broader World Bank portfolio of projects project in Burkina Faso. Its work has covered investment and is a major component of the Bank’s research, dissemi- lending, policy and regulatory reforms, privatization, nation, and policy development work in Africa. This work and e-government and applications projects. The Bank has included major outreach events, such as an innovations has provided technical assistance to 27 Sub-Saharan conference day held as a part of the Africa Union Summit in African governments to strengthen policy and regulatory Ethiopia in 2010 and a seminar on “The Transformational frameworks and build institutional capacity in the Power of ICT for Africa” at the 2010 World Bank Group sector, including capacity for regulatory design and Spring Meetings. implementation. The World Bank is also supporting African govern- In 2007 the World Bank Group launched an initiative to ments in their efforts to use the growing ICT infrastruc- enhance broadband connectivity in Eastern and Southern ture to develop the IT and ITES-BOP industry. Many of Africa through the Regional Communications Infrastruc- the ICT sector projects in Africa include components that ture Program (RCIP). The program is currently operational support the industry through training, regulatory reforms, in seven countries (Burundi, Kenya, Madagascar, Malawi, and, in some cases, investment in supporting infrastruc- Mozambique, Rwanda, and Tanzania) and may be expanded ture. A related initiative is the New Economy Skills for to other countries in the region. Similar projects are under Africa Program (NESAP), a joint effort by World Bank Box 19.6 Expanding Business Process Outsourcing in Ghana and South Africa In Ghana the government’s proactive policies of sector outsourcing agents is expected to further strengthen reform has created a competitive telecommunications Ghana’s position as one of the most attractive loca- industry with a telephone penetration of more than tions for ITES business in Africa. 60 percent. This industry is providing a platform on In July 2010 Amazon.com, the world’s largest Internet which an IT and IT-based services industry is growing. retailer, announced that it would open a new customer One of the efforts, an $84.4 million World Bank service call center in Cape Town, South Africa, where it project, is supporting is the development of a busi- will operate a software development center. According to ness process outsourcing center. More than 1,000 jobs the local media, this business development was brought were created in the industry in 2009 and 2010. Ghana about by the strong commitment by the central and estimates that it will create some 37,000 jobs by 2011, Western Cape governments for the development of the increasing the sector’s contribution to GDP by about ITES-BPO industry. It is estimated that the call center $750 million. The 2009 AT Kearney Global Services will create 1,000 new jobs (600 permanent positions and Location Index ranks Ghana 1st out of 50 countries in 400 seasonal positions). Servicing will be provided in the world in terms of financial attractiveness and 15th English and German. Local government officials hope in terms of location attractiveness. The partnership that the deal will boost the local economy not only by between the public and private sector to train some creating jobs but also by demonstrating that Cape Town 50 training providers and 6,000 business process has the capacity to host world-class clients. Source: http://www.busrep.co.za/index.php?fSectionId=561andfArticleId=5554024 http://retail.bizcommunity.com/Article/196/458/ 49992.html. Notes: (http://www.ites.gov.gh/IT-Industry.aspx). 348 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES Box 19.7 Supporting Skills Development in Africa through the New Economy Skills for Africa Program–Information and Communication Technologies (NESAP–ICT) Initiative NESAP–ICT is the World Bank’s cross-sectoral initia- Implementation of Window I began in 2008. The tive to support the development of ICT skills. various skills required for the IT/ITES sector were iden- Launched in 2008 in eight Sub-Saharan African tified and segmented to meet each country’s needs. countries (Ghana, Kenya, Madagascar, Mozambique, Fifty-four delegates from eight countries participated in Nigeria, Rwanda, Senegal, and Tanzania), the pro- a study tour of India. Pilot projects, guided by interna- gram seeks to bridge education and industry gaps tional best practice, were launched in four countries in the field, create a globally benchmarked assess- (Ghana, Kenya, Nigeria, and Tanzania). The program ment talent pool for the ITES-BPO industry, and established partnerships with the world’s leading ICT strengthen collaboration with IT industry leaders and companies (including Microsoft, Intel, Cisco, Oracle, associations. IBM, Nokia, and EMC); with learning institutions (for The initiative consists of two focus areas. Window I example, Carnegie Mellon University); and with IT- focuses on developing ICT skills for a new emerging BPO industry associations (for example, NASSCOM in ICT-based sector. Window II focuses on ICT use in India). Implementation of Window II, expected to begin education. in 2011, will build on the Window I accomplishments. education, private sector development, and ICT sector investment to 10 ICT projects in Africa during fiscal 2010. It units to provide skills development for the ICT industry is also investing in innovative new ICT businesses, such as (box 19.7). WIZZIT in South Africa, which is providing mobile-based In parallel with World Bank activities, IFC has been banking services, and Helios Towers Africa Ltd., which is closely involved in the sector since the beginning of the providing tower infrastructure services to mobile operators mobile revolution in Africa. It committed a $261 million in Nigeria. CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES 349 350 Table 19A.1 Privatizations of Telecom Incumbents in Sub-Saharan Africa, 1995–2010 Percent Initial privatization transaction Percent Amount private Country Operator Date sold ($millions) 2008 Note Burkina Faso ONATEL December 2006 51 295 51 Private sale to Maroc Telecom. Cape Verde Cabo Verde Telecom December 1995 40 20 59 Private sale to Portugal Telecom. Subsequent distribution to employees (5 percent of total), national private investors (14 percent), and government social security system (38 percent). Central African Rep. Socatel — — — — France Cable and Radio owned 40 percent of shares at one point. Current status not available. Côte d’Ivoire Côte d’Ivoire Telecom January 1997 51 210 51 Private sale to France Telecom. Equatorial Guinea Getesa 1987 40 40 Private sale to France Telecom. Gabon Gabon Telecom February 2007 51 79 51 Private sale to Maroc Telecom. Gambia, The GAMTEL January 2007 50 35 50 Private sale to Spectrum Investment Holding (Lebanon). Ghana Ghana Telecom December 1996 30 38 70 Original private sale to G-Com consortium headed by Telekom Malaysia. In 2002 the government of Ghana abrogated the management contract with G-Com and bought back shares. Subsequent private sale of 70 percent to Vodafone (UK) in August 2008 for $900 million. Guinea Sotelgui December 1995 60 45 0 Renationalized in 2008 following private sale to Telekom Malaysia. Guinea-Bissau Guinée Telecom 1989 51 3 0 Renationalized following private sale to Marconi (later assumed by Portugal Telecom). Kenya Telkom Kenya December 2007 51 390 51 Sale to consortium led by France Telecom (78.5 percent) with Alcazar Capital Ltd. (21.5 percent). Lesotho Telecom Lesotho November 2000 70 — 70 Private sale to Mountain Communications (Econet) (Zimbabwe), Mauritius Telecom, and Eskom (South Africa). Sale price not disclosed. Madagascar TELMA August 2003 34 12.6 68 Private sale to Distacom (Hong Kong, China), which also purchased France Telecom’s ownership. Malawi Malawi Telecom (MTL) February 2006 80 30 80 Private sale to Telecom Holdings Ltd. (THL) consisting of PCL (50.1 percent), Old Mutual (16.1 percent), NICO (5.0 percent), Detecon (Germany) (2.6 percent) and Press Trust (6.2 percent). Percentages refer to MTL ownership. Mali SOTELMA July 2009 51 384 51 Private sale to Maroc Telecom. Mauritania Mauritel April 2001 54 48 54 Private sale to Maroc Telecom, which subsequently engaged in a series of sales with local investors. Its ownership stood at 51 percent in 2008. Mauritius Mauritius Telecom November 2000 40 261 40 Private sale to France Telecom. Niger SONITEL November 2001 51 16 51 Private sale to Chinese and Libyan consortium. Government has announced intention to renationalize. Nigeria NITEL July 2006 51 500 Private sale to TransCorp (Nigeria). Government has rescinded the sale and was in process of reprivatizing in 2010. Rwanda Rwandatel June 2005 99 20 80 Initial private sale to Terracom (United States), which government later repurchased. Libya Arab Portfolio later acquired an 80 percent interest for $100 million. São Tomé and Principe CST (Companhia Santomense de 1989 51 1 51 Private sale to Portugal Telecom. Telecomunicações) Senegal Sonatel July 1997 33 90 73 Initial private sale to France Telecom. Subsequent additional sale to France Telecom and listing on regional stock exchanges. Seychelles Cable and Wireless Seychelles 1981 49 n.a. 100 Government granted the British company Cable and Wireless the right to operate the telephone network, inaugurated in 1954. Cable and Wireless was subsequently privatized by British government in three tranches (1981, 1983, and 1985). South Africa Telkom May 1997 30 1,261 34 Initial private sale to Thintana Communications (SBC [United States] 60 percent) and Telekom Malaysia (40 percent). Global initial public offering in March 2003 of 47 percent. Thintana Communications sold 14.9 percent interest in Telkom to South African and international institutional investors in June 2004 and its remaining interest to the Public Investment Corporation, wholly owned by the South African Government in November 2004. Subsequent Telkom share repurchases have altered its level of private shareholding. Sudan Sudatel 1997 n.a. n.a. 36 Multiple share offerings on local and regional stock exchanges. Tanzania TTCL (Tanzania Telecommunica- February 2001 35 65 35 Private sale to a consortium of MSI (Netherlands) and tions Company Ltd.) Detecon (Germany). Uganda Uganda Telecom May 2000 51 34 69 Initial private sale to UCOM consortium consisting of Detecon (Germany), Telecel (Switzerland), and Orascom (Egypt). UCOM ownership subsequently purchased by Libyan Arab Portfolio (LAP), which then increased its ownership through a capital increase. Zambia Zamtel June 2010 75 394 n.a. Privatization of 75 percent stake sold to LAP Green (Libya) for $394 million in 2010. Source: AICD 2009; Business Monitor International (BMI) database (http://www.businessmonitor.com/). Note: n.a. = Not applicable. — = Not available. 351 REFERENCES Qiang, C. Z-W., and A. Pitt. 2004. “Contribution of Infor- mation and Communication Technologies to Growth.” Agrawal, M. 2010. “M-Pesa: Transforming Millions of Lives – I.” World Bank Working Paper 24, Washington, DC. http://www.telecomcircle.com/2010/01/m-pesa/. AICD (Africa Infrastructure Country Diagnostic). 2009. Qiang, C. Z-W., and C. M. Rossotto, with K. Kimura. 2009. Africa Infrastructure Country Diagnostic Background “Economic Impact of Broadband.” In Information and Paper 10. Information and Communications Technology in Communication for Development: Extending Reach and Sub-Saharan Africa: A Sector Review. Washington, DC. Increasing Impact, 35–50. Washington, DC: World Bank. ———. 2010. Africa’s Infrastructure: A Time for Transforma- Roller, L-H., and L. Waverman. 2001. “Telecommunica- tion. Washington, DC. tions Infrastructure and Economic Development: A Simultaneous Approach.” American Economic Review ———. Forthcoming. ICT Sector Volume. Washington, DC. 91 (4): 909–23. Gaasbeck, V., and A. Kristin. 2008. “A Rising Tide: Measuring Safaricom Ltd. 2010. http://www.safaricom.co.ke/ the Economic Effects of Broadband Use across California.” Social Science Journal 45 (4): 691–99. Sudan, R., S. Ayers, P. Dongier, A. Muente-Kunigami, and C. Zhen-Wei Qiang. 2010. The Global Opportunity in International Telecommunications Union (ITU). 2010. IT–Based Services: Assessing and Enhancing Country World Telecommunication/ ICT indicators Database. Competitiveness. Washington, DC: World Bank. Geneva, Switzerland. TeleGeography CommsUpdate. 2009. July 8. http://www Koutroumpis, P. 2009. “The Economic Impact of Broadband .telegeography.com/cu/index.php. on Growth: A Simultaneous Approach.” Telecommunica- tions Policy 33 (9): 471–85. Tolkoff, S. 2007. “Increased Broadband Use Could Add Lobo, B. J., A. Novobilski, and S. Ghosh. 2008. “The Economic 186,000 Jobs.” Orange County Business Journal 30 (50): 9. Impact of Broadband: Estimates from a Regional Input- Williams, Mark D. J. 2010. Broadband for Africa: Developing Output Model.” Journal of Applied Business Research 24 Backbone Communications Networks. Washington, DC: (2): 103–14. World Bank. Morawczynski, O., and M. Pickens. 2009. Poor People Using World Bank. 2010a. Public Private Infrastructure Indicators Mobile Financial Services: Observations on Customer Usage Database. Washington, DC. and Impact from M-PESA. Consultative Group to Assist ———. 2010b. World Development Indicators. Washington, the Poor (CGAP), Washington, DC. DC. 352 CHAPTER 19: ICT IN SUB-SAHARAN AFRICA: SUCCESS STORIES CHAPTER 20 Mobile Payments Go Viral M-PESA in Kenya Ignacio Mas and Dan Radcliffe -PESA is a small-value electronic payment and M cash to and withdraw cash from their accounts by exchang- store-of-value system in Kenya accessible from ing cash for electronic value at a network of retail stores. ordinary mobile phones. It has seen exceptional Second, it allows users to transfer funds to others, to pay growth since its introduction in March 2007. Now in bills, and to purchase mobile airtime credit. Retail stores are use by 9 million customers—40 percent of Kenya’s adult paid a fee by Safaricom each time they exchange cash for M- population—the system processes more transactions PESA credit on behalf of customers. All M-PESA transac- domestically than Western Union does globally. M-PESA’s tions are authorized and recorded in real time using secure market success is the result of the interplay of three fac- short messaging service (SMS) and are capped at $500. tors: preexisting country conditions that made Kenya a M-PESA registration is free, as is making deposits into conducive environment for a successful mobile money the system. Customers are charged flat fees of approximately deployment; a clever service design that facilitated rapid $0.403 for person-to-person (P2P) transfers and bill pay- adoption and early capturing of network effects; and a ments, $0.33 for withdrawals (for transactions under $33) business execution strategy that helped M-PESA rapidly and $0.013 for balance inquiries. Individual customer reach a critical mass of customers, thereby avoiding the accounts are maintained in a server that is owned and man- adverse chicken-and-egg (two-sided market) problems aged by Vodafone, but Safaricom deposits the full value of that afflict new payment systems. its customers’ balances in the system in pooled accounts in two regulated banks. Thus, while Safaricom issues and man- ages the M-PESA accounts, the value of the accounts is fully M-PESA IN A NUTSHELL backed by highly liquid deposits at commercial banks. M-PESA (“M” for mobile and “PESA” for money in Rather than paying customers interest on the balance in Swahili)was developed by mobile phone operator Vodafone their M-PESA accounts, Safaricom sets aside a small per- and launched commercially by its Kenyan affiliate Safari- centage of account balances in a not-for-profit trust fund. com in March 2007.1 To access the service, customers must The purpose of these funds has not yet been decided. first register at an authorized M-PESA retail outlet. They are M-PESA’s function as a retail payment platform is then assigned an individual electronic money account important because it reaches a large number of people com- linked to their phone number and accessible through an pared with other financial services outlets in Kenya (figure application stored on the subscriber identification module 20.1).4 There are now nearly five times as many M-PESA (SIM) cards of their mobile phones.2 The application has outlets in Kenya as there are PostBank branches, post two main functions. First, it allows customers to deposit offices, bank branches, and automated teller machines 353 Figure 20.1 Outlets Offering Financial Services in Kenya M-PESA trials. DFID’s role in spotlighting the need for mobile payments and funding the early risk demonstrates 100,000 good practice for donor funding. 100,000 16,900 Number of outlets A snapshot of M-PESA after three years 10,000 The speed and extent to which M-PESA has been deployed 1,510 in Kenya is remarkable (figure 20.2). The number of cus- 1,000 800 840 440 tomers hit the 9 million mark in November 2009, less than three years after the service was launched. This number of 100 customers represents 60 percent of Safaricom’s customer base, 40 percent of Kenyan adults, and 23 percent of the cie s Ba fice t es s s rs of pos en he M re lle ch AT To s s to ag c country’s total population.5 se d ran an l as ta re br an k b es e nk -P By other measures, too, M-PESA has a deep reach in im n ba M rt st Ai Kenya.6 The number of retail stores at which M-PESA users Po Sources: Central Bank of Kenya, Kenya Post Office Savings Bank, and can cash in and cash out now totals 16,900, of which nearly Safaricom. half are located outside urban centers. M-PESA now han- dles $320 million in P2P transfers per month. On an annu- alized basis, this is equal to roughly 10 percent of Kenya’s (ATMs) combined. M-PESA’s presence in rural areas is par- gross domestic product (GDP).7 An average of $650 million ticularly important, because access to financial services in in cash deposit and withdrawal transactions is made at such areas is limited, and the ability to use existing retail M-PESA stores every month, with an average transaction stores as M-PESA cash-in/cash-out outlets reduces deploy- size of approximately $33. Nearly one-fifth of Safaricom air- ment costs, provides greater convenience, and lowers the time purchases are now conducted through M-PESA. cost of access compared with other financial services outlets. M-PESA’s bill pay function, launched in March 2009, has Importantly, both private and public actors were been popular: 75 companies now use M-PESA to collect pay- involved in creating and enabling M-PESA. The idea of ments from their customers. The biggest user, the electric M-PESA was conceived by a London-based team within utility company, reports that roughly 20 percent of its 1 mil- Vodafone. This team believed that mobile phones could play lion customers now pay through M-PESA. At least two banks, a central role in lowering the cost of access to financial ser- Family Bank and Kenya Commercial Bank, are now using vices for poor people. The idea was then developed by the M-PESA as a mechanism for customers to either repay loans Safaricom team in Kenya, which customized it and oversaw or withdraw funds from their banks accounts. And 27 com- a very focused execution. The Central Bank of Kenya, in par- panies are using M-PESA for bulk payment distribution. ticular the Payments System Group, helped to enable the launch of M-PESA by allowing a mobile operator to take the Customer perspectives on M-PESA lead in providing payment services to the general population. Following the first FinAccess survey in 2006, which showed A survey of 3,000 M-PESA users and nonusers conducted in very low levels of bank penetration in Kenya, the central bank late 2008 shed considerable light on the profile of M-PESA’s was determined to explore all reasonable options for correct- early adopters and customer usage patterns. Compared with ing the financial access imbalance. It worked in close partner- nonusers, the average M-PESA user is twice as likely to have ship with Vodafone and Safaricom to assess the opportunities a bank account (72 percent versus 36 percent) and is wealth- and risks involved prior to the launch of M-PESA. Conscious ier (65 percent higher expenditure levels), more literate, and that premature regulation could stifle innovation, the Central better educated (Suri and Jack 2008). Early adopters of the Bank of Kenya chose to closely monitor and learn from early service also appear to be experienced with banking services M-PESA trials and to formalize regulations later. and fairly savvy with technology, which probably makes Finally, the United Kingdom’s Department for Interna- them more keenly aware of the convenience of M-PESA rel- tional Development (DFID) played an instrumental role in ative to alternative financial services. the creation of M-PESA within Kenya, first by funding the Figure 20.3 highlights the ways customers report they use organizations that made the FinAccess survey possible, and M-PESA. More than half the sample use the service primar- then by providing seed funding to Vodafone for early ily for sending and receiving money, a use consistent with 354 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA Figure 20.2 Growth of the M-PESA Customer Base 10 Number of customers (in millions) 9 8 7 6 5 4 3 2 1 0 7 07 7 8 8 8 8 08 8 9 9 9 9 09 9 l-0 -0 -0 -0 -0 l-0 -0 -0 -0 -0 l-0 -0 p- p- p- ov Jan ar ay ov Jan ar ay ov Ju Ju Ju Se Se Se M M M M N N N Source: Safaricom. Figure 20.3 How M-PESA Customers Use the Service Ninety-eight percent of M-PESA users are happy with the service, according to the 2008 survey, and 84 percent Pay bills, 2% claim that losing M-PESA would have a large, negative effect Store/save money for emergencies, 14% None of these, 1% on them (Suri and Jack 2008). Figure 20.4 illustrates how customers compare M-PESA with alternative services. In all Buy airtime for categories—speed, convenience, cost, and safety—customers someone else, reported better service with M-PESA than with other forms 8% of financial services. Receive money, 28% M-PESA’s service evolution M-PESA’s original core offering, the P2P payment, enabled Buy airtime for myself, 14% customers to send money to anyone with access to a mobile phone. Importantly, it opened up a market for transactions that previously were handled largely informally—through Store/save money for Send money, personal trips, friends, and public transport networks (“per- 25% everyday use, sonal networks” in figure 20.5). Although many transactions 14% carried out under M-PESA, such as sending a portion of salary earned at the end of the month to relatives, can be characterized as scheduled payments, others allow people Source: Suri and Jack 2008. to draw on a much broader network of family members, friends, and business associates to access money when required. Thus, in addition to providing a large measure of M-PESA’s broad market positioning. Though 21 percent convenience for transactions that were already occurring, report using M-PESA for storing money, the survey revealed M-PESA also provides a basic form of financial protection that less than 1 percent of accounts had balances of more for a large number of users by enabling a network for than K Sh 1,000 ($13), and a government audit (Okoth instant, on-demand payments. 2009) of M-PESA in January 2009 showed that the average In recent months, Safaricom has increasingly opened up balance on M-PESA accounts was only $2.70. The survey M-PESA to institutional payments, enabling companies to also found that 52 percent of customers use the service on pay salaries and collect bill payments. In the future, Safari- only a monthly basis, suggesting that customers have yet com envisions increased use of M-PESA for e-commerce to incorporate M-PESA into their daily lives (Suri and and in-store purchases, a strategy represented by the down- Jack 2008). ward arrow in figure 20.5. CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 355 Figure 20.4 User Ratings of M-PESA Compared with Alternatives Slower Less 2% convenient 4% Speed Convenience More Quicker More convenient 98% Less safe expensive 96% 2% 4% Safety Cost Safer Cheaper 98% 96% Source: Suri and Jack 2008. Note: At the time of the survey, alternatives consisted of 876 bank branches, 1,025 Post Office branches, 1,424 ATMs and 6,104 M-PESA agents. Figure 20.5 Potential Range of Transactions Supported by M-PESA FORMAL FINANCIAL M-PESA role PRODUCTS in promoting savings, credit, insurance fuller financial inclusion? INFORMAL SERVICE PROVIDERS pawnbroker, moneylender PUSHING & PULLING MONEY ACROSS TIME “On-demand” payments PERSONAL NETWORKS “JUST “scheduled” payments PAYMENTS” REMOTE B2C/C2B INSTITUTIONAL PAYMENTS M-PESA as salaries, bill pay, G2P, a fuller retail online/e-commerce payments platform IN-STORE MERCHANT PAYMENTS for goods & services Source: Bill & Melinda Gates Foundation analysis. As represented by the upward arrow in figure 20.5, products that balance customers’ preference for liquidity another of M-PESA’s goals is to become a vehicle for delivery and commitment and that will connect to a broader range of a broader range of financial services for the Kenyan pop- of financial institutions. This is the path M-PESA must take ulation. Thus far, evidence that people are willing to use the to deliver on its promise of addressing the challenge of basic M-PESA account as a store of value is limited. There is financial inclusion in Kenya. A key precondition is regula- likely, however, to be a need to develop more targeted savings tion: the Central Bank of Kenya is in the process of finalizing 356 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA regulations that will allow nonbank outlets and platforms developing countries have developed a usage-based revenue such as M-PESA become channels for formal deposit-taking. model, selling prepaid airtime to poor customers in small Beyond that, Safaricom will need to develop appropriate increments, so that each transaction is profitable on a stand- service, commercial, and technical models under which alone basis. This is the magic behind the rapid penetration M-PESA can interface with the systems of other financial of prepaid airtime into low-income markets: a card pur- service providers. chased is profit booked, regardless of who buys the prepaid card. This usage-based revenue model is directly aligned The broader lessons from M-PESA’s success with the model needed to sustainably offer small-value cash-in/cash-out transactions at retail outlets and would In addition to the compelling marketing, cold business make possible a true mass-market approach, with no incen- logic, and consistent execution behind M-PESA, the success tive for providers to deny service based on minimum bal- of the service is also a vivid example of how great things can ances or intensity of use. happen when public and private organizations rally around common challenges and innovative solutions. Three top- A low-cost transactional platform that enables line lessons emerge from M-PESA’s success. First, M-PESA customers to meet a range of payment needs. Once shows the promise of leveraging mobile technology to customers are connected to an e-payment system, they can extend financial services to a large number of unbanked use this capability to store money in a savings account, send poor people. Second, it demonstrates the importance of and receive money from friends and family, pay bills and designing usage-based rather than float-based revenue monthly insurance premiums, receive pension or social wel- models for financial services for poor customers. And third, fare payments, or receive loan disbursements and repay M-PESA demonstrates the importance of building a low- them electronically. In short, when customers are connected cost transactional platform that enables customers to meet to an e-payment system, their range of financial possibilities a broad range of payment needs.8 expands dramatically. Putting these elements together, M-PESA has prompted Leveraging technology to extend financial services a rethink on the optimal sequencing of financial inclusion to unbanked poor people. Mobile phone technology is strategies. Whereas most financial inclusion models have quickly becoming ubiquitous, even among poor segments of employed “credit-led” or “savings-led” approaches, the the Kenyan population. Mobile penetration in Africa has M-PESA experience suggests that there may be a third increased from 3 percent in 2002 to 48 percent in 2010, and approach—focusing first on building the payment “rails” on is expected to reach 72 percent by 2014, according to Wire- which a broader set of financial services can ride. less Intelligence. The mobile device mimics some of the key ingredients needed to offer banking services. The SIM card inside GSM phones, for example, can be used to authenti- KENYA COUNTRY FACTORS: UNMET NEEDS, cate users, thereby avoiding the high costs of distributing FAVORABLE MARKET CONDITIONS separate bank cards to poor customers, who typically do not The growth of M-PESA is a testament to Safaricom’s vision generate significant profits for banks. The mobile phone can and execution capacity. But Safaricom also benefited from also be used as a point of sale terminal to initiate financial launching the service under several enabling conditions for transactions and securely communicate with the appropriate successful deployment of a mobile money service. These server to request transaction authorization, thus obviating the include strong latent demand for domestic remittances, need to deploy costly dedicated devices in retail environments. poor quality of available financial services, a banking regu- lator that permitted Safaricom to experiment with different A usage-based model for reaching poor customers business models and distribution channels, a mobile com- with financial services. Because banks make most of munications market characterized by Safaricom’s dominant their profits by collecting and reinvesting deposits, they tend market position and low commissions on airtime sales, and to distinguish between profitable and unprofitable cus- a reasonable amount of banking infrastructure. tomers based on the likely size of their account balances and their ability to absorb credit. Banks thus find it difficult to Strong latent demand for domestic remittances serve poor customers because the revenue from reinvest- ing small-value deposits is unlikely to offset the cost of Safaricom based its launch of the M-PESA service on the serving these customers. In contrast, mobile operators in very brief, but powerful, phrase: “send money home.” In CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 357 this, it capitalized on the fact that demand for domestic to mobile payments—an extensive and efficient semiformal remittance services is greater in locations where migration retail network of pawnshops offering domestic remittance has occurred, separating families, particularly when the services at commissions of 3 percent. breadwinner moves to an urban center and the rest of the In Kenya, by comparison, the most common channel for family remains home. This was certainly the case in Kenya, sending money before M-PESA was informal bus and where 17 percent of households depended on remittances as matatu (shared taxi) companies. Because these companies their primary income source as of 2006 (FSDT 2007a). are not licensed to transfer money, there is considerable risk A recent study of M-PESA suggests that latent demand that the money will not reach its final destination. Mean- for domestic remittances is related to urbanization ratios while, Kenya Post, Kenya’s major formal remittance (Ratan 2008), and that the most propitious domestic remit- provider, is perceived by customers as costly, slow, and tances markets are those in which the process of rural- prone to liquidity shortages at rural outlets. M-PESA’s pop- urban migration is sufficiently rooted to produce large ularity was also bolstered by the fact that Kenya’s bank migration flows but not so advanced that rural communi- branch infrastructure (currently, there are 840 branches) is ties are hollowed out. The study also finds that countries far too sparse to compete with M-PESA’s 16,900 cash- with mid-range urbanization ratios (20 to 40 percent), espe- in/cash-out outlets. Figure 20.6 illustrates how Kenyan cially those that are urbanizing at a rapid rate, are likely to households sent money before and after M-PESA. Of note is exhibit strong rural-urban ties requiring substantial transfer the dramatic reduction in the use of informal bus systems of value. This is the case in many African countries, such as and Kenya Post to transfer money between 2006 and 2009. Kenya and Tanzania, where the share of the population liv- As noted, M-PESA’s early adopters were primarily ing in urban areas as of 2008 was 22 percent and 26 percent, banked customers, suggesting that M-PESA did not acquire respectively, according to the World Bank. On the other its initial critical mass through competition with the formal hand, where urbanization ratios exceed 50 percent, such as sector but rather as a complement to formal services for in the Philippines and several Latin American countries, clients who were wealthier, more exposed to formal finan- remittances are likely to be more closely linked to interna- cial service options, and less risk averse. As M-PESA services tional rather than domestic migration patterns. move deeper into the Kenyan market, however, unbanked Further, the study shows that in locations where entire users are increasingly driving M-PESA’s expansion because nuclear families move, remittances are strongest when there of the competitive advantages of mobile banking offers over is cultural pressure to retain a connection with one’s ances- other options. This is one reason why Africa, with its high tral village. In Kenya, migrants’ ties with rural homes are proportion of unbanked people, is seen as such a promising reinforced by an ethnic (rather than a national) concept of market for mobile money applications. citizenship. These links are expressed through burial, inher- itance, cross-generational, social insurance, and other ties, A supportive banking regulator even in cases where migrants reside more or less perma- nently in cities.9 In countries where migrants have a Regulation of mobile money can help secure trust in new stronger connection to national as opposed to local or eth- mobile money schemes. At the same time, regulation may nic identity, rural to urban migration may have diminished constrain the deployment of a mobile money application by the significance of the rural “home” and hence dampened limiting the scheme operator’s freedom in structuring the domestic remittance flows. business model, service proposition, and distribution chan- nels. In the case of M-PESA in Kenya, Safaricom had a good working relationship with the central bank and was given Poor quality of existing alternatives regulatory space to design M-PESA in a manner that fit its Demand for mobile e-payments must be examined in the market. Together, the Central Bank of Kenya and Safaricom context of the accessibility and quality of alternative pay- worked out a model that provided sufficient prudential ment methods. If there are many good alternatives to comfort to the central bank. mobile payments, as is typically the case in developed coun- The Central Bank of Kenya insisted that all customer tries, it is difficult to convince users to switch to the new ser- funds be deposited in a regulated financial institution and vice. In the Philippines, for example, the G-Cash and Smart reviewed the security features of the technology platform, Money mobile payment services experienced low take-up in but it also allowed Safaricom to operate M-PESA as a pay- part because of the availability of a competitive alternative ments system outside the provisions of the banking law.10 358 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA Figure 20.6 Money Transfer Behavior before and after M-PESA 50 45 40 35 30 Percent 25 20 15 10 5 0 A d s ce sit r e e er Bu sfe qu els an ES th fi po of H he an -P O ne de tr M C st eo Po ct ey m ire on So D M 2009 2006 Source: FSDT 2007a, 2009b. Safaricom has paid a certain price for this arrangement. For be converted into cash-in/cash-out agents, stronger brand instance, interest earned on deposited balances must go to a recognition and trust among potential customers, and not-for-profit trust and cannot be appropriated by Safari- larger budgets to finance the heavy up-front investments com or passed on to customers. To address anti-money- needed to deploy a new service. With a market share of laundering concerns, there are also limits on the size of around 80 percent, Safaricom enjoyed all of these benefits M-PESA transactions. Fundamentally, however, Safaricom when it launched M-PESA. was able to design the M-PESA service without having to Successful deployment of a mobile money application contort its business model to fit within a prescribed regula- also has a greater chance of success in countries where the tory model. commissions that mobile operators pay airtime resellers are The Central Bank of Kenya has continued to support relatively low. If commissions are too high, resellers will not M-PESA’s development, even in the face of pressure from be attracted by the incipient cash-in/cash-out business. In banks. In late 2008, following a lobbying effort by the bank- Safaricom’s case, airtime commissions total 6 percent, of ing industry to shut down the service, the Central Bank of which 5 percent is passed on to the retail store. A commis- Kenya performed an audit of the M-PESA service at the sion of 1–2 percent on a cash-in/cash-out transaction is request of the Ministry of Finance and declared it safe and plausibly attractive—the store need only believe that the in line with the country’s objectives for financial inclusion volume of the cash business will be five times the size of (see Okoth 2009 for information). Thus far, the central bank the airtime business. This seems reasonable, considering appears justified in its confidence in M-PESA—there have that the bulk of airtime sales are of low denominations been no reports of major fraud. Although system downtime (around $0.25). remains frequent, it has not been catastrophic. A reasonable base of banking infrastructure A dominant mobile operator and Finally, the ability of M-PESA stores to convert cash to low airtime commissions e-value for customers depends on how easily they can rebal- The chances of a mobile money scheme taking root also ance their liquidity. If bank branch penetration is too low, depend on the strength of the mobile operator within its rebalancing will be more difficult to achieve, because the market, because a large market share is associated with a agent channel is forced to develop alternative cash transport larger potential customer base for cross-selling the mobile mechanisms. Thus, a mobile payment agent network money service, a larger network of airtime resellers that can must have at its disposal at least a minimal retail banking CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 359 infrastructure. There appears to be a branch penetration years later. Although people have proved creative in using “sweet spot” for mobile money, where penetration is not so M-PESA for their own needs, sending money home con- high that it hampers demand for mobile money services but tinues to be one of the most important uses, and the num- not so low that agents are unable to manage their liquidity. ber of Kenyan households receiving money transfers has Because of the branch networks of Equity Bank and other increased from 17 percent to 52 percent since M-PESA was banks and microfinance institutions, Kenya is reasonably introduced (FSDT 2009a). well supplied with rural liquidity points. Even so, shortage of cash or electronic value for M-PESA agents is a problem in both rural and urban areas in Kenya. Mobile payment A simple user interface operators in some other Sub-Saharan African countries face The simplicity of M-PESA’s message is matched by the sim- more serious liquidity constraints, especially in rural areas. plicity of its user interface. The service can be launched Such constraints are likely to be a major factor affecting the right from the main menu of a mobile phone, making it success of mobile services in specific country contexts. easy for users to find. And because the service resides on the phone and does not need to be downloaded from the net- work each time it is used, the menu loads very quickly and M-PESA’S SERVICE DESIGN: GETTING prompts the user for information step-by-step. For instance, PEOPLE ONTO THE SYSTEM for a P2P transfer, the user is asked to enter the destination While M-PESA’s rapid growth was fueled by certain country- phone number, the amount of the transfer, and the personal specific enabling conditions, the success of such an innova- identification number (PIN) of the sender. Once the infor- tive service hinged on the design of the service. Conducting mation is gathered, it is fed back to the customer for final financial transactions through a mobile phone is not an confirmation. Once the customer hits OK on a mobile intuitive notion for many people, just as walking to a cor- phone, the data is sent to the M-PESA server in a single text ner shop to make cash deposits and withdrawals may not message. Consolidating all information into a single mes- at first seem natural to many. To overcome this adoption sage reduces messaging costs as well as the risk that only barrier, Safaricom designed M-PESA in a way that helped part of the transaction data will be sent to the server. A final people immediately grasp how they might benefit from the advantage is that the application can use the security keys in service; removed barriers that might prevent people from the user’s SIM card to encrypt messages end-to-end, from experimenting with the service; and fostered trust in retail the user’s handset to Safaricom’s M-PESA server. outlets that would be tasked with promoting the service, registering customers, and facilitating cash-in/cash-out Removing adoption barriers: free to register, services. free to deposit, no minimum balances Safaricom designed M-PESA to make it as easy as possible A simple message targeting a big point for customers to try the service. Customer registration, of concern among the population which can be done at any M-PESA retail outlet, is quick, In very early phases of conception, Vodafone developers simple, and free. First, the retail clerk provides a paper reg- thought that M-PESA would be used as a way for istration form, where the customer enters his or her name, customers to repay microloans. However, as Safaricom ID number (from Kenyan national ID, passport, military ID, market-tested the mobile money proposition, they shifted diplomatic ID, or alien ID), date of birth, occupation, and the core proposition from loan repayment to helping people mobile phone number. The clerk then checks the ID and make P2P transfers to their friends and family. In its com- inputs the registration information into the customer’s mercial launch, M-PESA was marketed to the public with mobile phone. SIM cards in Kenya are now preloaded with just three powerful words: “send money home.” In an envi- the M-PESA application. If the customer’s SIM card is too ronment in which families were geographically split, this old and does not contain the application, the clerk replaces message tapped into a major concern for many Kenyans— it. The customer’s phone number is not changed even if the the risks and high costs associated with sending money over SIM card is. long distances. Thus, a simple marketing message turned a After Safaricom receives the application, it sends both the basic “e-remittance” product into a must-have “killer” customer and the retail outlet an SMS confirming the trans- application and remains the main marketing message three action. The SMS provides the customer a four-digit start key 360 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA (one-time password), which they use to activate their strong corporate brand. As the mobile operator in Kenya account. Customers enter the start key and their ID number, with the dominant share of the market (more than 80 per- after which they are asked to input a secret PIN of their cent at M-PESA’s launch and almost as much in 2010), choice. This completes the registration process. In addition Safaricom was already a broadly respected and trusted to leading customers through this registration process, retail brand, even among low-income customers. M-PESA retail clerks explain how to use the application and discuss the outlets are required to paint their store “Safaricom green,” a costs associated with each service. This customer support tactic that not only gives customers confidence that the store early in the process is particularly important in rural areas, is acting on behalf of Safaricom but also makes it easier for where a significant percentage of the potential user base is customers to locate cash-in/cash-out points. unfamiliar with the functioning of mobile phones. Second, by investing heavily in store training and on-site The minimum M-PESA deposit amount is approxi- supervision, Safaricom ensured that customers can have a mately $1.25, but there is no minimum balance require- remarkably similar experience in any retail authorized out- ment. And because customers deposit money for free, there let they use. This “sameness” has helped to build trust in is no immediate barrier to taking up the service. M-PESA both the platform and the outlets and gives customers a charges customers only for “doing something” with their consistently positive view of the service. Rather than relying money, such as making a transfer, withdrawing money, or on its channel intermediaries to carry out these functions in buying prepaid airtime. retail shops, Safaricom chose to centralize the functions through a single third-party vendor, Top Image. Quality is maintained through a rating process. A Top Image repre- The ability to send money to anyone sentative visits each outlet at least once per month and rates M-PESA customers can send money to any GSM mobile each store on a variety of criteria, including visibility of phone subscriber on the Safaricom, Zain, Orange, or YU branding and the tariff poster, availability of cash and M- networks in Kenya, regardless of whether the receiving party PESA electronic value to accommodate customer transac- is an M-PESA customer. When a transfer is sent, money is tions, and quality of record-keeping. debited from the sender’s account and the recipient receives Third, the fact that M-PESA customers receive instant a code by SMS, which is used to claim the monetary value at SMS confirmation of their transaction has helped them any M-PESA store. M-PESA is thus an account-to-cash ser- learn by experience to trust the system. Because the receipt vice, with the receiver’s experience being similar to the way confirming a money transfer includes the name and num- Western Union works today. M-PESA’s pricing, however, is ber of the recipient and the amount transferred, it allows quite different: customers pay a higher (roughly triple) P2P the sender to confirm instantly that the money was sent to charge when sending money to a non-M-PESA customer, the right person—the most common source of error. In the but at the other end the noncustomer is not charged to case of an error, the receipt can then be used to resolve the receive the cash, whereas registered customers pay a cash- situation. out fee of at least $0.30. Safaricom developed this pricing Fourth, Safaricom requires its outlets to record all cash- scheme with the understanding that the sender has power in/cash-out transactions in a paper-based, Safaricom- over the recipient, and so it chose to put pressure on the created logbook. For each transaction, the store clerk enters sender to require the recipient to register with M-PESA. the M-PESA balance, the date, agent ID, transaction ID, Furthermore, Safaricom hoped that providing noncus- transaction type (customer deposit or withdrawal), value, tomers with a good, no-cost first experience with M-PESA customer phone number, customer name, and the customer’s would lure them to register for M-PESA. national ID number. Customers are then asked to sign the log for each transaction, which not only discourages fraud but also gives agents a way to offer first-line customer care Building trust in the retail network for customers querying previous transactions. Each entry in Recognizing that M-PESA would not be rapidly adopted by the log is written in triplicate. The top copy is kept by the the Kenyan population unless customers had enough trust retail outlet for its own records, a second is passed on to the in the M-PESA retail network that they were willing to store’s master agent, and the third is sent to Safaricom. conduct cash-in/cash-out transactions through those out- Because all information contained in the logbook (except lets, Safaricom employed several measures to build that for the customer signature) is captured electronically by trust. First, it closely linked the M-PESA brand with its own Safaricom when the transaction is made and is available to CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 361 the master agents through a Web management system, the (through their master agents) rather than from customers. main purpose of the agent log is not for record-keeping but This arrangement reduces the potential for agent abuses. rather to provide comfort to customers who are accustomed Customer fees are uniform nationwide, and they are promi- to having transactions recorded on paper. nently posted in all outlet locations in the poster shown in figure 20.7. M-PESA chose to specify its fees in fixed currency terms Simple and transparent pricing rather than as a percentage of the transaction. This makes it M-PESA pricing is transparent and predictable. There are easier for customers to understand the precise cost of each no customer charges for the SMSs that deliver the service. transaction and helps them think of the fee in terms of the Instead, fees are applied to the actual, customer-initiated transaction’s absolute value (for example, sending money to transactions. All customer fees are subtracted from the cus- grandmother). It also helps them compare the transaction tomer’s account, and outlets cannot charge any direct fees. cost against alternative and usually costlier money-transfer Thus, outlets collect their commissions from Safaricom arrangements (for example, the matatu fare plus travel time). Figure 20.7 M-PESA Tariff Structure Transaction type Transaction range (KShs) Customer charge Minimum Maximum (KShs) Value movement transactions Deposit cash 100 35,000 0 Send money to a registered M-PESA user 100 35,000 30 100 2,500 75 2,501 5,000 100 Send money to a nonregistered M-PESA user 5,001 10,000 175 10,001 20,000 350 20,001 35,000 400 100 2,500 25 2,501 5,000 45 Withdraw cash by a registered M-PESA user at an M-PESA agent outlet 5,001 10,000 75 10,001 20,000 145 20,001 35,000 170 200 2,500 30 2,501 5,000 60 Withdraw cash by registered M-PESA user at PesaPoint ATM 5,001 10,000 100 10,001 20,000 175 Withdraw cash by a non-registered M-PESA user 100 35,000 0 Buy airtime (for self or other) 20 10,000 0 Pay Bill Transactions - - 0–30 Information transactions Show balance 1 Change secret word 0 Change PIN 20 Update menu 0 Change language 0 SIM replacement 20 Source: Safaricom. 362 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA M-PESA deposits are free to customers. Withdrawals of M-PESA’s liquidity system is not without its challenges, less than $33 cost around $0.33. Withdrawal charges are however. Because of cash flow constraints, M-PESA retail “banded” (that is, larger transactions incur a larger cost) so outlets cannot always meet requests for withdrawals, espe- as not to discourage smaller transactions. ATM withdrawals cially large ones. Furthermore, the agent commission using M-PESA are slightly more expensive than withdrawals structure discourages outlets from handling large transac- at a retail outlet ($0.40 versus $0.33). tions. As a result, customers are sometimes forced to spread P2P transfers using M-PESA cost a flat rate of around large withdrawals over several days rather than withdraw a $0.40. These transfers are where Safaricom makes the bulk lump sum, at an added cost and inconvenience. Cash flow of its revenue. Thus, for a purely electronic transfer, cus- constraints also undermine customer trust in M-PESA as a tomers pay more than double the price they pay for the mechanism for high-balance, long-term saving. Use of bank average cash transaction ($0.17), even though the cost of branches and ATMs to give customers a sort of liquidity providing purely electronic transactions is lower than cash mechanism of last resort has bolstered the credibility of the transactions. This model reflects a notion of optimal pricing M-PESA system, however, that is based less on cost and more on customer willingness to pay. M-PESA is still cheaper, though, than the other avail- M-PESA’S EXECUTION STRATEGY: QUICKLY able mechanisms for making remote payments, such as REACHING CRITICAL MASS money transfer by the bus companies, Kenya Post’s Postapay, or Western Union.11 Although strong services design has been a major factor in Notably, M-PESA has maintained the same pricing for the success of M-PESA, an appropriate execution strategy transactions in its first three years of operation. This strat- has also been a key factor. Importantly, Safaricom recog- egy has helped establish customer familiarity with the ser- nized that it would be difficult to scale M-PESA incremen- vice. Safaricom has changed, however, the price of two tally, because it had to overcome three significant obstacles types of customer requests that do not involve a financial common to any new electronic payment system, namely, transaction: balance inquiries (because the initial low price adverse network effects, the chicken-and-egg trap, and generated a burdensome volume of requests) and PIN trust. First, in regard to adverse network effects, the value to changes (because customers are far more likely to remem- the customer of a payment system depends on the number ber their PIN if the fee to change it is higher). The volume of people connected to and actively using it. The more peo- of both types of requests decreased substantially after ple on the network, the more useful it becomes.12 While these price changes. As previously noted, the SMS confir- network effects can help a scheme gain momentum once it mation of a transaction contains the available balance, reaches a critical mass of customers, they also make it diffi- which also helps cut down on the number of balance cult to attract early adopters of the new technology. Sec- inquiries. ond, to grow, M-PESA had to attract customers and stores in tandem. It is difficult, however, to attract customers when there are few stores to serve them, and equally hard to Liquidity of last resort at bank convince stores to sell the service when there are few branches and ATMs customers to be had (thus, the chicken-and-egg trap). From very early on, M-PESA signed up banks as agents, so Thus, M-PESA needed to drive both customer and store M-PESA customers could walk into any branch of those acquisition aggressively. Third, a company will be success- banks to conduct cash-in/cash-out transactions. One year ful in attracting customers only when prospective cus- after its launch, M-PESA went further and partnered with tomers have confidence in the reliability of the new system. PesaPoint, one of the largest ATM service providers in In the case of M-PESA, customers had to be comfortable Kenya. The PesaPoint network includes more than 110 with three elements that were new at the time in Kenya: a ATMs scattered in all eight provinces of the country. Cus- payment system operated by a mobile operator, using a tomers can now retrieve money from any PesaPoint ATM. nonbank retail outlet to meet cash-in/cash-out needs, and To do so, they select “ATM withdrawal” from the M-PESA using mobile phones to access account information and menu on their mobile phone, after which they receive a one- initiate transactions. time ATM authorization code. They then enter that code on In the early stages of development of a payments system, the ATM keyboard to make the withdrawal. No bank card is the three problems described above reinforce each other, needed for the transaction. creating a significant hurdle to growth. In many cases, this CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 363 hurdle helps explain why many other mobile money deploy- aimed at the poor. Over time, the marketing moved from ments remain subscale. M-PESA, however, overcame this young, up-market urban dwellers with desk jobs to more hurdle through very forceful execution on two key fronts: ordinary Kenyans with lower-paid professions. Safaricom made significant up-front investments in build- M-PESA’s launch was associated with significant up- ing a strong service brand for M-PESA, and it effectively front investment in television and radio marketing,14 but leveraged its extensive network of airtime resellers to build there was also intense outreach through road shows, in a reliable, consistent retail network that served customers’ which company agents traveled around the country signing liquidity needs. people up, explaining the product, and demonstrating how to use it. Over time, as people became more familiar with M-PESA, it was no longer necessary to do this kind of Aggressive up-front investment in hands-on outreach. Television and radio marketing was promoting the M-PESA brand largely replaced by the omnipresent M-PESA branding at From the beginning, Safaricom sought to foster customer all outlets, supported by a few large billboards. Newer ads trust in M-PESA and relied on existing customers to be the feature a general emotional appeal, with a wider range of prime mechanism for drawing in new customers. The task services indicated. was all the more difficult because Safaricom was not only introducing a new product, but an entirely new product category, to a market that had little experience with formal A scalable distribution channel financial services. Safaricom’s initial target for M-PESA From the time Safaricom launched M-PESA, it under- was about 1 million customers within one year of the stood that the primary role of the mobile phone is to launch of the service, equal to 17 percent of Safaricom’s enable the creation of a retail outlet–based channel for customer base of about 6 million customers at the time, cash-to-digital value conversion. It also understood that according to Safaricom company results for the year end- for this cash-to-digital conversion to be broadly available ing March 2007. to the bulk of the population, Safaricom had to develop a channel structure that could support thousands of National launch at scale. After small pilots involving M-PESA stores spread across a broad geographical area. less than 500 customers,13 M-PESA launched nationwide, To achieve this, Safaricom built four elements into its increasing the likelihood that the service could reach a channel management execution strategy: engaging inter- critical mass of customers in a short time frame. At launch, mediaries to help manage the individual stores, thereby Safaricom had 750 stores that covered all of Kenya’s 69 dis- reducing the number of direct contacts it had to deal trict headquarters. The launch was a massive logistical with, ensuring that outlets received sufficient incentives challenge that led to a great deal of customer and store con- to actively promote the service, maintaining tight control fusion and, during the first few months, delays of several over the customer experience, and developing several dif- days in reaching customer service hotlines. User and store ferent methods for stores to rebalance their stocks of cash errors were frequent since everyone was new to the service. and e-value. But the gamble paid off. Logistical problems subsided after a few months, leaving strong brand recognition and top- Two-tier channel management structure. Safaricom of-mind awareness among large segments of the popula- created a two-tier structure with individual stores (sub- tion. The service outran first-year growth targets, quickly agents, in Safaricom’s parlance) that depends on master turning network effects in its favor as new customers begat agents (referred to by Safaricom as agent head offices). more customers and turned M-PESA into a compelling Agent head offices maintain all contact with Safaricom, and business proposition for more stores. perform two key functions: liquidity management (buying and selling M-PESA balance from Safaricom and making it An appropriate marketing mix. Initial M-PESA market- available to individual stores under their responsibility), ing featured and targeted urban, relatively wealthy city and distribution of agent commissions (collecting the com- dwellers with a need to “send money home.” The choice mission from Safaricom based on the overall performance of this demographic as the initial customer created an aspi- of the stores under them and remunerating each store). rational image for prospective customers and avoided giv- Individual stores are either directly owned by an agent head ing the impression that M-PESA was a low-value product office or working for one under contract. 364 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA Giving stores incentives. Retail outlets will not maintain elements of the customer experience: store selection, train- sufficient stocks of cash and electronic money unless they ing, and supervision. are adequately compensated for doing so. Hence, Safaricom pays commissions to agent head offices for each cash- Developing multiple store liquidity management in/cash-out transaction conducted by stores under their methods. By far the biggest challenge faced by M-PESA responsibility. Although Safaricom does not prescribe the stores is maintaining enough liquidity, in terms of both commission split between agent head offices and stores, cash and e-float,16 to be able to meet customer requests. If most agent head offices pass on 70 percent of commissions they take too many cash deposits, stores will find them- to the store.15 For deposits under $33, Safaricom pays a selves running out of e-float with which to facilitate fur- $0.133 commission, of which $0.074 goes to the store after ther deposits. If they do too many withdrawals, on the tax. For withdrawals, Safaricom pays $0.200, of which other hand, they will accumulate e-float but will run out of $0.111 goes to the store. So, assuming equal volumes of cash. Thus, stores frequently undertake liquidity manage- deposits and withdrawals, the store earns $0.092 per trans- ment efforts. action. Assuming the store conducts 60 transactions per day, The M-PESA channel management structure was con- it earns around $5.50—almost twice the prevailing daily ceived to offer stores three methods for managing liquid- wage for a clerk in Kenya. ity. Two of these place the agent head office in a central Recall that Safaricom charges customers $0.333 on a role, with the expectation that it will “recycle” e-float round-trip savings transaction (free deposit plus $0.33 for between locations experiencing net cash withdrawals (that the withdrawal), which is in fact equal to what the channel is, accumulating e-float) and locations with net cash receives ($0.13 on the deposit plus $0.20 for the withdrawal). deposits (that is, accumulating cash). In the first method, So, assuming equal volumes of deposits and withdrawals, the agent head office provides direct cash support to stores. Safaricom does not make money on cash transactions. It The store clerk comes to the agent head office to deliver or merely “advances” commissions to the channel when cus- offload cash, or the agent head office sends cash runners to tomers make deposits and recoups them when customers the store to perform these functions. withdraw. As noted, Safaricom generates the bulk of its rev- In the second method, the agent head office and stores enue from services for which customer willingness to pay is use their respective bank accounts. If the store has excess the highest—electronic, P2P transactions. cash and wants to buy M-PESA e-float from the agent head Because store revenues are dependent on the number of office, the store deposits the cash into the account of the transactions they facilitate, Safaricom has been careful not agent head office at the nearest bank branch or ATM. Once to flood the market with too many outlets, lest it depress the the agent head office confirms receipt of the funds into its number of customers per agent. Instead, it has maintained account, it transfers M-PESA e-float to the store’s M-PESA balanced growth in the number of outlets relative to the account. If the store wants to sell e-float to obtain cash, the number of active customers, resulting in an incentivized store transfers M-PESA e-float to the agent head office. The and committed agent base. agent head office then deposits (or transfers) money into the store’s bank account, after which the store can withdraw Maintaining tight control over the customer the cash at the nearest branch or ATM. experience. Safaricom also recognized early on that cus- In the third method, stores interact directly with a bank tomers need to have a good experience at the retail stores that has registered as an M-PESA “superagent.” Under this offering M-PESA services, where the bulk of transactions method, the agent head office does not get involved in liq- take place. To ensure that it maintained control over the uidity management. Instead, stores open an account with a customer experience, it has not relied on agent head offices participating superagent bank. To rebalance their cash, to perform all channel management functions. Instead, it stores deposit and withdraw cash against their bank account concentrated the evaluation, training, and on-site supervi- at the nearest branch or ATM of the bank. The store then sion of stores in a single outsourcing partner, Top Image. electronically buys and sells e-float in real time against its The more routine and non-customer-facing store support bank account. From a store’s perspective, one drawback of activities, such as liquidity management and distribution of using a bank-based superagent mechanism is that it can store commissions, are left to a large pool of agent head only be used during banking business hours. offices. Through its contract with Top Image, however, In all cases, the e-float–cash nexus will remain the key Safaricom retained direct, centralized control over key constraint to further development of M-PESA, because it CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 365 requires the physical movement of cash around the country for example, is approximately 7 percent of the amount of and is thus the least scalable part of the system. the transaction ($0.40 for the transfer plus $0.33 for the withdrawal). Adjusting M-PESA’s current pricing model to account for smaller-denomination transactions would have M-PESA’S FUTURE EVOLUTION two advantages. First, it would make the service more acces- The experience of M-PESA demonstrates how powerful a sible to the poor, for whom pricing is now too high given payment network that offers convenience at an affordable their transactional needs. Such a reduction would allow cost can be once a critical mass of customers is reached. It also Safaricom to maintain customer growth once saturation shows that achieving critical mass requires a service design starts to set in at current pricing. Second, a pricing adjust- that removes as many adoption barriers as possible, together ment would allow customers to use M-PESA for their daily with significant investment in marketing, branding, and transaction needs, and in particular to save on a daily basis, agent network management. The Kenyan experience also sug- which would be beneficial to those who are paid daily. gests that several country-level factors must be aligned to set A reduction in customer prices could come about in sev- the scene for successful mobile money development, eral ways. There is room for “tranching” the P2P fee of including the labor market profile (demand for remittances $0.40, for example, so that the price of smaller, or more fre- generated by rural-urban migration), the quality of available quent, transactions becomes more affordable. For cash financial services, support from the banking regulator, and transactions, one way to enable lower fees would be to the structure of the mobile communications market (domi- establish street-level M-PESA subagents who would offer nant mobile operator and low airtime commissions). lower costs and commissions than store-based agents. Sub- While M-PESA has been more successful than anyone agents would be a kind of “e-susu collector,” operating with could have imagined at its launch, the model still has sub- small working capital in order to aggregate small customer stantial room for development. A threefold wish list for transactions. Subagents would use normal M-PESA retail M-PESA could be delineated as follows: further main- outlets to rebalance their cash and M-PESA stored value. streaming of M-PESA’s regulatory treatment, pricing that The key principle here is that segmentation of customers opens up a much larger market of microtransactions, and would go hand-in-hand with segmentation of agents. expanding M-PESA so that customers have access to a broader range of financial services. Linking with banks and other institutions While some customers use M-PESA as a saving device, the Mainstreaming M-PESA’s regulatory treatment service still falls short of being a useful method of saving M-PESA’s regulatory treatment as a payments vehicle needs for most poor people. The fact that the average balance of to be formalized so that it can become regulated in the most M-PESA accounts was less than $3 in early 2009 is partly a appropriate way. To this end, the Central Bank of Kenya is “large number” problem: if 900,000 people used M-PESA backing a new payments law that would cover M-PESA trans- to save, that means 10 percent of users use the service to actions (as of this writing, the draft had not yet been save, and that the average savings balance is diluted because approved by the Kenyan parliament). The Central Bank of it takes into account all M-PESA users rather than only Kenya is also in the process of finalizing agent banking regu- users who save. But the fundamental problem is that there lations that would allow commercial banks to use retail out- is still a lot of conversion of electronic value back into cash. lets as a delivery channel for financial services. Banks are quite This can be attributed to a combination of factors: reasonably complaining that they could not replicate the M-PESA service themselves because they are not currently ■ Lack of marketing. Safaricom does not want to publicly allowed to undertake customer transactions through agent promote using M-PESA as a saving tool for fear of pro- networks on their own. Allowing both banks and M-PESA to voking the Central Bank Kenya to regulate it more operate such agent networks would level the playing field. tightly. ■ Customer pricing. The flat fee of around $0.33 for with- drawals under $33 means that small withdrawals carry a Pricing that enables smaller payments large relative fee. M-PESA’s current pricing model is not conducive to small ■ Product design. M-PESA works very much like an elec- transactions. The fee for a $10 P2P transfer plus withdrawal, tronic checking account and does not offer structured 366 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA saving products that may help people build discipline times at the branch, a trip to the bank and back may take around savings. two hours—approximately a quarter of your working day. A ■ Inflation. M-PESA does not pay interest. In an environ- bus fare of only $0.50 to get to the bank may well represent ment with 15 percent inflation (during its first full year one-quarter of your income on a good day. With the bank of operation in 2008), saving may be too onerous for fees included, each banking transaction costs you the equiv- much of the Kenyan population. alent of almost half a day’s wages. It would be like charging ■ Trust. M-PESA deposits are not supervised by the Cen- someone with an average income in the United States tral Bank of Kenya. And unlike payments, where trust something like $50 for each ATM transaction. Then, imag- can be validated experientially in real time, saving ine a world without credit instruments or electronic pay- requires garnering the trust of customers over a longer ments. No checks, no credit cards, no money orders, no period of time. direct debits, no Internet banking. All your transactions are ■ Privacy. People may want more privacy for their saving done in cash or, worse, by bartering goods. All exchanges behavior than an agent provides. are physical, person-to-person, hand-to-hand. Consider ■ Excess liquidity. The 16,000 M-PESA cash-in points in the hassle and the risk of sending money to distant rela- Kenya are also 16,000 cash-out points. The ubiquity of tives, business partners, or banks. M-PESA agents may make it too easy for customers to How would you operate in such a world? A recent book, cash out their funds, thus limiting their ability to accu- Portfolios of the Poor, documents how poor people cope mulate large balances. (Collins et al. 2009). Some people save to “push” excess money from today to tomorrow, some people borrow to Rather than expecting Safaricom to develop and market “pull” tomorrow’s money to fund necessary expenses today. more comprehensive savings services, M-PESA should sup- They store cash in the home to meet daily needs, they leave port saving by linking to banks. M-PESA could then become it with a trusted friend for emergencies, they buy jewelry a massive transaction acquisition network for banks rather because that represents a future for their children, they pile than an alternative to them. That said, Safaricom is begin- up bricks for the day when they can build an extra room in ning to connect with banks. In September 2009, for exam- their house. They make contributions to a savings group ple, Family Bank and M-PESA established a connection with a circle of friends to build up a pot of money, and one allowing customers to transfer money from M-PESA to day it will be their turn to take that pot home to buy new their Family Bank account using M-PESA’s bill pay func- clothes. They borrow from friends, seek advances from their tion. This connection follows a successful pilot of loan employers, pawn their jewelry, and borrow from a high- repayments via M-PESA’s bill pay function. interest moneylender. M-PESA would also benefit from establishing further Lack of good financial options is undoubtedly one of the links with institutions beyond banks, such as billers, dis- reasons why poor people are trapped in poverty. In many tributors, and employers. By promoting M-PESA as a cases, poor people cannot sustain themselves or aspire to mechanism for distributing salaries and social welfare pay- earn higher incomes because they are not able to invest in ments, enabling payments across supply chains, and paying better farming tools and seeds to enhance their productiv- bills, the need for cash-in and cash-out services would be ity, start a microenterprise, or even take the time to search minimized, and, as a result, a key component of transaction for better-paying employment opportunities. Their income costs could be reduced. Savings balances may also be higher is volatile, often fluctuating daily, and without reliable ways if people received payments directly into their accounts of pushing and pulling money between good days and bad rather than in cash, and if they had other useful things to days, they face stark choices such as pulling their children do with their money in electronic form. out of school or putting less food on the table during bad patches. Without good financial tools, they also may not be able to cope with shocks that periodically set them back. CONCLUDING THOUGHTS: HOW M-PESA Most of these shocks are foreseeable, if not entirely pre- CAN REINVIGORATE FINANCIAL dictable: a drought, ill health, and lifecycle events such as INCLUSION EFFORTS marriage and death. Imagine a world where there are no banks where you live. Cash is the main barrier to financial inclusion. As long as The nearest branch is 10 kilometers away, and it takes you poor people are able to exchange value only in cash—or almost an hour to get there by foot and bus. With waiting worse, physical goods—they will remain too costly for CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 367 formal financial institutions to address in significant 10. Although the Central Bank of Kenya Act was amended numbers. Few banks are willing to build the costly infra- in 2003 to give the central bank broad oversight of pay- structure necessary for collecting low-value cash deposits ment systems, the operational modalities for this oversight and redeeming savings back into small sums of cash in have not been implemented; they are pending approval of low-income or rural areas. But once poor people have access a new National Payments System Bill that has languished to cost-effective electronic means of payments such as in parliament. M-PESA, they could, in principle, become profitable to 11. Morawczynski and Pickens (2009) find that sending financial institutions. Although M-PESA itself does not K Sh 1,000 through M-PESA is 27 percent cheaper than the post office’s PostaPay and 68 percent cheaper than sending constitute financial inclusion, it does provide a glimpse of a it by a bus company. commercially sound, affordable, and effective way to offer 12. Network effects are commonly illustrated with reference financial services to all. to fax machines: the first set of people who bought a fax machine did not find them very useful, because they were NOTES not able to send faxes to many people. But as more people bought fax machines, everyone’s faxes became increasingly 1. See Hughes and Lonie (2009) for a historical account of useful. Network effects are sometimes called demand-side the M-PESA service, Mas and Morawczynski (2009) for a economies of scale to emphasize that scale affects the value fuller description of the service, and Mas and Ng’weno of the service to each customer. This distinguishes it from (2009) for the latest accomplishments of M-PESA. supply-side economies of scale, which refer to situations in 2. A SIM card is a smart card found inside mobile phones which the average cost per customer fall as volume increases. that are based on the Global System for Mobile communi- Davidson (2009) discusses implications of network effects cations (GSM) family of protocols. The SIM card contains for mobile money. encryption keys, secures the user’s persona identification 13. Safaricom’s earliest pilot project conducted in 2004–05 number on entry, and drives the phone’s menu. SMS is a revolved around microloan repayments and involved the data messaging channel available on GSM phones. Commercial Bank of Africa, Vodafone, Faulu Kenya, and 3. These amounts use an exchange rate of $1 to 75 Kenyan MicroSave. shillings. 14. In a survey of 1,210 users in late 2008, 70 percent of 4. Kenya has a population of nearly 40 million. GDP per respondents claimed they first heard about M-PESA from capita is $1,600, 78 percent of people live in rural areas, and television or radio advertisements (FSDT 2009b). 19 percent of adults have access to a formal bank account. 15. Safaricom would like the split to be 20 percent/80 per- See FSDT (Financial Sector Deepening Trust) (2009a) for cent, meaning that more of the commission is passed on to financial access data derived from the FinAccess survey, a the retail outlet. nationally representative survey of 6,600 households con- 16. E-float is the balance of money that an agent has in his ducted in early 2009. M-PESA account, which he can electronically transfer to 5. The 2009 FinAccess survey (FSDT 2009a, p. 16) con- customers in exchange for cash. firms that 40 percent of adults have used M-PESA. 6. See Safaricom (2009) for key monthly M-PESA statis- REFERENCES tics. Additional figures are taken from Safaricom’s published half-year results for the period ending September 2009 and Camner, Gunnar, and Emil Sjöblom. 2009. “Can the Success of Central Bank of Kenya reports. M-PESA be repeated? A Review of Implementations in 7. Although the number of P2P transactions per customer Kenya and Tanzania.” Valuable Bits. http://mobileactive.org/ has been rising steadily, it remains quite low, probably still files/file_uploads/camner_sjoblom_differences_ke_tz.pdf. less than two transactions per month. Collins, Daryl, Jonathan Morduch, Stuart Rutherford, and 8. For more detailed accounts of the M-PESA story, see Orlanda Ruthven. 2009. Portfolios of the Poor: How the Heyer and Mas (2009) on the country factors that led to World’s Poor Live on $2 a Day. Princeton, NJ: Princeton M-PESA’s success, Mas and Morawczynski (2009) on University Press. M-PESA’s service features, Mas and Ng’weno (2010) on Davidson, Neil. 2009. “Tactics for Tipping Markets: Influ- Safaricom’s execution, and Mas (2009) on the economics ence Perceptions and Expectations.” Mobile Money for underpinning branchless banking systems. the Unbanked Blog, GSM Association, November 15. 9. For a fuller analysis of the use of mobile money for Financial Sector Deepening Trust (FSDT). 2007a. “Financial domestic remittances in Kenya, see Ratan (2008) and Access in Kenya: Results of the 2006 National Survey.” Morawczynski (2008). FSDT, Nairobi. 368 CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA ———. 2007b. “Key Findings of the FinScope Survey in ———. 2009. “Exploring the Usage and Impact of Trans- Tanzania in 2006.” FSDT, Nairobi. formational M-banking: The case of M-PESA in Kenya.” ———. 2009a. “FinAccess National Survey 2009: Dynamics Unpublished paper. of Kenya’s Changing Financial Landscape.” FSDT, Nairobi. Morawczynski, Olga, and Gianluca Miscione. 2008. “Exam- ———. 2009b. “Research on Mobile Payments Experi- ining Trust in Mobile Banking Transactions: The Case of ence: M-PESA in Kenya.” Unpublished paper, FSDT, M-PESA in Kenya.” In Social Dimensions of Information Nairobi. and Communication Technology Policy: Proceedings of the GSM Association. 2009. “Wireless Intelligence Database.” Eighth International Conference on Human Choice and http://www.wirelessintelligence.com. Computers (HCC8), IFIP TC 9, Pretoria, South Africa, September 25–26, 2008. ed. C. Avgerou, M. Smith, and Heyer, Amrik, and Ignacio Mas. 2009. “Seeking Fertile P. van den Besselaar, 287–298 (Boston, MA: Springer). Grounds for Mobile Money.” Unpublished paper. Morawczynski, Olga, and Mark Pickens. 2009. “Poor People Hughes, Nick, and Susie Lonie. 2009. “M-PESA: Mobile Using Mobile Financial Services: Observations on Money for the Unbanked.” Innovations (special edition Customer Usage and Impact from M-PESA.” CGAP for the Mobile World Congress 2009). Brief, CGAP, Washington, DC. http://www.cgap.org/p/ Isaacs, Leon. 2008. IAMTN presentation. MMTA Confer- site/c/template.rc/1.9.36723/. ence, Johannesburg, May. Okoth, Jackson. 2009. “Regulator Gives M-PESA a Clean Bill Jack, William, and Tavneet Suri. 2009. “Mobile Money: the of Health.” The Standard, 27 January. Economics of M-PESA.” Unpublished paper. Okuttah, Mark. 2009. “Safaricom Changes Method of Recruit- Juma, Victor. 2009. “Family Bank Offers New Service Link- ing M-PESA Agents.” Business Daily, December 223. ing Accounts to M-PESA.” Business Daily, December 18. Ratan, A. L. 2008. “Using Technology to Deliver Financial Kimenyi, Mwangi, and Njuguna Ndung’u. 2009. “Expand- Services to Low-Income Households: A Preliminary ing the Financial Services Frontier: Lessons from Mobile Study of Equity Bank and M-PESA Customers in Kenya.” Phone Banking in Kenya.” Washington, DC: Brookings Microsoft Research Technical Report. Institution. Safaricom. 2009. “M-PESA Key Performance Statistics.” Kinyanjui, Kui. 2009. “Yu Launches New Mobile Cash http://www.safaricom.co.ke/fileadmin/template/main/ Transfer Platform.” Business Daily, 16 December. images/MiscUploads/M-PESA%20Statistics.pdf. Mas, Ignacio. 2008a. “M-PESA vs. G-Cash: Accounting for Suri, Tavneet, and William Jack. 2008. “The Performance their Relative Success, and Key Lessons for other Coun- and Impact of M-PESA: Preliminary Evidence from a tries.” Unpublished paper, CGAP, Washington, DC. Household Survey.” Unpublished paper. CHAPTER 20: MOBILE PAYMENTS GO VIRAL M-PESA IN KENYA 369 CHAPTER 21 Independent Power Projects in Sub-Saharan Africa: Determinants of Success Anton Eberhard and Katharine Nawal Gratwick t the beginning of the 1990s, virtually all major A had the potential to benchmark state-owned supply to power generation providers in Sub-Saharan gradually introduce competition (APEC Energy Working Africa were financed by public coffers, including Group 1997). concessionary loans from development finance institutions In 1994 Côte d’Ivoire became one of the first African (DFIs). These publicly financed power generation assets countries to attract a foreign-led IPP to sell power to the were considered one of the core elements of state-owned, national grid under long-term contracts with the state util- vertically integrated power systems. A confluence of fac- ity. Ghana, Kenya, Nigeria, Senegal, Tanzania, and Uganda, tors, however, brought about a significant change in the among others, opened their doors to foreign and local ensuing years. investors in their power sectors shortly thereafter. With insufficient public funds for new power generation Although IPPs were considered part of larger power sec- and decades of poor performance by state-run utilities, tor reform programs in Sub-Saharan Africa, the reforms African countries began to adopt a new model for their were not far-reaching. In most cases, state utilities remained power systems, influenced by pioneering reformers in vertically integrated and maintained a dominant share of Chile, Norway, the United Kingdom, and the United the power generation market, while private investors were States.1 Urged on by multilateral and bilateral develop- allowed to operate only on the margin of the sector.2 Policy ment institutions, which were withdrawing funding from frameworks and regulatory regimes, necessary to maintain a state-owned projects, a number of countries adopted competitive environment, were limited. International com- plans to unbundle their power systems and introduce pri- petitive bids (ICBs) for IPPs were often not conducted vate participation and competition. Independent power because of tight time frames, resulting in limited competi- projects (IPPs)—privately financed, greenfield generation tion for the market and, because of the long-term PPAs, no supported by nonrecourse or limited-recourse loans and competition in the market. These long-term PPAs, along with long-term power purchase agreements (PPAs) with with government guarantees and security arrangements the state utility or another off-taker—thus became a pri- such as escrows and liquidity facilities, exposed countries to ority in overall power sector reform (World Bank 1993; significant exchange rate risks. Finally, while Africa has seen World Bank and USAID 1994). IPPs were considered a continued private participation in greenfield electricity proj- solution to persistent supply constraints, and they also ects, that progress has been erratic, with 2007 representing Anton Eberhard is a professor and Katharine Nawal Gratwick is a PhD graduate at the University of Cape Town Graduate School of Business. 371 the zenith, largely because of the financial close of one large largely state-run utilities, have been developed in Sub- project, Bujagali. Saharan Africa as of early 2010 (table 21.1). In total, about Several factors explain the recent trends in investment four gigawatts (GW) of IPP capacity has been added. With in Africa’s power sector. First, private sector firms were few exceptions,4 these IPPs represent a small fraction of deeply affected by the Asian and subsequent Latin Ameri- total power generation capacity and have mostly comple- can financial crises in the late 1990s and early 2000s. The mented incumbent state-owned utilities. Enron collapse and its aftershocks also influenced U.S.- Nevertheless, IPPs are an important source of new and European-based firms to reduce risk exposure in investment in the power sector in a number of African developing-country markets and to refocus on core activi- countries. In Togo, for example, Centrale Thermique de ties at home. The financial crisis of 2008 and 2009 has also Lomé, the country’s first IPP, will triple the country’s had a toll. Furthermore, DFIs began to reconsider their installed capacity. position of restricted infrastructure investment, a model The majority of IPP contracts in Sub-Saharan Africa that was predominant throughout the 1990s.3 As conces- have been upheld (namely, CIPREL and Azito in Côte sionary funding became available again, many countries d’Ivoire, Takoradi II in Ghana, Iberafrica, Tsavo, OrPower 4, opted for a hybrid solution—part public, part private. and Rabai in Kenya, Afam VI and Aba Integrated in Nigeria, Kenya represents one of the clearest examples of such a and Namanve in Uganda). Although the contracts of the hybrid, with KenGen, the state-owned generator, building two Senegalese projects, GTi Dakar and Kounoune I, remain alongside IPPs, with support from DFIs. largely intact, there are reports of changes in fuel supply Despite this revival of concessionary lending, power arrangements. A number of other IPPs, such as Bui Hydro sector investments have been insufficient in addressing in Ghana, Bujagali in Uganda, and Centrale Thermique de Sub-Saharan Africa’s power needs. Only 25 percent of the Lomé in Togo, have reached financial closure and are under population currently has access to electricity, and poor construction.5 Kenya is in the process of negotiating three supply is the rule, not the exception. The cost of meeting more IPPs after an international competitive tender. Africa’s power sector needs is estimated at $40.8 billion per For all of this progress, however, there have been sev- year, equivalent to 6.35 percent of Africa’s 2005 gross eral high-profile IPP mishaps in Sub-Saharan Africa. Two domestic product (GDP). Approximately two-thirds of the projects—AES Barge in Nigeria and Independent Power spending needed is for capital investment ($26.7 billion Tanzania Limited (IPTL) in Tanzania—are in arbitration. per year), and the remainder for operations and mainte- The costs of Songas, in Tanzania, meanwhile, have nance (O&M). Current spending on power infrastructure escalated as a result of the unplanned, and later disputed, totals approximately $11.6 billion per year. Approximately contracting of IPTL; Songas’s capacity charges were later 80 percent of existing spending is domestically sourced reduced after the government agreed to buy down the from taxes or user charges. The remainder is split among accumulated allowance for funds used during construction official development assistance (ODA) financing (6 percent costs. A dispute about escalating investment costs also of the total); other sources, mainly China (9 percent); and marked the Okpai project in Nigeria. In Kenya changes private sector investment (4 percent). Tackling existing util- may be made in the contracts of OrPower 4, which ity inefficiencies, including system losses, underpricing, reduced its tariff for the second phase of the plant. undercollection of revenue, and overstaffing, would make Another Kenyan project, Westmont, had an initial seven- an additional $8.24 billion available, but a funding gap of year contract that was not renewed. The other early IPP in about $21 billion would still remain (Eberhard and Kenya, IberAfrica, had its contract renewed but with Shkaratan 2010). much lower capacity charges. Closing Africa’s power infrastructure funding gap Following contract changes, IPPs have generally gone on inevitably requires undertaking reforms to reduce or elimi- to make a significant contribution to the countries’ power nate system inefficiencies. This will help existing resources generation—the main exceptions being Westmont, which go farther and create a more attractive investment climate ceased operation, and IPTL, which has operated intermit- for external and private finance, which still has growth tently during its arbitration proceedings. Another high- potential. With the original drivers for market reform still profile failure was the nontransparent procurement process present, future private sector involvement appears inevitable. surrrounding the Richmond/Dowans plant in Tanzania, Approximately 20 grid-connected IPPs, each in excess of which has not been allowed to operate since corruption 40 megawatts (MW) and with long-term PPAs with the charges were filed. Furthermore, there is evidence of stalling 372 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS Table 21.1 General Project Specifications of Sub-Saharan African IPP Projects Country/ Length of Commercial project Size (MW) Fuel/cycle Contract type contract (years) Project tender operating data East Africa Kenya Westmont 46 Kerosene/gas condensate/gas Turbine (barge-mounted) BOO 7 1996 1997 Iberafrica 109a HFO/medium-speed diesel engine BOO 7/15/25 1996,1999, 2008 1997, 2000, 2009 OrPower 4 48 Geothermal BOO 20 1996 2000, 2009 Tsavo 75 HFO/medium-speed diesel engine BOO 20 1995 2001 Rabai 90 HFO BOOT 20 2006 2009 Tanzania IPTL 100 HFO/medium-speed diesel engine BOO 20 1997 1998 Songas 180 Natgas/open cycle BOO 20 1994 2004 Uganda Namanve 50 HFO BOOT 6 — 2008 Bujagali 250 Hydro BOT 30 2005b — West Africa Côte d’Ivoire CIPREL 210 Natgas/open cycle BOOT 19 1993 1995 Azito 288c Natgas/open cycle BOOT 24 1996 2000 Ghana Takoradi II 220d Light crude/single cycle BOOT 25 1998 2000 Sunon Asogli 200 Combustion engine BOO 20 2007 — Bui Hydroe 400 Hydro BOO — 2005 — Nigeria AES Barge 270 Natgas/open cycle (barge-mounted) BOO 20 (2 parts) 1999 2001 Okpai 450 Natgas/combined cycle BOO 20 2001 2005 Afam VI 630 Natgas/combined cycle BOO 20 2000 2007 Aba Integratedf 140 Natgas/open-cycle BOO 20/15 2005 — Senegal GTi Dakar 52 Diesel/Nafta BOOT 15 1996 1999 Kounoune I 68 HFO BOO 15 2003 2008 Togo Centrale Thermique de Lomé 100 Triple fuel (natgas/HFO/diesel) BOOT 25 — 2010 Source: Authors. Note: Projects included here are greater than 40 MW; have reached financial close; and are under construction, operational, or concluded. BOO = build-own-operate contract; BOOT = build-own- operate-transfer contract; HFO = Heavy fuel oil. — = Not available. a. Iberafrica has been developed in three stages, with 44, 12 and 52.3 MW, totaling 109 MW at time of writing. b. The first phase of Bujagali’s conceptualization, spanning the mid-1990s until 2003 and involving AES, is not covered in this article. It should be noted that the project did not reach financial close during this time. Authors report only on the project from its second phase, starting in 2005. c. The initial project concept included specifications to raise capacity to 420 MW. d. Although the initial project concept included specifications to add a second phase of 110 MW and convert to combined cycle, lack of funding has limited the completion of this phase. e. The Bui project was initiated in the 1960s but aborted after the coup in Ghana in 1966. The project was reconsidered several times in the decades that followed. In 2005 the government of Ghana signed a memorandum of understanding with the Chinese firm Sino Hydro, and the plant is expected to be online in 2012. f. Aba Integrated is privately financed but is not a classic IPP in that ownership will extend to the off-taker. 373 in Takoradi II’s second phase and in Sunon Asogli, which as benefited from both customs and value added tax (VAT) of early 2010 had no gas supply—although efforts to rectify exemptions during construction, as well as the right to full this situation are under way. repatriation of profits. Currency conversion was also pro- vided for virtually all projects. In East Africa, Tanzania pro- vided a tax holiday of five years, but Kenya extended this COUNTRY-LEVEL FACTORS AFFECTING benefit only until plants were commissioned. Although THE IPP MARKET one would expect investment incentives to increase with Several elements have contributed to the success of IPP perceived risk, such a pattern is not apparent. projects in Sub-Saharan Africa: a favorable investment cli- mate; new policy frameworks and regulation; the linking of New policy frameworks and regulation planning, procurement, and contracting; and low-cost fuel and secure fuel contracts. While all eight Sub-Saharan African countries examined in this chapter have introduced legislation allowing for pri- vate sector power generation, few have actually formulated Favorable investment climate and implemented a clear and coherent policy framework Even though a investment climate is not the only factor in for procuring IPPs. Thus, there is abundant evidence of influencing IPP outcomes, it sets the stage for negotia- tentative experimentation with private power that does tions and contract terms and helps explain the initial not always lead to a sustained opening of the market for imbalance in some of the Sub-Saharan African cases. Host private investment. Furthermore, long-term power pur- countries with a strong investment profile attracted more chase agreements have the potential to constrain future investors and ultimately were able to negotiate deals with wholesale competition, although means to transition to more favorable terms than countries with weak invest- wholesale competition with IPPs have been identified ment conditions. Because all countries and entities com- (Woolf and Halpern 2001). In addition, state-owned utili- pete for capital, a risk-reward balance needs to be offered ties are rarely exposed to market costs of capital, and direct that will attract investors and lenders. That “balance” comparisons of state-owned utilities’ costs and IPPs’ costs starts with a stable and predictable investment environ- are often difficult to discern. ment (Rudo 2010a). The standard reform model for power sector reform— Of the Sub-Saharan countries whose IPPs are examined namely, unbundling of generation, transmission, and dis- in this chapter, none has an investment-grade sovereign tribution; as well as the introduction of competition and credit rating. Of the five countries that have received a spec- private-sector participation at both the generation and ulative rating (Ghana, Kenya, Nigeria, Senegal, and Uganda), distribution level—is not being fully adopted anywhere in all except Ghana received their rating after the first IPP Sub-Saharan Africa (UNEP and UNECA 2006; Malgras, deals were signed. Kenya’s investment climate was defined, at Gratwick, and Eberhard 2007a; Gratwick and Eberhard the time, by its aid embargo in the mid-1990s. Tanzania is also 2008b). Most incumbent national utilities are state-owned6 worth mentioning in this context. Throughout the 1990s, all and in a dominant position. However, elements of the export credit agencies were off-cover in Tanzania. Foreign reform model have been adopted: for example, Kenya has commercial banks were not willing to lend, because there unbundled power generation from its national transmission was no track record of successful repayment of commercial and distribution utility; Uganda and Ghana have unbun- loans. Consequently, traditional, project-financed IPP deals dled generation, transmission, and distribution; and in this climate were limited. In contrast in North Africa, the Uganda and Côte d’Ivoire have introduced private conces- Arab Republic of Egypt, Morocco, and Tunisia have all sions. The private sector has also invested in IPPs in many attained either a credit rating of investment grade or one countries. In general, there has been competition for the notch below and have experienced successful IPPs market, but not ongoing competition in the market in (Gratwick and Eberhard 2008a). While credit enhance- terms of customer choice. In effect, hybrid power markets ments and security arrangements have differed broadly have emerged across Africa, and in other developing between North and Sub-Saharan Africa, interestingly, regions. As such, the incumbent state-owned utility typi- incentives offered to investors in IPPs were relatively similar cally continues to play a key role in the sector, but because in both regions, though there was some variety with regard of inefficiencies and inadequate investment resources, IPPs to tax breaks. For instance, nearly all projects appear to have are gradually being introduced. These hybrid power 374 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS markets are giving rise to new challenges that need explicit scrutiny, with project sponsors and policy makers alike able attention if private investment is to be accelerated. to point to the benefits of the commercialized gas and the In addition to introducing legislation allowing for pri- reduction in fuel imports. vate power generation, the eight countries examined in this Finally, DFIs are behind many countries’ power sector chapter also have established independent regulators. In policy frameworks; notably the World Bank has had a hand Kenya the presence of the regulator, together with the adop- in nearly all power sector reform programs in Africa. DFIs tion of International Competition in Bidding (ICB) prac- have been particularly instrumental in advancing private tices, helped reduce PPA charges radically between the first sector participation in generation. And as many of those set of IPPs and the second. A similar trend may be observed same institutions began reconsidering publicly funded in Senegal, where the first IPP (GTi Dakar) was not overseen infrastructure investments at the end of the 1990s, coun- by the independent regulatory body and the second (Kou- tries’ policies have followed—from state to market and back noune I) was (Regulatory Commission of Senegal 2010; IFC again, albeit with some changes to accommodate the more 2010b). Kenya’s Energy Regulatory Commission (ERC) has hybrid market that has emerged. also been instrumental in helping set tariffs and manage the overall interface between private and public sectors. In Linking planning, procurement, and contracting Uganda sponsors note the benefits of having the regulator involved from project inception, particularly in helping to Intricately connected to sound policy frameworks are increase overall transparency, especially in the case of coherent power sector plans. Ideally, planning, procure- Bujagali. An ERC staff member affirms that the commis- ment, and contracting follow from sound policy frame- sion’s “presence has helped to focus minds on the require- works and include a number of core components: setting a ments for setting up power supply projects so that investors reliability standard for energy security, completion of coming in are clear of what is expected of them from the detailed supply and demand forecasts, development of a beginning and hence align their bids to these require- least-cost plan with alternative scenarios, clarification of ments. As a result, we have increasing numbers of investors how new generation production will be split between the applying to set up IPP projects” (ERC 2010a). private and public sectors, and identification of requisite In Côte d’Ivoire, Ghana, Nigeria, and Tanzania, however, bidding and procurement processes for new projects. One regulatory agencies have come into force only after IPPs important aspect of coherent power sector planning is vest- have been negotiated, and as yet the agencies have had little ing planning and procurement in one empowered agency to impact on new investment. In general, the mere presence of ensure that implementation takes place with minimal a regulator does not appear to be a defining factor in attract- mishaps (Eberhard and Malgas forthcoming). ing IPPs. An independent regulator may have positive, neg- While that planning and procurement arrangement may ative, or no impact on IPP outcomes. If, however, regulatory be ideal, the reality is often quite different. As one stakeholder governance is transparent, fair, and accountable, and if reg- at Ghana’s Public Utilities Regulatory Commission (PURC) ulatory decisions are credible and predictable, there is notes about the recent past, “A crisis arises, and everybody greater potential for positive outcomes for both the host panics; anybody who comes in [to propose generation] is country and investor alike. Evidence also suggests that effec- listened to” (PURC 2010). Indeed, in the eight countries tive regulatory oversight may lead to a reduction in the examined here, there have been several noteworthy plan- stated capital costs of projects and improved efficiency for ning mishaps that subsequently affected procurement and selectively bid projects (Phadke 2007; Eberhard and contracting. In some cases, demand and supply were not Shkaratan 2010).7 being accurately forecast partly because of weather condi- In Nigeria and Tanzania efforts have been made to tions such as extended droughts, which in turn necessitated exploit stranded gas as part of the IPP program.8 In Nigeria fast-tracking IPPs—that is, plans for IPPs were sped a reduction in gas flaring is central to the push for gas-fired through to meet immediate power shortages. The first power, while in Tanzania the IPP program commercialized two plants in Kenya (Westmont and Iberafrica), the first previously stranded (although not flared) gas through plant in Nigeria (AES Barge), and Ghana’s Takoradi II IPP Songas and Mtwara (a small private concession in the south were all negotiated during drought conditions. Although of the country). Although these two countries have a dis- both the Westmont and Iberafrica plants came online tinct set of challenges,9 in general the larger policy of involv- within 11 months, they were later investigated for suspicious ing stranded gas has insulated projects from intense public bidding practices and what were perceived as unnecessarily CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 375 expensive charges. Furthermore, public stakeholders’ Box 21.1 Poor Planning Led to Significant Delays unwillingness to make concessions over a tariff resulted in for Tanzania’s Projects Westmont not securing a second PPA. The AES Barge proj- ect in Nigeria took nearly two years to come online because The quick and obscure process through which IPTL of a renegotiation of the PPA, despite the project being (Tanzania) was negotiated resulted in the highest- fast-tracked. In Ghana, no agreement was reached over the profile IPP mishap on the continent to date. The second phase of the Takoradi II project. project’s total investment costs were estimated at Inability to estimate demand and supply accurately or to $150 million ($163 million including fuel conver- set a clear reliability standard has also led to several cases in sion), an amount the government and the World which emergency power units needed to be leased for one to Bank later argued was inflated by 40 percent. This two years with the purpose of plugging a short-term crisis. dispute led to a three-year arbitration process during In 2009 approximately 750 MW of emergency power was in which another project in Tanzania, the Songo Songo operation in Sub-Saharan Africa (Foster and Briceno- gas-to-electricity project, was also put on hold, Garmendia 2010). The governments of Ghana, Kenya, mainly through pressure from the World Bank, its largest donor, because of alleged corruption. Tanzania, and Uganda all have been forced to address Although the IPTL arbitration reduced its estimated drought and black-outs in recent years, often turning to costs it was still well above the price the government emergency power suppliers (see box 21.1 for an example). sought to pay. In addition, since both IPPs were Kenya harnessed 100 MW of emergency power twice: in unavailable until 2002 and 2004 respectively, the state 1999, 2000, and 2001, and again in 2006 (supplied by Aggreko, had to negotiate with another company, Rich- Cummins, and Deutz in the first instance and Aggreko alone mond/Dowans, for emergency power, a contract that in the second). In 2007 Aggreko’s contract was extended for came with its own inflated costs and controversies. an additional two years, and in 2009 Aggreko was selected to Furthermore, because of the delays, Songo Songo provide a further 140 MW for a total of 250 MW of emer- accumulated $100 million in interest charges on gency power.10 In Ghana emergency power was instrumental owner’s equity. in reducing the impact of the 1998 drought, but as drought In this project, critical planning elements are miss- conditions reversed, the state failed to honor its contracts ing, namely, a clear reliability standard, an accurate demand and supply forecast, a detailed plan for pri- with SIIF Accra; as of 2007 those contracts were still in dis- vately powered and publicly powered generation, and, pute. The cost of emergency power, at approximately 30 to most important, timely initiation of procurement and 40 cents a kilowatt hour (kWh), is high (IFC 2005). Tanzania, effective conclusion of contracts. for example, has estimated that it saves around $1 for every Source: Authors. kWh of outage averted (or about 5 to 10 times the ordinary cost of generating electricity).11 Competitive International Bidding Practices. Although it is easy, in hindsight, to accuse IPP stakehold- West Africa (Azito, GTi Dakar, and Kounoune I). Of the ers of acting imprudently in the face of emergencies, the projects examined here that have faced renegotiation, four actual conditions of load-shedding and shortages in the were not bid through an ICB (IPTL, Iberafrica, AES Barge, power sector appear to have provided few alternatives. Two and Okpai); the two exceptions were Songas and OrPower studies highlight the importance of the international com- 4. Absence of regulatory scrutiny is also noteworthy in petition in bidding practices in reducing up to 60 percent each of these four projects. Furthermore, Westmont, of the capital cost of plants (Deloitte Touche Tohmatsu which was selectively bid, quit the country after its first Emerging Markets Ltd. and Advanced Engineering Associ- seven-year PPA expired. Other non-ICB projects have also ates International 2003; Phadke 2007). There is also evi- encountered difficulties. For example, Ghana’s Takoradi II dence for selective bidding proving effective in certain has not been able to raise financing for the second phase of instances, provided there is regulatory scrutiny. Of the 21 the plant. Although reasons for these stumbling blocks IPPs examined here, ICBs are known to have been con- may be traced well beyond the presence or absence of an ducted for nine. Six of the projects in East Africa (OrPower ICB, the correlation is nonetheless revealing. 4, Tsavo, Rabai, Songas, Namanve, and Bujagali) con- Furthermore, the success of the ICB process is intricately formed to these bidding practices, as did three projects in linked to the number of bids received, with more bidding 376 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS driving down prices,12 as evidenced by IPPs in North Africa, with these planning, procurement, and contracting chal- where the number of bids submitted to ICBs has been gen- lenges (box 21.2). erally double to triple those submitted to ICBs in East and West Africa. That said, the time and cost required to com- Abundant, low-cost fuel and plete an international competitive bid should not be under- secure fuel contracts estimated. In Sub-Saharan Africa, drought-related energy crises are often cited as the reason why ICBs have not been The availability of competitively priced fuel supplies has pursued. In Togo, for example, the drought conditions of also emerged as a key factor in how IPPs are perceived and 2006 prompted a move to discontinue a rehabilitate-own- ultimately whether there is public appetite for such projects, transfer contract with Electro Togo to manage the Centrale in large part because fuel is generally a pass-through cost to thermique de Lomé. Rather than launching an ICB, the gov- the utility and, in many cases, also to the final consumer. ernment chose to negotiate directly with an existing player, Thus, if the IPP uses a different fuel than other plants, and Contour Global, which was already in discussions with the if that fuel is more expensive, there is greater potential for utility. In this case, time and project familiarity proved more disputes surrounding the new project. important than complying with international bidding prac- In several Sub-Saharan African countries—Ghana, Kenya, tices, which risked extending the project development time- Senegal, Tanzania, and Togo—low-priced hydropower was line (Ministry of Energy of Togo 2010).13 the dominant fuel source for power production, but IPPs The reason for these planning and contracting mishaps were thermal powered, using a combination of imported fuel seems to lie in the changing nature of power markets across Africa and other developing regions. In addition, as power markets continue to include private sector partici- Box 21.2 Planning, Procurement, and Contracting pation, it is not always clear who has responsibility for Challenges in Kenya maintaining security of supply. Often, the planning func- tion is shifted from the state-owned utility company to Recognizing that it does not have the internal capacity, resources, or planning tools to develop detailed and the energy ministry, which does not have the capacity, up-to-date electricity plans, Kenya’s Energy Regulatory resources, or experience to undertake detailed power sector Commission (ERC) has delegated this function to the planning. Alternatively, planning is contracted to consult- Kenya Power and Light Company (KPLC), guided by a ants who produce master plans that quickly become government committee chaired by the ERC. KPLC outdated as global equipment, fuel costs, and other key ceased being an electricity-generating company in parameters change (Eberhard and Malgas forthcoming). 1997, and thus has a neutral stance between the state Absence of regulatory control and institutional capacity is utility, Kenya Electricity Generating Company evident in more areas than just planning. In many (KenGen), on one hand and private IPPs on the instances, there are no clear criteria for allocating power other. KPLC has also been assigned responsibility plant building opportunities to either the incumbent for managing the procurement and contracting state-owned utility or the private sector. process for IPPs. As described by one project sponsor Often, it is not clear whether plans are merely indicative, in Kenya, commenting in May 2010, “Kenya has an IPP structure that is working. It has a track record. It can whether unsolicited proposals may be considered, or structure new projects based on experience gained whether plans have legal force in determining which plants from previous projects. And it has a very capable set of the regulator may license. Too often, plans do not translate teams working in KPLC, the Ministry of Energy/ into timely initiation of competitive bid processes for new Finance and KenGen. It understands project finance plants. Similarly, there is often insufficient capacity to nego- and is not surprised when a developer requests a com- tiate with winning bidders or to conclude sustainable con- fort letter, as one example.” The fruits of this approach tracts. Transaction advisers may be appointed, but often are evident. Kenya is about to add three new IPPs to its there is little continuity and the overall policy framework is existing five, maintaining its lead position in Sub- lacking information about security packages or credit Saharan Africa in terms of IPP investments.a enhancement measures being offered by government. a. Kenya also has been cited by public stakeholders in Ghana Despite past planning deficiencies and forced reliance on and Tanzania as having processes that should be emulated emergency power generation, Kenya provides an interesting (PURC 2010; EWURA 2010). recent example in which progress is being made in dealing CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 377 oil and domestically procured natural gas. These same coun- guaranteed through contracts well into the future. Fuel tries witnessed a series of debilitating droughts during the sourcing and other country-level elements that contribute 1990s and 2000s, during which existing hydropower infra- to successful IPPs are summarized in table 21.2. structure proved insufficient, and thermal power, provided almost entirely by IPPs, was increasingly integrated into the PROJECT-LEVEL FACTORS CONTRIBUTING fuel supply mix (from 10 percent to 60 percent in Tanzania, TO SUCCESSFUL IPPS for example), forcing up the price of power. Although more thermal power may be required, the public perception has Investors in IPPs in Sub-Saharan Africa must navigate been that IPPs drive up the price of power. OrPower 4, changing investment climates, national policies, and plan- Kenya’s geothermal IPP, deserves special mention in the con- ning frameworks. Starting with an evaluation of equity text of hydro-dominant systems that have been diversified arrangements, this section examines trends in investor with (largely imported) fuel oil (box 21.3). behavior and how investors secured revenue to service debt The number of IPPs in Sub-Saharan Africa with secure, and reward equity, particularly in the face of exogenous low-cost fuel sources is still relatively few. In Tanzania, stresses. although natural gas from the Songo Songo field is cheaper than the imported fuel oil currently powering IPTL,14 dis- Favorable equity arrangements putes continue to surround its use, and IPTL’s diesel units have yet to be converted to gas. In Ghana the Sunon Asogli In assessing the project-level factors contributing to the suc- Power Plant is completed but awaits fuel. Ghana’s Volta cess of IPPs, it is useful to ask several questions. Does the River Authority (VRA), the state-owned generator, is a presence of local equity shareholders make a difference in foundation customer for the West African Gas Pipeline project outcomes? Are projects with such participation less (WAGP), from which it has been receiving gas since Febru- likely to face pressure from host-country governments to ary 2010, but the recently completed Shenzhen IPP will change their contract terms? How does a firm’s previous receive gas only when the pipeline is pressurized and higher experience with a country play out in making or breaking volumes of gas flow. In short, an IPP’s access to abundant, deals? Does the presence of DFI or firms with international low-cost fuel is not only a matter of whether such fuel is development experience, such as Industrial Promotion Ser- available in a country, but whether the supply of such fuel is vices (IPS), Globeleq, and Aldwych International, affect the likelihood of an IPP’s success? Table 21.3 shows the country origin of Sub-Saharan African IPP sponsors, along with their respective equity share, whether the project faced a Box 21.3 Petroleum-Fired Versus Natural change in contract terms, and whether there was turnover of Gas-Fired IPPs the majority equity partner. Unlike in China and Malaysia, where local investors in As indicated in a press release from Ormat (2009), “At IPPs abound, foreign firms have long been dominant play- a price of US$29 per barrel of petroleum crude oil, [a] ers in Sub-Saharan African IPPs (Woodhouse 2005). Given 48 MW geothermal plant is cheaper to operate than a the limited availability of capital in Sub-Saharan African heavy fuel oil fired plant. This means that at the cur- countries in the sample, this is not surprising. Foreign rent oil prices still above US$40 per barrel, OrPower 4 involvement, however, raises the issue of foreign exchange is providing cheaper electricity to the national grid than any existing oil fired plants in Kenya.” In con- exposure.15 Only 3 of the 21 projects examined here have trast, Nigeria has relied entirely on domestically pro- local majority stakeholders: Nigeria’s Okpai, Afam VI, and cured natural gas to fuel its IPPs, and gas is the incum- Aba Integrated. In two of these cases, however, the stake- bent fuel. Until recently, although a series of issues holder was either the national utility or the Nigerian affected project outcomes, most notably the invest- National Petroleum Company (NNPC). ment climate and bidding procedures, fuel had not Several observers (for example, Hoskote 1995; Wood- been an issue; however, civil unrest in the Niger Delta house 2005) have cited local participation in IPPs as a led to a disruption in the fuel supply in 2007–09, possible means of reducing risk. Ten of the 21 projects albeit with improved conditions as of 2010. have local equity participation—namely, Sunon Asogli, Source: Authors. Iberafrica,16 IPTL, Songas, Takoradi II, AES Barge, Okpai, Afam VI, Aba Integrated, and Bujagali. As noted, six of these 378 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS Table 21.2 Elements within Purview of Host Governments Contributing to Successful IPPs Element Details Favorable investment climate ■ Stable macroeconomic policies ■ Legal system that allows contracts to be enforced, laws to be upheld, and arbitration ■ Good repayment record and investment-grade credit rating ■ Requires less (costly) risk mitigation techniques to be employed, translating into lower cost of capital, lower project costs, and more competitive prices ■ Potentially more than one investment opportunity Clear policy framework ■ Policy framework enshrined in legislation ■ Framework clearly specifies market structure and roles and terms for private and public sector investments (generally for single-buyer model, as wholesale competition does not yet exist in African context) ■ Framework is led and implemented by reform-minded “champions,” concerned with long-term power sector conditions Clear, consistent, and fair regulatory oversight ■ Oversight improves general performance of private and public sector assets ■ Transparent and predictable licensing and tariff framework improves investor confidence ■ Cost-reflective tariffs ensure revenue sufficiency ■ Consumers protected Coherent power sector planning linked to procurement ■ Energy security standard in place; planning roles and functions and contracting are clear ■ Power planning vested with lead, appropriate (skilled, resourced, and empowered) agency ■ Power sector planning takes into consideration the hybrid market (public and private stakeholders and their respective real costs of capital) and fairly allocates new building opportunities among stakeholders ■ Planning has built-in contingencies to avoid emergency power plants or blackouts ■ Responsibility for procurement is clearly allocated; plans are linked to procurement and bids are initiated in time ■ Procurement process is transparent and competition ultimately drives down prices ■ Capacity is built to contract effectively Abundant, low- cost fuel and secure contracts ■ Cost-competitive with other fuels ■ Contract safeguards affordable and reliable; fuel supply sufficient for duration of contract Source: Authors. projects have encountered some form of change to their on to negotiate a second 15-year PPA, in contrast to West- contract, and in four of the six, either the state utility or mont, which stopped production after failing to come to another government entity held an equity share, indicating an agreement on a second PPA. Although the presence of a that the existence of a local partner might not be critical in local partner may have helped Iberafrica create a long- preserving the original financing balance. term solution, with just one example, the evidence is not As for renegotiating contract terms, it is unclear conclusive. Togo’s Centrale thermique de Lomé, due to be whether having a local partner makes a difference. Kenya’s online in July 2010, may provide a more sustainable Westmont and Iberafrica, for example, were negotiated at method for balancing investment and development out- the same time under similar policy frameworks, although comes. In this project, 25 percent of the project equity Iberafrica had local participation and Westmont did not. must be sold to local investors within the first five years Iberafrica first voluntarily reduced its tariff and then went (Ministry of Energy of Togo 2010). CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 379 380 Table 21.3 Equity Participation in IPPs in Sub-Saharan Africa Majority equity Change in partner turnover Project Equity partners (country and percent of equity held) contract terms (number of times) Westmont (Kenya) Westmont (Malaysia, 100%); has sought to sell plant since 2004 — — Iberafrica (Kenya) Union Fenosa (Spain, 80%), KPLC Pension Fund (Kenya, 20%) since 1997 yes 0 OrPower 4 (Kenya) Ormat (United States/Israel, 100%) since 1998 yes 0 Tsavo (Kenya) Cinergy (United States) and IPS (international) jointly owned 49.9%, Cinergy sold to Duke Energy no 1 (United States) in 2005; CDC/Globeleq (United Kingdom, 30%), Wartsila (Finland, 15%), IFC (international, 5%) have retained remaining shares since 2000 Rabai (Kenya) Aldywch-International (Netherlands, 34%), BWSC (Denmark, but owned by Mitsui of Japan, 25.5%), no 0 FMO (Netherlands, 20%), IFU (Denmark, 20%) IPTL (Tanzania) Mechmar (Malaysia, 70%), VIP (Tanzania, 30% in kind); both have sought to sell shares, and actual equity yes — contribution is currently disputed Songas (Tanzania) TransCanada sold majority shares to AES (United States) in 1999, and AES sold majority shares to yes 2 Globeleq (United Kingdom) in 2003.a All preferred equity shares were converted into “loan notes” in June 2009, only common shares remain Bujagali (Uganda) Sithe Global (United States, 58%), IPS-AKFED (32%), government of Uganda (10%) no 0 Namanve (Uganda) Jacobsen (Norway, 100%) no 0 CIPREL (Côte d’Ivoire) SAUR International (88%). This is a joint venture between French SAUR Group owned by Bouygues no 1 (65%), and Electricité de France (35%). The remaining 12% was owned by a combination of BOAD (West African Bank for Development); the Investment and Promotions Company for Economic Cooperation, a subsidiary of AFD; and IFC. In 2005 98% of the shares were sold to Bouygues (France); BOAD retained 2%. Azito (Côte d’Ivoire) Cinergy (65.7%; joint venture between Swiss ABB and Globeleq), CDC/Globeleq (11%), and IPS-AKFED (23%) no 1 Takoradi II (Ghana) CMS (United States, 90%), VRA (Ghana, 10%); CMS sold shares to TAQA (UAE, 90%) in 2007 yes 1 Sunon Asogli (Ghana) Shenzhen (China), Togbe Afede XIV (Ghana/local strategic investor) no 0 Bui hydro (Ghana) Sinohydro (China) no 0 AES Barge (Nigeria) Enron (United States, 100%); sold to AES (United States, 95%) and YFP (Nigeria, 5%) in 2000 yes 1 Okpai (Nigeria) Nigerian National Petroleum Corporation (Nigeria, 60%), Nigerian Agip Oil Company (Italy, 20%), yes 0 and Phillips Oil Company (United States, 20%) have maintained equity since 2001 Afam VI (Nigeria) Nigerian National Petroleum Corporation (Nigeria, 55%), Shell (United Kingdom/Netherlands, 30%), no 0 Elf (Total) (France, 10%), Agip (Italy, 5%) Aba Integrated (Nigeria) Geometric Power Limited (Nigeria) no 0 GTi Dakar (Senegal) GE Capital Structured Finance Group (SFG) (USA), Edison (Italy), IFC no 0 Kounoune (Senegal) Mitsubishi (Japan), Matelec S.A.L (Lebanon) no 0 Centrale thermique Contour Global (United States, 80%) IFC (20%) no 0 de Lomé (Togo) Source: Authors. Note: — = Pending developments. a. Turnover within the Songas project has been complex: Ocelot (Canada), TransCanada (Canada), Tanzania Petroleum Development Corporation, TPDC (Tanzania), TANESCO (Tanzania), Tanzania Development Finance Company, TDFL (Tanzania, sponsored by European Investment Bank), IFC (multilateral), DEG (Germany), and CDC (United Kingdom) were shareholders by 1996, with TransCanada the majority shareholder. IFC and DEG sold their shares to CDC in 1997/98; TransCanada sold its shares to AES (United States) in 1999; Ocelot/PanOcean sold shares to AES in 2001; and AES sold majority shares to Globeleq (United Kingdom) and FMO (the Netherlands) in 2003. After the AES sale, equity shares and associated financial commit- ments in Songas were as follows: Globeleq, $33.8 million (56 percent); FMO, $14.6 million (24 percent); TDFL, $4 million (7 percent); CDC, $3.6 million (6 percent); TPDC, $3 million (5 percent); and TANESCO, $1 million (2 percent). The amount given for Globeleq does not reflect the additional $50 million that the company committed for the expansion of the project. Origins, Experience, and Mandates of Partners. From with development origins examined here has had any a global perspective, IPP investments during the 1990s were changes in contract terms, which may signal that they were led by a host of American and European investors who saw better balanced from inception in terms of investors and returns in their home markets diminishing. A wave of had a better ability to withstand public pressure. Further- investors originating from developing countries, however, more, in terms of the Songas change, although the particularly from Malaysia, was also present. Although it $103 million buy-down of the allowance for funds used would be inaccurate to say that investors based in develop- during construction resulted in a reduction in the capac- ing countries overlooked the risk involved in African coun- ity charge, the firm received full payment upon the buy- tries (or did not ultimately charge higher returns), they may down—a different case from many of the contract changes have had a greater willingness to consider IPP investments cited earlier. Alongside companies with development ori- in Sub-Saharan Africa in the first place. gins, foreign investors remain directly involved in IPP While the number of developing-country-based investors investments. Indeed, DFI equity shares have increased in in IPPs appears to be growing, three such firms are trying to recent years for at least five IPPs. With the exception of sell their shares (Mechmar, VIP, and Westmont). Thus, the Songas, no project with DFI involvement has seen any con- home country of the firm does not mean that project equity tract changes. Box 21.4 provides further insights on IPP is permanent, or that firms based in developing countries are projects concerned with their developmental impacts. best positioned to service debt and reward equity. An aspect more important than the nationality of the Equity Turnover. Of the 46 original equity partners in investing firm appears to be a firm’s experience and mandate. the 21 projects considered here, 7 have exited (from 5 pro- Across the pool of IPPs examined in table 21.3, several firms jects). This ratio, however, tells only part of the story. First, were actively involved in the country before making an IPP as previously indicated, shareholders in Westmont (Kenya) investment. Union Fenosa, for example, the parent company and IPTL (Tanzania) have been trying for several years to of Iberafrica, had previous experience in Kenya through an information-technology contract. IPS, a major shareholder in Tsavo (Kenya), Azito (Côte d’Ivoire), and Bujagali (Uganda), Box 21.4 Two IPPs with Development and has operated in Kenya since 1963 and in Côte d’Ivoire since Investment Mandates 1965. For certain projects, it may be argued that long-term relationships, particularly those with strong local manage- Development is perhaps more important for Indus- ment, appear to have contributed to the staying power of trial Promotion Services (IPS) than for any other firm firms and often the rebalancing of contract terms. investing in IPPs in Sub-Saharan Africa. As a rule, IPS In addition to the terms of the deal, the mandate of the invests only in projects with a substantial develop- investing firm appears to play a central role in firms’ deci- mental impact and a reasonable internal rate of sions about whether to invest in IPP projects. Although return. For the Tsavo project, that rate is 17–18 per- Globeleq, IPS, and Aldwych, for example, are commercial cent, and for Bujagali, 19 percent—considerably lower than typical internal rates of return for IPP projects in entities, they emerged from agencies with a strong commit- Sub-Saharan Africa (IPS 2010). ment to social and economic development. Until recently, Aldwych International, on the other hand, requires the two firms that were increasing their stakes were Glo- that its project investments both make commercial beleq and IPS. Globeleq holds a 43 percent share in Côte sense and serve a clear developmental function for the d’Ivoire’s Azito, 30 percent equity in Kenya’s Tsavo, and country and local community. The company’s experi- 56 percent in Tanzania’s Songas. IPS holds a 23 percent ence in the Rabai in Kenya helps to illustrate this share in Azito, and together with Duke Energy, a 49.9 percent point. The firm has faced three major hurdles as part share in Tsavo. IPS is also leading development of of its involvement in the project—legal trouble during Uganda’s Bujagali project and is a 35 percent shareholder the tendering process, national postelection civil in equity. Although a smaller player than either Globeleq unrest in late 2007, and the global financial crisis. or IPS, Aldwych International has also made significant Despite the challenges, the project reached financial inroads via Rabai in Kenya and is evaluating further close successfully in 2008; it even won the Project Finance International’s deal of the year award. expansion. With the exception of Tanzania’s Songas, none of the Source: Authors. Sub-Saharan African IPP projects with involvement of firms CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 381 sell their assets. In the case of IPTL, Mechmar, the lead nonrecourse project financing is the norm for privately shareholder, has indicated that the sale has been driven by financed electric power plants in developing regions, the an arbitration settlement that hurt its equity partners. 21 projects examined here represent several notable excep- VIP, the minority shareholder, cites oppression by the tions: Nigeria’s Okpai and Afam VI, which were 100 percent majority shareholder, fraud by Mechmar in inflating the financed by the balance sheet of equity partners, and the IPTL capital cost, and failure by Mechmar to pay its second phase of Songas, which was largely refinanced by a equity contribution (that is, the project was 100 percent World Bank loan in 2009 (Globeleq 2010a). Until recently, debt financed) as reasons behind its desire to sell its stake. Westmont, Iberafrica, and OrPower 4 were all financed For this and other projects, investors repeatedly say that entirely with the balance sheets of their sponsors. For West- their motivation to sell stakes in IPPs is driven primarily mont and Iberafrica, the reason cited for this arrangement by changing circumstances in home markets or corporate was insufficient time to arrange project finance, because strategy—that is, the desire to sell has little to do with plants had to be brought online within 11 months. For host-country actions and reactions or with poor invest- OrPower4, the reason was linked to lack of a security pack- ment outcomes, namely, the ability to service debt ade- age, which was not finalized until 2006. quately and reward equity. While investment outcomes may be partially motivating The Impact of DFIs on Projects. With limited appetite sales, turnover does not in and of itself appear to be chal- for IPPs among commercial banks, DFIs provide a sub- lenging the long-term sustainability of contracts, since in stantial amoung of credit to projects. Indeed, DFIs such as nearly all cases sellers have found willing buyers to take over the World Bank, International Finance Corporation (IFC), the original or recently renegotiated PPAs. The two excep- European Investment Bank, DEG, FMO, African Develop- tions, again, are Westmont, in which the first PPA expired ment Bank, PROPARCO, the Emerging Africa Infrastruc- and was shrouded in controversy, and IPTL, which has been ture Fund, and European Financing Partners, KfW, and embroiled in lawsuits. Under such circumstances, it may Agence Française de Développement have participated in therefore be understandable that the plants have not nearly every IPP. Beyond these institutions’ long history of attracted buyers. One stakeholder went so far as to assert activity in Sub-Saharan Africa, their involvement is attrib- that “[equity turnover is a] healthy factor in a maturing utable to real and perceived risks by private investors, market. It is a good sign when investors come and go—not which keep them from filling the financing gap, and to the a bad or threatening thing.” The return of the government DFIs’ interest in participating in the broader mandate of as a shareholder, as planned in the case of Tanzania’s IPTL, power sector reform. would, however, signal that some markets are less mature Although projects with DFI funding tended to take than initially expected. longer to reach financial closure than those financed through private sources, project sponsors in Kenya say that multilateral and bilateral development institutions helped Debt arrangements: Global and local them maintain contracts and resist renegotiation in the face With debt financing often representing more than 70 percent of external challenges such as droughts, when developers of total project costs, competitively priced financing has were pressured to reduce tariffs. A particularly revealing emerged as a key factor in successful projects. How and where contrast is between the two Kenyan plants, OrPower 4 and to get this low-cost financing is a challenge, but possible Tsavo. Although the two projects were negotiated under the approaches in the case of Sub-Saharan Africa are DFIs, same policy framework, the former initially had no multi- credit enhancements, and flexibility in repayment terms and lateral involvement in either its equity or debt,18 whereas the conditions, including possible refinancing.17 The goal for IFC arranged all the debt for Tsavo and took a 5 percent sustainable financing should be that the risk premium equity stake. Tsavo has since resisted KPLC pressure to demanded by financiers or capped by the off-taker matches reduce its tariffs. OrPower 4, on the other hand, ultimately the actual country risk and that project risks and risk pre- reduced its tariff for the second phase of the plant. Tanza- miums are not inflated. nia’s Songas project, for which the World Bank and Euro- While there is no uniform pattern in the debt financ- pean Investment Bank financed all project debt, also ing for the projects considered here, observation of sev- deserves special mention here. The project took almost a eral trends in terms of how investors handled costs may decade to reach financial closure; the World Bank played an contribute to the success of other projects. Although instrumental role in, among other things, pressuring the 382 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS IPTL arbitration, which ultimately led to more balanced generally also addressed. Nearly all the contracts specified contract terms. some form of international dispute resolution and mini- mum power availability. Locally Sourced Finance. While observers argue that For each project, sponsors negotiated or were granted local financing is key to increasing sustainable foreign PPAs in U.S. dollars, thereby reducing their exposure to investment, capital markets in many Sub-Saharan African currency devaluation. Over time, the bulk of project countries are underdeveloped and illiquid, hence unable contracts have been upheld, but not without changes. to provide financing for all projects. Three exceptions in Box 21.5 reviews PPA renegotiations in Kenya, Nigeria, the project sample are Kounoune I of Senegal, Geometric and Tanzania. of Nigeria, and Namanve of Uganda. The Kounoune I project received financing from both the West African Credit enhancements and security arrangements Bank for Development (BOAD), based in Togo, and the Banking Company of West Africa (CBAO), a private bank The underlying credit risk of IPP projects in Sub-Saharan based in Senegal. Funding for Geometric has been pro- Africa has been dealt with largely through a suite of credit vided by Diamond Bank and Stanbic IBTC Bank Plc, both enhancements, namely, guarantees, insurance, and cash (the of Nigeria. Similarly, for Uganda’s Namanve project, financ- latter of which took the form of escrow accounts, liquidity ing was provided by Standard Chartered Bank of the United facilities, and letters of credit of varying amounts and Kingdom through Stanbic Bank of Uganda. By contrast, in tenures). The Tsavo project in Kenya, for example, has an North Africa, all of the €213 million debt of Tahaddart, a escrow account equivalent to one month’s capacity charge 384 megawatt combined-cycle gas turbine plant in and a standby letter of credit from KPLC covering three Morocco, was financed by local banks. Local financing was months’ billing (approximately $12 million). At least 12 of aided by a number of factors, including the state utility’s the 21 projects examined in this chapter had some form of prominent role in the plant (it holds nearly 50 percent of cash security arrangement, with typical terms of between total equity) and the fact that Morocco’s commercial banks one and four months’ capacity charge in reserve. have a significant degree of state involvement. With or with- Not surprisingly, the number of credit enhancements out state involvement, however, no other country in Sub- appears to diminish as a country’s risk profile improves. Saharan Africa has yet been able to manage to arrange this There are, however, noticeable exceptions such as the first level and depth of local financing for IPPs. wave of IPPs in Kenya (Westmont and Iberafrica), for which One main drawback for IPPs without local finance is the risk appears to be entirely reflected in the higher capacity impact of macroeconomic shocks and local currency deval- payments negotiated. That said, corruption was also alleged uation. Since the late 1990s, Ghana, Kenya, and Tanzania in both these plants.19 Thus, the “security arrangement” have experienced substantial depreciation, with their cur- may lie not in a letter of credit but in an informal agreement rencies losing more than 100 percent, 200 percent, and among sponsors. 400 percent of their value against the U.S. dollar, respec- Of the many different credit enhancements used in tively. Inevitably, these currency shifts have led to pressure connection with IPPs, sovereign guarantees are most to reduce capacity charges and to countries reconsidering common. Such guarantees are known to have been IPP development. extended for at least nine of the pool of 21 projects: Tanzania’s IPTL, Nigeria’s AES Barge, Côte d’Ivoire’s Azito, Ghana’s Takoradi II (phase I), both GTi Dakar and Securing revenue: The PPA Kounoune I in Senegal, Togo’s Centrale thermique de Lomé, All 21 of the projects examined here had a long-term power and Bujagali and Namanve in Uganda. Several of the proj- purchase agreement with the incumbent state-owned utility ects without guarantees—Tsavo and Rabai, for example— to ensure a market for the power produced and to secure were, however, given assurances by the government in the revenue flows for debt and equity providers. In addition to form of comfort or support letters, through which political stating who would buy the power, the PPAs detailed how risk is assumed. In the case of the Okpai plant in Nigeria, much power capacity would be available and how much security was extended in the form of the state-owned oil buyers would be charged. Provisions concerning fuel meter- company’s revenues. Thus, if the off-taker defaults, NNPC, ing, interconnection, insurance, force majeure, transfer, ter- among the most liquid firms in the country, is liable. Partial mination, change of legal provisions, and refinancing were risk guarantees (PRGs) issued by the World Bank were used CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 383 Box 21.5 The Experience of PPAs with Renegotiations in Kenya, Nigeria, and Tanzania In Kenya’s first wave of IPPs, costs were inflated in part due to the short duration of contracts—only seven years. With Iberafrica facing ongoing pressure to reduce its tariff, coupled with an interest in negotiating another contract, the sponsor voluntarily reduced the capacity charge enshrined in the PPA. (Iberafrica’s second and third PPAs are for a considerably longer time frame than its first, and tariffs have been reduced significantly.) Westmont, on the other hand, did not negotiate a second contract after it failed to obtain the same terms—in particular, those related to capac- ity charges—spelled out in its first PPA. Likewise, changes in Kenya’s OrPower 4 and Tanzania’s Songas PPA agree- ments are related in part to the final amount of the capacity charge (as originally spelled out in the PPA).a In the case of Nigeria’s AES Barge, initially sponsored by Enron, renegotiations in 1999–2000 brought about sev- eral changes in the PPA, including a change in the fuel specifications (from liquid fuel to natural gas), which led to a major reduction in the fuel charge for the off-taker. The current arbitration with AES Barge involves, among other things, reconsideration of the availability-deficiency payment as well as the tax exemption. In each of the cases reviewed here, the original terms of the PPA have been viewed as unsustainable for the host country and therefore challenged. The case of Tanzania’s IPTL is slightly different, however. Although the contract was considered initially unsus- tainable due to the added capacity of Songas, the IPTL arbitration was prompted by what was deemed a breach in the PPA, namely, the project sponsors’ substitution of medium-speed for slow-speed engines without passing on the cap- ital cost savings to the utility, as per the PPA.b a. It is, however, worth reiterating in this context that failure to agree on the security package and the capacity charge contributed to delays in the development of OrPower 4’s additional 36 megawatts of production capacity. b. Although this dispute was resolved, a subsequent and prevailing dispute relates to the level of actual equity in the project, which in turn affects WACC and the allowed rate of return. for two of the projects, Azito and Bujagali.20 In these the government of the country must request a PRG; thus, instances, the PRG covered all debt of the commercial the project must be significant in the eyes of both the gov- banks. In the case of default by the project company, the ernment and the World Bank. For projects without PRGs PRG (backed by World Bank) would pay the commercial security arrangements and credit enhancements are lenders, and the World Bank would then claim repayment similar, with the DFIs generally accepting the political risks, from the government (World Bank 1997, 1999). For other such as in the Azito and Songas projects. IPPs, political risk insurance was provided by the Overseas In Kenya, the only country among those in Sub-Saharan Private Investment Corporation (OPIC), whereas guaran- Africa examined here to extend sovereign guarantees to tees relating to currency inconvertibility, expropriation, IPPs, stakeholders in Tsavo indicated that, without such a and political violence were issued by the World Bank’s guarantee, the presence of the IFC became critical, both to Multilateral Investment Guarantee Agency (MIGA). help arrange debt and share in equity. In Ghana, lack of sov- In several cases, these credit enhancements have ereign guarantees has been cited as the main obstacle to improved the sustainability of projects and attracted or developing the second phase of the Takoradi II project. assuaged lenders. The Bujagali project’s PRG, for example, Other credit enhancements have been used in abundance was instrumental in motivating and solidifying the involve- in Kenya, including a suite of escrow facilities, which have ment of four commercial banks, which contributed a com- contributed to KPLC’s cash-strapped position. Although bined $115 million at very competitive pricing. Some have government guarantees were recently debated by officials, likened the PRG to a hammer effect, with the World Bank it looks as if the government will retain its no-guarantee guaranteeing what the government has already guaranteed policy going forward, providing only letters of support to and thus making the government’s commitment twofold. IPPs. KPLC cites the absence of sovereign guarantees as That said, PRGs are not appropriate for all Sub-Saharan hampering its ability to raise private finance, while ERC African IPPs because they are typically used for large proj- counters that IPPs have been introduced to help commer- ects in countries that are in an early stage of reform and cialize the sector. Government guarantees work against when commercial lenders are also present. Furthermore, this goal, however. 384 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS Notably, sovereign guarantees, political risk insurance 95 percent between 2004 and 2006, versus 60 percent for policies, and PRGs have not been invoked in any of the 21 KenGen’s thermal plants. When asked generally about his projects examined here, including in those projects that most favorable IPP experience to date, one public stake- ultimately faced a change in the contract (AES Barge, IPTL, holder simply indicated, “their technical performance” OrPower 4, and Takoradi II). Recourse to international arbi- (ERC 2010). tration has occurred only in the case of IPTL. Additionally, it is helpful to reflect on the overall application of security arrangements and credit enhancements. Although there is Strategic management and relationship building some variation in the project sample, by and large the vari- Once 20-year contracts are in place, it would seem that the ation is limited, with the size of the project, the track record deal is set and the revenue secured, with clear provisions to of the regulatory regime (including its stability and credi- ensure debt repayment and reward equity. Several other bility), and the creditworthiness of the off-taker being the interrelated actions deserve mention, however. One involves main determinants. While there is resistance to government relationship building, including with local partners and guarantees on the part of some country stakeholders, proj- communities. Another relates to how sponsors handle the ect developers and multilaterals typically support them. As onset of stresses, including those presented by capacity one World Bank (2010) official noted, “the first level of sup- charges and refinancing. port has to come from the government.” Numerous IPP sponsors have adopted outreach pro- Furthermore, there has been very little evolution in the grams to improve relations with local communities. In credit arrangements used for IPPs in Sub-Saharan Africa. Kenya, for example, Tsavo set up a $1 million community All projects are supported by a PPA and their credit risk development fund for the duration of the 20-year PPA, largely carried by a government guarantee. In countries in from which grants of $50,000 each are disbursed each year which the market for power is not developed, a PPA, along to benefit environmental and social activities in Kenya’s with a government guarantee, is generally required, partic- Coast Region. Iberafrica maintains a social responsibility ularly where the off-taker is not creditworthy (IFC 2010a; program, and IPTL also is an active donor to its immedi- Rudo 2010a). This situation is quite different than that in ate community. In Ghana CMS’s social responsibility other developing regions. In Latin America, for example, involvement (before it sold its shares to TAQA in 2007) PRGs and other credit enhancements and security included providing scholarships for secondary and tertiary arrangements are virtually nonexistent because power education and supporting medical clinics and the con- markets are operational and local lenders are “in the struction of drainage systems. Bujagali also has a suite of driver’s seat and generally very comfortable with local social outreach programs. Although the sums are not sig- developers and regulation.”21 nificant, these programs, particularly when well adver- tised, have the potential to win allies and counter the stereotype of IPPs. Positive technical performance Another, perhaps more significant, action is how spon- Virtually all IPPs among the 21 considered here have sors cope with stresses such as macroeconomic shocks, cur- shown positive technical performance. Exceptions include rency depreciation, and pressure from host governments to Nigeria’s plants (which have had problems with fuel sup- reduce costs. Anecdotal evidence suggests that strategic ply) and, more recently, Kounoune I in Senegal (Nigeria management helped put Kenya’s Iberafrica back on track, in Electricity Regulatory Commission 2010; IFC 2010b). In contrast to Westmont, where there is no evidence of such general, IPPs’ technical performance is superior to that of action. Iberafrica, in fact, has had to cope with two major state-owned plants. An argument for extending gas from stresses, drought and alleged corruption. According to the West African Gas Pipeline to the Sunon Asogli Power stakeholders at Iberafrica, the IPP voluntarily reduced its Plant in Ghana, for example, is being made in part because capacity charge when KPLC was operating at a loss (due Sunon’s technical performance is superior to that of the in part to a drought-related recession), to show its sup- Volta River Authority (PURC 2010). The Ministry of port for the country and signal its interest in a second Energy of Togo (2010), meanwhile, has indicated that Cen- contract. Iberafrica later secured a second contract, albeit trale thermique de Lomé is expected to be more efficient after even further reductions were negotiated and passed than its state-owned counterpart. In terms of availability, by the electricity regulator.22 For Rabai, another project in IPPs in Kenya had an average availability of approximately Kenya, sponsors chose to continue work on the project CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 385 in the midst of countrywide protests following the elec- are expected to finance projects with the understanding tion in December 2007 (Aldwych International 2010a). that periodically it will be necessary to have a restructur- ing, the outcome of which is uncertain, the result will be to eliminate the availability of nonrecourse financing.” Project refinancing Given the already low level of availability of financing in A final area in which IPPs may yield greater balancing in Africa, this should be avoided. terms of development and investment outcomes is in the A host-country government’s willingness to share risks refinancing of projects, evidence for which can be seen in over the life of an IPP may thus be pivotal to the long-term the Songas project in Tanzania. Possible refinancing in the sustainability of projects. That said, strategic management case of IPTL, with the government of Tanzania proposing does not occur in a vacuum. Often, the government is not to buy the project’s outstanding debt equity, could also only an active counterparty but may even, as evidenced in lead to what may be perceived as a more balanced out- the refinancing of IPTL, initiate strategic management. come. A government buyback of IPTL would make a one- Other government-led initiatives include the government of time payment on behalf of the utility and then transfer Tanzania’s buying down of Songas’ allowance for funds used ownership of the asset to TANESCO, which may subse- during construction. The myriad project-level factors are quently decide to convert the plant to run on natural gas. summarized in table 21.4. Through this transaction, the capacity charge will fall to a token amount, and following conversion to gas, the energy LESSONS LEARNED FROM IPPs IN charge will drop from $9 million–$12 million to $1 mil- SUB-SAHARAN AFRICA lion–$1.5 million a month. The PPA would also be termi- nated and a new agreement drafted, one that would offer Despite numerous challenges, a number of Sub-Saharan customers discounted tariffs. African countries have managed to attract and sustain pri- Project refinancing has limited application and must vate investments in IPPs. More than 20 large IPPs have be dealt with carefully during the project negotiation. As taken root in eight countries, and a number of smaller, pri- one banker candidly indicated, “If project finance bankers vate projects have also been developed. While some IPPs Table 21.4 Elements Contributing to Successful IPP Investments Element Details Favorable equity partners ■ Local capital/partner contribution, where possible ■ Risk appetite for project ■ Experience with developing-country project risk ■ Involvement of a DFI partner and/or host-country government ■ Reasonable, fair return on equity ■ Firms with development origins Favorable debt arrangements ■ Competitively priced financing or involvement of DFIs ■ Local capital/markets to mitigate foreign exchange risk ■ Flexibility in terms and conditions (possible option to refinance) Secure and adequate revenue stream ■ Commercially sound metering, billing, and collections by the utility ■ Robust PPA (stipulates capacity and energy charges as well as dispatch, fuel metering, interconnection, insurance, force majeure, transfer, termination, change of law provisions, refinancing arrangements, dispute resolution) Credit enhancements and security arrangements ■ Guarantees: sovereign guarantees, partial risk guarantees ■ Insurance: political risk insurance ■ Cash: escrow accounts, letters of credit, liquidity facilities Positive technical performance ■ High level of technical performance (including availability) ■ Sponsors anticipate and mitigate potential conflicts (especially related to O&M and budgeting) Strategic management and relationship building ■ Sponsors work to create good image in country through political relationships, development funds, and effective communications and strategically manage their contracts, particularly in the face of exogenous shocks and other stresses Source: Authors. 386 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS have encountered some contract changes, nearly all have and DFIs were less likely to unravel, leading to one of two survived and are countributing to social and economic conclusions: these projects may have been more balanced development. from the outset and when an exogenous stress struck, or At the country level, factors such as a favorable invest- they may have been better equipped to resist host-country ment climate, clear policy and regulatory frameworks, and government pressure. local availability of cost-competitive fuels clearly help create In Sub-Saharan African countries in which an imbalance successful IPPs. Of growing importance are effective plan- between development and investment outcomes is per- ning, procurement, and contracting policies and practices. ceived, there is evidence that IPP contracts are more Kenya provides an example of how responsibilty for these unlikely to unravel. The incidence of such unraveling, how- functions can be effectively allocated and institutionalized. ever, does not necessarily signal the end of a project’s oper- Although evidence is not conclusive, strategic manage- ation. In all cases, efforts to close the initial gap between ment on behalf of IPP sponsors and governments, as well as investors and host-country governments’ perceptions and strong technical performance, have been used to strengthen treatment of risks must continue. Finally, the means of clos- projects. The role of firms with development origins, such ing this gap may not be only, or mainly, through increasing as Aldwych, Globeleq, and IPS, and development finance new protections such as PRGs or political risk insurance. institutions such as the IFC is increasingly important in the Rather, the solution may lie in systematic treatment of the successful development and operation of new IPPs in Sub- numerous elements contributing to success at the country Saharan Africa. Projects with participation of these firms and project level defined in this chapter. CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 387 ANNEX 21A PROJECT SPECIFICATIONS each of the regions are presented in alphabetical order. Under each country, IPPs are then presented in chrono- Detailed specifications are presented for each IPP discussed logical order, based on the date when the IPP first came in this chapter. Information is first grouped into regions online. with East Africa followed by West Africa. Countries within East Africa Kenya (1 of 5 projects) Project Westmont Size 46 MW Cost US$65 million $ per kW US$1,413 Fuel/technology Kerosene/gas condensate/gas turbine (barge mounted) ICB None, but selective international tender conducted Contract BOO, 7 years Debt/equity — DFI in equity and debt None Local participation in equity and debt None Equity partners (country of origin; % of each shareholder) Westmont (Malaysia, 100%) has sought to sell plant since 2004 Lenders — Credit enhancements, security arrangements None Project tender, COD 1996, 1997 Contract change Not a contract change per se, but firm failed to negotiate a second contract after its 7 year contract ended in 2004 due to failure to agree on tariffs Fuel arrangement Originally Westmont to procure fuel and then pass through to utility, however, following dispute with fuel supplier about taxes after the first year of operation, utility took over procurement Kenya (2 of 5 projects) Project Iberafrica Size 108.3 MW (44, 12, 52.3 brought on, respectively) Cost US$35 million (only for first 56 MW) $ per kW NA Fuel/technology HFO/medium speed diesel engine ICB None Contract BOO, 7 years, 15 years, 25 years Debt/Equity 72/28 DFI in equity and debt None Local participation in equity and debt Yes (equity and debt) Equity partners (country of origin; % of each shareholder) Union Fenosa (Spain, 80%), KPLC Pension Fund (Kenya, 20%) since 1997 Lenders Union Fenosa (US$12.7 million in direct loans and guaranteed US$20 million); KPLC Staff Pension Fund (US$9.4 in direct loans and guaranteed US$5 million through a local Kenyan bank). Credit enhancements, security arrangements None Project tender, COD 1996, 1997/1999, 2000/2008, 2009 Contract change Yes, Iberafrica reduced the capacity charge of its first PPA by 37% in April 2002 and then to 59% of the original PPA in September 2003.a Fuel arrangement Iberafrica buys fuel and passes cost through to KPLC based on the units generated and specific consumption parameters agreed on in the PPA Kenya (3 of 5 projects) Project OrPower 4 Size 13 MW + 35 MW Cost US$105b million $ per kW NA Fuel/technology Geothermal ICB Yes (continued next page) 388 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS East Africa (continued) Contract BOO, 20 years Debt/equity NA DFI in equity and debt DEG Local participation in equity and debt None Equity partners (country of origin; % of each shareholder) Ormat (100%) from 1998 to 2008. Lenders — Credit enhancements, security arrangements MIGA guarantee, a stand-by letter of credit, covering several months billing (although only finalized at end-2006) Project tender, COD 1996, 2000–2009 Contract change Yes, tariff for the second phased (35 MW) reduced Fuel arrangement The only fuel arrangement per se is that OrPower 4 was granted a Geothermal Resource License from the government, to which it pays a royalty of sorts (of US$0.004/kWh or USc 0.4/kWh) Kenya (4 of 5 projects) Project Tsavo Size 75 MW Cost US$86 million $ per kW US$1,133 Fuel/technology HFO/medium speed diesel engine ICB Yes Contract BOO, 20 years Debt/equity 78/22 DFI in equity and debt Yes (equity and debt) Local participation in equity and debt None Equity partners (country of origin; % of each shareholder) Cinergy and IPS jointly owned 49. 9%, Cinergy sold to Duke Energy in 2005, CDC/Globeleq (U.K., 30%), Wartsila (Finland, 15%), IFC (5%) retain remaining shares since 2000 Lenders IFC own account (US$16.5 million), IFC syndicated (US$23.5 million); CDC own account (US$13 million), DEG own account (€11 million), DEG syndicated (€2 million) Credit enhancements, security arrangements Letter of comfort provided by government, and escrow account, equivalent to 1 month capacity charge, and a stand-by letter of credit, equivalent to 3 months billing Project tender, COD 1995, 2001 Contract change None, but was pressured to lower tariff Fuel arrangement Tsavo buys fuel and passes cost through to KPLC based on the units generated and specific consumption parameters agreed on in the PPA Kenya (5 of 5 projects) Project Rabai Size 90 MW Cost US$155 million $ per kW US$1,722 Fuel/technology HFO ICB Yes Contract BOOT, 20 years Debt/equity 75% debt, (5% subordinated debt, 25% equity) DFI in equity and debt Yes Local participation in equity and debt No Equity partners (country of origin; % of shareholder) Aldywch: 34%, BWSC (Danish, but owned by Mitsui of Japan): 25.5%, FMO: 20%, IFU (Danish bilateral lender): 20% Lenders DEG: 15%, FMO: 25%, EAIF: 25%, Proparco: 25%, EFP (European Financing Partners): 10% Credit enhancements, security arrangements Support letter from government of Kenya (covers political risk but falls short of being an outright guarantee), and KPLC-issued letter of credit equivalent to 5 months of capacity (debt service, fixed costs and equity returns) payments and 2 months of fuel payments Project tender, COD 2006, 2009 Contract change None Fuel arrangement Fuel supply agreement with Kenol of Kenya (continued next page) CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 389 East Africa (continued) Tanzania (1 of 2 projects) Project IPTL Size 100 MW Cost US$120 million $ per kW US$1,200 Fuel/technology HFO/medium speed diesel engine ICB None Contract BOO, 20 years Debt/equity 70/30 (albeit ratio disputed) DFI in equity and debt None Local participation in equity and debt Yes (in-kind equity participation) Equity partners (country of origin; % of each shareholder) Mechmar (Malaysia, 70%), VIP (Tanzania, 30% in kind), both have sought to sell shares Lenders Two Malaysian-based banks (Bank Bumiputra Malaysia Berhad—now Bank Bumiputra Commercial Bank—and SIME Bank) auctioned debt to Standard Chartered Bank Credit enhancements, security arrangements Sovereign guarantee, liquidity facility equivalent to 4 months capacity charge (but not yet established) Project tender, COD 1997, 2002 Contract change Yes, postarbitration, monthly capacity charges lowered from US$3.6 million to US$2.6 million Fuel arrangement IPTL imports fuel, which is a pass-through to the utility Tanzania (2 of 2 projects) Project Songas Size 180 MWc Cost US$316 milliond $ per kW US$2,313 (for first 115 MW) and US$769 (for 65 MW expansion) Fuel/technology Natural gas/open cycle ICB Yes Contract BOO, 20 years Debt/equity 70/30 for 115 MW 100% equity financed for 65 MW expansion, which has since been refinanced DFI in equity and debt Yes Local participation in equity and debt Yes (TANESCO, TDFL) Equity partners (country of origin; % of each shareholder) TransCanada sold majority shares to AES (USA) in 1999 and AES sold majority shares to Globeleq (UK) in 2003.e All preferred equity shares were converted into “Loan Notes” in June 2009. Only common shares remain. Lenders World Bank (US$136 million), EIB (US$55 million), Sida (US$15 million) Credit enhancements, security arrangements Escrow account: for first 115 MW, with the government matching every US$1 spent by the project company; liquidity facility equivalent to 4 months capacity charge for the first 3 years, declining to 2 months starting in year 4 through the remaining years of the contract Project tender, COD 1994, 2004 Contract change The changes to Songas’ PPA include: US$103 million AFUDC buy- down; the use of escrow funds to buy down the AFUDC; freezing the liquidity facility, which was meant to be replenished and accessed only in the case of partial or nonpayment by TANESCO; and preferred equity conversion to “loan notes” Fuel arrangement Songo Songo gas provided to project company at a rate of US$0.55/ MMBtu for turbines I–V and at US$2.17 MMBtu for turbine VI Uganda (1 of 2) Project Namanve Power Plant/Jacobsen Uganda Power Plant Limited Size 50 MW Cost US$74 million at inception $ per kW US$1,488 Fuel/technology HFO ICB Yes (continued next page) 390 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS East Africa (continued) Contract BOOT, 6 year Debt/equity — DFI in equity and debt Yes (debt) Local participation in equity and debt Yes (debt) Equity partners (country of origin; % of each shareholder) Jacobsen Elektro (Norway, 100%) Lenders Loan from Standard Chartered Bank UK, partly guaranteed by GIEK (the Norwegian guarantee agency), commercial loan from Stanbic Bank (Uganda) Ltd. (backstopped by Standard Chartered Bank UK); subordinate loan from parent company Jacobsen Elektro AS; NORAD grant Credit enhancements, security arrangements GIEK guarantee backed up by government of Uganda Project tender, COD COD 2008 Contract change None Fuel arrangement Openly bid for contract; not part of license (plant availability should be >90%) Uganda (2 of 2) Project Bujagali Size 250 MW Cost US$860 million $ per kW US$3,440 Fuel/technology Hydro ICB Yes Contract BOT, 30 years Debt/equity 78/22 DFI in equity and debt Debt Local participation in equity and debt Yes, government of Uganda Equity partners (country of origin; & % of each shareholder) Sithe Global (58%), IPS/Aga Khan (32% but 50.1% voting), government of Uganda (10%) Lenders African Development Bank, EIB, PROPARCO, AFD, DEG, KfW, FMO, and IFC as DFIs, NedBank, Absa Capital, Standard Chartered Bank, and Fortis as commercial banks covered by the PRG Credit enhancements, security arrangements Government guarantee, MIGA, PRG/IDA Project tender, COD 2005, expected 2012 Contract change None Fuel arrangement — West Africa Côte d’Ivoire (1 of 2) Project Compagnie Ivoirienne de Production d’Electricité (CIPREL) Size 210 MW Cost US$105.6 millionf $ per kW US$503 Fuel/technology Natural gas/open cycle ICB None Contract BOOT, 19 years Debt/equity — DFI in equity and debt Yes (equity and debt) Local participation in equity and debt None Equity partners (country of origin; % of each shareholder) SAUR International, with 88% joint venture (JV) between French SAUR Group owned by Bouygues, 65% and EDF, 35%) BOAD, PROPARCO, and IFC holding the remaining 12%; in 2005 all shares sold to Bouygues (France, 98%), except BOAD (2%) Lenders World Bank Credit enhancements, security arrangements None Project tender, COD 1993, 1995 Contract change None Fuel arrangement Government procures fuel (continued next page) CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 391 West Africa (continued) Côte d’Ivoire (2 of 2) Project Azito Size 288 MWg Cost US$225 million $ per kW US$781 Fuel/technology Natural gas/open cycle ICB Yes Contract BOOT, 24 years Debt/equity 70/30 DFI in equity and debt Yes (debt) Local participation in equity and debt None Equity partners (country of origin; % of each shareholder) Cinergy (JV between Swiss ABB, 50% and French EDF, 50%) Holds 65.7% of shares, CDC/Globeleq (11%), and IPS (23%) Lenders IFC, CDC, FMO, DEG, AfDB, Societe Generale, and other European commercial banks Credit enhancements World Bank partial risk guarantee Sovereign guarantee Security arrangements Escrow account equivalent to 1 month capacity charge Project tender, COD 1996, 2000 Contract change None Fuel arrangement Government procures fuel Ghana (1 of 3) Project Takoradi II Size 220 MWh Cost US$110 million $ per kW US$500 Fuel/technology Light crude oil/single cycle ICB None Contract BOOT, 25 years Debt/Equity Financed exclusively with balance sheet of sponsors DFI in equity and debt No Local participation in equity and debt Yes (equity) Equity partners (country of origin & % of each shareholder) CMS (U.S., 90%), VRA (Ghana, 10%), CMS sold shares to TAQA (UAE, 90%) in 2007 Lenders Financed exclusively with balance sheet of sponsors Credit enhancements Sovereign guarantee Security arrangements US$3 million letter of credit provided by government Project tender, COD 1998, 2000 Contract change Failure to develop second phase (110 MW), investors cite the lack of government guarantees (granted in the first phase of Takoradi II), meanwhile government has indicated that the EPC costs are too high and further development remains at standstill Fuel arrangement Government procures fuel Ghana (2 of 3) Project Sunon Asogli Power Plant Size 200 MW Cost — $ per kW — Fuel/technology Combustion engine ICB No Contract BOO, 20 years Debt/equity — DFI in equity and debt None Local participation in equity and debt Local strategic investor, Togbe Afede XIV Equity partners (country of origin; % of each shareholder) Shenzhen (China, 100%) Lenders — Credit enhancements, security arrangements — Project tender, COD — (continued next page) 392 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS West Africa (continued) Contract change Related to fuel Fuel arrangement Inability to reach conclusion about fuel agreement, which has prevented plant from coming online Ghana (3 of 3) Project Bui hydro Size 400 MW Cost US$622 million $ per kW US$1,555/kW Fuel/Technology Hydro ICB No Contract — Debt/Equity — DFI in equity and debt None Local participation in equity and debt — Equity partners (country of origin & % of each shareholder) Sinohydro (China, 100%) Lenders — Credit enhancements and security arrangements — Project tender, COD 2005, expected online in 2012 Contract change None Fuel arrangement Hydro Nigeria (1 of 4) Project AES Barge Limited Size 270 MW Cost US$240 million $ per kW US$888 Fuel/technology Natural gas/open cycle (barge mounted) ICB None Contract BOO, 13 years Debt/equity — DFI in equity and debt None Local participation in equity and debt Yes (equity) Equity partners (country of origin; % of each shareholder) Enron (U.S., 100%) sold to AES (95%) and YFP (Nigeria, 5%) in 2000 Lenders — Credit enhancements, security arrangements OPIC political risk insurance Sovereign guarantee, US$60 million letter of credit from Ministry of Finance Project tender, COD 1999, 2001 Contract change Yes, initial plant size increased from 90 MW to 270 MW (9 units of 30 MW each) and change in the fuel from liquid fuel to natural gas, both of which had the effect of reducing the capacity charge, current contract changes under discussion involve among other things the availability of deficiency payment, meanwhile tax exemption certificate has been withheld by government for the duration of the project Fuel arrangement Utility arranges fuel Nigeria (2 of 4) Project Okpai Size 450 MW Cost US$462i $ per kW — Fuel/technology Natural gas/combined cycle ICB None Contract BOO, 20 years Debt/equity 100% equity financed DFI in equity and debt None Local participation in equity and debt Yes (equity and debt) Equity partners (country of origin; % of each shareholder) Nigerian National Petroleum Corporation (Nigeria, 60%), Nigerian Agip Oil Company (Italy, 20%), and Phillips Oil Company (U.S., 20%) maintained equity since 2001 (continued next page) CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 393 West Africa (continued) Lenders Provided by equity partners Credit enhancements, security arrangements PPA backed by Nigerian Petroleum Development Company’s oil revenues Project tender, COD 2001, 2005 Contract change Ongoing negotiations related to investment costs which rose by US$150 million, to US$462 million; although plant is producing power, due to the dispute, full payment is not being made by utility Fuel arrangement Project company provides fuel Nigeria (3 of 4) Project Afam IV Size 630 MW Cost — $ per kW — Fuel/technology CCGT ICB No Contract BOO, 20 years Debt/equity 100% equity DFI in equity and debt None Local participation in equity and debt Yes, NNPC Equity partners (country of origin; % of each shareholder) NNPC (Nigeria, 55%), Shell (U.K./Netherlands, 30%), Elf (Total) (France, 10%), Agip (Italy, 5%) Lenders Credit enhancements, security arrangements PPA backed by Nigerian Petroleum Development Company’s oil revenues Project tender, COD 2000, 2007 Contract change No Fuel arrangement Project company provides fuel Nigeria (4 of 4) Project Aba Integrated Power Project Size 140 MW Cost US$385 $ per kW US$2,750/kW (includes distribution investment) Fuel/t\Technology Gas fired open-cycle power ICB No Contract BOO Debt/equity DFI in equity and debt No Local participation in equity and debt Yes (equity) Equity partners (country of origin; % of each shareholder) Geometric Power Limited (Nigeria) Lenders Diamond Bank, Stanbic IBTC Credit enhancements, security arrangements Project tender, COD 2005, anticipated COD end-2010 Contract change No Fuel arrangement Senegal (1 of 2) Project GTi Dakar Size 52 MW Cost US$65 million $ per kW US$1,250/kW Fuel/technology Complete cycle/diesel/natfa ICB Yes Contract BOOT, 15 years Debt/equity 75/25 DFI in equity and debt Yes (both) Local participation in equity and debt None Equity partners (country of origin, % of each shareholder) GE Capital Structured Finance Group (U.S.), IFC, Edison (Italy) Lenders IFC and Credit Commercial de France (continued next page) 394 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS West Africa (continued) Credit enhancements, security arrangements Government guarantee, credit insurance through a guarantee program of SACE, the Italian export credit agency, and a partial interest subsidy through the Mediocredito Central Subsidy Department, escrow account Project tender, COD 1996, 1999/2000 Contract change None Fuel arrangement Senegal (2 of 2) Project Kounoune I Size 68 MW Cost US$110 million $ per kW US$1617/kW Fuel/technology HFO diesel ICB Yes Contract BOO, 15 years Debt/equity DFI in equity and debt Yes (debt) Local participation in equity and debt Yes (debt) Equity partners (country of origin; % of each shareholder) Mitsubishi (Japan), Matelec S.A.L (Lebanon) Lenders IFC, Proparco, AfDB, BOAD, and CBAO Credit enhancements, security arrangements Government guarantee, a PRG although never signed by the government, a letter of credit from Senelec Project tender, COD 2003, 2008 Contract change None Fuel arrangement During the project negotiations, the structures of the FSA and PPA were changed to turn the PPA into a tolling agreement. Togo (1 of 1) Project Extension of Central Thermique de Lome (CTL) Size 100 MW Cost US$196 million (including 18 km transmission lines rehabilitation and soil decontamination investment) $ per kW US$1,960 Fuel/technology Triple fuel ICB No Contract BOOT, 25 years Debt/equity 75/25 DFI in equity and debt Yes (debt) Local participation in equity and debt None (but contract stipulates up to 25% of equity must be sold to local/Togolese) Equity partners (country of origin; % of each shareholder) Contour Global (U.S., 80%), IFC (20%) Lenders OPIC Credit enhancements, security arrangements Government guarantee, OPIC guarantee, escrow account for 1 month of full operation, letter of credit of 2 months of full operation, (which become one month after 2 years of no incident of payment, and ceases to exist after 4 years of no incident) Project tender, COD Due online in July 2010 Contract change No Fuel arrangement The buyer (utility provides fuel) Source: Authors. Notes: — = Not available. BOT = build, own, transfer. a. Furthermore, although not a contract change per se, the value of the capacity charge for Iberafrica’s second PPA was 50 percent that of the first PPA. b. US$105 includes only the loan portion for the 35 MW part of the plant (Ormat, personal communication, 2010). c. There was considerable evolution in terms of the planned capacity for the plant, from 60 MW to the current 180 MW. d. Songas project costs include refurbishment of gas wells, a new gas processing facility, pipeline construction and fuel conversion of the existing power station (Ubungo), in total amounting to US$266 million, and an additional US$50 million for expansion in terms of two additional turbines (total 65 MW) and related infrastructure. The expansion was financed entirely by equity. A rough estimate for the electricity generation component would be 40 percent of project costs or US$130 million, based on US$35 million for refurbishment and fuel conversion of existing turbines, US$45 million assumed loans on existing turbines, and US$50 million for expansion. (continued next page) CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 395 West Africa (continued) e. Due to complexity, turnover is detailed in this footnote: Ocelot (Canada), TransCanada (Canada), Tanzania Petroleum Development Corpora- tion, TPDC (Tanzania), TANESCO (Tanzania), Tanzania Development Finance Company, TDFL (Tanzania, sponsored by EIB), IFC (multilateral), DEG (German), CDC (UK) were shareholders by 1996, with TransCanada the majority shareholder; IFC and DEG sold shares to CDC in 1997/8; TransCanada sold shares to AES (USA) in 1999; Ocelot/PanOcean sold shares to AES in 2001; AES sold majority shares to Globeleq (UK) and FMO (Holland) in 2003. After the AES sale, equity shares and associated financial commitments (expressed in US$ million) in Songas were as follows: Globeleq: US$33.8 (56%); FMO: US$14.6 (24%); TDFL: US$4 (7%); CDC: US$3.6 (6%); TPDC: US$3 (5%) and TANESCO: US$1 (2%). This does not reflect the additional US$50 million that Globeleq committed for the expansion. f. Investment cost €87.8m or 57.6 billion CFA, with the average 1994 conversion of US$ to CFA, 545,100. g. The initial project concept included specifications to raise capacity to 420 MW. h. The initial project concept included specifications to add a second phase of 110 MW and convert to combined cycle, however, lack of funding has limited the completion of this phase. i. Project costs include the gas infrastructure. NOTES continuous power producers, privately owned by the sugar mills. With installed capacity of 40 MW, roughly equal to 1. The standard model for power sector reform has been the IPP shortfall, these continuous producers also have roughly defined as a series of steps that move vertically long-term take or pay contracts with the state (Bergesen integrated utilities toward competition, generally includ- 2007; World Bank 2007a). ing the following activities: corporatization, commercial- ization, passage of the requisite legislation, establishment 5. Centrale thermique de Lomé is a preexisting 90 MW of an independent regulator, introduction of IPPs, facility for which a rehabilitate-own-transfer contract was restructuring/unbundling, divestiture of generation and provided to Electro Togo in 2001. In 2006, however, when distribution assets, and introduction of competition a power crisis intensified and the rehabilitation had not (Bacon 1999; Adamantiades and others 1995; Besant-Jones yet been completed, the government of Togo opted to ter- 2006; Williams and Ghanadan 2006). Although this model, minate Electro Togo’s contract and initiate a BOOT based largely on the early power sector reforms carried out directly with Contour Global, which had previously been in Chile, Norway, and the United Kingdom, came to repre- involved in discussions with Electro Togo (Ministry of sent a standard, it is arguable that not all steps were relevant Energy of Togo 2010). to conditions on the ground in most developing countries 6. A portion of the shares of the two Kenyan utilities, KPLC (Gratwick and Eberhard 2008b). and Kengen, however, are listed on the Nairobi Stock 2. Exceptions are Côte d’Ivoire and Tanzania, where IPPs Exchange. Côte d’Ivoire’s state utility has been under a pri- are contributing more than 50 percent to overall electricity vate concession contract since 1990—an arrangement that is production. Togo’s first IPP, Centrale thermique de Lomé, expected to continue until 2020. TANESCO in Tanzania and will also supply the majority of electricity for that country. KPLC in Kenya also have had private management contracts. 3. Two different categories of DFIs should be distinguished, In Senegal, there was an attempt (albeit aborted) to privatize namely, those that lend on commercial terms and largely to Senelec in 1999 through a concession with Hydro-Québec private companies (for example, FMO, PROPARCO, DEG, and Elyo (the concession was annulled by the president in and IFC), and the multilateral development banks (such as 2001). Uganda unbundled its national utility into separate the World Bank and the African Development Bank) that generation, transmission, and distribution companies and lend on concessionary terms and primarily to public sector entered into long-term private concession in generation projects. It is the latter that refocused on infrastructure (Eskom) and distribution (Umeme-Globeleq). Finally, (Rudo 2010a). Nigeria has expressed an intention to unbundle and priva- tize but has not yet fully realized this ambition. 4. Although Mauritius has four IPPs (which, at approxi- mately 200 MW combined, account for about 37 percent of 7. Alternatives to strictly independent regulation that may installed capacity and a little less than 25 percent of pro- provide a better match to a country’s regulatory commit- duction, as of end-2005), the country has not been ment and institutional and human resource capacity (for included in this sample. The IPPs, which are all cogenera- example, regulatory contracts, outsourcing of regulatory tion plants, provide power and steam to the country’s sugar functions, expert panels, and regional regulators) are mills throughout the crop season, reducing their contribu- increasingly being considered (Eberhard 2007). tion to the state-owned utility by about 30 percent. During 8. Domestic gas reserves were used for IPPs in Côte this time, the shortfall in production is made up by seven d’Ivoire. Unlike for the other countries, however, this use 396 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS did not represent the establishment of a new gas infra- Briceno-Garmendia (2010) report on the general phenom- structure. An attempt, led by the government, was also enon of non-OECD funders, including the emergence of made to exploit stranded gas reserves in Ghana’s Osagyefo increasingly significant funding flows from China to Sub- Barge project. Saharan Africa. 9. Stakeholders in Nigeria have seen costs escalate, causing 18. In 2009, 10 years after financial close, DEG arranged a the utility to withhold payments. In Tanzania costs have also $105 million loan (Ormat 2009). escalated, but for reasons unrelated to the project itself. 19. In 2005, it was found that the then managing director of 10. At the time of writing, Kenya had only 60 MW still in KPLC, Samuel Gichuru, received $2 million from Westmont. service, and that was expected to be retired by November 20. In addition, a PRG for Kounoune I was approved by the 2010. World Bank but was ultimately not signed by the govern- 11. In terms of international norms, however, it should be ment of Senegal, because the lender that the PRG would noted that Tanzania’s cost of unserved energy (CUE) is low. have covered was willing to fund the project without the South Africa’s CUE is approximately US$10/kWh, which is PRG (IFC 2010a). in line with the CUE in many industrialized countries 21. World Bank 2010. For IPPs in Central Asia, however, (Global Energy Decisions 2007). and for cross-border projects such as those in Laos and 12. On the other hand, given the high cost of mounting a Thailand (Nam Theum II), a PRG or other “strong” credit bid, having an excessive amount of bidders would also not enhancements would be employed. In the middle-income be desirable as some will not bid if they perceive their Latin American countries cited above, privatization trajec- chances to be low (Rudo 2010a). tories meant there were few IPPs, and in those few cases (for 13. Togo’s next IPP, a 24 MW wind farm, is being procured example, Colombia), bilateral and multilateral institutions through an ICB (Ministry of Energy of Togo 2010). helped back investments (Rudo 2010a). 14. Tanzania used the net-back calculation method to estab- 22. The fact that Iberafrica was not project-financed meant lish the gas price for the first 142 MW owned by Songas. that the company had greater flexibility to change the pay- Although it is viewed to be low (“US$0.55/Mcf ”), most of ment streams. the costs associated with the gas infrastructure are paid sep- arately as part of the capacity charge. In typical projects, these would have been part of the gas price (i.e., REFERENCES US$4.20/Mcf). The same gas price is not offered to IPTL. Adamantiades, A. G., et al. 1995. “Power Sector Reform in Under the IPTL PPA, fuel is a pass-through. TANESCO has Developing Countries and the Role of the World Bank.” negotiated with PanAfrican Energy and TPDC to book 100 Paper prepared for 16th Congress of the World Energy MMcfd at negotiated wellhead prices (in favor of the addi- Council, Tokyo. World Bank, Industry and Energy tional 40 MW owned by Songas, 145 MW owned by Department, Washington, DC. TANESCO, and the next 100 MW plant, which may be IPTL Africa Electra. 2005. “The Role of Local Banks in Financing or any other plant) starting at US$1.98/Mcf, escalated at 2.2 Power Projects.” Africa Electra 30. percent annually for the first five years from July 2007, then Aldwych International. 2010a. Personal communication, to US$3.5/Mcf thereafter, escalated at USCPI. In summary, May 5–6. the gas prices are discounted market prices of an alternative fuel (EWURA 2010a). ———. 2010b. Personal communication, June 9. 15. Projects such as Osagyefo Barge in Ghana and Ibom APEC (Asia-Pacific Economic Cooperation) Energy Working and Omoku in Nigeria are sometimes loosely termed Group. 1997. “Manual of Best Practice Principles for Inde- IPPs. In Ghana, private participation was expected (but pendent Power Producers.” APEC Secretariat, Canberra. not achieved), and in Nigeria, projects, although inde- Bacon, R. 1999. “A Scorecard for Energy Reform in Devel- pendent of the national utility, have been led entirely by oping Countries.” World Bank, Finance Private Sector the Rivers and Akwa Ibom state governments, respec- and Infrastructure Network, Washington, DC. tively. All of these projects are excluded from the analysis Bergesen, C. 2007. Mauritius. Columbus, OH: McGraw Hill. presented here. Besant-Jones, J. E. 2006. “Reforming Power Markets in 16. At the time of writing, the level of equity (including Developing Countries: What Have We Learned?” Energy local equity) is under dispute in Tanzanian court. and Mining Sector Board Discussion Paper 19, World 17. The emergence of two Chinese firms, Shenzhen and Bank, Washington, DC. Sinohydro, in Sunon Asogli and Bui hydro, respectively, in Deloitte Touche Tohmatsu Emerging Markets Ltd. and Ghana could represent a new element and be a significant Advanced Engineering Associates International. 2003. source of investment for other African IPPs. Foster and “Energy Program South Asia Generation Pricing Study.” CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 397 South East Asia Regional Initiative for Energy, USAID, IPS (Industrial Promotion Services). 2010. Personal com- Washington, DC. munication, May 24. Nairobi, Kenya. Eberhard, A. 2007. “The Independence and Accountability Hoskote, M. 1995. “Independent Power Projects (IPPs): An of Africa’s Infrastructure Regulators: Reassessing Regu- Overview.” Energy Note 2. In Energy Themes, ed. R. English. latory Design and Performance.” Paper presented at Washington, DC: World Bank. AFUR Annual Conference, Livingstone, Zambia, April International Energy Agency. 2006. World Energy Outlook 25–26. 2006. Paris: OECD/IEA. Eberhard, A., and I. Malgas. Forthcoming. “Hybrid Power Malgas, I., K. N. Gratwick, and A. Eberhard. 2007a. “Moroc- Markets in Africa: Generation Planning, Procurement can Independent Power Producers: African Pioneers.” and Contracting Challenges.” Energy Policy. Journal of North African Studies 13 (1): 15–36. Eberhard, A., and M. Shkaratan. 2010. “Africa’s Power Infra- ———. 2007b. “Two of a Kind: Lessons from Tunisian structure: Improving Investment, Connectivity, Reliabil- Independent Power Projects.” Journal of North African ity and Efficiencies.” Africa Infrastructure Country Diag- Studies 12 (4): 395–415. nostic, World Bank, Washington, DC. Matsukawa, T., R. Sheppard, and J. Wright. 2003. “Foreign EDF (Électricité de France). 2005. Personal communication, Exchange Risk Mitigation for Power and Water Projects January 11. in Developing Countries.” Energy and Mining Sector ERC (Energy Regulatory Commission of Kenya). 2010a. Board Discussion Paper 9, World Bank, Washington, DC. Personal communication, May 18. Nairobi, Kenya. Ministry of Energy of Togo. 2010. Personal communication, ———. 2010b. Personal communication, May 26. Nairobi, June 10–11. Lome, Togo. Kenya. Nigeria Electricity Regulatory Commission. 2010. Per- sonal communication, May 21. Abuja, Nigeria. EWURA (Energy and Water Utilities Regulatory Author- ity). 2010. Personal communication, May 18. Dar es Ormat. 2009. “Ormat Technologies Announces Closing of Salaam Tanzania. $105 Million Long-Term Senior Debt Financing for the Olkaria III Geothermal Power Project in Kenya.” Press Foster, V., and C. Briceño-Garmendia, eds. 2010. Africa’s release, March 6, Reno, Nevada. Infrastructure: A Time for Transformation. Washing- ton, DC: Agence Française de Développement and ———. 2010. Personal communication, May 17. World Bank. Phadke, A. 2007. “Mark-ups in the State Capital Cost of Private Power Projects in Developing Countries.” Global Energy Decisions. 2007. “Electrical Resource Needs Energy and Resources Group, University of California, Analysis: Adequate Reserve Margin for Development of Berkeley. Third National Integrated Resource Plan for South Africa.” Pretoria. PURC (Public Utilities Regulatory Commission of Ghana). 2010. Personal communication, May 19–21. Accra, Ghana. Globeleq. 2006. “Enabling Electricity Growth in Africa: Some Experiences.” In Financing Electricity for Growth in Regulatory Commision of Senegal. 2010. Personal commu- Africa. Tunis: Infrastructure Consortium for Africa. nication, June 11. Dakar, Senegal. Rudo, D. 2010a. Personal communication, May 14. Washington, ———. 2010a. Personal communication, May 5. DC. ———. 2010b. Personal communication, June 1. ———. 2010b. Personal communication, June 10. Washington, Gratwick, K. N., and A. Eberhard. 2008a. “An Analysis of DC. Independent Power Projects in Africa: Understanding Sinclair, S. 2007. “The World Bank Group Risk Mitigation Development and Investment Outcomes.” Development Instruments.” Paper presented at Private Sector Partici- Policy Review 26 (3): 309–38. pation in Power Conference, Dead Sea, Jordan. ———. 2008b. “Demise of the Standard Model for Power UNEP (United Nations Environment Programme) and UN Sector Reform and the Emergence of Hybrid Power Economic Commission for Africa. 2006. Making Africa’s Markets.” Energy Policy 36: 3948–60. Power Sector Sustainable: An Analysis of Power Sector IFC (International Finance Corporation). 2005. Personal Reforms in Africa. New York: United Nations. communication, January 24. Williams, J. H., and R. Ghanadan. 2006. “Electricity Reform ———. 2010a. Personal communication, May 21. Washington, in Developing and Transition Countries: A Reappraisal.” DC. Energy 31: 815–44. ———. 2010b. Personal communication, June 12. Washington, Woodhouse, E. J. 2005. “The Experience with Independent DC. Power Projects in Developing Countries: Interim 398 CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS Report.” Program on Energy and Sustainable Develop- ———. 1999. “Sub-Saharan Africa Benefits from the first ment, Stanford University, Stanford, CA. IDA Guarantee for Azito.” In Project Finance and Guar- Woolf, F., and J. Halpern. 2001. “Integrating Independent antees. Washington, DC: World Bank. Power Producers into Emerging Wholesale Power Mar- ———. 2007. Private Participation in Infrastructure Data- kets.” Policy Research Working Paper 2703, World Bank, base. http://ppi.worldbank.org. Washington, DC. ———. 2010. Personal communication, May 17. World Bank. 1993. The World Bank’s Role in the Electric World Bank and USAID. 1994. “Submission and Evaluation Power Sector: Policies for Effective Institutional, Regulatory, of Proposals for Private Power Generation Projects in and Financial Reform. Washington, DC: World Bank. Developing Countries.” IEN Occasional Paper 2. World ———. 1997. “Morocco’s Jorf Lasfar Power Station.” In Proj- Bank, Washington, DC. ect Finance and Guarantees. World Bank, Washington, DC. CHAPTER 21: INDEPENDENT POWER PROJECTS IN SUB-SAHARAN AFRICA: DETERMINANTS OF SUCCESS 399 PA R T V I Improving Human Development Outcomes with Innovative Policies CHAPTER 22 Innovative Financing for Health in Rwanda: A Report of Successful Reforms C. Sekabaraga, A. Soucat, F. Diop, and G. Martin wanda has been moving very rapidly to expand R 2008, ranking it among the best-performing countries in health service delivery. It has dramatically acceler- the world (figure 22.2). This rate of decline far exceeds the ated the trend of progress on health indicators, 5 percent rate needed to meet the MDG target of reducing putting it back on track to reach the health Millennium the maternal mortality ratio by three-quarters between 1990 Development Goals (MDGs). Three interrelated innovative and 2015. reforms contributed to improvement: community-based These achievements have been the result of innovative health insurance, performance-based financing within a strategies to address some of the key challenges affecting broader framework of reform of management of human maternal mortality. The share of women delivering their resources for health, and fiscal decentralization. This chap- babies in health facilities has steadily increased, rising from ter examines these reforms, showing how they worked 28 percent of pregnant women in 2000 to 45 percent in 2008. together to improve health outcomes in Rwanda. Many challenges remain, but preliminary data from the Health indicators in Rwanda improved dramatically in Ministry of Health for 2010 suggest that this figure has risen recent years, and projections for the coming years are pos- to two-thirds of all pregnant women; the finding should be itive. After an initial surge following the genocide, under- validated by the 2010 Demographic and Health Survey. five mortality (the probability of death per 1,000 live These successes can be attributed largely to an increase in births) significantly decreased, falling from 196 in 2000 to the use of essential health interventions, particularly high- 152 in 2005 and 103 in 2007 (figure 22.1). The infant mor- impact interventions that are critical in reducing disease tality rate decreased from 107 per 1,000 live births in 2000 burden in developing countries, including immunization, to 86 in 2005 and 62 in 2007. Improvements were particu- assisted deliveries, family planning, and the use of insecticide- larly significant among the poor, with the under-five treated bed nets to prevent malaria. Rwanda has maintained mortality rate among the poorest quintile declining by a very high and equitable coverage of vaccination against 50 deaths per 1,000 between 2005 and 2008 compared with avoidable childhood diseases since 2000: immunization rates, 38 deaths per 1,000 live births among the richest quintile. at 95 percent in 2008, are among the highest in Sub-Saharan The annual rate of decrease achieved between 2005 and Africa. Major progress has also been made in extending the 2008 was 12.2 percent—far greater than the 9.7 percent coverage of vitamin A supplementation among children and needed to meet the MDG 4 target of 50 in 2015. Rwanda is women, through a mass campaign and integration into rou- thus back on track to reach the health MDGs. tine health facility services. Treatment of acute respiratory Since 2000 the maternal mortality ratio has declined at infections of children has also increased, including among an annual rate of 12.1 percent to reach 383 per 100,000 in the poor. 403 Figure 22.1 Under-five Mortality Rate in Rwanda, 1990–2015 250 200 196 Deaths per 1,000 151 152 150 MDG 103 100 50 50 0 90 92 94 96 98 00 02 04 06 08 10 12 14 19 19 19 19 19 20 20 20 20 20 20 20 20 Sources: Measure DHS and ICF Macro 1995, 2000, 2005, 2008. Figure 22.2 Maternal Mortality Ratio and Facility-based Deliveries in Rwanda, 2000–15 100 Maternal mortality ratio (per 100,000) 1,200 1,071 90 Facility-based deliveries (%) 1,000 80 750 70 800 66 60 600 50 45 40 400 383 26 28 268 30 200 20 10 0 0 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Maternal mortality ratio Facility-based deliveries Sources: Measure DHS and ICF Macro 2005, 2008. Improvements in the use of women’s health services are HIV/AIDS epidemic. Malaria incidence and mortality have also evident, with significant increases in the proportion of declined dramatically, largely as the result of increased use assisted birth deliveries and the number of emergency of insecticide-treated nets, which among children under obstetrical cases referred. The use of modern contraceptives five rose from 11 percent in 2005 to 56 percent in 2007. As increased from 3 percent in 2000 to 27 percent in 2007, one a result, malaria-specific mortality was cut in half. Rwanda of the fastest increases ever observed. The proportion of has thus moved from being a country where malaria was women having at least one antenatal consultation rose from endemic to one focusing on eliminating malaria as a pub- 58 percent in 1995–2000 to 96 percent in 2000–07. The pro- lic health problem. The HIV/AIDS epidemic has been con- portion of assisted deliveries increased from 39 percent in tained, with 3 percent of the population affected and more 2005 to 52 percent in 2007. than 60 percent of patients needing treatment receiving Major progress has also been made in controlling com- highly active antiretroviral therapy. Knowledge of municable diseases—including malaria, a prime cause of HIV/AIDS is better in Rwanda than in any other Sub- morbidity and mortality in Rwanda—and containing the Saharan African country: nationwide 54 percent of women 404 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS and 58 percent of men had comprehensive knowledge of planning and ensuring free access to information, educa- HIV/AIDS in 2005.1 tion, and contraceptive services. This progress occurred in a context in which annual total per capita health expenditures doubled, from $17 REFORM STRATEGIES: COMMUNITY-BASED to $34 between 2003 and 2006. In 2006 total health HEALTH INSURANCE, PERFORMANCE-BASED expenditures reached 10.7 percent of gross domestic FINANCING, AND FISCAL DECENTRALIZATION product (GDP) (one of the highest levels of health expen- ditures observed in low-income countries), up from Rwanda has pioneered major programmatic, organiza- 3 percent in 2002. tional, and health financing reforms, which are increasing The overall share of public expenditures allocated to the accountability of major actors in the health sector. health has remained stable, with most of the increase com- Rwanda has a long history of centralized management ing from donors. Private expenditures were about $9.40 per structures with a clear hierarchy and a relatively low level of capita, and domestic public expenditure about $6.30 per corruption. It has progressively moved toward a modern capita, with donors contributing $17.70 per capita, one of health system design, including full autonomy of facilities, the highest levels of donor dependency in Sub-Saharan decentralization, third-party payment, and strategic pur- Africa. Much of this donor funding is earmarked funding chasing through performance-based financing. It has tran- for HIV/AIDS. sitioned from a faith-based service delivery model in the Despite the increase in funding, resources are not suf- colonial era to a model guided by the Bamako Initiative, ficient to meet the country’s health care needs. Rwanda which sought to expand access to health services through has therefore pioneered profound reforms, including an the development of local models of primary health care that innovative health system and a financing model grounded are managed and financed by communities. Both public and in grassroots initiatives and institutions. Three prominent private not-for-profit health facilities charge fees that are reforms were adopted to boost both the demand for and locally retained and managed to cover the costs of health the supply of health services: health microinsurance services and improve the quality of care. (mutuelles), performance-based financing, and fiscal To improve financial access, the government pioneered a decentralization. Those reforms have transformed the fis- microinsurance scheme and supported its expansion and sub- cal space landscape. Revenues generated by health facili- sidization. It then introduced a mechanism of performance- ties have increased significantly as a result of increased based financing to provide incentives to health facilities to use of health services and health insurance coverage, and deliver high-impact interventions and ensure quality of ser- an increasing share of domestically generated revenues is vices. In 2006 it established a fiscal decentralization policy captured by health centers, strengthening front-line and a legal framework that delineated a clear role for central providers. and local governments and service providers. In 2007 the Rwandan government adopted its second Together these three reforms constitute strategies to Poverty Reduction Strategy Paper (the Economic Develop- strengthen accountability for services to citizens as part of ment and Poverty Reduction Strategy). Its goals for the Rwanda’s 2006 national decentralized service delivery policy health sector are to maximize preventive health measures (Government of Rwanda 2006). Fiscal decentralization has and build the capacity to provide high-quality and accessi- been accompanied by measures to strengthen citizen partic- ble health care services to the entire population in order to ipation and accountability, including mechanisms for estab- reduce malnutrition, infant and child mortality, and fertility lishing accountability links between citizens and local and to control communicable diseases. The strategy also government officials, contractual performance between supports strengthening institutional capacity, increasing the health services providers and local governments or national quantity and quality of human resources, ensuring that policy makers, and contractual approaches between com- health care is accessible to the entire population, increasing munities and health providers. This policy can be visualized geographical accessibility, increasing the availability and by using the accountability framework laid out in the World affordability of drugs, improving the quality of services in Development Report 2004 (World Bank 2004). Accountabil- the control of diseases, and encouraging the demand for ity of health providers to clients (“client’s power”) is strength- such services. It also sets ambitious targets for slowing pop- ened through micro–health insurance funds that claim and ulation growth, calling for innovative measures in the fund health services on behalf of households. Accounta- strengthening of reproductive health services and family bility of providers to the government (the “compact”) is CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 405 strengthened thanks to performance-based financing mech- which have propelled these strategies forward. The nation- anisms. Accountability of government to citizens (“voice”) wide implementation of mutuelles was closely followed by is strengthened through decentralization, citizen report the implementation of performance-based financing. cards, and the possibility of recourse to the ombudsman Mutuelles revealed the need for improvements in staff (figure 22.3). motivation and incentives to deliver high-quality health Rwanda has finally settled into its current decentralized services, which in turn served as the driving force for per- model of care, in which health facilities are fully formance-based financing. As the implementation of autonomous entities responsible for the management of performance-based financing in the pilot districts proved financial resources, health service delivery, and human successful in improving staff motivation and health out- resources for health. Community-based health insurance comes, performance-based financing became a major pillar schemes, which have been established and scaled up nation- within the Ministry of Health Strategic Plan (2005–09). In ally, have evolved in response to low levels of utilization of 2006 the government, with the support of external part- health services. Partly to encourage community members to ners, expanded performance-based financing to the entire buy in to these health mutuelle schemes, the government health sector (table 22.1). developed performance-based financing as a complemen- Rwanda has continuously learned from and adapted tary scheme to boost the performance and motivation of its health service delivery strategies. The ability of the staff to deliver higher-quality services as well as to increase government to adapt strategies—as evident in the scaling the delivery of preventive services. up of successful pilot schemes to the national level, in Mutuelles and performance-based financing are two com- light of the changing macroeconomic and health sector plementary schemes. Both aim to shift health financing environment—was also essential to strengthen health mechanisms from inputs-based mechanisms toward output- services. Independent controls and quality checks under or results-based contractual mechanisms. The interactions contractual arrangements have been essential for the between the two strategies are therefore strong. Whereas monitoring and evaluation of health facility performance. mutuelles emphasize personal curative care services, perfor- Performance-based financing is a prime example of a sys- mance-based financing emphasizes high-impact preventive tem in which independent controls on performance are in services and the quality of services. place to ensure proper monitoring and reporting of The nationwide expansion of both strategies has health centers and district hospitals on the quantity and occurred rapidly since the pilot schemes were launched, quality of services delivered, which in turn drive facilities’ thanks largely to government ownership and commitment, reimbursement. Figure 22.3 Decentralization of and Accountability for Health Services in Rwanda National government Fiscal decentralization Performance-based Citizen report Imihigo transfers cards, ombusdman Local government Compact Voice Client power Clients/citizens Autonomous facilities providers Health insurance (mutuelles) Source: Government of Rwanda 2006. Note: Imihigo are performance contracts in which the region and its districts promise the president of Rwanda that they will implement the measures outlined in the annual plans. 406 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS Table 22.1 Health Financing Reforms in Rwanda, 1999–2008 Reform 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Enrollment in mutuelles (percent) 0 1.6 2.6 4.7 7 27 44 73 75 85 Number of districts with performance- based financing 0 0 0 2 5 7 7 14 23 30 Assisted deliveries (percent) 39 52 Key policy Rwandaise Fiscal decentralization Health insurance milestones d’Assurance law passed; law passed Maladie (RAMA) decentralized created; fiscal service delivery decentralization policy includes policy performance-based financing and mutuelles Key Community-based RAMA Performance- Integration of Fiscal Full autonomy implementation health insurance established based performance- decentralization given to public milestones pilots launched financing based financing health facilities, in three health pilots and subsidies to including over districts since launched mutuelles in hiring and firing July 1999 in four national budget National guarantee districts fund established Per capita health expenditures (National Health Account) $15 $37 Source: Authors. Note: RAMA = Rwandaise Assurance Maladie. 407 Community mobilization and intersectoral collabora- Figure 22.4 Use of Health Services in Rwanda in Event of tion have contributed significantly to the implementation of Illness, by Health Insurance Coverage, 2005 health reforms. Cultural and social factors, particularly sol- idarity within communities, have contributed to the success 60 Use of health services (%) of several health service delivery innovations. The rapid 50 proliferation of health mutuelles was made possible by the 17 40 9 strong solidarity within Rwandan society, in which commu- 30 20 9 13 nity members encouraged one another to join. This solidar- 11 20 ity has been long-standing in Rwanda, evident even before 11 35 25 10 the start of health mutuelles. In the case of health mutuelles, 10 16 cultural barriers initially thwarted implementation, particu- 0 None RAMA Health All larly because Rwandans were used to seeking care from tra- mutuelle ditional healers and because patients of traditional healers Public services Private services were able to pay in kind rather than making cash payments. Traditional practitioners Source: Government of Rwanda 2005. Community-based health insurance Health insurance has been scaled up at the national level on the basis of community-based health insurance schemes, health services, including prepayment mechanisms through with strong support from the government of Rwanda. Cov- health insurance schemes. In view of the significant scope of erage of health mutuelles increased dramatically between poverty following the war of 1994, the reintroduction of 2003 and 2008, rising from less than 7 percent to 85 percent user fees in the health sector in 1996 was accompanied by an of the population in just five years. exemption policy that allowed for free health care coverage Health mutuelles have had a significant positive impact in health facilities for people identified by the local admin- on health service utilization, income protection, and house- istrative authorities as indigent. Coverage of the poor and hold health behaviors (Sekabaraga, Diop, and Soucat 2010). vulnerable groups has been integrated in the development Health insurance coverage increased the use of modern of health mutuelles since they were first piloted in 1999. health services by children under five between 2000 and In 2000 Rwandaise d’Assurance Maladie, the first health 2007. In the general population, use of modern services by insurance scheme for the formal sector, was established for the insured population is nearly twice as high as use by the civil servants. Membership in RAMA soon became compul- uninsured (figure 22.4). In addition, health mutuelles sory in order to increase coverage. To complement insur- seem to protect against expenditure caused by unexpected ance schemes such as RAMA and a few private insurance illness, because uninsured households spend directly twice schemes that target the formal sector, health mutuelles cov- as much as insured households for illness-related expendi- ering rural communities and the informal sector were ture. Health insurance coverage has thus significantly expanded to promote equitable access to quality health ser- reduced household out-of-pocket health expenditures vices. Health mutuelles were designed to pool or spread the (Ministry of Health of Rwanda and World Bank 2011). financial risk of seeking care across their membership base. Among women who gave birth during the period 2000–05, The goal was to respond to the low use of health services 77 percent were affiliated with Rwandaise d’Assurance (caused in part by user fees) by improving financial access Maladie (RAMA). About 42 percent of women affiliated to health services, particularly for underserved populations. with a health mutuelle were assisted by a skilled health pro- The package of services reimbursed by health mutuelles to fessional during delivery compared with 35 percent of women health facilities has expanded over time and currently cov- with no insurance (Sekabaraga, Diop, and Soucat 2010). ers all services delivered within a health center as well as Before 1996 health services in Rwanda were free. Given drugs from the national essential drug list. Health mutuelles problems associated with the quality of services and the also cover most costs for health services and drugs delivered financial sustainability of the health system, cost-recovery at district and referral hospitals when mutuelle members mechanisms were reintroduced in 1996, while policy receive referrals for these higher levels of care. debates continued over alternative financing mechanisms These schemes have been extended to empower citizens for reconciling internal resource mobilization and access to and communities in the health sector and to change their 408 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS interactions with health care providers. Through contrac- coordinating the initial efforts of adaptation and expan- tual relations between health mutuelles and health care sion, a variety of local policies and incentives proliferated, providers, communities and citizens have a tool with initiated by local authorities of all categories (political which to hold health care providers accountable for ser- groups, associations, and so forth) to motivate the popula- vices provided. In essence, health mutuelles in Rwanda tion to join health mutuelles (by, for example, linking reflected a bottom-up strategy, driven largely by local membership to civil status services, microcredit, and so communities. The success of the initial pilot schemes forth). These initiatives helped develop the health mutuelles motivated the central government to scale up this strategy. and contributed to the growth of social demand for their Currently, the central government sets national guidelines expansion nationwide. and policies, including benefit packages and contribution Current levels of contributions to health mutuelles are policies. In 2008 it instituted a national guarantee fund for affordable for all but the poorest 10 percent of Rwandans. providing general subsidies to support the extension of the Affordability of enrollment in health mutuelles is assessed benefit packages of health mutuelles. Since 2009 the gov- based on the percentage of contributions in household total ernment has also built on the national network of health expenditures and household nonfood expenditures. The mutuelles to elaborate and implement demand-based tar- cost of membership rises with family size (for many reasons, geted subsidies, through which the government, donors, including high disease burden in large families and the and nongovernmental organizations (NGOs) are providing externalities of insurance benefits). Male-headed house- health insurance coverage to poor people, vulnerable holds have a higher proportion of enrollment than female- groups, and people living with HIV/AIDS. At the opera- headed households, partly because of the income difference tional level, health mutuelles are run and organized by between the two groups. Family heads that completed pri- community representatives and local health care providers. mary school or received some vocational training tend They also serve as a forum for promoting dialogue between to have the highest rate of enrollment among the least- the community and providers on the quality and range of educated and best-educated households. This trend is also health services offered. In this way, community members reflected in the enrollment rate by income (expenditure) are better able to hold providers accountable for services category, so that middle-income and middle-rich people delivered. tend to have the highest participation rates in health For the majority of the population employed in the rural mutuelles. Out-of-pocket illness-related expenditures and informal sectors, an incremental approach was followed among households enrolled in health mutuelles are twice as in developing mechanisms for pooling health risk. The high as those of members. Households that live very close to process of cumulative building of national capacities is a health centers spend more than those living far away. hallmark of the incremental development of mutuelles in Households that are not covered by health mutuelles spent Rwanda. Capacity building for local actors involved in set- nearly twice as much for illness-related services as people ting up, managing, and monitoring health mutuelles began who were insured. in 1999, with the establishment of mutuelles in three pilot Information has played an important role in the opera- health districts.2 The geographic extension of the mutuelles tional management and monitoring of the development of was accelerated in 2004, after the adoption of a national health mutuelles at the local level and in their strategic man- strategic framework for their development. The number of agement at the central level. Indeed, Rwanda is one of the health mutuelles grew by a factor of 2.5 in a single year, few African countries where an information system to sup- climbing to 226 nationwide. In 2007 each of Rwanda’s 403 port health mutuelle management has been developed to health centers had a partner health mutuelle or “health permit monthly monitoring of performance. It is also one mutuelle section,” and all of the country’s 30 districts had a of the few countries that has adapted training manuals on district health mutuelle, which, on average, was linked to 13 health mutuelle development, management, and monitor- health mutuelle sections. ing in line with the local context, including availability in The period of experimentation, which started in 2001, the local language. Numerous other activities to promote was followed by attempts to adapt the institutional and raise awareness of health mutuelles are carried out arrangements for health mutuelles to the environment of around the country and at the national level. The Ministry administrative and political decentralization and by early of Health occasionally organizes an annual event on mutual efforts to expand the mutuelles to other districts of the organizations at which prizes are awarded to the best per- country. In the absence of an explicit policy framework for forming health mutuelles. CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 409 Performance-based financing and reforms government purchases 13 quantitative indicators and 13 of human resources management qualitative measurements from health facilities (tables 22.2 Based on a positive evaluation of a three-year pilot phase in and 22.3).4 At the hospital level, performance is assessed two provinces, Rwanda has implemented a national pro- through a peer-evaluation mechanism. gram since 2005; it has scaled up performance-based Low intake of preventive services and poor quality of care financing since 2006 (Rusa et al. 2009).3 Performance-based in health centers served as the impetus for introducing per- financing is currently implemented at three levels: health formance-based financing strategies. Quality of care became centers, hospitals, and community levels. a particularly salient issue as expansion of health mutuelles The performance-based financing model is based on the increased utilization rates dramatically at all health centers, principle of separating purchaser and provider functions in adding to the workload of health personnel who, at the time, health service delivery (figure 22.5). By distinguishing had little or no incentive to take on the additional work. The between and maintaining a split between bodies purchasing impetus for the performance-based financing strategy in services and bodies providing services, this model promotes Rwanda first came from external actors; additional resources accountability and avoids conflicts of interest. The model and incentives were provided to health workers to improve consists of a family of methods and approaches that aim, efficiency and outcomes, under pilot schemes implemented through differing levels of intervention, to link incentives to in 2001–05. As a result of the pilots’ success, performance- performance. In the national model for health centers, based financing became a major pillar within the Ministry of payments for performance are based on the quantity of Health strategic plan and was implemented nationally. outputs achieved conditional on the quality of services Rwanda’s institutional performance-based financing delivered. Through performance-based financing, the central model can be classified as “output-based financing,” because Figure 22.5 Rwanda’s Performance-Based Health Care Financing Model Purchaser-provider split in the national model The Controllers District Hospital (public, FBOs) District Health Departments (USF) Technical Assistance Agencies Special Interest Groups Steering Committee Grass-root Organizations (piloted) Rwanda PBF The Purchasers The Providers GOR District Hospital (public, FBOs) WB MAP Health Center (public, FBOs) USG Second Tier: (GF) by Health Centers 1. CHWs 2. TBAs 3. Private for-profit Providers 4. Private not-for-profit Providers Source: Rusa et al. 2009. Note: Purchasers are those who pay for services. They include NGOs, which purchase services with their own funds and act as fundholders or pass-through mechanisms for other donors; the government of Rwanda (in the case of the national model); the U.S. government, through collaborative agencies such as Management Science for Health and Family Health International; World Bank MAP funding, NGOs in the Cyangugu and Butare pilots and purchasing performance through the new national model; and the Global Fund to Fight AIDS, Tuberculosis and Malaria, which is expected to start purchasing through the new national model. Providers can include public and faith-based managed health facilities (health centers and hospitals) and private for-profit health facilities. Controllers are those who control the level of performance, such as district health teams in the Ville de Kigali pilot, who certify a mix of quantity and quality deliverables in health centers and hospitals. District health teams were used in Butare for random quantity control and in Cyangugu l for a qual- ity measure in health centers. A peer-evaluation mechanism for district hospitals was piloted in Cyangugu and partially in Butare. 410 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS Table 22.2 Output Indicators and Unit Payments under Performance- Based Financing Formula Amount paid Output indicators per unit (US$) Visit indicators (number of) Curative care visits 0.18 First prenatal care visits 0.09 Women who completed 4 prenatal care visits 0.37 First time family planning visits (new contraceptive users) 1.83 Contraceptive resupply visits 0.18 Deliveries in the facility 4.59 Child (0–59 months) preventive care visits 0.18 Content of care (number of) Women who received tetanus vaccine during prenatal care 0.46 Women who received malaria prophylaxis during prenatal care 0.46 At-risk pregnancies referred to hospital for delivery 1.83 Emergency transfers to hospital for obstetric care 4.59 Children who completed vaccinations (child preventive care) 0.92 Malnourished children referred for treatment 1.83 Other emergency referrals 1.83 Source: Basinga et al. 2009. Table 22.3 Quality Indicators Services and Weights Used in the Performance-Based Financing Formula Share of weight Share of weight allocated to allocated structural to process Means of Service Weight components components assessment General administration 0.052 1.00 0.00 Direct observation Cleanliness 0.028 1.00 0.00 Direct observation Curative care 0.170 0.23 0.77 Medical record review Delivery 0.130 0.40 0.60 Medical record review Prenatal care 0.126 0.12 0.88 Direct observation Family planning 0.114 0.22 0.78 Medical record review Immunization 0.070 0.40 0.60 Direct observation Growth monitoring 0.062 0.15 0.85 Direct observation HIV services 0.090 1.00 0.00 Direct observation Tuberculosis service 0.028 0.28 0.72 Direct observation Laboratory 0.080 1.00 0.00 Direct observation Pharmacy management 0.060 1.00 0.00 Direct observation Financial management 0.050 1.00 0.00 Direct observation Total 1.000 Source: Basinga et al. 2009. it pays on a fee-for-service or case reimbursement basis to providers to provide more services while also increasing improve outputs. Although performance-based financing financial revenues at the health facility level. incentives are generally meant to induce providers (the sup- As of 2006 the government transferred about $1.80 per ply side), supply-side incentives in Rwanda work through capita from the Treasury directly to health facilities at the supplier-induced demand, whereby suppliers actively seek basic health service level on the basis of a performance- to convince people to use more of certain kinds of services. based formula. The program channels funds directly from Such incentives are necessary in Rwanda, where the goal is Treasury to the bank accounts of the more than 400 health not to limit excessive supply and unnecessary demand (as is clinics in Rwanda (40 percent of them faith based, 60 percent the case in richer health systems) but rather to induce of them public) on the basis of performance agreements. CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 411 These funds are flexible and may be used for facility expen- It is the policy of the Ministry of Health, in collaboration ditures, including performance-linked salary bonuses, with development partners, to harmonize the framework partially substituting for revenues from user fees. Rwanda for compensation packages of health professionals. The implemented a three-tier performance-based financing objective is to avoid introducing distortions in the distribu- model—including hospitals, health centers, and ultimately tion of health workers, which occurs when health workers the community level—in order to make health services move from the public sector to donor projects where the pay more community oriented. is higher. In light of this concern, donors, such as the Global One of the key objectives of performance-based financing Fund, have begun to use national pay scales and to fully was to introduce bonuses to health workers as incentives for integrate staff within the health system at large. good performance, based on a range of agreed indicators. Results-based block grants in Rwanda have contributed This system was designed to allow for better monitoring of significantly to the increase in assisted birth deliveries as health personnel activities and hence to enable district and well as the intake of child health services; the grants have central levels to track staff performance over time. Although also increased the quality of services. Clinics that received the central government determines the overall performance- performance-based financing (of about $1.80 per capita based financing budget envelope that the health facility per year) performed more assisted deliveries and more receives, based on a formula involving the quantity and qual- post-natal visits than clinics receiving the same funding ity of services provided, it is the committee within the health without a performance contract (figure 22.7). The quality facility itself that determines how these funds should be used. of care of antenatal services was 15 percent higher in per- In 2008 Rwanda decentralized wages. As a result, formance-based financing clinics than in control clinics financing and payments for health personnel are increas- (figure 22.8). ingly linked to performance in which block grants from The results achieved—in service supply and the enthusi- the government and donors can be used as salary. Direct astic participation of all stakeholders—after a few years of spending on wages and salaries by the central administra- experience point to a promising future. However, because it tion and transfers to public institutions for salaries of is a dynamic strategy, performance-based financing adjusts health workers have declined. In contrast, funds chan- to innovative ideas that benefit the population and health neled to human resources for health through provinces care providers. The remaining challenges are related to the and districts that come from both the government permanent oversight requirements, the accuracy of data, (including performance-based financing) and user fees and the delicate balance of the pricing of the various indi- collected directly by facilities have risen dramatically in cators. The future of performance-based financing will recent years (figure 22.6). depend on finding appropriate solutions to these issues. Figure 22.7 Number of Assisted Deliveries in Figure 22.6 Financing for Human Resources for Health Rwanda under Performance-Based and in Rwanda, 2005–08 Nonperformance-Based Financing, 2006 and 2008 25 22.4 60 55.6 Number of assisted deliveries 20 18.5 7.3 % increase US$ (millions) 15 49.7 due to PBF 50 11.2 10 8.2 8 6.7 5 3.0 40 36.3 0.5 0 34.9 05 06 07 08 20 20 20 20 30 Baseline (2006) Follow up (2008) Basic salaries Performance-based financing Linear (basic salaries) Linear (performance-based financing) Control facilities Performance-based facilities Source: Authors. Source: Basinga et al. 2009. 412 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS Figure 22.8 Quality of Care in Rwanda under government under the decentralization framework and Performance-Based and Nonperformance- further strengthen district authority while allowing for Based Financing, 2006 and 2008 greater community participation and facilitating resource 0.20 allocation to local government. Central government respon- 0.15 sibilities in this phase remained regulation and development 0.15 of policy frameworks, capacity building of local govern- 0.10 Quality of care 15 % Standard deviation ment, and monitoring and evaluation. 0.05 increase due to PBF Finally, phase three, which began in 2008, granted 0.00 0 autonomy to health facilities and transferred fiscal respon- –0.05 sibilities and financial resources from the central and local –0.10 –0.10 government to facilities. This reform has resulted in relative –0.13 autonomy in budgeting and financial management within –0.15 Baseline (2006) Follow-up (2008) facilities, because health care providers are now contracted Control facilities Performance-based facilities with and managed by health facilities. Fiscal decentralization in Rwanda was government Source: Basinga et al. 2009. owned and driven, with strong support and collaboration of development partners. The objective was to bring ser- vices closer to the people, and to improve the financial via- bility of districts. The infrastructural changes needed may Fiscal decentralization not have been in place (until June 2007 districts lacked A strong commitment to bring services closer to the people accounting software to manage financial transactions, and resulted in rapid fiscal decentralization, increased citizen local capacity in managing financial and human resources participation, and increased autonomy of health facilities. still remains limited), but overall success was striking. The Fiscal decentralization (adopted as policy in 2001 and strategy was organized by the central government, which enacted into law in 2006) served as an essential component from the outset determined the degree of authority delegated of Rwanda’s decentralization agenda to devolve authority to to local levels and delineated relevant policies and standards. the district level. As was to be expected, given the presence In these efforts, the central government received significant of such a strong state, a mindset change was needed to move technical assistance and guidance from development part- forward on many of these reforms. As in the case of decen- ners, which organized their support in the form of a sector- tralization, some officials at the central level felt disempow- wide policy to ensure government ownership over decision ered and were initially unwilling to relinquish their control making and policy setting; increase coherence between pol- at the outset. Officials at the local level had to adapt to icy, spending, and actual results; reinforce the government’s accept their new responsibilities, and donors had to adjust management systems; and harmonize donor support. to working with local governments. Decentralized units at local levels were given the author- The decentralization of authority across sectors was ity to manage the flow of funds (once received) as well as the planned through an incremental, three-phased approach. delivery of health services. Decentralization transformed The first phase (2001–05) focused on administrative and health facilities into autonomous entities, with the ability to political decentralization; it aimed to institutionalize manage financial and human resources as they deem most decentralized governance by establishing democratic and appropriate, according to local needs. The process gave community development structures, delineating policies, them complete control over the hiring and firing of health establishing legal frameworks, and strengthening institu- personnel. tional capacity at local levels. Phase two, which began in The accountability links between local governments 2006 and ran through 2008, focused on making local gov- and national policymakers are strengthened through ernments responsible for bringing health services closer to inspections, audits, and Imihigo—performance contracts beneficiaries. The devolution of responsibilities in health in which the region and its districts promise the president services and the transfer of resources under fiscal decentral- of Rwanda that they will implement the measures outlined ization are the backbone of relationships between the in the annual plans.5 Decentralization reforms have national government and districts in the health sector. It resulted in increased responsibilities of local governments aimed to reorganize roles and responsibilities within local in many areas and are increasing space for community CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 413 participation and community-driven development initia- the building of grassroots, community-based microinsur- tives. Satisfaction with service delivery is measured ance. Key reforms included provisions for financial protec- through citizen report cards. tion and other support for poor people. Lessons learned from the Rwandan experience provide a strong base for future action. As financial barriers to access CONCLUSION to health services are being reduced significantly through Fiscal decentralization, performance-based financing, and health care financing reforms, improving the quality and the expansion of health insurance have led to a dramatic sustainability of health services will remain among the main increase in resources for frontline providers. Between 2002 challenges facing Rwanda in the coming years. and 2007, public resources flowing to health facilities more The success of Rwanda in improving health outcomes, than tripled (figure 22.9). The increase in resources took particularly for women, children, and the poor, can be place at all levels, showing no higher priority given to the linked to both increased resources and implementation of primary care level. An increasing number of donors are reforms. Rwanda used the inflow of resources to strengthen channeling their assistance through on-budget support, and its country system, including its public finance and health major efforts are under way toward coordinating and har- systems. It designed its own brand of reforms, staying away monizing aid. Most of the increase in publicly managed from donor fads and looking realistically and opportunisti- resources flowed to human resources and to performance- cally at the balance between sustainability and equity. based financing; resources directly managed by donors Reforms focused on results, and results attracted more funded HIV/AIDS activities. On the domestic front, inter- funding, both domestic and external. Increased resources nally generated revenues of health facilities increased signif- and reforms mutually reinforced each other as part of a vir- icantly as a result of increased utilization of health services tuous cycle. Efficient and equitable use of resources required and health insurance coverage; an increasing share of inter- reform, and the success of reforms needed resources. nally generated revenues is captured by health centers, The success of the health financing innovations is criti- strengthening frontline providers. cally linked to the institutional context. Decentralization Rwanda chose to develop a mixed health care financing reforms coupled with performance-based financing ensure model, combining decentralization and performance-based that health facility managers have not only the incentives financing with a strategy to pool private spending through but also the power to ensure that these innovations translate Figure 22.9 Financial Transfers to Districts for Priority Programs, 2006–10 9 8.4 8 7 6 5.3 Local currency 5 4.8 4 4 3.7 3 2.5 2.3 2.2 2.3 2 2 1.7 1.3 0.9 1 0.4 0.4 0 Community-based health Performance-based Community health insurance financing 2006 2007 2008 2009 2010 Source: World Bank 2010. 414 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS into changes in the delivery of services. The performance- country, means knowing that use of condoms and having a based financing system is now being extended to provide single, uninfected, faithful partner can reduce the chances incentives to community health workers providing outreach of contracting HIV, knowing that a healthy-looking person services and demand-side incentives to women to continue can have HIV/AIDS, and rejecting the two most common the increased utilization of key maternal health services. local misconceptions about HIV/AIDS transmission and Strong government leadership, vision, and the step-by- prevention. step building of a policy and regulatory framework at all 2. At the time, the districts were called Byumba, Kabgayi, levels have fostered the short- and long-term sustainability and Kabutare. They are now called Gicumbi, Muhanga, and Save. of health sector reforms. The Rwandan government showed flexibility and was able to adapt strategies in light 3. This program has been supported by a broad consor- tium of donors, including the World Bank, the U.K. of the changing macroeconomic and international health Department for International Development, the European environment. Government coordination of donor funding Union, Sweden, the African Development Bank, the was critical to ensure that aid was used effectively and Netherlands, and Germany. aligned with national priorities. Systems for improved 4. A separate contract channels earmarked funds of global accountability, including contractual arrangements, inde- HIV/AID programs for another 16 indicators. pendent controls, and quality checks, were essential for 5. Imihigo are also monitoring instruments designed to monitoring and evaluating health facility performance. help local authorities plan and act realistically. Cultural and social factors, particularly solidarity within communities, also contributed to the success of several health service delivery innovations. REFERENCES A particularly important feature of the Rwanda experi- Basinga, P., P. Gertler, A. Binagwaho, A. Soucat, J. Sturdy, and ence is the integration of strong mechanisms to evaluate the C. Vermeersch. 2010. Paying Primary Health Care Centers impact of its policies. Rwanda’s experience, as well as that of for Performance in Rwanda. World Bank, Washington, DC. other low-income countries, in introducing pro-poor Government of Rwanda. 2005. “Enquete Integrale sur les financing policies needs to be systematically evaluated using Conditions des Menages.” Kigali. rigorous metrics and standardized benchmarks. It is possi- ———. 2006. Decentralized Service Delivery. Kigali. ble for countries to nest impact evaluation designs when May, J. F., and A. Kamurase. 2009. Demographic Growth and introducing new policies at scale. The donor community Development Prospects in Rwanda: Implications for the needs to support these evaluations if it wants to improve aid World Bank. World Bank, Washington, DC. effectiveness. Measure DHS and ICF Macro. 1995. Demographic and A key issue for the future is sustainability and the neces- Health Survey, Rwanda. Calverton, MD. http://www sary evolution of the institutional support for the health .measuredhs.com/. financing approach of Rwanda. Like many Sub-Saharan ———. 2000. Demographic and Health Survey, Rwanda. African countries, Rwanda is highly dependent on aid and Calverton, MD. http://www.measuredhs.com/. will remain so for the next decade. Sustained support from ———. 2005. Demographic and Health Survey, Rwanda. the donor community is therefore needed to support the Calverton, MD. http://www.measuredhs.com/. health system strengthening agenda. Mutuelles provide one ———. 2007. Demographic and Health Survey, Rwanda. way to ensure more sustained domestic funding. They rep- Calverton, MD. http://www.measuredhs.com/. resent an efficient way to pool private out-of-pocket spend- ———. 2008. Demographic and Health Survey, Rwanda. ing, but there is a need for the government—with the help Calverton, MD. http://www.measuredhs.com/. of donors—to subsidize the enrollment of the poorest Ministry of Health of Rwanda, and World Bank, 2011. Rwandans and to regulate the package of benefits as well as Country Status Report on Health, Health Systems and the provider payments mechanisms to ensure equitable Poverty. Human Development Series. World Bank, access to quality services. Washington, DC. Rusa, Louis, Miriam Schneidman, Gyuri Fritsche, and Laurent Musango. 2009. “Rwanda: Performance-Based NOTES Financing in the Public Sector” In Performance Incen- 1. Comprehensive knowledge of HIV/AIDS, which can be tives for Global Health: Potential and Pitfalls, eds. Rena used as a tracer indicator of general health knowledge in the Eichler, Ruth Levine, and the Performance-Based CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS 415 Incentive Working Group. Washington, DC: Center for Health MDGs? Evidence from Rwanda.” Communica- Global Development. tion to the International Symposium on Health Systems Sayinzoga, K., A. Soucat, A. Kjellgren, and P. Musafiri. 2010. Research, Montreux, Switzerland. From Reconstruction to Development: Budgeting for Per- World Bank, 2004. World Development Report 2004: Making formance in Rwanda. Human Development Series, World Services Work for Poor People. Washington, DC: World Bank, Africa Human Development, Washington, DC. Bank. Sekabaraga, C., F. Diop, and A. Soucat. 2010 “Can Innova- ———. 2010. Country Status Report on Health, Health tive Health Financing Policies Turn the Tide on the Systems and Poverty. Washington, DC: World Bank. 416 CHAPTER 22: INNOVATIVE FINANCING FOR HEALTH IN RWANDA: A REPORT OF SUCCESSFUL REFORMS CHAPTER 23 The Malaria Control Success Story Anne-Maryse Pierre-Louis, Jumana Qamruddin, Isabel Espinosa, and Shilpa Challa alaria is both preventable and treatable. Yet M Treatment of severe episodes, which can cost up to one- more than 220 million cases of malaria are esti- quarter of a household’s monthly income, accounts for up mated to occur each year, and approximately to 40 percent of public sector health expenditures in the 785,000 people die from the disease annually. Half of the worst-affected countries (WHO 2007). Other direct costs world’s population—some 3.3 billion people living in 109 include the costs of insecticide-treated nets and indoor countries—are at risk of malaria (Roll Back Malaria 2008). residual house spraying to prevent malaria, antimalarial Worldwide, malaria is the fifth-leading cause of death drugs, transportation to clinics, and hospital fees. Operating from infectious diseases (after respiratory infections, in a vicious cycle, malaria is both a cause and a consequence HIV/AIDS, diarrheal diseases, and tuberculosis). The dis- of poverty. ease is life threatening and needs early, accurate diagnosis Malaria keeps countries as well as households in poverty: and treatment, which can be difficult in remote areas that annual economic growth in countries with high malaria lack clinics, trained health care providers, technical assis- transmission has historically been lower than in countries tance, or medicine. without malaria. Economists have estimated that malaria is Malaria is caused by Plasmodium parasites, which are responsible for an “economic growth penalty” of up to 1.3 spread to humans through infected Anopheles mosquitoes, percent a year in malaria-endemic African countries (Gallup called malaria vectors, which bite mainly between dusk and and Sachs 2001). The disease discourages local and for- dawn.1 Ninety percent of the world’s malaria deaths occur eign investment and tourism, affects land use patterns and in Sub-Saharan Africa, where the most severe form of the crop selection (resulting in suboptimal agricultural produc- disease prevails, making malaria the second-leading cause of tion), and reduces labor productivity through lost work days death in the region after HIV/AIDS (Roll Back Malaria and diminished on-the-job performance. It affects learning 2008). The disease is one of the leading causes of death of and academic achievement through frequent absenteeism of children under the age of five in Sub-Saharan Africa and a teachers and students. In children who suffer severe or fre- major cause of complications, including maternal death and quent infections, it can cause cognitive impairment and in low birth weight, in pregnancy. some cases permanent neurological damage. Other indirect Malaria is not only a health problem in Africa; it is also a costs include loss of income and work, including the unpaid development problem. Death and disability (both short- work carried out largely by women who take care of the sick and long-term) from malaria have enormous social and and support them at home or in the hospital. economic costs, costing African countries an estimated $12 Expenditures on malaria control drain already fragile billion a year in lost productivity (Gallup and Sachs 2001). economies through the deterrent effect on investment 417 (private businesses are reluctant to expand in areas where high transmission, 100 percent of pregnant women the disease is affecting the workforce and creating gaps in receive intermittent preventive treatment. the production line). Economic and social decision mak- ■ The global malaria burden is reduced by 50 percent from ing at all levels is affected. Travelers are hesitant to visit 2000 levels, to less than 175 million–250 million cases countries with a high incidence of malaria, reducing and 500,000 deaths annually. tourism revenues and leading to losses of job opportuni- ties and income. This section looks at progress on key malaria indictors in A global effort to help countries control the disease Africa based on an analysis of nationally representative sur- that was revitalized in 2005 is showing signs of success. In veys in countries in which the data were available for a par- countries in which control measures have been intensi- ticular indicator. A subset of countries for which data were fied, there have been clear and positive results. Eleven available is used to illustrate progress on key indicators. countries and one territory in Africa show reductions of more than 50 percent in either confirmed malaria cases or malaria admissions and deaths in recent years (WHO Insecticide-treated nets 2010b). This progress is a direct result of the scaling up The promotion and distribution of insecticide-treated nets and acceleration of measures against the disease. Other has long been recognized internationally as a key cost-effec- positive outcomes include parallel declines in child mor- tive intervention to control malaria (Statesmen’s Forum tality in some countries. 2010 (box 23.1). This intervention has been the primary This chapter examines progress in Sub-Saharan African focus for scale-up over the past few years. Beginning in countries on key malaria indicators as well as in funding for 2008, some countries in Africa made a policy shift from pro- the fight against malaria between 2005 and 2010. The first viding insecticide-treated-net coverage only for populations section examines the four main interventions for preventing at greatest risk (including children under five and pregnant and treating malaria (insecticide-treated nets, antimalarials, women) to seeking coverage for the entire population at artemisinin-based combination therapy, indoor residual risk. This shift has largely been the result of a global effort to spraying). The second section looks at prevention among increase progress toward the Millennium Development and treatment of pregnant women and children under five. Goals (MDGs) and a call for action from the UN Secretary The third section discusses partnership and coordination in General in 2008. Given that the policy changed fairly the fight against malaria. The fourth section describes suc- recently, the shift to implementation has not yet been seen cess stories in four African countries (Eritrea, Ethiopia, on a large scale across the continent. Rwanda, and Zambia). The last section identifies lessons Figure 23.1 illustrates progress toward coverage with learned and discusses the way forward. insecticide-treated nets across a number of countries in Africa. At first glance, the story appears discouraging, as many countries are well below the 80 percent coverage INTERVENTIONS FOR PREVENTING AND target for this key intervention. Many of the data used, TREATING MALARIA however, come from representative surveys in countries Between 2000 and 2008 malaria control efforts in Africa conducted in 2006 and 2007—around the same time that were measured against targets set in the 2000 Abuja Decla- malaria control efforts on the continent started ramping up. ration.2 Since 2008 progress has been measured against the Among countries in which coverage is 50 percent or greater universal coverage targets set forth in the Roll Back Malaria (Ethiopia, Madagascar, Rwanda, Senegal, Zambia, and Global Malaria Action Plan 2010. These targets include the Zanzibar), most surveys were conducted between 2007 and following: 2009. The most recent 2010 Malaria Indicator Survey in Zambia (not illustrated in the figure) shows an increase in ■ Eighty percent of people at risk from malaria are using the share of households with at least one insecticide-treated locally appropriate vector control methods, such as net rising to 64 percent. long-lasting insecticidal nets, indoor residual spraying, Since 2006 coordinated efforts have supported the and, in some settings, other environmental and biologi- “catch-up, keep up” approach to bed net distribution in cal measures. countries with low coverage. During the “catch-up” phase, ■ Eighty percent of malaria patients are diagnosed and bed nets are distributed through either stand-alone or treated with effective antimalarial treatments; in areas of integrated bed net campaigns to increase coverage 418 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY Box 23.1 Cost-Effective Tools in the Fight against Malaria A variety of antimalarial tools are starting to pro- Indoor residual spraying duce results in Africa. They include artemisinin- Indoor residual spraying is the application of long- based combination therapies (ACTs), long-lasting insecticidal-treated nets, indoor residual spraying, acting chemical insecticides on the walls and roofs of intermittent preventive treatment, and rapid diag- all houses and domestic animal shelters in a given area nostic tests. to kill the adult vector mosquitoes that land and rest on these surfaces. Indoor residual spraying shortens the life span of vector mosquitoes so that they can no longer Artemisinin-based combination therapies transmit malaria parasites from one person to another, and reduces the density of the vector mosquitoes. Artemisinin-based combination therapies are pro- duced by combining compounds from the Artemisia annua plant with various antimalarial partner drugs. Intermittent preventive treatment They are recognized by the World Health Organiza- tion (WHO) and the scientific community as the Intermittent preventive treatment given to pregnant most effective therapy for treating Plasmodium falci- women at routine antenatal care visits has been shown parum malaria, the predominant form in Africa and to promote healthier pregnancies and yield benefits for the most deadly. ACTs are thought to delay the devel- both mothers and developing fetuses. It can signifi- opment of drug-resistant strains of the disease. They cantly reduce the proportion of low birthweight infants are also active against gametocytes, the sexual stage of and maternal anemia. Intermittent preventive treat- the parasite cycle, effectively reducing disease trans- ment in infants may also play a major role in malaria mission. For these reasons ACTs have become the first line of therapy for many countries struggling to prevention at the public health level; however, addi- control malaria infection and transmission. The tional research is needed to confirm this hypothesis. treatment is fast acting and produces minimal side effects, making it possible for patients to return Rapid diagnostic tests quickly to their daily routines. The challenge is to ensure that prompt and effective treatment with ACT Rapid diagnostic tests for malaria, also known as dip- is available as a critical intervention to complement sticks or malaria rapid diagnostic devices, have the prevention activities. potential to greatly improve the quality of management of malaria infections when high-quality microscopy is not readily available. The tests detect specific antigens Long-lasting insecticidal nets (proteins) produced by malaria parasites that are pres- WHO policy recommends long-lasting insecticidal ent in the blood of infected individuals. Some rapid mosquito nets (a type of insecticide-treated net), diagnostic tests detect only one species (P. falciparum), which offer protection against malaria. If used by at others also detect other species of the parasite. Blood for least 80 percent, in an affected area, these nets can the test is commonly obtained from a finger prick. help break the malaria transmission cycle, thus WHO (2010) now recommends that all cases of sus- reducing the risk for all who live nearby. Woven from pected malaria be confirmed with a diagnostic test insecticide-bound fibers, the nets kill the potential before treatment. As the incidence of malaria continues disease-carrying mosquitoes on contact and offer to decrease in Africa, the need to differentiate between protection from bites. Unlike insecticide-treated malaria from nonmalarial fevers becomes more pressing, nets that require redipping every few months, long- especially given the need to avoid drug resistance and lasting insecticidal nets remain effective for at least increase efficiency of resources. Implementation of the three years. The nets provide protection at night, WHO recommendations will require a more integrated when individuals are most vulnerable. They are safe approach to childhood illnesses (to address the question for children and simple to hang up. of what to do if the fever is not caused by malaria). CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 419 Figure 23.1 Coverage of Insecticide-Treated Nets in Africa, by Country 100 90 Percent of households with at least 80 one insecticide-treated net 70 60 50 40 30 20 10 0 IS 7 a Bu N on M ICS 06 Rw opi S 2 6 a F ia S 2 6 6 go l G 06 ad n M 006 6 f C ria S 7 m go S 6 er a M S 2 6 Sie rin wi M IS 2 a H 06 Re nd IS 06 au S 06 e IS 6 AI IS 08 nz iop IS 08 Za gal S 20 8 ni CS 07 m MIS 08 08 a- ger ICS 08 ric U tan HS 07 M bw H 008 S 06 bi eny CS 06 aL e S 7 Se ar D S 2 7 ali S 9 bl H 07 d ala M uine hi DH 200 IC 200 ua H 00 as DH 00 . o ige IC 00 Zi on DH 200 am an IC 0 Th M 200 ne H 00 rr cip IC 00 c H 0 M IC 200 0 M 0 An oria S 20 C Gh c M S 20 To e D 20 an ga ia M 20 ss H 0 ar M 20 Za Eth ia M 20 Be MI S 20 20 0 ri D 0 eo MIC 20 am K MI 20 as D 20 pu a D 20 S/ 20 Eq n D 2 ep N re M S 2 Bi D 2 ui N o M S 2 au e S 2 ba D 2 ag da IS H Ivo D M an M la M a a a, a ib ia d’ d o b t e lan i n in ib g i o rk am ne i ôt zi Et C Swa P G G .R é Af m em To al tr D o en Sã C Source: WHO 2010b. Note: DHS is Demographic Health Survey. MICS is Multiple Indicator Cluster Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. Data for Equatorial Guinea are from a national survey. a. Updated since original DHS. quickly. During the “keep-up” phase, bed nets are distrib- USE OF INSECTICIDE-TREATED NETS AMONG uted through mechanisms such as routine antenatal care VULNERABLE GROUPS visits, child wellness weeks, routine immunization com- Two of the many vulnerable population groups in the malaria pletion, and school health programs. This approach has epidemic are pregnant women and children under age five. To been proven to work in countries such as Ethiopia, Zam- help safeguard these populations, a number of countries in bia, and Zanzibar, where high coverage of bed nets is cor- Africa made the strategic choice to focus key prevention and related with sharp declines in malaria prevalence. Other promotion activities on these two groups. Despite the recent countries that are lagging behind, such as the Democratic policy shift to universal coverage, many countries are still tar- Republic of Congo (where just 4 percent of the popula- geting these groups, given constrained budget envelopes for tion has bed nets) and Nigeria (3 percent), have been health in general and for malaria in particular. Examining use focusing efforts on increasing coverage of bed nets. The of LLINs among pregnant women and children under five Democratic Republic of Congo plans to launching a thus still provides a good indicator of progress on use of bed large-scale bed net campaign in 2011 to ramp up coverage nets (figure 23.2). Among these groups, the rates of usage are rates. In Nigeria the government has been working with more encouraging, with Equatorial Guinea, Ethiopia, The partners to implement a “catch-up, keep-up” approach to Gambia, São Tomé and Principe, Zambia, and Zanzibar increasing bed net coverage. The catch-up phase has been reaching at least 40 percent usage. under way since 2008, through a phased-in nationwide Analysis of usage rates among pregnant women indicates mass campaign whose ultimate goal is to distribute that, despite some progress, there is much room for 70 million nets (providing every household with at least improvement. Figure 23.2 highlights two key issues in this two long-lasting insecticidal nets) by the end of 2011. In context: the lack of data for this indicator and the relatively Nigerian states in which the bed net campaigns have been low level of use among pregnant women compared with completed, data from a Lot Quality Assurance Sampling children under five. Ethiopia and Zambia4 are the highest (LQAS) survey conducted in 2010 show coverage of bed performers with 43 percent of pregnant women sleeping nets of more than 70 percent.3 under an insecticide-treated net. 420 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY Figure 23.2 Use of Insecticide-Treated Nets in Africa by Children under Five and Pregnant Women, by Country 100 90 Percent of households with at least 80 one insecticide-treated net 70 60 50 40 30 20 10 0 Zi ong HS 06 M bwe HS 8 ric U tani S 2 7 Re nd IS 6 bl H 07 6 rk oo IC 06 as ICS 06 6 to S 6 go Gu 6 Pr wi IS a rr cip ICS 07 eo MIC 006 go HS 06 8 bi au HS 06 e S2 6 M ICS 06 Rw iopi HS 6 d IS 06 ne HS 07 8 nz pia IS 8 AI S 2 8 M 7a 08 a M ne 0 H 0 an ga a M 00 ha IC 00 Be MIC 200 ua H 00 An rial 200 ne Nig MIC 200 Th IC 00 0 Za al M 200 Za thio ia M 00 0 D 0 ba D 20 D 20 pu a D 20 Bu er a M S 20 a F M 20 M 20 To D 20 0 0 20 a M 20 Se a D 20 ib M 20 20 S/ 00 i f C ia S 2 G ic M S 2 Eq n D S 2 2 G a-Bi er D S 2 M 2 b IS 2 in n S ne S IS . o iger IC h D ar I M M an Mal la m o, ali g ire o aL e ni am n m an am ss Ivo ri Sie in Et au a, N d’ E e d C ôt ui ep C G .R é Af m em To al tr D o en Sã C % of HH with >1 ITN % of children <5 years who slept under an ITN % of pregnant women who slept under an ITN Source: WHO 2010b. Note: DHS is Demographic Health Survey. MICS is Multiple Indicator Cluster Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. Data for Equatorial Guinea are from a national survey. a. Updated since original DHS. Antimalarials use combination therapies, preferably those containing artemisinin derivatives for p. falciparum malaria. In 2006 it Fever is internationally recognized as one of the main indi- developed standard guidance on artemisinin-based combi- cators of malaria. Prompt access to treatment of malaria, nation therapy for countries to adopt into policy. As of especially among children under five, is one of the key indi- 2010, 43 countries in Africa had adopted ACTs, several as cators of effective malaria treatment.5 The percentage of first-line treatment and a few as second-line treatment. The children under five receiving any antimalarial treatment rollout of this policy is being implemented in stages across (including artemisinin-based combination therapy, or Africa (table 23.1). ACT) through public and private channels varies by coun- ACTs retail for $6–$10 a course in the private sector, try (figure 23.3), but across the board there is vast room for making the drug prohibitively expensive; even the $1 paid improvement. In several countries, access to antimalarial by the public and nonprofit is 10 times the price of treatment is much higher in the public sector than in the cholorquine. In addition, a number of factors, including private sector, although many seek treatment first in the pri- procurement and supply chain issues, theft, and counter- vate and informal sectors. feiting have made the availability of quality-assured ACTs through public and private sectors challenging Artemisinin-based combination therapy (figure 23.4). But innovative solutions to access to ACTs In response to increasing levels of antimalarial resistance, in are being implemented at the global and country level. 2001 WHO began recommending that countries experienc- The Affordable Medicines Facility for Malaria mechanism ing resistance to conventional monotherapies, such as aims to reduce the unit cost of ACTs (box 23.2). In Zambia, chloroquine, amodiaquine, or sulfadoxine-pyrimethamine, a supply chain intervention has dramatically increased the CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 421 Figure 23.3 Percentage of Children under Five with Fever Receiving Any Antimalarial Medication in the Public and Private Sector, by Country 50 45 40 35 30 25 20 15 10 5 0 ila H 06 S 8 oz er H 07 rr que S 5 Zi Leo H 008 bw H 07 S 8 H 05 ha IS 7 ib H 07 w S 4 . o eg HS 7 on H 04 ny HS 5 7 Be DH 008 rk han H 06 U so D S 2 6 nd HS 08 ne HS 03 S 6 oo H 05 Ta bia S 2 7 an HS 4 S 7 04 hi D 00 Sie mbi DH 200 on H 0 C a M 200 ala H 0 ep Sen i D 200 Ke o D 200 N DH 200 a H 0 Li a D 200 H 0 nz D 00 D 200 az r D 20 M Nig a D 20 ba D 0 C e D S 20 Rw o D 20 am D 20 M ia D S 20 f C al D 20 Bu G in D 20 a F D 20 0 G a D 20 er a D 20 Za n D S 20 20 Et nd S 2 D 2 m ne S 2 r S2 Sw sca HS ia S d S S S in a S op H C be H H ag i D N d a ia ad Mal ige i am ri an g g n m ui ga a a a M .R em D Public sector Private sector Source: WHO 2010b. Note: DHS is Demographic Health Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. availability of ACTs and other essential drugs in health Table 23.1 Adoption of Policies for Malaria Treatment in Africa facilities (box 23.3). Policy Number Indoor residual spraying Number of endemic countries and territories 43 Number of P. falciparum endemic countries Indoor residual spraying is recommended for control of and territories 42 malaria in 71 countries, 32 of them in Africa. It is the pri- Artemisinin-based combination therapy (ACT) used for treatment of P. falciparum 42 mary vector control intervention in Botswana, Mozam- ACT provided free of charge in public bique, Namibia, South Africa, Swaziland, and Zimbabwe. sector for all age groups 24 The number of people benefiting from indoor residual ACT provided free of charge in public sector only for children under five 5 spraying rose from 10 million in 2005 to 73 million in 2009, ACT delivered at community level 25 or about 10 percent of the population at risk for malaria Pre-referral treatment with (WHO 2010b). In countries that have seen a rapid decline quinine/artemether IM 32 in malaria within the past five years, bed net distribution has Artesunate suppositories 25 been complemented with indoor residual spraying in the Source: WHO 2010. areas of the countries where it is epidemiologically justified. 422 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY Figure 23.4 Percentage of Children under Five with The data also point to deficiencies in the use of intermit- Fever Receiving Artemisinin-Based tent preventive treatment of pregnant women. In three Combination Therapy in the Public and countries (Malawi, Senegal, and Zambia), at least 70 percent Private Sector, in Selected Countries of pregnant women reported having received SP/Fansidar during an antenatal care visit (figure 23.5).6 In contrast, less 14 than 10 percent of pregnant women in Benin, Niger, and 12 Nigeria and less than 20 percent of women in the Demo- 10 cratic Republic of Congo reported receiving intermittent preventive treatment during an antenatal care visit. 8 6 PARTNERSHIP AND COORDINATION 4 FOR RESULTS 2 During the past decade governments, communities, donors, 0 and individuals have been working under growing consen- 06 20 o 08 08 07 06 07 S ng sus and coordinated action. Since 2005 partners and 20 20 20 20 20 20 07 H o D fC S S S S S S H H H H H H malaria-endemic countries have placed greater emphasis on .o D D aD D aD aD ep n na ria ny nd bi ni strengthening coordination of technical and human .R ha be m Be Ke ga G em Za Li U resources at global, regional, and country levels. This D Public sector Private sector intense collaboration, under the umbrella of the Roll Back Malaria Partnership (box 23.4), has been a key component Source: WHO 2010b. Note: DHS is Demographic Health Survey. of the success on the ground in Africa. These partnerships have helped reduce duplication in efforts, increase resource mobilization, raise awareness of the problem at the global and regional level, and create a network of technical and implementation experts for countries to draw on as they Box 23.2 The Affordable Medicines Facility implement strategies to control malaria. for Malaria Funding for malaria control has increased dramatically since 2005 (figure 23.6). Evidence-based advocacy efforts Most malaria cases are treated with an over-the- have led to a number of notable achievements, including counter drug such as chloroquine. Resistance to this drug has increased, however, making it increasingly increased funding from partners such as the President’s ineffective. As a result, in 2004 WHO recommended Malaria Initiative, the World Bank, and the Global Fund to that all countries use ACTs as first-line treatment for Fight AIDS, Tuberculosis and Malaria. As a direct result of malaria. ACTs are not widely available at private phar- these efforts, in 2008 the Global Fund increased its funding macies and other medical distribution centers to malaria to the historic level of $1.62 billion worldwide. In because of their high cost. addition, the Roll Back Malaria Partnership formed the Har- The Affordable Medicines Facility for Malaria monization Working Group, which seeks to harmonize (AMFM) is a financing mechanism, managed by the donor resources to support countries with costed national Global Fund to Fight AIDS, Tuberculosis and Malaria, malaria control plans. Partners also joined forces to develop designed to subsidize ACTs. The initiative negotiates a the Global Malaria Action Plan, which provides a framework lower price with pharmaceutical producers. It aims to for current and future malaria control effort worldwide. make ACTs even more affordable through a co- Aid harmonization for malaria has improved, and payment and subsidy system that will allow first-line buyers, including in-country private sector whole- malaria disbursements rose to $1.5 billion in 2009. But new salers, hospitals, and nongovernmental organizations, commitments for malaria control plateaued in 2010 at $1.8 to pass on the benefit to patients, who will pay billion. Government spending for health in high-burden $0.20–$0.50 for ACTs—comparable to what they countries remains low, with far less than 15 percent of budg- currently pay for less effective alternatives. ets spent on the health sector in more than 90 percent of Source: Authors. countries. The level of spending for malaria falls far short of the resources needed to control the disease, which the RBM CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 423 Box 23.3 Improving Access to ACTs in Zambia Recognizing that access to proper treatment needed to districts, while the remaining 4 districts relied on phar- be improved, the World Bank, the United Kingdom’s macy technologists. Department for International Development, and the Model B eliminates the intermediate storage of drugs U.S. Agency for International Development (USAID) at the district level. The district store is converted into a joined forces to try out new drug distribution methods “cross-dock,” a point of transit that receives shipments in 16 districts in Zambia. The results from this pilot, from MSL that are pre-packed for individual health facil- evaluated through a rigorous impact evaluation, have ities. Under this option, the district does not carry any been exceptionally encouraging. Simple but smart stock or perform any secondary picking and packing. steps to grease the supply chain for lifesaving drugs— This system has the potential to reduce the scope for pil- including hiring district-level planners to help manage ferage and leakages because it enables better shipment orders and deliver them more efficiently—have proved tracking.2 However, it hinges on the health facilities very effective. transmitting their order requisitions to MSL every Two models were tested. Under Model A drug stock month to allow for the assembling of packages to indi- continues to be held at the district level. The new posi- vidual health facilities. As in Model A, a CP was added to tion of commodity planner (CP) at the district level, the district store under this option to ensure the smooth designed to enhance planning capacity, is responsible flow of order information from the health facilities to for coordinating orders from the health facilities and MSL. stock management at the district. The CP also ensures Remarkable improvement in access to essential drugs that requisitions requests are sent every month by each has occurred at the health facility level, particularly in dis- health facility to the district store and performs picking tricts in which Model B was implemented. In districts and packing operations at the district level to fulfill the implementing Model A (panel a of figure B23.3), 38 per- order requisitions of health facilities under that dis- cent of health facilities were out of stock of DepoProvera trict. The CP also estimates the overall requirements (a contraceptive) at baseline. At the end of the pilot pro- and places orders to Medical Stores Limited (MSL)1 for gram, the stock-out rate had fallen to 17 percent. Reduc- the stock needed to maintain the desired inventory tions in stock-out rates of the same magnitude are level at the district store. CPs were recruited to hold the observed for amoxicillin and ACT for adults. Reductions logistics responsibilities in 12 of the intervention in stock-out rates were even more dramatic in districts Figure B23.3 Baseline and Endline Stock-Out Rates following Implementation of Pilot Program a. Districts implementing Model A b. Districts implementing Model B SP (malaria prevention) 58 SPa (malaria prevention) 52 46 16 DepoProveraa (contraceptive) 38 DepoProveraa (contraceptive) 1 45 17 CTXa (antibiotics) 42 CTXa (antibiotics) 40 37 33 Amoxicillin (antibiotics) 74 Amoxicillina (antibiotics) 72 31 16 ACT adulta (malaria treatment) 43 ACT adulta (malaria treatment) 48 22 6 ACT pediatric (malaria treatment) 34 ACT pediatrica (malaria treatment) 43 30 12 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 Percent Percent Baseline Endline Source: Authors. Note: Figures show the percentage of health facilities that were out of stock of the drug indicated. The intervention had similar impacts on almost all essential drugs tracked. a. Reduction in stock-out rate is statistically insignificant compared with observed change in control districts. 424 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY Box 23.3 (continued) implementing Model B (panel b of figure B23.3), where and over-five mortality by 25 percent. These reductions decreases in stock-out rates exceeded 40 percentage points would avert 3,320 under-five deaths and 448 over-five for SP, DepoProvera, amoxicillin, and ACT for adults. deaths from malaria each year. In economic terms, the Zambia is in the process of scaling up Model B aggregate household income loss averted as a result of nationwide. A conservative estimate suggests that this the national scale-up of Model B is estimated to exceed effort could decrease under-five mortality by 21 percent $1.6 million a year. Source: World Bank 2010. 1. The receipt, storage, and primary distribution of drugs and other health commodities from Lusaka to approximately 120 districts stores and hospitals are managed by a parastatal agency called Medical Stores Limited (MSL), currently under the management of Crown Agents. Improvements of MSL have been observed in recent years as a result of large capital investments and management reforms. Consequently, primary distribution is quite reliable and effective. 2. Zambia already had a very similar arrangement in place that has eliminated stock-outs of antiretrovirals at almost all antiretro- viral therapy sites. Figure 23.5 Treatment of Pregnant Women with SP/Fansidar during Antenatal Visit in Selected African Countries 80 70 60 50 Percent 40 30 20 10 0 S S s S S S S IS IS H H H H H AI H M M D D D D D D 8 8 09 07 –0 06 08 09 08 4 00 00 8– 20 07 20 20 8– 20 a2 i2 00 20 00 go bi n ria na w l2 ni m a2 on ha ia ala ige Be ga Za an G ny fC M ne N nz Ke Se .o Ta ep .R em D Source: Authors’ compilation based on data from national household surveys. Note: DHS is Demographic Health Survey. MIS is Malaria Indicator Survey. AIS is AIDS Indicator Survey. Global Malaria Action Plan estimates at about $5 billion a show that with a combination of effective prevention and year in 2010–15 and $4.75 billion in 2020–25. The shortfall treatment tools, country leadership, and partner coordina- underscores the need for policy makers and development tion, controlling malaria in Africa is possible. partners to continue to work collaboratively to ensure that adequate resources are mobilized to scale up and sustain control efforts over the long term. Eritrea In 1997 and 1998 Eritrea experienced a series of malaria epi- SUCCESS STORIES demics that produced more than 424,000 cases; more than This section highlights four countries that are showing sig- 500 inpatient malaria deaths were recorded in 1998 alone nificant progress in their malaria control efforts. The profiles (RTI 2005). A 2002 Demographic Health Survey revealed CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 425 Box 23.4 The Roll Back Malaria Partnership The Roll Back Malaria (RBM) Partnership is the global RBM’s overall strategy aims to reduce malaria framework for implementing coordinated action against morbidity and mortality by reaching universal cover- malaria. It mobilizes resources and forges consensus age and strengthening health systems. The Global among more than 500 partners, including malaria- Malaria Action Plan defines three stages of malaria endemic countries, their bilateral and multilateral devel- control: scaling up for impact of preventive and ther- opment partners, the private sector, nongovernmental apeutic interventions, sustaining control over time, and community-based organizations, foundations, and and elimination. research and academic institutions. The RBM Partnership was launched in 1998 by RBM’s strength lies in its ability to form effective part- WHO, UNICEF, the United Nations Development Pro- nerships globally and nationally. Partners work together gramme, and the World Bank, in an effort to provide a to scale up malaria control efforts at the country level, coordinated global response to the disease. It is served by coordinating their activities to avoid duplication and a secretariat, hosted by WHO in Geneva, that works to fragmentation and to ensure optimal use of resources. facilitate policy coordination at the global level. Source: Authors. Figure 23.6 Commitments and Disbursements for Antimalaria Activities in Malaria-Endemic Countries in Africa, 2000–09 1,629 1,800 1,472 1,600 1,400 1,200 1,055 1,037 Millions of dollars 1,000 745 800 692 599 600 518 393 388 400 204 200 175 99 108 35 45 44 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Commitments Disbursements Source: WHO 2010b. that less than 20 percent of households had at least one antimalarial treatment for the last birth. Twelve percent of mosquito net, just 7 percent of pregnant women slept under children under five were reported as having slept under a a mosquito net the night before the interview, and less than mosquito net the night before the interview, and just a third half of those used an insecticide-treated net. The report also of them used an insecticide-treated net; 30 percent had a indicated low use of antimalaria drugs by pregnant women, fever in the two weeks preceding the survey, but only 4 per- with just 5 percent of women who had given birth in the cent were treated with antimalarial medication—mostly five years preceding the survey reporting having received chloroquine (NSEO and ORC Macro 2003). 426 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY Efforts. The alarming condition in the country, coupled with implementation of prevention activities, diligent environ- the rise of HIV/AIDS in Africa, triggered the Eritrean Ministry mental management of malaria vector breeding sites, and of Health to establish a new strategy and program to control widespread diagnostic and treatment services, it has HIV/AIDS, malaria, sexually transmitted diseases, and tuber- reduced overall malaria morbidity from about 100,000 cases culosis in 2000. The program, dubbed HAMSET, supple- in 2000 to about 8,000 in 2008 (Ministry of Health 2008) mented the National Malaria Control Program (NMCP), These achievements have led to an overall decrease in established in 1998. The technical strategy of NMCP sought to maternal and infant mortality and a healthier population. reduce and better target indoor residual spraying to the high- est-risk areas and scale up rapid diagnosis and effective Ethiopia treatment for fever cases, environmental management activi- ties, and use of insecticide-treated nets. At the core of this Malaria was responsible for the deaths of nearly 29,000 strategy was the strengthening of disease surveillance and children—about 80 children a day—in Ethiopia in 2000 operational research activities, data from which would be used (WHO 2002). In 2003 Ethiopia had more than half a mil- to select and refine the mix of strategies and geographically lion reported malaria cases. According to a 2005 Demo- target control activities. During 1999–2005, the NMCP graphic Health Survey, malaria was the primary cause of introduced a first-line antimalarial drug and implemented health problems, accounting for 17 percent of outpatient an integrated vector control program using insecticide- visits, 15 percent of hospital admissions, and 29 percent of treated nets, indoor residual spraying in selected areas, and in-patient deaths (Central Statistical Agency [Ethiopia], and other interventions. In 2005 the Ministry of Health articu- ORC Macro 2006). During this period only 5.7 percent of lated a clear strategic plan, with the goal of “reducing malaria households owned at least one bed net, and less than 4 per- to such low levels that it is no longer a public health problem cent had an insecticide-treated net. The situation for chil- in the country” (Ministry of Health 2008). dren under five years of age was dire, with only 2.3 percent Through support from the World Bank, the Global Fund, sleeping under any net the night before the survey and less and other donors, Eritrea has been able to employ a com- than 2 percent sleeping under an insecticide-treated net. prehensive strategy against malaria through a centralized During the two weeks preceding the survey, about 19 per- approach. This approach involves targeting the use of cent of children under five were reported as having had a indoor residual spraying in the highest-risk areas, carrying fever, and only 3 percent of these received antimalarial out environmental management of malaria vector breeding drugs. Among pregnant women, only 1.6 percent slept sites, increasing distribution of insecticide-treated mosquito under any bed net the night before the survey and only nets, and expanding diagnostic and treatment services. about 1 percent slept under an insecticide-treated net. Less than 5 percent took any antimalarial drug, and only 0.5 per- Results. The fact that about 60 percent of Eritrea’s popula- cent received any intermittent preventive treatment. Just tion of 3.5 million people live in malaria- endemic or epi- one-tenth of all households reported occupying a dwelling demic-prone areas where disease is seasonal, highly focal, that was ever sprayed with insecticide to prevent malaria. and unstable puts the country in a disadvantageous position in terms of malaria control. Despite these odds, through the Efforts. Ethiopia has been trying to control and eliminate support of the World Bank and other donors, in only a few malaria for more than 60 years. The Malaria Eradication years Eritrea has been able to introduce new first-line anti- Service was established in 1959, making Ethiopia one of the malarial drugs and implement an integrated vector control pioneering countries in Africa to embark on a malaria erad- program using insecticide-treated nets, indoor residual ication effort. In 2000 the country became a cosignatory to spraying in selected areas, and other interventions. Through the Abuja Declaration, committing itself to the declaration’s these efforts, Eritrea has reduced malaria mortality by more aims of increasing coverage of malaria interventions and than 80 percent, malaria inpatient cases by 90 percent, and reducing malaria mortality by half by 2010. malaria outpatient cases by 68 percent (WHO 2009). It also The Ethiopian government has implemented a central- met the target set at the Africa Summit on Roll Back Malaria ized approach to bringing the disease under control, with in April 2000 of reducing the incidence of malaria by at least significant success. In 2009, after analysis of the results of 60 percent before 2010. the 2007 Malaria Indicator Survey as well as the discussions As a result of the strong commitment of the govern- and recommendations that followed a consultative meeting ment, Eritrea has come a long way in 10 years. Through held in Ethiopia in 2009, Ethiopia developed a six-year CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 427 (2010–15) National Strategic Plan for Malaria Prevention, reported having slept under any mosquito net the night before Control, and Elimination. In the new strategic plan, top the survey. Only 5.8 percent of women reported having taken priorities among malaria control strategies are given to any antimalarial drug to prevent or treat malaria during an community empowerment and social mobilization. These antenatal care visit during their last pregnancy, and just priorities were based on the results of the 2007 Malaria 0.3 percent of women reported having received intermittent Indicator Survey, which showed substantial differences preventive treatment during their last antenatal care visit. between the coverage and utilization of key malaria inter- ventions by the population at risk of malaria. Malaria diag- Efforts. Acting on the recommendation of its ministry of nosis, case management, disease surveillance, and epidemic health, the Rwandan government adopted two major strate- control will all be geared to serve Ethiopia’s goal of shrinking gies to reduce malaria: promoting the use of insecticide- malaria-endemic areas by 2015 and eliminating the disease treated nets for prevention, particularly among vulnerable throughout the country by 2020 (USAID and PMI 2010). groups, including pregnant women and children under five, Donors such as the World Bank, the Global Fund, the and providing early diagnosis and timely treatment with U.K. Department for International Development, UNICEF, combination therapy of amodiaquine and sulfdoxine/ and the Canadian International Development Agency have pyrimethamine (Concern Worldwide, International Rescue been providing coordinated support to the Ethiopian gov- Committee, and World Relief 2004). The prevention com- ernment in its fight against malaria by offering flexible and ponent of malaria transmission was moving along steadily. pragmatic funding to ensure that the country’s efforts have In contrast, as a result of rapid widespread resistance to been scaled up and sustained. Together they have con- older first-line treatment antimalarial drugs, the country tributed more than $151 million—for malaria control was no longer able to effectively combat the disease, forcing efforts in the country. Partner resources have supported the it to change its national treatment policy. In 2005 the procurement and free distribution of long-lasting insectici- Rwandan government made ACT the official first-line anti- dal nets, antimalarial drugs, and indoor residual spraying as malarial drug (Friends of the Global Fight against AIDS, well as the strengthening of the supply chain. Other compo- Tuberculosis, and Malaria 2008). To support the intense nents have included capacity building for other functions, efforts of the Rwandan government to control malaria and especially procurement and monitoring and evaluation. strengthen health systems, development partners such as the World Bank, the Global Fund, the President’s Malaria Results. The effort by the Ethiopian government and its Initiative, the Roll Back Malaria Partnership, WHO, Malaria partners has resulted in remarkable improvements in No More, the Bill & Melinda Gates Foundation, nongovern- malaria control and overall health in Ethiopia. Recent data mental organizations, and others contributed grants and suggest sharp declines in the number of malaria cases as loans to step up funding of programs across the country. well as declines in malaria outbreaks and deaths (Global In addition to providing assistance for prevention activi- Fund 2009). These efforts correlated with an overall ties such as bed net distribution and indoor residual spray- decrease in under-five mortality by more than 20 percent ing and funding for antimalarial treatment, some donors between 1990 and 2008, and maternal mortality declined have provided grants to finance the overall strengthening of by nearly 60 percent, from 968 maternal deaths per 100,000 Rwanda’s health system. In 2005 the Global Fund approved live births to 590. a grant of $33.9 million to finance health systems Given the scope of the problem in Ethiopia—about 68 strengthening activities. The grant reduced the cost of health percent of the population of 73 million lives in areas at risk services, improved the quality of care offered at health of malaria—the government has accomplished extraordi- centers, and financed training for more than 100 healthcare nary results. administrators in health financing and management for health information systems. Rwanda Results. Rwanda was once a country with extremely high Rwanda has made remarkable progress in malaria control. In malaria prevalence and large numbers of deaths from 2005, according to the Demographic Health Survey, only one- malaria, particularly among children. As a result of its fifth of households reported having at least one once-treated efforts, the number of malaria illnesses attended at health mosquito net. Sixteen percent of children under five, facilities declined from 1.5 million people in 2005 to 20 percent of pregnant women, and 13 percent of all women 800,000 in 2008, and the mortality rate from malaria 428 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY dropped from 41 percent in 2006 to 16 percent in 2008. In 43 percent of children under five suffered from fever or con- 2008 malaria was the third-leading cause of morbidity and vulsions in the two weeks preceding the survey, among mortality, after pneumonia and diarrhea. The 2008 interim whom about half received antimalarial drugs for their Demographic Health Survey indicated a child mortality rate symptoms. Among pregnant women, just 8 percent slept of 103 deaths per 1,000 live births, a 32 percent reduction under an insecticide-treated net, and only 36 percent from the 2005 rate of 152 (USAID and PMI 2009). received antimalarial drugs during pregnancy (DHS 2002). The percentage of households with at least one ever- treated net increased by nearly 40 percent, from 18.1 percent Efforts. To control and reverse the deteriorating situation, in 2005 to 57.2 percent in 2007–08. The percentage of chil- the Zambian Ministry of Health identified malaria control dren under five that slept under an ever-treated net the as one of its main public health priorities and included it in night before the Demographic Health Survey increased by both the National Development Plan 2006–10 and the 42 percentage points, from 15.7 percent in 2005 to 58.0 per- National Health Strategic Plan 2006–10. Through the cent in 2007–08, and the share of children under five with National Malaria Control Centre, the Ministry of Health fever in the two weeks preceding the survey decreased by developed a detailed national malaria strategic plan, which nearly 5 percentage points, from 26.2 percent in 2005 to aimed to drastically reduce the malaria burden—by scaling 21.4 percent in 2007–08. By 2008 the use of health centers up malaria control interventions—with a vision of achiev- had increased by 10.0 percent and the use of district hospi- ing a malaria-free Zambia. This plan was the backbone for tals had increased by 16.8 percent over 2005 (Malaria Free support by the country and its partners. Future 2008). The strategy adopted a two-pronged approach, aiming to Rwanda—a country whose entire population was once at bring malaria under control with key malaria control inter- risk of malaria—is now one of Africa’s frontrunners in the ventions while at the same time supporting broader fight against the disease. With increasing use of bed nets, improvements in health systems, including decentralizing long-lasting insecticidal-treated mosquito nets, antimalarial budgeting and planning, building capacity throughout the drugs, and intermittent preventive treatment during preg- supply chain for procurement and forecasting of commodi- nancy, Rwanda has significantly reduced the number of ties, and strengthening monitoring and evaluation and pro- malaria transmissions, malaria cases being reported at gram management. As a result of these combined efforts, 3.6 health facilities, and deaths from malaria. These achieve- million long- lasting insecticide-treated bed nets were dis- ments have contributed to a decrease in infant and maternal tributed throughout the country between 2006 and 2008, mortality and a healthier Rwanda. The Rwandan National raising the percentage of households with one long-lasting Malaria Control Program is continuing to pursue its very insecticidal-treated net from 48 to 72 percent. Indoor resid- ambitious goals, which include universal coverage of long- ual spraying coverage increased from 15 districts in 2006 to lasting insecticidal-treated mosquito nets, expansion of 36 districts in 2008. indoor residual spraying, subsidization of ACTs nationwide, and better monitoring and evaluation to tackle epidemics Results. The impact of Zambia’s efforts has been felt coun- (Malaria Free Future 2008). trywide. In just two years the number of malaria deaths declined 47 percent, and nationwide surveys showed a decline in parasite prevalence of 53 percent (from 21.8 per- Zambia cent in 2006 to 10.2 percent in 2008). The percentage of In 2001–02 Zambia had about 3.5 million cases and 50,000 children with severe anemia declined 68 percent (from 13.3 deaths a year from malaria, making the disease the leading percent in 2006 to 4.3 percent in 2008). The percentage of cause of morbidity and the second-highest cause of mortal- children under five with malaria parasite prevalence fell by ity. Disease prevalence was highest among pregnant women half (from 22 percent in 2006 to 10.2 percent in 2008) and and children under five, making Zambia one of the coun- the percentage of children under five with severe anemia fell tries with the highest malaria-related maternal mortality in from 14 percent in 2006 to 4.3 percent in 2008 (MIS 2008). Africa and one with an extremely high under-five mortality These successes reduced the overall under-five mortality rate (168 deaths per 1,000 live births) (DHS 2002). During from 168 in 2002 to 148 in 2008 (UNICEF 2008). The suc- this time, only 14 percent of households in Zambia owned cessful scale-up of Zambia’s malaria program contributed to an insecticide-treated net, 7 percent of children under five a 29 percent drop in the overall mortality rate for children slept under a mosquito net the night before the survey, and under five between 2002 and 2008, saving an estimated CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 429 75,000 lives over that period. The capacity of the malaria nets and scale up other proven interventions (access to program as well as supporting critical health systems inter- effective diagnosis and treatment, indoor residual spraying). ventions was strengthened for better management. Gaps in increased coverage and use of malaria control activ- ities remain substantial in many countries. To continue and Fragile progress. Zambia’s success illustrates one of the intensify the fight against malaria, increased funding is most important issues associated with malaria control required for activities that prevent and treat malaria. In efforts. Delayed external financing slowed execution of all of countries where there has been significant progress, the necessary malaria control interventions between June resources must be sufficient to, at the very minimum, main- 2009 and December 2010, causing a significant decline in tain efforts, so that the disease does not resurface. malaria outcomes in some provinces of the country. Data It is critical that political will and financing be sustained. from the 2010 Zambia National Malaria Indicator Survey Failure to protect and expand on the fragile progress taking indicates a resurgence in malaria between 2008 and 2010. place in Africa will lead to a resurgence of malaria cases and Most of this increase was in rural areas, among the poorest deaths. Learning from early successes is critical to avoiding segments of the population. In three of the country’s nine such a prospect. Several lessons emerge from African coun- provinces, about half of the decrease in malaria cases tries’ experience: between 2006 and 2008 was lost by 2010. In Luapula and the Northern Provinces, there were marked declines in owner- ■ The high level of external funding, coupled with in- ship and use of mosquito nets between 2008 and 2010, country ownership, has proven to be extremely success- which alone could explain the increase in malaria cases. ful in combating malaria. Government ownership and Zambia’s experience shows that once the disease is under initiative to work with partners in support of established control, efforts have to be sustained if gains are not to be policy in the malaria control programs at the country reversed or even worsened, because malaria can have an level is essential. even deadlier impact on a population that has lower immu- ■ Bed net coverage has rapidly increased, but other key nity to the disease. interventions are lagging, and the gaps are becoming increasingly more critical. There needs to be more focus on use of key interventions, monitoring of drug resis- LESSONS LEARNED AND THE WAY FORWARD tance, deployment of rapid diagnostic tests, and access to Effective tools for malaria prevention and treatment exist. ACTs. With concerted and collaborative efforts, implementation ■ At a minimum, maintaining current levels of funding in of these tools is starting to show results in Africa. Progress order to ensure prevention and treatment activities is in the past few years has shown that success in Africa is pos- essential in the short and medium term to protect the sible. The results of WHO’s 2010 World Malaria Report are progress that has been achieved to date. To reach the Abuja very encouraging. Since 2000, 11 countries have shown targets of the Roll Back Malaria Partnership—coverage of decreases in malaria cases after widescale implementation of 80 percent for insecticide-treated nets for people at risk, malaria control activities in the population at high risk. In treatment with appropriate antimalarial drugs for people Botswana, Cape Verde, Eritrea, Madagascar, Rwanda, with probable or confirmed malaria, indoor residual Zambia, and other countries, the number of cases has spraying for households at risk, and intermittent treat- fallen by more than half. By the end of 2010, about 289 ment in pregnancy in high-transmission areas—funding million insecticide-treated nets were to have been deliv- needs to be about four times the current level. This esti- ered to Sub-Saharan Africa, enough to cover 76 percent of mate does not take into account the resource needs to the 765 million people at risk of malaria there. Global reach the universal coverage target of the population at malaria control efforts have resulted in a reduction in the risk by the end of 2010 set by the Secretary-General of estimated number of deaths from nearly 1 million in 2000 the United Nations and highlighted in the RBM Global to 781,000 in 2009. The international community has played Malaria Action Plan. an important role in supporting government efforts to ■ Without continued and additional resources in the short reduce the impact of malaria through substantial financial and medium term, countries may risk a resurgence of and technical support in the past decade. malaria cases in the coming years. Zambia highlights this These emerging successes notwithstanding, more issue by showing that significant progress is possible but remains to be done. Sustained support is needed to replace that it is fragile and needs a sustained effort. 430 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY ■ To achieve long-term financial sustainability in national that there was a high probability that the lots accepted con- health systems, national malaria control funding as well tained no more than a specified proportion of defective as the national health budget of countries will need to goods and that the lots rejected contained a relatively high increase. In the long run, malaria control efforts will proportion of defective goods. Since the 1980s, use of LQAS need to be covered more substantially by those sources. in health assessment has increased. In 1991 WHO identi- fied LQAS as one of the more practical rapid assessment Both governments and partners need to address financial methods and encouraged further study. (For examples of sustainability. LQAS application in the health field, see Robertson 1997.) ■ Financing should focus on supporting countries that have LQAS has been employed for a range of purposes, includ- not yet scaled up their prevention and treatment efforts ing monitoring immunization program performance in while maintaining support to those that have. Southeast Asia, Peru, and the United States and assessing ■ Support needs to be provided to countries ready to move immunization coverage. To monitor immunization, toward eliminating malaria. In this context, more atten- supervisors sample records to assess compliance with tion will have to be placed on cross-border and regional immunization protocols. The number of “defective” or activities. incorrect procedures allowed per lot is set to determine ■ Strengthening health system aspects such as supply whether a facility (or lot) is accepted or rejected. LQAS chains and human resources as well the interface can also be used to assess compliance with a policy, such as between health systems and disease control will be patient screening practices or immunization administra- tion; identify areas of high incidence of specific diseases; needed to sustain gains. or assess program impact. 4. According to the Zambia MIS carried out in 2010, the NOTES percentage of women sleeping under an insecticide-treated net rose from 43 percent to 46 percent. 1. There are four types of human malaria: Plasmodium fal- 5. Once a diagnosis of malaria is established, effective ciparum, Plasmodium vivax, Plasmodium malariae, and treatment should be started within 24 hours of the onset of Plasmodium ovale. Plasmodium falciparum and Plasmodium symptoms, to avoid progression to severe malaria, for which vivax are the most common. Plasmodium falciparum, the the case fatality rate is high. most deadly form of the disease, is the predominant type of malaria in Sub-Saharan Africa. 2. The original Abuja targets were set and agreed to by 53 REFERENCES African heads of state in the Abuja Declaration on Roll Back Central Statistical Agency (Ethiopia), and ORC Macro. Malaria in Africa on April 25, 2000. They included the fol- 2006. Ethiopia Demographic and Health Survey 2005. lowing targets: at least 60 percent of those suffering from Addis Ababa and Calverton, MD. malaria have prompt access to and are able to use correct, Concern Worldwide, International Rescue Committee, and affordable, and appropriate treatment within eight hours of World Relief. 2004. Baseline Survey of the Rwanda Com- the onset of symptoms; at least 60 percent of those at risk of munity Distribution of Anti-Malarials Pilot Program. malaria, particularly pregnant women and children under http://www.concernusa.org/media/pdf/2007/10/Rwanda five years of age, benefit from the most suitable combination _Baseline_Malaria. of personal and community protective measures, such as insecticide-treated mosquito nets and other materials to Friends of the Global Fight against AIDS, Tuberculosis and prevent infection and suffering; at least 60 percent of all Malaria. 2008. Country Profile: Rwanda. http://www.the pregnant women who are at risk of malaria, especially those globalfight.org/view/resources/uploaded/Rwanda_Coun in their first pregnancies, have access to chemoprophylaxis try_Profile.pdf. or presumptive intermittent treatment. Gallup, J. L., and J. Sachs. 2001. “The Economic Burden of 3. Lot Quality Assurance Sampling (LQAS) was developed Malaria.” American Journal of Tropical Medicine and in the 1920s as a quality control technique for goods pro- Hygiene 64 (Suppl. 1): 85–96. duced on a factory assembly line. It was used to examine a Global Fund to Fight AIDs, Tuberculosis and Malaria. 2009. small number of units randomly selected from each lot. Early Evidence of Sustainable Impact on Malaria. If the number of defective items in that small sample http://www.theglobalfund.org/documents/publica exceeded a predetermined number, the lot was examined tions/onepagers/Malaria.pdf. more closely and either repaired or discarded; otherwise the Hogan, Margaret C., Kyle J. Foreman, Mohsen Naghavi, lot was accepted. The number of units tested and the maxi- Stephanie Y. Ahn, Mengru Wang, Susanna M. Makela, Alan mum allowable number of defects were determined to ensure D. Lopez, Rafael Lozano, and Christopher J. L. Murray. CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY 431 2010. “Maternal Mortality for 181 Countries, 1980–2008: A UNICEF (United Nations Children’s Fund). 2008. Zambia. Systematic Analysis of Progress towards Millennium http://www.unicef.org/infobycountry/zambia_statis Development Goal 5.” Lancet 375 (9726): 1609–23. tics.html. Malaria Free Future. 2008. Rwanda: Winning the Fight USAID (United States Agency for International Develop- against Malaria. http://www.malariafreefuture.org/ ment), and PMI (President’s Malaria Initiative). 2010a. rwanda/success.php. Country Profile Rwanda. http://www.fightingmalaria Ministry of Health of Rwanda. 2008. Integrated Approach to .gov/countries/profiles/rwanda_profile.pdf Reducing Malaria Related Morbidity and Mortality in ———. 2010b. Malaria Operational Plan Ethiopia FY 2010. Rwanda. Kigali. http://www.fightingmalaria.gov/countries/mops/fy10/ NSEO (National Statistics and Evaluation Office Eritrea), ethiopia_mop-fy10.pdf. and ORC Macro. 2003. Eritrea Demographic and Health WHO (World Health Organization). 2007. Fact Sheet 94. Survey 2002. Calverton, MD. http://www.who.int/mediacentre/factsheets/fs094/en/. Robertson, S. E., M. Anker, A. J. Roisin, N. Macklai, K. ———. 2009. World Malaria Report 2009. Geneva: World Engstrom, and F. M. LaForce. 1997. “The Lot Quality Tech- Health Organization. nique: A Global Review of Applications in the Assessment ———. 2010a. Basic Malaria Microscopy. Part 1, Learner’s of Health Services and Disease.”World Health Statistics Guide, 2nd ed. Geneva: World Health Organization. Quarterly 50: 199–209. ———. 2010b. World Malaria Report 2010. Geneva: World Roll Back Malaria Partnership. 2008. The Global Malaria Health Organization. Action Plan—for a Malaria-Free World. http://www.rbm .who.int/gmap/. World Bank. 2007. The World Bank Booster Program for ———. 2010. “Roll Back Malaria Report Confirms Aid for Malaria Control in Africa: Scaling Up for Impact. A Two- Malaria Is Working But Predictable Long-Term Funding Year Progress Report. Washington, DC. www.world- Is Essential to Achieve Millennium Development Goals.” bank.org/afr/malaria. http://www.rollbackmalaria.org/globaladvocacy/pr2010- ———. 2010. World Bank Policy Note: Enhancing Supply 03-18.html. Chain Management in Zambia. Africa Health Unit, RTI International. 2005. Combating Malaria in Eritrea. Washington, DC. http://www.rti.org/brochures/rti_Eritrea_malaria_broch Zambia NMCC (National Malaria Control Center). 2008. ure.pdf. Zambia Malaria Indicator Survey 2008. Ministry of Statesmen’s Forum. 2010. Global Health Initiative: Integra- Health, Lusaka. tion and Innovation for Better Outcomes. June 29. ———. 2010. Zambia Malaria Indicator Survey 2010. Min- UNICEF (United Nations Children’s Fund), New York. istry of Health, Lusaka. 432 CHAPTER 23: THE MALARIA CONTROL SUCCESS STORY CHAPTER 24 Health Extension Workers in Ethiopia: Improved Access and Coverage for the Rural Poor Nejmudin Kedir Bilal, Christopher H. Herbst, Feng Zhao, Agnes Soucat, and Christophe Lemiere ealth systems in Sub-Saharan African countries H rate, together with poor access to basic health services, con- often suffer from weak infrastructure, lack of tributed to one of the highest maternal and child mortality human resources, and poor supply chain manage- ratios in the world. Moreover, their use of available health ment systems. Access to health services is particularly low in facilities was low. In 2005 almost all births took place at rural areas, where the majority of the population still lives. home, with only 6 percent of women delivering in a clinic The few private outlets that are available usually favor urban or hospital (CSA and Macro International 2005). Major or wealthy areas. Together with an uneven distribution of causes of morbidity and mortality for children under age health workers, this pattern often results in little availability five were preventable. Extrapolations from the 2005 Demo- and poor quality of health services in rural areas. graphic and Health Survey showed malnutrition to be the Ethiopians’ access to services was particularly low before underlying cause of more than half of deaths of children the government came up with innovative ways of scaling under five (CSA and MACRO International 2005). In 2005 up the delivery of essential health interventions, in particu- only 1 percent of households owned a bed net, of which less lar through its Health Extension Program (HEP). The HEP than 18 percent were insecticide treated. was designed and implemented in recognition of the fact The brunt of poor health falls on the rural poor, most that the major factor underlying poor health services in of whom live out of reach of health providers. In 2005 only Ethiopia is the lack of empowerment of households and 40 percent of the population lived within 10 kilometers of communities to promote health and prevent disease. This a clinic or other health service delivery point. The number chapter reviews Ethiopia’s experience in producing and of trained health workers has historically been inadequate deploying health extension workers and summarizes some of in Ethiopia, with shortages of almost all cadres of workers, the key factors that made the program a success.1 particularly in rural areas. Throughout the early 1990s, universities and health professional training colleges focused on clinically oriented training rather than on THE PROBLEM: SHORTAGE OF HEALTH more relevant rural-oriented community health training. CARE PROVIDERS IN RURAL AREAS Although more than 85 percent of the population lives in With a population of 80 million people, Ethiopia is the these rural areas, doctors and nurses preferred to work in second-most populous country in Africa. Throughout the urban hospital settings, where professional opportunities 1990s poor nutritional status, infections, and a high fertility were better. 433 ADDRESSING THE PROBLEM: THE HEALTH and female health extension workers and female workers are EXTENSION PROGRAM thought to be more culturally acceptable. In addition, their selection is seen as empowering women. Selection is done by To address these problems, Ethiopia launched the HEP in a committee made up of members nominated by the local 2003. The program’s objectives were to reach the poor and community and representatives from the district (woreda) deliver preventive and basic curative high-impact interven- health office, the district capacity building office, and the dis- tions to all of the Ethiopian population. The HEP is an trict education office (FMOH 2007b). Upon completion of ambitious government-led community health service deliv- training, pairs of health extension workers are assigned as ery program designed to improve access to and utilization salaried government employees to kebeles (neighborhoods), of preventive, wellness, and basic curative services. where they staff health posts and work directly with individ- At the heart of this program is the production and ual households. Each kebele has a health post that serves deployment of more than 30,000 front-line community 5,000 people and functions as an operational center for a health workers. These health extension workers are posted health extension worker. Five health posts and a health cen- to rural communities across Ethiopia, where they provide ter work in collaboration and for the Primary Health Care better and more equitable access to health services for the Unit (PHCU) that serves 25,000 people. The health center poor, women, and children in a sustainable manner (Assefa serves as a referral center and logistic hub for a health post et al. 2010; Ghebreyesus 2010). The program focuses on and also offers technical support. The health post is under four major areas and provides 17 different packages to reach the supervision of the district health office and the kebele the poor and address inequities (table 24.1). administration and receives technical and practical support from the nearby health center. Training and deployment of health extension workers Health extension workers are trained to manage opera- Since 2003 the HEP has been rolled out step by step to reach tions of health posts; conduct home visits and outreach ser- full coverage of all rural villages by the end of 2010 (table vices to promote preventive health actions; refer cases to 24.2; figure 24.1). As a result of the program, the ratio of health centers and follow up on referrals; identify, train, and health extension workers to people increased from 1: 23,775 collaborate with voluntary community health workers; and in 2004/05 to 1: 2,437 in 2008/09. provide reports to district health offices. Health extension workers are recruited from the commu- Upon assignment, health extension workers conduct a nities in which they will work according to specific criteria: baseline survey of the village, using a standardized tool. they are female (except in pastoralist areas), at least 18 years They map households and the population by age category. old, have at least a 10th grade education, and speak the local They also prioritize health problems of the village, set language. Females are selected because most of the HEP targets with respect to the 17 packages of services, and draft packages relate to issues affecting mothers and children; thus a plan of action for the year. The draft plan of action is then communication is thought to be easier between mothers submitted to the village council and approved. The plans are Table 24.1 Major Areas and Packages of Health Extension Program Hygiene and environmental Disease prevention and Health education and sanitation control Family health services communication 1. Proper and safe excreta 1. Prevention and control of 1. Maternal and child health 1. Health education and disposal system HIV/AIDS communication 2. Proper and safe solid and 2. Prevention and control of 2. Family planning liquid waste management tuberculosis 3. Water supply safety measures 3. Prevention and control of 3. Immunization malaria 4. Food hygiene and safety 4. First aid 4. Adolescent reproductive health measures 5. Healthy home environment 5. Nutrition 6. Arthropod and rodent control 7. Personal hygiene Source: FMOH 2005. 434 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR Table 24.2 Number of Health Extension Workers Trained and Deployed in Ethiopia, by Region, 2005/06–2009/10 Region 2005/06 2006/07 2007/08 2008/09 2009/10 Total Afar 0 0 164 148 196 572 Amhara 3,500 2,631 680 382 330 7,342 B. Gumuz 0 59 120 315 403 924 Dire Dawa 0 33 0 0 63 142 Gambella 0 47 0 410 0 457 Harari 0 0 0 0 8 47 Oromia 1,296 3,524 2,884 4,526 524 13,487 SNNPR 1,500 2,666 2,650 800 627 8,542 Somali 0 0 420 545 327 1,427 Tigray 840 0 0 134 73 1,442 National 7,136 8,960 6,918 7,260 2,551 34,382 Source: FMOH 2009. Figure 24.1 Number of Health Extension Workers recognition. Health extension workers also work with Deployed, 2004/05–2009/10 communities through traditional associations, such as idir (community-level volunteer organizations that collect Number of trained health money on a regular basis to cover funeral costs and give extension workers 34,382 some money to the family of the deceased), mehaber (pro- 31,831 fessional and religious organizations), ekub (a scheme in 24,571 which people regularly contribute money, which at a speci- 17,653 fied time is given to one group member; this process is 9,900 repeated until all members receive a contribution), schools, 2,737 women’s associations, and youth associations. These institu- 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 tions help communicate health messages and mobilize the Source: HSDP Annual Performance Report 2009/2010. community to help with environmental cleanup, health post construction, and other efforts. also disseminated to the district and regional councils and The health extension workers’ remaining time is spent health offices. providing services, including immunizations and injectable The basic philosophy of the HEP is to transfer ownership contraceptives, at the health posts. They are trained to pro- of and responsibility for maintaining their own health to vide first aid; conduct safe and clean deliveries; diagnose individual households by transferring health knowledge and treat malaria, diarrhea, and intestinal parasites. In 2010 and skills to households. Health extension workers spend the government added the diagnosis and treatment of pneu- 75 percent of their time visiting families in their homes and monia to the HEP, following an evidence-based analysis of performing outreach activities in the community. The the potential impact of different packages of high-impact house-to-house activity starts by identifying households to interventions. The addition represents a significant step serve as role models. These households have earned the toward tackling a primary cause of child mortality. respect and credibility of the community because of their Health extension workers also participate in local politics extraordinary performance in other social aspects, such as and are part of the multisectoral local decision-making agricultural production. They are willing to change and, process. One of the two health extension workers in a village upon completion of the training, are able to persuade and sits on the village council, along with an elected village convince other households to follow appropriate health chairman, a teacher, an agricultural development agent, and practices. The model households are considered early a community representative. The council is the political adopters of health practices in line with heath extension administration of the village that serves to prioritize the packages. They help diffuse health messages, leading to the work of the HEP, provide support to health extension work- adoption of the desired practices and behaviors by the rest ers, and review their regular performance reports. of the community. Various institutions have well-defined roles in sup- Two health extension workers are expected to train 360 porting the work of health extension workers once they model households a year. The training lasts 96 hours, after are deployed. The Federal Ministry of Health provides which the household “graduates” receive a certificate as medical equipment and supplies, such as vaccines, cold CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 435 chain equipment, contraceptives, and insecticide-treated preventing HIV increasing by 78 percent in HEP villages bednets. Regional health bureaus and district health and 46 in control villages. offices pay the salaries of the health extension workers Vaccination coverage improved significantly. A study by and provide technical support and political leadership. Admassiea, Abebaw, and Woldemichael (2009) finds that a Health centers play a crucial role in providing referral care significantly larger proportion of children in villages where and technical and practical support. In addition, volun- health extension workers were deployed were vaccinated teer community health workers work closely with health against diphtheria, polio, and tetanus (DPT); measles; polio; extension staff. tuberculosis; and main antigens. A study by the Center for National Health Services indicates that more than 96 per- cent of health posts in the three largest regions of the coun- Impact of the intervention try provide immunization services. A routine report from More than 9 million households in Ethiopia (some 63 per- the Ministry of Health indicates that child vaccination is cent of all households) have completed their training on the increasing in Ethiopia: by June 2010, 86 percent of children 17 packages of the HEP. Training and “graduation” of had received Penta 3/DPT 3 vaccine, 82 percent had households is an output indicator for monitoring the per- received measles vaccine, and 62 percent had been fully formance of health extension workers (table 24.3). immunized (FMOH 2010a), an average annual increase in Health conditions, including the proportion of house- the number of fully immunized children of 15 percent holds with access to sanitation, improved disproportion- since 2006. A household survey in the four largest regions ately in HEP villages. A case control study conducted in finds that 64 percent of children received Penta 3 and HEP and non–HEP villages between 2005 and 2007 indi- 68 percent of children 12–23 months had been vaccinated cated that the proportion of households with access to against measles, one of the indicators in the Millennium improved sanitation reached 76 percent in the intervention Development Goals (MDGs) (The Last Ten Kilometers villages (from 39 percent at baseline). In contrast, access to Project 2009). Although the routine report and the house- improved sanitation in the control villages increased from hold survey differ in terms of immunization coverage, both 27 percent at baseline to just 36 percent during the follow- results indicate significant improvement. up survey period (Center for National Health Development Coverage of maternal health services also improved: in Ethiopia 2008.) Awareness of HIV/AIDS also improved, 85 percent of health posts could provide family planning with the level of knowledge of condoms as a means of services, 83 percent could provide antenatal care, 59 per- cent could perform clean deliveries, and 47 percent could provide postnatal care (Center for National Health Services 2007). Health posts had also become the major source for Table 24.3 Number of Households “Graduating” current users of contraceptives, as reported by about from the HEP, as of June 2010 60 percent of women from the four largest regions; 22 per- Number of cent of women cited heath centers. A household survey households Number of eligible, as of households conducted by the Last Ten Kilometers Project in the four September “graduating,” Coverage largest regions documents that 32 percent of married Region 2009 as of June 2010 (percent) women were practicing family planning; about 54 percent Addis Ababa 0 n.a. 0 of these women received antenatal care services and 20 per- Afar 258,572 130 0 Amhara 4,209,129 2,508,472 63 cent received focused antenatal care (four or more visits). Benishangul Some 42 percent of women received at least two doses of Gumuz 158,156 15,604 10 Tetanus Toxoid vaccine (TTII), and 54 percent were inocu- Dire Dawa 80,041 2,400 9 Gambella 72,304 – 0 lated against tetanus. The increase in the use of any contra- Harari 49,488 2,159 11 ceptive method among currently married women was also Oromia 6,011,967 4,300,287 72 higher in HEP villages (where it rose from 31 percent to SNNP 3,272,573 2,417,012 74 46 percent) than in control villages (where it rose from Somali 708,028 30,490 17 Tigray 1,030,199 703,152 66 30 percent to 34 percent). National 15,850,457 9,072,040 63 Ethiopia has made significant efforts in expanding cover- Source: FMOH 2010a. age of key malaria interventions. Major scale-up efforts began Note: n.a. = Not applicable. – = Not available. in 2004/05, with the introduction of artemisinin-based 436 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR combination therapy (ACT) as the first line of treatment, A report published in 2009 indicates that in-patient expanded use of rapid diagnostic tests, and enhanced vector malaria cases fell by 73 percent and deaths in children under control and prevention through the wide distribution of five fell by 62 percent (Otten and others 2009). Another long-lasting insecticide-treated nets coupled with targeted study reports a 48 percent reduction in morbidity, a 54 per- indoor residual spraying. Coverage with insecticide-treated cent reduction in admission, and a 55 percent reduction in nets has increased significantly since 2004/05 (figure 24.2). mortality related to malaria. The 2007 Malaria Indicator Survey documents that overall In contrast to the progress made fighting malaria, post- coverage increased by a factor of 15 between 2004/05 and natal care and assisted delivery coverage remain low. The 2009/10, with 68 percent of households in malaria-affected recently completed third round of the HEP evaluation areas protected by at least one net or indoor residual spray- shows that less than 10 percent of women in Ethiopia ing. Use of insecticide-treated nets by children under five receive postnatal care. Despite some improvements, many of and pregnant women increased to nearly 45 percent in them associated with the increased availability of health malaria-affected areas and to more than 60 percent in house- extension workers, more than 90 percent of women still give holds that owned at least one net (FMOH 2007a). A case birth at home (Last Ten Kilometer Project 2009). The main control study indicates that from roughly similar levels of underlying factor for the low rate of attended births is the coverage at baseline, ownership of nets increased more in low skill levels of health extension workers in assisting deliv- HEP villages (87 percent) than in control villages (62 per- eries and to some extent the cultural barriers to delivering at cent) during the follow-up period (Center for National a modern health facility (FMOH 2009). To deal with the Health Development in Ethiopia 2008). With this momen- problem, in 2009 HEP provided practical training to 5,000 tum, reaching universal coverage appears feasible. health extension workers on performing clean and safe Residents in HEP and control villages showed a marked deliveries. difference in seeking treatment for malaria. In HEP villages, about 53 percent of patients with fever or malaria sought ELEMENTS OF SUCCESS FOR HEP treatment with antimalaria drugs the day of or the day after the onset of symptoms. In control villages, only 20 percent Several factors contributed to the success of the HEP. This of patients sought treatment under similar conditions. The section examines these factors. baseline values were 33 percent and 26 percent. Government leadership and political commitment Figure 24.2 Number of Health Extension Workers and Ethiopia’s national health policy and strategies clearly indi- Number of Insecticide-Treated Nets cate that accelerated training of human resources for health Distributed in Ethiopia, 2004/05–2009/10 and expansion of infrastructure and supplies are the prior- ity vehicles for achieving the health MDGs. The national 40,000 40 Health Sector Development Plan is a mechanism for trans- lating Ethiopia’s health policy into action. It serves as a 35,000 35 guidebook for all stakeholders regarding priorities and Number of workers (thousands) 30,000 30 operation modalities in the health sector. Number of nets (millions) Graduated HEWs Ethiopia’s health policy focuses on providing quality pro- 25,000 25 motive, preventive, and basic curative health care services in 20,000 20 an accessible and equitable manner to reach all segments of Distributed ITNs the population, with special attention to mothers and chil- 15,000 15 dren. The policy emphasizes establishing an effective and 10,000 10 responsive health delivery system for rural residents. The health strategy indicates the need for an adequate 5,000 5 number of midwives and other required staff in order to 0 0 tackle the leading causes of neonatal and maternal mortality. The Health Sector Strategic Plan III (HSDP-III) set a target 5 6 7 8 9 0 /0 /0 /0 /0 /0 /1 04 05 06 07 08 09 of training and deploying 30,000 health extension workers at 20 20 20 20 20 20 Source: FMOH 2010a. the community level and 5,000 health officers at health CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 437 centers and primary hospitals to address both the preventive of more than 2,500 health centers. MWUD’s support in and curative aspects of morbidity and mortality. providing contractors, supervision, and quality assurance of The HEP features at the top of the agenda of health construction has been one of the key critical success factors sector leaders at all levels of the health system and in their in this area. discussions with stakeholders. In the bimonthly meeting The most important success factor has been the increase of the Federal Ministry of Health (FMOH) with develop- in expenditure allocated to implement the HEP (figure ment partners—known as the Joint Consultative Forum 24.3). Total per capita health expenditure in Ethiopia was (JCF) between the Ministry of Health and development $7.13 per capita in 2005 (FMOH 2005), 28 percent of it partners—issues related to these priority areas are at the borne by the government. The MDG needs assessment con- top of the agenda. The forum also regularly reviews the ducted in 2004 recommended an additional $3.48 per capita status of implementation of priorities and collectively per year on average over the period 2005–15 from all tackles challenges. sources, peaking at $4.55 per capita in 2015 just for scaling The other important governance structure is the up the HEP. To meet this guideline, the government com- bimonthly meeting of the FMOH and the regional health mitted to at least double its share of health expenditure over bureaus. The meeting is chaired by the minister of health the life of HSDP-III (2005–10). By 2008 total health expen- and includes the heads of the bureaus and the agencies and diture had reached $16.10 per capita, following an increase departments in the Ministry of Health. The main purpose in government expenditure to health of 77 percent over of this steering committee is to align federal and regional three years, indicating that the government is on track to stakeholders around priorities, review progress, identify fulfill its commitment over five years (FMOH/Abt 2010). challenges, share best practices, and agree on mitigation In addition to increasing expenditure, the federal gov- measures and actions. The HEP always comes at the top of ernment negotiated with the regional governments to share the agenda of the steering committee. The priorities also the cost of implementing the national health strategy. Roles appear as key messages in the opening remarks made by and responsibilities were clarified and agreed upon by the leaders and officials from the minister of health to district FMOH and the subnational health authorities (FMOH health officers during health sector–related meetings. 2009). Local governments (regions, zones, and districts) The key principle underpinning the scaling up of the took responsibility for covering the full cost of constructing HEP was ensuring ownership by the local community for health posts, partially covering the cost of construction of meaningful and sustainable change. Local governments at health centers ($230 million), and fully covering the salaries the village and district levels, together with community rep- of health extension workers and health officers ($143 mil- resentatives, formed a committee that selects candidate lion) over five years. The government supports these activi- health extension workers for training based on nationally ties, through a commitment of Br 275 million (about $21 agreed upon criteria. Local communities also contribute to million) a year for the salaries of health extension workers. the construction of health posts, by providing construction The agreement defining the roles of different levels of the materials and labor. One of the two health extension work- health system in scaling up health interventions is signed by ers sits on the village council, which discusses the perfor- the minister of health and the heads of the 11 regional mance and challenges facing HEP and provides support to mitigate challenges on a regular basis. For the construction of health centers, local governments were given the author- Figure 24.3 Total Health Expenditure in Ethiopia, ity to choose the site of construction so that it is within 1995/96–2007/08 walking distance to about 25,000 people. 20 Effective intersectoral collaboration has been demon- 16.1 strated in the last few years. Scaling up of health extension 15 workers and health officers would not have happened had it dollars not been for the strong collaboration between the Ministry 10 7.13 of Health and the Ministry of Education, which provided the 5.6 5 4.09 technical and vocational educational training (TVET) and resources, including teachers. Strong collaboration with the 0 1995/96 1999/2000 2004/05 2007/08 Ministry of Works and Urban Development (MWUD) and its offices at the subnational level facilitated the construction Source: FMOH/Abt Associates 2010. 438 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR health bureaus each year at the annual review conference. Figure 24.4 Cumulative Number of Health Posts This system set a model for using development partner Constructed in Ethiopia, EFY 2004–09 funds as a catalytic fund to provide incentives to different levels of health authorities to allocate more resources to No. of health posts 16,000 agreed-upon priorities in the health sector. 14,000 14,192 12,000 12,488 Multifaceted, systemwide approach 10,000 10,621 Number A multipronged strategy that would address systemic weak- 8,000 8,528 nesses was at the heart of the development of the HEP. Sim- 6,000 6,191 ply increasing the number of health extension workers was 4,000 not viewed as likely to be effective in isolation. The increase 2,000 had to be implemented together with improvements in 0 other system requirements. 5 5 8 9 0 /0 /0 /0 /0 /1 One important factor in the success of the HEP in rural 04 06 07 08 09 20 20 20 20 20 areas was the strategic commitment of regions and districts Source: FMOH 2009. to address complementary and supportive factors, including the availability of other health cadres; the availability of health infrastructure (health posts and health centers) Table 24.4 Ratio of Health Human Resources in together with adequate supplies of equipment and pharma- Public Sector to People in Ethiopia, ceuticals; and health information systems. All of these 2001/02–2008/09 strategic elements have been addressed, with various Category 2001/2002 2004/2005 2008/2009 degrees of success. All physicians 1:35,603 1:35,604 1: 34,986 General practitioners 1: 54,385 1:58,203 1:76,302 Infrastructure. One of the components of the sector Health officers 1:138,884 1:104,050 1: 20,638 Nurses (excluding strategy was the construction and rehabilitation of health midwives), BSc and facilities. The number of health posts staffed by health diploma 1:5,613 1: 4,980 1: 4,895 extension workers increased from 6,191 in 2004 to 14,192 Midwives (senior) 1:77,981 1: 55,782 1: 57,354 in 2010 (FMOH 2009) (figure 24.4). Funds were secured Health extension n.a. 1: 23,775 1: 2,437 workers from development partners to equip all health posts. As of 2009, health posts were present in 70 percent of the kebe- Source: FMOH. les of Ethiopia’s four largest regions (Last Ten Kilometer Note: n.a. = Not applicable. Project 2009). As of 2007 about 94 percent of villages had health posts built by the local communities or the government (Center increased, boosted by the proliferation of private training for National Health Development in Ethiopia 2007). The institutes. The number of universities involved in training majority of these health posts were equipped with basic fur- of medical doctors also increased, from three to six, over a niture. However, access to other important services (clean decade, raising the average annual output of doctors from water, electricity, telephone, and means of transportation) 90 to 350. One of the most significant achievements of was generally low in all regions. these universities was the production of cadres with inter- mediate skill mixes adapted to the Ethiopian context. Medical cadres for supervision, support, and refer- In 2005 the government began preparing a new midlevel ral of health extension workers. The training of health cadre called health officers. Health officers provide health extension workers was part of a larger government most of the curative care services at the first level of clinical effort to increase the number, distribution, and perfor- services (health center) level. The program was launched mance of health workers. Over the past few years, there has with the financial support of the U.S. Agency for Interna- been a significant increase in the intake and output of tional Development and the technical support of the Carter health training colleges and universities for other categories Center. A network of 5 universities (Jimma, Haramaya, of health workers (table 24.4). The output of nurses has Hawassa, Mekelle, and Gondar) and 21 hospitals in CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 439 7 regions was strengthened to provide theoretical and prac- perform minor surgical procedures (wound and trauma tical training to trainees. The accelerated training of health management). officers lasts five years, including three years at university ■ 80–95 percent feel confident to provide antenatal and and two years in training hospitals. After completion of postnatal care and to diagnose and manage or refer nor- their training, health officers are supposed to provide sup- mal labor and common gynecologic problems. port to the HEP by working in health centers with other ■ 55–65 percent feel confident to perform common obstet- health team members in providing curative, preventive, ric and gynecologic procedures, such as manual removal promotive, and rehabilitative health care. Their duties of placenta, instrumental delivery (vacuum), safe abor- include the following: tion, and postabortion care. The low level of confidence is attributable mainly to inadequate on-the-job training ■ Assess community health needs caused by a shortage of obstetricians, surgeons, and sur- ■ Plan, implement, and evaluate activities and resources of gical equipment and supplies. the primary health care unit ■ Collect, organize, and analyze health and health-related Accountability structure for health extension data from health institutions, communities, and other workers. A supportive accountability mechanism was relevant areas and use and disseminate the information established to support health extension workers. Supervi- to the community and other concerned bodies sors were trained and deployed in 3,200 health centers. Each ■ Conduct and provide on-the-job training to the staff of supervisor supports 10 health extension workers in 5 satel- the primary health care unit and to community health lite health posts, which together form a primary health care workers unit. An independent survey indicates that two-thirds of the ■ Provide comprehensive outpatient care and in-patient health posts were supervised during the three months pre- services ceding the survey and that 75 percent of these health posts ■ Perform minor surgical procedures received feedback from supervisors (Carter Center 2009). ■ Refer difficult cases and follow up on return to ensure continuity of care Adequate supplies and equipment. Ensuring continuous ■ Mobilize individuals, families, and communities for logistics supply—equipment, contraceptives, vaccines, health action insecticide-treated nets, delivery kits, and so forth—is a cru- ■ Promote and be engaged in intersectoral activities cial area of support to health extension workers. At the fed- ■ Undertake essential and operational health research eral level, FMOH engaged in a continuous dialogue with ■ Document and report all primary health care unit development partners to improve aid effectiveness by sup- activities porting priorities (such as the HEP) and improving harmo- ■ Oversee the equity and efficiency of health services nization and alignment to reduce transaction costs. Two pooled fund arrangements, Protection of Basic Services Public health officers can go on to become physicians or and the MDG Performance Fund, were established and public health specialists, or they can pursue master’s level made functional at the federal level. The establishment of training on emergency surgery. these funds contributed significantly to financing the pro- The health officers program has more than tripled the curement of contraceptives, vaccines, insecticide-treated uptake and production of health officers training in nets, and equipment. In addition, funds from global health Ethiopia (see table 24.4). More than 5,500 health officer stu- initiatives, particularly the Global Fund to Fight AIDS, dents have entered the training program. By August 2010, Tuberculosis and Malaria (GFATM) and the Global Alliance 3,871 (71 percent) of the overall target had graduated. for Vaccines and Immunisation (GAVI), contributed to the An independent study of graduating students (Carter financing and procurement of equipment and pharmaceu- Center 2009) reveals the following: tical supplies as part of their support to health system strengthening. ■ 90–97 percent feel prepared to diagnose and manage or Ensuring timely and continuous logistic supply requires refer common adulthood infectious diseases and com- an efficient procurement and distribution system. Building mon chronic illnesses. such a system takes time. Ethiopia had to come up with a ■ 80–90 percent feel prepared to diagnose, manage, or refer quick transitional mechanism while building sustainable surgical emergencies (acute abdominal problems) and capacity of the health sector in supply chain management 440 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR through the implementation of the logistic master plan. health extension workers grew every year, reaching 36 by One mechanism it adopted to do so was the use of develop- 2007. The same year, 140 TVET tutors were given trainers ment partners with comparative advantage in procurement instruction in HEP teaching methodology as well as an inte- and distribution of commodities to the health posts. FMOH grated refresher training course. Appropriate education negotiated with UNICEF and the United Nations Popula- materials were developed, printed, and distributed to the tion Fund (UNFPA) for the procurement and distribution TVET institutions. of vaccines, cold chain equipment, insecticide-treated nets, contraceptives, and health post kits. This partnership Mobilizing financial support from development provided a remarkable short-term solution to the logistic partners challenges associated with the HEP. The progressive increase in domestic resource allocation to Information systems. Information systems that facilitate priorities was key to ensuring sustainability. With regard to the collection, analysis, use, and dissemination of data the HEP, an agreement was reached between FMOH and were perceived as significantly improving the support pro- regional health bureaus under which the ministry mobilizes vided to the HEP as well as the quality and relevance of the funds from development partners to provide support to the HEP to beneficiary communities. Accordingly, the FMOH TVET institutions for printing the HEP training manuals designed a robust, simplified, and standardized health and tools and for procuring and distributing medical equip- management information system contextualized to the ment, insecticide-treated nets, contraceptives, and other Ethiopian setting. Family folders were developed based on supplies; subnational governments allocate domestic the 17 packages of health interventions, and health exten- resources for stipends to health extension workers during sion workers and HEP supervisors were trained on the sys- training, pay their full salary on deployment, and cover the tem’s application and use. Each household has a family costs of building the health posts. folder that records the status of its members (for family FMOH mobilized resources from development part- planning, antenatal care, expanded program of immuniza- ners to provide teaching materials, transportation ser- tion, and so forth) and the household in general (ownership vices, and other relevant supplies for the accelerated and use of a latrine, clean water supply and use, waste dis- training of health officers. To put in place 3,200 health posal, and so forth) in terms of completing the desired centers by the end of 2010, the regional health bureaus changes indicated in the HEP. The Ministry of Health is agreed that subnational authorities would construct one printing family folders to make sure that all households in matching health center for every health center the Ethiopia have a formal medical record. FMOH constructed using donor funds. As part of this agreement, FMOH covers the required medical equip- ment while the regional health bureaus cover the cost of Innovative training strategy furnishing the centers. Training more than 30,000 health extension workers could Significant support was mobilized from various devel- not have been done through traditional means. Innovative opment partners to support the new initiatives. GFATM approaches were applied through the use of existing TVET and GAVI provided financial support for the construction for theoretical training and health centers for practical of 512 health centers and procurement of medical equip- training. FMOH provided training materials; regional ment for 7,340 health posts and 300 health centers (table health bureaus provided the stipend and transportation ser- 24.5). Development partners that contributed to the Pro- vices for the students. tection of Basic Services program—the World Bank, the Health extension workers must complete a 12-month United Kingdom’s Department for International Develop- course of theoretical and field training. One-quarter of the ment (DFID)], RNE, Irish Aid, Canadian International period is allocated to theoretical teaching at TVET institu- Development Association (CIDA) , Italian Cooperation— tions; three-quarters of the period is spent in a practicum in also provided support for the procurement of medical the community. The average monthly cost of the program equipment for 2,295 health centers and 7,000 health posts, per health extension worker was about $234 for training, contraceptives worth $37 million, essential drugs worth $178 for apprenticeships, and $83 for salaries. $10 million, and insecticide-treated nets worth $7.9 mil- The TVET institutions are run by the Ministry of Educa- lion. These contributions helped provide the required tion. The number of institutions involved in the training of inputs to maintain HEP basic services. CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 441 health services by the rural poor. Investment in health Table 24.5 Support from Development Partners in Key Areas of Scaling Up of Health extension workers has been part of a wider package of sup- Interventions, 2006–2010 port services that is showing promising results. Some weaknesses remain, and not all regions show a sim- Support Development partner/area (millions of dollars) ilar level of success. Incentive packages for health human Global Alliance for Vaccines and Immunisation (GAVI) resources are lacking, and the link between compensation Health center construction (212) 25.8 and performance is weak. The capacity of the district health Health center equipment (300) 6.8 offices to provide supervision, monitoring, and evaluation is Health post equipment (7,340) 20 low. These and other constraints need to be addressed if gaps Subtotal 52.6 Global Find to Fight AIDS, Tuberculosis and Malaria (GFATM) in coverage are to be filled and access scaled up even further. Health center construction (300) 55.4 A number of lessons can be drawn from the imple- Protection of Basic Services I and II (World Bank, DFID, mentation of the HEP, which can be replicated in other RNE, Irish Aid, Canadian CIDA, Italian Cooperation) Health center equipment (1,999) 31.2 countries: Health post equipment (7,000) 5.6 Contraceptives 43.6 First, political leadership and champions at higher levels Insecticide-treated nets 31.6 are a critical success factor in improving health out- Vaccines and cold chain equipment 9.8 Artemisinin-based combination comes. Adoption of the HEP as a major political agenda therapy (ACT) 2.4 of the government at various levels of the health system Essential drugs 18.5 tightened the focus and galvanized the involvement of Subtotal 142.7 various stakeholders. Beyond the general increase in its MDG Performance Fund fiscal space to finance the HEP, the government made Health center construction 15 Health management Information sure that such increments happened at local levels. system printing 6.5 Accordingly, salaries of health extension workers, con- Contraceptives 3.1 struction of health posts, and the basic running cost of Vaccines 5 the HEP are financed mainly by subnational health Advocacy and training 2.3 authorities (regions, districts, and zones), creating the Insecticide-treated nets 4.8 Subtotal 36.7 foundation for local ownership and sustainability of the USAID program. Accelerated training of health officer Second, delivery of services and management of programs (5 universities and 21 teaching should be integrated into existing systems. Vertical hospitals) 11.8 programs and projects can be successful in the short Total 287.4 term, but they are often unsustainable. What HEP has Source: Authors’ compilation, based on data from FMOH and demonstrated is that vertically mobilized resources can Financial Resource Mobilization Directorate. be used for systemwide interventions that make dis- ease-specific programs successful while strengthening health systems. Adopting this approach avoids creating In addition to their own financial support, UNICEF and parallel systems and procedures in the delivery of ser- UNFPA helped in the bulk procurement and distribution of vices and management of programs, averting unneces- health commodities to the service delivery points by using sary administrative burdens, transaction costs, and the funds provided by the Global Fund, GAVI, the World inefficiencies. Bank protection of basic services (PBS), and the MDG Per- Third, community ownership is a key to sustainable formance Fund. UNICEF provided support in procuring impact. The major principle underpinning the HEP is and distributing insecticide-treated nets, vaccines, and transferring the right knowledge and skills to communi- health post kits to health facilities. UNFPA supported the ties and households so that they are able to adopt behav- iors that improve their own health. Households are government by procuring and distributing contraceptives. trained and certified, after which they take responsibil- ity for promoting behaviors that lead to positive health LESSONS LEARNED outcomes. Fourth, all components of the health system need to be The HEP has been successful largely because of the strong addressed to make a program work. The HEP does not political commitment to strengthening health systems, with merely train and deploy health extension workers. Sig- the ultimate goal of improving coverage of and access to nificant investment is made in setting up and equipping 442 CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR health posts to serve as formal institutional hubs for the BIBLIOGRAPHY program. A health information system has been adapted, Admassiea, A., D. Abebawa, and A. Woldemichael. 2009. and HEP supervisors have been trained and deployed to “Impact Evaluation of the Ethiopian Health Services enhance supportive supervision and continuous Extension Program.” Journal of Development Effectiveness improvement in quality of program management and 1 (4): 430–49. service delivery. Continuous assessment and in-service training have been conducted to fill the gap in capacity of Assefa, A., Alebachew, H. Fassil, N. Haniko, M. Ihalainen, health extension workers. Referral levels (to health cen- A. Zwandor, and B. Kinfegebriel. 2010. “HIV/AIDS and ters, health officers, and so forth) have been expanded to the Health-Related Millennium Development Goals. The ensure delivery of a complete package of essential ser- Experience in Ethiopia.” Lancet. vices. A major shift in the amount and modality of Carter Center. 2009. Evaluation of the Quality of Training of financing of the sector has been undertaken to support Health Officers and Quality and Utilization of Health the key components of the program. Learning Materials. Report of an Independent Assess- Fifth, buy-in and involvement of key stakeholders is crucial. ment. Addis Ababa. A unique model for partnership and collaboration has Center for National Health Development in Ethiopia. 2008. evolved in Ethiopia between the government and various Ethiopia Health Extension Program Evaluation Study, actors in the health system, including the community, 2005–2007, vol. I. Household Health Survey. Addis Ababa. development partners, and other sectors. The growing Center for National Health Services, 2007. Functioning of trust among these stakeholders has resulted in harmo- Health Posts. Addis Ababa. nization of financing, program implementation, moni- CSA (Central Statistical Authority), and MACRO Interna- toring, and evaluation, leading to further strengthening tional. Ethiopia Demographic and Health Survey, 2005. of health systems. Addis Ababa. Sixth, a program needs to be flexible and adaptable to vari- FMOH (Federal Ministry of Health). 2005. Third National ous contexts. The HEP has been implemented in settings Health Account Report. Addis Ababa. with significant diversity in socioeconomic, cultural, and geographic conditions without compromising the basic ———. 2007a. Ethiopian National Malaria Indicator Sur- principles that led to its success. Ethiopia designed three vey. Addis Ababa versions of the HEP (agrarian, urban, and pastoralist) to ———. 2007b. Health Extension Program in Ethiopia Pro- modify and fit key aspects of program implementation file. Addis Ababa. in these widely varying contexts. This flexible nature of ———. 2009. Report of the Joint Review Mission of HSDP the program provides key lessons that are unique to the III. Addis Ababa. different environments, facilitating replication in other ———. 2010a. Annual Performance Report of HSDP. Addis countries. Ababa. Seventh, success of a program or an intervention should be ———. 2010b. Ethiopian National Malaria Indicator Sur- assessed by concrete and measurable improvements in vey. Addis Ababa. health outcomes. Implementation of HEP has shown FMOH (Federal Ministry of Health), and Abt Associates. encouraging results in a short time in increasing cover- 2010. Fourth National Health Account Report. Addis age of essential interventions and reducing morbidity Ababa. and mortality related to communicable health diseases. Ghebreyesus, A. T. 2010. “Achieving the Health MDGs: Country Ownership in Four Steps.” Lancet 376 (9747): NOTE 1127–28. http://www.lancet.com/journals/lancet/article/ 1. Certain criteria were applied to identify an intervention PIIS0140-6736(10)61465-1/fulltext?_eventId=login. as successful. An intervention is considered successful if it Last Ten Kilometers Project. 2009. Baseline Household addresses diseases of public health importance; is owned Health Survey: Amhara, Oromiya, SNNP and Tigray. JSI and financed by the government (to ensure sustainability); Research & Training, Inc., Addis Ababa. fits into the country’s conventional health system without Otten, Mac, Maru Aregawi, Wilson Were, Corine Karema, creating parallel structures and systems; is flexible enough Ambachew Medin, Worku Bekele, Daddi Jima, Khoti to be applied in different socioeconomic, cultural, and geo- Gausi, Ryuichi Komatsu, Eline Korenromp, Daniel Low- graphic settings; is embraced and supported by develop- Beer, and Mark Grabowsky. 2009. “Initial Evidence of ment partners, nongovernmental organizations, and other Reduction of Malaria Cases and Deaths in Rwanda and stakeholders; is delivered at low cost; and shows concrete Ethiopia Due to Rapid Scale-Up of Malaria Prevention results in terms of improving health outcomes. and Treatment.” Malaria Journal 8 (14). CHAPTER 24: HEALTH EXTENSION WORKERS IN ETHIOPIA: IMPROVED ACCESS AND COVERAGE FOR THE RURAL POOR 443 CHAPTER 25 Family Planning Trends in Sub-Saharan Africa: Progress, Prospects, and Lessons Learned Mona Sharan, Saifuddin Ahmed, John May, and Agnes Soucat ub-Saharan Africa has the highest average fertility S data from Demographic and Health Surveys in 45 coun- rate in the world. In 2009 the total fertility rate tries (all countries with more than one survey during the (TFR), or the average number of births per woman, reference period were included). Data from the World was 5.1—more than twice that in South Asia (2.8) or Latin Bank’s World Development Indicators and the United America and the Caribbean (2.2) (World Bank 2009). The Nations were also used for additional analyses. The chapter average contraceptive prevalence rate (22 percent) is less discusses the impact of micro and macro level factors, than half that of South Asia (53 percent) and less than a that have contributed to changes in the dynamics of con- third that of East Asia (77 percent) (World Bank 2009). traceptive use and fertility decline. It highlights national Population in the region continues to grow at a faster rate policies, institutional frameworks, and service delivery (2.3 percent) than in other regions, including both Asia and strategies in selected countries in which the greatest Latin America (1.1 percent each) (UN DESA 2008). progress has been made. These dismal indicators at the aggregate level conceal ongoing and imminent fertility transitions that are taking TRENDS IN CONTRACEPTIVE USE place at the country level. Contrary to the popular percep- tion, there is evidence of progress in fertility decline in many Contraceptive use indicators—including the contraceptive countries across Africa (Cohen 1998). Emerging research prevalence rate (CPR), method mix, and unmet need—are shows that fertility transition has already begun in Sub- showing encouraging progress in some countries of the Saharan Africa and that some countries are undergoing region. Such changes are often precursors of fertility decline. dynamic and unprecedented changes in fertility patterns. Acceptance of family planning in the region has tradi- Changes in the contraceptive prevalence rate tionally been low and cultural resistance to family planning high (Caldwell and Caldwell 1987). Nevertheless, over the The modern contraceptive prevalence rate—the proportion past two decades, contraceptive use increased in several of women of reproductive age who are using a modern con- countries. Its impact, along with that of changes in other traceptive method—varies widely across the region. Among determinants of fertility, is leading the onset of fertility women of reproductive age, CPRs for modern methods decline in the region. ranged from 1.2 percent in Somalia to 60.3 percent in South This chapter examines contraceptive use and fertility Africa (figure 25.1). Geographic variations in family plan- trends in Sub-Saharan African countries between 1986 ning use were apparent in the findings, with countries in and 2009. It is based on an analysis of household survey Southern Africa reporting the highest levels of contraceptive 445 Figure 25.1 Modern Contraceptive Prevalence Rates in Sub-Saharan Africa, by Country Somalia 1.2 Chad 1.7 Guinea 4 Angola 4.5 Niger 5 Eritrea 5.1 Congo, Dem. Rep. of 5.8 Benin 5.9 Sierra Leone 6 Guinea-Bissau 6.1 Mali 6.3 Côte d’Ivoire 8 Mauritania 8 Burundi 8.5 Central African Republic 8.6 Nigeria 9.1 Senegal 10 Liberia 10.3 Togo 11.1 Gabon 11.8 Mozambique 11.8 Cameroon 12 Congo, Rep. 12.7 Gambia, The 12.7 Burkina Faso 13.3 Ethiopia 13.7 Ghana 16.6 Madagascar 16.7 Djibouti 17.1 Uganda 17.9 Comoros 19.3 Rwanda 26.1 Zambia 26.5 São Tomé and Princípe 27.4 Kenya 31.5 Lesotho 35.2 Malawi 38.4 Mauritius 39.3 Botswana 42.1 Swaziland 46.8 Namibia 53.5 Zimbabwe 57.9 South Africa 60.3 0 10 20 30 40 50 60 Percent East Africa Central Africa West Africa Southern Africa Source: United Nations Population Division 2009. use, followed by countries in East Africa. With a few excep- Levels of fertility were high throughout Sub-Saharan tions, West and Central African countries report very low Africa during the 1970s. Some indication of fertility transi- rates of family planning use. Some of the lowest contracep- tion began to emerge in the 1980s in some parts of Africa. tive prevalence rates in the world exist in these two subre- Evidence was accumulating that fertility was falling or gions of Africa. expected to fall and contraceptive prevalence was high in 446 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED Southern Africa, at least in comparison with the levels for Analysis comparing African countries with countries in Africa as a whole in the 1970s (Lucas 1992). Parts of East other regions shows that the rate of percentage change in Africa were also showing signs of change. Surveys carried CPRs in Malawi, Mozambique, Namibia, and Zambia was far out in the mid-1980s reported pockets of high contraceptive greater than in many South Asian, East Asian, and Latin prevalence in Botswana, Kenya, and Zimbabwe and an American countries. Although fertility transition began late increasing desire among more than one-third of women in in Africa, the increase in CPR occurred from a very low base, these countries to stop childbearing—a proportion that was thus there was scope for a greater magnitude of change rela- above the 10 percent average for the region (Way, Cross, and tive to countries in other regions that had already attained Kumar 1987). high levels of coverage (figure 25.3). More recent data corroborate the onset of fertility decline in parts of the region. Wide variation in contraceptive preva- Changes in the choice of contraceptive method lence persists across countries, suggesting that country-level contexts and policies may underlie these differentials. An indicator of progress in family planning adoption is the Although overall progress is only modest, the experience change in the type of contraceptive methods used by family of a few countries in increasing contraceptive prevalence planning acceptors. The use of traditional methods tends to stands out. An analysis of fertility trends in 23 countries of be higher in settings where acceptance of family planning is Sub-Saharan Africa from 1980 to 1995 showed evidence of low and use of family planning programs is weak. Tradi- fertility decline in two-thirds of the countries, with a par- tional methods have a high failure rate compared with mod- ticularly rapid decline in Kenya and Zimbabwe (Kirk and ern methods and are therefore not considered an effective Pillet 1998). mode of contraception. Trend data on modern CPR over the past 20 years show Trends in contraceptive choice show that in many that some countries have made remarkable progress countries of the region, use of traditional methods has (figure 25.2). Countries such as Namibia and Zimbabwe declined and use of modern methods increased started out with high levels of contraceptive prevalence in (figure 25.4). The use of modern methods has increased the 1990s and saw their rates climb steeply over the next most markedly in countries that had the greatest increases two decades. Other countries, such as Malawi, Madagas- in CPR (Madagascar, Malawi, Namibia, Zambia, and Zim- car, and Mozambique, began with relatively lower CPRs in babwe). Use of traditional methods in these countries has the early 1990s, but these rates sharply increased in the either remained stagnant or has decreased. Ghana, Kenya, following years. Progress was apparent not only in South- Tanzania, and Uganda showed increases in use of modern ern Africa but also in countries in East Africa, where methods while maintaining use of traditional methods. In increases in Zambia, Uganda, and Tanzania were particu- West African countries such as Benin, Burkina Faso, larly noteworthy. Cameroon, Senegal, and Togo, traditional method use CPRs between the first and the most recent Demo- declined and relatively modest gains in modern method graphic and Health Survey during the study period were use were observed. compared to examine the rate of change over time. The Family planning programs that have been successful in findings reveal that some countries experienced dramatic Africa have promoted birth spacing. Marriage patterns in increases in contraceptive prevalence within relatively Africa differ from those in Asia, possibly accounting for a short periods of time. The increases in CPR in Namibia, cultural preference for spacing methods. Various studies Tanzania, Zambia, and Zimbabwe were particularly rapid. in the region document African cultural preferences for Mozambique experienced the steepest increase in modern spacing rather than limiting births (Cohen 1998). In con- CPR within the shortest time frame in the region: trast to Asian family planning programs, which have between 1997 and 2003 its CPR increased more than four- emphasized permanent contraceptive methods, such as fold, from 5.6 to 25.5 percent. Although other countries sterilization and abortion, programs in Africa rely on reported larger increases, changes there occurred over temporary methods, such as pills, injectables, and longer time periods. Data show that between the first and implants (Caldwell and Caldwell 1988). It has been sug- the most recent Demographic and Health Survey, all gested that successful program strategies in Africa must countries reported increases in CPR. Malawi, Mozam- promote methods that are temporary, can be used bique, Namibia, Zambia, and Zimbabwe made the great- covertly by women, and do not have to be stored at home est progress. (Caldwell and Caldwell 2002). CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 447 448 Figure 25.2 Trends in Modern Contraceptive Prevalence Rates in Selected Countries in Sub-Saharan Africa Angola Benin Botswana Burkina Faso Burundi Cameroon Central African Republic Chad 60 40 20 0 Comoros Côte d’Ivoire Congo, Dem. Rep. of Djibouti Eritrea Ethiopia Gambia Ghana 60 40 20 0 Guinea Guinea-Bissau Kenya Lesotho Liberia Madagascar Malawi Mali 60 40 20 0 Mauritania Mauritius Mozambique Namibia Niger Nigeria Rwanda Senegal 60 40 20 0 Sierra Leone Somalia South Africa Swaziland Togo Uganda Zambia Zimbabwe 60 40 20 0 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 East Africa Central Africa West Africa Southern Africa Source: Demographic and Health Surveys, various years. 449 Percentage change 0 1 2 3 C ha In d d N ia M ige or r Vi occ et o n Li am b Se eria ne ga M l Health Surveys Er ali i G trea ui ne Be a N ni Az ige n er ria ba ij Source: Demographic and Health Surveys, various years. Ph T an ilip og pi o n G es Bu Rwhan rk an a in da aF a G Yemso ua e te n G ma H eor la Sub-Saharan Africa o g El ndu ia Sa ra Zi lva s m do ba r Asia bw Ke e n C Uga ya am nd Ba er a ng oo la n Pa des kis h t Jo an rd a H n N i ait e Bo pal In liv do ia North Africa/Central Asia n U esia Takrai C nza ne ol ni o a Ec mbi N uad a ica o r C E ragu Latin America D ôt th a e io om d p in C ’Ivo ia ica am ir n bo e Re d pu ia Eg yp M Za blic t, ad mb Ar ag ia ab as Re car p. o Pe f M ru ala B wi Pa raz Figure 25.3 Annual Rate of Change in Modern Contraceptive Use in Selected Countries between First and Last Rounds of Demographic and ra il g M Na uay oz m am ibi bi a qu e 450 Figure 25.4 Trends in Modern and Traditional Contraceptive Prevalence Rates in Selected Countries in Sub-Saharan Africa Benin Burkina Faso Cameroon Chad Côte d’Ivoire 50 25 0 Eritrea Ethiopia Ghana Guinea Kenya 50 25 0 Liberia Madagascar Malawi Mali Mozambique 50 Percent 25 0 Namibia Niger Nigeria Rwanda Senegal 50 25 0 Tanzania Togo Uganda Zambia Zimbabwe 50 25 0 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 80 90 00 10 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 19 19 20 20 modern methods traditional methods Source: Demographic and Health Surveys, various years. The use of family planning methods depends not just on TRENDS IN TOTAL FERTILITY RATES users’ preferences but also on health system characteristics. Another set of analyses examined trends in total fertility Strong family planning programs rely on effective family rates, or the average number of children per woman. In planning service delivery strategies, such as those that most countries the TFR declined over time; in some coun- offer methods tailored to the needs of users, provide family tries it remained stagnant (figure 25.7). The steepest planning counseling and medical expertise for administer- declines in average fertility were observed in Ghana, Kenya, ing methods, and follow up on users’ response to the Liberia, Namibia, and Zimbabwe. Other countries, such as method. Countries in the region with frail health systems Madagascar, Senegal, and Togo, also showed promising are faced with the challenge of improving contraceptive declines. The TFR increased or remained constant in a few method choice within existing constraints. There is growing countries, including Mozambique, Niger, and Nigeria. evidence that new methods such as injectables are being Despite increases in contraceptive prevalence in many readily accepted by women in the region; these methods countries, fertility decline has been slow. As there tends to accounted for 62 percent of modern contraceptive users in be a time lag between changes in contraceptive use behavior Malawi and 66 percent in Ethiopia (National Statistical and a corresponding decline in average fertility, it is likely Office Malawi and ORC Macro 2005; Central Statistical that subsequent rounds of surveys will show greater fertility Agency and ORC Macro 2006). declines in countries in which CPRs have risen. Fertility decline also tends to be correlated with demographic and socioeconomic factors, such as the level of urbanization, Changes in unmet need and satisfied demand women’s education, women’s labor force participation, and Unmet need measures the gap between demand for family economic growth. Studies in Africa have shown that differ- planning and use of contraception. Expressed as the per- entials in fertility trends across countries are associated with centage of sexually active women who do not want addi- women’s education, child survival (Kirk and Pillet 1998), tional children but are not using any family planning and exposure to modern roles and behaviors linked with method, this measure is often considered a precursor of fer- growing urbanization (Garenne and Joseph 2002). tility decline, because it indicates that demand for family Trends in actual and wanted TFR were examined in planning services exists but is not being met. each of the countries for which data were available. Actual Changes in unmet need can be influenced by a variety of TFR exceeded wanted TFR in all countries (figure 25.8), factors related to fertility preferences or family planning indicating that women were not able to regulate their fer- acceptance, which may or may not be related to the effec- tility preferences. In countries such as Uganda, that have tiveness of family planning programs. Nevertheless, when high unmet need, the gap between the actual and wanted examined in relation to contraceptive prevalence, these TFR widened, suggesting a growing demand for small changes provide an estimate of the gap between demand family size and the failure of family planning programs to and utilization of family planning. meet the latent demand for services. Countries, such as Trends in modern CPR and unmet need indicate that in Kenya and Zimbabwe, that have had strong family plan- countries such as Kenya, Madagascar, Malawi, and Zambia, ning programs and have improved contraceptive preva- decline in unmet need has corresponded with an increase in lence showed a narrowing of the gap between desired and family planning (figure 25.5) suggesting a convergence of actual fertility. demand and supply of family planning. In contrast, in other Some countries that made progress in the 1990s started to countries, such as Eritrea, Ghana, Mali, and Senegal, the gap falter at the beginning of the 2000s. Stagnation on contra- has remained wide and consistent. ceptive prevalence and total fertility rate, was evident (see Satisfied demand for contraception is defined as the per- figures 25.2, 25.7). Uganda, Tanzania, Malawi, Kenya, and centage of sexually active women who do not want addi- Ghana underwent large increases in contraceptive preva- tional children and are practicing family planning. Increases lence in the 1990s but stagnated after 2000. Stagnation in sat- in satisfied demand corresponded with a decrease in the isfied demand was evident in Kenya, Malawi, Tanzania, and unmet need for family planning in certain countries in the Zambia. Total fertility rates also stagnated in the 2000s, espe- region. The percentage change in satisfied demand cially in Cameroon, Ghana, Kenya, and Zimbabwe. In some increased most in countries such as Madagascar, Mozam- countries, such as Rwanda, Tanzania and Uganda, the gap bique, Tanzania, and Zambia where contraception preva- between actual and wanted fertility widened after 2000, a lence rates are increasing (figure 25.6). CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 451 452 Figure 25.5 Unmet Contraceptive Need and Modern Contraceptive Prevalence Rate in Selected Countries in Sub-Saharan Africa Benin Burkina Faso Cameroon Chad Côte d’Ivoire 50 25 0 Eritrea Ethiopia Ghana Guinea Kenya 50 25 0 Madagascar Malawi Mali Mozambique Namibia 50 Percent 25 0 Niger Nigeria Senegal Tanzania Uganda 50 25 0 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 Zambia Zimbabwe 50 25 0 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 Modern contraceptive prevalence rate Unmet need Source: Demographic and Health Surveys, various years. Figure 25.6 Satisfied Demand for Contraception in Selected Countries in Sub-Saharan Africa Benin Burkina Faso Cameroon Chad Côte d’Ivoire 80 60 40 20 Eritrea Ethiopia Ghana Guinea Kenya 80 60 40 20 Madagascar Malawi Mali Mozambique Namibia 80 Percent 60 40 20 Niger Nigeria Rwanda Senegal Tanzania 80 60 40 20 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 Uganda Zambia Zimbabwe 80 60 40 20 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 East Africa Central Africa Southern Africa West Africa Source: Demographic and Health Surveys, various years. 453 454 Figure 25.7 Total Fertility Rates in Selected Countries in Sub-Saharan Africa Benin Burkina Faso Cameroon Chad Côte d’Ivoire 7 6 5 4 Eritrea Ethiopia Ghana Guinea Kenya 7 6 5 4 Liberia Madagascar Malawi Mali Mozambique Total fertility rate 7 6 5 4 Namibia Niger Nigeria Rwanda Senegal 7 6 5 4 90 95 00 05 10 19 19 20 20 20 Togo Uganda Zambia Zimbabwe 7 6 5 4 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 East Africa Central Africa Southern Africa West Africa Source: Demographic and Health Surveys, various years. Figure 25.8 Actual and Wanted Total Fertility Rates in Selected Countries in Sub-Saharan Africa Benin Burkina Faso Cameroon Chad Côte d’Ivoire 8 6 4 2 Eritrea Ethiopia Ghana Guinea Kenya 8 6 4 2 Liberia Madagascar Malawi Mali Mozambique Total fertility rate 8 6 4 2 Namibia Niger Nigeria Rwanda Senegal 8 6 4 2 Tanzania Togo Uganda Zambia Zimbabwe 8 6 4 2 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 90 95 00 05 10 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 19 19 20 20 20 Actual total fertility rate Wanted total fertility rate Source: Demographic and Health Surveys, various years. 455 probable consequence of the weakening of the family plan- tion) (Bongaarts, Frank, and Lesthaeghe 1984). In countries ning program or a shift in desired family size (figure 23.8). in which fertility reduction is most pronounced, there is evi- Although conclusive evidence on the reasons for the dence that fertility-reducing variables such as age of mar- stagnation remains elusive, some explanations include riage have risen (Cohen 1998). However, comparison of the changes in the international policy arena and contextual proximate determinants of fertility in countries in which changes at the country level. A key policy factor is purported the fertility transition is more advanced and those in which to be the reduced priority of reproductive health and family it is delayed indicates that contraceptive use is by far the planning after its exclusion from the Millennium Develop- most important factor accounting for intercountry differ- ment Goals as well as competition for resources from dis- ences (Kirk and Pillet 1998). eases such as tuberculosis, malaria, and HIV (Gillespie Increases in induced abortion—suspected to be a major 2004). Country-level studies (for example, in Kenya) attrib- method of contraception in urban areas of Africa—are ute the faltering of the fertility decline to dwindling donor associated with recent declines in fertility (Garenne and support for family planning and a greater emphasis on Joseph 2002). Abortion is the most likely explanation for the HIV/AIDS and other sexually transmitted diseases (Blacker drop in fertility, from 6.9 in 1980 to 5.5 in 2010, in Western and others 2005). A regional study on the stagnation of fer- Africa, where contraceptive use remains very low (Cleland, tility decline in Eastern Africa concludes that changes in Ndugwa, and Zulu 2011). An estimated 14 million unin- socioeconomic variables, the family planning program envi- tended pregnancies occur in Sub-Saharan Africa every year ronment, and reproductive behavior models are associated (Hubacher, Mavranezouli, and McGinn 2008). Conse- with the decline in contraceptive use and increases in unmet quently, the demand for medical abortion is expected to be need, preferences for larger families, and adolescent fertility very high. Because abortion remains illegal in all but a few (Ezeh, Mberu, and Emina 2009). countries in the region, women have to seek unsafe abor- The HIV/AIDS epidemic has also impacted fertility lev- tions from illegal practitioners. It is estimated that more els in Sub-Saharan Africa. The region has the highest preva- than 4 million unsafe abortions are performed in Africa lence of HIV/AIDS and the largest number of people living every year (Brookman-Amissah and Moyo 2004). Abortion with HIV/AIDS in the world. Stagnation in fertility decline is a major risk factor underlying the high levels of maternal over the past 10 years has been related to the increase in HIV mortality in Africa. prevalence. In Zimbabwe, for example, estimated total fer- The main socioeconomic determinants of fertility tility was 8.5 percent lower than it would have been in the include socioeconomic status, women’s education, and absence of HIV, and HIV-associated changes in fertility urban residence. The negative association between women’s behavior accounted for one-quarter of the drop in fertility education and fertility level observed in other settings is since the 1980s (Terceira, Simon, and Gregson 2003). In apparent in Africa as well (figure 25.9). Analysis of the rela- South Africa, where the prevalence of HIV is among the tionship between economic growth and fertility indicates highest in the region, the spread of HIV is expected to accel- that increases in GDP are associated with higher rates of erate fertility decline (Moultrie and Timaeus 2003). contraceptive prevalence (figure 25.10). The direction and pathways of causality between fertility and economic growth remain debatable. In general, socioeconomic change DETERMINANTS OF FERTILITY DECLINE is believed to modify the incentives to have children, diffuse Fertility patterns tend to be influenced by proximate and new ideas about childbearing through society, and provide socioeconomic determinants of fertility. Findings indicate women with better access to contraceptive methods (Bryant that changes on both fronts are taking place in Africa. 2007). Emerging economic growth prospects in the region The proximate determinants of fertility are the biologi- indicate potential for future fertility decline. cal and behavioral factors through which socioeconomic Other covariates of fertility, such as infant mortality and environmental variables operate to influence the rate rates, have declined in the region. A plot of the relation- of childbearing in a population (Bongaarts 1987). These ship between contraceptive prevalence rates and infant determinants have been classified into two broad categories: mortality rates indicates the negative correlation between fertility-enhancing trends (shortening of breast-feeding the two variables (figure 25.11). Fertility decline in coun- periods and postpartum abstinence, decline in pathological tries such as Botswana, Kenya, and Zimbabwe, which had sterility) and fertility-reducing trends (rise in age at first lower levels of infant mortality than other countries in the union, higher prevalence and effectiveness of contracep- region, provides evidence that improved rates of child 456 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED Figure 25.9 Relationship between Women’s Secondary Education and Contraceptive Prevalence Rates in Sub-Saharan Africa 50 Y = 7.97 + 0.47X 2 Namibia R = 0.61 Congo, Rep. of 40 Zimbabwe Contraceptive prevalence rate (percent) Swaziland Madagascar Kenya 30 Zambia Lesotho Mozambique Malawi Cameroon Tanzania 20 Uganda Congo, Dem. Rep. of Ghana Benin Burkina Faso Nigeria Liberia Guinea 10 Ethiopia Sierra Leone Niger Mali Senegal Ethiopia Chad 0 0 20 40 60 80 percent of women with secondary education Source: Demographic and Health Surveys, 2000–08. Figure 25.10 Relationship between Annual Percentage Change in GDP and Contraceptive Prevalence Rate in Sub-Saharan Africa, 1990–2009 100 Contraceptive prevalence rate (percent) 80 60 40 20 0 –5 0 5 10 15 Annual percentage change in GDP Source: World Bank 2009. CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 457 Figure 25.11 Relationship between Contraceptive Prevalence Rates and Infant Mortality Rates in Sub-Saharan Africa 50 South Africa 1998 Namibia 2006 Congo, Rep. of 2005 40 Zimbabwe 2005 Swaziland 2006 Gabon 2000 Contraceptive prevalence rate (percent) Zimbabwe 1994 Kenya 2008 30 Madagascar 2008 Zambia 2007 Lesotho 2004 Cameroon 2004 Kenya 1993 Malawi 2004 Mozambique 2003 Rwanda 2007 Togo 1998 Namibia 1992 Tanzania 2004 Burkina Faso 1993 Cameroon 1991 Côte d’Ivoire 1998 20 Uganda 2006 Congo Democratic Republic 2007 Ghana 2008 Ghana 1993 Benin 2006 Côte d’Ivoire 1994 Nigeria 2008 Benin 1996 Comoros 1996 Burkina Faso 2003 CAR 1994 Liberia 2007 Rwanda 1992 Uganda 1995 Madagascar 1992 Ethiopia 2005 Guinea 2005 Zambia 1992 Malawi 1992 10 Sierra Leone 2008 Senegal 2005 Niger 2006 Tanzania 1991 Nigeria 1990 Guinea 1999 Mali 1995 Senegal 1992 Eritrea 2002 Mali 2006 Mozambique 1997 Eritrea 1995 Ethiopia 2000 Mauritania 2000 Niger 1992 Chad 1996 Chad 2004 0 40 60 80 100 120 140 mortality rate Source: Demographic and Health Surveys, 1986–2009. survival may be a necessary condition for fertility decline major strides in improving family planning uptake can be in Africa (Caldwell, Orubuloye, and Caldwell 1992). made if political commitment exists (box 25.3). Some observers have argued that high rates of fertility in the region can be linked with the lack of policy level FAMILY PLANNING POLICIES AND PROGRAMS commitment for family planning programs. During the Some studies have tried to identify reasons why certain 1960s and 1970s, African governments were reluctant to countries in the region underwent fertility decline institute effective family planning programs; political sup- whereas others did not. One study compares Kenya, where port for family planning in the public sector was weak total fertility fell about 40 percent between 1980 and 2000, throughout the continent. Since the 1974 and 1984 World with neighboring Uganda, where fertility declined by Population Conferences, however, governments in several 10 percent. It finds that both economic development and African countries have acknowledged high levels of fertil- a strong national family planning program were associ- ity and initiated family planning programs (Kalipeni ated with lower fertility in Kenya (Blacker and others 1995). Africa has lagged other regions on fertility decline 2005) (box 25.1). because family planning programs were introduced rela- A comparative analysis of Zimbabwe, where the fertility tively late in the region. Family planning programs in rate fell more rapidly than in Zambia, reveals that a strong Africa are not as strong or as old as those in other parts of family planning program in Zimbabwe backed by high- the world, but as the experience of many African countries level political commitment and institutional and financial reveals, if strong and high-quality family planning stability were key ingredients of success (Lee et al. 1998) programs are developed, people will use them and fertility (box 25.2). Emerging evidence from Rwanda suggests that will decline (Mbacke 1994). Relatively better progress on 458 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED Box 25.1 Kenya Case Study Kenya was the first country in Sub-Saharan Africa to social marketing of contraceptives have been effective adopt a population policy, with the launch of a in increasing coverage. national family planning program in 1967. Initially, The combined program efforts of public and private the promotion of the program was halfhearted and agencies facilitated Kenya’s transformation from the ineffective: fertility appeared to be rising rather than country with the highest fertility level in the world in falling. In the 1980s, however, the government rapidly the late 1970s to one in which significant fertility expanded the number of outlets from which contra- decline has been achieved. The rate of contraceptive ceptives could be obtained and the number of health use among married women increased from 17 percent workers trained in family planning. By the end of the in 1984 to 39 percent in 1998, one of the highest rates in century, more than half of the country’s 3,500 (gov- Sub-Saharan Africa (Magadi and Curtis 2003). The main ernmental and nongovernmental) health facilities driving force behind the success was the government’s were offering maternal and child health and family effort in increasing the number of family planning service planning services. delivery points and an intensified and focused informa- In 1996, the National Council for Population and tion, education and communication strategy (Aloo- Development published its National Population Advo- Obunga 2003). The leveling off in the fertility decline after cacy and IEC Strategy for Sustainable Development, 2000 may have been caused by problems in the supply of 1996–2010, which aimed to promote the use of mod- contraceptives, weaknesses in the quality of care, and ern contraceptive methods among less educated changes in the contraceptive method mix (Pathfinder women (Blacker and others 2005). Modern methods International 2005). Another factor may have been the of contraception have been available in Kenya since HIV/AIDS program, which gradually pushed family plan- 1957 through the facilities of the Ministry of Health ning off the agenda as it became a priority for funding and and the private sector, including nongovernmental strategic programming considerations (Aloo-Obunga organizations. Community-based distribution and 2003; Pathfinder International 2005). Source: Authors. family planning indicators in Eastern Africa compared to separation of institutions responsible for policy formula- Western Africa has been attributed to stronger family tion and implementation resulted in a weaker family plan- planning efforts that ensured wider availability of modern ning program (Lee and others 1998). contraceptive methods (Cleland, Ndugwa, and Zulu 2011). A few countries have tried community-based distribu- Although family planning programs in the region have tion of contraceptives to extend family planning to hard- been weak overall, some encouraging progress in program to-reach populations, particularly in rural areas. Commu- implementation began to emerge in the 1980s. A study of nity depots, mobile clinics, women’s groups, and both paid family planning program effort finds that the greatest and volunteer village health workers are some modes of improvement among all regions of the world between service delivery used by such programs. Countries such as 1982 and 1989 occurred in Sub-Saharan Africa, where Ghana, Kenya, Nigeria, and Zimbabwe have implemented there was a sharp increase in family planning program large community-based distribution programs at the effort indicators, albeit from a low base (Mauldin and Ross national level. Although conclusive evidence on the effec- 1991). Policy-level support in countries in which strong tiveness of such programs is not available, these programs commitments existed translated into successful national provide good examples of successful bottom-up approaches family planning programs. that have been applied in the region (Phillips, Greene, and Certain family planning program management strate- Jackson 1999). gies have been found to be particularly effective in the region. Countries such as Kenya and Zimbabwe had strong LESSONS LEARNED family planning associations that spearheaded policy changes and program implementation. A unified institu- Notwithstanding the high levels of aggregate fertility in tional structure responsible for program implementation Sub-Saharan Africa, some countries in the region have was found to be effective in Zimbabwe. In Zambia the made significant progress on fertility decline. Ongoing CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 459 Box 25.2 Zimbabwe Case Study After independence, in 1980, the government of Zim- of a large national information, education, communi- babwe reformed and expanded the family planning cation campaign that promoted family planning with program with great success. The family planning pro- messages targeting men. The number of service deliv- gram was spearheaded by the Zimbabwe National Fam- ery points was increased, particularly in rural areas; ily Planning Council, a body created in the early 1980s the number of family planning personnel more than and backed by high-level leadership from Sally Mugabe, doubled in some units; and the government made the president’s wife, and Ester Boohene, his sister-in- health care free to lower-income groups, thereby law. The Zimbabwe National Family Planning Council removing a major barrier to contraceptive use (Zinaga built consensus for family planning among opinion 1992). leaders including religious groups, the business com- The availability and quality of family planning and munity, mass media, nongovernmental organizations, health services in the community was a key determi- and civil servants (Lee, and others 1998). nant of higher rates of adoption of modern contra- Initially, Zimbabwe’s family planning program was ceptives (Thomas and Muvandi 1994). The impact of clinic-based. A community-based distribution system community-based distributors was associated with was launched in 1983 and it was considered among the increased adoption of modern methods. Mobile fam- most successful programs of its type in the region ily planning clinics had a powerful impact on adop- (Way, Cross, and Kumar 1987). Distributors were tion, as did the presence of a general hospital in the selected by the communities they served and paid area. These two investments in infrastructure had an government salaries and benefits. They were respon- above-average impact on women with little education. sible for making household visits to deliver modern The program also provided a range of contraceptive contraceptives, recruit new acceptors, follow up on methods, including Norplant, the female condom, and dropouts, and make referrals where necessary (Phillips, emergency contraception (Koblinksy 2003). Family Greene, and Jackson 1999). Mobile clinics covered planning was positioned as an integral part of the about 29 percent of the rural population (Koblinksy maternal and child health program. The primary 2003). Groups of men and women were recruited to health care strategy adopted by the government motivate and educate people in communities about included both maternal and child health and family family planning. Other innovative approaches included planning. All service delivery units were instructed to the mobilization of farmers’ wives to provide contra- provide family planning as an integral part of their ceptives to workers on their farms and the launching maternal and child health services (Zinaga 1992). Source: Authors. transformations in contraceptive use and fertility behav- mitment is not enough, however; the leadership must pro- ior signal the onset of fertility declines in more countries vide contraceptives and appropriate outlets for obtaining in the years to come. Despite tumultuous political situa- them and create an environment that is conducive to tions several success stories in family planning policy adoption of family planning (Caldwell, Orubuloye, and formulation and program implementation have emerged. Caldewell 1992). Lessons drawn from countries that have made progress The existence of strong family planning programs is a attest to the importance of political commitment, institu- prerequisite to reducing fertility. Family planning pro- tional arrangements, and service delivery strategies in grams that have delivery points throughout the country; increasing the use of family planning methods and lower- provide a range of contraceptive methods; ensure easy ing fertility. availability of contraceptives; adopt a reproductive health Many countries that were successful in reducing fertility approach; and reach adolescents, men, and unmarried peo- adopted population policies and instituted family planning ple are most likely to accelerate progress toward fertility programs relatively early. Programs in Botswana, South decline in Africa (Caldwell and Caldwell 2002). Some spe- Africa, and Zimbabwe have been considered particularly cific service delivery strategies that have been found to be successful in this regard (Lucas 1992). High-level policy effective in Africa are those that promote spacing methods, commitment and political ownership of the population give women the means to assume responsibility over con- program was a key ingredient for success. Political com- traceptive adoption, and allow women to use contraception 460 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED Box 25.3 Rwanda Case Study In the 1980s Rwanda had one of the highest levels of improve access and quality of health services. Some fertility in the world (8.6). The availability of family successful strategies included the introduction of a planning service was very limited, and cultural atti- performance-based financing mechanism at various tudes dictated the desire for a large family. As a result, levels. Performance contracts were established contraceptive prevalence rate was extremely low, at between the presidency and district mayors (Imihigo) 3–4 percent in 1988 (May, Mukamanzi, and Vekemans that included an indicator on family planning. Perfor- 1990). mance incentives were also given to health workers to The National Office of Population (ONAPO) was improve their motivation levels and to health centers established in 1981 to increase demand for and improve for providing quality care. Universal health insurance the supply of family planning services. It made some schemes (mutuelles de santé) further enhanced the initial progress, but its gains were lost during the geno- coverage of health care and encouraged community cide of 1994. After the recovery, the government intro- involvement in health provision. Other strategies duced sweeping reforms that have resulted in dramatic included decentralizing health services, strengthening progress on health and family planning indicators, contraceptive supply systems, training health workers with the contraceptive prevalence rate increasing for family planning provision, and establishing sec- several-fold, skyrocketing from 4.3 percent in 2000 to ondary posts as an alternate means for providing 26.1 in 2007. modern contraceptives to circumvent that religiously The importance of providing family planning ser- affiliated health facilities were not allowed to do so vices as a rationale for national development found (Solo 2008). Rwanda’s remarkable turnaround from strong support at the top leadership level of the govern- its conflict-torn past to current achievements serves ment. A number of innovative strategies were applied to as a model for other countries in the region. Source: Authors. without their partners’ knowledge (Caldwell, Orubuloye, policy and resource commitments can easily reverse the and Caldwell 1992). gains that have been made in the past. Policies that go beyond simply increasing contraceptive Family planning remains an unfinished agenda in the prevalence to address the proximate determinants of fertil- region, because high fertility and rapid population growth ity can accelerate fertility decline in significant ways. present a great threat to the achievement of the Millennium Increasing the age of marriage through legislation and Development Goals (Cleland and others 2006). A greater behavior change, encouraging natural child spacing thrust in this direction will be required to sustain and through promotion of exclusive breastfeeding, and reducing improve the prospects for health and development in Sub- the risk of unsafe abortion by removing legal restrictions Saharan Africa in the coming decades. will be key factors underlying fertility decline in Africa (Guengant and May 2002). Improvements in health, educa- tion, and socioeconomic factors can also catalyze fertility REFERENCES reduction in the region. For example, experiences from Aloo-Obunga, C. 2003. Country Analysis of Family Plan- Botswana, Kenya, and Zimbabwe show that countries that ning and HIV/AIDS: Kenya. Policy Project, Washington, are making improvements in child survival and women’s DC. education have the greatest potential for reducing fertility Blacker, J., C. Opiyo, M. Jasseh, A. Sloggett, and J. Ssekamatte- (Caldwell, Orubuloye, and Caldwell 1992). Ssebuliba. 2005. “Fertility in Kenya and Uganda: A Com- Although fertility transition has begun in Africa, it is lim- parative Study of Trends and Determinants.” Population ited to a few countries; fertility decline across the region Studies 59 (3): 355–73. remains a distant goal. Even in countries in which contra- Bongaarts, J. 1987. “The Proximate Determinants of Fertil- ceptive prevalence is increasing, huge differentials exist by ity.” Technology in Society 9 (3–4): 243–60. socioeconomic strata, urban-rural residence, and correlates Bongaarts, J., O. Frank, and R. Lesthaeghe. 1984. “The Prox- such as female education and autonomy. The stagnation in imate Determinants of Fertility in Sub-Saharan Africa.” recent years provides compelling evidence that faltering Population and Development Review 10 (3): 511–37. CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 461 Brookman-Amissah, E., and J. B. Moyo. 2004. “Abortion Kalipeni, E. 1995. “The Fertility Transition in Africa.” Geo- Law Reform in Sub-Saharan Africa: No Turning Back.” graphical Review 85 (3): 286–300. Reproductive Health Matters 12 (24): 227–34. Kirk, D., and B. Pillet. 1998. “Fertility Levels, Trends, and Bryant, J. 2007. “Theories of Fertility Decline and the Evi- Differentials in Sub-Saharan Africa in the 1980s and dence from Development Indicators.” Population and 1990s.” Studies in Family Planning 29 (1): 1–22. Development Review 33 (1): 101–27. Koblinksy, M. 2003. Reducing Maternal Mortality: Learning Caldwell, J. C., and P. Caldwell. 1987. “The Cultural Context from Bolivia, China, Egypt, Honduras, Indonesia, Jamaica, of High Fertility in Sub-Saharan Africa.” Population and and Zimbabwe. Washington, DC: World Bank. Development Review 13 (3): 409–37. Lee, K., L. Lush, G. Walt, and J. Cleland. 1998. “Family Plan- ———. 1988. “Is the Asian Family Planning Program ning Policies and Programmes in Eight Low-Income Model Suited to Africa?” Studies in Family Planning Countries: A Comparative Policy Analysis.” Social Science 19 (1): 19–28. & Medicine 47 (7): 949–59. ———. 2002. “Africa: The New Family Planning Frontier.” Lucas, D. 1992. “Fertility and Family Planning in Southern Studies in Family Planning 33 (1): 76–86. and Central Africa.” Studies in Family Planning 23 (3): Caldwell, J. C., I. O. Orubuloye, and P. Caldwell. 1992. 145–58. “Fertility Decline in Africa: A New Type of Transition?” Magadi, M. A., and S. L. Curtis. 2003. “Trends and Determi- Population and Development Review 18 (2): 211–42. nants of Contraceptive Method Choice in Kenya.” Stud- Central Statistical Agency, and ORC Macro. 2006. Ethiopia ies in Family Planning 34 (3): 149–59. Demographic and Health Survey 2005. ORC Macro, Mauldin, W. P., and J. A. Ross. 1991. “Family Planning Pro- Calverton, MD. grams: Efforts and Results, 1982–89.” Studies in Family Cleland, J., S. Bernstein, A. Ezeh, A. Faundes, A. Glasier, and Planning 22 (6): 350–67. J. Innis. 2006. “Family Planning: The Unfinished May, J. F., M. Mukamanzi, and M. Vekemans. 1990. “Family Agenda.” The Lancet 368 (9549): 1810–27. Planning in Rwanda: Status and Prospects.” Studies in Cleland J., R. P. Ndugwa, and E. M. Zulu. 2011. “Family Family Planning 21 (1): 20–32. planning in sub-Saharan Africa: Progress or Stagna- Mbacke, C. 1994. “Review: Family Planning Programs and tion.” Bulletin of the World Health Organization 89: Fertility Transition in Sub-Saharan Africa.” Population 137–143. and Development Review 20 (1): 188–93. Cohen, B. 1998. “The Emerging Fertility Transition in Sub- Moultrie, T. A., and I. M. Timaeus. 2003. “The South African Saharan Africa.” World Development 26 (8): 1431–61. Fertility Decline: Evidence from Two Censuses and a Ezeh, A. C., B. U. Mberu, and J. Emina. 2009. “Stall in Fer- Demographic and Health Survey.” Population Studies tility Decline in Eastern African Countries: Regional 57 (3): 265–83. Analysis of Patterns, Determinants and Implications.” National Statistical Office Malawi, and ORC Macro. 2005. Philosophical Transactions of the Royal Society Biological Malawi Demographic and Health Survey 2004. ORC Sciencies 364 (1532): 2991–3007. Macro, Calverton, MD. Garenne, M., and V. Joseph. 2002. “The Timing of the Fer- Pathfinder International. 2005. Community-Based Family tility Transition in Sub-Saharan Africa.” World Develop- Planning in Kenya: Meeting New Challenges. Nairobi and ment 30 (10): 1835–43. Watertown, MA. Gillespie, D. G. 2004. “Whatever Happened to Family Phillips, J. F., W. L. Greene, and E. Jackson. 1999. Lessons Planning and for That Matter, Reproductive Health?” Learned from Community-Based Distribution of Family International Family Planning Perspectives 30 (1): Planning in Africa. No. 121, Population Council, 34–38. New York. Guengant, J. P., and J. M. May. 2002. Impact of the Proximate Solo, J. 2008. “Family Planning in Rwanda: How a Taboo Determinants on the Future Course of Fertility in Sub- Topic Became Priority Number One.” IntraHealth Inter- Saharan Africa: Prospects of Fertility Decline in High Fer- national, Chapel Hill, NC. tility Countries. New York: United Nations. Terceira, N., G. Simon, and S. Gregson. 2003. “The Hubacher, D., I. Mavranezouli, and E. McGinn. 2008. Contribution of HIV to Fertility Decline in Rural “Unintended Pregnancy in Sub-Saharan Africa: Mag- Zimbabwe, 1985–2000.” Population Studies 57 (2): nitude of the Problem and Potential Role of Contra- 149–64. ceptive Implants to Alleviate It.” Contraception 78 (1): Thomas, D., and I. Muvandi. 1994. “The Demographic 73–78. Transition in Southern Africa: Another Look at the 462 CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED Evidence from Botswana and Zimbabwe.” Demography Way, A. A., A. R. Cross, and S. Kumar. 1987. “Family Plan- 31 (2): 185–207. ning in Botswana, Kenya and Zimbabwe.” International UNDESA (United Nations Department of Economic and Family Planning Perspectives 13 (1): 7–11. Social Affairs). 2008. World Population Prospects, the World Bank. 2009. World Development Indicators. Washing- 2008 Revision. New York: United Nations. ton, DC: World Bank. United Nations Population Division. 2009. World Contra- Zinaga, A. F. 1992. “Development of the Zimbabwe Family ceptive Use. Nw YorK; United Nations Department of Planning Program.” Policy Research Working Paper, Economic and Social Affairs. World Bank, Washington, DC. CHAPTER 25: FAMILY PLANNING TRENDS IN SUB-SAHARAN AFRICA: PROGRESS, PROSPECTS, AND LESSONS LEARNED 463 CHAPTER 26 Achieving Universal Primary Education through School Fee Abolition: Some Policy Lessons from Uganda B. Essama-Nssah n 1997 the government of Uganda abolished school I In 1996 the government committed itself to providing fees for primary education. Immediately, primary UPE. It abolished school fees at the primary level starting in enrollment rates skyrocketed, with gross enrollment January 1997. The policy was not adopted overnight. It is an rates rising from 77 percent in 1996 to 137 in 1997 and net integral part of a broader reform program, which can be enrollment rising from 57 percent to 85 percent.1 Enroll- traced back to the creation in 1987 of the Education Policy ment remained high over the following decade, with poor Review Commission (see discussion later in chapter). children, girls, and rural residents benefiting dispropor- Education contributes directly to human development tionately from the increase in access. and is an investment in economic prosperity. In recogni- The quantitative success of the government’s universal tion of the fact that it has both intrinsic and instrumental primary education (UPE) initiative put significant stress on value for development, the government placed education at the country’s educational infrastructure, however, with a the center of its Poverty Eradication Action Plan (PEAP). consequent toll on the quality of primary schooling. The current 2010–15 five-year National Development Plan Increased enrollment led to overcrowding, multiple shifts, also recognizes the importance of education for sustained shortages of teachers and material, and an increase in over- economic growth and social transformation. age students.2 This chapter analyzes the outcomes of UPE in Uganda in order to draw policy lessons from the experience. OUTCOMES OF REFORM The goal of Uganda’s primary education policy is to ensure THE NEED FOR REFORM that every child enrolls at the appropriate age and success- fully completes the full cycle of education.3 Achievement of Uganda’s education sector achieved considerable progress fol- this objective requires expanding educational opportunities lowing independence, but that progress was reversed as a and improving educational outcomes. The review of the result of various sociopolitical crises in the 1970s and 1980s. results of reform presented in this section focuses on four In 1985 the level of government expenditure on education was outcomes: school enrollment and its cost, age at enrollment, about 27 percent that of the 1970s (Appleton 2001). Partly as completion rates, and equity considerations. a result of the high costs of education to families, the gross pri- mary enrollment rate stood at 50 percent in 1980—the same School enrollment rate as in 1960. Significant improvement was observed in 1985, when the gross primary enrollment rate increased to Between 1996 (the year of the policy announcement) and 73 percent, but it remained stagnant at that level until 1995. 1997 (the first year of implementation), primary school 465 enrollment in Uganda increased from about 3.1 million The surge in enrollment associated with the reduction in (a gross enrollment rate of 77 percent) to 5.3 million (a gross the cost of primary schooling is evidence that this cost was enrollment rate of 137 percent), a 58 percent increase. The a significant impediment to primary school enrollment by net enrollment rate rose from 57 percent in 1996 to 85 per- the poor (Deininger 2003). Indeed, lack of interest and the cent in 1997 and 90 percent in 1999 (MOES 1999). cost of enrollment are the two major determinants of To what extent are these changes in enrollment a result nonenrollment. of the UPE policy? The 1997 education policy reform can As a result of these reforms, the importance of cost as a be viewed as a natural experiment in which a sharp change reason for not enrolling in school declined significantly: in in government policy is an exogenous variation in a treat- 1992, 42 percent of households cited cost as a factor. This ment variable to which only a segment of the population figure fell to just 13 percent by 2005 (Essama-Nssah, Leite, (that is, primary school students) is exposed. Deininger and Simler 2008). This constraint did not change in urban (2003) estimates a linear probability model for primary areas, and it declined by less in the central region, where it school enrollment among three age groups: 6–12, 6–8, and fell from about 46 percent in 1992 to about 36 percent in 9–12 (see annex 26A for technical details). To control for 2005, than elsewhere in Uganda. Lack of interest, which unobserved effects of factors associated with a common increased from 49 percent in 1992 to 56 percent in 2005, economic environment, he estimates the same equation remains a severe constraint to enrollment. for secondary school enrollment (the 12–18 age group). To what extent have increases in enrollment been sus- All primary school regressions show a positive and signif- tained over time? Total enrollment increased steadily, from icant time trend, implying that the probability of primary 5.4 million in 2000 to about 7.4 million in 2010 (figure school enrollment increased for everyone in the sample. 26.1). The gross enrollment ratio dropped from about 128 When all other variables are set to their mean levels, the percent in 2000 to about 104 percent in 2004 (figure 26.2). probability of a child being enrolled in primary school in After a slight increase in 2005, it has been hovering around 1999 is about 60 percent higher than it was in 1992. 115 percent since 2006. After peaking at 100 percent in 2003, Deininger finds no significant effect of UPE on secondary the net enrollment ratio has shown an upward trend, rising school enrollment. Essama-Nssah, Leite, and Simler from 90 percent in 2004 to 96 percent in 2010. These trends (2008) replicate this analysis using data for 1992 and 2005. suggest that the sustained efforts of the government to They find similar results. improve education policy led to and maintained an increase Figure 26.1 Boys’ and Girls’ Enrollment in Public Primary Schools in Uganda, 2000–10 8 7 6 Millions 5 4 3 2 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Total Girls Boys Source: MOES 2010. 466 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION Figure 26.2 Gross and Net Primary Enrollment Rates in Uganda, 2000–10 140 130 120 Percent 110 100 90 80 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Gross primary enrollment rate Net primary enrollment rate Source: MOES 2010. Note: The gross primary enrollment rate (GER) is the ratio between the number of children enrolled in primary school and the number of children of pri- mary school age. The net primary enrollment rate (NER) is the ratio between the number of children of primary school age enrolled in primary school and the number of children of primary school age. See note 1 for more information. in enrollment not only for over-age but for under-age chil- that over time there has been a significant increase in school dren as well. fees paid by families at all levels of schooling. However, stu- Essama-Nssah, Leite, and Simler (2008) study the impact dents in public primary schools were paying less in 2005 than of the policy shift on the cost of enrollment using a linear they paid in 1992, suggesting that UPE may have reduced the regression model with interactive terms (analogous to the cost of enrolling in public primary schools. A comparison of linear probability model discussed above). The dependent the policy impact at the primary and the secondary levels sug- variable is the logarithm of school fees paid (as reported in gests that the policy shift may indeed have moderated the the 1992 and 2005 surveys). The specification accounts for increase in school fees at the primary level but that there was the fact that school fees are observed only for students who no spillover effect to the secondary level. Their results show are enrolled and not for those who are not enrolled in that in 2005, fees paid at the secondary level of education school. This potential selection bias calls for the application were higher in public than in private schools. UPE did lower of Heckman’s (1976) two-step estimator. This approach, the cost of primary education. which is analogous to instrumental variable estimation, requires simultaneous modeling of both exposure to policy Age at enrollment and completion rates intervention and the associated outcome. Identification therefore hinges on an exclusion restriction—that is, the Deininger (2003) finds that mother’s education is the most variation of at least one exogenous variable that affects par- important factor influencing timely enrollment (that is, ticipation (enrollment) but not the outcome (fees paid). enrollment between the ages of 6 and 8). It does not affect Essama-Nssah, Leite, and Simler apply this methodology children who enroll between the ages of 9 and 12. to both the primary and secondary levels of education, using Grogan (2009) uses data from the 1995 and 2000 Demo- community-level costs and peers’ enrollment rates in the graphic and Health Surveys (DHS) along with the 2001 identifying restriction. Thus at the household level, the con- DHS EdData survey for Uganda to assess the impact of UPE struction of these variables excludes the information pertain- on the likelihood of enrolling in school before the age of ing to the household under consideration. This construction nine—a key factor in preventing drop-out.4 She shows that ensures that the instrumental variables used for identification starting school before age nine is associated with a 16–26 are not directly related to the outcome variable. They find percent increase in the probability of completing at least CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 467 seven years of schooling. Using the regression discontinuity rates for boys consistently above those for girls. Improving framework from the program evaluation literature, she completion rates thus remains a challenge to policy makers. shows that elimination of school fees associated with the UPE policy had a positive albeit small effect on the propen- Equity sity to enroll in primary school before age nine (the overall effect is estimated at 3 percent). This finding is consistent The achievement of equity is increasingly seen as an impor- with the conclusion reached earlier, on the basis of the evo- tant goal of socioeconomic development. Identifying win- lution of enrollment rates, that UPE led to an increase in the ners and losers is therefore an important concern for policy enrollment of under-age children. makers. Doing so depends fundamentally on the response Essama-Nssah, Leite, and Simler (2008) analyze the of socioeconomic agents to the incentives of the policy impact of the policy shift on primary school completion. In under consideration. particular, they focus on the probability of completing the The distributional implications of an intervention seventh year of primary school (P7), the last grade of pri- depend on the variation in the impacts across groups or mary school in Uganda. As in the case of enrollment, house- individuals. Such impact heterogeneity stems from both hold income, age, parental education, and urban residence treatment heterogeneity (if the dose received depends on are key determinants of the likelihood of completing pri- some unit attributes) and heterogeneity in individual cir- mary school. These factors have a positive and significant cumstances (which determine the response to treatment). effect on the probability of completing P7. Essama-Nssah, The same heterogeneity that must be controlled for in iden- Leite, and Simler find that the policy shift may not have tifying causal effects must be accounted for in identifying improved the chances of completing P7. They also find that winners and losers. Models that include interaction between the income constraint on the chances of completing P7 has the treatment indicator and covariates offer a framework for become more severe. the analysis of systematic variation in mean impacts across The overall completion rate fell continuously between socioeconomic groups. Interactive models focus on differ- 2004 and 2007, from 62 percent to 47 percent (figure 26.3). ences in subgroup means. This trend seems to support the results of the analysis based The regression framework used by Deininger (2003) is on the 2005 household survey discussed earlier. The com- suitable for the analysis of the distributional implications of pletion rates for girls and boys follow the same pattern, with the UPE policy. In 1992, before the introduction of UPE, Figure 26.3 Boys’ and Girls’ Primary School Completion Rates in Uganda, 2001–10 75 70 65 60 Percent 55 50 5 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Overall Boys Girls Source: MOES 2010. 468 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION there was a strong bias against girls in Uganda, and parental The achievement gap between boys and girls narrowed as income had a strong impact on the probability of enrollment. well. In 2001 the primary school completion rate was 71 The probability of enrollment, which increases with parental percent for boys and 63 percent for girls. The two groups income, was 5 percent higher for boys than for girls. The rela- almost achieved parity in 2008, when the completion rate tionship between parental education and the probability of was 51 percent for boys and 49 percent for girls. For 2010 primary school enrollment was also positive and significant. the gap is estimated at 6 percentage points. Fixing all variables at their mean level and increasing the Convergence in achievement by boys and girls is evident father’s education by one year would lead to an increase in the from figure 26.4, which shows the success rate on the Pri- probability of enrollment of about 3 percent; an additional mary School Leaving Examination over the past decade. In year of education by the mother would increase the probabil- 2000, 90 percent of male candidates but only 63 percent of ity of her child enrolling in school by 4 percent. The baseline female candidates passed the examination. In 2010, 92 per- results show significant regional disparities. cent of boys and 90 percent of girls who take the exam were Interacting the policy variable with a variety of house- expected to pass. hold characteristics, Deininger (2003) shows that UPE has Another dimension of gender inequality deserves special been pro-poor. The positive impact of parental income on attention. The 2010–15 five- year national development the probability of enrollment is significantly lower after plan notes that the largest proportion of out-of-school chil- reform. Comparing 1992 and 2005, Essama-Nssah, Leite, dren are girls. Girls are also more likely than boys to drop and Simler (2008) find a significant increase in enrollment out of school or repeat grades. of the poor. At the bottom quintile, enrollment increased by The World Development Report 2006 argues for the pur- more than 28 percentage points (from 50.2 percent in 1992 suit of equity on both intrinsic and instrumental grounds. It to 78.8 percent in 2005). However, the gap between the bot- defines equity in terms of a level playing field on which indi- tom and the top quintile does not seem to have narrowed. viduals have equal opportunity to freely pursue chosen life The gap between boys and girls that was evident in 1992 dis- plans and are spared from extreme deprivation in outcomes. appeared in 2005, with the gap in the gross enrollment rate The equitable distribution of educational resources is one of for girls (23.5 percent) somewhat higher than that for boys the best ways to try to equalize opportunity across socioe- (20.4 percent). These results confirm the trends observed by conomic groups. Observed inequality of outcomes among Deininger (2003). groups defined on the basis of circumstances beyond their Figure 26.4 Boys’ and Girls’ Success Rates on Uganda’s Primary School Leaving Examination, 2000–10 95 90 85 80 Percent 75 70 65 60 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Overall Boys Girls Source: MOES 2010. CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 469 control is one potential indicator that current policies have Vision and political commitment not fully achieved equality of opportunity. A clear strategic vision focused on the role of education in Regional differences in resource allocations are wide in poverty reduction as well as strong political commitment Uganda (figure 26.5). Of the country’s four regions, the at various levels of government backed by a sound policy eastern and the northern have the highest pupil-teacher and framework underpinned the “big bang” approach adopted pupil-classroom ratios. These data suggest that children in in Uganda and contributed to its success. In 1996–97 the these regions do not have the same learning opportunities top political leadership in Uganda showed its commitment as their counterparts in the central and western regions. to poverty reduction by spearheading the formulation of the Poverty Eradication Plan. Uganda was among the first POLICY FRAMEWORK low-income countries to prepare a comprehensive and participatory national strategy for poverty reduction; its The abolition of primary school fees and the associated experience inspired the design of the Poverty Eradication measures adopted in the context of the UPE reform led to a Plan in Uganda and the Poverty Reduction Strategy Pro- significant and lasting expansion of educational opportuni- gram (PRSP) approach (Mackinnon and Reinikka 2000). ties and an improvement in equity in Uganda. The policy The government placed education at the center of this process that underpins these achievements has been held up action plan, in recognition that education has both intrin- as a leading example of the crucial role played by country sic and instrumental value for development. To show its ownership and donor cooperation within a sectorwide determination, the government quickly translated the approach to policy reforms. action plan into a budget and medium-term expenditure The success or failure of a policy is usually assessed in framework. terms of its objectives. But to understand the observed out- comes one needs to consider the policy-making process. The UPE reform sought to expand educational opportuni- Nature and scope of education policy reform ties and improve teaching and learning outcomes. There is The World Development Report 2004 (World Bank 2003) evidence that the gains in access and equity have not been argues that the effectiveness of social services in developing fully matched by improvement in educational outcomes. human capital depends fundamentally on the method of This section considers the policy framework in order to delivery and the behavior of key actors, including policy identify key determinants of these outcomes and draw pol- makers, service providers, and potential beneficiaries.5 Out- icy lessons from the Ugandan experience. comes are thus jointly determined by supply- and demand- side factors and the interactions among them. Supply-side interventions seek to increase the level and quality of ser- Figure 26.5 Pupil-Classroom and Pupil-Teacher Ratios in vices provided. Such interventions usually entail building Uganda, by Region, 2008 and staffing facilities, providing inputs, implementing insti- tutional reforms, and strengthening the incentives facing 100 services providers. On the demand side, typical interven- tions seek to improve the ability of participants to benefit 90 from the service provided. Some interventions include 80 incentives for potential beneficiaries to provide the requisite level of effort to achieve the desired outcome. One way of Percent 70 stimulating demand for services is to make them more 60 affordable. Other incentive-based interventions make 50 resource transfers conditional on some desired behavior. For example, a conditional transfer program may seek to 40 promote human capital accumulation by making cash or in- 30 kind transfers to poor families provided that their school- Central East North West National age children stay in school and young children and pregnant Pupil-classroom ratio Pupil-teacher ratio and nursing women participate in some health-enhancing Source: MOES 2008. activities. 470 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION The education policy reform in Uganda is broadly consis- facilities grants. The purpose of the capitation grant is to tent with this framework. The sudden increase in the demand shift some of the burden of school fees from parents to the for public primary education created a series of challenges government and to provide schools with some of the for the government, related mainly to financing the reform, resources necessary to run the school and support teaching improving the quality of primary education, ensuring and learning.6 The capitation grant is meant to cover tuition equitable access to primary education, and planning beyond only; families remain responsible for writing materials, uni- primary school. To improve the quality of primary educa- forms, and lunches. This grant is paid to schools in nine tion, the government has focused on five areas: curriculum monthly installments, at an annual rate of U Sh 5,000 development, provision of basic learning materials, teacher (about $3.00) per pupil in P1–P3 and U Sh 8,100 (about development, language of instruction for lower primary $4.75) per pupil in P4–P7 (Penny and others 2008).7 pupils, and establishment and maintenance of standards. The school facilities grant is designed to assist schools in It took some time for reform in these areas to signifi- the neediest communities in building classrooms, latrines, cantly improve the quality of teaching and learning in and teachers’ houses and procuring furniture. After a favor- Uganda. Test results from the National Assessment of able review in 1999, it became the only mechanism through Progress in Education between 1996 and 2000 showed which public funds are channeled for the construction of deterioration in student performance in math, reading, sci- school facilities. Both grants are conditional, to the extent ence, and social studies (Bategeka 2005). The key factors that funding is given to districts or municipalities under explaining the decline outcome were lack of coherence and strict guidelines and regulations and under the supervision consistency within the system and changes in teaching and of the Ministry of Education and Sports (MOES).8 learning methods (Penny and others 2008). Over time the situation has improved, for both boys and Planning girls. Between 2003 and 2010, the overall numeracy rate for students in P3 rose from 42 percent to 72 percent, with sim- The success of a policy initiative like Uganda’s UPE requires ilar increases for girls and boys (figure 26.6). These gains background analytical work, identification of financing were more dramatic than were gains in literacy (figure 26.7). sources, and the development of implementation and Uganda uses two types of grants to increase equitable monitoring capacity. Analytical work could take the form access to primary education: capitation grants and school of an overall assessment of the performance of the current Figure 26.6 Numeracy Rates for Boys and Girls in P3 in Uganda, 2003–10 75 70 65 60 Percent 55 50 45 40 35 2003 2004 2005 2006 2007 2008 2009 2010 Overall Boys Girls Source: MOES 2010. CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 471 Figure 26.7 Literacy Rates for Boys and Girls in P3 in Uganda, 2003–10 60 55 50 Percent 45 40 35 30 2003 2004 2005 2006 2007 2008 2009 2010 Overall Boys Girls Source: MOES 2010. system. Such an assessment would identify constraints that schools. One school in each cluster serves as the coordi- must be addressed and consider the feasibility and desir- nating center school, in the sense that it had a trained ability of a variety of policy options. In an environment in teacher (the outreach tutor) in charge of helping parents, which parents value education, as is the case in Uganda, community leaders, teachers, and principals in each school abolition of school fees will normally lead to a surge in in the cluster coordinate their efforts for improving teach- enrollment. It is therefore important to assess the expected ing and learning outcomes. Outreach tutors are supervised increase in demand for schooling and plan for the addi- and supported by outreach administrators from core pri- tional resources (teachers, facilities, and teaching and learn- mary teacher colleges. By 1998, 18 core primary teacher ing material) needed to cope with the surge. colleges were supervising about 550 outreach tutors The UPE reform in Uganda followed the “big bang” (MOES 1999). approach to policy reform, but adoption of the reform fol- In 1994 the National Curriculum Development Center lowed a long gestation period, and it benefited from many lost its monopoly on the production of instructional mate- critical prior actions, starting with the creation of the Edu- rials. The role of the Instructional Material Unit is to help cation Policy Review Commission in 1987. In its 1989 districts acquire instructional materials (including books) report to the government, the commission set an ambitious and to ensure that the process is consistent with government goal for primary education, recommending that education regulation designed to ensure transparency. policy ensure that every child enroll in school at the appro- In 1995 a monitoring and evaluation program, the priate age and successfully complete the full cycle (P1–P7) National Assessment of Progress in Education was set up (Grogan 2009).9 The first response to this recommendation to attempt to determine the amount of and type of knowl- came in 1993, in the form of the Primary Education and edge acquired by pupils relative to the objectives of the Teacher Development Project, supported by the World Bank curriculum. and the U.S. Agency for International Development (USAID). The project, which sought to introduce a coun- Partnerships trywide evaluation framework for overall progress in educa- tion, led to the establishment of a Teacher Development and The outcome of a policy intervention hinges critically on Management System and the creation of an Instructional individual behavior and on the rules that govern interaction Material Unit. Under the Teacher Development and Man- among stakeholders. In the context of the provision of pub- agement System, all public schools (also known as govern- lic services, the World Development Report 2004 argues that ment-aided schools) were organized into clusters of 18 there is a need for a governance framework that defines 472 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION accountability for results between policy makers and ply and demand constraints to service delivery are providers, between policy makers and beneficiaries, and addressed within a common sector policy and planning between providers and beneficiaries. The UPE reform was framework supported by well-functioning mechanisms for embedded in Uganda’s national Poverty Eradication Plan, the flow of funds and information. which is consistent with the key principles underlying the The same cannot be said of the partnership between the PRSP approach of accountability based on partnership. In central government and domestic stakeholders. Higgins the context of education reform, now a strategic component and Rwanyange (2005) present results from a qualitative of the National Development Plan, the central government assessment of the education reform process in Uganda is engaged in a two-tier partnership involving development focusing on ownership and accountability in the sector. partners and domestic stakeholders. Based on data collected between 1997 and 2004, the study The government of Uganda and its development part- finds that officials at the district level acknowledge ners adopted a sectorwide approach to deal with resource improvements in the transfer of funds but feel that they mobilization and aid coordination. The common vision lack the autonomy to adjust the use of funds to the partic- binding this partnership stems from both the adoption of ular needs of the districts and that the process is designed the PRSP approach toward the end of the 1990s and the to cater to the needs of the central government and the principles voiced at the World Education Forum held in donor community. Dakar in 2000. The government agreed with its develop- At the school level, according to the same study, parents ment partners that 31 percent of the recurrent discretionary and teachers are excluded from the decision-making process budget would go to education and that at least 65 percent of for the utilization of funds allocated to schools. Some that amount would be allocated to primary education (Hig- observers consider this lack of involvement and alienation a gins and Rwanyange 2005). The World Education Forum serious constraint to the performance of schools and stu- led to a commitment by the international community to dents (Higgins and Rwanyage 2005). In the survey underly- support any developing country that seriously engaged in ing the study, respondents cited the lack of integrity of head the pursuit of policies aimed at securing quality education teachers among the factors that impede effective learning for all (Eilor 2004). and school management. Some head teachers reportedly fail The sectorwide approach started with the Education to post capitation grants on public notice boards, as Strategic Investment Plan (ESIP) 1998–2003. At the end required by the rules. of the ESIP funding agreement, the Medium-Term Budget All stakeholders are concerned about the quality of Framework became the common budget support modal- entrants to primary teachers’ colleges and the training they ity (Penny and others 2008). The sectorwide approach receive there. Teachers themselves feel marginalized, emphasized actions designed to enhance local leadership because they have little say in decisions that affect their and the integration of development partner and govern- working conditions. These considerations suggest that ment efforts. Budget support was thus conditioned on a domestic partnerships for education in Uganda may not be few key outputs, and outcomes related fiduciary integrity as effective as they could be, because of imbalance in influ- and progress toward improving service delivery and equi- ence and capacity. Education reform in Uganda continues to table access. The Heavily Indebted Poor Country Initia- evolve, however, and coordination mechanisms have been tive led to an increase in funds available to the education improving. Starting in 2007, for instance, the participation sector. The effectiveness of this institutional arrangement in sector reviews and discussions of the Education Sector was significantly enhanced by the use of sector reviews Consultative Committee have been opened up to represen- and budget working groups. These mechanisms, along tatives of teachers and head teachers, local authorities, and with an effective and reliable Education Management parent teacher associations.10 Information System, are now playing a fundamental role in the education planning and budgeting processes CONCLUSION (Penny and others 2008). The partnership between the government of Uganda and Uganda’s success in increasing access and equity in primary the aid community is often cited as a success story. This education stems from the following institutional factors: partnership is based on a common vision about the effec- tiveness of poverty reduction strategies. It pursues a holistic ■ Strong political commitment to a development strategy approach, putting the country in the driver’s seat. Both sup- centered on building human capital CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 473 ■ A comprehensive approach to policy making based on care- Reform initially put enormous stress on the country’s ful planning and implementation of critical prior actions educational infrastructure, reducing the quality of educa- ■ Effective and sustained domestic and international part- tion as well as completion rates. Significant improvements nerships supportive of country ownership and donor have been achieved over the past few years. Policymakers in cooperation within a sectorwide approach Uganda must now consolidate and expand these improve- ■ Efficiency gains from measures designed to improve ments if they are to successfully link the country to the transparency and accountability at the school level in the global economy, which pays a premium for knowledge and use of available resources. skills. 474 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION ANNEX 26A DIFFERENCE-IN-DIFFERENCES The interaction terms in this equation play a crucial role in IDENTIFICATION OF POLICY the assessment and interpretation of various effects of the shift IMPACT to the UPE regime. One can think of the dummy variable T (representing the shift in policy regime) as having a moderat- When repeated cross-section data are available, the impact ing influence on the effects of the individual and household of a policy can be identified and estimated using a two-step characteristics on the schooling outcome (here the probability procedure known as difference-in-differences (DID) or of enrollment). The dummy variable T is a moderator vari- double difference. The first stage involves the reflexive (or able.13 The coefficient (g) associated with the moderator vari- before and after) change in average outcomes of the treat- able represents the effect of any secular trend captured by T. ment and control groups. This difference eliminates the The other coefficients in the model are easily interpreted selection bias associated with permanent group differences, within the logic of difference-in-difference estimation. leaving intact the bias as a result of the time trend. The bias that remains after the first step is removed by the difference in the second stage, in which the change in the average out- NOTES come of the comparison group is subtracted from that 1. The net primary enrollment ratio is the ratio between of the treated group.11 Alternatively, the treatment effect the number of children of primary school age enrolled in can be obtained by measuring the difference in the change primary school and the number of children of primary in average outcomes across groups (that is, comparing school age. By definition, this ratio cannot exceed 100 per- with- and without-treatment groups) before and after cent. The gross primary enrollment ratio is the ratio treatment. The DID method can thus be said to be path between the number of children enrolled in primary school independent, because the same result is achieved regardless and the number of children of primary school age. It can of the sequence in which these two differences are com- exceed 100 percent if children above (or below) primary school age are in primary school (because, for example, of puted. DID estimation can be implemented within a repetition or delayed entry). regression framework, which also offers a convenient way 2. The pupil-teacher ratio rose from 38 pupils per teacher to obtain the relevant estimates and the associated standard in 1996 to about 52 in 1997. The situation has improved errors. This framework also makes it easier to add more since 2000, with the pupil-teacher ratio falling steadily from groups and time periods using dummy variables (Angrist 65 in 2000 to an estimated 47 in 2010, the pupil-classroom and Pischke 2009).12 ratio falling from 106 in 2000 to an estimated 66 in 2010, Deininger (2003) applies a regression model that is con- and the proportion of untrained teachers declining contin- sistent with the double difference approach to two nation- uously (MOES 2010). ally representative household surveys: the 1992 Uganda 3. This section relies heavily on earlier work by the author Integrated Household Survey (UIHS) and the 1999 Uganda and two of his colleagues (Essama-Nssah, Leite, and Simler National Household Survey (UNHS). Given that UPE took 2008). effect in 1997, observations for 1992 represent the base case; 4. The 2001 DHS EdData survey collected information on schooling outcomes observed in 1999 presumably reflect the age at which children started and finished schooling, edu- continued implementation of UPE. cational attainment, reasons for nonenrollment, the extent to To see clearly what is involved, let Sit stand for an indica- which parents and guardians were aware of the UPE program. tor of enrollment status (enrolled or not enrolled) of child i The survey also contains information about the assessment by in year t = 1992 or 1999. Think of this outcome as a func- parents and guardians of the quality of local schools. tion of individual, household, and community characteris- 5. Such methods include central government provision, tics (Xit) such as gender, income, and parental education. contracting out to the private or nongovernmental This indicator also depends on a time trend (T) and unob- sector, decentralization to local governments, community participation, and direct transfers to households (World servables (eit) assumed to be independently and identically Bank 2003). distributed. The time variable T is a dummy variable that is 6. Parents paid up to 90 percent of recurrent and capital equal to one for 1999 (to mark exposure to UPE regime) expenditure for primary education in 1991 (Oketch and and zero for 1992 (indicating no exposure). Deininger’s Rolleston 2007). (2003) regression is of the following form: 7. For students in P7, the government also pays the regis- tration fees for the National Examination Board (Bategeka Sit = a + bXit + g T + dXitT + eit . 2005). CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 475 8. For instance, guidelines from the MOES require pri- Avenstrup, Roger. 2006. “Reducing Poverty through Free mary schools to spend capitation grant as follows: 50 per- Primary Education: Learning from the Experiences of cent on instructional material, 30 percent on co-curricular Kenya, Lesotho, Malawi, and Uganda.” In Attacking activities such as sports and clubs, 15 percent on utilities Africa’s Poverty: Experience from the Ground, ed. Louise and maintenance, and 5 percent on school administration Fox and Robert Liebenthal. Washington, DC: World (Bategeka 2005). Bank. 9. According to the commission, primary school comple- Avenstrup, Roger, Liang Xiaoyan, and Nellemann Soren. tion is the minimum level of education that all citizens need 2004. “Kenya, Lesotho, Malawi and Uganda: Universal to be able to live a full life. It noted that society would ben- Primary Education and Poverty Reduction.” Paper pre- efit a great deal from this outcome, which would probably sented at the conference “Scaling Up Poverty Reduction: increase national unity, moral standards, and prosperity A Global Learning Process and Conference,” Shanghai, (MOES 1999). May 25–27. 10. Communication with Luis Benveniste, World Bank, May Baron, Reuben M., and David A. Kennedy. 1986. “The 24, 2010. Moderator-Mediator Variable Distinction in Social Psy- 11. The first stage relies on the assumption of time invari- chological Research: Conceptual, Strategic and Statistical ance of the group effects; the second is justified by assuming Considerations.” Journal of Personality and Social Psychol- that, in the absence of treatment, the average outcome of the ogy 51 (6): 1173–82. treated would follow the same time path as the average out- Bategeka, Lawrence. 2005. Universal Primary Education come of the control group. When combined with the fact (UPE) in Uganda: Report to the Inter-Regional Inequality that both groups are observed before treatment, the Facility. Institute of Development Studies, University of assumption of common time trend makes it possible to Sussex, UK. frame the DID method within the logic of the traditional Deininger, Klaus. 2003. “Does Cost of Schooling Affect counterfactual approach. The treatment effect can thus be Enrollment by the Poor? Universal Primary Education in identified and measured by comparing the observed out- Uganda.” Economics of Education Review 22: 291–305. come of the treated with the counterfactual predicted on the Devarajan, Shantayanan, and Ritva Reinikka. 2004. “Making basis of the evolution of the average outcome of the control Services Work for Poor People.” Journal of African group. Economies 13 (1): i142–i166. 12. The double difference approach (in which impact is Dixit, Avinash, K. 1996. The Making of Economic Policy: A defined by the difference in the change of outcomes for the Transaction-Cost Politics Perspective. Cambridge, MA: treated and untreated group) can be implemented by a lin- MIT Press. ear regression that includes a constant, a group dummy Eilor, Joseph. 2004. Education and Sector-Wide Approach in variable, a time dummy, and an interaction term between Uganda. International Institute for Educational Plan- the two dummies. Impact is measured by the coefficient ning, UNESCO, Paris. associated with this interaction term. Essama-Nssah, B., Phillippe G. Leite, and Kenneth R. Sim- 13. By definition, a moderator is a qualitative or quantita- ler. 2008. “Achieving Universal Primary and Secondary tive variable that affects the direction or strength of the rela- Education in Uganda Access and Equity Considerations.” tionship between a response variable and a predictor or World Bank, Poverty Reduction and Equity Group, independent variable (Baron and Kennedy 1986). For Washington, DC. instance, the shift in policy regime (the moderator) may affect the impact of gender (a predictor) on schooling out- Government of Uganda. 2010. National Development Plan, comes (response variables). 2010/11–2014/15. Kampala. Grogan, Louise. 2006. “Who Benefits from Universal Pri- mary Education in Uganda.” Department of Economics, University of Guelph, Ontario. REFERENCES ———. 2009. “Universal Primary Education and School Angrist, Joshua D., and Jörn-Steffen Pischke. 2009. Mostly Entry in Uganda.” Journal of African Economies 18 (2): Harmless Econometrics: An Empiricist’s Companion. 183–211. Princeton, NJ: Princeton University Press. Heckman, James J. 1976. “The Common Structure of Statis- Appleton, Simon. 2001. “What Can We Expect from Univer- tical Models of Truncation, Sample Selection and Lim- sal Primary Education?” In Uganda’s Recovery: The Role ited Dependent Variables and a Simple Estimator for of Farms, Firms and Government, ed. Ritva Reinikka and such Models.” Annals of Economic and Social Measure- Paul Collier. Washington, DC: World Bank. ment 5 (4): 475–92. 476 CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION Higgins, Liz, and Rosemary Rwanyange. 2005. “Ownership 10, Consortium for Research on Educational Access, in the Education Reform Process in Uganda.” Compare: A Transitions and Equity, Institute of Education, University Journal of Comparative and International Education 35 of London. (1): 7–26. Penny, Alan, Michael Ward, Tony Read, and Hazel Bines. Ingram, Gregory K. 2004. “Overview.” In Evaluation and 2008. “Education Sector Reform: The Uganda Experi- Development: The Partnership Dimension, ed. Andres ence.” International Journal of Educational Development Liebenthal, Osvaldo N. Feinstein, and Gregory K. 28: 268–85. Ingram. New Brunswick, NJ: Transactions Publishers. Picciotto, Robert. 2004. “The Logic of Partnership.” In Mackinnon, John, and Ritva Reinikka. 2000. “Lessons from Evaluation and Development: The Partnership Dimen- Uganda on Strategies to Fight Poverty.” Policy Research sion, ed. Andres Liebenthal, Osvalso N. Feinstein, and Working Paper 2440, World Bank, Washington, DC. Gregory K. Ingram. New Brunswick, NJ: Transactions MOES (Ministry of Education and Sports, Uganda). 1999. Publishers. The Ugandan Experience of Universal Primary Education. Reinikka, Ritva, and Jakob Svensson. 2004. “Local Capture: Kampala. Evidence from a Central Government Transfer Program ———. 2010. The Education and Sports Annual Performance in Uganda.” Quarterly Journal of Economics 119 (2): Report. ESAPR. for the Financial Year 2008/09. Kampala. 679–705. Nelson, Jane, and Simon Zadek. 2000. Partnership Alchemy: ———. 2007. “The Returns from Reducing Corruption: New Social Partnerships in Europe. Copenhagen Centre, Evidence from Education in Uganda.” CEPR Discussion Copenhagen. Paper 6363, Center for Economic Policy Research, Wash- Nishimura, Mikiko, Yamano Takashi, and Sasaoka Yuichi. ington, DC. 2008. “Impacts of the Universal Primary Education Pol- World Bank. 2003. World Development Report 2004: Making icy on Educational Attainment and Private Costs in Services Work for Poor People. Washington, DC: World Rural Uganda.” International Journal of Educational Bank. Development 28: 161–75. World Bank, and UNICEF (United Nations Children’s Oketch, Moses O., and M. Rolleston Caine M. 2007. “Poli- Fund). 2009. Abolishing School Fees in Africa: Lessons from cies on Free Primary and Secondary Education in East Ethiopia, Ghana, Kenya, Malawi and Mozambique. Wash- Africa: A Review of the Literature.” Research Monograph ington, DC: World Bank. CHAPTER 26: ACHIEVING UNIVERSAL PRIMARY EDUCATION THROUGH SCHOOL FEE ABOLITION 477 Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The Office of the Publisher has chosen to print • 12 trees Yes Africa Can: Success Stories from a Dynamic • 369 pound of solid waste Continent on recycled paper including 50% • 5,820 gallons of water post-consumer recycled fiber in accordance • 1,290 lbs Co2 equivalent with the recommended standards for paper of net greenhouse gases usage set by the Green Press Initiative, a • 5 million BTU’s of total nonprofit program supporting publishers in energy using fiber that is not sourced from endan- gered forests. For more information, visit www.greenpressinitiative.org. rawing on the existing knowledge of African development from previous D publications Yes Africa Can takes an in-depth look at 26 economic and social development successes in Sub-Saharan African countries—20 from individual countries and 6 that cut across the region. These stories manifest at the project, provincial, sub-national, national, or regional level and across themes, programs, and sectors. The book aims to address how Sub-Saharan African countries have overcome major development challenges. The main components of each case study include a description of the achievement and the elements that qualify the outcome as successful; an assessment of the main policies, interventions, actions, and other factors that contributed to the positive outcome; a presentation of the lessons learned and the contribution to the discourse on African development; and insights on the usability or applicability of the achievement in terms of the potential for scaling up the interventions and actions. The studies can be classified into four categories: overcoming or avoiding massive government failure, rebuilding or creating a government, rationalizing government involvement in markets, and listening to the people. ISBN 978-0-8213-8745-0 SKU 18745