Report No. 37860-RW Rwanda Toward Sustained Growth and Competitiveness (In Two Volumes) Volume I: Synthesis and Priority Measures October 12, 2007 Poverty Reduction and Economic Management 3 AFCC2 Africa Region Document of the World Bank GoR Government of Rwanda GACACA Specially ElectedTribunals GNFS Goods and NonFactor Services GNI Gross National Investment GNP Gross National Product GNFS Goods and NonFactor Services GoR Government o f Rwanda Hv HighValue HI Herfindahl Index PDL HIM0 Labor Intensive LocalDevelopment Program ICT Information Communication. Telecommunication IDA InternationalDevelopment. Association. IFMIS Integrated FinancialManagement System IMF InternationalMonetary Fund ISAR National Agricultural ResearchInstitute ISDN Integrated Services DigitalNetwork ISPS Internet Service Providers LT Low Technology M 2 Narrow Money MBB Marginal Budgetingfor Bottlenecksmodel MDG MillenniumDevelopmentGoals MDRI Multi-Lateral DebtReliefInitiative METR Marginal Effective Tax Rate MINAGRI MinistryofAgriculture MINECOFIN MinistryofFinance andEconomic Planning MINEDUC MinistryofEducation MINICOM MinistryofCommerce MINISANTE MinistryofHealth MFI Microfinance Institutions MSSE Micro and Small Scale Enterprises MT MediumTechnology NBR NationalBank o f Rwanda NGO Non-GovernmentalOrganization NHA National Health Accounts NICI National Information and Communication Infrastructure NUR National Universityof Rwanda NTBs National Tender Board ODA Official Development Assistance OCIR-Cafk Office des Cultures Industrielles Rwandais-CafC OCIR-The Office des Cultures IndustriellesRwandais-ThC ORTPN Rwandan Office of Tourism and NationalParks ODA Official Development Assistance OCIR-CafC Office des Cultures Industrielles Rwandais-Cafe PEPFAR The President's Emergency Planfor Aids Relief PPPS Public-PrivatePartnerships PIL Partnership inLearning PPIAF Public-Private Infrastructure Advisory Facility PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper PTA Parent Teacher Association RE3 Resource Based 11 RER RealExchangeRate REER RealEffectiveExchangeRate RIA ResearchICTAfrica RITA RwandaInformationandTechnology Authority RIA ResearchICTAfrica R&D ResearchandDevelopment RER RuralEconomyReview(RealExchangeRate) R I M S RwandaIndustrialandMining Survey RJNEX RwandaInternet Exchange RPA RwandaPatrioticArmy RRA RwandaRevenueAuthority RURA RwandaUtility RegulationAgency SSA Sub-SaharanAfrica SOPYRWA InsecticideProcessingPlant SOWARTI-IE TeaAssociations STABEX ImportTenderingBoard TFP TotalFactorProductivity TOT Terms ofTrade UBPR Unionde BanquesPopulairede Rwanda UK-DFID UnitedKingdomDepartmentfor InternationalDevelopment UNDO UnitedNationsIndustrialDevelopmentOrg. UTEXIRWA TextileCompany VAT Value Added Tax VECM VectorErrorCorrectionModel VPP Village PayPhone VSAT Very SmallApertureTerminalSatellite WDI WorldDevelopmentIndicators ... 111 PREFACE This Country Economic Memorandum (CEM) for Rwanda provides the analytical basis for determiningwhich priority interventions are neededto achieve the Government's long-term goals for growth and poverty reduction, as articulated in the Government's Vision 2020. Those goals were to increaseper capita income from US$230 to US$900 and reduce the incidence o f poverty by 50 percent bythe year 2020.' Two years after those targets were established, Rwanda's first Poverty Reduction Strategy Paper (PRSP)? issued in2002, projected that GDP growth would needto be inthe range o f 6 to 7 percent over the long term for those targets to be realized. The principal sources of growth inthe medium-term were to be the primary and manufacturing sectors, with growth of the primary sector projected to start at 5.2 percent and accelerate over the period, due to improved soil productivity from increased use of agricultural inputs.Manufacturing growth was projected to rise sharply to 11.5 percent, based on the expansion o f manufacturing capacity in agro- processing, and then slow to a more sustainable level of 7 percent. To achieve this level of growth, the PRSP focused on six priority areas: (i)rural development and agricultural transformation; (ii)human development; (iii)economic infrastructure; (iv) good governance; (v) private sector development; and (vi) institutional capacity development. Overall economic growth has been strong over the past four years, in line with an upward trend over the past decade. Real GDP growth averaged 10 percent over the period 1995-2005, and average inflation was maintained around 10 percent. The poverty incidence declined from a level of 78 percent just after 1994, to 60 percent o f the population by 2000, and is currently estimated at 56.9 percent. With agriculture as the main source o f growth over the past decade, growth has been broad-based, in terms o f coverage o f the majority o f the population. Per capita real GDP rose from US$245 in 2000 to US$260 in 2005. In addition, net primary school enrollment has reached 95 percent. Under 5 mortality fell by 30 percent, and immunization coverage, at 90 percent, i s one o fthe highest inSub-Saharan Africa (SSA). Since 2002, average growth rates in manufacturing and agriculture, however, have been below the projected targets, and improvements inpoverty have been marginal, due to a number of factors: (i) lack of investment in infrastructure during the recovery and stabilization phase, to complement the reforms undertaken to improve the business environment; (ii) lack o f investments incapacity, institutions, and land/water management inthe agricultural sector; (iii) continued low use of inputs; (iv) instability within the region; (v) delays in Rwanda's accession to the East Africa Community (EAC); and (vi) a slower than expectedpace of reform in key sectors such as the tea industry. The slower than expected growth in these sectors has highlighted the need to examine sector-specific issues related to agriculture, industry, and services, in order to identify the most 1 * Government The Vision 2020 document was issuedby the Government ofRwandain2000. ofRwanda, 2002. bindingconstraints to growth and improved competitiveness. The examination of those issues in this report is particularly relevant at this time, as the Government is now poisedto enter the phase o f sustained shared growth, with increased emphasis on private sector investment. The analysis will inform the preparation o f the country's second-generation PRSP, appropriately titled the Economic Development and Poverty Reduction Strategy (EDPRS), which has recently been launched. Itwill also inform the preparation of a new Country Assistance Strategy (CAS), to be preparedjointly with DFIDinFY08. In addition to the CEM, the Bank has undertaken other analytical work on economic management and growth to inform the EDPRS and the CAS, some in collaboration with other development partners. These studies include a Diagnostic Trade Integration Study (DTIS), an Agricultural Policy Note, a Financial Sector Assessment Program (FSAP), a study on Private Solutions for Infrastructure in Rwanda: A Country Framework Report, a fiscal baseline study on infrastructure- Fiscal Costs of Infrastructure Provision inAfrica: Creating a Baseline - Rwanda, and an upcoming Investment Climate Assessment, which will be used in the formulation of a private sector development strategy. This Country Economic Memorandum consists o f two volumes. Volume 1 is targeted at policymakers and other development partners, and provides an overview of the main issues, with suggested interventions in priority areas. Volume 2 provides more detailed and technical coverage of the issues outlined inVolume 1. Volume 1begins with an overview, inChapter 1,of reforms and policy issues to date. Chapter 2 then discusses the management o f foreign aid revenues to achieve maximum impact. Chapter 3 examines the sources of growth and constraints in the agriculture sector, and Chapter 4 does the same for the non-agriculture sectors. Chapter 5 provides an overview of issues involved in export diversification and competitiveness. Chapter 6 concludes with an outline o f prioritized measures to sustain growth. w--4 Acting CountryO'ConnorDirector for Rwanda Ju M. Africa Region ACKNOWLEDGEMENTS This report and the accompanying Volume I1draw from a series of background papers prepared by a team of World Bank staff members and consultants, and staff from Rwanda's National Institute of Statistics and the National Bank. Much o f the work, (especially in support o f the National Institute o f Statistics' implementation o f the informal and manufacturing surveys), would not have been possible without the financial support received from the Belgian Trust Funds, the Poverty Reduction Strategy Trust Fund, and World Bank consultant trust funds from the Governments of the Netherlandsand Denmark. Volume Iand I1of this report were produced under the overall guidance of Yvonne Tsikata (Sector Manager, AFTP3) and Chukwuma Obidegwu (Lead Economist, AFTP3). Conceptualization o f the work benefited from the guidance o f Brendan Horton (Lead Economist, AFTP3), and Cadman Mills (former Sector Manager, AFTP3). The main author and task team leader o f the report is Kene Ezemenari (Senior Economist, AFTP3). The work draws heavily from a series o f background studies, prepared by the following team members as follows: Kalamogo Coulibaly (sources of growth and growth diagnostics) and with Rwanda Revenue Authority staff and Tembo Maburuki (cross-border trade); Ephraim Kebede (constraints to growth, fiscal effects of aid), and with Jean Musoni Rutayisire, from the National Bank o f Rwanda (sources o f inflation); Mireille Linjouom (impact of energy crisis on the economy); Professor Kigabo Rusuhuzwa Thoma of the Free University of Kigali with coordination support from Oscar Masabo (energy crisis and manufacturing firms); Mary Kamari (assisted in the study coordination and provided inputs on business environment and private sector development); Rahel Kassahun (tourism); Tilahun Temesgen with Jean Bosco Iyadema (design o f the RIMs), and with staff o fNational Institute of Statistics (analysis and drafting o f the R I M s and the the informal sector survey); Patrick Nugawela, with inputsfrom Viateur Bicali and Viateur Ndagijmana (micro- and small- enterprises), and with Prosper Mutijima (design of the informal sector survey); Jiro Tominaga (ICT), Ibrahim Elbadawi and Linda Kaltani (exchange rate alignment); Vandana Chandra with inputs from I.Osario, M. Desai, and Y. Li (export competitiveness); Arsene Nkama (domestic resource cost and competitiveness of the manufacturing sector); Stephan von Klaudy and Daniel Benitez (Infrastructure spending efficiency); Xinshen Diao and Bingxin Yu (multi-market model with focus on agriculture growth); and Michael Morris and Liz Drake, who made available the results of the background studies for the Agriculture Policy Note, and commented on drafts of the agriculture chapter. The report also benefited from discussions with Bank colleagues including Celestin Monga, and Paul Brenton, as well as comments and insight from the reviewers: Jeffrey Lewis, Vera Songwe, Mathew Verghis and Stefan0 Paternostro. It also greatly benefitedfrom a Quality Enhancement Review panel chaired by Edgardo Favaro (PREMED), and consisting o f panel members: Stephen Mink (EASRD), Magdi Amin (AFTPS), and Constantino Lluch (consultant); along with feedback from Susan Opper, and Lucy Fye. Comments and suggestions from outside the Bank are also gratefully acknowledged, in particular from KristinaKostial (IMF); as well as from participants (including government and other development partners), during an in-country validation workshop which took place on March 20, 2007. These also have been used to revise the report. Thanks are also due to Lucia Chuo and Paula White who processed the document, to Marie Uwanyarwaya, for logistical support, and Deborah Davis for editorial assistance. vi EXECUTIVE SUMMARY 1. After a decade of strong growth and successful reforms, Rwanda i s now poised to move into the next phase o f development with its second PRSP. However, the country faces a number o f challenges given that the rapid growth o f the past decade i s beginning to slow. First, resource mobilization i s low, although the country is well placed to receive scaled up aid. Second, the high cost o ftrade is compounded by the landlockedposition o f the country and the poor state o f infrastructure. Third, the country's high dependence on rain fed agriculture contributes to its vulnerability to terms of trade shocks. Fourth, there is a shortage o f skilled labor needed to support development of the private sector. The CEM analyzes the sources, opportunities and constraints to growth and comes to the following conclusions and strategic priorities. Inthe short to medium term, the focus should be on improving productivity in the agriculture sector. In the medium to long term, increased capital investment to reduce the costs o f energy, water, and transport and to increase the supply o f skilled labor are needed. 2. The CEMidentifies four Priority measures for sustained growth: Priority Measure 1:Invest to ease the in9astructure constraints to growth and export diversijkation; Priority Measure 2: Put inplace supportivepoliciesfor improvedspendingeficiency; Priority Measure 3: Transform the agriculture sector to be more market oriented; Priority Measure 4: Provide support and incentivesfor private sector development. Priority Measure 1: Invest to ease the infrastructureconstraintsto growth and export diversification 3. The C E M finds that the poor state o f infrastructure is a major constraint to growth. In particular, frequent power outages and unreliable supply o f electricity has had a significantly negative effect on the productivity o f both medium sized and large manufacturing firms. The low level o f electricity coupled with low investment contributes to low capacity utilization for manufacturing firms. For micro-enterprises, a smaller share o f firms report electricity as a constraint compared to the larger firms. However, for those micro and small enterprises which report electricity as a major constraint, the effects are more severe. Aside from electricity, the poor state o f unpaved roads limits the access o f rural households to the market. To address the problem o f high maintenance costs for roads, a systematic, planned approach to rehabilitation and road maintenance, with a focus on communal and non-classified roads and options for cost recovery is needed. Priority Measure2: Put in placesupportivepolicies for improvedspending efficiency 4. The macroeconomic management o f aid will needto focus on improving the composition and efficiency o f spending, to sustain growth and lay the foundations for transitioning away from an excessive dependence on aid. To sustain growth, greater attention and resources are needed for improved delivery o f economic and infrastructure services. Improved efficiency in the social sectors should help to consolidate and build on progress made in this area to date. In the case o f education, the focus should be on the e-examination o f relative subsidies between primary and secondary education relative to tertiary education. In the case o f health, there needs to be greater focus on strengthening community based health delivery systems. In the case o f infrastructure, vii reduction intechnical losses related to supply o f electricity and water, and rehabilitation o f rural roads i s needed. The private sector i s already engaged inthese sectors, and identifyingmeasures to strengthen and build on the involvement o f the private sector will service delivery in key sectors andcontribute to private sector development. PriorityMeasure3: Transformthe agriculturesector to be more market oriented 5. Sector specific policies to sustain growth and reduce poverty will need to focus first on transforming the agriculture sector to make it more market oriented. This requires investmentsto improve productivity, and the value added o f output. Measures to provide support to farmer organizations and ensure they are linked to existing and potentially new value chains will be needed. Inaddition, this will require strengthening the extension system, including increasing the numberof agronomists andveterinary services. However, developingnon-farm employmentwill be equally important for generating alternative income sources, particularly for households with extremely small landholdings. It would also facilitate the process of moving labor out of agriculture. Provided the needed investments are made to substantially increase the use o f improved inputs and water management, government can achieve the agriculture and overall growth and poverty reductiontargets. PriorityMeasure4: Providesupport and incentivesfor privatesector development 6. Development of the private sector therefore needs improvement in the business and investment climate through increasing the access o f small f m s to finance, infrastructure investment and skills development. The importance o f these constraints vary for different firms Access to finance poses a greater constraint for micro- and small enterprises which could be addressedthrough the development of financial products that are specifically geared to the needs of smaller firms. Investmentsin infrastructure would help to increase the scale of exports by easing supply constraints. The lack of skilled labor poses a constraint for all firms. Inthe case of manufacturing firms, it puts them at a less competitive position compared to neighboring countries. Inthe case of micro-enterprises, it limits their access to information and know-how for business development, and prevents them from becoming more formalized. Programs for the development o f a skilled labor force would include support to on-the-job training, development o f vocational training, and improvement in the overall education system to align with demands of the labor market. This will require a comprehensive strategy and action plan for skills development, inclose partnership with the private sector. ... V l l l TABLE OF CONTENTS 2. THE MACROECONOMICSOFAID MANAGEMENT .............................................................. 5 2.1 The Composition and Efficiency of Spendingto ImproveAbsorption.................... 8 2.1.1 Efficiency of SpendinginHealth.............................................................. 10 2.1.2 Efficiency of Spending inEducation........................................................ 13 2.1.3 Efficiency of Spending on Infrastructure.................................................. 14 2.2 Role of The Private Sector ..................................................................................... 17 2.3 Conclusionand Priority Measures ......................................................................... 19 3. POTENTIAL SOURCES OF GROWTHAND CONSTRAINTSTO 2020 ............................... 20 3.1 20 Recent Sources of Growth, Constraints, and Opportunities................................... Growth inAgriculture - Strong but insufficient to meet Demand......................... 3.2 21 3.3 FutureDeterminants of Growth, Opportunities, and Constraints inthe Sector Agriculture's PotentialContribution to Growth andPoverty Reduction...............28 (2007-2020) ............................................................................................................ 26 3.4 3.5 Balance of Trade inAgriculture............................................................................. 29 3.6 Sourceso f Growth for Poverty Reduction............................................................. 30 3.7 Income Inequality andProspects for Poverty Reduction....................................... 30 3.7.1 The Target GrowthRate of 6 percent inAgriculture is Feasible and Critical for Poverty Reduction.................................................................. 31 3.7.2 Reducing Poverty by Halfby 2015 Will Require 8 Percent Growth........31 4. GROWTHPROSPECTSAND CONSTRAINTSINNON-AGRICULTURE ........................... 33 4.1 Micro and Small-scale Enterprises ........................................................................ 33 4.2 Growth Prospects and Constraints inManufacturing ............................................ 37 4.3 Tourism and InformationCommunications Technology ....................................... 41 4.4 Conclusionand Priority measures.......................................................................... 43 5. EXPORTGROWTH AND DIVERSIFICATION ........................................................................ 45 5.1 Pattern of Export Discoveries -Many with Little Scaling-Up .............................. 46 5.2 Factors Constrainingthe Scaling up of Exports..................................................... 46 5.3 Sector-Specific Strategic Focus to Promote Diversification.................................. 48 5.4 Conclusions and Strategic Approach ..................................................................... 50 6. CONCLUSIONAND PRIORITY MEASURESFOR SUSTAINEDGROWTH ...................... 51 LIST OF BOXES Box 2.1: Overview of Economic and Structural Reforms inRwanda. ....................... 1961-1994 2 Box 2.1: Measuringthe Use o fMarginal IncreasesinAid Flows.................................................. 6 LISTOFFIGURES Figure2.1: Rwanda's Monetary Indicators. 1995-05..................................................................... 6 Figure2.2: Distribution o fDonor Fundingby Strategic Objective.............................................. 12 Figure 3.1: Average Increase in CultivatedArea andYield. 2000-2005 (percent) ...................... 22 Figure3.2: Trend inCultivatedLandArea andLandProductivity inRwanda............................ 22 Figure 3.3: GDP Growth Projections. 2005-15 ............................................................................ 29 i Figure 3.4: Poverty Reduction inthe MDG Scenario (AgGDP gr: 9.0%; NonagGDP gr: 7.2%; GDP gr: 8.0%) .................................................................................................................. 32 Figure4.1: InfrastructureIndicators: Electricity .......................................................................... 39 Figure 4.2: Infrastructure: Electricity and Transport as Major Concerns..................................... 39 LISTOFTABLES Table 2.1: Analysis o f Two Aid Surges for Rwanda...................................................................... 7 Table 2.2: Rwanda-Public Recurrent Spending on Priority Sectors. 1998-2005 (share o f GDP) 8 Table 2.3: Progress in Selected Social Indicators(Percent of relevant Population) ....................... 9 9 Table 2.5: Rwanda -On- and Off-budget Public Expenditures................................................... Table 2.4: InternationalBenchmarking of InfrastructureSpendingLevels ................................... 16 Table 2.6: InfrastructureQuasi-Fiscal Deficits inRwanda.......................................................... 16 Table 2.7: Maintenance SpendinginInfrastructure inRwanda (on and off budget) ................... 17 Table 3.2: Distributiono f LandHoldings Among RuralHouseholds inRwanda. 2006..............23 Table 3.1: Crop Yield Comparison, 1999-2003 (average) ........................................................... 21 Table 3.3: Use of ImprovedInputsby Province and Income Quintile. Rwanda.......................... 24 Table 4.1: Structure o fthe MSSE Sector, Classifiedby Rwandan Systemof Production...........25 Table 3.5: Improved Seeds Productionand DemandCoverage (%). 200-2005 ........................... Table 3.4: Fertilizer Responsefor SelectedCrops ....................................................................... 25 Table 4.3: Employment inSmall-scale MiningFirms, Regional Comparison ............................ Table 4.2: The Structure of GDP According to Degree o f Informality. 1995 .2003 ...................33 34 38 Table 4.4: Average Education. Experience, and Weekly Earnings inRwanda............................ 41 .. 11 1. SETTINGTHE STAGE-OVERVIEW OFREFORMS AND RESULTSTO DATE 1.1 Achievements in growth and macroeconomic stability have been impressive in the face of politicaland social challenges, and the shock that occurred in 1994. For a number of years leading up to the genocide, the Rwandan economy had been stagnant. The average annual growth rate between 1985 and 1993 was 0.8 percent. The 1994 genocide led to a 50 percent contraction in the economy (Box 1.1). It took 6 years o f growth, between 1994 and 2000, at an average annual rate o f 14 percent, for the economy to recover its pre-1994 level. Once the economy recovered, growth averaged 5.2 percent annually between 2001 and 2005, and inflation averagedjust over 6 percent. 1.2 The progress in growth after 1994 arose not just from the catch-up effect, as labor and capital increasingly became re-engaged in productive activities, but also from a continuationof reforms that had been initiated prior to the crisis. A flexible exchange rate was introduced in 1995. Tariff rates were reduced from 34.8 to 18 percent and the number of bands was reduced from 5 to 4, ranging from 0-30 percent, by 2003. Liberalization of the monetary and financial sectors led to the adoption of new currency exchange regulations, the creation o f private commercial banks, and the privatization o f state-owned banks.3 Current account operations (imports, exports, services) were liberalized, and certain restrictionson capital flows were eliminated. These included the transfer o f capital and revenues related to foreign direct investment (FDI), and free withdrawal on foreign exchange accounts incommercial banks.4 More recently, Rwanda has joined the East African Customs Union (COMESA) and the East African Community (EAC). 1.3 Intheareaofstructuralreforms, theRwandaUtilitiesRegulatoryAgency andaone-stop- shop center were established in the Rwanda Investment and Export Promotion Agency. In addition, an export promotion strategy and action plan has been recently elaborated. The Government has also reduced its role in the market by privatizing numerous state-owned enterprises, including two commercial banks (Banque Commerciale du Rwanda and BACAR), Rwandex (the largest coffee exporter), three tea factories, and Rwandatel (the telecommunication company). However, the privatization o f tea factories has progressed more slowly than planned; and less progress has been made in transforming agriculture and improving economic infrastructure inrural areas. 1.4 Substantial progress has also been made in the area of public finance and expendituremanagement. The MediumTerm Expenditure Framework (MTEF) was introduced in 2001, leading to improved budget predictability and a more rational approach to budget preparation, implementation, and monitoring. The new Organic Law of State Finances and Assets, which is expected to modernize the public financial management system, was adopted in 3Rwandahadthree commercial banksprior to 1995; since that time, the number of commercialbankshas doubled. The countryhas currently6 commercial banks. KanimbaFrancois, 2004. Deepeningthe financial market inRwanda. Power point presentation, September. 1 March 2006; and a Public Procurement Code is currently being reviewed in parliament. Despite the gains in institutional and legal reforms, however, capacity continues to be a major challenge that affects overall economic governance. Box 2.1: Overview of Economic and Structural ReformsinRwanda - 1961-1994 During the period 1961-1990, Rwanda had an administered economy, characterized by the imposition of severe restrictions on trade and foreign exchange transactions, as well as a fixed exchange rate. In the early 1990s, the average tariff rate was 34.8 percent, with 5 tariff bands ranging ffom 0-60 percent. Each imported product, as well as importer, was subject to a quota. Further, licenses to export were authorized only by the Banque Nationale du Rwanda (BNR), and the purchase o f foreign exchange was subject to a ceiling. This over-protection o f the economy adversely affected fm incentives and competition, which in turn prevented local companies from innovating and bringing about technological change. As a result, real GDP per capita grew by -0.29 percent per year, on average, during the 1980-89 period, as shown inthe table. Indicesof Trade Reforms 1980-89 1996-04 Indices of trade reform Real GDP per capitagrowth -0.29 2.00 Imp/GDP 21.00 25.00 Imp(cap)/GDP 3.90 6.50 Exp/GDP 10.00 8.00 Trade ratio 26.00 36.00 Trade-tax ratio 46.60 21.00 Indices offinancial reforms Real GDP per capita growth -0.29 2.00 M2/GDP 14.00 15.30 Priv.Cr./GDP 7.20 9.60 Real Int. 9.13 8.50 Indices of exchange rate reforms Real GDP per capita growth -0.29 2.00 Real Effective ExchangeRate 109.00 78.00 Type of exchangerate Fixed rate Flexiblerate Trade ratio: exports and importsofgoods and servicesover GDP. Trade-tax ratio: Internationaltrade tax over total revenue. Source: World BankDevelopmentIndicators,2005. The period 1991 until 1994 marks the beginning of substantial reductions in trade restrictions and foreign exchange transactions, and the gradual establishment of a market economy. While this period covers the fist phase o f trade reforms, it was also characterized by macroeconomic and political crisis in Rwanda, which culminated in the genocide o f 1994, leading inevitably to the destruction o f manpower' and capital stock, and a total absence o f the state. The genocide also brought uncertainty to trade reforms and unpredictability to the incentive structure. With uncertainty, the few businesses that survived were reluctant to incur the sunk cost o f adjustment caused by the f i s t wave o f trade reforms. The inability o f businesses to adjust, and the uncertainty o f the macroeconomic and political environment, contributed to the poor economic performance. Over the period 1991-1994, real GDP per capita contracted at the rate of 8 percent per year. With the end o f the crisis, the Government resumed its liberalization, leading to the strong growth observed over the next 10 years that followed. 1. It is estimatedthat 800.000 ueoulelost their livesbetweenAuril andJune 1994. See UN 1999. 1.5 The move toward an outcome-focused budget and MTEF has been accompanied by reforms in education and health. In the health sector, reforms have focused on piloting performance-based contracting mechanisms to improve incentives and engage the private sector. 2 Inthe case of education, there is increased involvement of communities; and capitation grants (i.e. transfers per student to school districts) have beenadopted. 1.6 Along with structural reforms, the Government initiated measures to promote reconciliation and peace, in recognition of the importance of these factors for growth. A system ofgacaca courts, modeledafter atraditional approachto settling disputes, was established as a means of engaging citizens inthe genocide trials. The current estimate is that 717,000 people will be tried in 12,100 gacaca courts throughout the country. Inaddition, a National Unity and Reconciliation Commission was established to redress Rwanda's legacy o f divisive politics through civic education initiatives. Further, about 3.5 million refugees have been repatriated and re~ettled,~and more than 37,000 ex-combatants have been demobilized, of which about 15,000, including the former commander-in-chiefof the (FDLR),6have been integrated into the Rwanda Defense Force at various command levels. 1.7 A decentralization program was undertaken in 2000. The first phase of the decentralization (2000-03) involved a number of legal, institutional, and policy reforms aimed at defining central and local roles and responsibilities; establishing accountability mechanisms; and ensuring grassroots participation in political, social, and economic development processes. The second phase (2004-08) i s emphasizing service delivery to communities through an integrated framework. 1.8 In addition, a new constitution was promulgated in 2003. At the same time as the decentralization, a new constitution provided the legal framework for managing, controlling, and resolving social conflicts, and for power sharing at the top level o f government. The president cannot appoint more than 50 percent of the cabinet from hisher own party, and the speaker of the parliament has to come from the minority party. 1.9 These reforms have helped to maintain stability and peace in Rwanda, which in turn has made possible the stronggrowth and social sector improvementsobserved over the past decade. Under-five mortality has been reduced by 30 percent, and now stands at pre- genocide levels. Immunization rates are at 90 percent, among the highest in SSA; and HIV incidence i s 3 percent of the population. The next challenge is to addressthe problem of maternal mortality and nutrition. In education, net enrollment i s 93 percent at the primary level. The challenge going forward will be to improve the quality of education. 1.10 However, the overall GDP growth rate has been slowing. Between 1995 and 2005, overall GDP growth averaged 10 percent per year, with a corresponding per capita GDP growth rate of 5.4 percent. Duringthe latter years of the decade (2001 to 2005), however, GDP growth slowed to 5.2 percent, with a corresponding per capita growth rate of 2.7 percent, driven mainly by the agriculture sector. Years of destruction and neglect arising from the civil conflict, coupled with insufficient allocationofresourcestowardthe delivery of economic and social services, have acted as barriers to continued highgrowth and increasedproduction, particularly inrural areas. Currently there are about 57,000 Rwandanrefugees in21 African countries. The FDLR (Forces Democratiquesde Liberationdu Rwanda), founded in2001, is a rebelgroupoperating ineasternDemocraticRepublicofthe Congo. 3 1.11 As a result, there has been only a marginalimprovementin income-based measures of poverty. Preliminary estimates from the second EICV, conducted in 2005-06, indicate that economic growth rates have resulted inonly a marginal reduction innational measures of income poverty- from 60.4 to 56.9 percent. Rural poverty declined from 66 to 62.5 percent; and poverty inKigali and other urban areas declined by 3 and 5 percentage points, respectively. Poverty in Kigali now stands at 13 percent, and in other urban areas at 41 percent. The main factor determining these trends has beenthe level o f productivity and limited progress inproduction for the market, inthe agriculture sector. 1.12 These trends pose a challenge to sustained growth, given the structure of the economy. The structure o f the economy is such that it is heavily dependent on agriculture and related primary products, which account for 40 percent of GDP. Low levels of irrigation means that there i s a high dependence on rain-fed agriculture systems and the accompanying high vulnerability to terms-of-trade shocks. In addition, the landlocked position o f the economy, coupled with the poor condition of infrastructure, leads to very high trade costs, and impedes export competitiveness and trade. Finally, the highrate o f population growth poses an important policy problem if the country i s to reduce the share o f the population engaged in agriculture. Measures to transform the agriculture sector to one that is market oriented, along with policies to support growth and employment in the non-agriculture sectors, will be critical for sustained growth. 1.13 Rwanda is also characterizedby a low private investment and savings rate and a low export base,which limits the degree of possible resource mobilizationand increasesthe reliance on external aid. The country has benefited from the recent Highly Indebted Poor Country (HIPC) and Multi-lateral Debt Relief Initiative (MDRI) initiatives', but will need to strengthen its export base. The continued low export base limits the level of external borrowing the country can undertake without a seriousdeteriorationinthe net presentvalue ofdebt to export indicator. Macroeconomic progress puts the country in a good position with donors to receive external grants. However, highlevels o f grants or aid will not be sustainable inthe longterm, and the Government needs to ensurethat aid funds are well investedso that when donors exit, growth will not be comprised. This will require attention to appropriate macroeconomic management o f aid funds and to the efficiency and composition of spending. 1.14 The CEM therefore begins with an examination of how aid funds have been used and in what ways efficiency and composition of spending can be improved to support growth. It stresses the need to re-examine spending allocations, to improve the delivery o f infrastructure and economic services while maintaining the gains made in the social sectors. Increased efficiency in spending will be the key to achieving this aim. In addition, there will need to be increased engagement of the private sector. Following on this, the CEM examines the prospects for and constraints to growth, and the implications for meeting the Vision 2020 targets. Based on this analysis, the report puts forth a set o f priority measures for consideration by policymakers. 7These two initiativeshave ledto a substantialreductioninRwanda's debt stock from 93 percent of GDP in2003 to 15 percentby end-2006. 4 1.15 The CEM identifies three main constraints to growth. First is the problem of low productivity in agriculture, arising mainly from low use of inputs. Second, the poor condition of infrastructure - particularly electricity, transport, and water - has led to stagnant growth in the manufacturing sector; limited the extent o f growth and transformation in the agriculture sector; and limitedthe development o f the service sector. Third, the low availability o f skilled labor i s an underlying constraint that affects the entire economy. A comprehensive strategy is needed to address this issue so that it does not become a bindingconstraint inthe future, particularly inthe area of export growth and diversification. The report concludes with a prioritized set of interventions to support growth and poverty reduction. 2. THE MACROECONOMICS OFAID MANAGEMENT 2.1 Over the period 2007 to 2020 aid flows could as much as double in total from the current level of around 15 percent of GDP, provided Government continues to maintain sound policies. For 2007, there are already some indicationsof scaling up with increased budget support from the African Development Bank (roughly an additional 1to 2 percent of GDP), there are prospects for more donors to move to budget support (with at least one joining in the next year), and the Education for All Fast Track Initiative (roughly a total o f 2 percentage point increase, expressed as a share o f GDP, for a period of two to three years). In the medium term, although the order of magnitude and timing is not yet clear, further scaling up could be forthcoming from the US. Millennium Challenge Account. Therefore, assuming an average annual increase of 2 to 3 percentage points per year, this could give rise to substantial inflows to the economy. In an optimistic scenario o f scaling-up, a 2 percentage point increase each year in the flow of aid would in 10 years give rise to aid flows of around 30 percent of GDP; while domestic revenue would still be below 20 percent o f GDP 2.2 The implications of this magnitude of aid inflows will depend on how the Governmentmanages the macroeconomic effects of the increased liquidity, and on how the aid funds are used. Since 1997, when the economy begun to stabilize following the genocide, the country experienced two episodes of aid surges: one in2000, when the ratio of ODA to GDP increased from about 16 percent in 1997-99 to about 20 percent in 2000-2002; and another in 2003, when it rose 25 percent over the period 2003-2004. However, high aid flows did not prevent depreciation inthe real exchangerate (RER) duringthis period, mirroringthe experiences o f Tanzania and Uganda duringtheir periods o f aid surge (IMF 2005). 2.3 The likely reason for the RER depreciation in Rwanda (as well as Tanzania and Uganda) after the aid surge was that very little of the additionalaid was absorbed or spent. The terms of trade (TOT) in Rwanda deteriorated by more than 10 percent in 2000-02, and improved by more than 8 percent in 2003-04, while aid flows as a share o f GDP rose by 11 and 27 percent, respectively. However, the RER depreciated by about 7 and 11 percent for the two periods. This suggests that the RER depreciation in the latter period was due less to the TOT effect, and more to the steady increase in reserves over the decade. More recently, the steady increase in grants over the period 2004-2006, has been accompanied by a real effective exchange rate appreciation of around 16 percent. 5 2.4 Over the period 1995 to 2005, core inflation-Le., excluding energy and food prices -remained stable at around 5 percent up to 2004. However, rising food and energy prices pushedoverall inflation to 10percent by end-2004, mainly becauseo f supply shocks, particularly related to shortages in electricity generation. By 2005, end-year inflation had dropped to 5.6 percent, even with an impulse of 2 percent from increased electricity tariffs. 2.5 These trends highlight the importance of monetary policy to manage the liquidity due to aid flows. Inthe case o f Rwanda, it suggests the need for an increased focus on the sales of foreign exchange (as opposed to sales of treasury bills which can lead to the crowding out of the private sector). There have been some concerns about the potential effect on exports of accommodating an appreciation o f the real effective exchange rate. However, an econometric analysis of determinants o f exports (discussed in Volume 2) indicates that the export response to the exchange rate is significantly outweighed by supply-side factors, and the effect of domestic prices. Accommodating a moderate appreciation of the exchange rate through foreign exchange sales will likely not have a strong effect on exports. The highly significant effect o f supply-side factors on the exchange rate therefore indicates that the focus needsto be placed on removing the supply-side constraints that are impedingexports. Box 2.1: Measuring the Use of Marginal Increases in Aid Flows The rate o f absorption is defined as: Absorption=A(non-aid current account deficit)/ AAid For a given fiscal policy, absorption is controlled by the monetary authority through its decision o n how much o f the foreign exchange associated with aid to sell and how much to sterilize through its interest rate policy, which affects the demand for private imports through aggregate demand. The central bank can choose to increase reserves or to make them available to importers. In the case when the central bank uses the full amount o f the aid to increase international reserves, none o f the aid increase i s absorbed. The rate o f spending i s defined as the increase inthe fiscal deficit net o f aid, divided by the increase inaid: Spending= A(G-T)/ M i d Spending captures the extent to which the Government uses the aid to increase expenditures or reduce taxation. The increase in expenditures can come ffom the purchase o f imports or domestically produced goods and services. Source: IMF,2005. Figure 2.1: Rwanda's Monetary Indicators, 1995-05 1995 1996 1997 1998 199s 2000 2001 2002 2003 2004 2005 I +resews/GDP - r e s e w s as months of mporls -A- domesec OebVGDP Sources: MINECOFINand Bank staff estimates, 2006. 6 2.6 For aid to be translated into real transfer of resources to the economy, it must be absorbed through its sale by the central bank in exchange for domestic credit or to retire publicdebt. Instead, inRwanda, a sizable share o f the marginal increase inODA has been used to build up reserves, with the rate o f reserve accumulation rising by 3.7 and 2.7 percent respectively, between 1997-99 and 2000-02 and between 2000-02 and 2003-04. As early as 2001, gross reserves were 12 percent o f GDP, reaching a peak o f more than 16 percent in 2004 before declining to about 13 percent in 2005 (see Figure 2.1). This level o f gross reserves accounted for five months o f imports in 2005. As Table 2.1 shows the marginal increase in aid was partially spent (at a rate o f 19.6 percent in the first surge and 53 percent during the second surge) and not absorbed. Table 2.1: Analysis of Two Aid Surges for Rwanda Pre- Suree 1 Difference Aid Surge 2 Difference Aid Surge 2000-02 from absorbed 2003-04 from absorbed 1997-99 previous previous Percent of GDP (unless otherwiseindicated) Aid flow 15.8% 20.3% 4.6% 24.7% 4.3% Non-aid CA -9% -5.9% 3.1% 0.0% -5.4% 0.5% 0.0% balance Not Not Absorbed Absorbed Non-aidKA 1.8% -2.1% -3.9% -4.2% -2.1% balance Change in -8.7% -12.4% -3.7% -15.o% -2.7% reserves (-=increase) Aid Spent Aid Spent Net fiscal aid 7.1% 14.1% 6.9% 18.5% 4.4% inflows Revenue (excluding 10.3% 11.1% 0.8% 13.7% 2.6% grants) Expenditure (excluding external 12.1% 14.3% 2.2% 19.3% 5.0% interest) Overallfiscal -1.8% -3.2% -1.4% 19.6% -5.5% -2.3% 53% balance Marginally Partially beforeaid Spent Spent Source: Bank staff estimates, 2006. Note: CA -Current Account; KA-Capital Account. 2.7 Reducing the dependence on aid in the long-term will require investing to promote private sector development and the broadening of the tax base; short-term measures can also be adopted. Aid flows may be predictable in the short- and medium-term but are not a permanent source o f government revenue. One strategy to reduce the dependence on aid is to consider over time a higher share o f aid flows as a financing source (below the line) rather than as a permanent source o f revenue. For instance, in the next five years 70 percent o f aid may be consider as permanent revenue, in the following five years it could be 50 percent. This would also help to keep the exchange rate at its appropriate or equilibrium level. Another strategy i s for 7 government to continue to improve revenue administration and efficiency to reduce costs and improve collection (inamanner that does not unduly constrain growth inthe private sector). 2.1 THECOMPOSITIONAND EFFICIENCYOFSPENDINGTOIMPROVEABSORPTION 2.8 Where the aim is to stimulate growth, aid funds should be used to expand the capacity for absorption. This requires improvements in the composition and efficiency of spending. Between 1998 and 2005, total spending increasedfrom 18.9 to 26.5 percent of GDP. Total priority spending increasedfrom less than a third to roughly 50 percent of the total budget, with half of the spending going to the social sectors. Increased aid inflows went disproportionatelyto recurrentpriority spending(Table 2.2). Table 2.2: Rwanda-Public Recurrent,%ending on Prioritv Sectors. 1998-2005(share of GDP) 1998 1999 2000 2001 2002 2003 2004 2005 TotalBudget 18.9 19.69 18.7 21.0 21.3 23.9 26.1 26.5 Recurrentpriority spending, ofwhich: 2.8 3.9 4.0 5.3 6.2 6.9 8.5 8.4 Education 2.2 3.2 3.2 3.5 3.9 4.2 4.0 3.9 Health 0.4 0.5 0.6 0.7 0.8 0.6 1.0 0.9 Agriculture 0.0 0.0 0.0 0.2 0.3 0.2 0.2 0.4 Exportpromotion 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.5 Transport and communication 0.0 0.0 0.0 0.1 0.3 0.4 0.3 0.9 Inftastructure(energy andwater) 0.0 0.0 0.0 0.0 0.2 0.0 1.1 0.3 CommonDevelopmentFund(CDF) 0.0 0.0 0.0 0.0 0.1 0.4 0.3 0.3 Other* 0.2 0.2 0.2 0.9 0.5 1.0 1.3 1.3 *This category includesspending on internal afairs, local government, commerce, andyouth and sports. Sources: Rwandese authorities and Bank staff estimates 2.9 While increased spending in the social sectors led to substantial improvements in outcomes (Table 2.3), there has been only limited spendingfor economic services, including investment to improve productivity in agriculture and manufacturing. Total infrastructure spending (including private and public sources) is estimated to be 7.1 percent of GDP, in 2005. This is relatively low by regional and international standards. For example, in2004, the spending level for Uganda is estimated to be 7.2 percent of GDP, 9.2 percent of GDP for Tanzania, and 10.6 percent for Kenya (see Table 2.4). As the remaining chapters of this volume will show, increasedspending on infrastructure, particularly to increase access to markets and increase the availability of energy, will be neededfor sustained growth. 2.10 Rwanda's average annual spending per capita of US$16.80 is very low by internationalstandards. Uganda's annual per capita spendingon infrastructure is higher by 13 percent, Tanzania's is higher by 84 percent, and that of Kenya by 185 percent. In international comparison, Rwanda's spending on infrastructure is less than half o f the lowest average annual per capita spending in Latin American countries at the end of the 1990s. It is comparableto what Indonesia was spending in infrastructure right after the 1997 financial crisis, while Indonesia's current spending level i s estimatedto be more than five times that ofRwanda(Table 2.4). 8 Table 2.3: Progressin Selected Social Indicators (Percent of relevant Population) Social Sector Outcomes Status in 2001 Estimate2005 Gross PrimaryEnrollment 117 140 Net PrimaryEnrollment 73 94 PrimaryCompletionRate 29.6 55 TransitionRateto Secondary 43 47 DPT3 coverage of 0 to 1year olds 77 85 InfantMortality Rate(per 1000births) 107 83 Use ofmalariatreatedbednets 4 30 Enrollmentinmutuelles 7 41 Populationwith access to improvedwater source 52 57 Source: DHS2005. Table 2.4: International Benchmarkingof Infrastructure SpendingLevels Infrastructure spending Income per capita Latest year (US$/person) (%GDP) (Current US$) available Uganda 2004 19.0 7.2 250.0 Tanzania 2004 20.9 9.2 297.3 Rwanda 2005 16.8 7.1 236.8 Kenya 2003 47.8 10.5 430.0 Indonesia 2004 93.2 7.9 1166.5 Turkey 2004 240.1 5.7 4209.7 Chile 1998 3 17.3 6.4 4923.5 Note: Infiastructure includes energy (oil, gas, andelectricity), water and sanitation, irrigation. Source: Benitez and von Klaudy (2006), and Infrastructure FiscalCost Baseline, 2006. 2.11 The need for scaled-up infrastructure spending becomes more pronounced when initial conditions are factored in. The current stock and condition of rural roads in Rwanda is low. Most Latin American and important Asian countries already had more developed road and electricity networks by the 1990s than Rwanda has today. For example, the worst electrification rate of the Latin American sample - 72 percent in Peru- is nearly twelve times higher than the electrification rate in Rwanda today, which is roughly 6 percent (von Klaudy and Benitez 2006). Moreover, the marginal cost of extending networks to isolated areas (which represents a large part o f expenditures in Rwanda) is significantly higher than it is for population clusters in the more urbanized Latin American and Asian countries. 2.12 At the same time, spending in the social sectorswill needto be maintained, giventhe need for a skilled and healthyworkforce. Although spending in education and healthhas been relatively higher than spending on infrastructure, it i s still below what is needed to meet the MillenniumDevelopment Goals (MDGs) and support growth. According to the MarginalBudget for Bottlenecks (MBB) model', an additional $20.00 per capita per annum, or an additional 7 percent of GDP per year, would be neededto meet the MDGs for health. Similarly, in education, based on the technical assessment of the Government's education strategy for the Education For *The MBBmodelandanalysis was conductedby ajoint UNICEF and World Bankteam ledby Netsanet Walelign. 9 All Fast Track Initiative (EFA-FTI)', it is estimated that an additional 5.5 percent of GDPwill be needed to fill the financing gap for measures needed to achieve the MDGs. In the case of the education sector, Rwanda will benefit from the EFA-FTI, which will provide, on average, $60 million to $70 million per year between 2007 and 2010 toward meeting the funding gap in education. Therefore, additional donor resources, as they materialize, can be focused on needs arising from other sectors. 2.13 Until additional resources materialize, the Governmentshould focus on improving the efficiency of spending, and identifying ways to involve the private sector. The Government has moved to improve efficiency in spending by focusing spending in those areas where eliminating constraints would yield the highest benefits. However, further efficiency gains are possible through a reduction in administrative costs, improved resource allocation to reflect long-term strategic priorities, and better coordination of interventions. 2.1.1 Efficiencyof SpendinginHealth 2.14 In healthspending, the main efficiency issues are (i)allocationof resources by level of care or type of facility; (ii)allocation of resources across programs that is not closely aligned with long-term needs and sustainability; and related to this, (iii) limited coordination of activities by the largenumberof donors and NGOsoperatingin the sector. 2.15 Allocation of resources between referral hospitals and primary care facilities is a key areawhere efficiency can be substantially improved. Referral hospitals receive 42 percent o f total health spending; the central administration receives 38 percent o f total expenditures, while district hospitals receive 10 percent of total health spending. National programs receive 8 percent of total health expenditures and other national agencies receive 6 percent of health spending. Spendingefficiency could be improved by increasing the focus on districts and health centers giventhat primary healthcare takes place inthe district facilities. 2.16 Results from the MBB model indicate that additional resources will have greatest impact if focused on community schemes, to strengthen the performance-based contracts and the safety net for mutuelle membership. The Global Fund" has recently committed resources to fund mutuelle" payments in six provinces. This commitment demonstrates that there i s scope for disease-specific funds to support system improvements linked to outputs relevant to the core mandate12. However, coordination across the main programs (i.e. Global Fund, MAP, PEPFAR, as well as government) needs to be strengthenedto ensure alignment with The EFA-FTI was launchedin2002 to galvanize countryreformsand catalyze donors to fund these to achievethe MDGsand universaleducationgoals. Donorsalignongoingfunding on endorsedplans, and targets, basedon soundcountrypolicies,to increasethe efficiencyandtransparencyof assistance. 10The GlobalFundwas createdto finance adramatic turn aroundagainst AIDS, tuberculosis, andmalaria, which kill over 6 millionpeopleeachyear. It is an innovativeapproachto health financingthat involves government, civil society, the private sector and affectedcommunities. l1Mutuelles are communitybasedhealthfinancingschemes developed inRwanda. l2See the Round5 proposals for HealthSystem Strengthening, www.theylobalf'und.org. The GlobalFund operates on aset ofprinciples: it is strictly afinancing mechanism, it does not implement;it leverages other resources;supports programsthat reflect nationalownership; promotes abalanceacrossregions, other diseasesand interventions, as well as preventiveversus curative treatment; proposals are evaluated through an independentprocessthat promotestransparency and accountability. 10 an overall health sector strategic focus and vision. Along these lines, incentives to strengthenthe performance contracting system that are tied to the improvement o f broad healthoutcomes would helpto promote a more integratedapproach across programs. 2.17 Increased coordination across programs would also help to reduce administrative and program delivery costs in health. Rwanda has 21 official donors that disburse funds through a network o f more than 40 NGOs; and there are many more civil society organizations (CSOs) and community-based organizations (CBOs) engaged in activities that need to be more strongly linked with an overall sector strategy and vision. In addition, there is a need for improved coordination to minimize gaps and duplication of activies. PEPFAR13alone disburses funds via more than 50 separate implementing organization^,'^ which at times results inthe same district or facility receiving support from several different agents, although the funding comes from a single donor. 2.18 Coordinationwould also help to improve the centralgovernment's ability to plan in an integratedand system oriented manner. The central government manages only 14 percent of donor support to the health sector, which seriously hampers its ability to plan and manage health services as a system - even though more than 75 percent o f health services are delivered through Government facilities or Government-assisted nonprofit service providers, who are viewed as part of the public system. As documented in the Government's 2006 assesssment of needs for scaling-up to achieve the health MDGs", this has compromised efficiency leading also to uneven spending across regions that i s not necessarily aligned with needs. In 2005, total per capita health spending was estimated to vary from $1.86 per capita in the least funded province (Ruhengeri) to $11.84 inthe highest funded province(Kibuye). 2.19 There is also a need to re-examine the distributionof donor funding across specific programs. For example, $18 million has been earmarked for malaria (which represents the biggest cause of mortality and morbidity), and $1 million for the integrated management o f childhood diseases, which stands in stark contrast to the $47 million allocated to HIV/AIDS (Figure 2.2). Even though, for some programs, a significant amount of funds are allocated to mitigation activities (such as income generation and training inthe case of the MAP), efficiency could be improved through greater coordination between programs that emphasizes an integrated approach at the level o f the overall health system (as opposed to the specific program level), in line with the overall incidence of disease and broad health needs as perceived by patients and health providers16. l3PEPFAR, The President'sEmergencyPlanfor Aids Relief(PEPFAR) is aU S government funded programthat haspledged $15 billion over 5 years (2003-2008) to fight the HIVIAIDS pandemic. The programhopes to reach 120countries inthe world, providing anti-retroviral treatment (ART) for 2 million, preventing7 millionnew infections,and supporting care for 10million. 14See http:llwww.state.povlslpaclcountriesl~clrwandaipartnersl. l5Governmentof Rwanda, 2006. MINISANTE. l6Forexample,theMAPprojectrecentlyinitiatedperformancebasedcontractingandhasplanstotiepaymentsto quality indeliveringbasic healthservices in order to support strengthening ofthe overall healthsystem. 11 Figure 2.2: Distribution ofDonor Funding by StrategicObjective Other healthservice $30 HIV/AID funding $20 $10 $0 I L L HR D,V, C Geogr. Financ. Health Referral Inst. Access Access Services Hosp. Capacity Sources: MINECOFINandMINSANTE,2006. 2.20 The proliferation of NGOs also puts additional strain on human iesources, which are already severely limited, and which in turn increases administrativecosts. Inparticular, physiciansemployed by NGOs to deliver HIV/AIDS services are paid almost six times as much as physicians in the Ministry of Health. Such differences in salaries pose a challenge to keeping well qualifiedhealthpersonnel inthe public sector.17 2.21 Finally, aid resource programming in the health sector raises the issue of sustainability in additionto cost effectiveness. For example, increasedresources for health are programmed annually inthe case of PEPFAR, and for 2 to 5 years inthe case o fthe Global Fund. However, greater emphasis i s needed toward strengthening national systems and institutions, to enable an eventual reduction indonor funding. At the present time, any interruption inaid would present a serious problem to service delivery, since the Government i s not in the position to take on the financing, given available resources, the level of needs, and competing demands. 2.22 The issuesof spendingefficiency inthe healthsector can be summarized as follows: Disproportionate allocation of resources to referral as opposed to primary care facilities where the needs are greatest; Insufficient coordination o f programs (both government and donors), at the overall health system level to ensure activities are better couched in a comprehensive and strategic sectoral context that is in line with long-term goals, thereby minimizing the attendant risks to sustainability of the health system; Distortion ofthe current system for allocating resourceswhere they are needed, due to limited Government control o f health sector resources, which are mainly off- budget; High transaction costs, coupled with the high proportion of funds absorbed by management and administrative costs rather than paying for service delivery. l7MINISANTE,HumanResourcesfor Health Strategic Plan2006-2010, 2"ddraft, April 2006. 12 2.1.2 Efficiencyof SpendinginEducation 2.23 In the case of education, the technical assessment completed in the context of the EducationFast for All FastTrack Initiative (EFA-FTI)'* highlighted: (i) the needto improve quality by attracting and keeping well-trained teachers, and improving access to school structures and learningmaterials; and (ii) the needto reduce the relative subsidyto higher education inorder to release more resources to improving quality at the primary and secondary education levels. In terms of higher education, the focus should be on support to producing graduates in those technical areas that are neededfor economic growth and employment. 2.24 The Government's policy of nine years of basic education has led to significant improvementsin primary enrollmentrates(which are now at 92 percent for netenrollment and 136 percent for gross enrollment), but quality needs to be improved. Increased enrollment has exacerbated the crowding o f classrooms; the number of students per class increased from 54 in 2000 to 65 in 2005. Simultaneously, the student-teacher ratio increased from 51:l in 2000 to 64:l in 2005. This i s compounded by the fact that although teachers are accredited, their methods and approaches are outdated; Le., they do not have effective competencies and skills. In addition, there are problems with the procurement o f books and training materials. Greater attention to these issues could help improve learning outcomes and the rates of transitionto higher levels. 2.25 Poor transition rates result in limited and inequitable access at the secondary and tertiary levels. The primary completion rate is less than 45 percent (in contrast to the SSA average o f 62 percent); and transition rate from primary to secondary is 45 percent (compared to 60 percent for SSA). Gross enrollment at the secondary level is 17 percent (compared to 28 percent for SSA). The problem i s worse for girls than boys, and stems from the quality of education and learningoutcomes at the primary level. 2.26 Efficiency in spending could be achieved through increased allocation to teaching materials, construction of schools, and training for teachers, particularly at the primary level. The funding for primary education has increased, and there are plans for it to increase further under the EFA-FTI. Given the need for learning materials, funds could initially be focused on imports of textbooks. This would help also to quickly absorb these additional aid funds, andminimize the needto sterilize the funds domestically. 2.27 Further efficiency gains could be realizedby reducing boardingin secondary school and re-examiningthe level and structureof subsidies to higher education. Higher education i s subsidized even though the system i s not producing the type of credentials that are critical for growth. It i s estimated that 63 percent of students are inthe general social sciences, 14 percent in the sciences, 11percent inmedical andrelatedfields, and 8.6 percent inengineering. 2.28 Expendituresfor higher education remain high, even though the share of recurrent education expenditure fell modestly from 35 percent in 2003 to 31 percent of recurrent l8 The EFA-FTItechnical assessment involved a review ofthe GoR's EducationSector StrategicPlanto identify priority areas to achieve the MDGs and determinethe level of financing neededto fill the financing gap. Identified funding will cover the period up to 2010 with additional commitmentsto 2015 where feasible. education expenditure in 2005. Student subsidies for tertiary and higher education have increased from 37 percent of tertiary recurrent expenditure in 2005 to 42 percent in 2006. The subsidy covers tuition payments (FRw 300,000 per year) and a monthly allowance that is more than double the average monthly per capita income. The gross salary o f teaching staff averages 41 times GDP per capita, compared to 3.8 times for primary school teachers. Yet, the private schools charge substantially less and have accounted for most o f the recent growth in higher education, particularly among women. These levels stand in stark contrast to the extremely low levels of spending on text books and construction of schools to ease overcrowding at the primary level. There is therefore a need to realign subsidies to support greater improvements in the quality of primary and secondary education. This can be achieved through, for example, placing increased emphasis on teacher training, with particular attention to improving student achievement inthe math and the physical sciences, as well as the social sciences. 2.29 A recent assessment of school infrastructurehighlightedthe benefits of moving to a more coordinated, community based approach". The study estimated that the project approach is producing about 240 classrooms per year, when the need is for 2,200.20 Scaling-up of school construction will require a decentralized approach, similar to the one adopted by Uganda. That approach uses pooled donor funding, harmonized norms and financial procedures, robust technical and financial monitoring, and active involvement of school communities, leading to the construction o f 19,000 new classrooms in a little over 4 years. The feasibility o f the approach i s highlighted by the fact that there are already communities inRwanda buildingtheir own schools, sometimes with the support of NGOs. However, giventhe differences incapacity across districts and communities, government would need to play a facilitating role. The Government i s currently preparing a school construction strategy that would integrate all sources of financing, and establish guidelines for construction. Government's role will be to set standards and facilitate the relationship between contractors and districts. 2.30 Finally, integrationof the recurrent and development budgets will help to ensure that the full costs of programs are capturedin the budget. Followingpassageo f the Organic Budget Law, in2005 Government has begunto integrate the development and recurrent budgets. This integration will help ensure that the recurrent cost implications o f project investments can more readily betaken into account. Inthe meantime, however, the donor project investmentsthat are outside the Government's education strategy -Le. the Education Sector Strategic Program (ESSP) - are continuing, particularly in the higher education sector, with potential recurrent cost implications that are not being captured. Overall, roughly 8 to 10 percent o f total government expenditure still remains off-budget. 2.1.3 Efficiencyof SpendingonInfrastructure 2.31 Efficiency gains in infrastructure spending can be realized by addressing: (i) the level of maintenance expenditures and low cost recovery; (ii) the need to have a project selection process based on economic returns and Government objectives; (iii) the need to reduce the large 19 See HumanResourcesDevelopmentProjectSupervisionMission Aide Memoire, October 2006; and Foster, Mick, J. Virtue, and S. Ndaruhutse, 2005. "Improving the Provisionand ManagementofExternal Supportto Education," January-February. MINEDUC. 2o See Foster et al.,2005, op. cit.. 14 quasi-fiscal deficits; and (iv) the need to increase the role o f the private sector. Table 2.5 summarizes the allocationo f costs across sectors in infrastructure. 2.32 In the road sector, maintenance of the entire network is constrained by the lack of road maintenance funds. This is reflected in the low ratio o f actual versus normative road maintenance expenditures in Table 2.5. In August 2002, the Government defined its strategic priorities for roads, stating that resources for maintenance should be distributed to national roads o f international interest, roads o f local (communal) usage, and the urban road network. However, this policy has yet to be fully implemented due to capacity and funding constraints. Instead, maintenance i s carried out ad hoc or inresponse to urgent needs for repairs. There i s a clear need to develop a systematic, planned approach to rehabilitation and maintenance, particularly if the private sector i s to be effectively involved. Further, only one third o f the funds needed for maintenance are generated from user levies and taxes. It will be necessary to define options for increasing cost recovery and a sustainable subsidy mechanism. 2.33 Maintenance expenditure for Electrogaz' electricity and water supply operations are below the normative level; in the case of electricity this is by a significant margin (see Table 2.7). These numbers are consistent with assessments and perceptions o f the level o f service provision in each sector. Inthe electricity sector, lack o f maintenance has resulted inhigh transmission losses, low available capacity o f hydro generation, and frequent outages. Inwater, while the estimated ratio ranks better, the existence o f unaccounted losses indicates potential sources o f inefficiencies in the usage o f maintenance expenditures. However, despite the recent reduction in the overall level o f leakages and unaccounted losses, high technical losses and unreliability o f the dilapidated network have led to extensive shortages and high explicit and implicit (hidden) cost o f supply (see Table 2.6). 2.34 The levels of expenditures for Electrogaz' electricity and water operations requires a re-examination of the tariff structure, as well as the application of maintenance expenditures, to reduce losses and improve cost-recovery. The flat tariff structure for electricity distorts economic incentives and limits the ability to effectively target cross-subsidies to existing and prospective low-volume customers and poor households. A recent tariff study2' proposed a new tariff structure with differentiated prices for different consumer categories, including a block structure for small and medium power users and a time-of-use tariff for large users. This overhaul, which requires changes to the metering systems, has not yet been introduced, and only a marginal change aiming at a straight tariff reduction for large users has been considered. Whatever the future tariff structure, it will need to be adjusted more frequently than inthe past, to take into account the effect o f volatile fuel import prices on the cost o fthermal power generation. In the water sector, large unaccounted losses in water supply suggest inefficiencies in the usage o f maintenance expenditures, with tariffs that are well below cost- recovery. Resolution o f the issues raised here would significantly redress the problem o f Electrogaz' low cost recovery and large quasi-fiscal deficits. Their resolution would require the design and implementation o f a comprehensive restructuring program which would need to focus on the merits and drawbacks o f a potential separation o f electric and water utility functions. It would also need to clearly elaborate the role government wants the private sector to play in 21ELECTROGAZ, Assessment of current and Proposalofnew Power and Water Tariffs, 2005. See also CORE International Incorporated, 2005. www.coreintl.com. 15 operating, managing and financing electricity and urban water supply. Efficiency gains in infrastructure service delivery can also be expected from greater involvement o fthe private sector in the management and financing of infrastructure operations. Options for private sector engagementare discussedbelow. Table 2.5: Rwanda-On- and Off-budget PublicExpenditures Nominal GDP Shares (%) 2002 2003 2004 2005 Infrastructure Public Spending 4.97% 3.80% 6.26% 6.59% Total Gross Investment 2.71% 1.58% 3.62% 3.37% Total O&M 2.26% 2.22% 2.65% 3.22% Electricity 1.19% 1.30% 2.79% 2.81% Gross Investment 0.34% 0.44% 1.54% 1.31% O&M 0.85% 0.87% 1.26% 1.51% Road Transport 0.99% 0.29% 0.83% 1.23% Gross Investment 0.96% 0.28% 0.83% 0.86% O&M 0.03% 0.00% 0.00% 0.37% Air Transport 0.49% 0.15% 0.22% 0.49% Gross Investment 0.46% 0.15% 0.19% 0.45% O&M 0.03% 0.00% 0.03% 0.04% Communications 1.17% 1.16% 1.24% 0.89% Gross Investment 0.13% 0.13% 0.17% 0.05% O&M 1.04% 1.03% 1.07% 0.83% Water Supply and Sanitation 1.11% 0.86% 1.17% 1.16% Gross Investment 0.82% 0.56% 0.89% 0.71% O&M 0.30% 0.30% 0.28% 0.45% Irrigation 0.00% 0.03% 0.00% 0.01% Gross Investment 0.02% 0.00% O&M 0.00% 0.00% 0.00% 0.00% Source: World Bank Infrastructure Fiscal Cost Baseline, 2006. Table 2.6: InfrastructureQuasi-FiscalDeficitsin Rwanda Percentage ofNomrnal GDP 2002 2003 2004 2005 Electricty UnaccountedLosses21 0.14% 0.20% 0.23% 0.27% Under-Pricing31 0.16% 0.20% 0.23% 0.16% Collection Inefficienices41 0.15% 0.14% 0.19% 0.25% ELECTRICTY HIDDENCOST 0.45% 0.55% 0.65% 0.68% Water Supply UnaccountedLosses21 0.15% 0.13% 0.06% 0.07% Under-Pricing31 0.15% 0.16% 0.12% 0.26% Collection Ineficienices 41 0.07% 0.07% 0.06% 0.06% WATER SWPLY HIDDENCOST 0.37% 0.36% 0.24% 0.38% TOTAL HIDDENCOST 0.82% 0.90% 0.89% 1.06% Sources: von Klaudy,and Benitez 2006; and World BankInfrastructure Fiscal Cost Baseline, 2006. Figures refer only to Electrogaz operations. 16 Table 2.7: MaintenanceSpending in Infrastructure inRwanda (on and off budget) Actual Unit RatioActual to Normative Maintenance spending/ MaintenanceCost(USD) Unit. Maintenance Cost asset value (%) NormativeUnit Maintenance Electricity Cost USD168per KW 2004 1,528,576 20.60% 2.40% NormativeUnit Maintenance Road CostUSD2,9141 per KW 2005 8,631,523 21.20% 2004 3,701,388 9.10% 2003 3,976,689 9.70% 2002 5,807,233 14.20% NormativeUnitMaintenance Water Supply & Sanitation CostUSDl2 per connection 2004 358 691 76 70% 1 70% Source: von Klaudy and Benitez, 2006; and World BankInfi-astructureFiscal Cost Baseline, 2006. 2.2 ROLEOFTHEPRIVATE SECTOR 2.35 The private sector could be engaged to improve the delivery of services in health, education, and infrastructure. 2.36 Health. The private sector is strongly involved in delivering services through contracting. In2005, spendingfrom both the public sector and donors came to between $12 and $14 per capita.22In addition to public sector and donor spending, the private sector in 2003 accountedfor 25 percent of total health expenditure.This amount of private spendingrepresented an additional expenditure of $3.47 per capita, of which $2.75 came from households, and the remainder from private employer contributions and NGOS.~~ health expenditure by 2005 Total may therefore have been on the order of $15 to $17 per capita, when spending by the private sector is included. Despitethis, it is estimatedthat an additional $20 per capitawould be needed to undertake the actions necessary to achieve the MDGs for health. Increased coordination of programs, led by government would greatly increase efficiency of spending by both the public and private sector and reduce duplication of activities. 2.37 Private costs, however, can be a barrier to access. Studies show that payments as low as $1 are beyond the financial reach of many households in rural areas,24where cash incomes are 22 See Scorraville, G., J. Nachtiga, and J. Munyenpenda, 2006. Rkpublic du Rwanda, Ministere de Santd, Revue de Dkpenses Publique dans le Secteur du Sante, 2002-2005; Janvier, page 18. This report gives a figure of $8.5 per capita, based on external aid of $4.7 per capita, or $45 million. . However, the same document quotes healthaid as $75 million in2005, $7.8 per capita, which would put total Government plus donor spending inthe range of $11.6 per capita. Coverage of health aid statistics i s incomplete, and health aid is believed to be more than $100 million, putting per capita spending in the range of $13 if user fees are excluded, and morethan $14 ifincluded. 23National Health Accounts Table 24 and Table A-I. Table 25 inthe same source quotes a much higher private sector share of 47 percent, presumably because off-budget spending by donors is included with the private sector. 24Smithson, P. and J. Martinez, 2006. Addressing Health MDGs inRwanda Progress, Gaps, Challenges and Opportunities, DFID resource Center, February. See also Kalk et. al. 2005. 17 low and more than 40 percent o f the population is classified as "food poor" (or comes from households with total expenditure that is insufficient to meet minimum caloric req~irements).~~ Government will therefore need to continue to play a strong role in ensuring that poorer households have access to health care and that resources are allocated where they are most needed. 2.38 Education. Efficiency improvements in school construction have been realized through decentralized contracting. The Government also plans to have around 30 percent o f the required classroom construction based on off-budget activities, through local communities and faith-based organizations. In addition, there i s a high level of involvement in parent-teacher associations (PTAs), which gives parents a voice in how capitation grants are used at the local level. The involvement of parents has also led to more transparent exams, with success dependent on performance. Further strengthening o f PTAs will help to ensure the efficient use o f scarce resources. Finally, the Ministry o f Education (MINEDUC)will continue contracting with private schools that have capacity to accommodate children intronc commun26,and to subsidize the cost for such students by offering the same amount of capitation grant as that given to public students. 2.39 Infrastructure. The share o f the private sector in infrastructure spending is relatively small, particularly when compared to other countries. Overall private investment i s estimated at about 0.5 percent o f GDP (annual average for 2001-2005), with average investment per project amountingto US$30 million. 2.40 Government has taken steps to increase private sector participation in certain areas:' particularly telecommunications, by: (i)mobilizing private sector capital and management skills; and (ii)creating the legal and regulatory basis for rapid market expansion through the entry of additional telecommunication service providers. Since reforms started later inRwanda relativeto comparable countries, however, its telecommunications sector has not kept pace with those of other countries. 2.41 Opportunitiesalso exist for privatesector investment inthe electricitysector, and to a lesser extent inwater supply and transport Inthe electricity sector, although the monopoly heldby Electrogaz has been formally abolished, major legal and regulatory reforms have not yet been enacted. The current energy crisis offers opportunities for private participation, particularly through the Lake Kivu Methane project, but also through the development o f additional private hydro and other generation facilities. In urban water supply, also managed by Electrogaz, private participation will also dependon the implementation of legal and regulatory reforms. 2.42 In the case of decentralized water supply and sanitation outside urban areas, a policy framework aiming at private sector involvement has been formulated, but very few (small) privateor NGOschemes have been implemented.While the sector offers opportunities 25Ahobamuteze, Dom, Harvey, and Purcell, Evaluationof generalbudget support, RwandaCountryReport (draft), November2004. 26 Troc communrefers to the first 3 year cycle of secondary schoolwhich caters to children fromthe ages of 13 to 16years of age. 27 A detailed assessment o f opportunitiesfor privateparticipationin infrastructureis providedinPrivate Solutionsfor Infrastructure in Rwanda, A Country Framework Report, PPIAF and the World Bank 2005. Eventhough that report is basedon informationobtainedup to 2004, most of its conclusions and recommendationsremainvalid. 18 for private sector participation, the scope for private funding i s limited, as it i s in other low- income countries, by cost recovery constraints. Intransport, experience from other countries that have adopted performance-based contracting indicates that this would yield the greatest efficiency gains. However, the scope for generating additional private sector funding for transport through this mechanism i s limited. There would be greater scope for private participation in airport infrastructure and aviation, as well as inrailways, ifa new railway connection were found to be feasible. 2.3 CONCLUSIONAND PRIORITY MEASURES 2.43 The macroeconomic management of aid will require continued attention to monetary policy and greater focus on sales of foreign exchange to limit reserve money growth. Itwill also require greater efficiency in spending. Given the focus on growth, there will need to be increasedspendingto improve delivery of economic and infrastructure services. At the same time, however, the level of spending in the social sectors will need to be maintained. Improvedefficiency o f spending could help achieve this goal. 2.44 Development partners need to coordinate their focus on infrastructure to ensure that it is given high priority in public expenditures. Infrastructure expenditures also need to be more efficient to achieve better service provision in both urban and rural areas with a given level of resources. In addition, options for involving the private sector in the management, ownership, and financing o f infrastructure operations needto be explored. 2.45 Efficiencygains in spending can be realizedin the social sector. Inhealth, efficiency gains can be realized through increased coordination of disparate activities, and increased support for community-based delivery systems, and alignment o f donor funding with the long-term goals of patients and health providers. In education, efficiency gains can be realized through (i) increased attention to quality, particularly at the primary level; (ii)increased spending at the primary level on learning materials; (iii)increasing the availability o f qualified teachers; and (iv) the construction of more classrooms. At the same time, subsidies to higher education should be gradually reduced, or more focused on those subject areas where skilled labor is lacking in the country. 19 3. POTENTIALSOURCESOF GROWTHAND CONSTRAINTSTO2020 3.1 In order for growth to be broad based, without leading to increasinginequality, it will be important to ensure sustained growth in both the agriculture sector which employs over 90 percent of the population, as well as the non-agriculture sector. In the past year, there are indications of an increasing contribution to growth from the service sector. Improvement in the state of infrastructure will help to strengthen this trend, and promote increased agriculture production for the market. However, the development o f non-farm income sources in rural areas will be needed to develop other sources o f income that will support the movement o f labor out o f agriculture. The promotion o f agri-business and strengthening the links between agriculture production and manufacturing will facilitate this development. Inaddition, the discussion in Chapters 4 and 5 show that mining could become an important source of non- farm income. 3.2 In the short-term, there will be a need to promote labor intensive sectors and technologies to engage the excess supply of labor. Growth accounting analysis indicates that factor productivity analysis explained over 60 percent o f GDP growth in Rwanda, in the past decade. For the medium term, improvements in factor productivity will continue to be the main source o f growth. Inthis regard, long-term sustained growth will require increased investment in physical and human capital to maintain and build factor productivity. Although current investments inhuman capital have ledto substantial improvements inthe social sector indicators, the overall level or stock o f human capital is still low in comparison to SSA. For example, the overall transition rate from the primary level is 42 percent, which is below the average of 60 for SSA. In addition, low investment in machinery and physical capital, poor maintenance o f machinery and outdated equipment has contributed to the low capacity utilization o f manufacturing firms. Therefore, emphasis will need to be on generating increased investment in physical capital and infrastructure while maintaining and building on the progress made in the social sectors. 3.3 The analysis in this section prioritizes measures to improve productivity in agriculture and increased production for the market through building the capacity of farmers and rebuilding the extension system to the promote use of improved technologies and handling methods. The section highlights the importance of the non-agriculture sector for growth. Inthe case o f the private sector, measures to support capacity development, as well as investments to improve the state of infrastructure to encourage investment are needed. These are discussed in Chapter 4. The analysis shows that removing the infrastructure constraints to production are the key measures needed to promote export diversification, and scaling-up o f the potential and newly developingvalue chains. 3.1 GROWTHINAGRICULTURE -STRONGBUTINSUFFICIENTTO MEETDEMAND 3.4 Over the period 1998-2004, agriculture GDP grew at an average 6.39 percent, and accounted for just under 50 percent of GDP growth. This growth was driven mainly by food crops, which accounted for 93 percent o f growth in agriculture and 44 percent of overall GDP growth between 1998 and 2004. Export crops accounted for less than 1 percent of GDP growth, 20 and livestock for 1.4 percent. Sixty percent o f the growth was due to yield increases, and 40 percent to the expansion of land. 3.5 Despite strong growth, the level of agricultural production still falls short of domestic and export demand. Rwanda remains a food-deficit country that imports at least 130,000 tons of food items such as edible oil, wheat, sugar, rice, beans, maize, cooking banana, and dairy products. In addition, indicators o f food insecurity show that most Rwandans are permanently deficient incalories, proteins, and lipids. Nutrition indicators for children under-five show that over the period 2000 to 2005, while height for age has increased, weigh for height and weight for age have declined. Low productivity of food staples has also been strongly correlated with a rise in inflation (which, inturn, adversely affects food insecurity). Productionof coffee, tea and pyrethrum (Rwanda's main exports), is also insufficient to support demand for Rwanda's exports. 3.6 The large gap between actual and potential yields for most crops highlights the opportunitiesand potentialof the sector. As Table 3.1 shows, yields o f major food staples in Rwanda are on par with the average for Africa, but significantly below the World average. In addition, Figure 3.1 shows that for most crops there have been very small increases inyield over time; and, as will be shown below, yield gaps remain large. Table 3.1: Crop Yield Comparison, 1999-2003 (average) Yield (Mt/Ha) Rwanda Burundi Ethiopia Tanzania Uganda Afiica World average average Maize 0.8 1.1 1.8 1.6 1.8 1.3 4.4 Sorghum 0.9 1.2 1.3 1.1 1.5 0.8 1.3 Cassava 6.1 9.0 10.2 13.2 8.9 10.6 Sweet Potato 5.8 6.5 9.6 1.9 4.4 4.6 14.9 Potato 8.0 2.6 9.1 6.9 7.0 7.7 16.3 Plantains 6.5 5.2 16.0 2.2 5.9 5.6 6.3 Beans 0.7 0.9 0.6 0.7 0.7 0.6 0.7 Peas 0.5 0.7 0.7 0.4 0.6 0.7 1.7 Coffee 0.7 0.9 0.9 0.4 0.7 0.5 0.7 Tea 1.3 0.8 1.o 1.3 1.9 1.9 1.3 Source: Calculatedfiom FAOSTAT,2005. 3.2 RECENT SOURCES OF GROWTH,CONSTRAINTS, AND OPPORTUNITIES 3.7 Increased production has come in equal share from increased land area under cultivationand higheryields. The increaseincultivated land has been significant for maize and wheat, and smaller for coffee, soybeans, and potato. For vegetables and fruits, the increased production has come mainly from higher yields (Figure 3.1). 21 Figure 3.1: Average Increase in Cultivated Area and Yield, 2000-2005 (percent 30 ~.~-- .-_^__I - -- ---_ _-__ ~ --x ~ ~ _* _I^ ~ I I E3 Cultivated area Meld Source: WorldBank Agriculture Policy Note, 2006 and FAO. 3.8 Land productivity has been increasing since 1994 (Figure 3.2). However, an examination o f the trend in productivity changes suggests that the rate of increase is slowing. This highlightsthe need for increased use of inputs, not only to reach production potential, but also to mitigate the effects of soil miningand erosion. Figure 3.2: Trend in Cultivated Land Area and Land Productivity in Rwanda I -," 1970 1973 1976 1979 1982 1985 1988 1991 1994 1991 2000 2003 , , I Source: FAO, 2006. 3.9 Many of Rwanda's soils are naturally fertile, but their fertility has been severely eroded by poor land and soil management practices. Soil nutrient losses in Rwanda are among the highest on the continent. The International Fertilizer Development Center estimates that only 23.4 percent ofthe country's soils have small or no erosion risks, that 37.5 percent need adjustments before cultivation, and that 39.1 percent have a high level o f erosion risk. Water erosion causes an annual countrywide loss o f organic material of 945,200 tons, and also causes annual losses of 41,210 tons of nitrogen, 280 tons of phosphorus, and 3,055 tons o f potassium. Productivity i s also constrained by late rains and incidents of localized drought. Sustainable productivity gains therefore depend on substantial improvements in water management and irrigation. 22 3.10 Increased pressure on the land from population growth is also contributing to the trends in soil fertility. Increasingpopulation density has led to increased farming on marginal lands, and the failure to use inputs has had the effect o f literally mining the soil. Land i s being farmed much more intensively without being allowed to lie fallow so the soil can rejuvenate. Further, pasture and woodlands are beingcut down for cultivation. Finally, the fragmentation of landhas been associatedwith lower returns.28 3.11 Households with more land are better off. Table 3.2 shows that poverty rates are higher for households with smaller landholdings. Inaddition, households with more land produce more for the market, while those with less concentrate more on food staples. Table 3.2: Distribution of Land HoldingsAmong Rural HouseholdsinRwanda, 2006 Rural Rural Rural All Group 1 Group2 Group 3 Rural <0.3 ha 0.3 1.0 ha - > 1.0 ha Groups Share o f rural households (%) 39.4 31.7 25.8 96.9 Average householdsize (number ofpersons) 4.5 4.9 5.6 5.3 Total landholdings (`000 ha) 64 274 750 1,087 Share oftotal landholdings (%) 5.9 25.2 68.9 100 Landholdings per household(ha) 0.11 0.58 1.94 0.75 Landholdingsper capita (ha) 0.02 0.12 0.35 0.15 Povertyrate (YOhouseholds< poverty line) 74 67 54 66 Source: IFPRI calculationsfiom EICV data, 2006. 3.12 Improving productivity and mitigating the loss of soil nutrientswill require a significant increase in the use of inputs. The current level of input use is extremely low, particularly for poorer households, and varies by region (Table 3.3). Overall, just under 7 percent of all households use improved inputs.29Households inthe most well-off quintiles utilize significantly more inputs than the poorest households. Across regions, the areas that grow cash crops and where institutions and support are available to farmers are those that use the most inputs. For example, Gikongoro i s the tea and potato-growing area; despite relatively poorer soils than other areas, farmers in Gikongoro are productive because they have received support from various projects to access fertilizer and improved inputs. 3.13 The low level of input use is more evident when one looks at individual inputs, namely: fertilizer, pesticides, and seeds. In the case offertilizer, for example, the number o f households using fertilizer has decreased from 10.5 percent in 1990 to 8 percent in the second growing period of 2005. Fertilizer use varies by agro-ecological zone, land use patterns, and degree of integration of the crop into a commodity chain. Due to lack o f pasture, lack o f fallow land, and low income, only 18 percent of rural households own cattle, and 36 percent of households own sheep, two o fthe most important sources of organic manure. Low availability of '*WorldBank, ''EICV, 2006, Table 5, page 11. 2001. 23 organic fertilizer makes chemical fertilizer particularly important for improving land productivity. Yet, only an estimated 5.3 percent of households that purchased chemical fertilizer used it on at least one plot of land.30 Table 3.3: Use ofImprovedInputsby Provinceand Income Quintile, Rwanda Proportionof householdsreportingimprovedinput use (YO) 1st 2nd 3rd 4th 5th quintile quintile quintile quintile quintile Total Butare 0.8 5.0 7.8 11.8 19.4 7.0 Byumba 2.9 1.7 5.7 5.1 20.2 6.9 Cyangugu 2.7 4.3 7.7 19.7 13.6 9.0 Gikongoro 10.2 15.1 26.2 33.4 40.4 20.5 Gisenyi 0.0 0.7 4.3 15.0 24.8 9.8 Gitarama 1.1 5.9 12.2 4.0 9.2 6.6 Kibungo 0.0 0.0 0.0 3.3 4.0 1.8 Kibuye 1.8 0.0 2.4 1.o 3.2 1.5 KigaliNgali 0.0 1.8 2.7 8.9 18.4 4.3 Kigali Ville 0.0 0.0 0.0 8.3 7.4 5.6 Ruhengeri 1.3 6.8 8.3 5.6 15.5 6.3 Umutara 0.0 0.0 0.0 4.0 2.8 1.7 Total 2.1 4.0 7.0 9.2 14.3 6.9 Source: 2001 EICV, as cited in WorldBank, 2006. 3.14 Low fertilizer use by farmers arises from their lack of knowledge of appropriate application methods, lack of purchasing power, and lack of access to output markets. On the other hand, for crops suchas tea, Irishpotato (involcanic zones), and rice, vertical integration of the supply chain facilitates access to: (i) inputs (seeds, fertilizer, pesticides); (ii) extension; and (iii)seasonalcredit. Marketing is also ensured by factories (tea and rice) and cooperatives (potato). Table 3.4 reports experimental results o f fertilizer use taken from the quantitative analysis conducted by Kelly and Murekezi (2000) on fertilizer profitability. The results show that usedcorrectly, fertilizer can more than double the yields for most crops. 3.15 Pesticides are used by only 8 percent of household^.^^ Pesticides are used mainly in coffee production (to control the coffee bug) and Irishpotato (to control phytophtora). Fertilizers and pesticides are pre-financed by OCIR-caf6, cooperatives, and farmers' associations. The low level of use relates to farmers' poor knowledge o f pest and disease control methods, which is exacerbatedby the lack o f research on chemical pest control practices. Inaddition, difficulties in procuringpesticides (which are mainly imported) have ledto low availability inthe market. 3.16 The use of improved seeds is also insignificant. In2005, only 12 percent o f households used improved seeds, covering only 2 percent o f cultivated land. According to the 2005 Agricultural Survey, 90 percent of the seeds used for food crops are produced on the farm. In recent years, however, the use of improved seeds has been increasing for potato, rice and maize. 30 EICV, 2001. op. cit.. 3 1GovernmentofRwanda, Agriculture Survey, 2005. MINAGRI 24 Table 3.4: Fertilizer Response for Selected Crops Increase Increase as percent National average yield in yield of national yield (1998-01) mtha Percent mtfha Maize 0.7-2.6 89.0 330.4 0.79 Rice 1.O-1.7 57.2 -- 97.3 1.75 Sorghum 0.6-1.9 62.2 - 196.8 0.97 Potato 5.0-9.3 60.8 113.1 8.22 Sweetpotato 4.2-7.4 73.5 129.5 5.72 Beans 0.4-1.2 92.6 277.7 6.48 Soybeans 0.4-0.8 77.1 ---- 154.1 0.52 Vegetables (cabbage) 28.3 506.1 5.59 Source: Calculatedfrom Kelly and Murekezi, 2000. 3.17 The main reason for low use of improved seeds is low availability. Reform of the National Seed Service, and passage of a Seed Law in 2005, to encourage increased private investment, have produced only limitedresponsefrom the private sector. Reasonsinclude lack o f knowledge of methods for propagating improved seeds, the low level of demand for these seeds, and limitedextension activitiesto distribute seeds. There have beensome recent initiatives related to maize, sorghum, rice, wheat, and beans, as well as improved virus-resistant planting materials for potato, cassava and banana, but despite these efforts, overall seed production remains small. Table 3.5: ImDroved Seeds Production and Demand Coverage (YO).200-2005 Crops 2001 2002 2003 2004 2005 supply Cover Supply Cover Supply Cover Supply Cover Supply Cover (kg) (%) (kg) ("/.I (kg) (%) (kg) ("/I (kg) ("h) Sorghum 495 7.22 64 0.94 58 0.86 367 5.35 206 3 Maize 1292 11.3 363 3.18 1228 10.74 ,1324 11.58 1127 9.85 Wheat 111 1.02 54 0.5 25 0.24 0 0 50 0.47 Beans 432 0.47 856 0.93 707 0.77 503 0.55 521 0.57 Soybean 379 4.02 284 3.01 345 3.66 172 1.83 80 0.85 Potato 1036 0.08 1020 0.08 1258 0.11 617 0.05 1172 0.1 Source: Compiledfrom AFSR Stu@ "Demande Solvable et RentabilitC de la Production SemenciBre" - Janvier 2006, as cited in Aertseen et.al. 2006. 3.18 Finally, agriculture in Rwanda remains non-mechanized, with limited use of draught animals or machinery. Most cropping operations are conducted with hand tools such as hoes, shovels, and rakes.32As a result, the average productivity of labor i s about halfthe level for SSA. 3.19 I n the livestock sector, herds were generally decimated during the genocide, but have since recovered, and the numbersof cattle, sheep, and goats are now approximatelyat the pre-genocide levels. However, productivity levels remain low for all species due to the low level of adoption of new methods and technology. Government is promoting the intensification of production, particularly dairy production, through the introduction o f improved breeds and 32World Bank, 2006. 25 conversionto contained feedingmethods (as opposed to the current traditional extensive grazing method). 3.20 Followingthe loss of many senior scientificstaff during the genocide, the Institut des Sciences Agronorniques du Rwanda (ISAR) gradually rebuilt its capacity, but further strengtheningis needed. The new Strategic Plan for ISAR focuses on more applied research; i.e., participatory on-farm trials o f imported technologies. However, the lack o f trained scientists remains a problem, and I S A R continues to operate with the help o f foreign experts while a cohort of Rwandanscientists is being trained outside the country. Further, responsibilitiesfor technology transfer have been devolved to the local level, and lack o f capacity and follow-up at the local level has ledto a gradual fading of public extension services.33 3.3 FUTURE DETERMINANTSGROWTH, OF OPPORTUNITIES,AND CONSTRAINTS INTHE SECTOR (2007-2020) 3.21 Future growthwill depend on policies gearedtoward transforming the sector to one that is market oriented. Because Rwanda is landlocked, transformation o f the sector will require overcoming the issue of high transport costs, particularly given the state o f transport infrastructure. It will also require larger-scale production, through increased productivity and the promotion of value addition. Growth will therefore be driven by increasing the scale o f production through land expansion, improving productivity through increased intensification, and higher value additionthrough improved quality and increasingly processedagricultural products. Scaleof Production 3.22 There is still scope to expand the land frontier in Rwanda. Between 1971 and 2003, the rate of land expansion averaged 1 percent per year; between 1996 and 2003 it averaged 3.3 percent per year; and it now appears to be falling back to 1980s levels o fjust below one percent, which was the rate when population growth was at 2.9 percent. This i s similar to experience in East and South Asia in the 1960s and 1 9 7 0 For~area under cultivation, the rate o f growth fell ~ ~ ~ by 14percent between 1993 and 1994, and started to increaseagain in 1996. To maintainthe rate o f expansion, new technologies, particularly for marshland reclamation and improved use of inputs, will be critical. 3.23 While total area cultivatedis likely to rise at the margin, it will only do so for those crops that yield the highest returns, in the context of increased land values. This is because farmers will switch from less to more profitable crops. There are already trends to this effect for sweet potato, cassava, and beans. 3.24 The rate of increase in cultivated areas is not likely to keep pace with population growth, and will lead to a reduction in the per capita availability of land. This again highlightsthe need to increase the productivity of land through the use o f improved inputs. It also impliesthat non-farm sources of employment will be needed, and that attentionto population policy and rural-urban migration will be important. Current estimates are that by 2020, the 33World Bank,2006. 34World Bank,2006. 26 population will double to 16 million people. If Government succeeds in reaching the target o f reducing the population engaged in agriculture from 90 to 50 percent, the number o f people engaged in the sector would remain more or less at current levels. The recent adoption o f the Land Law in 2005 should help to increase the security o f tenure, which will improve long-term investment in land. The Land Law will also facilitate the development of a rural land market, which would enable households to sell, lease, or rent out their land, to promote the movement of labor. Overall, measures focused on improving infrastructure and economic services in rural areas, and generating non-farm employment, will be o fparamount importance. Increased Productivity 3.25 The Government has set targets for short and medium-term productivity growth rangingfrom 3 to 5 percent per year for most crops. An in-depth assessment o f the technical and economic feasibility o f these targets35indicates that they are achievable, provided that the issue of inputs is addressed. Value addition 3.26 There is limited scope for value addition in the short run. However, opportunities do exist in the medium to long term, particularly for market-oriented crops. These crops can be categorized into those that are domestically consumed (Irish potatoes, maize, wheat, rice, soybeans); new high-value crops (horticultural products, fruits, medicinal plants); and traditional export crops (coffee, tea, pyrethrum, and hides and skins). 3.27 There will be limited scope for the expansion of food staples as urbanization increaseswith populationgrowth, and greater demand for the more market-oriented crops leads to decliningprofitability of traditional and subsistence food crops. There will be more scope for increasedvalue added with the market-oriented crops, if issues relatedto quality and to the scale of processing plants are addressed. In the case of maize, for example, the two maize milling plants are operating below capacity, while demand (for both human and animal feed consumption) is likely to continue growing. Additional factors include the challenge of moving to a uniform variety of maize, to yield uniform quality in flour production. Post-harvest management is largely traditional and labor intensive, and fumigation i s poor, which affects the quality of production. Nevertheless, the high farm gate price of maize in Rwanda (US$55.6/ton compared to Uganda at US$15.7/ton) could be an additional source of incentive to farmers. Issues related to the capacity o f processing units also apply in the case of soybeans, while issues of quality dominate inthe case of wheat. 3.28 In the case of new high-value crops, a high enough scale and quality of production will need to be established before there can be substantial value addition. For example, in the case of passion fruit, the low level of production due to a disease outbreak is insufficient to supply the five existing processing plants. Other crops, such as macadamia, have a two-year gestation period before the trees begin to bear fruit. In general, these new high-value crops are targeted at niche markets in Europe and the Middle East, and production i s generally both capital and labor intensive. Successful development of value chains for these crops will require an 35World Bank, 2006. 27 integrated approach that focuses on production and process, transportation, and marketing through dedicated contracting arrangements. Giventhe highbarriers to entry, based on standards and quality, as well as price and perishability of horticulture products, a significant amount o f up- front investmentwill be needed.36 3.29 Nevertheless, opportunities exist, particularly with the newly emerging regional supermarkets across eastern and southern Africa. Rwanda could position itself to take advantage ofthese regional supermarkets. Opportunities also exist to export to European markets. However, certain key technical constraints would need to be addressed: (i)infrequent cargo flights with limited capacity and high transport cost; (ii)lack o f cold storage and handling facilities; (iii) lack of technical knowledge at all levels; (iv) absence of certification and quality control services and laboratories; (v) absence of specific inputsand cheap packaging.37 3.30 Opportunities also exist to improve value addition for hides and skins. The first stage involves increasing the proportion o f hides and skins exported in "wet blue" form. This will require strengtheningthe system for collecting unprocessed hides and skins, by buildingthe transport and logistics systems of tanneries and educating livestock producers about the improved handling of hides and skins at the farm level. The second stage involves increasing the quantity and quality o f finished leather goods. This would require the development of a leatherwork industry, which would likely require significant capital investments in facilities, training of skilled workers, and establishment o f links to international leather distribution networks. Achieving this latter stage would be a long-term endeavor. 3.4 AGRICULTURE'S POTENTIALCONTRIBUTIONTO GROWTH AND POVERTY REDUCTION 3.31 A multi-market model of the Rwanda economy was used to assess the potential contribution of agriculture to growth in relation to the non-agriculture sectors. The model captures the underlying distribution and structure o f the sector (see Volume 2), and the distributional implicationso f agricultural growth. 3.32 Four growth scenarioswere considered: (i) baseline scenario assumes that the level the of input use, institutional development, and research remains at current levels; (ii)the commodity-specific growth scenario assumes exogenous increases in the rate o f production growth for each commodity, in line with Government targets, with all other crops remaining at the baseline level of growth; (iii)in the agriculture-led growth scenario, the production of all crops i s at target growth levels until 2010, and then returns to historical levels; and (iv) the diversified growth scenario assumes that growth is strong in both the agriculture and non- agriculture sectors. 3.33 Under the agriculture-led growth scenario, agriculture GDP would grow at 6.09 percent annually between 2005 and 2015, or just under twice the level of growth in the baseline scenario. At these levels, total GDP grows at 5.04 percent, compared to 3.88 percent under the baseline scenario. 36 World Bank2006; pg. 52 and 53. 37 World Bank2006, and DTIS 2005. 28 3.34 However, higher growth and poverty reduction are achievable when growth is diversifiedand strong in both the agriculture and non-agriculturesectors. The simulations show that if, in addition to agricultural growth, there is also more than 6 percent annual growth in the non-agriculture sector through linkages with agriculture, then growth in agriculture further increases, to 6.17 percent, and the GDP growth rate rises to 6.24 percent. At this growth rate, per capita GDP increases by 3.44 percent annually, almost tripling the annual growth rate in the baseline. As a result, the national poverty rate falls to 42.4 percent, 17 percent lower than the rate in2005. Figure 3.3: GDP Growth Projections, 2005-15 I - -- B - 3.44 3.3t 3.49 GDP AgGDP NonAgGDP GDP pc AgGDP pc Non-AgGDP PC Source: Diao and Yu, 2006 (model simulation results). 3.35 Significant growth in agriculture will still be critical, over the period, in order to reach overallGDP growth targets. Projectionsunderthe model indicate that agriculturewould contribute to half of overall GDP growth (under the diversified growth scenario), and contribute 60 percent to overall poverty reduction. This contribution would be highest for the poorest households. The poverty rate for rural group 3 will be reduced by I S percentage points (from 52.5 percent in 2005 to 34.4 percent by 2015); while the poverty rate for rural group 2 will decline by close to 20 percentage points (from 66 percent in 2005 to 46.2 percent in 2015); and that of rural group 1 will be reduced by 16 percentage points (from 73 percent in 2005 to 57 percent in2015Q8. 3.5 BALANCEOFTRADE AGRICULTUREIN 3.36 The trade surplus in agriculture is projectedto increase, since agriculture imports are smaller than exportsin absoluteterms, and exports are projectedto increaseat a higher rate than imports. This increase is highest under the full growth scenario - a 6.17 percent 38These rural groups are defied according to the size of landholding (see Table 4.2): rural group 1 are households with landholdings less than 0.3 hectares; rural group 2 are households with landholdings between 0.3 and 1hectares; and rural group 3 are households with landholding greater than 1hectare. 29 growth in agriculture from meeting all crop-specific growth targets, coupled with 6.24 percent growth innon-agriculture. The simulated agriculture trade balance in the full growth scenario is estimated to be just over $47 million, or 3.5 percent times higher per year than the baseline scenario. 3.6 SOURCES OF GROWTHFORPOVERTYREDUCTION 3.37 The simulations show that staple food growth (cereals, root crops, oilseeds, and livestock) will continue to be more pro-poor, and will contribute to 40 percent of overall reductioninpovertybetween2005 and 2015. Foodstaples are projectedto account for roughly 35 percent of overall GDP growth, which will increase food availability. Growth in agriculture exports also contribute to poverty reduction, but more so for rural income group 3. The main benefit o fexport crops is that they earn foreign exchange. 3.38 Overall, food staple-led growth contributes most to poverty reduction, with a 1 percent increase in growth translatingto 1.3 percent reduction in poverty both nationally and in rural areas. This contrasts with export-led growth, which results in a 0.85 percent reduction inpoverty nationwide, and a 0.87 percent reduction inruralpoverty. 3.7 INCOMEINEQUALITY AND PROSPECTSFORPOVERTYREDUCTION 3.39 Poverty rates are inversely associatedwith cash crop productionand linkagesto the market economy. Poverty rates are projected to decline faster nationally than in rural areas (falling from 59.2 to 42.2 and from 64.5 to 46.6 percent, respectively). 3.40 Income growth is unequal between male and female-headed households. Income growthfor male-headed rural householdsaverages 6.37 percent per year, compared to 6.18 percent for female-headed rural households. Thus, the poverty rate for male-headed households was falls from 61.9 percent in2005 to 43.6 percent in2015; while the poverty rate for female-headed households falls from 71.4 percent in2005 to 54.7 percent in2015. 3.41 Poverty rates between households engaged in cash crop production and those that are exclusivelysubsistence orientedwill widen by 2015. The poverty rate for rural households that are subsistence oriented or not engaged in cash crop production falls modestly, from 88.5 percent in2005 to 78.3 percent by 2015, due to the slower rate of increase in income. Incontrast, income growth among rural households engaged in cash crop production (roughly 90 percent o f households) will average 6.33 percent per annum.39 Lack of opportunities to be involved in cash crop production determine the low income level and high poverty rates across non-cash crop or subsistence households, and also partially explains why these households do not benefit equally from rapid agricultural growth compared to households that produce cash crops. 39This contrastswith the average o f 6.01percent for households(estimated at 10 percent of the rural households)that engage exclusively insubsistenceproductionof food crops. Onthe other hand, the povertyrate for rural householdswith cash crop productionfalls to 43.6 percentby 2015, from 62.2 percent in2005. 30 3.7.1 The Target Growth Rate of 6 percentin Agriculture is Feasibleand Critical for PovertyReduction 3.42 Ifthe Government's cropspecificagricultural growthtargetsare achieved, Rwanda could achieve its 6 percent per year agricultural growth target. If accompanied by equally strong growth in the non-agriculture sector, significant reductions in poverty and growth in income could be achieved. A recent study4' notes that the agriculture growth target i s achievable for the following reasons: (i) yield gaps can be bridged through improved use o f inputs, and marshland reclamation is also providing an impetus to use improved inputs; (ii) the targeted yields (on which the simulations were based) are still modest by international standards; (iii) significant value can be realized for coffee and tea plants, based on a strategy o f targeting specialty markets, and can occur even with modest increases inexports of these products. 3.7.2 ReducingPoverty by Half by 2015 Will Require 8 PercentGrowth 3.43 Reducing poverty by half by 2015 is the target stated in the Millennium DevelopmentGoals. The feasibility o f this target is tested in the fifth scenario, which assumes much more rapid growth for most agricultural commodities, coupled with the assumption o f a higher growth rate inthe non-agricultural sector. Based on these assumptions, the model yields an annual growth rate for agriculture GDP o f 8.7 percent, and a growth rate o f 7.2 percent in non- agricultural GDP, between 2005 and 2015. At these growth rates, this poverty reduction target is feasible. 3.44 Even if the national poverty rate is halved by 2015, poverty reduction will still be modest among the poorest household groups. For these groups, more targeted growth and poverty reductionpolicies will be needed, to enablethem to share inthe growth gains. 3.45 Some priority issues need to be addressed in order to meet the growth and poverty reductiontargets for 2010: farmer productivity needs to be increased through improved input use; significant increases in value addition need to be developed; and non-farm income sources needto be generated, particularly for households with the smallest landholdings. 40World Bank,2006. 31 Figure 3.4: Poverty Reductionin the M D G Scenario (AgGDP gr: 9.0%; NonagGDP gr: 7.2%; GDP gr: 8.0%) ........................ National Rural Rural hh with Rural hh without Rural mala Rural female- Ruralgroup 1 Rural group2 Rural group 3 cash MPS cash crops headed headed Source: Diao and Yu, 2006 (model simulation results). 3.46 First, input use needs to be improved. As noted above, the major causes o f low input use are: (i)lack o f awareness or knowledge of appropriate methods o f application; (ii) low the availability of inputs; (iii)low purchasing power or lack of access to credit to purchase inputs; and (iv) the high cost o f transport. The evidence also shows that a well-defined, vertically integrated production system for certain crops facilitates access to credit for improved inputs. Therefore, measures to support the linkage of farmers to an identifiedvalue chain should help to improve their knowledge and access to finance. In addition, promoting the development o f an extension system that i s more participatory and private sector-led would increase the use of inputs. On the supply side, measures need to be adopted to strengthen private sector-led procurement and distribution systems for fertilizer and other improved inputs. 3.47 Second, measures are needed to promote value addition. The first step will be to invest in meetinggrades and standards for food safety and phytosanitary requirements along the supply chain. Specific investments are neededto improve quality assurance and related training and skills development. This is particularly important for coffee and tea, which are increasingly going to be targeted to specialty markets. In the case o f horticulture, an effective public quality certification system (accompanied by relevant support to producers) needs to be developed. This applies as well to the hides and skin sector, which also needs to finalize and implementits sector strategy. 3.48 Third, policies are needed to stimulate non-farm income. Given Rwanda's large population, non-farm income sources, in both rural and urban areas, will promote the absorption of labor outside of agriculture. In the short term, non-farm income sources can help to supplement farm income. Inthe long term, such sources will also help to reduce the share of the population engaged inagriculture production. 32 4. GROWTHPROSPECTSAND CONSTRAINTSINNON-AGRICULTURE 4.1 Growth prospectsand constraints in non-agriculturewill depend very much on the speed and degree to which the issueof highinfrastructurecosts (particularlyelectricityand transport), access to finance and finance costs, particularly for small enterprises, and low availability of skilled labor can be addressed. The issue o f skills and cost of finance is most severe for smaller enterprises. In the case of larger manufacturing firms, the cost of electricity poses a greater constraint. Lowering the high cost of infrastructure services in electricity and transport is also a prerequisite for development ofthe tourism and ICT service sectors. 4.1 MICRO SMALL-SCALE ENTERPRISES AND 4.2 An assessment of the micro and small-scale enterprise (MSSE) sector completed in 2004 identified 70,000 formal and informal micro and small-scale enterprises, collectively employing more than 300,000 people:' Including mining workers, mobile venders, informal caterers, and other types of economic activity, the number of enterprises could be as high as 80,000. The units are both formal and informal, or in transition from informal to formal operation, and range from individual family-owned businesses to registered small companies. The small and medium enterprises (SMEs) as well as the large-scale enterprises are generally organized factory-type operations. These are distinguished by categories of enterprises, by the numbers employed, organization, system of production (factory-type or non-factory-type), value of turnover, and investments(Table 4.1). Table 4.1: Structure of the MSSE Sector, Classified by Rwandan System of Production Home-basedproduction1 Dispersedproduction -small scale Factory type - organized and structured Informal Informal/ formal enterprises - intransition to Formal- organized enterprises small and medium Family production Micro entreprises I Small entreprises Medium-scale I Large-scale Estimated 700,000 800,000 MSSEs - Estimated 200 SMEs IEstimated 30 - 50 firms I Artisans (i.e. basketry), Organized artisans, Organized operations Organized, structured, formal enterprises; home-based food formal and informal infactory-type processing, banana beer, units structure Registered Family labor Fewerthan 10 10-30employees 30-100 100 or more employees employees employees 41Nugawela, P., V. Bicali, V. Ndajigimana. 2004. Review and Assessment ofMicro and Small Enterprises (MSSEs) inRwanda. Backgroundpaper to the CEM, World Bank. 33 access to electricity, 56 percent have a water source nearby, and just 42 percent have fixed telephone lines, although 54 percent do have access to cell phones. 4.4 Indicationsare that the MSSEs, particularlyinthe informalsector, are growingand that there is potentialfor these to impact growth and poverty reduction. As shown inTable 4.2 below, the informal sector had the highest rate of growth between 1995 and 2003. The data also show that there has been an increase in the share o f the informal sector in GDP over time, and that next to the non-market component of GDP, it has contributed the most to GDP growth. However, it is important to note that the table below also includes informal activities in the agriculture or farm sector, which accounted for close to halfof the informal activities in 1995 and by 2003 accounted for 32 percent o f all informal activities in GDP. Most of the increase in informal activity has been in the services sector, followed by the construction sector, both o f whose shares doubled between 1995 and 2003 (from 5 to 10.8 percent and 2.1 to 4.6 percent respectively). Thesetrends highlight the potential o fthe service sector. Table 4.2: The Structureof GDPAccording to Degree of Informality, 1995 2003 - GDP Estimate Share in Total Annual Percentage GDP Growth Rate Contributionto Total Growth 1995 2003 1995 2003 1995-2003 Total 326.0 641.4 100 100 8.8 8.8 I Non-market 130.6 254.3 40.1 39.6 8.7 3.5 Informal 77.4 202.0 23.7 31.5 12.7 3.5 Formal 111.0 173.3 34.1 27.0 5.7 1.7 Indirect taxes 6.9 11.8 2.1 1.8 6.9 .14 Sources: MINECOFIN, national accounts, and World Bank staff estimates. 4.5 Among the MSSEs surveyed in 2004, about 69 percent were created after 1995, and 14 percent were created between the period 1985 and 1994. Results from a more recent survey completed inJune 2006 estimated the average age o f enterprises surveyed to be 5.2 years, with formal enterprises one year younger, on average, than partially formal enterprise^^^. The study suggested that this result may indicate that enterprises tend to formalize as they become more certain oftheir cost base and profitability. 4.6 There are also indications that more productive firms are increasingly being created, as opposed to retail firms. Most of the artisansal, agro-based, and food-processing enterprises were created recently. Private consultancy companies are also a recent development. A large share of these newer enterprises is located in Kigali City (11 percent), Umutara (11 percent), and Gitarama (7 percent), with the rest evenly distributed across the remaining regions. However, many of these MSSEs lack innovative ideas, and produce the same products or services that have already saturatedthe market. This situation is common in all provinces except for some new investment areas inKigali City, Umutara, and a few urban centers. 4.7 The 2004 assessment of MSSEs also identified potential new investment opportunities. These include fish farming, new agri-businesses related to processing, 42The Foreign InvestmentAdvisory Service (FIAS). 2006. "Government o fRwanda Survey o f Informal Economic Activity inRwanda." International Finance Corporation and the World Bank. 34 consultancy services, training centers, and new handicraft sectors, new products in the furniture industry(cane and bamboo furniture), recycling, and new service sectors inthe provinces. These needto be supportedwith training. 4.8 Earnings in the informal sector are higher than in subsistence agriculture, where the daily rate was estimated at FRw 300 per day in 2004. Estimates are that 75 percent o f permanent personnel in the informal sector earn around FRw15,OOO per month, or FRw 500 per day. The 2004 assessment also estimated that roughly 73 percent o f the firms had a turnover of less than FRw 500,000 per month, and that 36 percent had a turn-over o f between FRw 100,000 and 500,000. 4.9 Although firms operatingin the informal sector provide employment opportunities for the poorest, higher growth and poverty reduction outcomes can be realized by promotingtheir transition into the formal sector. Firms operating in the informal sector pay lower wages than those in the formal sector; and their employees have little job security and are not covered by the social security system. Informality also limits investment, lowers participation in public incentive schemes and public procurement, lowers productivity, and reduces access to capital and finance, all resultinginhigher costs o f doing business. 4.10 The 2006 assessment found that gender and level of education are strongly correlated with the degree of informality, as is the geographical scope of the enterprise. Women are more represented in the informal enterprises (58 percent o f women compared to 42 percent of men). Inpartially formal enterprises, 60 percent were owned by men, compared to 40 percent by women. This result is supported by data from the 2001 EICV, which found women to be overly represented inthe informal service sector. Interms o f education, 22.2 percent of those in formal or partially formal enterprises had post-primary education, compared to 12 percent in informal enterprises. In terms of geographic scope, 34 percent of enterprises operating on a provincial level or higher have an operator with primary education compared to 17.5 percent o f enterprises operating at the district or local level. These results indicate that informal firms are more likely to be very local in scope, and runby women with lower education. Moreover, the low level of education of MSSE owners and operators limit their ability to access information that could help them to expand their business. 4.11 Firms in the informal sector rank the lack of informationabout where and how to register almost as high as the tax burden imposed by registration as a barrier to doing business. Firms stay in the informal sector to avoid payment o f labor taxes and social security contributions, income or profit taxes, and compliance with labor regulations, while taking advantage of a more flexible work force. However, the actual financial burdenof complying with the taxes or social security contributions is less of a concern than the time and effort needed to find the relevant information about how to comply (cited by 36 percent ofpartially formal and 46 percent of informal enterprises). The reasons given by firms for not obtaining legal licenses including those needed to operate are primarily related to the high cost o f obtaining the licenses and the lack of information about what, how and where to go get the necessary licenses. In addition, firms also indicatedthat difficulty inmeeting the requirementsto get licensed, and weak government enforcement contributed to their decision to remain informal. 35 4.12 Unlike manufacturingfirms which cite the cost of electricity as the main constraint to business, MSSEs cite access to finance as a major disadvantage to informality and the most common constraint to business operations and growth. However, access to finance is not seen as a major consideration for becoming more formal. Firms in the informal sector have fewer assets than relatively more formal firms, which may also contribute to their lack of access to finance. Overall, for micro-enterprises that are either informal or partially formal, access to finance, transportation cost, macroeconomic instability, the cost o f financing, and access to land (in that order) are ranked as the highest constraints to business development and growth. Nevertheless, the FIAS study shows that those firms that cite electricity costs and cost o f finance as a major constraint also assessedthese to be extremely severe. 4.13 The most important reason cited for transitioning from informality to being partially formal was a realization that taxes were not as high as they thought and that administration was fair. However, 30 percent of the firms indicated that the major factor in their decision to transition to more formal operation was government pressure or the high cost of penalties for remaining informal. Another 20 percent cited access to information, together with a reduction in administrative burden, as the main factors in their decision. The 2006 study concluded that reducing the bureaucracy involved in becoming official could facilitate the transition of firms to the formal sector. 4.14 Many SME leaders felt that their staff were just barely competent enough to keep the company operating, but not trained to the level needed to develop a successful export business. Both technical and managerial skills needed to fill productive roles in export value chains are in short supply. S M E leaders are willing to pay a reasonable price for business services, particularly training, but currently there are a limited number o f options, and skilled technicians are in short supply.As a result, the private sector still imports skilled technicians from neighboring countries, despite Rwanda's high level of unemployment, and at a high cost to the economy. The lack o f access to facilities for technical and managerial training i s highlightedas a bindingconstraint to SME development in Rwanda. The Kigali Institute o f Science, Technology and Management (KIST) i s involved in training for SMEs, including support for young entrepreneurs, but the school i s based in Kigali and can cover only a small proportion of training needs across the country. Conclusionand Priority Measures 4.15 Increase the number of management training programs for private entrepreneurs and managers; and develop a clear policy for vocational and technical training to support MSSEgrowth. Actions inthese areas canyield immediate benefits for small enterprises. 4.16 Providestronger incentivesfor MSSEs to formalize and reduce the associated costs. Some measures could include: (i)providing information, training, and business development services; (ii)reassessing the social security contributions o f MSSEs; (iii)reducing the bureaucracy-related opportunity costs of becoming formal; (iv) simplifyingland registration; and (v) increasing access to finance and capital through the development o f different products, which could be delivered through the Union de Banques Populaire de Rwanda (UBPR). Despite the numerous microfinance organizations operating in Rwanda, only 55 percent o f SMEs have accessed micro credit. Multiple surveys have indicated that the high cost o f financing i s still a 36 constraint to firm-level investment and growth. The development o f alternative financing products geared to the needs of smaller and more informal firms, as has been done in Ghana (see Volume 2, Box 3.1) would helpto address this issue43. 4.17 Create incentives for increasing the quantity and quality of on-the-job training programs. Ina recent survey o fRwandan businesses, 72 percent said that they provided training at the place o f business, while only 28 percent said that they used external sources. About 50 to 70 percent of MSSEs in rural areas are organized into associations or economic interest groups, yet there still appear to be gaps inthe supporting industries and services for many businesses. Of the associations that are formed, the majority do not engage in long-term planning or are ineffective in serving members. In order to fully integrate MSSEs into export value chains, improved institutional support and business services are needed. Supplier inputs, quality control regulatory bodies, efficient transport and logistics services that can be accessed by MSSEs and networksto facilitate marketing activities are critical. 4.2 GROWTH PROSPECTSAND CONSTRAINTS INMANUFACTURING 4.18 The 2005 RwandaIndustrialand ManufacturingSurvey (RIMS) shows that growth prospects in the manufacturing sector depend on removingthe constraints posed by high costs of electricity,transport, and finance. The data show that productivity (measured both as output per worker and as value added per worker) i s within range for the region. Median value added per worker is estimated at US$2,910, which i s below the value for Kenya (US$3,457), but above the values for Tanzania (US$2,061) and Uganda (US$1085). Across industries inRwanda, the median value per worker is highest in construction and mining (US$6,094 and US$6,237 respectively), compared to the chemical industry (US$5,488), food and beverages (US$3,23 8), and furniture making (US$1,533) Productivity per worker inRwanda is also higher for exporters (US$9,800) than for non-exporters (US$1,700). 4.19 An examination of the determinants of productivity shows that a firm's stock of capital, the proportion of educated or trained workers, the level of education and experienceat the manageriallevel, and engagementin exportingall have a positiveeffect on firm productivity. The data indicate that strengtheningthe export orientation o f firms may be an important means o f stimulating growth inthe sector. 4.20 Firms in the manufacturing sector are already relatively export-oriented - 42 percent of firms surveyed export 36 percent of their output. A sectoral breakdown o f exporters reveals that food and beverage firms are more likely to be exporters than firms in other sectors; and firms inthis sector export a larger share o ftheir output than do firms inother sectors. About 61 percent o f all food and beverage firms are exporters - well above the average o f 42 percent for all firms inRwanda. This is followed by the miningand metal or furniture producers, of which about 38 percent are exporters. The average share of export value to total sales in Rwanda's food and beverage industry i s about 54 percent, compared with an overall average of 36 percent across all firms in Rwanda. About 20 percent o f firms in the chemical industryhave 43See www.newsroom.barclavs.co.uWcontent/detail.as~?ReleaseID=533&NewsAreaID=2 which outlines the fiamework and partnership between Barclays Bank, UK and local micro finance institutions in Ghana to deliver financial services to small and informal enterprises. 37 engaged inexport, and there is room for growth. There are also prospects for growth inthe fruits and vegetables sector (provided the issues related to low agriculture productivity discussed in section 4 are addressed), as well as inwood working and artisanal activities. 4.21 There are particularly good prospects in the mining sector, which exports mainly to Indonesia and other parts of East Asia, including China. Given Asia's rapid growth, expanding the processing and export of minerals could potentially be a strong source o f growth for Rwanda. Although the mining sector i s small, it contributes significantly to GDP growth. Much of the sector's growth has been due to the rising price of coltan, a mineral widely used in cell phones and other electronic devices. Other ores are mined for use in construction. However, the industry is dominated by labor-intensive, small-scale artisans. In2004, an estimated 50,000 people were involved inmining. Growth inthis sector could be a significant source o f non-farm income for households inrural areas. 4.22 The Government has been working to attract foreign investment into the mining sector. Privatization of the state-owned mining company (REDEMI) is currently under review. The prospects for privatization have drawn interest from mining companies from Australia, the United States, and South Africa. However, it is important that Government not overlook the small-scale miner and the potentialbenefits o f small-scale miningrelatedtojob creation, off-farm employment, and poverty reduction. Table 4.3 shows the estimated numbers o f small-scale miners inkey comparator and competitive countries. Table 4.3: Employmentin Small-ScaleMiningFirms,RegionalComparison Country Numberemployed Rwanda 50,000 BurkinaFaso 100,000 -200,000 Ghana 200,000 Malawi 40,000 Mozambique 60,000 Tanzania 550,000 Zambia 30,000 Zimbabwe 350,000 Source:InternationalInstitutefor Environment andDevelopmentof Small-ScaleMining, andRwandaMinistry of Mining andLandAffairs (MINITERRE), 2003. 4.23 Through the privatization of REDEMI, the Government aims to identify foreign firms that are interested in also engaging the small-scale miner. Large-scale investors can play an important role in developing the sector and transferring knowledge by: (i) helpingsmall- scale miners obtain access to finance; (ii) improving health, safety, and working conditions and reducing exposure to harmful chemicals; (iii) providing training and transfer of skills to artisan miners, to improve productivity and reducethe price o fproduction. 4.24 The overall growth potential of the manufacturing sector, however, is constrained by the conditionof infrastructure (particularly roads, as discussed in chapters 3 and 6, and electricity); lack of access to finance; and low levels of skilled labor. Cross-country comparative data reveal that infrastructure-related problems, particularly electricity and transport issues, weigh more heavily on Rwandan firms than on firms in neighboring countries. Across all 38 surveyed firms, 81 percent said that electricity is their single biggest obstacle; for export firms, the figure was 85 percent. Losses due to unreliablepower supply averaged 10 percent of the value o f outputs for all firms, although some reported losses o f up to 85 percent. Duringthe fiscal year preceding the interview, which was 2003, Rwandan establishments reported an average of 23 power outages, with the outages lasting an average o f 19 hours. The problem i s more severe in Kigali due its higher level o f connectivity. Figure4.1: Infrastructure Indicators:Electricity 30 29 20 10 0 All Rwanda non-kigali Kigali B No.of outages e outage duration (hrs.) m output lost ( o h ) Source: RIMS, 2005. Figure4.2: Infrastructure:Electricityand Transport as Major Concerns (percent of firms reportingconcern) 90 1 81 Rwanda Tanzania Uganda China Electricity L9 Transportation Source: RIMS, 2005. 4.25 The majority of Rwandan businesses cope with power problems by using a generator. Generator ownership in Rwanda is surprisingly high compared to other countries. Close to 75 percent o f respondents reported owning or sharing a generator to cope with unreliable power supply from the public grid. For these firms, an average o f 45 percent o f their power comes from these generators. Due to costs o f fuel and upkeep, however, fewer generators are owned by smaller firms, domestically owned firms, and firms that do not participate in international markets (non-exporters). Overall, electricity shortages are estimated to increase the average cost o f operating a business by 16 percent. 4.26 The cost of finance (interest rate, fees, etc.) was identified as a constraint by 54 percentof firms surveyed, and as a major or severe obstacle to doing businessby 34 percent of firms. The percentage o f firms identifying cost o f finance as an obstacle was larger (i)for 39 firms in Kigali than for those outside o f Kigali; (ii) non-exporters than for exporters; and (iii) for for medium and small firms than for micro and large firms. 4.27 Smaller firms are more credit constrained than larger firms, given the need for collateralto obtain a loan. The data show that the majority o f loans to Rwandan establishments were provided against one or a combination o f the following types o f assets: land and buildings (reported by 86 percent of surveyed firms); immovable plant and machinery (58 percent); movable machinery and equipment, including vehicles (32 percent); other tangible assets such as accounts receivable and inventory (29 percent), personal assets o f the ownedmanager, such as hisher house (37 percent); and other types of assets (19 percent). The approximate average value ofcollateralrequiredto get a loaninRwandais about 105 percent of the loanvalue. 4.28 Loan terms for small enterprises may also contribute to their credit constraints. In 2004, the average annual interest rate on loans was a reported 17 percent, and the average term of the loan was 36 months. The corresponding average interest rates on loans to micro and small establishments were not much different - respectively 16.3 and 18 percent per year. Average terms, however, were much shorter - 18 months and 43 months for micro and small firms, compared with 51 months for larger firms. Longer-term financing is therefore a constraint for micro and small enterprises. Further, larger firms are charged an average o f 14 percent for overdraft facilities or lines of credits, while micro and small firms are charged an average of about 17 percent. 4.29 Generally, retained earnings account for 52 percent of total financing for both working capitaland new investments,and commercialbank loans account for 20 percentof financing. Other sources of finance include trade credits (2 percent), equity sales (8 percent), and informal sources (1.2 percent). For working capital, 53 percent o f financing comes from retained earnings, compared to 80 percent o f firms in Uganda and 46 percent o f firms in Kenya. For new investments, 51 percent of financing comes from retained earnings, compared to 71 percent in Uganda and 45 percent inKenya. 4.30 Firms identifiedthe tax rate as an important constraint, yet the marginal effective tax rate (METR) compares favorablywith countries inthe region, and are not likely to be a significant barrier to investment. MET& in Rwanda are generally lower than or roughly equivalent to those in the neighboring countries, except for taxes on inventories incountries with lower rates of inflation, and taxes on some investments in Zimbabwe. METRs on investments in inventories in Rwandawould be reduced dramatically iffirms used the last-in-first-out inventory accounting system.44 4.3 1 The availabilityof skilled labor is reportedto be a major constraintby 40 percentof firms. The high returns to education, by level completed (Le.18 percent increase in income from completion of primary school), arising from employment in the manufacturing sector45indicates that educatedlabor is scarce inRwanda. The increases by level are higherfor women (48 percent for primary education) than men (8 percent for primary education). Overall, women have, on average, ten years o f education compared with nine years for men. They also have comparable 44FIAS, 2005. 45RwandaIndustrialandManufacturing Survey, 2005. 40 years of experience inthe labor market and an equivalent number o f hours worked per week, on average, across all education levels, but their average earnings are well below the average for men. These results may arise from the fact that women tend to be concentrated in low-paying occupational and that very few have high-paying managerialpositions. Table 4.4: Average Education,ExDerience, and Weeklv Earnings in Rwanda Average Average work Union Average Hours Average weekly education Experience membership worked weekly earnings (years) (years) ("/.I (number) (FRw) Men 9.2 20.0 23.0% 42.1 42,956 Women 10.3 16.0 24.1% 41.3 22,239 Total 9.6 21.8 23.3% 41.9 37,665 Source: RwandaIndustrialandManufacturingSurvey (RIMS), 2005. 4.32 Women are disproportionately under-represented in managerial positions in manufacturingfirms. Women's share in the sample is about 30 percent, but they occupy only 17 percent of management positions. Onthe other hand, there are relatively more women inthe professional, services, health, and other non-production categories. Within the female sub-sample of workers, the largest share of women (19 percent) are concentrated in the productive occupations, followed by the technical and services categories (16 percent each). Only 2 percent of women are inthe management category, compared with 5 percent for men. 4.33 Providing education and training opportunities for workers at both the firm level and the national level - particularly for women - is important for skills development. An assessment of the skill mix of the labor force, in relation to the needs of the industrial sector, i s needed to ensure that skills of graduates, the school curriculum, and education policy are appropriate for supporting growth. Further, a national strategy to develop technical and vocational education i s needed as an important aspect ofthe overall improvement inquality and access to the broader educational system. 4.3 TOURISMAND INFORMATIONCOMMUNICATIONS TECHNOLOGY 4.34 Since the re-launchingof the tourism sector in 2002, the sector has achieved steady growth in arrivals and receipts. The number of tourist visits to the national parks increased from 1,663 in 1995 to a peak of 27,000 in 2004 (compared to the target of 20,000); and rose further, to 24,120, in2005 (compared to the target of25,000). These figuresrepresentthe number of visitors to the three national parks, not the total number o f tourist arrivals. In 2004, the immigration department recorded 35,000 international visitors, which i s below the 1984 high of 39,000. 4.35 In terms of tourism receipts, the targets for 2004 and 2005 have been surpassed. Rwanda's revenue from tourism amounted to US$17 million in2004 (compared to a target o f $12 million); and reached US$26 million in 2005 (compared to a target o f $22 million). Average spendingper tourist inRwanda in 2005 was US$485, compared with US$365 inKenya, US$868 inUganda, andUS$1,280 inTanzania. 46The 2001HouseholdLiving ConditionsSurvey (EICV) also found that women inRwandaare engaged more inlow-payingoccupations. 41 4.36 Progress made over the last two years indicates that the tourism goals set for 2010 will be achieved. The national tourism strategy has set three main goals for 2010 - increasing tourist arrivals from 27,000 in2004 to 70,000, increasing the number o f days per visitor from 4 to 7, and raising the amount spent per day per tourist from roughly $120 in 2004 to $200. Achievement o f these goals would increase overall earnings from the tourism sector from $26 million in2005 to close to US$lOO million in2010. Infact, the sector has the potential to exceed US$lOO million, providedthat the necessary investmentsare undertaken intime. 4.37 Considering that most of the investments planned for 2003-2010 have not yet materialized, the fact that the country has already surpassedits goalswith the current level of tourism infrastructure suggests that the targets can be revised upwards. Rwanda can attract more tourists and a higher level o f spending, which would make the tourism sector one of thetop foreign exchangeearners. 4.38 The Governmenttourism agency (ORTPN) and investment and promotion agency (RIEPA) have a key role to play in stimulating investment in tourism. The continued improvement o f the overall investment climate i s also critical. The Primate Discovery Tours are an example o f the kinds of coordination needed between the two agencies. The eight-day tours were designedinline with the sector development objective o f promoting ecotourism. The circuit -consistingofNyungweForest,LakeKivu,andParcNationaledesVolcans(PNV) -featuresa primate safari that includes the mountain gorillas o f PNV, 13 primate species inNyungwe Forest, and other unique fauna and flora. However, plans for the tour have not yet been operationalized, due to lack of adequate accommodation inNyungwe, lack of a boat on Lake Kivu, and the poor condition of the road from Cyangugu to Gisenyi (link between Nyungwe and PNV). This year, ORTPN secured funding from USAID in 2006, to implement the Nyungwe development plan. To be on track for meeting the targets in tourist arrivals and spending, it will be crucial that at least the minimumlevel o f investment-an accommodation inNyungwe and a boat on Lake Kivu -beinplacebyendof2007. A legalandregulatoryframework isalso neededto ensurethat zoning and development, along with the links to local communities, preserve the environment while also creatingthe high-levelnichetourism Rwanda is seeking. 4.39 In the case of ICT, the entry point for Rwanda would be lower in the value chain, while the countryinvestsin buildingthe neededinfrastructure.Rwandaneeds to aggressively improve the quality, availability, and affordability o f telecommunications infrastructure, and the electricity networks that power that infrastructure. To this end, the National Information and Communications Infrastructure (NICI) agency has a comprehensive plan to improve infrastructure, access, affordability, and standards; and to strengthen private sector participation and local contents. 4.40 Favorable policies and positive developmentsin recent years have reduced costs in the sector. Rwandatel, the old telephone and internet Government monopoly, was sold to TerraCom Communications in2005. TerraCom i s implementinga fiber optics network backbone with wireless local loop access to reach users in remote areas. It has already deployed a fiber optics network inside Kigali and is extending it outward.47 TerraCom has also started to offer ADSL (Asymmetric Digital Subscriber Line) service inselected parts ofKigali, with much lower 47 It now goes as far as Butare. 42 rates than RwandaTel used to offer. The basic residential package is 128 kbps at a flat rate of FRw 45,000 per month and initial installation fee of FRw 100,000.48The Rwanda Internet Exchange (RINEX) was created in June 2004 and allows Internet Service Providers (ISPs) to exchange domestic internet traffic without having to send data across multiple European and U.S. hubs. This will help increase broadband speed inside the country and reduce overall internet costs. 4.41 There have been other encouraging signs in ICT. For example, a Chinese telecom firm specializing in the production of low-cost handsets has reportedly agreed to invest US$2 million to set up an assembly plant at the proposed Kigali ICT Technology Park (techno-park):' The techno-park is expected to house software development, call centers, hardware assembly, and other ICT-related firms. There also are a number o f domestic software service providers inthe technology park that can cater to local needs. 4.42 To support the growth of ICT, broad-based measureswill be needed in the areas of education, science, and technology. Government can begin to explore areas for immediately improving ICT literacy and usage among the population. Small-scale local interventions with well-defined objectives can generate positive value for citizens. Although the benefits o f these interventions would be limited, they could serve as quick wins that lay the groundwork for projects with larger benefits. Many projects indeveloping countries are successfully implemented in an environment with constraints similar to those in Rwanda. Those projects seem to generate success by maintaining a level o f technical sophistication appropriate to the countries' limited infrastructure and technical capacity. The integration of ICT i s typically achieved by settingclear goals and then usingICT as one o f the ways to achieve these goals. For example, applications in rural areas could include (i) transmitting information on prices o f agricultural products in nearby markets; (ii) weather forecasting for fishermen; (iii) provision o f expert advice on agricultural the practices through an interactive system; (iv) interactive tools for e-commerce directed at farmers; and so forth. Results from field surveys in rural China note the importance o f integrating these ICT forms of communicationwith more traditional means. 4.4 CONCLUSIONAND PRIORITYMEASURES 4.43 The broad-based actions needed to stimulate investment and growth in the non- agriculturesector pertainto the issues of low skills of the labor force, the poor conditionof electricity and transport infrastructure, and the high cost of finance, particularly for smaller firms. Training in business development is particularly important for MSSEs. In manufacturing, increased investment to upgrade machinery and capital will be needed. For some sectors (tea factories, mining), this will involve privatization. Finally, the problem of low access to and high cost of finance may requirethe development of financial instrumentsthat are tailored to the specific needs of differenttypes o f firms. 4.44 In the case of ICT, the entry pointfor Rwandawould be in the segment lower in the value chain, such as product assembly for domestic or regional markets, while investments are made to ease the infrastructure constraints to developing higher-level ICT capacity. In the 48 TerraComCommunications: http://www.terracom.rw/ 49 "Rwanda to Make Mobiles," East African BusinessWeek, May 22, 2006. 43 tourism sector, a result-oriented plan i s needed that identifies short, medium, and long-term results, and the specific interventions and resources needed to achieve the desired goals. A legal and regulatory framework to ensure that development will be sustainable and engage communities, andwhich reflects the neededzoning requirements. Finally, improved coordination between government and non-government entities, including private sector operators, will be necessaryto ensure effective implementation. 44 5. EXPORT GROWTHAND DIVERSIFICATION 5.1 Analysis of the links between diversification(measured by the Herfindahl Index, HI) and growth in SSA indicates that diversificationhelped SSA to transition from 3-4 percent export growth rate in the 1980s to 9-10 percent in the 2000s. It follows that in addition to liberalization having a positive effect on export growth, policies that spur export diversification are also imp~rtant.~' Yet, Rwanda is far less diversified compared to neighboringcountriesthat are similar to Rwandaintopography, climate,and soil. Since the 1960s, the Herfindahl Indices (HI) inRwanda, Uganda and Burundihave ranged between 0.4 and 0.9, reflecting strongly concentrated export baskets that were dominated by coffee and ores. In contrast, Tanzania, the only coastal country inthe set and a coffee exporter, consistently enjoyed a well-diversified export basket with the HIinthe range o f 0.3 and 0.1. 5.2 The low level of export diversification in Rwanda and its neighbors arises from socio-political events, and limited investment toward diversification. Socio-political events eroded the export base in several countries, leaving little else than natural vegetation on the ground. Examples include Uganda during the Idi Amin era (1971-85), Rwanda during the genocide o f 1994, and Tanzania during Nyerere's socialist regime. The annihilation of the productive sectors and human capital duringthese periods o f social and political upheaval limited exports to sectors that survived the shocks, such as coffee that grew naturally or minerals that did not need new investments in Rwanda and DRC. Other reasons for weak diversification are economic policies and institutions, or, as in Rwanda, limited investment in improving economic and infrastructure services, which left the export mix5'to be determinedby natural comparative advantage. The trends inthe HIalso indicate a structural shift around 1995, when the momentum o f economic reforms inRwandaand its neighbors beganto stimulate export diversification. 5.3 In the past five years, the emergence of mineral products has led to significant export diversification in Rwanda. The predominance o f coffee in Rwanda's export mix diminished dramatically as the share of minerals, metals, and chemicals increased, from 6.5 percent in 1976-89 to 40 percent in 2003-04. Global demand was a key driver o f the diversification, signaling Rwanda's sensitivity to global forces and its diversification towards new export markets. The country's diversification from coffee to minerals has helped to reduce its vulnerability to coffee price shocks. Diversification has also introduced a potential new source of growth for at leastthe nextdecade, or as long as China continues to grow at double digit rates. 50 The Herfmdahl index i s used to measure the degree o f diversification. It is computed as the sum o f the N squared shares o f exports from each industryout o ftotal exports:, or HERFINDAHL= s," r=l where s, is the share oftotal exports attributed to the ith industry. 5' InUganda, for instance, the military regime o fthe IdiAmin era led to mass destruction o f the economic base, causing export concentration in coffee, the cropthat grows naturally inUganda. 45 5.1 PATTERNOFEXPORT DISCOVERIES -MANYWITHLITTLESCALING-UP 5.4 There have been many export discoveries, but at a low volume and scale?2 Export discoveries have been identified for high-value agricultural products (such as horticulture products), unprocessed commodity exports, and manufactured goods. Bananas have been the largest agricultural high-value discovery, increasing from $5000 to $75,000 between 1976-89 and 1995-04. Inthe case of hides and skins, which are locally producedfor leather exports, the export value increased from $23,000 in 1976-89 to $166,000 in 1995-2004, illustrating success inexport diversification along the supply chain, from raw hides and skins to a manufactured export. The textiles chain represents the most sophisticated discovery, involving cross-sectoral synergies. An example is the use of chemicals to produce fabrics made from yarn and imported cotton, and exports of lighttechnology-based garments and outer wear made from locally available materials. Some ofthese product lines have exports inexcess of $100,000. 5.5 The processingof mineralsand metals-related commoditiesinto mediumtechnology manufactures (metal scrap, salts, anthracites, aluminum alloys) has contributed to the increaseinexportvalues of non-ferrous ores and concentrates, from $2.5 millionduring the 1976-1989 period, to $12.1 million in 1995-2004. The large variety of export discoveries in resource-based minerals and metals suggests that there i s potential in this area that has been largely unexploited. The emergence o f new minerals such as coltan, used in the manufacture o f cellular phones, can be a good source of growth in the short to medium term. Other natural resourceswhose exports exceeded $100,000 are anthracite and aluminum and its alloys. 5.6 The small scale of discoveries highlights two crucial points: (i)the problem of commodity export diversification inRwanda is the problem o f scaling upthe diverse discoveries; and (ii) every export discovery, even ifrelatively small in scale, has been important. Eventhough two traditional products (coffee and tea) accounted for more than 51 percent o f total exports in 2004, the large number o f non-traditional export discoveries indicates that the dilemma of export diversification in Rwanda is not the lack o f new and viable products, but their small values. For example, the remaining 49 percent o f non-traditional exports comprised tin ores and concentrates and non-ferrous ores (40 percent), while all other products accounted for only 9 percent o f total exports. Rwanda's total exports in2004 were only US$97 million. 5.2 FACTORS CONSTRAINING THE SCALING UP OF EXPORTS 5.7 The scaling up of exports has been constrained by the low base from which the country is starting, and by barriers to trade that have significantly increased costs. The destruction of property and assets in 1994 put the country at a lower starting base, which was compounded by years o f neglect and lack of investment in and maintenance of economic services. For example, many coffee trees were destroyed, but investments to replace destroyed and low-producing trees, andto develop a stock o f improved varieties, were not made until2004. The destruction that took place also ledto declines inexports from other sectors. 52 Export discoveries are definedhere as products whose export value doubled betweenthe two periods studied, 1976-89 and 1995-04.Productswith near negligible values were excluded, giventhe small size of Rwanda's economy. Inthe literature, a cut-off point of $10,000 hasbeenusedfor larger economies. 46 5.8 The main barriersto trade have been energy, transport, and ICT. Energy costs have been a significant deterrent to the expansion of manufacturing and processing a~tivities.~~ 5.9 In the case of transport, reduced cost would increase access and significantly increasethe returnsto farmers. Measuresto improve roads and generate sufficient competition among transport providers should lead to lower transport costs and translate to higher returns for farmers. Discussions in Volume 2 o f this report also show that for the case o f coffee, improved access to rural roads would greatly increasefarmers' income. 5.10 The cost and reliabilityof using corridor routes to reach gateway ports is a severe impedimentto Rwanda's capacityto accessworld markets. The cost perton o ftransport from Kigali to Mombasa can be as much as 70 percent higher than between Kampala and Mombasa. Overall, transportation costs translate to a 30 to 50 percent increase in the cost o f trade. In addition, there are extreme delays inthe transit o f goods from Rwanda along the corridors. The average transit time between Kigali and Mombasa i s four weeks. The long delays arise from congestion inthe ports as well as procedural obstacles, which account for upto halfo f the time in transit. The rehabilitation of roads and adoption o f measures to reduce congestion and administrative procedures would substantially reduce transit times and costs. Other measures could include development of a transportation network with neighboring countries, which would lower cost and increaseefficiency by reducing border transactions. 5.11 For the transport of perishableand high-valuegoods such as horticulturalproducts, the lackof regularflights to major marketsand poor handlingfacilities are major obstacles. Problems with the screening o f airfreight can lead to serious delays; when a scanner breaks down due to electricity surges, a repair crew hasto be flown infrom South Africa. This leadsto manual inspection of cargo, which can mean days o f delays for a large container. There are also insufficient cold storage facilities at the airport. 5.12 ICT costs will be an important factor in providinghouseholdsand firms with access to market information,directcontact with buyers, and linkagesto financial institutionsand government. These links are particularly important for export diversification. Measures to improve access to low-cost ICT services will be important in determiningthe flow o f information into and out of rural areas, as well as the extent to which rural households can effectively participate intrade and shift into market-oriented and commercial activities. Such measures will depend on the availability of energy in rural areas and the development o f skills relevant to the sector. 5.13 Finally,the lackof research and extension posesa significant barrier to the adoption of new methods and approaches to increase agriculture production. Farmers will need advice, based on sound research, on ways to move into new initiatives that require new techniques and skills. In the case o f coffee, for example, roughly half a million farmers will require extensive training in all aspects o f production, from planting through care and maintenance to harvesting, handling, and transportation. 53Diagnostic Trade Integration Investment Survey (DTIS), 2005. See also section 5 of this volume. 47 5.3 SECTOR-SPECIFICSTRATEGIC FOCUSTO PROMOTEDIVERSIFICATION 5.14 Each sector or commodity chain would benefit from a sector-specific export promotion plan to identify and link directly to external buyers - thus bypassing local auction markets, such as those for coffee and tea. This approach has been adopted for coffee for Starbucks and handicrafts for Macy's, and could be exploredfor the main export discoveries - minerals, chemicals, textiles, pyrethrum, wood, livestock, hides and skins, flowers, and horticulturalproducts. 5.15 In the case of tea and coffee, productivitymust first be through improvedaccess to finance and inputs; dissemination of information; and establishment and enforcement of standards. For the coffee sector, measures are also neededto improve the capacity of washing stations, which often operate below capacity due at times to water and electricity shortages. Further, the current installed capacity of washing stations can process only half of the beans produced, and more washing stations will needto be built. 5.16 Increased productivity in the tea sector will require training (related to pruning, plucking,etc.), and a pricingand privatization policy that will provide the right incentives to farmers to improve quality. There is also a need to improve the business environment to encourage investments inplants operating below capacity. 5.17 In the case of horticulture,poor soils and disease remain a constraintto scaling-up. Measures will need to include support or extension to farmers in the appropriate application o f inputs. Capacity support would also be needed in handling and packaging to ensure consistent quality. Support to grower associations could greatly facilitate training and increased access to finance for farmers inthis sector. 5.18 In the hides and skins sector, a national strategy is needed to support production, and to invest in upgrading slaughterhouses and collection systems to ensure consistent quality. A key component o f the strategy would include training in handling, to ensure consistent quality. Organizing small leather producers into production units would also help in establishing a value chain and access to market information and finance. 5.19 In the mining sector, which is characterized by small producers, a number of actions can betaken to attract investment. First,organizing producers into cooperatives would enable them to take advantage o f economies of scale. It would also promote investment in the sector and facilitate the provision of training in modern mining and processing techniques. Facilitating the formation of partnerships between domestic producers and foreign investors would also help to modernize the sector, given the shortage of adequately trained workers. Forming public-private partnerships to support prospecting would improve the information base on the stock of mineral resources. Partnerships would also be o f benefit in establishing the profitability of manufactured metal products, and would mobilize private investment to rehabilitate dilapidated and outdated machinery. Improving the transport and electricity infrastructure is also important for attractinginvestment. 5.20 In the case of pyrethrumand chemical exports, Rwanda's pyrethrum concentrateis exported mainly to the U.S. under AGOA, and to Kenya and South Africa. Presently, 48 Sopyrwa, the pyrethrum producing company, exports crude pyrethrum extract to be refined abroad. Rwanda could also use its flowers for oil distillation or solvent extraction of other products, such as paprika. Expansion of the pyrethrum sub-sector i s constrained by a dormant refining unit - due partly to electricity supply - which, if operational, could allow Sopyrwa to complete the refiningprocess and capture significant additional value added. To partly address the electricity issue, a critical sector-specific action would be to introduce solar54dryers to dry flowers.55 Below-capacity production of pyrethrummeans that Rwanda i s not taking advantage o f high demand inthe world market and existingpreferential trade arrangements with the United States andthe European Union. 5.21 Exports from the textile sector comprise one of the larger product chains, with newly discovered exports of nearly US$ 500,000 in 2000-04. The potential of this sector for export diversification far exceeds that of other sectors, because (i) it produces manufactured goods for which there i s a stable and growing demand; (ii) it enjoys special preferences in the U.S. market; and (iii) growth i s not constrained by Rwanda's relatively less skilled labor force. its While export discoveries of almost US$500,000 in 2000-04 have triggered rapid growth in the sector, Rwanda i s not yet fully exploiting its textile and garments exports potential which would enable itto take full advantage of AGOA and scale up.56 5.22 One of the potentialexports being considered in Rwanda is silk textiles. The case of this nascent industrymay have lessons for other sectors. Silk is not native to Rwanda, but with the help of FA0 and in cooperation with the Institut des Sciences Agronomiques du Rwanda (ISAR), under the Ministry of Agric~lture,~'mulberry trees have been planted in the Butare region, as a first step toward silk production. A similar approach to developing other new commodities could be adopted for other export sectors. 5.23 I n 2000-04, Rwanda exported about US$270,000 worth of newly discovered second and third-tier manufactured wood products such as pulp, cellulose, and paper, most of which were either resource based or light technology manufactures. Exports o f furniture, a second-tier, low-tech product, have recently exceeded US$100,000 per year, signaling the potential for rapid growth in that sub-sector. While the sector is vertically diversified, however, scaling up is stalling. The low and unreliable supply of electricity is one important factor affecting the sector, an equally important factor is the availability o f skilled labor. 5.24 Handicrafts production and wickerwork exports are a perfect example of how, because of Rwanda's small size, a single firm's entry can turn into a large (for Rwanda) export industry. There has already been a move to pursue the U.S. market through the African Growth and Opportunity Act (AGOA). Until 2004, wicker basket and hat exports averaged US$24,000 per annum. Through support from a USAID project, one enterprising woman's contact with the Macy's department store causedthe industry's exports to increase to US$25,000 54This would eliminatethe needfor harvestingfuelwoodto fire the ovens, thus contributingto improved andsustainableenvironmentalmanagement. j5ADAR "Agribusiness DevelopmentAssistance" Task OrderNo. 807.2004 SecondQuarterProgress Report, Submittedto USAID by Chemonics,July 2004. j6"The AfricanGrowthand OpportunityAct: Achieving Successthroughthe AfricanGlobalCompetitive Initiative", June2006, www.agoa.gov 57"Rwanda: Diagnostic Trade-IntegratedStudy", November 1,2005 49 in 2005. Special public interventions to create networks and informational externalities have significantly increased exports and investments in new production centers. The recent development o f a handicraft production training program should help to further expand this promising export industry. There should also be effort put toward identifyingother markets both in the U.S. and Europe. The challenge will be to meet growing demand while maintaining quality. 5.25 Sector-specific strategies to promote export diversificationshould therefore aim to: (i) identify external markets for specific products; (ii) the organization of producers inthe support hides and skin sector and the mining sector, to facilitate access to finance and training, and the development o f standards or certification; (iii)implement measures to support increased productivity inspecific sectors. 5.4 CONCLUSIONSAND STRATEGICAPPROACH 5.26 First, eliminatethe barriersto trade. Foremost among these is the electricity problem, which constrains the capacity of processing plants. The recent investment inLake Kivu Methane gas development should help to reduce the cost and increase the reliability o f electricity in the longterm. Second, invest inrural roads to link rural households and small producers to markets. Third, identify commodity-specific constraints to productivity, to help develop the value chain and increase producers' access to training and finance. Fourth, facilitate and promote engagement o f the private sector. Fifth, provide assistance to producers' organizations, including training in the use of improved inputs and microfinance, to promote the transition to increased market production or new activities. There would need to be coordination across agencies, with primary responsibility for organizing agricultural producers coming from MINAGFU; and for others from MINICOM. RIEPA would havethe role ofpromotion outside the country. 5.27 Putting in place measures to link producers to local, national, regional, and international markets will require identifying viable export chains and organizing producers to link them to these value chains. Sectors with growth potential that can benefit from value chains include tourism, textiles and silk production, hides and skins, horticulture, fruit juice manufacturingand processing, furniture making, and mining. 50 6. CONCLUSIONAND PRIORITY MEASURESFOR SUSTAINED GROWTH 6.1 The analysis in this Country Economic Memorandumsuggests a number of priority actions for transforming the economy and sustaining growth. Given the country's high dependenceon aid, there needs to be a focus on improvingthe efficiency o f spending to yield the greatest benefits in terms o f growth. Spending over the past decade has focused on the social sectors, with limited spending on improving productivity in agriculture or rehabilitating and maintaining the country's infrastructure base. As a result, there has been substantial progress in the social indicators, but little progress in the delivery of economic and infrastructure services. To generate growth and poverty reduction, there will need to be increased focus on private investment and growth. Spending should go to those areas where absorption can take place quickly. In the short to medium term, this means a focus on improving productivity in the agriculture sector. Inthe mediumto longterm, this means increased capital investment to reduce the costsofenergy, water, andtransport, andto increasethe supply of skilled labor. 6.2 Annex 1outlines the specific measures prioritized in the CEM and discussed in more detailin Volume 2. Four mainPriority Measures are identifiedto support growth as follows: Priority Measure 1:Invest to ease the infrastructure constraints to growth and export diversijication; 0 Priority Measure 2: Put inplace supportivepoliciesfor improved spendingeflciency; Priority Measure 3: Transform the agriculture sector to be more market oriented; Priority Measure 4: Provide support and incentivesfor private sector development. Priority Measure 1: Invest to ease the infrastructure constraints to growth and export diversification 6.3 In the medium to long-term, investments in energy supply, water supply and irrigation,and transportwill be the most crucial to easing the infrastructureconstraints to sustained growth. The high transaction costs resulting from poor infrastructure have seriously undermined the productive capacity o f firms, and have limited market access for rural households. To address the immediate energy shortage, the Government has taken action to import diesel generators. Immediate action is also needed to elaborate a comprehensive strategy for developing alternate sources of energy, particularly for the rural sector. For water supply, a detailed analysis of fundamental water sector and operational issues i s needed before strategic investment plans can be defined. For ICT, producers can develop a working knowledge of ICT through small projects to support increased productivity, particularly inrural areas - for example, the development of applications for the transmission ofprice information, weather forecasting for fishermen, and expert Q&A interactive systems on agricultural practices. For transport, investments in infrastructure would help to increase the scale of exports by easing the supply constraints. In particular, there needs to be a systematic, planned approach to rehabilitation and maintenance of roads, with a focus on communal and non-classified roads and options for cost recovery. In addition, investmentsto improve standards and quality, and the construction of cold storage facilities at airports, would facilitate the development of perishable high-value exports such as horticulture products. 51 Priority Measure2: Put in place supporting policies to improveabsorptive capacity 6.4 Rwanda is poised to benefit from increased aid flows. The macroeconomic management of this aid to promote growth should focus on minimizing volatility in spending, re-examining the composition of spending, and improving its efficiency. To minimize volatility in spending, the Government should first aim to limit the non-aid deficit, to levels that will not cause shocks to spending. Second, the composition o f expenditure needs to be focused on improving the infrastructure network and the support system for agriculture, through investments in irrigation and transport. Further spending efficiency gains can be derived from strengthening community health delivery systems and re-examining the relative subsidies to primary versus tertiary education. These measures should also help to maintain the outcomes achieved in the social sectors. Improved investments in infrastructure, along with increased spending efficiency, should helpto increase private investment and the absorption o f aid. Inthe case of exchange rate management, the limited response of exports to changes in the exchange rate, coupled with the strong effect of prices and supply-side factors, highlightsthe need to focus on easing supply-side constraints. 6.5 Additional measures to support increased absorption include wide-scale skills development and formulationof a populationpolicy. The analysis inthe CEMhighlights the shortage o f technical and managerial skills to support business development. To address this issue, the Government could play a facilitating role in strengtheningpartnerships between higher education and the private sector, to ensure that the education system i s producing graduates that can meet the demand o f the economy. A country-wide skills assessment could be the basis o f a comprehensive approach to employment and skills development, including through vocational training. Sector-specific issues related to skills and capacity development are outlined under Priority Measure 4. Education of girls and women will be particularly important to improve their earning potential. This latter measure, coupled with other policies to reduce early marriage and fertility (such as the promotion of family planning), would help to reduce the population growth rate. Priority Measure3: Transform the agriculturesector to be more market oriented 6.6 In the short to mediumterm, the focuswill need to be on increasing productivityin agriculture. Between 80 and 90 percent of the population is engaged in the agriculture sector, which is currently the main source of growth and foreign exchange. It i s therefore important to increase the market orientation of the sector, and the degree o f value addition, in order to raise incomes and reduce poverty. Measures will be needed to raise awareness about the appropriate use of improved inputs. One approach would be to support the organization of farmer associations and cooperatives around clearly definedvalue chains. The Government could play a facilitating role in terms o f providing information, setting standards, and regulation. This would require investments in developing viable commodity chains (in horticulture, leather, or maize for domestic market) or strengthening existing ones (coffee, tea, pyrethrum). Investments in strengthening the extension system, and particularly increasing the number o f agronomists and veterinary services, will be important, as will investments inimproving infrastructure. Priority Measure4: Providesupport and incentivesto privatesector development 6.7 Excess unskilled labor in rural areas, coupled with the high degree of land fragmentation, means that there will be a need to pull labor out of agriculture. To achieve this, it will be particularly important to strengthen the links between the agriculture sector and those non-agriculture sectors that rely on agricultural inputs.Establishing those linkswill require 52 the relevant investmentsininfrastructure to support private investment. Itwill also require sector- relevant skills development to support the integration o f farmers and MSSE owners into market and export-oriented value chains (basket weavers, furniture makers). The tourism sector also provides an opportunity to develop non-farm sources o f income. However, to ensure the development of a high niche tourism market that i s environmentally sound and supports community development, a comprehensive development plan i s needed that specifies standards, regulations, and zoning. 6.8 The problem of low availability of skilled labor cuts across all types and levels of firms. Inthe case of micro and small enterprises, the results from the CEM indicatethat owners of these firms are less likely to be educated, which constrains their ability to access information. This finding is supported by both the 2001 household consumption survey (EICV 1) and the assessment o f the micro-enterprise sector, discussed in section 4. Training support for business development and to facilitate access to information should help to reduce some o f the constraints these firms face. Training could also be used as a means to provide incentives for firms to move to the formal sector. Many firms remain informal due to lack o f knowledge or information on the requirements for formal registration of businesses and property, or exporting, or taxes. Therefore, a first step would beto streamline registrationprocedures and provide information to firms. 6.9 The private sector, especially SMEs, do not have adequate management and technical skills to grow their businesses, and lack access to training facilities to progressively improve their skills. The Government, academia, and the private sector should collaborate in taking measures to raise the level of technical and management competence of Rwandan businesses. This should involve the combination o f full-time training, continuing and part-time education in academic tertiary institutions and vocational training centers, and occasional management and technical training workshops in specialized private and public institutions. The Kigali Institute o f Science and Technology (KIST) has initiatedvarious levels of full and part-time technical and management training programs that can serve as a model for other training institutions. It will be important for Government and private employers to create incentives for continuing education and on-the-job training programs. 6.10 Skills development is particularly important for competitiveness of the manufacturing,ICT, and tourism sectors. Inthe immediateterm, Government should explore options for supporting on-the-job training and sector-relevant skills development. In the medium to long term, the Government needs to adopt policies to improve primary completion rates, to facilitate increased enrollment and completion rates at the secondary school level. These measureswould be important for reducingthe difference inearnings between men and women. 6.11 Another key barrier, aside from poor infrastructureand lack of training, is lack of access to long-term finance by small firms. The main reason is the collateral requirements neededto gain access to finance. Measuresto support enterprises inregisteringproperty, and the development of financing instruments that are more suited to smaller firms, perhaps in collaboration with the Union Banque Populaire du Rwanda, would helpto ease these constraints. Support for integrating firms into export value chains would be another means of increasing their accessto finance. 53 -w 8L n d a 5 - a a 5 7. REFERENCES Aertssen, G., A. Mutijima, and S. Mbarubugeye. 2006. Rwanda Agricultural Policy Note, Background Study #1: Prospects for achieving sustainable agricultural growth in Rwanda.Unpublished. CORE International Incorporated. 2005. Rwanda Electricity and Water Tariff Analysis, November; www.coreintl.com, andUSAID. Diao, X and B. Y. 2006. Agricultural Growth and Poverty Reduction Options in Rwanda. Background note for CEM, and Background note for the Agriculture Policy Note. InternationalFoodPolicy ResearchInstitute. Fast Track Initiative. 2006. Republic of Rwanda Assessment of the Government's Education Strategyand Financial Framework, 2006-2015. September. Governmentof Rwanda. 2006. Scaling up to Achieve the Health MDGs inRwanda. Background paper for the HighLevel Forum Meeting. Tunis, June 12-13. Government of Rwanda. 2002a. IntegratedHouseholdLiving Conditions Survey, EICV (2000- 2001). Finalreport. MINECOFIN. Governmentof Rwanda. 2002b. Profile of Poverty in Rwanda. (An analysis based on the results of the HouseholdLivingCondition Survey-EICV-1999-200 1). MINECOFIN. Governmentof Rwanda.2001. Poverty ReductionStrategyPaper. MINECOFIN. Governmentof Rwanda.2000. Vision 2020. FinalReport.MINECFIN. IMF, 2005. "The Macroeconomics of Managing IncreasedAid Inflows: Experiences of Low- Income Countries and Policy Implications" Available at http://www.imf.org/external/np/pp/eng/2OO5/0808O5a.pdf. IMF, Staff reports, various years InternationalTelecommunicationUnion. Kalk, A., J.K. Mayindo, L. Musango, and G. Foulon. 2005. Paying for Health in two Rwandan Provinces: FinancialFlows and Flaws, Volume 10, no. 9:pp. 872-78, September. Kelly, V.A., E. Mpyisi, A. Murekezi, and D. Neven, with assistance from E. Shin giro. 2001 Fertilizer Consumption in Rwanda: Past Trends, Future Potential, and Determinants. Paper preparedfor the Policy Workshop on Fertilizer Use and Marketing, organized by MINAGRI and USAID,Rwanda, 22-23 February. PPIAF and the World Bank. 2005. Private Solutions for Infrastructure in Rwanda: A Country FrameworkReport. W,1999. Report ofthe IndependentInquiry into the Actions ofthe UnitedNations duringthe 1994 GenocideinRwanda. UnitedNations, December. 1 Von Klaudy, S. and Benitez, D. 2006. Rwanda Infrastructure Public Expenditure Efficiency - Contribution to the CEM. December. WorldBank. 2006. Fiscal Space inRwanda-Health Study. World Bank. 2006. Fiscal Costs of Infrastructure Provision in Africa: Creating a Baseline - Rwanda, AFTPI. WorldBank. 2006. RwandaAgriculturalPolicy Note, December. WorldBank 2005. Diagnostic Trade Integration Study. WorldBank. 2005. Financial Sector Assessment Program. 2 3