Report No. 37860-RW Rwanda Toward Sustained Growth and Competitiveness (In Two Volumes) Volume II: Main Report October 12, 2007 Poverty Reduction and Economic Management 3 AFCC2 Africa Region Document of the World Bank G2C Government-to-Citizen G2G Government-to-Government GDP Gross Domestic Product GoR Government o f Rwanda GNFS Goods and Non Factor Services GNI Gross NationalInvestment GNP Gross NationalProduct HCAP Labor Force Hv HighValue HI Herfindahl Index ICT InformationCommunication Telecommunication IDA InternationalDevelopmentAssociation IMF InternationalMonetary Fund IFMIS Integrated FinancialManagement System ISAR NationalAgricultural ResearchInstitute ISDN Integrated Services DigitalNetwork ISPS Internet Service Providers ITU InternationalTelecommunications Unit LT Low Technology M 2 NarrowMoney MDG MillenniumDevelopmentGoals MINAGRI MinistryofAgriculture MINECOFIN MinistryofFinance andEconomicPlanning MINICOM MinistryofCommerce MFI Microfinance Institutions MSSE Micro and Small Scale Enterprises MT MediumTechnology NBR National Bank of Rwanda NGO Non-Governmental Organization NICI National Information and Communication Infrastructure NIS National Instituteof Statsitics NUR National University of Rwanda NTBs National Tender Board ODA Official DevelopmentAssistance OCIR-CafC Office des Cultures IndustriellesRwandais-CafC OCIR-ThC Office des Cultures Industrielles Rwandais-ThC ORTPN RwandanOffice of Tourism andNational Parks PDL HIM0 Programme de dkveloppement local a haute intensite de main- d'ceuvre, or Labor-IntensiveLocalDevelopment Program PPPS Public-Private Partnerships PIL Partnership inLearning PRS Poverty Reduction Strategy PRSC Poverty Reduction Support Credit RB Resource Based PRSP Poverty Reduction Strategy Paper RIA ResearchICT Africa RITA RwandaInformation and Technology Authority R&D Researchand Development RER Rural Economy Review (Real Exchange Rate) R I M S RwandaIndustrial and MiningSurvey .. 11 RINEX Rwanda Internet Exchange SINELAC Societe Internationale d'Electricite des Grands Lacs SOPYRWA Insecticide ProcessingPlant SOWARTHE Tea Associations STABEX Import TenderingBoard TFP Total FactorProductivity TOT Terms of Trade UBPR Union de Banque Pop. de Rwanda UK-DFID UnitedKingdom Department for InternationalDevelopment UNDO UnitedNations Industrial Development Organization UTEXIRWA Textile Company VAT Value Added Tax VECM Vector Error CorrectionModel VPP Village Pay Phone VSAT Very Small Aperture Terminal Satellite WDI World Development Indicators ... 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 increase per 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 need to be inthe range of 6 to 7 percent over the long term for those targets to be realized. The principal sources of growth inthe medium-termwere 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 o f 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 maintainedaround 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 of 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, is one o fthe highest in Sub-SaharanAfrica (SSA). Since 2002, average growth rates in manufacturing and agriculture, however, have been below the projectedtargets, and improvements inpoverty have been marginal, due to a number o f factors: (i) lack o f investment in infrastructure during the recovery and stabilization phase, to complement the reforms undertakento improve the business environment; (ii) lack o f investments in capacity, 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 expected pace o f 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 'The Vision 2020 document was issuedby the Government ofRwandain2000. Governmentof Rwanda, 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 poised to 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 DFID inFY08. 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 inRwanda: A Country Framework Report, a fiscal baseline study on infrastructur+ 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 of two volumes. Volume 1 is targeted at policymakers and other development partners, and provides an overview o f the main issues, with suggested interventions in priority areas. Volume 2 provides more detailed and technical coverage o fthe issues outlined inVolume 1. Volume 1beginswith an overview, in Chapter 1,of reforms and policy issues to date. Chapter 2 then discusses the management of foreign aid revenues to achieve maximum impact. Chapter 3 examines the sources of growth and constraints inthe agriculture sector, and Chapter 4 does the same for the non-agriculture sectors. Chapter 5 provides an overview of issues involved inexport diversification and competitiveness. Chapter 6 concludes with an outline of prioritized measuresto sustain growth. :$>L c: I. .O'C Acting Country Director for Rwanda Africa Region ACKNOWLEDGEMENTS This report and the accompanying Volume 2 draw 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 of the National Institute of 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 Netherlands and Denmark. Volume 1and 2 ofthis report were produced under the overall guidance of Yvonne Tsikata (Sector Manager, AFTP3) and Chukwuma Obidegwu (Lead Economist, AFTP3). Conceptualization of 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 i s 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 o f aid), and with Jean Musoni Rutayisire, from the National Bank of Rwanda (sources o f inflation); Mireille Linjouom (impact of energy crisis on the economy); Professor Kigabo Rusuhunva Thoma of the Free University o f 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 of the RIMS),and with staff o f National Instituteo f Statistics (analysis and drafting o f the RIMSand the informal sector survey); Patrick Nugawela, with inputs from 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); Xinshen Diao and Bingxin Yu (multi-market model with focus on agriculture growth); Stephan von Klaudy and Daniel Benitez (Infrastructure spending efficiency); and Michael Morris and Liz Drake, who made available the results o f the background studies for the Agriculture PolicyNote, and commented on drafts ofthe agriculturechapter. 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 benefited from a Quality Enhancement Review panel chaired by Edgardo Favaro (PREMED), and consisting of panel members: StephenMink (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 Kristina Kostial (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. Comments received have beenusedto revise the report. Thanks are also due to Lucia Chuo and Paula White who processed the document, to Marie Uwanyanvaya, for logistical support, and DeborahDavis for editorial assistance.. vi TABLE CONTENTS OF 1. INTRODUCTION ............................................................................................................ 1 1.1 Objective. Scope and Structure of the CEM.......................................................... 3 2. SETTING THE STAGE-POLITICAL AND ECONOMIC CONTEXT ..................................... 5 2.1 The Political andHistorical Context ..................................................................... 2.2 Overview of Key ReformsUndertaken and Outcomes. 1994 .present................ 78 2.3 The Economic Context ........................................................................................ 10 2.3.1 Structure ofthe Economy....................................................................... 10 2.3.3 Fiscal Policy and the Compositionof Spendingto Support Growth...... 18 2.5 Summary.............................................................................................................. 30 3. GROWTH:SOURCES, CONSTRAINTSAND CHALLENGES .............................................. 32 3.2 Analysis ofthe Constraints to and Sources of Growth........................................ 33 3.3 Analysis ofthe Sources of Growth...................................................................... 58 3.4 Was Growth All Post-Conflict Bounce-Back, Or Did Government Policy Contribute? .......................................................................................................... 66 3.5 Conclusion........................................................................................................... 68 4. MANAGINGAID FLOWSFORGROWTHAND POVERTYREDUCTION .......70 4.1 4.2 Marginal increases inaid flows were mainly spent and not absorbed.................72 A Framework for AssessingAbsorption ofMarginal IncreasesinAid ..............71 4.3 Increased Aid Flow DidNot Translate Into Dutch Disease ................................ 76 4.4 Efficiency and Composition of Spendingto Improve Absorption ...................... 79 4.5 Looking Forward- StrengtheningInstitutional Arrangements To IncreaseThe 89 Conclusion and Policy Implications .................................................................... Efficiency OfAid ................................................................................................ 4.6 92 5. AGRICULTURE-THEDRIVEROF GROWTH AND POVERTY REDUCTION ...................94 5.1 Sources o f Growth inAgriculture ....................................................................... 95 5.2 Overview o f Constraints to Growth..................................................................... 99 5.3 Growth Prospects for Key Commodities........................................................... 103 5.3.1 Growth ofFood Crops.......................................................................... 104 110 5.4 Agriculture's PotentialContribution to Growth and Poverty Reduction - 5.3.3 Summary of Growth Prospects............................................................. Overview o f the Multi-Market model ............................................................... 110 5.4.1 Possible Terms-of-TradeEfSectfor Key Agricultural Products...........117 5.4.2 Agriculture Growth Needs ToIncrease by Almost 9 Percent a Year To Cut Poverty in Halfby 2015................................................................. 118 5.5 Summary and Conclusion.................................................................................. 119 6. NON-AGRICULTURESECTOR-THEPULLFACTOR ECONOMIC FOR GROWTH 121 ...... 6.1 Introduction ....................................................................................................... 121 6.2 Microenterprise Development and Informality ................................................. 122 6.3 Overall Business Environment.......................................................................... 129 7. COMMODITY EXPORT DIVERSIFICATION INRWANDA - ............................................. 160 7.5.1 TechnologicalContent of Export Discoveries...................................... 174 7.5.2 Market Failure is aKey Factor Constraining Scaling Up...................175 7.6 Sector-Specific Strategic Focus To Promote Diversification............................ 178 7.6.1 WhichSectorsHave the GreatestExport Potential?............................ 179 7.6.2 WhichSectorsShouldGovernmentSupport?....................................... 181 7.6.3 Sector-SpeciJicPolicy ObstaclesMay Be the Critical Constraint .......188 i 7.6.4 General Constraintsto Export Diversijkation .................................... 189 7.7 Conclusion......................................................................................................... 190 8. PRIORITIZATION OF CONSTRAINTS: WHAT I S THE MOSTEFFECTIVEOFAID FUNDS? ........................................................................................................................... USE 191 9 ANNEX 1: NEW ESTIMATES GDPREBASEDTO2001- OF .......................................... 198 10 .. ANNEX 2: OVERVIEW OF THEDETERMINANTS EXPORTS OF .................................. 201 11. ANNEX 3: ECONOMIC REFORMSAND TOTAL FACTOR PRODUCTIVITY ...................205 12 ANNEX 4: DETERMINANTS OFPRODUCTIVITY AND EMPLOYMENT .......................... 208 13.. ANNEX 5: EXPORTDIVERSIFICATION ......................................................................... 210 REFERENCES ............................................................................................................................... 218 LIST OF TABLES Table 1.1 GDP Growth2002-2005 ................................................................................................ 2 11 Table 2.2: Trend and composition o f GDP. 2002-2005 .............................................................. Table 2.1: Sectoral Shares inGDP. 1995-2004........................................................................... 12 Table 2.3: Revenue inselected Sub-SaharanAfrican countries .................................................. 15 Table 2.4: Selected MacroeconomicIndicators. 2005-2009 ....................................................... 17 Table 2.5: Rwanda: Priority Spending. 1998-2005 .................................................................... 19 Table 2.6: Factors Influencing Inflation inRwanda.................................................................... 23 Table 2.7: Inflationand Net Credit to Government. 1998-2005.................................................. 24 Table 3.1: GDP GrowthRate and Investment inRwanda.......................................................... 38 Table 3.2: Cost o f Starting a Business. 2006............................................................................... 39 Table 3.3: Comparisono f Tax Burden across Countries............................................................. 40 Table 3.4: Rwanda- Share o f ClassifiedRoads in Good............................................................ 41 Table 3.5: Extento f Classified RoadNetwork inRwanda.......................................................... 42 Table 3.7: Public PrivateInvestment Infrastructure- Rwanda and InternationalBenchmarks...48 Table 3.6: Basic indicators for Electricity (Electrogaz) .............................................................. 52 Table 3.8: Returns to Educationby Level (In percent) ............................................................... 55 Table 3.9: Select Social Indicators-Rwanda 55 Table 3.10: Regional Macroeconomic Indicators.......................................................................... Post-Conflict ....................................................... 57 Table 3.11: Sectoral Contributionsto GDP growth, 1998-2004 ................................................... 60 Table 3.12: Demand-Side Decomposition of Growth, 1998-2004 ................................................ 62 Table3.13: The Geographical Structure of RwandanExports....................................................... 63 Table 3.14: Decomposition o f Real GDP Growth inRwanda,...................................................... 65 Table 3.15: TFP and GDP Growth Rate, by Type o f Regime....................................................... 67 Table 4.1: Two Episodes of Aid Surges in Rwanda.................................................................... 73 Table 4.2: Trends inthe External Accounts, 1998-2005 (In unit indicated)............................... 75 Table 4.3: Change inComposition of Imports, 1995-2005 ......................................................... 76 76 Table 4.5 : International Benchmarkingof Infrastructure............................................................ Table 4.4: Trade and Exchange Rate Indicatorscovering aid surges over .................................. Table 4.6: Rwanda - On-and Off-Budget Public Expenditures, 2002-2005 ................................ 79 85 Table 4.7: Rwanda-Maintenance Spendingin Infrastructure ................................................... 86 Table 4.8: Rwanda- Quasi-fiscal deficit in SOEs 11(Percentage of Nominal GDP) ................87 Table 5.2: Fertilizer Consumption(`000) for Major Crops, 2000-2005 ...................................... Table 5.1: Improved Seeds Production and Demand Coverage, 2000-2005 ............................... 97 Table 5.3: Total Imports of Pesticides and Values (In TonesandBillion of RWJ ..................... 97 97 11 Table 5.4: Crop yield comparison. 1999-2003 average (Mt/Ha)............................................... Table 5.5: Proportion o f Loans taken by HouseholdQuintiles (Inpercent) ............................. 100 102 Table 5.6: Sources o f Household Income.................................................................................. 102 Table 5.7: Fertilizer responsefor selectedcrops (In unit as indicated)..................................... 103 Table 5.8: Yield Responseto Fertilizer usage and Improved Seeds,......................................... 103 Table 5.9: Rwandanexports of fruits, vegetables and flowers, ................................................. 106 Table 5.10: Traditional agricultural exports (In million USD)................................................... Table 5.11:Rwandan coffee production, 2000-2005 (In unit indicated) .................................... 107 108 Table 5.12: Performance ofthe tea sector, 2000-2005 ................................................................ Table 5.13: Number of Animals, Hides and skins, 2000-2005 (In tons) ..................................... 108 Table 5.14: HouseholdAggregation by Landholding and Region.............................................. 110 Table 5.15: Crop distribution by household groups .................................................................... 111 Table 5.17: Level of initialyield and area, and annual growth rates used inthe model..............112 Table 5.16: Distribution of households growing coffee and tea by household groups................ 112 113 Table 5.18: Income growth and poverty reduction inthe model................................................. 114 Table 5.19: Sources o f Income Growth and Poverty Reduction inthe Model............................ 115 Table 5.20: Simulated Imports o fAgricultural Products (In unit indicated)............................... Table 7.1: Fixed Effects -Export GrowthModel for the Sub-Region, 1975-2004 ..................116 Table 5.21:Poverty Reduction-growth elasticity - PercentageReduction inPoverty Rate......115 Table 7.2: FixedEffect Model-Pattern o f Export Diversification inAfrican Countries,........ 164 166 Table 7.3: DifferentMeasures o f Export Diversification in Six Countries ............................... 169 Table 7.4: Growing Exports and New Discoveries inRwanda, 1976-2004.............................. 175 Table 7.5 : Diminishingand Disappearing Exports inRwanda, 1976-2004 .............................. 176 Table 7.6: Rwanda's Export Destinations, 1978-2004.............................................................. 179 Table 8.1: Priority Interventions for SustainedEconomic Growth ........................................... 195 LISTOFFIGURES Figure 1.1. GDP and GDP Per capita Growth. 1995-2005............................................................ 2 Figure 2.1: Historical Trend inthe Components of GDP. 1970 .2004....................................... 11 13 Figure2.3: Volatility inthe MainComponents o fthe Balance................................................... Figure 2.2: Terms o f Trade. Export volume and export value .................................................... 15 Figure2.4: The Savings-Investment Balance. 2004 .................................................................... 17 Figure 2.5: FDIis extremelyLow. even by................................................................................. 18 Figure2.6: Workers' Remittances exceed................................................................................... 18 19 Figure 2.8: Infant Mortality Rates inG-11 Countries. 2004 ...................................................... Figure 2.7: Reccurent and Development Expenditures. 1970-2004............................................ 20 Figure 2.9: Adult literacy rate in2004 and Trends inthe illiterate population ........................... 20 Figure 2.10: Trends inGrants. ForeignLoans............................................................................... . 22 22 Figure2.12: Rate o fInflation inRwanda. 1995-2005................................................................... Figure2.11: Total Aid Inflows 1995-2005.................................................................................. 24 Figure 2.13: Annual Percent Change inDomestic Credit. 1998-2005 .......................................... Figure 2.14: Actual and Equilibrium Exchange Rate -Comparison of Rwandaand Uganda......25 26 Figure2.15: Public Investmentis Definedby Availability ........................................................... 27 29 Figure 2.17: Public Investment. 1995-2005 .................................................................................. Figure 2.16: EducationalAttainment of 15-19 year olds Still....................................................... 30 Figure 2.18: Capital Formation Drivenby the Private Sector. 1995-2005.................................... 30 Figure3.1: Diagnostic DecisionTree -Constraints to Growth inRwanda ................................ 33 Figure3.2: Interest Rates inRwanda. 1999-2004 ....................................................................... 35 Figure 3.3: GDP Growth and Trends inCredit to the Economy ................................................. 38 Figure 3.4: Evolution of Charcoal Prices. 1975 -2004............................................................... 45 ... 111 Figure 3.5: Rwandaelectricity production. 1990-2004(in GWh)............................................... 45 Figure 3.6: Average Productivity inAgriculture Among G-11Countries .................................. 54 Figure 3.7: Average GDP per Worker amongthe G-11Countries ............................................. 54 Figure 3.8: 57 Figure 3.9: GrowthRate of Exportsto Asia andNeighboring African countries. ...................... GNIPer Capita inRwanda's Neighborhood. 2005 ................................................... 64 Figure 4.1: Rwanda's ReserveRatio andDomesticDebt 1995-2005......................................... 74 Figure 4.2: Rwanda's Monetary Indicators ................................................................................. 74 75 Figure 4.4: RwandaandNeighboringCountries .RealEffective............................................... Figure 4.3: Shares of Domestic Financing inGDP ..................................................................... Figure 4.5: Dynamics ofRERMisalignment.GDP per Capita and Export Growth...................77 Figure 4.6: Distribution of Donor Fundingby Strategic Objective............................................. 77 82 95 Figure 5.2: Increase inCultivated Area andYield. 2000-2005................................................... Figure 5.1: Trendsin Cultivated LandArea andLandProductivity ........................................... 96 Figure 5.3: Yield Trends inFoodCrops. ..................................................................................... 98 Figure 5.4: Yield Trends inthe MainExport Crops, ................................................................... 99 Figure 5.5: RelativePricesfor Selected Crops.......................................................................... 118 Figure 6.1: Sectoraldistribution of Manufacturing Firms (In Percent) .................................... Figure 5.7: Poverty reduction inthe MDG scenario.................................................................. Figure 5.6: RelativePricesfor Selected .................................................................................... 118 119 Figure 6.2: Distribution of Establishmentsby Legal Status and Size Group (Inpercent) ........132 133 Figure 6.3: MedianValue addedper worker in......................................................................... 134 Figure 6.5: CapacityUtilization-Rwanda and NeighboringCountries.................................... Figure 6.4: MedianValue added per worker- Rwandaand comparators (in USdollars) .........134 134 135 Figure 6.7: Percent of exporters and average exportsby sector ................................................ Figure 6.6: RegionalParticipation ininternationaltrade........................................................... 136 Figure 6.8: Direction ofTrade: Share of exports(by destination)............................................. 137 Figure 6.9: Ten Major Business Obstacles inRwanda.............................................................. 138 Figure 6.10: Major Obstacles. by Local andExport Status......................................................... . 139 Figure 6.11:Infrastructure indicators:Electricity........................................................................ 140 Figure 6.13: Infrastructure: Electricity andTransport as major concerns ................................... Figure 6.12: ComparingShare of Energy Cost ........................................................................... 140 140 Figure 6.14: Generatorownership by various establishment ...................................................... 141 Figure 6.15: Interest ratespaid (Inpercent) and terms................................................................ 142 Figure 6.16: Establishmentswith overdraft facility or bank loans.............................................. 142 Figure 6.17: Establishmentswith either or both overdraft .......................................................... 143 Figure 6.18: Average unusedshare of overdrafts or lines of....................................................... 143 Figure 6.19: RelativeCost ofthe MainConstraintsto DoingBusiness...................................... 144 Figure 6.20: Workers education: Share ofworkers by................................................................ 145 Figure 6.21:Availability of Skills and other Obstacles to........................................................... 146 Figure 6.22: Regulationissues as major concerns....................................................................... 147 Figure 6.23: Average Tourism SpendinginSelected.................................................................. 151 Figure 6.24: TourismReceipts inRwandain2004 and 2005, .................................................... 152 156 Figure 7.2: Export Diversification Dampensthe Volatility ofExport Growth ......................... Figure 7.1: Export Diversification Causes Export Growth in SSA ........................................... Figure 6.25: Cellular mobile subscribersin2005........................................................................ 162 165 Figure 7.3: Rwandaand Its LandlockedRegionalNeighbors................................................... 168 Figure 7.5: The Structure of Exports - Rwandaand NeighboringCountries, ........................... Figure 7.4: Rodrik-HausmanScore of Export Sophistication- Rwanda................................... 171 172 Figure 7.6: Sector and Technological Composition ofExports................................................. 174 iv LISTOFBOXES Box 1.1: Overview of SelectedKey Reforms since 1995 ............................................................. 4 Box 2.1: Overview ofEconomic and Structural Reforms. 1961-1990.......................................... 6 Box 3.1: A New FinancialInstrumentBuilds on Traditional Credit Savingsand Schemes in Ghana........................................................................................................................ 37 Box 3.2: The Impact ofthe Electricity Crisis onFirmCompetiveness....................................... 46 Box 4.1: Defining The Impact of anAid Surge........................................................................... 72 Box 4.2: Fromthe Priority Sectorsto Outcomes......................................................................... Box 6.1: Various Stages of Tourism.......................................................................................... 91 148 Box 7.1: The Case for RefinedTin Exports.............................................................................. 182 Box 7.2: RwandanHandicraftsExports.................................................................................... Box 7.3: ScalingUp andDiversification inCoffee................................................................... 185 187 Box 7.4: The African Growth and Opportunity Act .................................................................. 189 V 1. INTRODUCTION 1.1 In its Vision 2020 document, published in2000, the Government of Rwandaestablished targets for GDP growth and poverty reduction, to be achieved by the year 2020; these were to (i) raise real per capita income from $230 to $900; and (ii) reduce the poverty incidence by half. To reach these targets, the Government projected in its 2002 Poverty Reduction Strategy Paper (PRSP) that GDP growth would have to be in the range o f 6 to 7 percent over the medium term. Growth would come from the primary sector and manufacturing. Within the primary sector, the PRSP projectedthat growth would start at 5.2 percent and accelerate over the period (2002-2004), as the effect o f increased input use ledto improved soil productivity. Manufacturing growth was predicted to rise sharply, to 11.5 percent in 2002 and then slow to a more sustainable level of 7 percent. The projections were based on the expansion of manufacturing capacity in agro- processing, startingintobacco and brewingand expanding to other sectors. 1.2 The PRSP anticipated that growth in the agricultural sector would proceed with progressive commercialization, with ensuing demand for agricultural and non-agricultural goods and services inrural areas, resulting inincreasing non-farm employment. Over time, there would be diversification beyond primary commodities to more sustained engines of growth - agro- processing, garment exports, commercial and information and communications technology (ICT) services (including re-exports), tourism, mining, and export o f skills within the region. To provide an enabling environment for this growth, a number of areas were identifiedas priorities: investment inhuman development and economic infrastructure; promotion of private investment and exports; financial, commercial, and legal reforms; privatization of public enterprises; and specific measures to stimulate growth in key sectors, including mining, tourism, arts and craft activities, and ICT. 1.3 In line with this vision, the Government undertook reforms to liberalize the economy, improve the business environment, and invest in human capital. Trade and input subsidies were eliminated, and a floating exchange rate regime was established. There were extensive privatizations and establishment of a utilities regulatory agency (see Box 1.1). In addition, the Government put inplace measuresto improve social sector indicators ineducation and health. 1.4 As a result, real GDP growth averaged around 10 percent over the period 1995- 2005. Average inflation during this time was maintained below 10 percent, suggesting a real increase in average income over the period, with an average per capita income growth o f 5.4 percent. However, indications are that the impact on poverty reduction has been extremely modest. Overall the national poverty rate declined from 60.4 to 56.9 percent, and rural poverty declined from 66 to 62.5 percent. Nevertheless, there were substantial improvements inaccess to education and health services. For example, primary net enrollment increased from 73 to 85 percent, according to the 2005 household data. 1 Table 1.1 GDP Growth 2002-2005 (Inpercent) Average Average 2002 2003 2004 2005 2002-04 2002-05 PRS Scenario 1 1.3% 6.1% 6.2% none 7.0% nla Vision2020 6.9% 6.9% 6.9% 6.9% 6.9% 6.9% Actual 10.0% 0.9% 4.4% 6.0% 5.0% 5.6% Source: PRS, Vision 2020 and MINECOFINestimates 1.5 A comparison of growth performance with PRSP targets reveals that despite overall positive trends in GDP growth, and while targets in the social sectors were either met or surpassed, the expansion of manufacturing did not occur as projected, and growth in agriculture was, on average, below the projectedtargets. More recent projections completed in preparation for the Economic Development and Poverty Reduction Strategy (EDPRS) - Rwanda's second PRSP, to be published in 2007, and to cover the following 3 to 4 years- indicate that at the current level of growth, per capita GDP would needto double inorder to meet the growth targets, by the year 2020. This implies a per capita GDP growth rate of around 4 percent and real GDP growth in excess of 1percent per annum (assuming the population growth rate remains at 2.9 percent). Yet, as Figure 1.1 shows, GDP growth appears to be slowing down, after a period of rapid recovery following the genocide. Without a rapid implementation of productivity- enhancing measures, there is a risk that growth will stabilize at the 3 to 5 percent level that prevailed prior to the war, and which would be insufficient to meet the Vision 2020 GDP growth target. Figure1.1: GDP and GDP Per capita Growth, 1995-2005 (Inpercentage) 4c GDP growth (left axis) CDPper capita growth 25 25 -2c 20 -15 15 -1c 10 5 0 I 1995 19% 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: World Development Indicators, 2006. 1.6 Government has more recently focused on addressing the constraints related to infrastructure and the costs of doing business. In addition to the Diagnostic Trade 2 Integration Study (DTIS) and Agricultural Policy Note (2006), analytical work has been completed on the growth potential o f particular commodities, and on issues related to trade facilitation, to inform the development o f a trade and export promotion strategy. Building on the results o f the Financial Sector Assessment Program (FSAP, 2005), an action plan for deepening the financial sector is under preparation. In addition, an Investment Climate Assessment will soon be launched, andwill be used as the basis for a private sector development strategy. These strategies and action plans need to be assessed within a consistent macro-economic framework that reflects the resource envelope, inorder to test the realism o f the strategies; and prioritize measures and clarify issues related to the sequencing o f reforms and interventions to achieve growth targets. Framingthe growth strategy within a consistent macro-economic framework will help to ensure that macroeconomic policies are in line with growth objectives and the fiscal policy to achieve those objectives. 1.1 OBJECTIVE, SCOPEAND STRUCTUREOFTHE CEM 1.7 This Country Economic Memorandum (CEM) aims to provide the analytical basis for the development o f strategies and action plans that will realize Government's vision for growth and poverty reduction. The aim is to devise a macro-framework that also takes into consideration the challenge of managing external shocks, and that will also examine the potential scaling up of aid inthe context ofthe country's already highaid dependence. 1.8 Results from the CEM will inform a further elaboration o f the Government's growth strategy in its EDPRS, currently under preparation. These results will also facilitate the identificationof growth-related measures, to be included inthe Poverty Reduction Strategy Credit (PRSC)3, as well as provide the specificity neededto generate donor grant financing. Inaddition, the CEM will inform the preparation ofa new Country Assistance Strategy (CAS), to be prepared jointly with DFJDinFY08. 1.9 The CEM focuses on identifyingthe sources of growth through an examination of past trends. The study also examines sector-specific issues related to agriculture, industry, and services, with a focus on identifying the most binding constraints to growth and improved competitiveness. The study concludes with some policy simulations to examine the trade-offs among key policy options. 1.10 The rest of the CEM is organized as follows: Chapter 2 provides an overview of the structure o f the economy and outlines the macro-economic context. Chapter 3 analyzes the sources and constraints to growth. Chapter 4 examines the impact of high aid flows on the economy, and discusses issues related to the macro-economic management o f aid. Chapter 5 outlines the sources of and constraints to growth in the agriculture sector, which has been the driver of growth over the past decade. It also shows that although growth in agriculture will be The PRSC i s the World Bank's policy lending instrument in support o f the PRSP. The past three PRSCs have focused on support to reforms inthe social sectors andto improvefinancialand expenditure management, with growth as a key pillar. There has been an increasedfocus on reforms inthe infrastructureandagriculturesectors. 3 necessary in the immediate term to achieve Government targets, it will not be sufficient. Strong growth in the non-agriculture sector will also be needed, especially to lay the foundation for transformation o f the economy. Chapter 6 outlines the sources of growth and constraints for the non-agriculture sector with a focus on private sector development issues and the service sectors. It examines issues from the perspective o f small, medium, and large enterprises. Chapter 7 examines the constraints to export growth and competitiveness over the past decade, prioritizing what measures to adopt for quick gains, with attention to commodity-specific constraints. Chapter 8 concludes with policy recommendations. Box 1.1: Overview of Selected Key Reforms since 1995 Source: Ministry of FinanceandEconomicPlanning. 4 2. SETTING THE STAGE-POLITICAL AND ECONOMICCONTEXT 2.1 Rwanda has made remarkable progress since the crisis of 1994 and is now moving toward a period of sustained growth. The average growth rate of the economy between 1985 and 1993 was 0.8 percent. GDP declinedby 30 percent following the 1994 genocide, and it took 6 years o f 14 percent average annual growth to reach pre-genocide levels. Between 2001 and 2005, the economy grew at an average annual rate o f 5.3 percent, with a 6 percent average annual rate of inflation. 2.2 Much o f the growth observed inthe past was due to the economy's catch-up to pre-1994 GDP levels; however, the continuation of reforms initiated prior to 1994 also contributed to the growth trends? Reforms included measuresto liberalize the trade and exchange regimes and the commodity and financial markets, and initiate the privatization of public enterprises. Reforms were also adopted to improve the management o fthe national budget, through the adoption o fthe Medium-TermExpenditureFramework (MTEF), and an updated Organic BudgetLaw, with the aim of modernizing the public financial management system. Recent reforms have increasingly focused on improvingthe results orientationof the MTEF. 2.3 However, growth is now slowing. Years o f destruction and neglect arising from the civil conflict, coupled with slow progress in rebuilding and extending economic infrastructure have hindered productive activities and economic growth, particularly in rural areas. Estimates from the second household consumption survey (EICV), conducted in 2005-06, show a slow process in the reduction o f poverty. Recent data indicate that the reduction in income poverty incidence between 2001 and 2005 was marginal - from 60.4 to 56.9 percent. Rural poverty declined from 66 to 61.5 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 been the level of productivity and limited progress inproduction for the market inthe agriculture sector. 4See boxes 1.1and 2.1. 5 6 2.4 Sustained growth and poverty reduction require that five key conditions be achieved: (i) maintenance of peace and stability, particularly givenRwanda's location in a region that has been prone to conflict; (ii)reduction in the economy's dependence on rain-fed agriculture, to limit its vulnerability to weather-related shocks; (iii)increased resource mobilization, both domestic and external, to finance economic growth; (iv) development of the private sector, with sustained improvements in the business climate; and (v) increased productivity in agriculture and the development o f non-farm rural economic activity. These issues are outlined beow, in the sections on the political and historical context, outcomes o f reforms, and the economic context. 2.1 THEPOLITICALAND HISTORICALCONTEXT 2.5 The Government of Rwanda's vision and strategy for growth and poverty reduction is very much grounded in the experience o f previous civil strife and conflict, its understanding of the issues that causedthe past conflict and violence, and the lessons learned. A brief review of the historical and political context is helpful for understanding the forces that influenced the choice of reforms. 2.6 Rwanda's colonial legacy deepenedpre-existingsocial and class distinctions, and laidthe foundation for incidents of civil strife and violence that have periodically occurred in the past. Colonial rule resulted in social and economic privileges for the Tutsi minority (about 15 percent o f the population). In 1959, three years before independence, the Hutus (accounting for over 80 percent of the population) revolted against Tutsi domination, killed many o f the chiefs, and overthrew the monarch. The killings and massive displacement continued in 1963, when Tutsis in exile launched an attack on the newly formed government, and then again in 1973 when the killings o f Hutus in Burundiled to anti-Tutsi violence in Rwanda. After a decade and a half of relative stability, a full-scale civil war erupted inOctober 1990with the invasion o f Rwandafrom neighboring Uganda by the Rwandese Patriotic Army (RPA), the military wing o f the Rwandese Patriotic Front, an organization o f Rwandese exiles seeking to returnto Rwanda. Efforts by the international community to end the fighting culminated in the signing of a power-sharing agreement in Arusha, Tanzania in 1993 by all political groups in Rwanda including the RPF. However, before, the Arusha Accord could be put into effect, a massive genocide erupted inApril 1994 during which an estimated 800,000 lives were lost, 2 million Rwandese sought refuge in neighboring countries and hundreds o f thousands more were internally displaced. The RPA, with forces in Rwanda was able to put an end to the genocide and following this, a Government o f National Unity (GNU) was formed inJuly 1994,within the framework o f the Arusha Accord. 2.7 As highlighted in Obidegwu(2003), the GNUfaced enormous challenges related to " ... restoring peace, resettling millions o f displaced people and refugees, facilitating the return, resettlement and reintegration of the refugees outside the country, assisting victims o f genocide, bringing perpetrators of genocide to justice, building institutions and capacities, rehabilitating economic and social infrastructure, and reviving the economy." This was clearly a large agenda 7 for a coalition of disparate forces with little experience in government and very limited financial resourcesand technical ~apacity,~which initially faced intermittent military challenges. 2.8 InDecember 1994, the GNUissued a Declaration of Principles that outlined its political, social, and economic agenda for a "New Rwanda". On the economic side, the Declaration emphasized the new Government's commitment to a market-based economy with a limited role for government, a strong presenceof the private sector; the privatization of public enterprises, the liberalization of the trade and exchange rate regimes, and the reduction of military expenditures. Subsequent Government policies and actions largely followed these commitments (see boxes 1.1 and 2.1 for an overview o f key reforms undertaken). Moreover, a political-economic analysis of the causes of the genocide, and grassroots consultations for the Government's 1998 Vision 2020, stressed the need for increased voice, inclusion, and widespread participation. Inparticular, the focus on decentralization, human resource development and capacity, and the broadening o f grassrootsparticipation and government responsivenessemerged from the lessons learned. 2.9 The direction o f Government policy was influenced by the experience of returnees from Uganda, where a transition to more liberal economic policies had occurred. Moreover, a political-economic analysis o f the causes of the genocide, and grassroots consultations for the Government's 1998 Vision 2020, stressed the need for increased voice, inclusion, and widespread participation. In particular, the focus on decentralization, human resource development and capacity, and the broadening o f grassroots participation and government responsivenessemerged from the lessons learned. 2.2 OVERVIEW OFKEY REFORMSUNDERTAKENAND OUTCOMES, 1994 PRESENT - 2.10 In parallel with efforts to buildpeace and stabilize the economy, a number of economic and structural reforms were undertaken to strengthen the market orientation of the economy. A flexible exchangerate was introducedin 1995. Tariff rates were reduced from 34.8 to 18percent; and the number of bands was reduced from 5 to 4, ranging from 0-30 percent, by 2003. Liberalizationof the monetary and financial sectors ledto the adoption o f new currency exchange regulations, the creation o f private commercial banks, and the privatization o f state-owned banks.6 Current account operations (imports, exports, services) were liberalized, and certain restrictions on capital flows were eliminated. These includedthe transfer o f capital and revenues related to foreign direct investment (FDI), and free withdrawal from foreign exchange accounts in commercial banks.' More recently, Rwanda has joined the East African Customs Union (COMESA) and the East African Community (EAC). 2.11 Inthe area ofstructural reforms, the RwandaUtilities RegulatoryAgency and aone-stop- shop were established inthe Rwanda Investment and Export Promotion Agency. In addition, an export promotion strategy and an action plan have been recently elaborated. The Government has ' See C. Obidegwu (2003), "Rwanda: The Search for Post-Conflict Socio-EconomicChange, 1995-2001." Africa RegionWorking Paper Series, Number 59, World Bank. This sectiondraws heavily from the paper. 6 Rwanda had three commercial banks prior to 1995; since that time, the number of commercialbanks has doubled. 'Thecountrycurrently has 6 commercialbanks. Kanimba, Francois(2004). "Deepening the financialmarket inRwanda". PowerPointpresentation(September). 8 also reduced its role inthe market by privatizing numerous state-owned enterprises, includingtwo commercial banks (Banque Commerciale du Rwanda and BACAR), Rwandex (the largest coffee exporter), three tea factories, and Rwandatel (the telecommunication company). However, the privatization of the public utility, Electrogaz, and the tea factories has progressed more slowly than planned. Less progress has been made intransforming agriculture and improving economic infrastructure inrural areas. 2.12 Substantial progress has also been made in the area of public finance and expenditure management. The MediumTerm Expenditure Framework was introduced in2001, leading to improved budget predictability and a more rational approach to budget preparation, implementation, and monitoring. The Organic Law o f State Finances and Assets, which is expected to modernize the public financial management system, was adopted inMarch 2006; and a Public Procurement Code is currently beingreviewedinparliament. 2.13 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. In the case of education, there is increased involvement of communities; and capitation grants (transfers per student to school districts) have been adopted. Despite the gains in institutional and legal reforms, however, capacity continues to be a major challenge that affects overall economic governance. 2.14 Along with structural reforms, the Government initiated measures to promote reconciliation and peace, in recognition o f the importance of these factors for growth. A system of gucucu courts, modeled after a traditional approach to settling disputes, was established as a means of engaging citizens inthe genocide trials. The current estimate is that 717,000 people will be tried for war crimes in 12,100 gucucu courts throughout the country. In addition, a National Unity and Reconciliation Commission was established to redress Rwanda's legacy of divisive politics through civic education initiatives. Further, about 3.5 million refugees have been repatriated and resettled8; more than 37,000 ex-combatants have been demobilized; and another 15,000, including the former commander-in-chief of the FDLR,' have been integrated into the RwandaDefense Force at various command levels. 2.15 A decentralization program was undertaken in 2000. The first phase o f the decentralization (2000-03) involved a number o f 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. 2.16 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 Currently there are about 57,000 Rwandan refugees in21African countries. The FDLR(Forces DCmocratiques de LibCration du Rwanda), founded in2001, is a rebel group operating ineastern Democratic Republic ofthe Congo. 9 resolving social conflicts, and for power sharing at the top level o f government. Under the new constitution, the president cannot appoint more than 50 percent o f the cabinet from hisher own party, and the speaker o fthe parliament has to come from the minority party. 2.17 These reforms have helped to maintain peace and stability in Rwanda, which inturn has made possible the strong growth and social sector improvements observed 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 o f the population. The next challenge i s to address the problem o f 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 o f education. Sustained growth inthe future will require more attention to improved delivery of economic and infrastructure services. This i s discussed further inthe followingsection. 2.3 THE ECONOMIC CONTEXT 2.18 The economic context reveals a number of challenges to growth, as well as opportunities. First,the structure o fthe economy following the 1994 crisis is suchthat there is a heavy reliance on agriculture. This makes the economy more vulnerable to weather-related and terms-of-trade shocks, and has had an adverse impact on economic growth. Second, the degree of resource mobilization is limited, given low domestic savings and investment, although revenue has increased steadily over the years. In addition, foreign direct investment i s extremely low. Given the low export base and the need to maintain sustainable debt indicators, the level of borrowing is limited, which has resulted in a highdependence on foreign aid. Yet, the potential for scaling up aid presents an opportunity for Rwanda to ensure that aid funds are managed in a way that supports growth and increased absorption. Finally, although the population growth rate poses a challenge, the high population density and small landholding may also provide an opportunity to make a relatively large and quick impact (compared to more sparsely populated countries with large land masses), through investments and policy measuresto reduce transaction costs and improve the condition of infrastructure. 2.3.1 Structureof the Economy 2.19 Rwanda's economy i s characterized by the high dependence on agriculture and primary products that materializedafter the 1994 crisis. Prior to 1994, the economy showed a strong and continual shift toward the service sector, while agriculture fell significantly as a share o f GDP. However, these trends were reversed after 1994, reflecting a significant structural change. Agriculture currently makes up 45 percent o f the economy, with more than 90 percent of production coming from food crops and most o f the remainder from livestock production (Table 2.1). The sector accounts for more than 71 percent o fthe country's export revenues, mainly from coffee and tea. 10 Figure 2.1: Historical Trend in the Components of GDP, 1970 2004 - (As share of GDP) I i 7 0 b72 L74 L 7 6 L78 L 8 0 L 8 2 L 8 4 k 8 6 L 8 8 L 9 0 L 9 2 L 9 4 A96 k 9 8 2000 2002 200; Source:World DevelopmentIndicators,2005. 2.20 The tertiary sector is the second largest contributor to GDP, accounting for an average over the past decade of 37 percent of GDP, followed by the secondary sector, which accounts for 18 percent o f GDP. Over the period 1995 to 2004, commerce and tourism accounted for 27 percent and public administration accounted for 19 percent of the tertiary sector. Transport and communications accounted for 16 percent. Table 2.1: SectoralShares in GDP, 1995-2004 (Inpercentage, based onyear 2000 relativeprices) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 est Primary 45.9 46.4 45.8 45.2 41.8 42.3 43.4 44.8 45.2 46.9 1. Food crop 38.1 37.6 38.5 39.0 35.4 36.2 37.1 39.1 39.7 40.7 2. Export Crop 2.6 1.5 1.5 1.1 1.2 1.O 1.1 0.7 0.7 0.9 Of which: Coffee 2.6 1.3 1.3 0.8 1.O 0.8 0.8 0.4 0.4 0.5 3. Livestock 3.4 5.4 4.5 3.8 3.8 3.5 3.6 3.5 3.3 3.2 4.Others 1 8 1.9 1.3 1.3 1.5 1.6 1.6 1.5 1.5 2.0 Secondary 16.6 17.7 18.3 18.5 19.3 19.0 18.8 18.6 18.1 18.5 1, Mining and quarrying 0.1 0.1 0.1 0.2 0.2 0.3 0.5 0.5 0.3 0.9 2. Manufacturing 10.6 11.3 11.7 10.9 10.4 9.6 9.6 9.3 8.8 8.9 Ofwhich. Food, beverages, & tobacco 9.1 9.8 10.2 9.2 8.6 7.7 7.7 7.3 6.9 7.1 3. Electricity, gas, & water 0.3 0.4 0.6 0.8 0.7 0.6 0.6 0.5 0.4 0.3 4. Construction 5.6 5.9 5.9 6.6 7.9 8.5 8.1 8.3 8.6 8.4 Tertisry 38.4 35.4 35.0 35.3 38.8 38.8 37.6 36.0 35.7 33.7 1. Wholesale & retail trade, rest. & hotels 11.3 10.6 10.4 10.3 10.9 10.5 10.0 9.6 8.3 8.4 of which Wholesale & retail trade 10.7 10.I 9.9 9.8 10.4 10.0 9.5 9.2 7.9 8.1 Restaurants & hotels 0.6 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.4 2. Transport & Communication 4.3 3.9 4.1 4.7 6.4 7.0 7.0 6.8 6.3 6.1 3. Service b/ 2.2 1.9 2.2 2.3 2.7 3.1 2.9 2.8 3.4 2.9 o f which Financial institutions 1.9 1.8 2.2 2.2 2.6 2.7 2.6 2.4 3.1 2.5 4. Public administration 6.8 6.9 7.0 6.6 7.3 7 3 7.3 6.8 6.5 6.0 5. Others 13.7 12.1 11.3 11.4 11.5 10.9 10.4 10.0 11.1 10.2 o f which Non-profit organizations 1.9 1.6 1.3 1.2 1.2 1.0 1.O 0.9 0.9 0.9 Less: Imputcd bank service charge 2 9 2.0 1.5 1.4 1.6 1.6 1.6 1.4 1.5 1.6 Plus: Import duties 2.1 2.5 2.4 2.4 1.7 1.6 1.8 2.0 2.4 2.4 Gross domestic Product d 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Ministry o f FinanceandEconomicPlanning, Direction of Statistics a/ Realvalue of GDP andits components are obtained by dividing nominalvalues by the overall GDP deflator b/. Includefinance, insurance, real estate, business c l Total may not sum perfectlydue to roundingup 11 2.21 As Table 2.1 shows, the secondary sector accounts for around 18 percent o f GDP, and manufacturing accounts for 54 percent of output in this sub-sector. This is followed by construction, which accounts for 48 percent o f the secondary sector. The manufacturing sector contributed, on average, 10 percent o f GDP between 1995 and 2004, and currently contributes about 5 percent to total exports. The sector is small and highly concentrated, with 20 firms contributing roughly 65 percent of total manufacturingoutput. Agro-industrial activities account for 80 percent of total manufacturing. The largest firms are engaged in the production of soft drinks, beer, and other food products. Key firms are also engaged in the production of cement, tobacco, textiles, aluminum products, plastic, and chemical products (soaps, paint, and matchsticks). Thus, the economy is heavily dependent on the primary sector, with industry strongly tied to the processing o f primary products from the agriculture sector. 2.22 Miningcurrently representsabout one percent of the secondary sector, and is dominated by the production of important commodities such as cassiterite, colombo tantalite, and wolfram. Between 2000 and 2003, miningproducts represented 26 percent o f total export value. In2001, they represented 45.5 percent oftotal export value. 2.23 The heavy dependence on the agriculture sector and related primary products, coupled with a high dependence on rain-fed production systems, makes Rwanda highly vulnerable to terms-of-trade and weather-related external shocks (see Table 2.2). Low levels o f input use, coupled with limited investment in irrigation and water management, have contributed to the strong link between exogenous shocks to agriculture and the impact on overall GDP. As a result, growth inagriculture has not materializedas foreseeninthe PRSP. 2.24 External drought conditions during late 2002 and early 2003, coupled with falling international commodity prices, ledto a fall inGDP growth to 0.9 percent." These trends ledto a worsening o f the current account deficit, which was financed through foreign reserves. By 2004, GDP growth had increased again to 4 percent, due to improved rains and the price of coffee. A similar trend occurred between 2004 and 2005, when agriculture growth led to overall GDP growth increasing from 4 to 6 percent. Table 2.2: Trend and compositionof GDP, 2002-2005 (In unit as indicated) 2002 2003 2004 2005 RwFbillion RwFbillion Growth ("A) RwFbillion Growth (YO) RwFbillion Growth(%) Agriculture 299.1 286.9 4.1 289.36 0.9 306.3 5.8 Ofwhich food crop 255.9 244.4 -4.5 241.8 -1.1 259.4 7.3 Industry 112.5 120.5 7.1 128.8 6.9 143.1 11.1 Services 224.1 234.7 4.7 253.2 7.9 266.9 5.4 GDP 635.6 641.4 0.9 668.4 4.2 710.6 6.3 Source: Ministry of FinanceandEconomicPlanning. 2.25 Inadditionto the impact of drought conditions on GDP, vulnerability to terms-of-trade shocks arises from the narrow export base and the concentrated nature o f exports. The export 10This contrastswith average 10percentgrowthover the period 1995-2002. GDP growth between 1996 and 1998 was 11.4 percent,and 8.1 percentbetween2000 and2002. 12 basket i s concentrated in a few commodities. In 2005, the volume o f coffee exports was $32 million, mineral ores $31 million, and tea $6 million. Jointly, these commodities accounted for 91 percent of total export volume. The remaining 9 percent comprised o f at least * *170 products, eachwith a near miniscule share, often no larger than US$2000to US$3000. Skins and hides, the fourth largest export item, for example, was valued at only $350,000. It is also important to note that although exports of services, as a share o ftotal export value, has increased from 29 percent in 1990 to 51 percent in 2004, the concentrated nature o f the commodity exports continues to strongly affect the terms oftrade. 2.26 In terms of geographic concentration, out of the 90 percent of exports stemming from agriculture and mining, 46 percent goes to the EU, 26 percent to East Asia and the Pacific - mainly Hong Kong and China, and the remaining 4.6 percent to COMESA. However, there appearsto be increasing diversification inthe destination o f exports. Figure 2.2: Terms of Trade, Export volume and export value 1995-2005 (Index Base 100=2000) 1300 250 1250 200 1200 I150 150 100 50 0 Source: NationalBankof Rwanda(NBR) 2.27 Over the past 5 years, the share of commodity exports to the EUand USA markets has declined. The emerging markets of China and other Asian countries are playing an increasingly important role. However, as Chapter 7 shows, the scale o f exports for these emerging markets has been small - exports to emerging markets are concentrated injust one or two commodities. For example, almost 100percent of exports to China and Hong Kong between 2000 and 2005 were ores; exports to Malaysia were also dominated by tin ores; and more than 90 percent of the exports to Pakistan were tea. Yet, the traditional exports (such as tea and coffee) still accounted for 51 percent of total exports over the period, while non-traditional exports (half of which comprised tin and non-ferrous ores) accounted for just 9 percent o f total exports by 2004. 11The number of products depends onthe level of disaggregation at which the data is analyzed. For SITCl - 3 digit, a country can have a maximum o f about 235 products; this number increases to more than 600 when a finer level of disaggregationis applied(SITC2 4 digit). - 13 2.28 Total exports has hovered around 10 percent of GDP over the period 1995 to 2005, with coffee and tea accounting for more than 50 percent of Rwanda's export volume. Terms-of-trade shocks for these commodities contributedto the deficit inthe balance o f payments. After a strong improvement in the world market prices for coffee during 1995-96, the collapse o f coffee prices in 1997 ledto a 42 percent (cumulative) deterioration inRwanda's external terms oftrade during 1995-2002, which was equivalent to a terms-of-trade shock of 19 and 20 percent, respectively, in 2001 and 2002. This, in turn, led to a 27 percent reduction in coffee production between 2002 and 2003. With the passing o f a short-lived boom in coltan (2000-Ol), and with the adoption o f an export promotion strategy for the coffee and tea sectors delayed by management limitations, exports dropped in 2002, leaving Rwanda with one o f the lowest merchandise export-to-GDP ratios inthe world (4 percent in 2003). Inthe case of imports, trends led to a largely unchanged trade balance, as a fall infood imports over the period was replaced by an increase in imports of capital goods to support the construction sector. 2.3.2 Financingand Investmentto Support Growth 2.29 In terms of financing, domestic revenues, including receipts from export and trade revenues, cannot currently support the level of public spending; but sustainedreforms have led to improved revenue performance over the period 1995 to 2005. Government revenue, estimated at about 13 percent of GDP in 1992, fell below 4 percent in 1994 and increased gradually to an average of around 10 percent per annum between 1996 and 2001. By 2002, revenues stood at 12.2 percent of GDP, reflecting the increase in the VAT rate from 15 to 18 percent, and the substantial strengthening o f revenue administration. Revenue as a share o f GDP rose further in the following years and is now at 15 percent, compared with the SSA average of around 22 percent (see Table 2.3). This increase in revenue reflects the impact of income tax reforms that went into effect in 2003, including the taxation of in-kind benefits. However, while revenue performance has improved over the past decade, the fiscal imbalance remains large, at 12.7 percent of GDP in2005. 2.30 Total government spending as a share o f GDP has steadily increased, over the period 1995 to 2005. The high costs of reconstruction and poverty reduction programs have increased spending levels to above 26 percent of GDP in 2005. The Government has made efforts to reallocate resourcesto priority sectors, where spending has almost doubled inthe past few years. Inparticular, spending onthe healthandeducation sectorsnow accountsfor roughly 30 percent of the total budget and 40 percent o f the recurrent budget. Increased spending levels have resulted in large fiscal deficits, financed mostly by a combination of external grants and concessional loans. 14 Table 2.3: Revenue in selectedSub-SaharanAfrican countries (as share of GDP) 2004 Rwanda 13.9 Kenya 15.8 Uganda 11.7 Benin 14.6 Ghana 21.6 Togo 14.8 Burundi 18.3 Nigeria 42.5 South Africa 23.7 Sub-SaharanAfrica 22.5 Source: World Development Indicators 2005. 2.3 1 Interms o fthe external accounts, the overall balance o f payments has averaged 1percent o f GDP, with some degree o f volatility driven by the current account balance (Figure 2.3). The trend in the current account has been driven by the trade balance. The relative movement in the terms-of-trade index, and the indices for exports and imports, suggest that exports underpin changes inthe trade balance. Figure 2.3: Volatility in the Main Components of the Balance k of Payments 1995-2005 (aspercentage of GDP) 10 6 9 9 *- 5 .-- 0 l p 5 1996 1997 1998 1999 2000 2001 -5 - 4. 0' \ -10 t 0 -15 - -.- * CapitalAccount - --Currentaccount -BOP -. Source:International Monetary Fund 2.32 Export earnings covered only a fraction of the import bill, estimated to be about 25 percent o f GDP, in 2005, leaving Rwanda highly dependent on external assistance to finance its substantial balance o f payments deficit. Although Rwanda received US$1.3 billion in public current transfers during 1995-2002 (10 percent o f GDP, on average), this fell short of spending and investment needs. Currently, the fiscal deficit (excluding grants) is around 13 percent, and the current account deficit isjust over 21 percent. 15 2.33 Exports make up a small share o f total GDP. Rwanda's export volume as a share of GDP, at 5.3 percent, is one o f the lowest in the world. With a significant increase in coffee volumes, and higher volumes and price increases for metals such as cassiterite and coltan, exports o f goods recoveredto US$125 million in 2005, compared to US$97.7 million in2004 and US$63 million in 2003. Coffee and tea continue to play a key role in Rwanda's exports, although their share has declined tojust over 50 percent, down from 90 percent inthe mid-1990s. 2.34 However, there has been a slight improvement in the trade balance over the past 5 years, arising from improvements inthe terms o f trade, and improved volume o f exports. The terms-of- trade index in 2004 was 65, compared to 100 in 1995. The traditional exports o f coffee and tea accounted for more than 67 percent of exports in 1996-99 and around 55 percent during 2001- 2003. There has been a significant increase in non-traditional exports, particularly crude materials and minerals. Crude materials accounted for about 17.8 percent of the total volume of exports in 1996-1999; they increased to 37.5 percent in 2001-2003, and are now an estimated 49 percent of merchandise exports. The total value of imported goods (fob) consisting mostly of consumer goods, were US$275.9 million in2004 and US$384.9 million in2005. 2.35 On the services side, exports o f services, includingreceipts from improved tourism and travel, increased to about US$92 million, in 2005. The decline in service payments mirrors the fall in payments for insurance and freight (associated with a decline in imports). Revenue from exports of services (mainly tourism), on the other hand, has been sluggish, but i s showing some improvement over the past two years. With US$247 million in imports of services, and net income of US$34 million (which includes investment income, interest payments, and remittances) the current account deficit as a share of GDP was 19 percent in 2005. Most of the current account deficit has been financed in recent years by grants and the rest by concessional loans, mainly from multilateral institutions. The relative improvements in the income accounts, when compared to the level in the 1990s is mainly due to the lower interest payments arising from debt relief, rescheduling, and a greater share of concessional debt. Over this period, gross official reserves have remained atjust over 4 to 5 months of imports, on average. 2.36 Despite the fact that actual borrowing fell below projected amounts, Rwanda's debt indicators worsened during the 1995-2002 period. However, achievement of the HIPC completion point in 2005, coupled with the country's eligibility for the Multi-lateral Debt Relief Initiative (MDRI), have reduced Rwanda's ratio o f net present value o f debt to exports to 58, from a pre-HIPC level o f almost 300 percent. In addition, the stock o f external debt has declined from 73 percent of GDP to 15 percent (Table 2.4), with substantial reductions in debt service payments. Future borrowing, however, continues to be severely limited due to the low export base, and can only be expanded as the private sector grows and develops further. This will require policies to increaseproductivity and exports. 16 Table 2.4: SelectedMacroeconomic Indicators, 2005-2009 (In unit indicated) 2005 2006 2007 2008 2009 Actual Projected Percent change Real GDP 5.0 4.3 4.5 4.9 4.9 Inflation (CPI average) 7.0 4.0 4.0 5.0 5.0 Export Volume -5.6 6.3 6.2 6.3 6.5 Terms o f Trade -4.5 -3.1 -4.9 -4.9 -2.4 Percent of GDP Public Finance Gov. Domestic Revenue 14.0 14.1 14.3 15.1 14.9 Grants 11.8 12.1 11.1 11.6 12.0 Total Expenditures 26.7 27.7 27.3 28.6 28.6 Fiscal deficit (excl. grants) -12.7 -13.6 -13.0 -13.5 -17.7 Fiscal deficit (incl. grants) -0.9 -1.6 -2.1 -1.9 -5.7 Balance of Payment Current Account Balance -19.4 -19.8 -19.0 -18.0 -17.5 Exports, GNFS 9.3 9.5 9.5 9.7 ... Imports, GNFS 31.4 29.6 29.5 28.9 ... Official Reserves (mos. o f imports) 6.3 4.8 4.6 4.6 4.6 External Debt 71.0 15.0 16.9 20.1 24.3 Memo Items: Exchange Rate RFR/US$ 570 580 585 596 ... GDP (US$ million) 2054 2215 2380 2512 2766 Source: Ministry of Finance, World Bank andFund StaffEstimates. 2.37 The importance of private sector development is demonstrated by the current low levels of investment, savings, and credit, which cannot be used to effectively manage shocks or substantially support growth. Following the 1994 conflict, private investment picked up gradually and by 2002 was 12 percent of GDP; this figure is estimatedat 12.6 percent for 2004. Public investment is estimated at 8.5 percent for 2004, and was 4.9 percent in 2002. The country's gross domestic investmentrate over recent years has been reasonable (about 20 percent of GDP), but the gross domestic savings rate remains extremely low, at less than one-tenthofthe average for Sub-SaharanAfrica, or below 2 percent of GDP (see Figure 2.4). Figure 2.4: The Savings-Investment Balance, 2004 (As share of GDP) GI1Average SSA Rwanda Uganda Tanzania Senegal Mozambique Mali Ghana Ethiopia Burkina Faso Benin I I 1 Investment ~9Savings 10 15 20 25 3c Source: World DevelopmentIndicators2006. 17 2.38 Rwanda has attracted a very small fraction of the FDI inflows that neighboring countries have been able to attract. Despite significant increases in FDI flows to SSA, Rwanda's share has been declining since the mid-1990s (Figure 2.5). In 1990, FDI inflows to Rwanda were 0.28 percent o f GDP. Although FDIhad increased to almost 0.4 percent o f GDP by 2004, remittances accounted for a slightly larger share, at just over 0.53 percent of GDP (Figure 2.6). Figure2.5: FDIis extremely Low, even by Figure 2.6: Workers' Remittancesexceed SSA Standards (aspercentage of GDP) FDI, 1995-2004 (asPercent of GDP) I X / I I990 1992 1994 1996 1998 2000 2002 2004 tFDIRnanda . t F D I . S S 4 A 1 1 +Private inv.. Rwanda +Private inv.. SS4 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Source: World DevelopmentIndicators2005 andInternationalMonetaryFund. 2.39 The country's location in an area that has historically been prone to conflict is likely a major factor inthe low levels of investment. Past conflicts and unrest inthe region have been a significant source of external shocks, which have affected foreign investment and constrained growth. Violent conflicts have contributeddirectly and indirectly to the progressive destruction of human, physical and social capital inRwanda. Inthe rural areas, the bedrock o f the economy and home for 80 percent of the population - the livestock herd, which accounted for about 5 percent of GDP -was almost wiped out by the genocide. Simultaneously, soil fertility declined as a result of environmental degradation and the loss o f animal manure. The violence resulted in the destruction of capital and a decline in investment, culminating in the loss of productivity and national income and a doubling o f income poverty from 40 percent in 1990 to an estimated 78 percent of the population, just after 1994. 2.40 Capital outlays have remained substantial throughout the post-genocide period, generating significant debt, which has been financed by a combination of external grants and loans. While revenue performance has improved over the past decade, the fiscal imbalance remains large. Given the level o f needs, the country will continue to be dependent on grants or highly concessional external financing for the foreseeable future. Trends in the composition o f spendingshow that the Government hasrecognized the importance of usinggrant funds to lay the foundation for growth. 2.3.3 FiscalPolicy and the Compositionof Spendingto Support Growth 2.41 Since 1994, fiscal policy in Rwanda has been geared toward stabilization and launching the country on a growth path. Domestic lending remained around 5 percent of GDP while government reforms focused on improving the level and efficiency of revenue 18 generation. External borrowing was largely on concessional terms and focused on reducing the debt burden. Fiscal policy remained expansionary given the extent o f destruction andthe reconstructionneeds. The pattern o f spending focused on investing in the depleted human resources and rebuilding institutions to facilitate reconciliation. Reconstructionefforts ledto highgrowth inthe construction sector, and at the same time, Government expenditure focused on improved delivery of services (Table2.5), duringthe mid-1990s an into the early 2000s. Table 2.5: Rwanda: Priority Spending, 1998-2005 (Inpercentage of GDP) 1998 1999 2000 2001 2002 2003 2004 2005 Total Budget 18.9 19.7 18.7 21.0 21.3 21.9 25.2 26.2 Total priority expenditure o f which: 2.8 3.9 5.3 5.3 6.4 6.9 7.0 8.4 Education 2.2 3.2 3.5 3.5 3.6 3.8 3.9 3.9 Health 0.4 0.5 0.7 0.7 0.8 0.8 0.8 0.9 Agriculture 0.0 0.0 0.2 0.2 0.2 0.2 0.2 0.4 Exportpromotion 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.5 Transport and communication 0.0 0.0 0.1 0.1 0.4 0.5 0.4 0.9 Infrastructure(Energy and water) 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.3 Common Development Fund 0.0 0.0 0.0 0.0 0.1 0.3 0.3 0.3 Other* 0.2 0.2 0.2 0.9 0.5 1.o 1.3 1.3 Source: Rwandese authoritiesand IDA staff estimates. * This categoryincludes, inter alia, spendingon internal affairs, local government, commerce, andyouth andsports. 2.42 After a downward trend inthe late 1990s,total Government spending increasedfrom 19 to 24 percent between 1998 and 2003, to 26 percentby 2004 (Figure 2.7). Recurrent expenditure has also beenon an upwardtrend since the 1990s. ' Figure 2.7: Reccurentand DevelopmentExpenditures, 1970-2004 1 2 0 . (as a share of GDP) 25.0 . . . ----e---- Recurrent expenditure -Development. . expenditure . I I A 15.0 10.0 5.O j 0.01!970 1973 1976 1979 1'982 1'985 1988 1991 1994 1997 2000 2003 1 Source: World Development Indicators 2005, Rwandese authorities, and Bank Staff calculations. 2.43 The composition of spending has been geared mainly toward the social sectors and services, as opposed to delivery of economic services. The increased expenditure partly went to priority spending, and within priority spending, at least 50 percent went to the social sectors in 2004. The trends and composition o f expenditures, given increases in official development 19 assistance (ODA), indicate that there was no issue o f fungibility. For example, between 1998 and 1999 priority expenditures increased by 43 percent, while net ODA increased by 6.6 percent. Further, between 2001 and 2002, priority expenditures increased by 26 percent while net ODA increased 19 percent. 2.44 The pattern in spending described above led to significant improvements ineducation and health outcomes, but there was little to no improvement in economic services related to infrastructure and agriculture. Health immunization rates have increased significantly, however mortality rates for infants (per 1000 live births) are significantly higher than the SSA average (Figure 2.8), and are still above the pre-1994 levels. For example, infant mortality in 1990 was 103 but rose to 118 in 2004. Primary school enrollment increased, but attainment has not yet reachedpre-1994 levels, as will be discussed inChapter 3. Literacy rates are slightly below those o f Uganda and Tanzania, although they are above the G-11 country group average (Figure 2.9.) The next generation o f reforms in education needs to focus on improvements in quality o f education. Figure 2.8: Infant Mortality Rates in G-11Countries, 2004 (per I000 live births) Mauritania Burkina Faso 20 40 60 80 100 120 140 Source: World DevelopmentIndicators,2005 Figure 2.9: Adult literacy rate in 2004 and Trends in the illiterate population (Percentage ofpopulation ages I 5 and above) 90 80 70 -BurkinaFaso 60 50 40 -Mauritania 30 -Mozambique 20 10 -Tanzania Tanzania 1 20 40 60 80 0 1970 1975 1980 1985 1990 1995 2000 Source: World Development Indicators 20 2.3.4 The ImportanceofAid Flows and MacroeconomicPolicyImplications 2.45 Aid flows have beena major factor influencingthe main macroeconomicindicators. Net official development assistance averaged 15 percent o f GDP in the 1980s, and by the 1990s had risen to roughly 50 percent of GDP. These trends are summarized in Figure 2.10, which shows foreign aid inflows, in the form o f grants, loans, and net official development assistance (ODA) as a share of GDP.'* Each form of financing (as a share o f GDP) has increased steadily since the 1970s. External debt was on an upwardtrend from the late 1970s, with foreign loans provided mainly on commercial terms. By the 1980s, loans were the largest share of total financing, and were becoming increasingly more concessional compared to the 1970s. As the level o f loans increased, the level of grants decreased, resulting in a large and growing gap between grants and loans beginning inthe late 1990s. 2.46 Since 1994, a substantial amount of aid has been transferred to the country. Between 1994 and 2004, net ODA averaged 29.7 percent o f GDP and went from 95 percent to 19.2 percent. Data on net official inflows in the national accounts also demonstrate the importance of aid flows (Figure 2.11). Total official inflows continued to increase over the period 1998 to 2003. These levels of aid flows have had an impact on inflation, the exchange rate, andthe level and pattern of spending. '*Concessional loans are defined as loans subsidized at a low fixed interest rate, with certain eligibility criteria. In principle, only low-income countries are eligible for concessional loans, which are at a fixed concessional international rate as long as the country's per capita income is below a certain level. For instance, the rates o f the World Bank's International Development Association (IDA) and of International Monetary Fund (IMF) concessional loans are 0.75 and 0.5 percent, respectively. 21 Figure 2.10: Trends in Grants, ForeignLoans Figure 2.11: Total Aid Inflows, 1995-2005 andODA, 1970-2005(In percent of GDP) (Aspercent of GDP) 140 5 1 +Exports - o f ods and services 120 - 4- -NetOfficia/?Iow 0 - -Total +FDIfinancialflowpluexports 100 - / 80 - 60 - 40 - 1980 1983 1986 1989 1992 1995 1998 2001 2004 I..__.. Foreignloans -Grant --A-- ODA 1 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: World Development Indicators and authors' Source: NationalBank of Rwanda,National calculation. Note: Grants to government include unrequited, Accounts non-repayable, noncompulsory government receipts from other governments or internationalinstitutions.Foreign loans represent debt outstanding and disbursed. Net ODA disbursements equal gross ODA disbursements less principal repayments(amortization)ofpreviousODA loans 2.47 Aid flows and inflation. After the 1994 crisis, inflation was driven by a large growth in the money supply, due to the high level o f liquidity from high aid inflows, coupled with supply shortages. Inflation peaked at 41 percent in 1995, fell to 7 percent by 1996 (Figure 2.12), and rose again in 2007 due to higher money growth, accompanied by a huge budget deficit. The containment o f money growth and steady GDP growth helped inflation to fall by more than 10 percent in 1999 (Table 2.6). Inflation increased again by about 6 percent in 2000, following significant money growth and considerable appreciation o f the Rwandan franc. During2001 and 2002, inflation declined again, supported by the strong growth in GDP and tighter fiscal and monetary policy. However, a larger budget deficit accompanied by slow GDP growth and high money growth ledto further increases in inflation between 2002 and 2004. These trends, coupled with the significant nominal appreciation ofthe Rwanda franc in2004 andthe expansionary fiscal policy in 2003, reflected the challenges to monetary policy posed by the economic context following 1994. 22 Table 2.6: FactorsInfluencingInflation in Rwanda (Annualpercentage change) 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 1980-89 1990-99 2000-04 Real GDP growth 35.2 12.8 13.9 8.9 7.6 ' 5.9 6.7 9.4 0.9 3.7 3.2 2.1 5.3 Inflation rate 41.4 7.4 12.0 6.2 -2.4 4.3 2.9 2.3 7.2 11.9 4.7 15.1 5.7 Food price index 62.6 4.2 23.9 8.9 -14.7 0.9 6.0 -1.2 13.8 n.a 2.6 14.8 4.9 Narrowmoneygrowth 54.6 16.3 21.4 -6.9 10.7 2.6 15.0 12.6 12.0 14.6 3.9 14.8 11.4 Exchangerate depreciation 115.2 2.2 0.2 8.6 5.7 23.2 3.1 6.4 22.9 -2.3 -1.5 20.3 10.7 Budgetdeficit as a shareofGDP -5.6 -4.3 -1.4 -1.9 -4.4 -3.8 -3.6 -3.6 -6.3 -3.9 -0.4 -3.5 -4.3 Source: World Bank, World DevelopmentIndicators 2.48 In 1995, foreign aid to Rwanda surged.13 Inaddition, increases in M 2 (narrow money) resulted intremendous changes in nominal GDP and inflation. These changes were due to several factors. First, the Government introducednew notes and coins, which had the effect o f devaluing the national currency by 45 percent. Second, there was a huge increase inthe amount o f currency in circulation in a situation where people were forced to keep their money in very liquid form. These factors contributed to an average rate o f growth of narrow money o f 14.8 percent during 1990-99. This later fell to 3.9 percent over the period 1980-89 (See Table 2.6). 2.49 Moreover, in this period o f emergency and shortages, too much money was chasing too few goods and services, limiting investment in other assets. This contrasts with the difficulties in monetary management, due to large fluctuations inbroad money, mainly from aid flows, which at times has caused Government to delay spending until inflation is under control. In 2000, broad money grew by only 0.8 percent, while in 2002 it grew by 17 percent. These fluctuations in broad money reflected a higher real level of growth, several large new infrastructure related projects, and new banking products. In2003 and 2004, domestic credit to Government increased because of two additional large projects, elections, and a huge hotel investment. These fluctuations, together with unusually strong seasonality in currency in circulation, translated into an unstable money multiplier. l3ODAasa share of GDP was 18 percent in 1993, rising sharply to 95 percent in 1994, and declining to 54 percent in 1995. 23 Figure2.12: Rate of Inflation in Rwanda, 1995-2005 (Percentagechange) 60 -1011995 1996 1997 1998 1999 2000 2001 2002 2003 2004 20051 Source: InternationalMonetary Fund 2.50 Core inflation remained low up to 2004, but rising food and energy prices pushed annual inflation to 10 percent by the end of the year. Output growth turned out somewhat lower than expected and inflation higher mainly because of exogenous shocks, particularly related to shortages in electricity generation. By end-2005, inflation had dropped to 7 percent, including an impulse of2 percent from increasedelectricity tariffs. 2.51 To controlinflation and the growth in the money supply, the Government resorted to the sales of treasury bills. The increase in expenditure contributed to an increase in the money supply (with the rate of growth inthe money supply peaking at 15 percent by 2003). This contributed to a significant increase in inflation in 2003 and 2004. To curtail this increase, the Government sold treasury bills, which resultedin an increase in interest rates. As shown inTable 2.7 and Figure 2.13 these changes resulted in some transfer of resources from the private sector. However, credit to the private sector rose significantly in2005, following the decline innet credit to the Government that occurred in2003 and2004 due to a large inflow o f aid. Table 2.7: Inflation and Net Credit to Government, 1998-2005 (In unit indicated) 1998 1999 2000 2001 2002 2003 2004 2005 Inflation (average) -6.0 2.1 5.8 -0.2 6.2 7.7 10.2 5.5 Domestic CredidGDP (percent) 13.0 15.0 13.0 13.0 11.0 13.0 8.0 7.0 Annualpercent change Domestic Credit 9.9 12.8 0.7 0.3 -4.0 19.7 -22.7 -53.0 of which: Government 1.0 8.0 -9.6 -5.1 -12.4 10.7 -28.9 -16.4 Economy 8.9 4.8 10.3 5.4 8.5 9.0 6.2 11.2 Source: International Monetary Fund 2.52 The heavy reliance on sales of treasury bills as a means to sterilize liquidity in the economy may have compromised economic growth. The central bank was reluctant to sell foreign exchange and appreciate the exchange rate due to concern about the volume and 24 competitiveness of exports. Consequently, the Government relied mainly on sales of treasury bills as the main mechanism for sterilization o f the liquidity injected into the economy. The above figures and tables, coupled with the contraction in credit to the economy when credit to the -w- Government increases (Table 2.7 and Figure 2.13), suggest that this resulted in some transfer of resourcesfrom the private sector, and is likelyto have reduced GDP growth. Figure 2.13: Annual PercentChange in Domestic Credit, 1998-2005 12 , 15 - 10 10 5 8 .t *.. . , ',.* , 6 -- 0-5 -10 ' 6 -15 4 -- -20 2 0 I I I I I 1 1 ' -35 2.53 Aid flows and the exchange rate. The real effective exchange rate has been depreciating gradually despite increased aid flows (Figure 2.14), and this has also had some favorable effects on exports. Rwanda has gradually introduced more flexibility in its exchange rate regime, through the use o f a managed float. Foreign exchange bureaus are authorized to buy and sell foreign exchange at freely negotiated rates. Consequently, the exchange rate premium in the parallel market has gradually declined. For example, based on estimates from the National Bank of Rwanda, the black market exchange premium, which stood at 9.20 percent over the period 1995-2000, decreasedto 2.55 percent duringthe period 2001 to 2004. The real exchange rate has depreciated by about 12 percent since 1995, owing primarily to the weakening Rwanda franc, vis-a-vis the U S dollar. The franc fell by about 20 percent in nominal terms. Indications are that this has had a favorable effect on exports. 2.54 Other factors, however, may have outweighed the favorable effect of depreciation on exports. Although the volume of exports has been increasingsince the mid 1990s (except for a big dip during 2002-03 due to a large drop in coltan exports), the improvement in external competitiveness through a favorable real exchange rate movement has been offset by the 26 percent decline inthe export price index since 1995, 2.55 Inadditionto the effect ofthe depreciation, part ofthe growth inexport volume has been through recovery from its big fall in 1994.For example, Rwanda exported about 40 thousand tons o f coffee, annually, in the late 1 9 8 0 ~before the start of civil conflict. By 1994, coffee exports ~ had declined to 13 thousand tons, and have only started recovering inrecent years. The increase intotal commodity exports starting in 1999hasbeendue mainlyto the coltan boom. 25 2.56 Moreover, an examination o f Uganda's experience indicates that Rwanda's focus on a few key export commodities may also have dampened export growth. As noted, the negative effect on exports of the terms-of-trade shocks to the coffee sector were countered in the earlier part of the 2000s by increased trade in minerals. This improved Rwanda's ability to cushion itself against a terms of trade shock. In similar fashion, Uganda was able to cushion itself from the decline in coffee prices because of the diversified nature of its exports, including non- traditional exports such as particular fish, gold, cut flowers, and maize.14By 2004 and 2005, fish exports had overtaken coffee exports as the largest merchandise export item in Uganda. More recently, tourism, which has ridden steadily in importance after the mid-1990s' has surpassed these new exports. Tourism i s now Uganda's single largest foreign exchange earner. Figure 2.14: Actual and Equilibrium Exchange Rate -Comparisonof Rwanda and Uganda Rwanda Uganda Source: ElbadawiandKaltani 2006 2.57 This discussionunderscores the importanceof aid. Aid is the main source of capital flows, as well as the main source of public financing for investments (given the extremely low levels of foreign direct investment, as shown in Figure 2.6). Inthe balance of payments, capital flows are primarily capital grants and net borrowing (on concessional terms), and aid is the major source of finance for the budget. Therefore, until alternate finance sources are developed, or private investment significantly rises, the country will continue to be highly dependent on external aid. As a result, it i s important to ensure that available aid is put to the best use in terms o f investment, and that the macroeconomic effects o f aid are managed in a way that promotes growth. This will also require buildingthe needed institutions and skills to enable an assessment of returns from alternative investment options. These issues are discussed inChap. 4. l4Coffee exports drove merchandise export growth in the first half of the 1990s. But by the late 1990s, Ugandan coffee exports declined due to falling international prices, the occurrence of coffee wilt disease, and an unsuccessful Governmentpolicy of replantingcoffee trees. 26 Figure 2.15: Public Investment is Defined by Availability of Grants, 1990-2004 (Inpercentage of GDP) 4 5 2 2 0 8 6 4 3 2 0 - - 0 1994 1996 1998 2000 2002 2004 1990 1992 1994 1996 1998 2000 2002 2004 -1990- .1992 - . .Grant . Public investment - - - - - .Grant - Budgetaryinv. (CentralGov. capital formation) Source: World DevelopmentIndicators and International Monetary Fund 2.4 LONG-TERM MACROECONOMIC TRENDS 2.58 Population growth and capital formation will be the two key factors defining Rwanda's ability to meet its Vision 2020 targets. The rate o f population growth will affect development of the work force and therefore the rate at which poverty can be reduced. Closely related to this will be the need for skills development to enable a large percentage of the labor force to transition out o f subsistence agriculture. In addition, increased investment and capital formation will be neededto expand the level of absorption of the economy and lay the foundation for sustainedlong-term growth. 2.4.1 PopulationGrowth will Influence the Rate ofPovertyReduction 2.59 Sustained poverty reduction is going to require attention to population growth and employment creation. Data from the World Development Indicators (2005) indicate that per capita GDP growth was extremely high, at more than 8 percent, over the period 1998 to 2000. Population growth fell to an average o f 2.38 percent between 2001 and 2005. This rate is likely to decline gradually inthe coming years. Population growth rates inyears prior to 2003 were far higher than normal, following the decimation o f population that occurred in 1994. Ifpopulation growth were to average 2.2 percent between 2006 and 2020, then real GDP would have to grow at an averageannual rate above 7 percent to reach the Vision 2020 targets. Since 2000, Rwanda's real GDP growth rate has averaged about 5.3 percent. A lower population growth rate would lower the target GDP growth rate neededto reduce poverty and achieve the Government's target o fbecominga middle-income country by 2020. 2.60 The rate at which population growth declines will depend on policies and measures put in place. Inan optimistic scenario, population growth could gradually decline from 2.5 to 1.5 percent over the 2006-2010 period, for an average o f about 2 percent. A less optimistic projection would have it decline gradually from 2.6 to 1.8, for an average o f about 2.2 percent. 2.61 Inaddition to the population growth rate, attention to the age profile ofthe population and labor force is needed to ensure that it does not become a binding constraint to growth and poverty reduction. As a result of the genocide, the majority o f the population now consists o f 27 young people, yielding a high dependency ratio. Rwandans who are 15 years or younger constitute 45 percent of the population, and those 20 years and under make up 60 percent of the population. This contrasts with the 10 percent o f the population that are over 50 (Census, 2002). Rwanda has a greater proportion of its population below 20 years o f age than all of Sub-Saharan Africa (57 percent and 54.8 percent, re~pectively).'~A recent assessment of youths found that they were generally employed in informal sector activities such as petty trading and selling inthe markets. Some are engaged in domestic labor, barbering, and menial work such as street sweeping and trash collection. Others engage in begging, petty theft, and idleness. Inrural areas, youth are often concentrated in subsistence agriculture. Selected focus groups conducted among youth revealed feelings o f destitution, hopelessness, and powerlessness. Addressing the issue of employment among youth will be necessaryto ensure that this does not lead to significant social problems that could potentially destabilize the economy inthe future. 2.62 Raising the education attainment of young people will be critical to increasingtheir employability. Gross primary enrollment has risen rapidly to 131 percent o f the school age population (reflecting enrollment of overage students), and net enrollment has increased from 71 percent in 2001 to the current level of 95 percent. However, the completion rate remains low, at 47 percent, and only 15 percent of children attend secondary school. Vocational education programs are practically non-existent. Moreover, educational attainment is still below pre-1994 levels (Figure 2.16). 2.63 The Government's recent policy of fee-free nine-year basic education (with the aim o f doubly the transition rate from primary to the first year o f secondary school by 20lo), should help to address the issue o fthe general level o f educational attainment. Government has also combined this with policies to improve the quality of education through reforms inteacher training and an updated curriculum. In addition, the National Science, Technology and Innovation policy, adopted in 2005, will eventually lead to a change in curricula to support the transition to a knowledge-based, technology-led economy. However, programs geared to on-the-job training and vocational training are critical to meet the current demand for skilled labor, the shortage of which, as shown will be inChapters 3 and 6, is already beginning to pose a bindingconstraint to firmproduction. l5 Datafor Sub-SaharanAfrica is basedon US.CensusBureau, GlobalPopulationProfile (2002), anddatafor Rwanda is from the 2002 PopulationCensus. 28 Figure 2.16: EducationalAttainment of 15-19year olds Still Below Pre-1994 Levels fPoDulation share) 1, I I 0.9 0.9 0.8 0.8 5 0.7 0.7 6 $ 0.6 .I 0.6 0.5 4n 0.4 0.5 0.4 2 0.3 3a8 0 0.3 L 0.2 0.2 0.1 0.1 0 X . - K . . Z " I 0 2 4 6 8 10 0 2 4 6 8 10 _. - Grade Grade -Urban1992 _ _ _ _ -n- Rural1992 Female 1992 Urbanzoo0 + - -- -.-Male2000 _ _ Male 1992 - ~ Rural2OOO Female 2000 Source: http:llww.worldbank.org/research/projectsledattaid. 2.64 Finally, labor mobility and rural-urban migration will also have an impact on the rate of growth and poverty reduction. Growth and poverty reductionwill occur at a faster rate iflaborisableto move into activitiesthat aremoreproductive, particularly inthenon-agriculture and urban sectors. According to the 2002 Population Census, 17 percent o f the population lives inurban areas. This contrastswiththe 6 percent ofthe populationthat lived inthe urban areas at the time of the 1991 Census. The more than doubling of the proportion of population living in urban areas in a decade is due to both rural-urban migration and changes in urban borders, leading to urban sprawl. Migration to urban areas has been predominantly a male phenomenon. The population growth rate for Kigali is estimated to be 6 to 7 percent, resulting in a tripling of the population between 1994 and 2002. The number o f males in urban areas exceeds that o f females by close to 13 percent. This phenomenon i s particularly strong inthe 20-54 age groups; infact, inthe 30-49 year old age group, males inurbanareas exceedfemales by 50 percent. 2.4.2 Investmentand CapitalFormationDriven by the Private Sector 2.65 Although public investment as a share of total expenditure has declined over the past decade (Figure 2.17), capital formation has been on an increasing trend, from its level of 13.4 percent of GDP in 1995 to 20.8 percent in 2004 (Figure 2.18). The increase in overall investment, particularly since 1997, has been drivenby the increase inprivate investment, giventhe decline in public investment duringthe same period. National accounts figures show a steady rise inprivate sector investment, from 5 percent o f GDP in 1995 to about 12 percent inrecent years, which is in line withthe SSA average (excluding South Africa) of 12 percent o f GDP. 29 Figure 2.17: Public Investment, 1995-2005 Figure 2.18: Capital Formation Driven by the (as a share of total expenditure) Private Sector, 1995-2005 (as share o f GDP) 145 25.0 20.0 .... . - ...,*...-......*. .*.-. ::: 30 15.0 10.0 202515 - 5.0 10- 0.0 5 - 1995 1997 1999 2001 2003 200: 0 .-. ...Grossdom.invest, . -Gross prkinvest. -Gross public invest. Source: InternationalMonetaryFund Source: World DevelopmentIndicators2006 2.66 Central government investment has also started to show an upward trend over the past two years. Public investmentbegan to increase in 1995,but fiscal restructuring and falling donor financing led to a slump in this investment between 1997 and 2002. More than 80 percent of Rwanda's development budget i s financed through foreign aid, and this aid drives public capital formation. In general over the period, the trend in capital goods imports has been in line with the trend in private investment, and appears to be associated with investments in equipment (Figure 2.19). Figure 2.19: Share of Capital Goods Import our of total Equipment Investment increasingover time (unit as indicated) x)O 90 80 70 60 50 40 30 20 x) 0 " .- """" "" " 1990 1992 1994 I996 1998 2000 2002 2004 890 1992 2994 2996 2998 2000 2002 2004 .-- - .hp o r t s ofcapitalgoods as a %oftotalinv. -hports ofcapitalgoods as a %oftotaleqipment mv. Source: World DevelopmentIndicators2006 2.5 SUMMARY 2.67 The discussion in this chapter has highlighted some key issues that will be examined further inthe remaining chaptersFirst, the discussion will show that peace is critical to continued economic growth, and that unrest and political instability would derail that growth. This highlights the importance of ongoing peace and reconciliation programs, as well as the importance of regional cooperation, investments, and projects to buildregional trade. Second, the 30 fact that the rate of economic growth has been falling over time suggests that part o f the growth observed during 1995-2005 was due to a surge in economic activity made possible by peace and political stability, and that the average growth rate for the next decade is not likely to be as high. Third, sustained growth will require diversification of the economy away from agriculture and low-value products, particularly given its landlocked location and the needto move labor out of agriculture. Fourth, the discussion highlights the importance of aid for financing the overall growth process. Inthe short term, this means using aid to lay the foundation for development of the private sector and the transition of labor out of agriculture; and to ensure that the inflow of liquidity does not destabilize the economy. Inthe long term, however, sustainable financing of growth will needto come from the private sector. 2.68 Interms offuture sourcesofgrowth, thetrends pointto the potential ofthe miningsector, as evidenced by its growing importance in exports, and the expanding demand from Rwanda's trade partners in Asia. Since the mining sector in Rwanda i s characterized by small producers, this would be an opportunity to develop alternate sources of rural income. Also, the service sector is beginningto show signs o f rebounding to its former growth trajectory. However, the risk is that benefits from growth may not accrueto households inrural areas. Currently, there is a danger that the benefits will remain mainly in the urban areas or contribute to worsening o f income inequality. The analysis in the rest o f the CEM will explore these issues further, beginningwith an examination ofthe sources o fgrowth. 31 3. GROWTH: SOURCES,CONSTRAINTSAND CHALLENGES 3.1 INTRODUCTION 3.1 This chapter provides an overview o f Rwanda's progress in generating growth over the past decade. The discussion focuses on constraints to growth, and on the supply and demand-side sources o f growth. It also examines the role of exports for growth, to 2020, and the implications o f factor accumulation and productivity. 3.2 The analysis indicates that the main constraints to growth are the poor state of infrastructure (related mainly to transport, energy, and water), the scarcity o f skilled labor, and lack of access to finance. The poor state of infrastructure poses the most binding constraint to growth. Followingthis, the scarcity o f skilled labor needs to be urgently addressedto ensure that itdoes notunderminelong-termgrowth. Investments inthe social sectors have beenwell placed, judging from the high rate o f returns. However, the increasing trend o f investmentas a share of GDP, in the face of declining GDP growth, particularly for the agriculture sector, suggests that investmentmay not be goingto areas that can producethe highest economic returns. 3.3 Analysis of sources of growth from the supply side shows that agriculture (primarily food crops) has been the main driver of growth. Exports have not contributed substantially to growth over the past 10 years, although they are a key source o f foreign exchange. Growth has therefore been broad based, in that it has focused on the sector where the majority o f the population is engaged. 3.4 On the demand side, growth has been driven by domestic absorption. External demand represented a residual amount to overall growth. Within domestic absorption, private consumption has been responsible for almost 60 percent of the growth, followed by real private investment. Growth in GDP overshadowed growth in private consumption by 2 to 3 percentage points. By the period 2002-2004, the growth rate of domestic absorption and external demand had fallen significantly, with growth inprivate consumption outweighing overall GDP growth. 3.5 Inthe case offactor productivity analysis, the results indicate that muchofthe growth has been due to a bounce-back effect following 1994, as the political and social situation improved and factors o f production, especially labor, moved back into productive activities. However, a significant amount of the growth was also due to liberalized trade and a deepenedfinancial sector. The analysis suggests that the binding constraints related to the poor infrastructure network have compromised the gains that could have beenreapedfrom trade and financial reforms. 3.6 The chapter therefore shows that investments for sustained growth to reach the Vision 2020 targets should focus on infrastructure and on human and physical capital accumulation. Growth has been declining over time, partly due to the once-and-for-all nature o f the economic recovery. However, part of the high growth has beena result of the first generation of reforms to liberalize the market. As the economy stabilized, it was inevitable that growth would begin to decline compared to levels observed in the late 1990s. Nevertheless, significant growth gains could still be realized by easing infrastructure constraints and strengthening the links between 32 farmers and manufacturing or export chains, while maintaining investments to upgrade the availability of skills. 3.7 The chapter begins with an analysis of the constraints to growth followed by an examination of the sources o f growth fiom four perspectives: demand side, supply side, growth accounting, and exports. The chapter concludes with a discussion o f the main results, which provide a rationale for the areas to be analyzed inChapters 4 to 7. 3.2 ANALYSIS OF THE CONSTRAINTSTO AND SOURCES OF GROWTH 3.8 The analysis of the constraints to growth uses the framework outlined in Haussman, Rodrick, and Velasco (2004), which outlines an approach for formulating hypotheses on what may be constraining growth. It aims at identifyingbindingconstraints to economic activity. The methodology can be conceptualized as a problem tree, which begins by asking what keeps the level of domestic investment and entrepreneurship low (Figure 3.1). I s it low returns to investmentor accumulation; or is it due to the highcost o f or inadequate finance? 3.9 In this framework, constraints to growth are derived by examining the contribution of private returns to domestic investment or accumulation, against the cost o f financing that investmentor accumulation. The private returnto accumulationcould be constrained due to a low rate of return to factors of production and low appropriability (i.e., the proportion o f returns investors are able to capture). A high cost of financing accumulation or a high effective rate of interest relevant for investment decisions could also lower the level of privatereturns. Figure 3.1: Diagnostic Decision Tree -Constraints to Growth in Rwanda domesticinvestmentlow7 L i Low privaterenunto Highcost offmanceor accumulation accumulation m Market failures F Coordinationeaemalities Source: Haussman, Rodrik, andVelasco (2004). Note: Factorsin bold denotethose areas identifiedto be relatively moreof aconstraintto growth inRwanda. 33 3.10 Given the above framework, a binding constraint i s defined as a distortion that has the greatest direct adverse impact on economic growth. A distortion i s seen as the most bindingto growth if its removal produces the largest economic gains compared to any other policy measure. The main goal o f identifying the most binding constraint is to facilitate the Government's prioritization o f reforms; it does not discount the fact that other equally important constraints may exist. Moreover, the binding constraint may not necessarily coincide with what economic agents perceive to be the largest distortion in the economy. The emphasis i s rather on the impact on economic growth of relaxing the most bindingconstraint. Evidence usedto determine the impact of the constraint can be direct or indirect. Direct evidence would involve demonstrating that relaxation or tighteningofthe constraint causes sizeable movements inthe growth rate. Ifthere is no substantial movement in the constraining factor, there should be limited movement in the growth rate, and the related shadow price would be high. Indirect evidence would involve demonstrating the existence of a high prevalence o f activities designed to circumvent the constraint, or ruling out other potential constraints based on their demonstrated limited effect on growth. 3.11 In the Rwandan context, there are a number of impediments to development of the private sector that undermine private returns or accumulation and increase the cost of finance. The DTIS identified the poor state of infrastructure and low production as major impedimentsto trade. Subjective measuresbased on firm surveys rank the state of infrastructure, access and cost o f finance, tax rates, and skill of workers as higher business obstacles compared to factors that affect appropriability, such as macroeconomic instability, corruption, and anti-competitive practices (see Figure 6.9 later in this volume). Furthermore, the Government is working to improve security o f property rights in both rural and urban areas, which could be a major deterrent to appropriability. Specifically, a land law was adopted in 2005 and Government has actively worked since then to make it operational. Therefore, the examination o f constraints to growth in Rwanda focuses on the areas related to the state of infrastructure, access to and cost of finance, the tax rate and administration, and availability of skilled labor. 3.12 The analysis to follow inthis chapter also shows that the poor state o f infrastructure and low productivity are the most bindingconstraint to growth. The analysis begins with an overview o f issues related to the cost and access to finance, which show that it is on par with regional levels. Therefore, the cost of finance i s not the key factor that i s puttingRwanda at a competitive disadvantage with respect to neighboring countries. By contrast, an examination o f indicators o f returns to private accumulation and the factors that determine returns to private accumulation strongly suggests that these have been the most constraining to growth in Rwanda. Inparticular, low productivity and technology adoption, poor geographic location, and the state o f infrastructure are identified as the major constraints. 3.2.1 Cost and Access to Finance as Constraintsto Growth 3.13 The high cost of financing accumulation or the effective rate of interest relevant for investment decisions could be high due to: (i)low domestic savings; (ii)poor financial intermediation; and/or (iii) lack of adequate access to international financial markets. Domestic savings are extremely low in Rwanda, at less than one-tenth of the average for Sub-Saharan Africa, or below 2 percent of GDP (see Chapter 2, Figure2.4). Low availability o f savings leads to low availability o f funds to lend, and therefore to a higher cost of finance. Access to 34 internationalfinancial markets i s limited by Rwanda's ability to earn foreign exchange, based on its low export base. This requires the country to rely on concessional lending and grants. The weak state o f financial intermediation is manifested inthe relatively low level o f access and high cost of finance. 3.14 Although interest rates do not appear to be exorbitant, firms perceive the cost of finance to be an obstacle to entrepreneurship and business development (see Figures 3.2 and 6.9). Interest rates have been generally reasonable and stable. However, the cost o f finance is higher for smaller firms and firms in the informal sector. For these firms, credit i s typically short term. Smaller firms that are able to access commercial loans face higher interest rates due to limited competition among commercial banks; segmentation o f loanable funds such that large firms pay 12.5 percent interest, while smaller firms pay up to 17 percent interest; a high central bank rate (of around 14.5 percent); a relatively high rate of nonperforming loans (which currently stand at 19 percent); and the high cost of external resources. Interest rates for long-term credit range between 15 and 16 percent per annum, in addition to a guarantee commission of 3 percent. In contrast, interest rates for smaller firms are 2 to 3 percent per month. Banks generally focus on short-term trade financing to established customers, and make only limited credit available to finance medium and long-term investments which accounted for approximately 20 percent of total bank credit in2004. Figure 3.2: Interest Rates inRwanda, 1999-2004 (In percent) Source:Ministryof Finance and Central Bank of Rwanda 3.15 Currently, access i s defined by the fact that the financial sector i s small, weak, and reluctant to finance commercial activities that are not directly related to the export sector or that do not provide hard currency. Only 4 percent of the population has bank accounts, and overall financial services account for 3 percent of GDP. A 2005 survey o f manufacturing firms (the Rwanda Industrial and Manufacturing Survey) showed that 54 percent of firms identify the cost o f finance as a constraint to business. Low access to finance was reported as a constraint by 34 percent of manufacturing firms in 2005, based on the Rwanda Industrial and Manufacturing Survey (RIMS, see Chapter 6). In addition, a qualitative analysis o f the SME sector found that 35 more than half o f the SMEs interviewed cited access to finance as a constraint. Most relied on family and personal savings to finance business ventures.16 3.16 In terms of sectors, loans go mainly into trade, tourism, property development, and manufacturing, and are focused on urban areas. The 2005 Financial Sector Assessment Program (FSAP) found that agriculture, by far the largest sector o f the economy, received only 2.3 percent of bank credit in 2003 (compared with about 10 percent in Kenya and 14 percent in Tanzania), notwithstanding the existence o f Union de Banques Populaire de Rwanda (UBPR), which is explicitly charged with provision of credit to the rural areas. Most bank loans in 2004 went to trade and tourism (38 percent), with 20 percent to the manufacturing sector. Furthermore, 80 to 90 percent of banking transactions take place inthe capital, Kigali. Only UBPR is represented in all of the eleven prefectures. Therefore, regions and sectors have benefitedvery unevenly. 3.17 Although access to finance i s currently limited to the formal banking sector, some agriculture producers are able to gain access to financing through contract farming or through commodity chains that have a well-defined or integrated supply chain. One example o f this is the tea sector, where farmer cooperatives and associations sell their production to processing plants. The contract arrangementenables farmers to gain accessto inputs. 3.18 Farmers and rural micro-enterprises are able to gain access to short-term credit from microfinance institutions. Since 1994, more than 65 microfinance institutions (MFIs) have been established in Rwanda, and offer micro-finance services to more than 200,000 micro- entrepreneurs. The FSAP reports that according to data available as o f December 31, 2004, the microfinance industry in Rwanda had reached about 590,000 clients with deposit services. This represents a penetration rate o f 13 percent of the active population, which i s high by international standards, and is growing.17 However, the industry requires coordination and structure. The central bank has recently initiated measures to register, supervise and regulate these institutions. While credit to the economy has grown rapidly in recent years, provision of credit to the private sector remains narrowly concentrated inKigali and inspecific sectors. 3.19 The FSAP shows that improved access in rural areas requires addressing a number o f issues, some o f which fall under the purview of the financial sector, and some o f which do not. First, within the formal sector, capacity building for financial institutions is a key area that could lead to better access for farmers and microenterprises. The majority o f micro-finance institutions in Rwanda seem to lack capacity in several areas that are important to manage a financial institution profitably. The FSAP notes that micro-finance institutions have expressed a need for capacity building in management, accounting, internal controls, development of new products, and setting up management information systems (MIS). A few institutions in Rwanda, such as Aquadev'* and the Rwanda Microfinance Forum (RMF), are able to provide technical assistance. l6 Nugawela et a1(2004). 17 There is anecdotal evidence from the registration o f micro-finance institutions at the central bank, that the economy i s becoming increasingly monetized. l8 Aquadev is a Belgiannon-governmental organization (NGO)created in 1987with the aim o f (i) engaging with local stakeholders in the elaboration o f poverty reduction strategies; and (ii)supporting the development o f interventions focused on the most vulnerable and poor inWest and Central Africa. Its areas o f focus are microfinance, food security and nutrition, environment, and water resources. See also mnv.adbanking.org. 36 RMF, inparticular, provides training and other capacity buildingopportunitiesto its membership o f about 44 micro-finance institutions, and i s planning to establish itself as an advocacy group in the absence of a professional association ofmicro-finance institutions. RMFhas been involved in discussions on the Government's microfinance policy and on taxation issues in microfinance. However, it remains a fragile institution with limited membership and donor support, which restricts its activities. 3.20 Some measures needed to improve access to finance fall outside of the purview of the financialsector. First, there is a need to develop bankable projects inthe agricultural sector, and to develop other types of collateral besides registered land and real estate. Second, the agriculture sector is prone to a high level of risk, which i s compounded by factors that undermine productivity (see Chapter 5), such as lack of skills for business development, and low use o f improved inputs and methods. 3.21 Despite these impediments, indications are that the microfinance industry i s becoming more competitive, and this should improve access. There are already indications that some microfinance institutions will reduce their interest rates inthe face o f increasing competition from a number of commercial banks, which are trying to reach lower segments o f the market. Measures to develop new financial products for microenterprises and small firms should help to improve access and reduce costs. Similar approaches have been used in other countries to reach microenterprises and firms operating in informal sectors (see Box 3.1). Source:www.newsroom.barclavs.co.uk/content/detail.asp?ReleaseID=~33&NewsArealD=2. For a discussion of other forms of susus, androtatingsavingsand financial schemes in Ghana, see also World Bank (2004). 3.22 These factors, coupled with a comparison o f trends in credit to the economy with that o f GDP growth, suggest that the most bindingconstraints may be supply constraints that undermine private accumulation. The sharp fall in GDP in 2003 coincides with a drought that occurred in 37 f i n * * vk 8 $ 6 38 4 1 3 a 2 3.23 The trend in the investment rate relative to GDP growth suggests that determinants o f private accumulation may be an important constraint to GDP growth. Gross investment as a share o f GDP has increased slowly but steadily since 1998 (Table 3.1). Much o f the increase came from the public sector, although over the past two years the private sector has marginally increased its contribution. This trend i s also supported by the percent change in domestic credit to the economy, which has declined over the two periods, with much o f the decline coming from the government sector. Credit to the economy has decreased over time, mainly because o f a reduction in credit to government. The increase in credit to the private sector has not fully offset this reduction. However, ithas enabled an increase inprivate investment between 1998-2000 and 2001-2005. Despite this, GDP growth has continued to decline. Since increased investment has been associated with a reduction in average GDP growth rate between the two periods, it implies that there either are other constraints hindering returns to investment, or investments were not put where the highest returns could be gained. The discussion here supports the view that the most bindingconstraints may be associated withthe low returns to private accumulation. Table 3.1: GDPGrowth Rate and Investment inRwanda Average Average 1998-2000 2001-2005 GDP growth 7.5 5.2 (Percent change) Domestic Credit 7.8 -11.9 Of which: Credit to Government -0.2 -10.4 Credit to Economy 8.0 8.1 (as share of GDP) Domestic Credit 13.6 10.4 Gross Investment 16.5 19.3 PrivateInvestment 10.2 12.2 PublicInvestment 6.7 7.4 Source: National Bank of Rwanda, WDI, and Staff estimates, 2005. 38 3.2.2 FactorsDeterminingReturnsto Private Accumulation 3.24 A low rate of return to accumulation could be due to low appropriability or a low rate of return to factors of production. Low appropriability is definedbythe following factors: (i) taxrates;(ii) high inefficient tax structure; and (iii) expected expropriation risk (insecure high property rights, corruption, and macroeconomic instability). Low rate o f return to factors of production is defined by: (i)large externalities, spillovers, or coordination failures related to institutions and servicesthat need to be inplace before growth can take off; (ii) low productivity, low technology adoption, or weak public incentives; and (iii)poor geographic location, inadequate infrastructure, hightransport cost, telecommunication, or shippingcosts. 3.25 InthecaseofRwanda, itcanbearguedthatthelowrateofreturntofactors ofproduction pose a greater constraint to growth than low appropriability. As previously discussed, the Government i s working to improve the security o f property rights through implementation o f the Land Law. In addition, macro-instability and anti-competitive behavior are ranked low among constraints faced by firms in the manufacturing sector. The main obstacle to the business environment i s related to the high transaction costs of starting and doing business. Costs of startinga business stemmainly from procedures requiredfor licensingand fees. Table 3.2: Cost of Startinga Business,2006 Region or Economy Procedures Duration Cost Min.Capital number days %GNI %GNI per capita per capita Rwanda 9 16 188.3 0.0 Tanzania 13 30 91.6 5.5 Uganda 17 30 114.0 0.0 Kenya 13 54 46.3 0.0 East Asia & Pacific 8 46 42.8 60.3 OECD 6 17 5.3 36.1 Sub-SaharanAfkica 11 62 162.8 209.9 Source:DoingBusiness 2007. 3.26 In the case of taxes, rates appear to be in line with regional levels, and administration is relatively effective. Sustained reforms have led to improved revenue performance. Government revenue, estimated at about 13 percent of GDP in 1992, fell below 4 percent in 1994 and increased thereafter. By 2002, revenues stood at 12.2 percent o f GDP, reflectingthe increaseinthe VAT rate from 15 to 18 percent, and the substantial strengthening o f revenue administration. Revenue as a share of GDP rose further in the following years, and is now at around 15 percent. This is in contrast with the SSA average of 22 percent in 2004. This increase in revenue reflects the impact o f reforms to the income tax system that went into effect in2003, includingthe taxation of in-kindbenefits. Finally, issues surroundingtax administration and corruption are ranked low among constraints faced by manufacturing firms (only 27 percent o f firms rank it as high, compared to 55 percent in Tanzania and 36 percent in Uganda). Table 3.3 also indicates a relatively low burden o f tax administration in Rwanda. Appropriability is therefore does not appear to be the most binding constraint leading to low returns on accumulation. 39 Table 3.3: Comparisonof Tax Burdenacross Countries Region or Economy Payments Time Total tax rate number hours %profit Rwanda 43 168.0 41.1 Tanzania 48 248.0 45.0 Uganda 31 237.0 32.2 Kenya 17 432.0 74.2 East Asia & Pacific 30 290.4 42.2 OECD 15 202.9 47.8 South Asia 30 304.6 45.1 Sub-Saharan Afkica 41 336.4 71.2 Source: DoingBusiness2007 3.27 The low return to accumulation is likely a result of the low returns to factors of production. Several factors may be contributingto this outcome, beginningwith the structure of the economy and the high dependence on subsistence agriculture, which increases the vulnerability to external shocks. Other factors are the overall state of infrastructure, which increasesthe cost o f doing business; and the geographic location of the country (landlocked, and in a neighborhood prone to conflict), which influences perceptions of the country and curtails foreign investment. 3.28 Land is a key factor of production, and its productivitycould be improved through better inputs and irrigation. The analysis to follow, on the sources o f growth, highlights the hugeweight of staple food production (and therefore land as an asset) inthe growth or production process. Yet, as Chapter 5 will show, irrigation and input use are very low in Rwanda. Untila year or two ago, there had been limited investment in land to improve its productivity. Thus, the data presented in Table 3.1 reflect, to a degree, the low level o f investment in land to improve productivity, which i s manifested by the low level of research, extension, and input use, and poor water management. The sources of growth analysis to follow, as well as the analysis in Chapter 5, strongly indicate that investment inthis asset could help to sustain the growth trends observed over the past decade. 3.29 There is a need to improve the overall business environment in order to improve investmentand growth. The presenceof a large informal sector, estimated to be between 30 and 40 percent of GDP, is also an indication of the high costs and structural rigidities inthe business climate that lead to higher costs o f doing business. Chapter 6 shows that the main factors influencing the business environment are related to three key factors. Specifically, the (i)overall poor state of infrastructure; (ii)low stock of human capital, in relation to labor supply and productivity; and (iii)geographic and landlockedlocation of the country. OverallPoor State ofInfrastructure 3.30 The poor state of infrastructure is manifested in the poor state o f transport, energy, and water supply, which imposes significant costs on the economy and represents a serious impediment to improving per capita income. The road network is generally in disrepair and poorly maintained, mainly due to low human capacity and resources. This translates into high 40 transport costs and therefore high trade costs. Inthe case of water, regional variability inrainfall, coupled with limitedirrigation and poor management capacity, have beenkey factors determining agriculture productivity. Low water supply has also curtailed agro-processing in some sectors, such as coffee washing stations. Finally, the current energy crisis has increased the price o f electricity generation, and that, in turn, has led to an increase in production costs o f up to 16 percent for some firms. 3.31 In the case of transport, the overall road network is in poor condition and largely unmaintained. A recent sector survey concluded that the share of main roads with international importance (basically all 1,100 km of non-urban asphalt roads) that were in good condition increased from 23 to 41 percent between 2002 and 2005.19 However, despite modest progress over that period, barely 6 and 5 percent of the secondary and communal networks, respectively, were in good condition in 2005, and had only seen negligible improvements over the past three years (Table 3.5 and Table 3.6). Table 3.4: Rwanda -Share of Classified Roads in Good Condition (Inpercent) 2002 2005 International asphalt roads 23% 41% Non-asphalt roads of national interest 5% 6% Non-asphalt roads o f communal interest 2% 5% Source: MININFRA. 3.32 There have been recent improvements on the asphalted network, but this will not be sufficient to reduce the disadvantages o f the country's landlocked position and facilitate market access of agricultural products through feeder roads. About 60 percent ofthe classifiedroads and a much higher percentage of non-classified roads require heavy rehabilitation or periodic maintenance, since most have not been maintained for more than a decade. In addition, the condition o f a significant part o f the communal network and of the non-classified roads has not been monitored and is therefore not known, which makes planning and prioritization very difficult. While priorities within road infrastructure have been defined, the road maintenance program so far has not been able to follow these priorities because o f insufficient resources and inadequate planning of road works. To improve planning, an initial pre-requisite would be to improve the database on the state o froad infrastructure. 3.33 The poor state o f the road network, coupled with the landlocked position of the country, which is roughly 2,000 km from the nearest seaport, means that Rwanda is burdenedwith high transport costs. The latest estimates are that transport costs represent nearly 50 percent of export and import values. This level of cost implies transport costs o f FRw 2,200 to 2,500 for an imported 20-foot container; and FRw 2,175 to 1,950 for a container exported to Dar Es l9 andWorldBank(2005).PrivateSolutionsforInfrastructureinRwanda,ACountryFrameworkReport. The PPIAF report provides road condition data for 2002, quoting earlier technical reports. Road condition data for 2005 are MININFRA. 41 Salaamhlombasa. With Rwanda's mountainous terrain and excessive rainfall, erosion i s particularly severe on the road network, the maintenance cost o f which is twice that of most Sub- Saharan African countries. The country's total road network of 14,000 km, spread over barely 27,000 square kilometers of nationalterritory i s among the densest in Sub-SaharanAfrica, and far exceedhuman and financial capabilities. Table 3.5: Extent of Classified Road Network in Rwanda Data Source Type ofRoad RIPA report by EUreDortbv Ministryof AIPA Scetauroute Infrastructure (mid-2002) (Jan. 2002) (Dec. 2002) Pavedmainroads 930 1,022 1,100 Unpavedmain roads 4,436 4,386 4,250 Unengineeredgravel roads 1.75 Total 5,366 5,408 5,350 Source: Africa Institutefor Policy Analysis and Economic Integration, 2002; Scetauroute, 2002; and Ministry of Infrastructure. 3.34 Reducing transport costs through increased maintenance and rehabilitation would reduce trade costs and promote growth, and has the potential to significantly reduce poverty. The Diagnostic Trade Integrated Study (DTIS) reports that in the coffee sector, rural transport costs alone; i.e., transport from the farm gate to Kigali (via the washing station), i s estimated at 40 percent o f the farm gate price. The report estimated that cutting these costs by half would translate into a 20 percent increase in coffee producer prices, and lead to a 6 percent reduction in the poverty rate among coffee producers. Another example of the potential impacts o f reducing transport costs pertains to poor maintenance of the road between Gitarama and Kibuye, which caused vehicle operating costs to rise from US$ l.OO/km in 1989 to US$ 3 . 4 0 h in 1996. Rehabilitation of the road ledto a 50 percent reduction in vehicle operating costs and an overall reduction intransport costs o f about 40 percent, which has facilitated the move from subsistence to market oriented production inthe area. 3.35 However, the ongoing decentralization reforms may make maintenance of roads more difficult. Responsibility for maintaining the unpaved national road and communal road network i s being transferred from the central government to the provincial governments. Maintenance o f paved roads is carried out by the private sector. Low capacity at the local level for road maintenance may therefore compound the problem. Transfer o f responsibility o f unpaved road maintenance to the district will needto be accompanied by a strategy for capacity development at the local level. 3.36 A comprehensive strategy for road maintenance and rehabilitation is neededto facilitate the setting of priorities, assigning of responsibilities, and development of a framework for engaging the private sector and raising needed resources. The Government's policy for the road sub-sector has not yet been fully articulated, but its core goals are clear. The central thrust is to (i) enhance Rwanda's integration into the regional economy by improving the condition - and subsequently the extent - o f the country's national road system; and (ii)improve the availability and quality of local road infrastructure, in order to provide better access to markets. The 42 promotion of private sector participation in road rehabilitation, maintenance, and development is a central objective in pursuit o f these goals. However, a direct consequence o f the current decentralization policy i s likelyto be that overall maintenance o f unpavedroads - and the quality -will becomevery difficultto control. 3.37 Broadly in line with the strategy, the current activities under the transport Medium Term Expenditure Framework cover the five modes o f transport (air, land, rail, river, and lake). However, these will needto be sufficiently prioritized, given scarce resources and limited human and technical capacity to implement. Inaddition, most of the country's transport legislation dates from the colonial era and needs to be aligned with new and emerging challenges, including the country's integration with the East African Community, and the increased move to decentralization. While priorities within road infrastructure have been defined, the road maintenance program so far has not been able to follow these priorities because o f insufficient resources and inadequate planning o f road works. In other transport sub-sectors, ambitious plans are still being pursued in airport improvement (construction of a new international airport), establishment of a railway link, and revitalization o f lake transport on Lake Kivu. However, the economic and social viability of large airport and railway investments should be clearly assessed to ensure that the highest growth benefits result from particular investments and costs. 3.38 Given that transport modes other than roads play a very minor role in Rwanda, it is critical that economic and social benefits, as well as the feasibility o f alternate transport investments, be rigorously assessed. The country has two international and three domestic airports, handling a passenger volume of about 150,000 and freight of about 5,600 tons annually. Some airport infrastructure improvements (increase of storage facilities, new taxiway) may be needed to support exports, particularly at the main airport of Kanombe at Kigali. However, the focus interms of transport infrastructure should be on the road network, particularly the unpaved network that i s critical to rural areas. In general, airport facilities will be sufficient for some time to come, even if tourism increases and some niche markets for air-freighted exports can be developed. The previous Airports Authority was transformed into a Civil Aviation Authority (CAA) in 2005, and the eventual separation o f airport infrastructure management and civil aviation oversight functions is planned. However, to date these functions are all combined inthe new CAA, and only ancillary activities such as restaurants have been outsourced. A study on re- activation of lake traffic on Lake Kivu will be launched in 2007, but i s unlikely to lead to any significant public expenditure requirements. However, two project concepts are under consideration that may require larger financial resources, potentially including public support. One is the construction of roughly 400 km rail link to the railhead Isaka in Tanzania, and the other one is a project to make the river Kagera navigable. This would link Rwanda to Lake Victoria. The viability and feasibility of these projects would need to be established through detailed studies. 3.39 The railway system could helpto facilitate integration into regional markets, with the aim of reducing the cost of bulk transport, but the costs and benefits need to be carefully assessed. Rwanda is currently not served by any railway system. The Government's key policy objective for rail is to investigate the feasibility of establishing a railway to connect the ports of Dar es Salaam or Mombasa through one o f the other regional railway systems. The Government i s also considering the possibility of establishing a southern transport corridor linking the countries of the Great Lakes region with South Africa, as part of the Great Lakes Railway Project. The main 43 concern, however, relates to traffic potential. While rail transportation offers lower rates, as much as 50 percent less than road transport, use has plummeted due to poor operational performance. Building a railway link to Kigali i s a potential long-term approach; but the investment required is only likely to be justified if trade volumes increase substantially and the efficiency of rail operators improves. There i s significant uncertainty about whether what would be a large investment could be justified given the current and projected volume o f Rwanda's exports and imports. 3.40 Insum, for the transport sector, rehabilitation and maintenance of the road infrastructure should be the main area of focus for immediate growth benefits. Emphasis should be on the elaboration of a road maintenance plan that takes into account existing capabilities and aims to improve access of rural households to markets. The plan should also helpto define the necessary role o fthe private sector to ensure sustainability. 3.41 In the case of energy infrastructure, Rwanda faces four fundamental issues (i) overwhelming dependence on wood fuels; (ii)very high transport costs o f petrol products; (iii) the unavailability of additional inexpensive hydropower resources; and(iv) the poor performance o f the national utility, Electrogaz, which supplied both electricity and water. There i s currently limited availability of alternative energy sources, resultingin increased costs for both households and firms, and insubstantial constraints to production. Traditional fuel supplies are inefficient for industrial and modernized production, and are quickly becoming depleted. Wood, peat, and charcoal make up 90 percent o f the energy used in Rwanda. To date, only 1 percent of energy use i s sourced from electricity and 9 percent from petrol. Results from the 2001 EICV dataset indicate that less than 2 percent o f Rwandan households have electricity, and o f that 2 percent, many households and business must rely on expensive generators as back-up for frequent blackouts. The majority of the population relies on charcoal for energy, and wood for construction material. 3.42 Rwanda has among the lowest per capita consumption of electricity and petroleum products in the world, resulting from the low availability and lack o f access to viable energy sources. Compared to estimates based on the EICV dataset, more recent data from Electrogaz indicate that about 75,000 customers (roughly 6-8 percent of households and firms) have access to grid-suppliedpower, almost entirely inthe main cities. The capital, Kigali, alone accounts for more than 70 percent o f total low-voltage electricity consumption. Diesel remains the primary fuel for self-generation in rural areas and for backup in urban areas, and kerosene i s used for lighting by the vast majority o f the population. In addition, rapid urbanization and industrial demand are stressing wood fuel supply systems, which, combined with the limited access to electricity and highcost of petroleum products, has recently ledto rapid growth inurban charcoal demand. Retail prices of wood fuel, and particularly charcoal, have increased rapidly, in nominal terms, during the past few years. In real terms, charcoal i s now more expensive than ever. Severe deforestation during the conflict period, in part by internally displaced persons and returning refugees, has been further compounded by the large-scale demand for charcoal. 44 Figure 3.4: Evolution of Charcoal Prices, 1975 -2004 - (In unit as indicated) FRw/kg 120 us%'Yoo - 180 100 inus9 - 80 140 8 I 40I 8 100 60 - 20 40 1975 I980 1985 I990 1993 2000 2004 Source: MINIFRAand Bank staff estimates 3.43 Investmentsin new generation and network capacities duringthe past decade have been very limited, and have been mostly for emergency repairs and rehabilitation. As a result o f the post-conflict reconstruction activity, electricity production and sales have grown at about 7 percent per annum on average since 1997, but peaked in 2003 due to supply constraints, while demand continued to grow (Figure 3.5). As a result, a demandhupply imbalance emerged that threatenedto seriously hinder economic growth and become a deterrent to potential investors. Figure 3.5: Rwanda electricity production, 1990-2004 (in GWh) 160 300 140 250 120 200 100 80 150 60 100 40 50 20 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 199920002001 20022003 20042005 Source:Electrogaz 3.44 The deficit in electricity generation poses a significant constraint to growth, due to its effect on the production costs and productivity of firms (Box 3.2) Total production o f electricity has recently declined compared to its trend over the period 1995 to 2003. Yet demand continues to increase, leading to an increased deficit in electricity generation. Between 2003 and 2004, production fell by 29 percent, coinciding with a doubling o f the tariff rate from FRw 42 to FRw 82/kWh, or US$ 0.16, compared with the cost in Uganda of 73.6 shillings, or US$ 0.04kWh. In 2003, average domestic bulk supply from hydro sources was estimated to cost about USc l/kWh, and power from thermal plants has recently been estimated to cost more than USc 20/kWh, depending on the highly variable price of imported diesel. This exceeds by far Electrogaz' cost 45 for imported electricity from SNEL and Sinelac;' which amounts to nearly USc 8kWh and has inthe past beenconsidered an expensive source compared to domestic production. 3.45 In early 2005, the tariff increased once again to FRw llZ/kwh. With the latest increase in price to FRw 112/kwh, the cost o f electricity in Rwanda is now estimated to be 5 times the cost inUganda. Yet available supply i s increasingly insufficient to meet demand. This affects the competitiveness o f Rwandan goods vis-&vis goods from other COMESA members (Uganda, Kenya, and Tanzania). 2oThe Societe Internationale d'Electricite des Grands Lacs (Sinelac) was set up in 1983 with joint and equal participation o f Rwanda, Burundi, and the Democratic Republic of Congo (DRC), for the purpose of building and operating the Ruzizi I1 power plant. Its sole clients are the three national public electricity utilities, SNEL (DRC), REGIDESO(Burundi), and Electrogaz(Rwanda).The state utility, Electrogaz, suppliesbothelectricity and water. 46 3.46 The increased deficit in electricity production arises from the combination o f strong growth in demand and unexpectedly low water levels in the lakes, which provide the source of hydropower. Factors that have led to increased demand include population growth, and in particular increased rural-urban migration, which has led to a 157.8 percent increase in Kigali's population between 1991 and 2002. Higher economic growth and demand from industry, particularly construction-related, has also contributedto the higher demand for electricity. 3.47 Low water levels inthe lake usedfor hydropower generation have constrained the full generation potential ofbothElectrogaz and Sinelac. The crisis was worsened in2004 and 2005 by strongly increasing demand fueled by economic growth. Moreover, high technical losses andunreliability o f Electrogaz' dilapidated network have exacerbatedthe crisis, and ledto extensive and lengthypower cuts starting inearly 2004. Technical losses (representing22 percent of total energy transmittedthrough the network (Table 3.8) arise from (i)normal technical losses; (ii) outdated network and lack of maintenance; (iii) defective meters; and (iv) illegal connections. As a result, in 2005, load shedding increasedto about 25 percent of peak demand, despite re-commissioning of an old diesel plant. The addition in 2004 of three thermal (diesel) power plants, with a combined capacity of 28 MW, brought only marginal relief. Power supply shortages have also worsened water supply problems inthe main urbanareas, leading Electrogaz to switch to diesel water pumping, thereby further worsening its financial position. 3.48 Rising demand in the face of limited supply caused Electrogaz to increase imports of power from the DCR's Rusizi Iplant (total capacity 39.6 MW), and from Rusizi I1(total capacity 44 MW), owned by Sinelac (Box 3.2). Electrogaz has been able to draw substantially more than its allotted quota of one third from Rusizi I1in the past, because o f low utilization by Burundi. Imports o f power generation in 2002 from Rusizi I1totaled 126 GWh, or 56 percent o f the total electricity available to Electrogaz for dispatch. 3.49 There is, however, limitedscope for further increasesin importsfrom Sinelac in the immediateterm. There has been a general lack o f maintenance and poor rehabilitation both on Rwanda's domestic hydro and Sinelac plants. In addition, Rwanda's off-take from the Sinelac unitshas beencurtailedto its contracted allocation, and amounted to only 65 GWh, or 32 percent of total electricity dispatched in 2005. This has come about due to the decline inthe water levels of Lake Kivu and other lakes in the region. Consequently, medium-voltage consumers have no choice but to rely on their own standby generators for lengthy periods in order to mitigate the effects o f frequent and lengthypower cuts. 47 Table 3.6: Basic indicators for Electricity (Electrogaz) (In unit indicated) 2002 2003 2004 2005 Capacity (Ma3 Installed 44.5 45.5 44.2 44.2 Available(% ofinstalledMW) 96.2% 77.7% 66.2% Energy (MWh) Netproduction (I) 225,740.4 235113.3 203,962.8 203,538.9 Nationalproduction 98,214.1 117,640.7 90,536.4 116,210.7 Hydro production 98,214.1 117,640.7 84,279.4 65,734.2 GIHIRA 6,912.3 6,574.4 5,548.6 5,908.8 GISENYI 5,699.9 4,782.7 4,539.6 4,380.6 NTARUKA 28,911.O 35,170.6 21,162.2 15,350.6 MUKUNGWA 56,690.9 71,113.0 53,029.0 40,094.2 Thermal production 6,257.0 50,476.5 GATSATA 2,663.9 14,071.9 JABANA 3,593.1 25,75 1.5 RENTAL.POWER 1 10653.1 Exports 8,164.9 3,307.6 2,214.2 I,822.7 GISORO(elg) 2,296.6 3,307.6 2,214.2 1,806.6 MURURUI1 GOMA (Elgz) 5,868.3 16.1 Imports 135,691.3 120,780.3 115,705.4 89,150.9 RUSIZI 7,437.0 2,522.0 20,090.0 20,891.8 RUSIZII1 126,736.0 116,063.O 91,422.0 64,564.0 W A L E (UEB) 1,518.3 2,195.3 4,193.4 3,594.3 GOMA (snel) 100.8 Losses and outages (% of netproduction) losses 24.9% 25.1% 28.0% 22.6% Technicallosses (2) 13.3% 10.8% Non-technicallosses (2) 14.7% 11.8% Connection Total 57,679.0 60,886.0 64,093.0 70,187.0 Lowvoltage 57,378.0 60,568.0 63,453 .O 69,486.0 Mediumvoltage 301.0 318.0 640.0 701.0 Prices ElectricityUS$/MWh 88.17 78.12 142.01 201.46 OilproductsUS$/L 1.03 0.98 0.93 0.99 CharcoalUS$/bag 3.78 3.72 5.22 7.2 Source: MININFRA-Revue Conjointe du Secteur des Infrastructures 2006 andELECTROGAZ Note: 1. Net production =Nationalproduction-exports+imports 2. Basedon estimated figures provided by ELECTROGAZ 48 3.50 In the short term, it may not be feasible for the region to continue to supply hydro power. Rwanda has a significant potential for generation o f electricity from hydro sources in numerous steep, fast-flowing rivers and streams. Many mini and micro-hydro plants existed prior to the 1994 conflict, butwere damagedor destroyed and have yet to be brought back into service. Until just two years ago, more than 90 percent of power generation in Rwanda came from domestic hydro sources. Although these covered barely half o f the country's electricity demand, the remainder, imported from neighboring countries, was also hydro-based. There is still considerable untapped hydroelectric potential on the Rusizi and Kagerarivers, which respectively constitute Rwanda's borders with the DRC and Tanzania. But this can only be developed on a bilateral or regional basis, requiringlong preparation and implementation periods, and this will be dependenton political and economic stability inthe region. 3.5 1 The above trends highlightthe urgentneedfor Rwandato identifyalternatesources of power. In the short term, there is no other solution than to revert to additional expensive thermal generation. Slower demand growth, more rapid rehabilitation of hydro plants, or greater than expected rainfall, leading to faster recovery of lake water levels, may lower thermal generation requirements to a degree. But Rwandawill, nevertheless, have to transition away from an all-hydro supply systemwith low average variable costs to a mixedhydro-thermal system with much higher variable costs. 3.52 In the mediumterm, Lake Kivu methane gas reserves could be used to fill a large part of the opening demand-supply gap. It is estimated that 100to 150 million cubic meters of methane gas are generatedannually in the lake from a proven reserve of 29 billion cubic meters. According to studies undertaken so far, part of this reserve could be captured to initially supply a 30 MW generation facility that eventually could be expanded. A small methane extraction unit was installed in 1963 to supply some 8,000 cubic meters o f methane per day to the BRALIWA brewery, and it is still operational. A pilot plant for future power generation facility is currently beingbuilt, and is expectedto become operational in late 2007 or early 2008. The Government is at an advanced stage of discussions with a private consortium on construction and operation of the facility. In addition, methane gas could also be marketed in the country as a direct energy source. 3.53 Geothermal, peat, and papyrus are also potential energy sources, but are much less well studied and therefore offer only longer-term prospects. There are an estimated 150 million (dry) tons of exploitable peat reserves. A small amount is currently extracted for use inUnitedNations refugee camps. Geothermal resources are believedto be sufficient to generate more than 170MW of electricity, but no geological or chemical investigations have yet been undertaken. As for papyrus, 150,000 tons of annual production i s believedto be possible. 3.54 The supply and demand conditions in the energy sector therefore pose a serious constraint to growth and competitiveness, particularly for the manufacturing and service sectors. To address these issues, a comprehensive strategy for the sustainable provision o f energy i s needed. The focus should be on developing alternative sources of energy, particularly for rural areas. It will not be possible to have development inthe rural areas without a viable, sustainable, affordable, and widely accessible source of energy. There should be a focus on ways to engage the private sector to promote sustainability, discussed in the next chapter. In the meantime, investmentsare neededto upgrade the dilapidated network o fElectrogaz. 49 3.55 A detailed assessment of fundamental sectoral and operational issues is neededto guide investments for sustainable improvements to Electrogaz. The assessment would guide the formulation o f strategic investment plans in electricity and urban water supply, including technical and commercial efficiency improvements, cost-recovery schemes, subsidization for low-income groups, appropriate tariff levels and adjustment mechanisms, and financing plans. More fundamentally, the assessment would also produce findings on (i) the merits and drawbacks o f a potential separation of electric and water utility functions; and (ii) the role the private sector could play inoperating, managing, and financing electricity and urban water supply. 3.56 In the case of water supply and sanitation, despite the country's abundant water resources, only about 57 percent of Rwanda's total population has access to safe water, which is low even by African standards.*' A much lower percentage i s connected to piped water sources, mainly in the urban areas and networks operated by Electrogaz. While 80 percent of the population uses latrines, a bare 8 percent is estimated to have access to hygienic sanitation facilities. The sector overall has not only suffered from considerable destruction duringthe period o f the civil war and the genocide, but has also been affected by years o f underinvestment and poor management. Inrural areas, increasing population density, pollution, and the dry weather o f the past few years have reducedthe quantity and quality of safe water traditionally available from natural sources. 3.57 In terms of productivity and growth, the management o f water resources has not contributed to irrigation and agro-processing, including in the manufacturing sectors. Water supply has on occasion limitedthe processing o f coffee, and also constrained investment incoffee washing stations in parts of the country (such as Gitarama) that are relatively drier.22In general, poor water supply arises from the currently high technical losses and unreliability of the dilapidated water supply network. As with electricity, this has contributed to shortages and the highcost ofsupply. 3.58 Policies are in place to guide delivery of water services. Government has undertaken substantial efforts to implement a national water sector policy aiming at rehabilitation and reconstruction o f infrastructure, rebuilding and enlarging the human resource base, and developing a long term vision for the sector. This policy, first issued in 1996, defines the framework for effective and efficient management o f the water cycle in its entirety, incorporating new aspects such as decentralization; community participation; program approaches; and private involvement in operations, ownership, and financing o f the sector. The policy supports the MDGs as well as the Government's overall development strategy, as laid out inthe Vision 2020 and PRSP documents. 3.59 Government's strategy consists of three programs. The first i s the national water resourceprogram, which encompasses components related to institutional, technical, and water resources development. The second i s the ambitious rural water supply and sanitation program, 21 This estimate is basedon Governmentpolicy andplanningdocuments, and is significantly lower thanthe 73 percent estimate of the World Bank's Africa DevelopmentIndicatorsand other sources used for internationalbenchmarking. The lower estimate would change Rwanda's relative standing in safe water access from "outperform" to "underperform" (see ). 22See GovernmentofRwanda(2005), Rwanda's CoffeeIndustry--CurrentSituationandProposalsfor Action. 50 projectedto cost nearlyUS$ 1 billion and designed to increase access to safe water and hygienic sanitation in rural areas to 100 percent by 2020. The third i s the urban drinking water and sanitation program which is primarily focused on Kigali and projected to cost US$ 0.5 billion. The other urban centers are, for the purpose o f the water sector strategy, considered rural areas andtherefore managedas part ofthe ruralwater supply program. 3.60 However, the technical inadequacies o f the distribution system and low capacity (both in terms o f availability and knowledge o f technical staff and resources), pose a major challenge to implementation of the policy, particularly given the context o f decentralization. District authorities provide some funding, but often this i s limited. The situation in rural areas i s worsened by the absence of rural-based service providers. Most o f the private contractors come from Kigali, and this makes it difficult to supervise them locally. There are also no good records and figures on rural water provision. However, Government i s working to develop a database that would facilitate monitoring service delivery in water supply and consumption, including maintenance by private operators. In urban areas, Electrogaz has been grappling with the preservation of its dilapidated networks, with little funding for new investments or rehabilitation. Moreover, the company has not been able to raise water tariffs to cost recovery levels, which would have permittedit to generatethe necessary fundingfrom its own cash flows. 3.61 Investments in water and sanitation infrastructure are considered equally important to enhance private sector development inurban and rural areas. A multi-sector utility regulator was established to support deregulation of and private involvement inutilityfirms and inthe transport sector. Substantial progress toward some o f these policy objectives has been achieved, including deregulation of telecommunications, establishment o f a regulator, and improvement o f international road links. But much more needs to be done to ensure adequate infrastructure provision. 3.62 These infrastructure constraints partly explain why past reforms to liberalize trade have not led to significant or sustained increases in trade and foreign direct investment. The DTIS cited supply-related factors as the major constraint to exports, and the poor state o f infrastructure has been a key factor defining productivity. Past reforms have ledto significant liberalization o f trade in Rwanda. For example, Rwanda has received a score o f 2 in the early 2000's on the InternationalMonetary Fund's trade restrictiveness index (on a scale o f 1to 10, with 1 indicating most liberal and 10 the most restrictive). Tanzania, for example, received a score o f 5, which i s around the average score across all COMESA countries. Yet, in 2002, Rwanda attracted only US$3 million dollars of net FDI-halfthe amount received by the Central African Republic, and 48 and 80 times less than Uganda and Tanzania, respectively. In terms of investment in infrastructure, with the exception of the water sector, Rwanda generally scores below international and regional levels. Table 3.7 summarizes Rwanda's investment performance in infrastructure compared to SSA and other low-income countries. 3.63 The evidence therefore strongly indicates that the poor state o f infrastructure has been a more binding constraint to growth than the cost o f and access to finance. The level o f past neglect has seriously depleted the stock of infrastructure, and current levels o f investment have not been sufficient to reverse the depletion. The result has been relatively higher direct costs to firms and the processof doing business. 51 3.64 Inthe case of water, regional variability in rainfall, coupled with limited irrigation and poor management capacity, have been key factors determining low agriculture productivity. Agricultural and rural activities have also been seriously hampered by insufficient maintenance of the non-classified road network, thus aggravating the transport constraints of the country, given its landlocked position. The current energy crisis has led to power shortages and frequent load shedding, as well as significant electricity tariff increasesto meet rising operating costs. This in turn, coupled with the need for back-up power facilities, has raised production costs up to 16 percent for some f m s , according to recent surveys. 3.65 The Government has taken significant steps to meet growing expectations for infrastructure service delivery. In transport, for instance, the Road Fund operations and management have been reinforced, and the transformation of the traditional ministerial Roads Department to a more independent Road Agency is in the final design stages. This will give further impetus to apriority-basedmanagement ofroadmaintenance. Inelectricity, the monopoly of the parastatal Electrogaz was formally abolished; its accounts were revamped and modernized and now permit revenue and cost allocation between electricity and water; and electricity prices have been adjusted to at least operational cost--recovery levels. Inthe water sector, in addition to improvements at Electrogaz (responsible for urban water supply), the decentralization program i s expected to channel resources more effectively to local governments. Moreover, the Government's intention i s to progressively transition to a program approach that would permit a more coherent implementation of priorities, as defined in the sectoral strategy and the Medium Term Expenditure Framework. Table 3.7: Public PrivateInvestment Infrastructure - Rwanda and International Benchmarks .. % Relative Sub- % Relative Rwanda L1c Deviation Performance tunca Deviation Performance 5-year Annual Average Private Investment ~~~~l (2001-2005) ("A GDP) 0.48% 0.96% -47.9% Underperform 0.81% -42.5% Underperform 5-year Annual Average Private Investment Infrastructure (1996-2000)(% GDP) 0.17% 0.40% -58.6% Underperfom 0.37% -40.2% Underperform Total # of OperationalProjects(2005) 20 7.25 -72.0% Underperfom 5.02 -60.0% Underperform 5-year Annual Average Private Investment (1996-2000)(??GDP) 0.00% 0.23% -100.0% Underperfom 0.09% -100.0% Underperform E~~~~ 5-yearAnnual Average Private Investment f2001-2005>(% GDP) 0.00% 0.18% -100.0% Underperfom 0.03% -100.0% Underperform Total # of OperationalProjects(2005) 1 2.59 -61.4% Underperfom 1.09 -8.3% Underperform 5-yearAnnual Average Private Investment (2001-2005) (% GDP) 0.00% 0.00% 0.00% OnAvg. 0.00% 0.00% OnAvg. WatersuPPb` 5-year Annual Average Private Investment and Sanitation (1996-2000)(% GDP) 0.00% 0.00% 0.00% OnAvg. 0.00% 0.00% OnAvg. Total # o f OperationalProjects(2005) 0 0.19 -100.0% On Avg. 0.26 -100.0% On Avg. 5-year Annual Average PrivateInvestment (2001-2005) (% GDP) 0.48% 0.61% -20.6% Underperfom 0.66% -24.2% Underperform Telecom 5-year Annual Average Private Investment (1996-2000)(% GDP) 0.17% 0.19% -11.0% Underperfom 0.30% -43.0% Underperform Total # o f OperationalProjects(2005) 1 2.81 -64.0% Underperform 2.62 -62.0% Underperform 5-year Annual Average Private Investment (2001-2005)(% GDP) 0.00% 0.05% -100.0% Underperform 0.07% -100.0% Underperform T~~~~~~ 5-yearAnnual Average Private Investment (1996-2000)(% GDP) 0.00% 0.03% -100.0% Underperform 0.04% -100.0% Underperform Total # of OperationalProjects(ZOOS) 0 1.66 -100.0% Underperfom 1.06 -100.0% Underperform Source:World Bank/PPIAFPPIDatabase.AFTPI Processed. Note:Annualizedlows of commitmentsbased on a 10-yearperiodspread, exceptfor telecominvestmentswhich are recordedas flows. Includescommitmentsmadeonandafter 1990.Operationalprojects as of2005 initiatedon or after 1990 * Represents a managementcontractwith ELECTROGAZ(Electricity and Water sanitation) **Includes only RwandaCell, however, inOctober2005 RwandaTelwas privatized PerformanceKey: Outperform benchmark;Average - Deviationbetween - Deviationgreaterthan +lo% from benchmark;Underperform Deviationless than -10% from - -10% and+lo% from benchmark 52 3.66 However, across the transport, energy, and water sectors, the information base needs to be improved to informthe development of comprehensive strategies. Intransport, this is needed to improve planning, and should focus on the non-classified road network in rural areas. Inthe case o f energy, the focus should be on developing alternate energy sources, particularly for rural areas. Inthe water sector, the focus should be on improving water management and irrigation. For all three sectors, capacity support i s needed, particularly at the decentralized levels, to improve maintenance and the efficiency o f spending. The strategic plans should also cover the viability of engaging the private sector, to further expand the current low level o f investments (Table 3.7). The Low Stock ofHuman Capital, SkilledLabor, andProductivity 3.67 The data show that the returns to labor as a factor o f production are generally low, reflecting the abundant supply o f unskilled labor and the scarcity o f skilled labor. In general, educational attainment is low in Rwanda, and firms identify the low availability o f skilled labor as a constraint to production. This is discussed further in chapters 6 and 7. The high returns to skilled labor also indicate its degree o f scarcity. 3.68 In the Rwandan case, the trends indicate that returns to labor in agriculture are at least two to three times lower than innon-agriculture. This reflectsthe fact that low-skilled agriculture labor i s abundant relative to the scarcity o f skilled labor inother sectors. It also points to the low productivity inthe agriculture sector, which poses a constraint to growth (Figure 3.6 and Figure 3.7). Since more than 85 percent o f the population is engaged in the agriculture sector (and the sector exhibits high labor and low capital intensity), a comparison of labor productivity across countries i s used here to examine the level of returns to this factor o f production. Overall, labor productivity in agriculture i s significantly below the trend for the region, highlightingthe level of labor surplus inrural areas. However, it i s important to note that land productivity i s significantly above the trend for the region. Yet, as Chapter 5 will show, even higher yields are potentially feasible through wider use of improved inputs. 3.69 Rwanda could learn from the Benin experience, where between 1998 and 2004, agriculture labor productivity increasedby 40 percent (from US$ 400 to US$ 575). The evidence suggests that increased use of improved inputs played a major role (Figure 3.6). For example, fertilizer consumption in Beninwas only 0.57 kgha in 1980, whereas it was 18.76 kgha in2002, suggesting substantial increase inthe use o f modern inputs and methods. This contrasts with the level of use in Rwanda, which was 0.13 kgha in 1980 and 0.3 kgha in 2001. For Senegal, the figures are 8.2 and 13.6 kg/ha; and for Tanzania, 11.45 and 17.9 kgha. The observed increasesin input use across these countries may have been accompanied by increased rates of urbanization, except for Benin, where the urbanization rate was only about 1.5 percent in 2001. The urbanization rate for Ethiopia is 5.3 percent; for Tanzania, 3.1 percent; and for Rwanda, 4.2 percent. 53 Figure 3.6: Average ProductivityinAgriculture Among G-11 Countries 1980-2003 (Constant 2000 US$) +Burkina Fasc 600 700 c a. Labor productivity b. Land productivity 500000 +Benin +Benin +BurkinaFaso 500000 -Ethiopia 500 /I Ethiopia -Ghana +Ghana 100000 +Kenya +Burundi 300000 -o- Mali 200000 Mozambique - +Mauritania 100000 I - Mozambtque 4Tanzania 1980 1984 1988 1992 1996 2000 Rwanda o -ItUganda I990 I993 1996 1999 2002 Source:World Development Indicators, 2005, and. FA0 Figure 3.7: Average GDP per Worker among the G-11Countries 1980-2005(constant 2000 US$) 600 --tBenin 400 -B- Burkina Faso Ethiopia 200 -W- Ghana 000 -m-Burunrn &Mali 800 +Mauritania Mozambique 600 400 Senegal Tanzania 200 1980 1983 1986 1989 1992 1995 1998 2001 2004 Source: World DevelopmentIndicators2005 3.70 The high rate o f return to education also points to the relative scarcity o f educated or skilled labor (Table 3.8). The returns to education for Rwanda range from about 19 percent for primary education, to 22 percent for secondary education, and 33 percent for higher education.23 Temesgenand Ezemenari (2006) estimate average returns of 6.8 percent for every additional year of education, for workers in the manufacturing sector. The rates o f returnalso vary by gender, as illustrated in (Table 3.8). Inthe case o f returns based on gender, the results show that the returns to education for females are significantly higher than for males. This highlightsthe importance of girls' education. Ezemenari and Wu (2005) also estimate the returns to education using the Household Survey dataset, based on the level completed, and disaggregated by gender, which 23 Lasbilleand Tan(2004). 54 mirror the results of the studies previously cited. Their results also confirm that, on average, men earn more than women, and that this difference declines at higher levels o f education, and i s significantly less for those under 35 years of age. The results suggest that the education of girls would help increasethe supply o f relatively more skilledlabor and reduce income disparities. Table 3.8: Returns to Education by Level (Inpercent) Data Source Primary Secondary Tertiary University Average EICV I2001 19 22 33 25.O Level Completed Male (RIMS 2005) 8 54 61 110 Female (RIMS 2005) 48 58 105 159 Average (RIMS 2005) 18 53 71 119 RIMS 2005(for additional yr. o f education.) 6.8 Male (RIMS 2005 additional year o f education.) 6.5 Female (RIMS 2005 additional year o f education.) 8.5 Source: Lasbille andTan (2004), basedon EICV; Temesgen and Ezemenari (2006), basedon RIMS. Note: RIMS-RwandaIndustrialand Manufacturing Survey 3.71 Completion rates for males and females are on par at the primary level; however at the post-primary levels, disparities are wider in the rural areas. At the national level, the disadvantage in access for girls emerges in higher education. Table 3.9 provides a regional comparison of human capital indicators that give an indication of the quality o f human capital in Rwanda. These results indicate that the low availability o f skilled labor may be an impedimentto growth, and can potentially become a constraint to sustained growth, if not addressed. Analyses inchapters 5 to 7 provide evidence to support this statement. The current situation with skilled labor, among micro-enterprises and manufacturing firms, i s discussedfurther inchapters 6 and 7. Table 3.9: Select Social Indicators-Rwanda Post-Conflict and other countries inthe Region(In unit indicated) Net primary Public expenditure on Public expenditure on Life expectancy Infantmortality enrollmentratio Education ("A ofGDP) Health(YOo fGDP) (Yews) (per 1,000) 1999-00 1998-2000 1995-2000 2004 2004 SSAAverage n.a n.a 6.0 46 101 Rwanda 93 2.8 5.2 44 118 Burundi 54 3.4 3.1 44 114 DR Congo 33 n.a 2.9 44 129 Central African Rep. 55 1.9 1.5 39 115 Ethiopia 47 4.8 4.6 43 110 Eritrea 38 7.4 4.3 54 52 Uganda 84 4.0 3.9 49 80 Kenya 69 6.4 8.3 48 79 Tanzania 58 2.5 5.9 46 78 Source: World Development Indicators, 2006. 55 3.72 Given the Government's aim to become an information technology (IT) and service- oriented economy, investments intraining and education o fthe labor force will be keyto ensuring its transformation from one that i s subsistent based and concentrated in agriculture to one that is market orientedand engaged inhigh-productive and value-added activities. The development o f a skilled labor force requires a healthy and well-educated population. The preceding discussion has highlightedthe importance o f education. However, the high mortality rates, as shown in Table 3.9, indicate that improvements are also needed in health and nutrition. Despite the progress made in immunization, mortality rates remain high and life expectancy remains low. Inaddition, much more needs to be done to address chronic malnutrition, which affects 45 percent of children underfive, as well as other childhood illnesses. Only about halfof children who had a respiratory infection or fever and only a quarter o f those who had diarrhea consulted a medical person, according to the 2005 Demographic and Health Survey (see Government o f Rwanda, 2006b). Improving nutrition is important for reducing extreme poverty, and contributes to increased productivity o f workers and learning outcomes of children. Therefore, while increased focus is placed on improving delivery of infrastructure services, it i s important to maintainthe gains inthe social sectors and improve nutritional outcomes. 3.73 The implementation o f policies to improve labor productivity through relevant training and provision o f social services, as well as the development of channels for sharing and diffusing knowledge and technology in the agriculture secctor, will help to reduce the gap between actual and potential yields, as well as the incidence of poverty. Increasing the availability o f skilled labor will be critical for developing high-value exports and services. Overall, the focus will need to be on strengthening the overall education system to enable learning and capacity buildingfor youth, to improve their employability and trainability. This i s particularly important since low levels o f education undermine the well-being of children and may also contribute to poor governance. GeographicLocation -theNeighborhoodEffect 3.74 Past conflicts and unrest in the region have been a significant source of external shocks, which have affected foreign investment and constrained growth. Violent conflicts have contributed directly and indirectly to the progressive destruction of human, physical, and social capital inRwanda. Inthe rural areas, the bedrock of the economy and home for 80 percent of the population, the livestock herd that accounted for about 5 percent o f GDP was almost wiped out by the genocide. Simultaneously, soil fertility declined because of environmental degradation and the loss of animal manure. With the destruction o f capital and decline in investment over time, productivity and national incomes fell, poverty increased, and government revenues declined. A World Bank poverty assessment completed in200424estimates that GDP growth was 30 percent lower than it would have beenhad the 1994 conflict not taken place. 24Rwanda:PovertyandHumanDevelopment,2004. Poverty ReductionandEconomicManagementUnit, Sub-SaharanAfrica Department. WorldBank. 56 Figure 3.8: GNI Per Capita in Rwanda's Neighborhood,2005 700 600 500 400 3 0 0 200 100 0 LIC Kenya Tanzania Rwanda Ethiopia Burundi Source: World DevelopmentIndicators,2006. 3.75 As noted inChapter 2, the Government has taken steps to addressthe causes and legacies of genocide, by mainstreaming dialogue and creating institutions that promote national reconciliation and build social capital. However, the Government's effort toward reconstruction, reconciliation, and peace i s made more difficult becausethe country i s ina neighborhoodthat has beenprone to conflict. Rwanda is located inthe Great Lakes Region east of Democratic Republic of Congo, with Burundito the south, Uganda to the north, and Tanzania to the east. Rwanda's social indicators are in line with those of neighboring countries (Table 3.9). It ranks low in a neighborhood o f very poor economies, with GNI in 2002 that was estimated to be half the average for low-income countries 3.76 The prospect for peace and social stability among countries in the Great Lakes is instrumental to Rwanda's growth and economic prospects. The private sector and civil societies inthese countries have a crucial role to play in fostering regional cooperation and integration2'. Peace would need to be followed by constructive actions to improve regional cooperation and intergovernmental relations, eliminating obstacles to the free flow o f factors o f production and other goods and services. Some needed actions are already being carried out under multilateral arrangements in the region, and could make a considerable socio-economic contribution toward maximizing the peace dividend for the impoverishedpeople inthe region. Table 3.10: RegionalMacroeconomicIndicators (In unit indicated) Real GDP Inflation Fiscal deficit Current account Net FDI grad<%) 3000 2 KIOOO 1500 u 1000000 2000 z 500000 1000 1000 5000 0 0 500 1997 1999 2001 2003 2005 0 0 I 1990 1993 1996 1999 2002 2005 I Irish Potato Vegetables and Fruits i 160000 I 1I2000 700000 1 ,8000 140000 1 YieI$ 9 10000 600000 ;l20000 -2 500000 100000 < 8000 pm 80000 5000 -. 9 6000 $ 400000 v s $ 5E 4000 Y -3 60000 300000 v Y 4000 3000 5 Y 2 40000 200000 > 20000 2000 2000 100000 0 0 1000 1990 1993 1996 1999 2002 2005 0 0 1990 1993 1996 1999 2002 2005 Source: World BankAgriculture Policy Note, 2006 andFA0 98 5.14 Given the low level of input use, as well as the limited adoption of improved technologies and methods, the increased expansion of land area and yields seems to have come about becauseof an increased shift of labor into agriculture. Duringthe 1994 conflict, there was a massive shift out o f agriculture production, as labor was engaged in non-productive activities and existing assets (such as land and coffee trees) were neglected or destroyed. Once the conflict ended, labor shiftedback into agriculture, leadingto growth inthe sector. The increased shift of labor into agriculture has also had an impact on improvingyields, as shown in Figure 5.3 below. An examination of average yields over time (figure 5.3 and 5.4) indicates that (with the exception of maize), most crops are either at or have surpassed historical yield levels. It would seem that the growth observed in the past decade has been mainly due to the post-conflict catch-up effect. Looking forward, sustained growth will require improved inputs and the adoption of new methods and technologies. Figure5.4: Yield Trends in the Main Export Crops, 1990-2005 (In unit indicated) Coffee Tea 60000 14000 71600 i!50000 = m I2000 Pedare% 10000 6 40000 5 - 500 f 8000 -0 6000 ,z- '2 6 30000 - 400 Y 3 z 20000 - 300f .-3 5 4000 - 200 0 10000 - 100 2000 0 - 0 0 1990 1993 1996 1999 2002 2005 I 1990 1993 1996 I999 2002 2005 Source: World BankAgriculture Policy Note, 2006 and FA0 5.2 OVERVIEW OF CONSTRAINTSTO GROWTH 5.15 An examination of the constraints to growth indicates that low levels of input use pose the most bindingconstraintto growth. Further constraints are the distance from farm gate to rural markets, poorly maintained rural feeder roads, and the absence o f rural transport equipment. 5.16 Domestic food markets in Rwanda are generally underdeveloped, and markets for locally produced food are largely informal in nature. Marketing chains for locally produced food are poorly integrated, with produce changing hands several times as it moves from the farm gate to the final consumer.Farmers sell produceto rural assemblersinthe field, at the farm gate, or along the road, or they sell to rural traders located in local assembly markets. Most rural assemblers operate on a very small scale. Lacking mechanized transport, they may arrange for the use of wheelbarrows or bicycles to transport produce from the farm gate to collection points, from where it can be sold on to rural traders. Ruraltraders typically loadproduce onto small trucks and transport it to urban wholesalers. Urban wholesalers in turn supply urban retailers, who break loads down into small lots for resale inopen-air market stalls or small neighborhoodshops. 99 5.17 Marketing margins (measured as a proportion of the final consumer price for domestically produced food crops) are high in Rwanda compared to other countries in Africa. The high marketing margins reflect more the presence o f high real costs, especially transport costs and than a lack of competition among intermediaries. However, there has been increased competition from greater numbers o ftraders operating at all stages along the marketing chain, but especially inrural assembly and retailing, where barriers to entry are low. This has helped to put downward pressureon marketingmargins.60 5.18 Despite these constraints and the informal nature o fthe marketing chain, the marketing of locally produced crops and livestock products appears to be relatively efficient. High and essentially unpredictable inter-annual price fluctuations discourage investment in long-term storage, but since many food staples are subject to significant storage losses, the intra-annual (seasonal) price fluctuations o f 25-30 percent for most crops seem reasonable. Spatial market integration i s good, as indicated by price correlations across major markets.61 This i s not surprising, considering that Rwanda is a small country with relatively good road links between major urban population centers. 5.19 In general, the degree of integration across regional markets allows for: (i)a total transaction time o f only 1-5 days, without speculative storage; (ii)relatively short distances between rural and urban markets (25-150 km); and (iii)a large number of traders and transporters, who buy and sell each day inrural and urban markets. The efficiency of the system i s also helped by: (i) widespread use of mobile telephones; (ii) availability of vehicles and the the fuel; and (iii)a reasonable network and quality of main roads and feeder roads. However, transport costs from farm gate to the market are relatively highdue to poor road maintenance, and the poor state of non-classified and localized rural roads. This has already been discussed in Chapter 3. 5.20 In the case of improved inputs, utilization and coverage rates continue to be extremely low, and are major factors in the low yield levels. Among the key food staples, coverage rates and production of improved seeds are below 12 percent (see Table 5.1). Table 5.4: Crop yield comparison, 1999-2003 average (Mt/Hu) Africa's World Yield (Mt/Ha) Rwanda Burundi Ethiopia Tanzania Uganda 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 n.a 10.2 13.2 8.9 10.6 Sweet Potatoes 5.8 6.5 9.6 1.9 4.4 4.6 14.9 Potatoes 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: Calculatedfrom FAOSTAT. 2005. 5.21 In the case of fertilizer consumption, the number of households using fertilizer has decreased from 10.5 percent in 1990 to 8 percent in the 2005 season B growing period. 6o Goessens andRwamasirabo (2006). 61Loveridge(1989) 100 Fertilizer use varies by agro-ecological zone, land use pattern, and degree of integration o f the crop into a commodity chain. 5.22 Fertilizer access and use is more developed for crops such as tea, Irish potato (in volcanic zones), and rice, where there is a vertical integration of the supply chain, which facilitates access to: (i)inputs (seeds, fertilizer, pesticides); (ii)extension; and (iii) seasonal credit. Marketing i s also ensured by factories (tea and rice) and cooperatives (potato). However, this is notthe casefor cropsthat do not havethe same level ofmarket integration. 5.23 Low yields can be explained by the use of traditional technologies and a lower level of modern inputs, especially fertilizer. Due to lack o f pasture and fallow land, as well as low incomes, only 18 percent of rural households own cattle, and 36 percent own sheep, two o f the most important sources of manure. The low availability o f organic fertilizer makes chemical fertilizer particularly important for improving land productivity. Yet, estimates from the Household Survey indicate that only 5.3 percent of households that purchased chemical fertilizer used it on at least one plot of land. 5.24 Related to the issue of fertilizer use is the poor state of soil and water conservation in the country. Only 23.4 percent o f the country's soils have little or no erosion risk, whereas 37.5 percent need adjustments before cultivation and 39.1 percent have a high erosion risk.62 Country-wide, water erosion causes a total annual loss in organic material o f 945,200 tons, and also causes annual losses of 41,210 tons o f nitrogen, 280 tons o f phosphorus, and 3,055 tons o f potassium. Late rains and incidents o f localized drought also significantly affect agriculture production and translate into lower overall GDP growth. 5.25 The 2005 Agricultural Survey shows that only 8 percent of households use pesticides. They are used mainly in coffee production (coffee bug) and Irish potato (phytophtora). Fertilizers and pesticides are pre-financed by OCIR-cafe, cooperatives, and farmers' association^^^. 5.26 Financeand credit. In2003, total bank loans inthe agriculture sector amounted to FRw 2.25 billion, representing2 percent of all bank loans. The amount varies by institution. In the case o f the Banque Rwandaise de Development (BRD), 27 percent o f its loan ortfolio in 2003 was in agriculture. For the Union des Banques Populaires du Rwanda (UBPR)6 ,11.3 percent o f 9 its portfolio was in agriculture. This figure i s likely higher for 2005 given the expansion o f UBPR and other micro-finance institutions. Estimates are that inJune 2005, Rwanda had almost 230 financial institutions, of which 170 had been approved by BNR, including 149 from UBPR. These institutions serve more than a million clients (of which 36 percent are with UBPR), with deposits totaling FRw 51 billion (of which 60 percent are at UBPR). Overall, UBPR has about 386,000 memberswith savings worth FRw21.8 billion. 5.27 Although the availability of credit has improved, access is still an issue. The 2001 Household Survey shows that the use of formal credit i s limited. More than 60 percent o f households rely on personal household savings for financing. An additional 10 to 12 percent rely on loans from parents.65As shown inTable 5.5, the highest quintile takes out almost three times 62 Aerssten et. al.(2006). 63 Specifically, the cost o f inputs are deducted from sales made by farmers to OCIR-cafe, or the cooperative. 64 UBPR is a network o f micro finance institutionsthat serve as micro-lenders and micro-banks for nearly two-thirds o f all depositors inthe country. UBPR controls over 97 percent o f the micro-finance sector in Rwanda. . 65 Dabalen et a1 (2004). 101 the amount of loans as the lowest quintile. The data indicate that the collateral requirements of formal institutions are one o f the main reasons for the lack of access by households in the lower income quintiles. Table 5.5: Proportion of Loanstaken by Household Quintiles (Inpercent) Loans with Formal Credit Institutions Loans for Business Expansion All areas Rural Areas All Areas Rural Areas 1st Quintile 10.0 10.1 2.2 2.3 2nd Quintile 13.3 13.3 2.3 2.1 3rd Quintile 15.2 15.4 2.0 2.1 4th Quintile 14.9 14.8 5.6 5.5 5th Quintile 28.1 22.1 9.8 5.8 Total 16.9 15.0 4.7 3.5 Source: Dabalen, Patemostro, andPierre (2004) basedon the 2001EICV. 5.28 Overall, the patterns of input use are defined by the level of institutional organization of farmers, which enables them to gain access. For example, in the case of the coffee sector, which i s completely privatized and characterized by many smallholders, OCIR- Caf6 plays a coordinating and facilitating role in organizing farmer associations to gain access to inputs. For those crops where there is limitedvertical integrationof supply chains, and/or farmers are not organized, inputuse i s consistently lower. 5.29 In the case of the tea sector, farmers gain access to inputs through contract arrangements with plantations. For example, the strong level o f organization and facilitation provided by SOWARTHE translates into significantly higher yields compared to other tea factories66.Aside from the improved institutional organization o f farmers, continued expansion and deepening of the micro-finance sector will be important for access to finance by rural households and farmers. 5.30 Low purchasing power is also a constraint to growth and highlights the need to develop the non-farm sector. Given the average low farm holding o f 0.7 hectares, off-farm income i s an important means of supplementing household income, particularly for less well-to- do households. However, for poorer households, a large proportion o f non-farm income comes from remittances, while these households continue to be engaged mainly in farm employment. Thus, there is a sharp distinction between the poor and non-poor in the rural areas, with poorer households engaged primarily in the agriculture sector, and better-off households more engaged inthe non-farm sector. Supporting the transition ofthe poor away from agriculture to other areas o f the economy will require removing the constraints to non-farm enterprise development, as discussedinChapter 3 and Chapter 6. Table 5.6: Sources of Household Income lStquintile zndquintile 3rdquintile 4'hquintile sthquintile Average Share out of Total Income: FarmIncome 1.31 0.78 0.74 0.81 0.33 0.794 Farm Employment 0.14 0.17 0.059 0.042 0.029 0.088 Non-Farm Employment 0.035 0.024 0.065 0.095 0.353 0.1144 Non-Farm Enterprise 0.038 0.045 0.071 0.075 0.209 0.876 Remittances -0.37 -0.067 0.018 -0.063 0.023 -0.0198 Other Income 0.028 0.044 0.039 0.036 0.051 0.0396 Source: EICV 2001. 66SORWATHE, SociCtC Rwandaise de ThB,or the Rwanda Tea Company is a privately owned tea plantation. 102 5.3 GROWTH PROSPECTSFORKEY COMMODITIES 5.31 To attain targeted growth levels, or at least technically feasible levels, given the quantity and quality of soil conditions, it will be necessary to ease constraints related to improved productivity and value addition, since the scope for increasing the scale of production through area expansion is limited. Land expansion would therefore imply reallocationfrom less to more lucrative crops (or from other uses). Aerssten et a1(2006) provide a detailed assessment o f commodity-specific growth prospects. 5.32 Studies indicatethat there is still a huge potentialto increaseyield based on proper and adequate application of inputs. For example, Table 5.7 reports experimental results of fertilizer use drawn from Kelly and Murekezi (2000), who synthesized findings from MINAGRI studies on fertilizer profitability. It shows that used correctly, fertilizer i s highly profitable, since yields can be more than doubledfor most crops. Table 5.7: Fertilizer response for selected crops (In unit as indicated) Increase Increase as percent National averageyield inyield(mtlha) o f national yield (YO) (1998-01) (mtlha) Maize 0.7-2.6 89 330.4 - 0.79 Rice 1.0-1.7 57.2 97.3 1.75 Sorghum 0.6-1.9 62.2 196.8 0.97 Potatoes 5.0-9.3 60.8 113.1 8.22 Sweet potatoes 4.2-7.4 73.5 129.5 5.72 Beans 0.4-1.2 92.6 6.48 Soybeans 0.4-0.8 77.1 ------ 277.7 154.1 0.52 Vegetables (cabbage) 28.3 506.1 5.59 Source: Calculatedfrom Kelly and Murekezi (2000). 5.33 A more recent study by Byakweli (2004) shows that yield gaps still exists for key crops, but also provides a more recent estimation ofthe yield potentialfor key crops. Table 5.8: Yield Response to Fertilizer usage and Improved Seeds, Selected Crops (In Tons/Ha) PotentialYield Average Yield IrishPotato 25-20 8.6 Rice 7.0 3.7 Tea 3.5 1.3 Coffee 2.4 1.6 Beans (dwarf) 1.5 0.8 Climbing Beans 3-4 0.8 Maize 3.5-6 0.8 Banana 25-35 5-10 Source: Byakweli, 2004, as cited in Aerssten, 2006. 5.34 In addition to increased productivity, growth can also be achieved through improved value addition for key crops. 103 5.3.1 Growth of Food Crops 5.35 Potential for value addition also exists for food crops. Following Aerssten el a1 (2006), food crops can be divided into traditional food crops, export-oriented and import-reducing food crops, and new high-value crops. a. Traditional Subsistence Crops 5.36 Among the key traditional crops are bananas, cassava, sweet potatoes, and beans. The demand for cooking bananas is high in Rwanda, but still depends on imports from the Democratic Republic of Congo and Uganda. There has also beena recent deterioration inplants due to diseases andpests, which presents an opportunity to introduce improvedvarieties. 5.37 The relatively higher production of brewing bananas arises from the expansion of cultivation onto increasingly marginal soils in the central part of the country. They are cultivated in subsistence systems with low inputs, resulting in low outputs. The decline in soil fertility i s one of the reasons for increased reliance on brewing bananas, although there are plans to increasingly substitutecooking for brewingbananas. 5.38 Value addition could be promoted through increased training of producers and traders in improved banana handling, and processors in new technologies. Targets have been set to train 10,000 producers and 100 traders in improved handling, and 100 processors in newtechnologies. Improved handling and market information will be neededto improve outlook inthe sector. 5.39 In the case of cassava, average yields are far below standards (potential of 40 tons/ha, compared to an average of 6.7 tons/ha) and have been on a declining trend since 2003, mainly due to cultivation on marginal soils and resistant strains of the mosaic virus. Production projections are based on a 3 percent average annual increase in the cultivated area with a constant annual yield of 8,500 k o a . However, the challenges o f the mosaic virus, poor spacing and weeding practices, rapid post-harvest deterioration, and hightransport cost to market will need to be addressed. There is currently limitedpotential for value addition due to the rapid post-harvest deterioration of tubers, as well as irregular supply, home consumption, and lack of adequate storage facilities. 5.40 Sweet potato production has been decreasing since 2003 (a 31.5 percent decrease between 2002 and 2005, from 195,727 ha to 148,525 ha), due to the promotion of rice and maize in some marshlandswhere sweet potato has traditionally been cultivated.Yields have beenrelatively constant but low, approximately 6 tonsha, whereas under researchconditions, 20- 30 tonsha can be obtained. However, and despite the technical possibility o f significantly increasing productivity by 10 tonsha, it is likely that the cultivated area will continue to decrease (to 100,000 ha by 2010, or an average decrease of 7 percent per year), due to replacement by rice and vegetables indeveloped marshlands. 5.41 Beans, like sweet potatoes, are also major food staple. Beans provide 38 percent o f proteins in Rwanda, and average consumption per capita i s 40 kg. However, production has decreased since 2003 (from 358,002 ha in 2002 to 313,014 ha in 2005), and yields remain excessively low (approximately 650 k o a , compared to potential yields of 3-4 tonsha for climbing beans). The main constraints have been over-exploitation of the soil (due to low rotation); attacks by pests (bean fly and aphids) and disease; minimal use of fertilizers and pesticides; and very limited use of organic manure on other crops included in the rotation o f 104 beans. It i s not likely that an increase in production can be realized through an increase in cultivated area, due to pests and the subsistence character o f the crop. Despite the technical possibility of increasing productivity through intensification practices, it i s likely that, as with sweet potato, the cultivated areas will continue to decrease due to a steady expansion o f climbing beans with a targeted yield of 1tonsha. b. Export-Oriented andImport-Reducing Food Crops 5.42 Inthe caseofIrishpotatoes,only31percentofthe cultivatedarea(42,830 ha) could be suppliedwith improvedseeds, in 2005, thereby seriously limitingproductivity.Inthe case o f value addition, there is potential to develop a processing plant for potato chips and French fries, at an estimated cost o f US$90,000.67There i s a growing urban population and emerging middle class that could form the market for this product. This could result in a potential increase incultivatedareaof 1.5 percent between2006 and 2010, and a 10percent increaseinproductivity inthe same period. 5.43 In the case of rice, it can be assumed that the current scheduled pace of marshlands development, combined with annual yield increases of 5 percent (or from 4.4 tonsha to about 6 tonsha), will result by 2020 in a total cultivated area o f about 36,000 ha (compared to 15,500 ha in2006), and atotal production of216,000 tons ofpaddy (compared to 74,400 tons in2006), thus satisfyingthe national needs, which are now estimated at 140,000 tons o f processedrice. 5.44 In 2005, maize was cultivated on 109,400 ha, resulting in a total production of 97,251 tons (an increaseo f 55.6 percent compared to 2000), with an average yield o f 889 kgha, whereas potential yields under intensified conditions range from 3.5 to 6 Tha. National production projections upto 2010 assume that areas under cultivation increaseto 207,388 ha and resultingin a total production o f 333,895 T, with an average yield of 1.6 Tha. However, the assumptions include the probable replacement o f banana, sorghum, and sweet potato by maize in some zones o f the country. Moreover, the majority o f recent new maize development has been done on marshlands, thus also limitingthe potential for growth through land expansion. 5.45 There is also potential for increased value addition in maize production. However, the two maize milling plants are operating below capacity, while demand (for both human and animal feed) i s likely to continue growing. Additional factors include the challenge to move to a uniform variety of maize to yield uniform quality in flour production. Post-harvest management i s largely traditional and labor intensive, and fumigation is poor and affects the quality o f production. Nevertheless, the high farmgate prices in Rwanda (at US$ 55.6/ton compared to Uganda at US$15.7/ton), could be an additional source o f incentiveto farmers. 5.46 In the case of wheat, yields remain at 30 percent of the potential levels, although land area has more than doubled between 2000 and 2005 (from 10,043 ha to 24,157 ha), resulting in considerable increases inproduction (from 6,444 tons to 21,942 tons) and yields (from 642 kgha to 908 kgha). The main constraints include: (i)high level of soil degradation and erosion; (ii) highsoil acidity inthe Congo NileDivide region; (iii) use ofimproved seeds and inputs;and poor (iv) inappropriate farming techniques. 5.47 In the case of value addition for wheat production, millers and bakers claim that delivered wheat is of poor quality; seeds used by farmers originate from different sources 67 Aerssten et a1(2006). 105 which makes it difficult to adjust the milling equipment. In addition, traditional harvesting techniques contribute to the reduction in the quality and value o f wheat. Finally, the protein content of local wheat, at 11percent, is below the minimumrequirement o f 12 percent. 5.48 In the case of soybeans, although there is considerable technical potential for its development as a cash crop, mainly through marshland development, there i s a need to examine the comparative advantage of cultivating soybeans in marshland, given the potential of other suitable crops to grow in marshland, such as maize. Soybeans have the potential to contribute significantly to improved nutrition due to its high content o f proteins and lipids. Also, there i s a dynamic growers' association inthe vicinity of existingprocessing units. Therefore, the increase inproduction should contributeto increasedjob creation. However, due to lack of interest on the part of the private sector, financing may be a constraint to developing the 15,000 hectares of marshland. Inaddition, there are insufficient quantities o f improved quality seeds. 5.49 In terms of value addition, the existing processing units are not able to process all that is produced, yet there are projected increases in production and demand. Therefore, Government aims to construct five additional processing units inthe main productionzones. The proposal is for the private sector to contribute 80 percent o f financing, and Government to facilitate the process by contributing 20 percent. This will lead to increased employment in the area. However, the existingprocessing units are managed by NGOs and their penetration o f the market i s limited. There are also no post-harvest storage facilities other than in Kigali and Gitarama. Giventhat soybeans are mainly cultivated on marginal soils previously cultivated with beans, the potential for yield increases may be lower than projected by the Government, at around 50 percentby the year 2010, with an area expansion of 5 percent per annum. c. New High-ValueCrops 5.50 Inthe caseofhigh-value food crops, horticultural productshave a future growth potential. However, export growth has been erratic inthis area. Table 5.9: Rwandan exportsof fruits, vegetables and flowers, 2000-2003 (Export Weight, Kg) Crop 2000 2001 2002 2003 Dessert Banana 64,585 36,045 43,248 41,615 Passion Fruit 9,263 12,342 9,024 13,885 Avocado 3,192 1,008 1,576 1,255 Pineapple ... 510 1,377 1,270 Roses 109,123 66,945 110,000 ... Draceana Plants 85,708 151,828 4,791 40,940 Eggplants 605 430 150 French Beans ... ... 57 ... Snap Peas ... 415 566 Cooking Banana ... 7,920 7,740 370 Source: MINAGRI 5.51 There is some potentialfor development of the passion fruit sector. The cultivated area for passion fruit i s estimated at 1,168 ha, providing a total annual production o f 11,211 tons with an averageyield of 10to 15 tonsha, whereas commercial irrigatedplantations are producing 20-35 tonsha. The current fruit supply i s largely insufficient for the five existing processing plants, mainly due to an outbreak o f virus diseases. 106 5.52 Macadamia nut has been identifiedas a potentialcrop for development. It takes two years to produce quality seedlings, and following planting, the macadamia tree starts producing after 5 years, with production increasing from roughly 3 kg/tree to 20-100 kg/tree after 15 years. The tree has no major pests anddiseases and is not highnutrient dependent; thus it can be mainly grown by use o f farmyard manure and does not require inorganic fertilizer. The root system i s lateral, andtherefore improves soil texture and prevents soil erosion. 5.53 In addition to macadamia nuts, other high-value products for which Rwanda has suitable agro-climatic conditions, and which are suitable for smallholders, include: (i) essentialoils; (ii)medicinal crops; (iii) vera; (iv) sunflower for oil; and (v) silk worms. aloe 5.3.2 Growth of Export Crops 5.54 Within the export sector, coffee accounts for 80 percent of growth in agriculture (or 1.08 percentof overall GDP growth). Ingeneral, export growth has been adversely affected by world prices and low production, which has left little for export. Background papers 1 and 2 of the Agricultural Policy Note6* suggest that in addition to the trend in world prices and low productivity, quality and value addition have also beenfactors constrainingexport growth. Table 5.10: Traditional agricultural exports (In million USD) 1992 1995 2000 2001 2002 2003 2004 Total exports 68.76 50.44 69.04 93.55 67.36 64.05 89.51 Agricultural exports 60.27 45.26 47.23 44.62 40.36 42.78 55.26 Coffee 35.12 38.75 22.52 19.36 14.65 15.21 29.45 Tea 20.76 3.84 24.28 22.71 22.02 22.50 22.11 Hides and skins 2.29 2.22 0.43 0.78 2.64 3.79 2.41 Pyrethrum 2.10 0.45 0.00 1.77 1.05 1.28 1.29 Source: MINICOM, Aertssen et al. 2006 5.55 Coffee production has been low due to outdated trees and low relative prices which have led farmers to switch to more lucrativeactivities. For example, the production o f green coffee oscillated between 20,000 and 40,000 tons during the period 1985-1992, but dropped to the range of 15,000-20,000 tons in the late 1990s because of low world market prices. The benchmark New York C-price for coffee fell below the Rwandan cost o f production, which resulted in a sharp decline in both production and quality. In addition, the low level o f input use (which i s currently at 1.8 percent of total needs), coupled with the cumbersome and slow procedures for import tendering by STAI3EX6' (which often cause delays to annual pest control campaigns), have ledto lower yields than technically feasible. Finally, farmers have had limited access to credit to enable them to purchase seeds. World Bank2006a. 69 STABEX stands for the old EuropeanUnionDevelopmentFund. 107 Table 5.11:Rwandan coffee production, 2000-2005 (In unit indicated) 2000 2001 2002 2003 2004 2005 Total coffeeproduction(tons) 16,098 18,267 19,426 14,175 29,000 18,000 Productionhlly washed (tons) 17 51 29 333 700 1,100 Value ofexports(million US$) 22,4 19,4 19,2 15 32 38 YOstandard 19,5 18,5 29,4 32,4 40,O 45 YOordinary 72,4 75,O 58,O 55,O 50,O 49 YOhlly washed 0,1 0,3 0 2 2 3 19 YOexport contribution 32,5 26,6 29,l 23,s 36,O --6 Source: OCIR-Cafe, 2005. 5.56 These factors explain the observed pattern in production, which currently indicates that only 65 percent of the 85 million coffee trees growing in the most favorable agro-ecological zones are producing beans. In total, roughly 33,000 ha o f coffee are cultivated by 500,000 farmers, with total production o f green coffee o f 18,000 tons in2005. 5.57 This productionlevel representsa 37 percent decline from 2004 levels,due mainlyto unfavorableclimaticconditions.Processing o f coffee beanswas also below installed capacity in 2005. Washing stations operated at 30 percent below their capacity in 2005 (and 15 percent below capacity in 2004). Inaddition, to the factors listed above, insufficient basic infrastructure inthe form ofwater supply, and access and feeder roads, have also posedconstraints. 5.58 Inthe case of tea, it is cultivatedon about 12,871 ha, with 66 percent of the area in industrial estates belonging to OCIR-ThB and some private investors, and 44 percent cultivated by small out-growers. Dry tea production has steadily increased between 2000 (14,391 tons) and 2005 (16,899 tons). Although yields o f green leaves remain low on small out- growers' estates, they are 2 to 2.5 times higher on private estates (which include the tea associations and SOWARTHE) than on government-operated estates. For example, recent estimates are that yields average 7 tonsha for OCIR-ThC, against 15 tonsha for SOWARTHE. Moreover, given the total cultivated area for tea, factories are producing 63 percent of what could potentially be produced on average, across all tea estates." Table 5.12: Performance of the tea sector, 2000-2005 (In unit indicated) 2000 2001 2002 2003 2004 2005 Production Black Tea (tons) 14,481 17,817 14,948 15,483 13,998 16,899 Average made tea price (US$kg) 1.77 1.5 1.56 1.59 1.69 __ Source: OCIR-ThC 5.59 The lower yields are mainly due to poor technical management, high energy costs for processing, and high transport costs for export. These factors also hamper the improvement in value added for the sector. In particular, insufficient manufacturing capacity 70 Aerssten et a1(2006). 108 (due to outdated and poorly maintained machines) and frequent power outages are major constraints to improvedvalue added. 5.60 Pyrethrum is a major cash crop in the highlandsofRuhengeri and Gisenyi, where it competes with Irish potato. Production of pyrethrumcame to a halt inthe mid-l990s, but has since increasedto 26,000 tons (from its level of 6,000 tons in 1990). In2004, the 26,000 farmers who grow pyrethrum produced 550,070 kg of dried flowers, which produced a 75 percent increase in their revenues. Total output is bought and processed in Ruhengeri by the company SOPYRWA and exported as crude extract, for which there are very few clients inthe world, and no domestic or regional market. However, the refined product has more direct applications, even inthe region. 5.61 Currently, the factory capacity i s 3,000 tons of dried flowers per year (based on 1kg o f crude extract from 25 kg o f dried flowers, it takes 5 kg of fresh flowers to produce 1kg o f dried flowers). However, frequent power outages have caused the factory to operate at only one-third capacity. 5.62 The frequent power outages and the lack of new and improved varieties (which is not sufficiently addressed by research) are constraints to production, as have been price incentives. The price of dried flowers was 500 FRw per kg in 2005. However, farmers will switch to Irishpotatoes ifprice conditions become unfavorable. 5.63 Hides and skins are Rwanda's fourth largest agricultural export product. In 2004, Rwanda exported 1,890 tons o f leather, of which 93 percent was dry leather and 7 percent was wet blue (which has a higher value). These were exported to Hong Kong (41 percent), Pakistan (3 1 percent), Kenya (11 percent) and Italy (6 percent). The hides and skins come mainly from domestic ruminants, i.e., cattle, sheep, and goats, the populations o f which are estimated at 1.1 million, 690,000, and 2.6 million, respectively. Only a small part o f these hides and skins are collected by tanneries and processed, due to a poorly organized collection system. Production amounted to 2,620 tons in 2005. Moreover, poor post-slaughter treatment often degrades the hides and skins to a point where they receive a significant discount on international markets. There i s a strong base from which the industry can develop and increase investment. However, at present, the industry is competing in fairly low-quality niches. It i s also hampered by insufficient numbersof skilledtechnicians, leadingto poor quality, and by low processing capacity. 5.64 Other by-products from livestock production include milk, meat, and eggs, production of which is estimated at 97,981 liters, 39,126 tons, and 2,432 tons, respectively. Fishproduction is at 7,612 tons. These levels are not sufficient to cover domestic consumption, estimated to be 12 liters of milk and 4.8 kg o f meat per person per year, which is lower than the nutritional recommendations o f 220 liters and 50 kg, respectively. Constraints to production are related to quantity and quality o f animal feed and animal nutrition. In addition, loss due to disease poses a problem and is compounded by weak veterinary services, legislation, and regulation. Related to this i s an inadequate link between research and extension. Finally, poorly organized farmers' associations constrain access to inputs, dissemination o f information, and marketing. 109 Table 5.13: Number of Animals, Hides and skins, 2000-2005 (In tons) 2000 2001 2002 2003 2004 2005 %Growth Cattle 755,123 814,124 960,450 991,697 1,006,572 1,079,206 42.9 Goats 756,502 916,753 919,785 1,270,903 1,263,962 2,663,551 252.1 Sheep 232,724 266,539 300,600 371,766 686,837 689,556 196.3 Pigs 177,220 197,081 207,783 211,918 326,652 456,041 157.3 Chicken 1,277,706 1,055,644 2,432,449 2,482,124 2,841,399 2,011,572 57.4 Rabbits 338,616 495,290 488,629 498,401 643,927 90.2 Hides/Skins 816 1,180 503 1,885 3,559 336.2 Source:MINAGRI, Statistics series, 2006. 5.65 This analysis indicates that some immediate growth response can be generated by improvingthe productivecapacity of the key processing plantsfor the export market. This is particularly the case for coffee, tea, and pyrethrum, where the impact of poorly maintained machines coupled with unreliable electricity supply, has significantly affected production. This in turn has affected the demand for farm output usedby processing plants, which inturn affects farm-level productivity. Investment in updated production methods and machinery will be particularly important for the development of the high-value niche market for the key export crops. For the tea sector, privatization was expected to help facilitate this process. However, there has been very slow progress. In the case of pyrethrum, increased demand for environmentally safe insecticides on the international market provides good prospects for SOPYRWA (the processing plant), to reach its target o f increasing the cultivated area to 7,000 ha (from 3,500 ha) within the next year. In the case o f hides and skins, farmer organization, improved extension, andresearchare critical to improving quality. 5.3.3 Summary of Growth Prospects 5.66 This section has outlined the sources of growth within the agricultural sector, both at the sector level (crop, livestock, exports) and by approach (land expansion, increased productivity, value added). To begin the process of prioritization among growth sources and approach, the following section will examine which sector contributes most to growth and poverty reduction. 5.4 AGRICULTURE'S POTENTIALCONTRIBUTIONTO GROWTH AND POVERTY REDUCTION OVERVIEW OF THE MULTI-MARKET - MODEL 5.67 An economy-wide multi-market (EMM) model of the Rwandan agriculture sector developed by Diao and Yu (2006), is used to explore sub-sectoral contributions to GDP growth. The model captures heterogeneity in 30 agricultural commodities and two aggregated non-agricultural sectors. The 30 commodities are from eight broad sub-sectors: (i)grains, (ii) bananas, (iii)roots and tubers, (iv) pulses and oilseeds, (v) export crops, (vi) other cash crops, (vii) livestock and products, and (viii) other food and non-food agricultural commodities. The EMM model also considers the level of regional heterogeneity such that production and consumption of the 30 agricultural commodities (and the two aggregate nonagricultural sectors) are further disaggregated into 11provinces plus Ville de Kigali City. 5.68 The model captures the distribution of land holdings, as defined by three categories of rural households. Rural group 1representsrural households holding land of less than 0.3 hectare, 110 including rural landless households; rural group 2 represents households with landholding between 0.3 and 1.O hectare; and rural group 3 represents households with land of more than one hectare. Table 5.14 summaries the distribution of rural households according to landholding categories and region, based on the 2001 Household Survey. About 40.6 percent of rural households belong to rural group 1, with 32.8 percent in rural group 2, and 26.6 percent in rural group 3. Table 5.14: Household Aggregationby Landholdingand Region KuralOroup 1 Kuraltiroup 2 KuralGroup3 perhhlandholding<0.3 ha 0.3ha<= landholding 1.0ha Urban Rural Female- Male- Female- Male- Female- Male- Female- Male- Total Total Number of household Butare 33,052 61,219 10,093 17,046 3,731 8,810 3,574 3,546 133,951 141,071 Byumba 13,594 38,308 11,693 42,779 9,353 34,298 619 1,830 150,025 152,474 QangUgu 19,195 44,223 8,307 15,782 4,076 18,077 796 2,055 109,660 112,511 Gikongoro 21,222 44,610 2,263 12,437 3,693 15,483 683 1,500 99,708 101,891 Gisenyi 18,413 37,028 17,322 31,673 12,344 29,424 837 5,348 146,204 152,389 Gitarama 27,420 30,068 20,230 40,477 16,745 30,777 1,504 3,764 165,717 170,985 Kibungo 4,959 6,939 16,848 39,099 15,737 45,209 2,282 2,250 128,791 133,323 Kibuye 16,721 27,048 10,002 20,907 6,122 12,577 292 2,640 93,377 96,309 Kigali 17,889 25,656 26,875 45,185 24,870 40,825 1,539 4,257 181,300 187,096 Ruhengeri 30,280 63,452 20,132 40,050 7,773 19,342 1,746 4,414 181,029 187,189 Umutara 3,730 4,533 9,874 16,302 7,604 19,243 305 610 61,286 62.201 Memorandum items: City of Kigali 30,362 82,349 0 112,711 National 206,475 383,084 153,639 321,737 112,048 274,065 14,177 32,214 1,451,048 1,497,439 % of nationalrural total Butare 2.3 4.2 0.7 1.2 0.3 0.6 9.2 Byumba 0.9 2.6 0.8 2.9 0.6 2.4 10.3 Cyangugu 1.3 3.0 0.6 1.1 0.3 1.2 7.6 Gikongoro 1.5 3.1 0.2 0.9 0.3 1.1 6.9 Gisenyi 1.3 2.6 1.2 2.2 0.9 2.0 10.1 Gitarama 1.9 2.1 1.4 2.8 1.2 2.1 11.4 Kibungo 0.3 0.5 1.2 2.7 1.1 3.1 8.9 Kibuye 1.2 1.9 0.7 1.4 0.4 0.9 6.4 Kigali 1.2 1.8 1.9 3.1 1.7 2.8 12.5 Ruhengeri 2.1 4.4 1.4 2.8 0.5 1.3 12.5 Umutara 0.3 0.3 0.7 1.1 0.5 1.3 4.2 Memorandum items: National 14.2 26.4 10.6 22.2 7.7 18.9 100 % of nationaltotalhh Butare 2.2 4.1 0.7 1.1 0.2 0.6 0.2 0.2 9.4 Byumba 0.9 2.6 0.8 2.9 0.6 2.3 0.0 0.1 10.2 Cyangugu 1.3 3.0 0.6 1.1 0.3 1.2 0.1 0.1 7.5 Gikongoro 1.4 3.0 0.2 0.8 0.2 1.0 0.0 0.1 6.8 Gisenyi 1.2 2.5 1.2 2.1 0.8 2.0 0.1 0.4 10.2 Gitarama 1.8 2.0 1.4 2.7 1.1 2.1 0.1 0.3 11.4 Kibungo 0.3 0.5 1.1 2.6 1.1 3.0 0.2 0.2 8.9 Kibuye 1.1 1.8 0.7 1.4 0.4 0.8 0.0 0.2 6.4 Kigali 1.2 I.7 1.8 3.0 1.7 2.7 0.1 0.3 12.5 Ruhengeri 2.0 4.2 1.3 2.7 0.5 1.3 0.1 0.3 12.5 Umutara 0.2 0.3 0.7 1.1 0.5 1.3 0.0 0.0 4.2 Memorandum items: City ofKigali 2.0 5.5 7.5 National 13.8 25.6 10.3 21.5 7.5 18.3 0.9 2.2 100.0 Source: Diao and Yu (2006), basedon the 2001 HouseholdSurvey 111 As shown in tables 5.15 and 5.16, more than 17 percent of grains and 13 percent of bananas, root crops, pulses, and oilseeds are produced by rural group 1,which holds less than 6 percent of the total farmland in the country. Eventhough these households allocate more land to staple crop production, food availability at the per capita level i s still extremely low among these households. Table 5.15 shows that per capita grain production i s 75 kg for rural group 3, which is almost four times the amount produced per capita for rural group 1 (19 kg). A similar situation is also evident for other staple crops, and on average, rural group 1's per capita food production is 50 to 70 percent below the national rural average level, and 75 to 85 percent below the per capita level for rural group 3. Table 5.15: Crop distributionby householdgroups (In unit indicated) Rural 1 Rural2 Rural 3 Ruraltotal Totalproduction (I000 mt) Grains 50.0 81.0 160.0 291.0 Roots and tubers 483.0 887.0 1716.0 3086.0 Bananas 236.0 712.0 1377.0 2324.0 Pulses and oilseeds 38.0 80.0 165.0 283.0 Share in rural total PA) Grains 17.1 27.9 55.0 100.0 Roots and tubers 15.7 28.8 55.6 100.0 Bananas 13.6 28.2 58.3 100.0 Pulses and oilseeds 13.6 28.2 58.3 100.0 Average per capitaproduction fig) Grains 19.0 35.0 75.0 41.0 Roots and tubers 183.0 384.0 804.0 455.0 Bananas 89.0 308.0 645.0 308.0 Pulses and oilseeds 15.0 34.0 77.0 40.0 Source: Diao and Yu (2006) based on the EICV Table 5.16: Distribution of householdsgrowing coffee and tea by householdgroups (In unit as indicated) Rural 1 Rural 2 Rural 3 Rural total Number ofrural households Coffee 53,171.0 45,701.0 48,766.0 147,638.0 Tea 2,789.0 4,790.0 5,415.0 12,993.0 Share ofhouseholds in rural total (5%) 4.5 4.9 5.6 5.3 Coffee 36.0 31.0 33.0 100.0 Tea 21.5 36.9 41.7 100.0 Source: Diao and Yu (2006), based on the EICV. 5.69 The model is also disaggregated according to gender of the household head. As a result, both production and consumption are further disaggregated into four sub-groups - rural female-headed, rural male-headed, urban female-headed, and urban male-headed households. Thus, for each province, there are eight household groups, two in urban and six inthe rural (with the three rural categories). All household groups are assignedto commodity-specific supply and 112 demand functions. Table 5.14 summarizes the distribution o f households by province, rural/urban, landholding size, and gender. In total, there are 3,072 (32 sectors by 12 provinces or regions by 8 household groups) production and consumption functions for each individual commodity in the model. Specifically, all sampled households in the Household Survey are linkedto their corresponding representative household group inthe EMMmodel. This allows the model to capture the distributional effects of exogenous changes.71 5.70 The base assumptions used in the modeldraw on 2003 data for land area and yields of individual crops. The model simulation is based on national growth projections for 17 agricultural sub-sectors or commodities based on the assessment o f growth prospects completed inAerssten et al. (2006), and outlined inthe previous sections of this chapter. Each scenario is constructed by assuming a base-run growth rate for a specific sub-sector between 2005 and 2015, while productivity growth in the other sub-sectors is held constant. After 2010, the base-run growth rate i s also appliedto the targeted sectors. However, inthe case of maize, rice, wheat, and soybeans (where the base-run growth rates are lower than the average growth rates for 2000- 2005), growth rates over the period 2010-2015 are assumed to be slightly higher. Table 5.17 reports the growth rates assumedby the model for each crop. 5.71 Ifall cropsjointly grew at the levelsassumedinTable 5.17 until2015 (scenario 22 in Diao and Yu), agricultural GDP would grow at 6.09 percent between2005 and 2015, which i s almost twice the level of growth in the base run. At that level o f agricultural growth, total GDPgrows at 5.04 percent, compared to 3.88 percent inthe baserun. Table 5.17: Level of initialyield and area, and annual growth rates used in the model at national level (In unit as indicated) Base-run Growth simulations (scenarios 1-19) Required annualgrowthrate Assumedgrowthrate Yield Area Targetedlevel by 2015 in2005-2010 (%) in 2005-2010 (%) Level in Assumed Level in Assumed 2003 (mtha) base-run 2003 base-tun Yield Area Yield Area Production Yield Area Roduction growthrate (%) (1000ha) growthrate ("A) (mtha) (1000ha) (mtiha) (1000ha) (mtha) (1000ha) maize 0.8 2.3 104.0 1.4 1.3 160.0 4.6 3.3 8.1 3.7 5.4 9.3 rice 3.7 2.9 8.0 1.8 6.3 19.0 3.7 9.2 13.2 3.5 9.7 13.5 sorghum 0.9 2.0 184.0 0.5 1.0 193.0 2.0 0.5 2.5 2.0 0.5 2.5 wheat 0.7 3.2 21.0 0.8 1.6 23.0 7.1 0.8 7.9 7.0 0.9 8.0 cassava 7.5 1.9 137.0 0.0 9.5 153.0 2.0 0.9 2.9 2.0 1 2 3.3 potatoes 8.2 2.3 134.0 1.2 12.1 170.0 3.4 2.0 5.4 4.0 2.8 6.9 sweetpotatoes 5.9 1.1 149.0 0.0 7.5 150.0 2.2 0.1 2.3 3.0 0.1 3.1 otherroots 5.1 1.8 28.0 0.0 6.8 28.0 2.5 0.1 2.5 3.0 0.1 3.1 beans 0.7 1.1 363.0 0.0 1.0 361.0 3.4 -0.1 3.3 5.2 0.0 5.2 peas 0.5 1.5 35.0 0.2 0.8 36.0 3.5 0.2 3.7 5.2 0.2 5.4 bananas 6.7 3.0 363.0 1.0 10.0 402.0 3.4 1.0 4.5 3.8 1.0 4.9 peanuts 0.6 1.6 17.0 0.7 0.8 18.0 2.9 0.7 3.6 3.9 0.7 4.7 soybeans 0.6 1.3 37.0 0.7 0.9 52.0 4.2 2.6 7.0 6.2 4.5 11.0 vegetables 12.3 6.0 45.0 0.7 27.0 47.0 7.0 0.4 7.4 7.9 0.4 8.3 h i t s 12.3 6.0 13.0 0.7 24.8 14.0 6.1 0.6 6.7 6.1 0.6 6.7 sugar 3.0 6.0 2.0 0.5 6.2 2.0 6.1 0.6 6.7 6.2 0.1 6.3 coffee 0.7 1.0 31.0 0.1 1.2 46.0 5.2 3.5 8.8 5.2 4.7 10.1 tea 1.2 1.0 12.0 0.1 2.8 25.0 3.0 6.7 9.9 2.9 8 9 12.0 p y r e h 0.2 6.7 3.0 0.5 0.4 15.0 6.7 14.7 22.4 6.7 18.5 26.5 tots1 1686.0 0.5 1914.0 1.0 1.5 Source:Diao andYu, 2006, based on the 2001 Household Survey 5.72 However, further growth and reductionsin poverty are achievable with growth in both the agriculture and non-agriculture sectors (Table 5.18). The model simulations show that if (in addition to the growth in the agricultural sector), there is also more than 6 percent 71A descriptionof the model is presentedinDiao andYu (2005 and2006). 113 annual growth in the non-agriculture sector, through the linkages between non-agriculture and agriculture (scenario 23, in Diao and Yu), growth in the agricultural sector further increases to 6.17 percent, while the GDP growth rate rises to 6.24 percent (Table 5.18, part 1). At this growth rate, per capita GDP grows at 3.44 percent annually, almost tripling the annual growth rate in the base run. As a result, the national poverty rate falls to 42.4 percent, 17 percent lower than the poverty incidence in2005. Table 5.18: Incomegrowthand poverty reductioninthe model (Resultsfrom base-run andgrowth scenario 23) Additional Annual growth rate ("A), Growth rate Growthrate growth from 2005-2015 inbase-run inScenario 23 base-run GDP 3.88 6.24 2.36 AgGDP 3.60 6.17 2.57 Part 1 NonA g GDP 4.08 6.28 2.21 GDP pc 1.15 3.44 2.29 AgGDP pc 0.87 3.37 2.50 NonAg GDP pc 1.34 3.49 2.15 Income for rural HHwith cash crop 3.89 6.33 2.44 Income for rural HHwithout cash crop (10% oftotal rural HH) 3.73 6.01 2.28 Income for ruralmale-headed HH 3.87 6.37 2.50 Part 2 Income for rural female-headed HH 3.90 6.18 2.28 Income for rural group 1 3.70 6.21 2.51 Income for rural group 2 3.89 6.33 2.45 Income for rural group 3 3.91 6.34 2.43 Staple production 3.76 Grainproduction 4.12 6.21 2.45 Root production 2.21 9.62 5.50 Part 3 Pulse and oilseed production 1.44 3.27 1.06 Livestock production 4.28 3.69 2.25 Export crop production 1.21 7.82 3.54 Poverty by 2015 from targeted growth Poverty rate Poverty rate Poverty rate at 2005 by2015 reduction National 59.20 42.40 -16.80 Rural 64.50 46.60 -17.90 Urban 13.80 43.60 -18.70 Rural hhwith cash crops 62.20 78.30 -10.20 Rural hhwithout cash crops 88.50 43.60 -18.40 Part 4 Rural male-headed 61.90 54.70 -16.70 Rural female-headed 71.40 56.90 -16.20 Ruralgroup 1 73.10 46.20 -19.80 Rural group 2 66.00 34.40 -18.00 Rural group 3 52.50 42.40 -16.80 Source: Diao andYu (2006), basedon the EICV 5.73 Table 5.19 summarizes the sources of growth and poverty reduction across the various sectors (agriculture, non-agriculture,crops, and commodities), based on scenario 23 -growth in all agriculture commoditiescoupledwith non-agriculturegrowth. The table illustrates that significant growth in agriculture will still be critical, over the period, in order to reach overall GDP growth targets. It also shows that the growth of staple food crops and livestock production will have the greatest impact on overall growth and poverty reduction. 114 Table 5.19: Sources of Income Growth and Poverty Reduction in the Model Agriculture Non-agriculture Staplescropsandlivestock Cereals Roots Pulses Livestock Subtotal Exportcrops Total Total &bananas & oilseeds Contribution to growth (total is 100) GDP 14.3 3.4 3.1 14.2 35.0 15.1 50.1 49.9 AgGDP 30.1 7.2 6.9 25.1 69.3 28.1 97.4 2.7 Incomefrom diffrereni rural household groups: with cash crop 19.3 4.5 4.1 17.2 45.1 21.2 66.3 33.7 without cashcrop (10% of rural) 26.3 4.3 7.4 7.6 45.6 5.9 51.5 48.5 male-headed 19.1 4.4 4.1 18.0 45.6 22.3 67.9 32.1 female-headed 19.7 4.5 4.5 17.1 45.8 15.5 61.3 38.7 rural group 1 27.7 6.0 5.2 13.6 52.5 19.7 72.2 27.9 rural group 2 20.6 4.6 4.1 14.6 43.9 23.5 67.4 32.6 rural group 3 16.9 4.0 4.0 20.1 45.0 19.8 64.8 35.3 Contribution topoverty reduction (total is 100) Incomefrom diffrerent rural household groups: National 11.5 4.2 9.6 15.0 40.3 20.4 60.7 39.4 Rural 11.8 4.4 9.8 15.6 41.6 21.4 63.0 37.0 with cashcrop 11.2 4.3 9.3 16.0 40.8 22.4 63.2 36.7 without cashcrop(10% of rural) 19.5 6.0 16.1 9.9 51.5 7.3 58.8 41.1 male-headed 12.1 2.9 9.4 16.5 40.9 22.6 63.5 36.5 female-headed 11 8.2 10.7 13.3 43.2 18.5 61.7 38.3 rural group 1 18.7 6.0 9.8 12.1 46.6 20.2 66.8 33.2 rural group2 7.5 2.3 10.3 15.7 35.8 21.8 57.6 42.4 rural group 3 8.1 5.1 8.9 20.9 43.0 22.8 65.8 34.3 Source:Diao andYu (2006), basedon the EICV 5.74 The trade surplus in agricultureis projected to increase, since agriculture imports are smallerthan exports in absolute terms, and exports are projectedto increase at a higher rate than imports. This increase is highest under the full growth scenario - a 6.17 percent 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 $47 million, or 3.5 percent times higher per year than the baseline scenario. The associated projected levels of imports for the scenario are summarized in table 5.20. Table 5.20: SimulatedImports of Agricultural Products(In unit indicated) Base year (2003) Growth simulation (Scenario 19) Production Imports Projected Projected Import annual production in imports in growth rate 2015 2015 2005-15 (1000 mt) (1000 mt) (1000 mt) (1000 mt) ( 7 0 ) Maize 81 11 189 42 15.0 Rice 28 13 108 0 -- Wheat 15 20 36 30 3.9 Beans 258 15 367 96 16.7 Veg. oil 1 6 2 9 3.O sugar 7 11 15 20 5.7 Milk 129 3 323 0 __ Industry (million US$) 286 100 552 124 1.9 Source: Diao andYu (2006), basedon the EICV 115 5.75 Table 5.21 also providesan indication of which sources of growth in agriculturewill contribute most to poverty red~ction.'~ simulation based on staple food-led growth - Le., A growth proceeds as outlined above for maize, rice, wheat, cassava, potato, and soybean, with all other crops left to grow at current trends - results in a 1.26 percent reduction in poverty nationally, and a 1.28 percent reduction inthe rural areas, with a 1 percent increase in GDP per- capita growth. This contrasts with an export-crop led growth, where coffee, tea, and pyrethrum grow at projected levels, with all other crops growing at current trends. In that scenario, the reduction in poverty (resulting from a one percent increase in GDP growth), i s 0.85 percent nationally, and 0.88 percent in the rural areas. For the staple-led growth simulation, soybeans contribute the most to poverty reduction - resulting in a 2.15 percent reduction nationally and a 2.20 percent reduction inthe rural areas, due to a 1 percent increase inGDP. This is followed by growth in the maize sector, which leads to a 1.7 percent reduction in poverty nationally, and a 1.69 percent reduction in rural poverty. Among export crops, coffee has the most poverty reduction potential. A 1 percent increase in GDP per-capita growth led by the coffee sector results in a 1.5 percent reduction inpoverty nationally, and a 1.54 percent reduction inpoverty in the rural areas. Table 5.21: Poverty Reduction -growth elasticity Percentage Reduction inPoverty Rate - due to 1%GDPper capita growthledby aspecific agriculturalsub-sector Poverty-growth elasticity Scenarios: National Rural Staple-led growth (scenario 22) -1.30 -1.31 Cereal-led growth (scenario 4) -1.22 -1.23 Maize-led growth (scenario 1) -1.70 -1.70 Rice-led growth (scenario 2) -0.49 -0.50 Wheat-led growth (scenario 3) -0.83 -0.83 Root crop-led growth (scenario 8) -1.56 -1.58 Cassava-led growth (scenario 5) -1.56 -1.50 Potato-led growth (scenario 6) -1.50 -1.53 Sweet potato-led growth (scenario 7) -2.26 -2.32 Banana-led growth (scenario 9) -1.03 -1.05 Pulses and oilseed-led growth (scenario 12) -2.36 -2.36 Bean-led growth (scenario 10) -2.37 -2.36 Soybean-led growth (scenario 11) -2.16 -2.20 Livestock-led growth (scenario 20) -1.05 -1.07 Poultry and egg led growth (scenario 17) -1.33 -1.35 Other meat and milk led growth (scenario 18) -0.98 -1.oo Export crop led growth (scenario 12) -0.85 -0.87 Coffee-led growth (scenario 13) -1.45 -1.49 Tea-led growth (scenario 14) -0.33 -0.34 Pyrethrum-led growth (scenario 15) -2.97 -3.05 Agriculture-led growth (scenario 22) -1.16 -1.18 Source: Diao and Yu (2006), basedon the EICV 5.76 Income growth will differ across household types according to landholding and whether the household produces cash or non-cash crops. In the case of household types categorized according to land size, the annual income growth rate is 6.21 percent for rural group 1, and 6.33 and 6.34 percent for rural groups 2 and 3, respectively (Table 5.18, part 2). Poverty 72The poverty-growth elasticity used inDiao and Yu (2006) measuresthe responsivenessof the poverty rate to changes inthe per capitaGDP growth rate. The formula uses average annual changes (from the base-year) inthe poverty headcount rate and level of per capita GDP; andthe base-yearpoverty headcount rate and per capitaGDP. The poverty-growth elasticity therefore measures the percentage change inthe poverty headcount rate caused by a one- percent increase inper capita GDP. This is not equivalentto apercentage point change inthe poverty headcount rate. 116 falls in all three rural groups. However, assuming the same rate o f poverty reduction across all groups, the poverty rate i s initially much higher in rural group 1; this group's poverty rate will still be as high as 56.9 percent by 2015, while the poverty rate for rural group 3 will fall to 34.4 percent. 5.77 Differences in income growth also seem related to whether rural households are engaged in cash crop production. For households engaged in such production, total income grows at 6.33 percent annually in scenario 23, while the growth rate for households not engaged incash production is 6.01 percent - a 0.32 percent difference. Due to the relatively slow growth in income, the poverty rate for rural households not engaged in cash crop production falls modestly in scenario 23 (78.3 percent by 2015, from the 2005 level of 88.5 percent). Onthe other hand, the poverty rate for rural households with cash crop production falls to 43.6 percent by 2015, from 62.2 percent in2005 (Table 5.18, part 4). More than 10 percent o f households are not engaged incash crop production, including export crops or crops for the domestic market such as potatoes, fruits, and vegetables. Among these households, 43 percent are female headed and 60 percent are in rural group 1, with a landholding size smaller than 0.3 of a hectare. Lack o f opportunities to be involved in cash crop production determines the low income level and high poverty rate among those households, and also partially explains why these households do not benefit equally from rapid agricultural growth compared to households that produce cash crops. 5.78 There are also differences among householdslinked to the gender of the household head, because women are primarily engaged in food crop production. Annual income growth is 6.37 percent for the rural male-headed households in scenario 23, but i s 6.18 percent for the rural female-headed households, a 0.19 percentage point gap in each year's growth rate. Considering that the poverty rate for female-headed households is higher than for the male- headedhouseholds, such differences inthe income growth rate translate into an enlarged poverty gender gap. The poverty rate for rural male-headed households falls to 43.6 percent by 2015, from 62 percent in 2005, while the poverty rate for the rural female-headed households falls to 54.7 percent by 2015, from 71.4 percent in2005 (table 5.18, part 4). 5.79 These results indicate that the growth rate of food staple production is more pro- poor than production of export crops, since it increases food availability. Growth in agriculture exports also contributes to poverty reduction, especially for rural group 3, but the main benefit of export crops i s their effect of earningforeign exchange. 5.4.1 PossibleTerms-of-Trade Effectfor Key Agricultural Products 5.80 Growth, especially when it is unbalanced, may not always benefit all producers, depending on the pattern of growth across sub-sectors and the related changes in relative prices. Given the high growth rate targeted for rice, potatoes, poultry, and eggs, the model simulation results illustrate the effects of a likely reduction in prices when growth in these commodities is too high compared with growth in the other agricultural sub-sectors and the nonagricultural sector (Figures 5.5 and 5.6). 117 Figure 5.5: Relative Prices for Selected Crops Figure 5.6: Relative Prices for Selected (base year 2005 = 1) Livestock Products (base 2005 = I) I-#-Maize -A- Potatoes -Sweet potatoes- mBanana/ 1.1 1.6 1.o 1.4 0.9 1.2 0.8 1.0 0.7 0.8 ..................... 0.6 0.6 0.5 %-*--Ir +-+ 0.4 0.4 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: Diao andYu, 2006; model simulationresult 5.81 For example, the models suggests that if rice production grows as projected, and accounting for an increase in the demand from rice, the rate o f growth in production will still be much higher than increases in demand. The result is that rice prices will fall by more than 20 percent, which will benefit consumers while decreasing the returns for rice farmers. 5.82 Potato prices will also fall (about 10 percent), given an annual growth rate of 5.4 to 6.9 percent between 2005 and 2015. Expanding markets or demand for potato (including opportunities to increase potato processing or exports), can help reverse these trends. In the livestock sector, 12 and 16 percent o f annual growth in poultry and eggs, respectively, between 2005 and 2015, will cause a decline in the prices o f these two commodities (Figure 5.6), even though their demand is highly income elastic. 5.4.2 Agriculture Growth Needs To Increase by Almost 9 Percent a Year To Cut Poverty in Half by 2015 5.83 Generating growth across all commodities in the agriculture sector, at the base-run level, will allow the country to meet the goal of 6 percent growth for the agriculture sector. The model shows that another scenario which combines growth in agriculture commodities with an assumption of modest growth in the non-agricultural sectors significantly reduces the poverty rate. However, it does not enable the country to meet the poverty reduction goal of halving poverty by 2015, even ifthe targetedgrowth for the non-agricultural sector is achieved. 5.84 A further scenario is developed in which much more rapid growth is assumed for most agricultural commodities. This is coupled with an assumption of higher growth in the non-agricultural sector. With these assumptions, the model yields an annual growth rate for agriculture GDP of 8.7 percent, and a growth rate of 7.2 percent in non-agricultural GDP, between 2005 and 2015. These results will contribute to overall GDP growth of 8 percent over the period. At these growth rates, the first MDGofhalvingthe poverty rate by 2015 is achievable. 118 Figure 5.7: Poverty reduction in the MDG scenario (AgGDP gr: 9.0%; NonagGDPgr: 7.2%; GDPgr: 8.0%) 90 80 70 60 50 40 30 20 10 National Rural Ruralhh Ruralhh Rural Rural Rural Rural Rural 'th cash without male- female- group 1 group 2 group 3 rn Poverty rate by 2015 crops Source:Diao and Yu, 2006; modelsimulationresults 5.85 It is important to note again, however, that even with such high growth rates at the national level, poverty reduction will still be modest among some household groups, especially those that have initial high poverty rates. For these groups, more targeted growth and poverty reduction policieswill be necessaryto enable them to share equally inthe growth gains. 5.5 SUMMARY AND CONCLUSION 5.86 Growth in the past has come about mainly through increased movement of labor into agriculture production, and this growth has been led by the food sub-sector. There is potentialto sustainand even improve the level of growth observed over the past decade; however, this will require an immediate increase in the application of improved inputs and production methods. Agriculture-led growth can lead to substantial poverty reduction, but the greatest poverty reduction will come from a combination o f agriculture and non-agriculture growth. In either case, some groups will benefit more than others, depending on their initial level o f poverty and the types of crops they grow. 5.87 Although food-crop driven growth in the agriculture sector is necessary for achieving strong GDP growthin the near future, it is not sufficient for long-termsustained growth. Growth in the non-agriculture sector will also be critical for sustained growth and poverty reduction. In this regard, measures to improve non-farm income would also have significant benefits for the poorer households. 5.88 Non-farm income is a key means of supplementing income for poorer households cultivatingsmall plots on marginal land. I t also provides a form of finance for agriculture production. Non-farm income is a critical pull factor to stimulate domestic demand, and also ties into increasing value added for export markets. In this regard, measures to strengthen the linkages between the farm sector and industries that use their inputs, such as policies to strengthen contract arrangements betweens farmers and manufacturing firms in the food and beverage industries, can helpmaintainand increasedemand for agriculture output. 5.89 Given limited resources,there will be a need to focus on broad-based measures that will have greatest impact in terms of improvingproductivityand the quality of output. To improve input use by farmers, it will be necessary to initially focus on institutional aspects, such as establishing farmer organizations, commodity chains, and distribution channels; the 119 dissemination of information, technology, and improved methods; and marketing. In addition, improved water management will be particularly important where crops are not sufficiently vertically integrated and input use is low. The PDL HMO'~ public works program could be used to implementthese measures, while also generating rural employment. 5.90 In the case of the non-farm sector, government's role should be to ensure an enabling environment for private sector development. To identify constraints to growth, it will be important to determine the size and characteristics of the informal sector, and the reasons why a large percentage of firms (and the population) chooses to remain in that sector. A key priority will therefore be to identify the main constraints in the non-agriculture, along with measures to eliminate the causes of informality, related to low capacity and information on businessdevelopment. 5.91 Looking forward, the Government has various growth scenarios to consider: (i) continue with previous trends, which would likely lead to lower growth or even a return to historical levels o f 3 percent growth, particularly if growth in the non-farm sector i s stagnant, leading to a limitedreduction in poverty; (ii) implement policies that would yield some increase in inputuse (even ifmoderate, due to constraints to delivering extension services) and a annual growth rate of 5 to 6 percent, provided that measures are adopted to address the issue o f water management. (iii)strive for the high-growth scenario, with overall growth of more than 7 percent, which can be achieved only with exceptional utilization o f improved inputs and technologies and an immediate yield response or quick improvement inthe quality of land. This latter scenario would need to be coupled with measures to improve the overall business environment and stimulate growth in the non-agriculture sector. The next chapter examines the constraints and growth potential for that sector. 73PDL HIM0stands for Programmede dkveloppementlocal a haute intensit6 de main-d'oxvre, or Labor-Intensive Local DevelopmentProgram. 120 6. NON-AGRICULTURE SECTOR-THE PULLFACTOR FORECONOMIC GROWTH 6.1 INTRODUCTION 6.1 Chapter 4 showed that inthe absence o f substantial growth inthe non-agriculture sectors, with strong links back to agriculture, the Government's Vision 2020 growth targets will not be achievable, and the reduction in poverty rates by 2015 will be modest. Growth in the non- agriculture sector can play a strong role in helping to increase domestic purchasing power and demand for agricultural products. It could also facilitate the transformation o f the economy from one that i s subsistence basedto one that is market oriented. An examination of the non-agriculture sector indicates that it i s largely informal, and that there are potential areas for growth and increased formalization ofthe private sector. 6.2 A reductionin the levelof informalityof the economy would promotegrowth in the non-agriculture sector and reduce poverty. This will require an improvement in the overall business environment. The analysis in this chapter shows that the informal sector has grown significantly, reflecting growth in the service and construction sectors. These trends indicate the potential of the service sector for income generation and poverty reduction. However, for the microenterprise sector, the need for improved skills in business development and the high incidence of informality i s an indication of the need to adopt measures to improve the business environment, particularly inareas related to the cost of doing business. 6.3 For example, small and medium enterprises (SMEs) report that low purchasing power and market demand, physical isolation, lack of market development activities, and lack of access to finance are constraints to business development. Inadequate access to and facilities for technical and managerial training are other binding constraints. Lack o f training translates to lack o f knowledge about services available to business, and ways to access finance and neededcapital; this increasesbusinesscosts and keeps firms inthe informal sector. 6.4 In the industrialsector, the high cost of energy and low availability of skilled labor are key factors limitingproductivityand growth.Giventhat manufacturing is largely basedon agriculture, addressing these issues would both help to increase demand for agricultural and facilitate the process of pulling labor out of agriculture. The manufacturing sector accounts for roughly 10 percent of GDP. Between 40 to 50 percent o f manufacturing firms are engaged inthe food and beverage industry, followed by 48 percent inmining. Roughly 60 percent of these firms export more than half of what they produce, followed by 38 percent o f mining firms, which export 32 percent of their output. However, the high cost o f electricity has increased production costs and limited capacity utilization of firms in the sector; these, combined with the scarcity of skilled labor and the high cost of finance, pose severe constraints to production. Therefore, as with the microenterprise sector, improvement ofthe overall business environment will be critical to growth ofthe sector. 6.5 Skills development needs attention across all sectors, but particularly in the service sector, given Rwanda's focus on the high-end tourism market and its aim o f becoming a service- oriented economy. Inthe case of ICT, Rwanda could beginto enter the market through segments lower in the value chain, such as product assembly for domestic or regional markets. In the meantime, aside from the broad-based measures that will be needed in the areas o f education, 121 science and technology, Government can begin to explore areas for immediately improving ICT use and literacy among the population 6.6 The rest of the chapter further elaborates on these issues and is divided into three sections. The chapter begins with an overview o f the microenterprise sector, which constitutes a large part of the non-agriculture sector, and pays particular attention to the issue o f informality and its relation to the business environment. This i s followed by an examination o f the potential of boththe manufacturingand services sectors to contribute to overall economic growth. 6.2 MICROENTERPRISE DEVELOPMENT INFORMALITY AND 6.7 There is an understandable urgency to develop export markets for Rwanda, but it should not overshadow the importance of developing domestic market linkages and integrating SMEs into export value chains. The two objectives are inter-dependent and re- enforcing. Only by increasing the efficiency and productivity o f domestic markets and increasing the availability of skills and services at the local level, will Rwandan value chains be able to eventually increase their competitiveness in regional and global markets. Consequently, due to its landlocked location, Rwanda needs to closely examine the options to develop efficient market linkages that reduce transport and transaction costs and facilitate improved understanding o f internationalmarkets by firms inthe microenterprise sector. The discussion inthis section shows that two key issues need to be addressed to achieve this task. The first issue i s the need to improve the skills and knowledge o f local firms on issues related to business development and management. The second issue pertains to ensuring improved access to finance. 6.8 The micro and small scale enterprise (MSSE) sector is largely informal and is a significantlygreater employer than the formal sector. An employment study completed by the Ministry of Planning in 1985 estimated that 105,000 are employed in the informal sector, compared to 12,442 in the formal industrial sector - a ratio o f eight to one. The 1992 National Labor Survey estimated that 92 percent o f the active population o f 2.9 million i s linked to the informal economy, with the non-agriculture informal sector employing 197,289 persons in rural areas; while employment in the formal sector for the whole country was 173,010. The study indicated that more than 50.3 percent o f urban employment was from the informal sector, with the formal sector employing 49.5 percent. In 1992, average employment in the informal sector was 2 persons per enterprise. 6.9 The World Bank's Doing Business database estimates that the informal sector accounted for more than 42 percent of GDP in 2005. However, this is most likely a lower level estimate. For example, evidence from the EICV indicates that over 95 percent o f non-farm enterprises are in the informal sector. Inaddition, around 85 percent of employees (particularly women), are engaged inthe private informal sector. An assessment of the microenterprise sector completed in 2004 identified 70,000 formal and informal micro and small-scale enterprises employing more than 300,000 people.74 The study estimated that the 70,000 identified firms represents a lower bound on the number of firms, since it does not include non-business activities such as home-based production of banana beer; also, mining workers, mobile venders, and informal caterers may also be excluded from this estimate. However, the size of these excluded groups would not significantly increasethe overall number of enterprises beyond 80,000. 74 P. Nugawela, V. Bicali, and V. Ndagijimana(2004), "Review andAssessment of Micro- and Small-scale EnterprisesinRwanda, Backgroundpaper for the CountryEconomicMemorandum. The study draws on primary data collectedfor the assessment as well as previoussurveysand studies ofthe sector. Sources are given inthe paper. 122 6.10 The majority of small businesses participate in the retail sector (47 percent); businesses in other sub-sectors include wood works (7.4 percent), tailoring and garments (7.3 percent), building materials (6 percent), handicrafts (9 percent), construction (4.7 percent), and services (9 percent). Of the current small businesses operating inRwanda, nearly 70 percent are less than 10 years old, and only 3 percent are formally registered as limited liability companies. The importance of infrastructure is highlighted by the fact that nearly all (94 percent) of identified SMEs are located near a road. However, only 61 percent have access to electricity, 56 percent have a water source nearby, andjust 42 percent have fixed telephone lines, although 54 percent do have access to cell phones. 6.11 The production system of microenterprises can be home-based cottage industries, artisans, activities that operate in the market place or in open areas, and enterprises operatingin a more formal or factory-type structure. 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 SME, as well as the large-scale enterprises, are generally organized factory-type operations. We can distinguish these sub-categories of enterprises by the numbers employed, organization, system o f production (factory type or non- factory type), value o fturnover, and investments (Table 6.1). Table 6. 1: Structure ofthe MSSE Sector, Classified by the RwandanSystem of Production Production System Home based Disperse production Small scale - Factory Type/organized and structured Production Informal InformaVformal enterprises - Units Formal organized Enterprises intransitionto small andmedium Family production Micro enterprises Small enterprises Medium Scale Large-Scale Enterprises Enterprises Artisans Organized Artisans Organized, Organized, Organized, Home base food Formal & Informal Operations ina structured, formal structured, formal Preparations/ units specific building- enterprises enterprises Processings Productions-open factory type Modern enterprise, Modem enterprise, Bananas,sorghum air, Market place Associations Registered companies Registered companies Beer, basketry etc. Retail Shops Formal Employment: Employment: less Employment: 10to Employment: Employment: than 10 30 30 to 100 100or more FamilyLabor Less than 10 10-30 employees 30 to 100 100 or more employees employees employees Source: P. Nugawela, V. Bicali, and V. Ndagijimana(2004). 6.12 Interms of organization,MSSEs are individually owned. At least 50 to 75 percent of these individual units group themselves into economic interest groups or associations, usually o f about seven individually owned units. They employ one or two workers, either hired or family labor, includingunremunerated family labor. 6.13 The sector is composed of associations, cooperatives, and individual production units. The procedures for registering MSSE associations at the provincial level are not cumbersome. However, provision o f services to improve incentives for registration, and some further streamlining of registration processes, could help facilitate the movement of more firms into the formal sector. An estimated 43 percent of the MSSEs identified were working as 123 associations; 39 percent were individually owned; 4 percent were managed by families; and 12 percent were incooperatives. Only 3 percent have formed limited liability companies. About 38 percent of MSSEssurveyedwere rural basedwith another 17percent located inor near secondary market areas. 6.14 The economic activities of microenterprise firms have shifted over time. A 1990 national survey o f 7000 micro and small enterprises conducted by MINICOM, in collaboration with UNIDO, indicates that after 1994, the sector shifted toward retail activities. However, more recent data suggestthat there is a gradual increase inproductive relative to retail activities. 6.15 According to the 1990 survey, MSSEs are engaged in the manufacturing and services sector. These units had more than 100,000 employees. The survey found that 35 percent of the all MSSEs were engaged in woodworking and furniture making, followed by 28 percent in textiles and garments, and 19 percent in minerals and mining. The remainder were distributed among agri-business or processing (3 percent), hides and leather products (1 percent), metal products (8 percent), and repair services (4 percent). 6.16 This contrasts with the 2004 assessment, which showed a relative reduction in the degree of productive compared to retail and trade-related activities. According to the 2004 assessment, 47 percent of microenterprise firms were engaged in retail sales, 9 percent were engaged inhandicrafts, another 9 percent inservices, 8 percent in carpentry and furniture making or woodworking, 11 percent inconstruction, and the rest in various agriculture based and mining activities. Overall, 2 percent of the firms were engaged in ago-business and food processing, with 35 percent o f the identified units located in Kigali City. Most agro-processing microenterprises have been recently created. The situation i s similar with private consultancy firms. The major proportion of recently created enterprises i s located inKigali city. The relative reduction in the degree of productive (as opposed to purely retail) activities may also reflect the degree o f disruption that occurred as a result ofthe events in 1994. 6.17 Indications are that the numbers of MSSEs, particularly in the informal sector, are growing and that there is potentialfor these to impact overall economic growth and poverty reduction. As shown in Table 6.2, 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 o f 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 half o f all 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 inthe services sector, followed by the construction sector (Table 6.3). Table 6.2: The Structureof GDP Accordingto Degreeof Informality, 1995 -2003 Annual Percentage GDPEstimate Share inTotal GDP Growth Rate Contribution to 1995 2003 1995 2003 1995 -2003 Total 326.0 641.4 100.0 100.0 8.8 8.8 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 0.1 Source:MINECOFIN,National accounts, and World Bank staffestimates 124 Table 6.3: Structure of Informal Component of GDP, 1995and 2003 (Inpercent) I 3s 199s 2003 Ip savicg 5.2 10.8 CJ Tmnsport and 1.9 3.3 Cbmrmmications IIl Q l andTouism I 0.7 1 1.2 I o>llstrrrtion I 2.1 I 4.6 I l ~ ~ a c t u - j l l s I 2.2 I 1.6 I Source: MINECOFIN,National accounts, andWorld Bank staff estimates 6.18 Among the MSSEs surveyed, about 69 percentwere created after 1995, compared to 14 percent created between 1985 and 1994. Most o f the artisan, agro-based and food- processing enterprises were created recently. Private consultancy companies are also recent developments. Major portions of these enterprises are located in Kigali City (11 percent), Umutara (11 percent), and Gitarama (7 percent), with the rest evenly distributed across the remaining regions. However, there are problems with product and market development. Many o f the MSSEs and associations lack innovative ideas on new products and services, quality improvements, packaging, market development, etc. Many firms produce the same products or services they are already have market saturation. This situation i s common in all the provinces, except for some firms located inKigali City, Umutara, and a few urban centers. 6.19 The 2004 assessment of MSSEs identified potential new sub-sectors that offer new investment opportunities. These include fish farming, new agri-businesses related to processing, consultancy services, training centers, new handicraft sectors, new products in the furniture industry (cane and bamboo furniture), recycling products, and new service sectors inthe provinces. These needto be identified and promoted. 6.20 The informal sector makes up a significant share of the urban economy. For example, the city o f Kigali is estimated to have a population o f just over 600,000 people (according to the 2002 Census), and it is increasing by 10 percent a year. About 45 percent o f the population i s female. InNyarugane district, inthe city o f Kigali, it is estimated that about 40,000 people, or 80 percent of the labor force, are employed inthe informal sector. This contrasts with employment inthe formal private sector, whichwas estimated at around 7,000, or 13.5 percent o f the total labor force for the 6.21 Earnings in the non-farm informal sector are higher than what would be typically earned in subsistenceagriculture, where the daily rate was estimated at FRw 300 per day in Recensement des activitis lucratives dans la Ville de Kigali (2002). 125 2004. Estimates are that 75 percent of permanent personnel in the informal sector, earn around FRw 15,000 per month or morethan FRw 500 per day. Interms ofturnover, the 2004 assessment estimated that roughly 73 percent o f firms had a inventory turnover of less than FRw 500,000, while 36 percent had aturnover o f between FRw 100,000 and 500,000. 6.22 Althoughfirms 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 bythe 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 o f which result inhigher costs o f doing business. 6.23 There is also a gender dimension to the MSSE sector. A study of women entrepreneurs conducted in 2002 found that they are overwhelmingly engaged in the informal sector. The study found that the enterprises are mainly inthe retail sector (82 percent). The rest are engaged in the services (16 to 17 percent) and manufacturing (1 to 2 percent) sectors. Most (64 percent) have an annual turnover of FRw 500,000. The level of education is generally low: 70 percent have a few years of primary education, or none at all. Only 17 percent have some level of secondary education. Overall management capabilities and accounting practices are very poor. Very few have the appropriate business training, and use their own savings or family support for financing. Financing requirements for startup average 50,000, with 58 percent needing less than FRw 1,000,000. Among the constraints identifiedby the women were limited market demand and low purchasing power o f clients; taxes paid to local authorities; inadequate access to funds for investment; and inadequatetechnical and advisory support. 6.24 A more recent assessment, conducted in 2006, 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 representedinthe informal enterprises (58 percent of women compared to 42 percent o f men). Inpartially formal enterprises, 60 percent are owned by men, compared to 40 percent by women. This result is supported by data from the 2001 Household Survey, which found women to be over-represented inthe informal service sector. In terms o f education, 22 percent of women in formal or partially form 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 of enterprises operating at the district or local level. These results indicate that informal firms are more likely to be localized in scope, and run by women with lower education. Moreover, adults with secondary or higher education tend to be less represented in the informal sector. Overall, 10 percent o f adults with secondary education work inthe private informal sector, and less than 2 percent of those with highereducation work inthe sector. The low level of education o f MSSE owners and operators limits their ability to access information that could help them to expand their business. 6.25 Regionaldifferences exist among MSSEs, with those in the urban areas being more modernized than their rural counterparts. Urban-based MSSEs use more power-generated modem equipment and hand tools compared to those in rural areas for the same sub-sector. For example, in carpentry and woodworking enterprises, those located in urban sectors use more power-generated machines, while their rural counterparts use traditional mechanized hand tools. Apart from Kigali City, the MSSE activities in all other areas are almost similar in terms of products and service sectors, level of organization, andmarkets. 126 6.26 Regardless of their regionallocation, firms in the informal sector rank the lack of informationabout where and how to register almost as high as the tax burden imposed by registrationas a barrier to doing business. Firms stay inthe informal sector to avoid payment of labor taxes and social security contributions, income or profit taxes, and compliance with labor regulations, while taking advantage o f 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 neededto find the relevant information about how to comply (cited by 36 percent o f partially formal and 46 percent o f informal enterprises). The reasons given by firms for not obtaining legal licenses, including those neededto operate, are primarily related to the high cost of obtaining the licenses and the lack of information about how and where to get them. Firmsalso indicated that difficulty in meeting the licensing requirements, along with weak government enforcement, contributes to their decision to remain informal. 6.27 Another potential constraint that leads to informality pertains to costly and moderately lengthy business registration procedures. Business registration is relatively costly, at FRw 50,000, and i s likely a disincentive to many informal firms. According to the Doing Business database for 2005/6, the cost of starting a business i s on average 315 percent of Rwanda's per capita income. Finally, there may not be an incentive to register, since small firms do not require trade licenses to use local inputs. Lengthy procedures to register land(354 days on average, according to DoingBusiness 2005) are also a deterrent to formality. 6.28 Further probingon the issue of tax burden and its effect on the decision to remain informalrevealedthat the tax rate is less of a problemthan the lackof knowledgeabout tax levels and the process for filing returns. The most important reason cited by firms for transitioning from informality to being partially formal was the realization that taxes were not as highas they thought andthat administrationwas fair. However, 30 percent ofthe 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 assessment concluded that reducing the bureaucracy involved in becoming official could facilitate the transition o f firms to the formal sector. 6.29 While manufacturing firms 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 commonconstraint to businessoperationsand 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 microenterprises 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. The 2006 FIAS study on informality, however, did find that firms citing the costs o f electricity and finance as major constraints also assessedthese costs to be extremely severe. 6.30 Results from the studies cited above generally supported the results of the 2001 Household Survey, which found that most MSSEs were mainly individually funded and were reluctant to apply for loans due to their lack o f collateral. However, many were not aware o f the available financing facilities and application procedures, or they proposed business ventures that were not considered creditworthy. 6.31 Despite this, the 2004 assessment finds that 55 percent of firms surveyed had accessed microcredit (65 percent consider this their main source o f funding), and have done so through the 127 Banque Populaire. Only 26 percent have attempted to access commercial bank financing, and 19 percent have applied to the National Deveopment Bank (Banque Rwandaise de Developement or BRD). The other banks are considered inaccessible dueto the requirements for collateral. 6.32 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. SME 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 inshort supply. As a result, the private sector still imports skilled technicians from neighboring countries, despite Rwanda's high level o f unemployment, and at a high cost to the economy. The lack of access to and facilities for technical and managerial training i s highlighted as a bindingconstraint to SME development inRwanda. 6.33 The following issues have been identifiedas key constraints to businessgrowth. Many o f themreinforcethe on-the-ground lessonsfrom private sector development projects: Access to large, profitable markets and market development facilities o Linkages to sufficiently large, profitable markets for most Rwandan SMEs are constrained by a low level of domestic demand and purchasing power; physical isolation; and low availability of market development facilities, such as affordable marketing channels, sales outlets, information and methods for pricing products and services, and mechanisms for product development and quality control. 0 Access to financing o Despite the numerous microfinance organizations operating in Rwanda, only 55 percent of SMEs have accessed microcredit. The high cost of financing is still a constraint to investment and growth at the firm level. Lack o f working capital and collateral are the most common problems mentioned in accessing financing. Poor knowledge of applicationprocedures i s also a big issue. Human resource capacity and accessto training facilities o Inadequate access to and facilities for technical and managerial training are highlightedas bindingconstraints in supporting SME development in Rwanda. The Kigali Institute of Science, Technology and Management (KIST) i s involved in training for SMEs, including support for young entrepreneurs, but the school i s based inKigali and can only cover so muchofthe training needs across the country. 0 Related and supporting industriesand institutions o About 50-70 percent of SMEs 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 to support SMEs, the major to not have long-term plans and/or are ineffective in serving members. In order to fully integrate SMEs 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 SMEs, and networks to facilitate marketingactivities, are critical. 128 6.3 OVERALLBUSINESSENVIRONMENT 6.34 Many first-generation reforms related to the removal of price distortions and subsidies have been implemented, but the high cost of doing business continues to encourage informality. While still extremely expensive, starting a business in Rwanda is relatively fast and efficient. As illustrated in Table 6.4, among peer group countries in sub- Saharan Africa, the average number of procedures for starting a business is 9 and takes an average of 21 days. Registering property, however, takes as long as 354 days in Rwanda, compared to 48 days in Uganda and 65 days, on average, for sub-Saharan Africa. Labor market flexibility also appears to be an issue in Rwanda. Rwanda's scores in the Doing Business 2005 report for difficulty o f hiring(89) and firing (60) are much higher than the regional average (36 and 39, respectively). The Doing Business 2005 data indicate areas of the business environment that still require reform; however, more information is needed on the actual monetary cost o f the constraints, including the cost of operating an existing business incomparison to other countries. 6.35 Rwanda needs to take additional steps to further improve its business environment and make its industrial sector competitive with those inother countries at a similar stage o f economic development. Table 6.4 summarizes some basic observations from the World Bank's Doing Business Databaseon Rwanda and selectedAfrican countries. Table 6.4 : Doing Business in Rwanda: Starting a Business (In unit indicated) No. o f Time it takes Cost (as % of Minimumcapital Country procedures (indays) P/C income) required (YOo f P/C income) Burundi 11 43 201 0 Congo (Dem.) 13 155 503 216 Ethiopia 7 32 65 1532 Kenya 13 54 48 0 Rwanda 9 21 280 0 Tanzania 13 35 161 6 Uganda 17 36 118 0 SSA average 11 64 215 297 Source: Doingbusinessdatabase, 2006 6.36 As shown in Table 6.4, in terms of the number of procedures involved in starting a business and the time it takes to do so, Rwanda is withinthe range and in some cases even better than most ofthe comparator countries inAfrica. There i s no minimumcapital requirement to start a business in Rwanda, whereas the average minimumcapital requirement in Sub-Saharan Africa is 297 percent of per capita income. Interms o f the number o f procedures and the time it takes to start a business, Rwanda rates favorably among most of the African countries, although it does not rate as well with regard to the cost of doing business. Starting a business costs an average o f 280 percent of per capita income in Rwanda, compared with 48 percent in Kenya, 65 percent in Ethiopia, and an average of215 percent for the whole of Sub-SaharanAfrica. 6.37 As shown in Table 6.5, the same data reveal that much needs to be done to further improve firms' participation in international trade. The export potential o f Rwandan establishments seems to be constrained, compared with neighboring countries, by current trade policy and practices. In fact, in cross-border trading, Rwanda underperforms in relative 129 comparison with the average in Sub-Saharan Africa. The number of documents required for exports from Rwanda stands at 14, requiring about 27 signatures. These numbers are high compared with the SSA averages o f 8.5 documents and 19 signatures. Also, the time it takes for exports, according to the Doing Business database, is 63 days in Rwanda, compared with 30 in Tanzania, 45 in Kenya, 58 in Uganda, and an average o f 49 days for the whole o f Sub-Saharan Africa. 6.38 Commercialand contractlaws and regulations are either outdatedand incomplete, or brand new and not yet tested, and both present barriersto investment. In addition, there are systematic legal and judicial impedimentsto the resolution of private sector disputes. The backlog in the judicial system restricts the ability to apply the law consistently and predictably. Rwanda only has about 100 lawyers, making contract enforcement much more costly than in neighboringcountries. Table 6.5: Doing Businessin Rwanda: Trading across Borders No. of documents Total number Time it takes country required for exports of signatures to export (days) Burundi 11.0 29.0 67.0 Congo (Dem.) 8.0 45.0 50.0 Ethiopia 8.0 33.0 46.0 Kenya 8.0 15.0 45.0 Rwanda 14.0 27.0 63.O Tanzania 7.0 10.0 30.0 Uganda 13.0 18.0 58.0 SSA average 8.5 19.0 49.0 Source:Doingbusiness database, 2006. 6.39 Corruption in Rwanda is considered to be relatively less of a problem compared to many of its neighbors. While this is an advantage for businesses operating inRwanda, exporters still face costly rent-seeking charges in transporting their goods to port, whether through the Northern or Central corridor. 6.40 However, the majority of the private sector, especially across SMEs, operates with low levelsof education, and without adequate training, experience, or skills to managetheir businesses competitively. A legacy of central planning has hindered development of an entrepreneurial culture and resulted inworkforce development challenges inRwanda such as: 0 Integrating a new generation of relatively more skilled management into the business community; 0 Rehabilitating S M E s after the loss of a critical mass o f skilled people throughout the country, and fostering enterprise learningina post-conflict environment; 0 Upgrading the technical expertise of the labor force to be able to take advantage of technology efficiencies; and 0 Understanding the needs of increasingly sophisticated buyers on domestic and internationalmarkets. 130 6.41 These challenges present a number of key issues and opportunities that should be addressedby the Government, academia, and the private sector collectively to raise the level o f sophistication of Rwandan businesses. Efforts should include: Increasing the supply of and access to management training programs for private entrepreneurs and managers, by o Transforming the old Centre de formation professionelle pour cadres de 1'administration de Murambi/Gitarama into a training institute offering courses inadministrationandmanagementfor boththe public andprivate sectors; and o Increasing the number o f and accessto technicaltraining programs. Creating incentives for increasing the quantity and quality of on-the-job training programs. o Ina recent survey of Rwandanbusinesses, 72 percent said they provided training at the place o f business, while only 28 percent said that they usedexternal aid in training. 6.4 THEMANUFACTURING SECTOR 6.42 The 2005 Rwanda Industrial and Manufacturing Survey shows that growth prospects in the manufacturingsector depend on removing constraints posed by the high costs of electricity, transport, and finance. The manufacturing sector represents about 10 percent of the GDP and contributes about 5 percent to total exports. Manufacturing i s dominated by 20 firms, which accounted for 65 percent of total manufacturing in2002. The sector is based on agro-industrial activities, which account for 80 percent of total manufacturing output. Thus, there is an opportunity to strengthen the links to the agriculture sector, thereby helpingto expand market opportunities for farmers. 6.43 However, high electricity, and transport costs have contributed to the stagnation of the sector. The sectoral share of manufacturing in GDP has gradually declined from a level of 10.6 in 1995 to 8.9 percent in 2004. The trends inthe share o f electricity, gas, and water, which intotal account for roughly 0.3 to 0.4 percent of GDP, reflect this stagnation. By contrast, the share of the construction sector in GDP has doubled over the same period. The manufacturing sector accounted for 11 percent o f overall GDP growth, with 96 percent o f this coming from the food and beverages sub-sector. Overall, the growth rate of the manufacturing sector has been on a declining trend, over the past decade. 6.44 This declining growth trend in manufacturing has posed a problem for export growth, since the manufacturingsector is relatively export oriented. For example, between 1995 and 2002, manufacturing exports experienced a negative annual growth rate o f about 8 percent, on average. This number is far below the 8.25 percent average growth of the manufacturing sector and the 6 percent average growth o f the country's total exports during the same period, highlighting the need to improve the sector's competitiveness in international markets. 6.45 This section summarizes the main trends in the manufacturing sector and examines the level of constraints posed by the poor state of infrastructure, low availability o f skills, and lack o f access to finance. The first section provides an overview o f the manufacturing and business 131 environment. The second section analyzes recent performance of the sector, in comparison with neighboring countries. This i s followed by an overview of the factors that have constrained performance. The final section examines the prospects for future growth and competitiveness, and attempts to prioritize among constraints. 6.4.1 Characteristics of the ManufacturingSector 6.46 An overview ofthe sector is basedonthe RwandanIndustrial and MiningSurvey. R I M S was financed by the Government and conducted between late 2005 and early 2006 by the National Institute of Statistics, with technical support from the World Bank. The survey covers 111 establishments, of which about 20 were from the mining sector. The remaining establishments represent a census of manufacturing firms, covering mainly food and beverages, construction materials, chemicals and plastics, metals, wood furniture, textiles, garments, and leather processing activities. The data cover the period2004-2005. 6.47 Results from the survey (Figure 6.1) show that the majority of establishments inRwanda are concentrated inthe food and beverages sub-sector (41percent), followed by the miningsector (26 percent). Figure6.1: Sectoral distributionofManufacturingFirms (In Percent) rText'Garmt'Leather 3% Metals Furniturdwood 23% Source:RIMS,2005 6.48 About 63 percent are in the micro and small size category, employing fewer than 50 workers; and only about 17 percent of establishments inthe data employ 100 or more workers. In terms o f their legal status, the majority, 47 percent, are privately owned limited liability companies, followed by sole proprietorship and partnerships, which are each about 14 percent (Figure6.2). 132 Figure 6.2: Distribution of Establishments by Legal Status and Size Group (Inpercent) DistniimufIXaby L@ Sam( JXttibtiCno f m aby SizeGroup(O/4 Wh O h W ? o 47?? 5 1% 14OA 14% 120h 0% Pubk P r i v a t e e m pumership Sols cthcr company (notbted), ( m s d pro+tonhip IhnaedbbGity EA-) (indi4wl)) company Source:RIMS, 2005 6.49 In terms of location and age of establishments, the data show that most of the industrial sector activities (63 percent of the establishments covered by the survey) are located in and around the capital, Kigali. The average age o f establishments located in Kigali i s 14 years (with a median age o f 9 years); while for those located outside Rwanda the average age is 13 years (witha median age ofjust 4 years). This impliesthat development o f the industrial sector in Rwanda, particularly in locations outside Kigali, is a very recent phenomenon. Another characteristic of the Rwandan businesses is that a good number o f them participate in international trade - 42 percent reported having exported at least 5 percent o f their total sales duringthe fiscal year prior to the survey interview. 6.4.2 Performanceand GrowthProspectsof IndustriesinRwanda 6.50 Performance of the sector is measured by productivity per worker, employment, and export growth. The data show that the sector is productive and export oriented; and that productivity (measured both as output per worker and as value added per worker) is within range for the region (Figure 6.3 and Figure 6.4). Median value added per worker i s estimated at US$2,910, which i s below the value for Kenya (US$3,457), but above the values for Tanzania (US$2,061) and Uganda (US$lOSS). Across industries in Rwanda, the median value per worker i s 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 S), and furniture making (US$1,533). Productivity per worker inRwanda is also higher for exporters (US$9,SOO) than for non-exporters (US$1,700). 6.5 1 The data also show that the miningand construction sectors registeredthe highest median value added per worker (respectively, at US$6,237 and US$6,094), equivalent to more than twice the average for the whole survey data set (see Figure6.3). Looking at the various size groups, the highest (median) value added per worker was registered by micro enterprises (those with fewer than 10 workers) compared to the larger firms. Productivity i s also higher for exporters than for non-exporters. The median value added per worker for micro enterprises was US$6,334 compared, for example, with US$3,883 for medium and US$2,146 for large enterprises. The value added differential between exporters and non-exporters is, in fact, very high by any standard. A preliminary comparison o f estimated median value added per worker suggests that productivity inthe exporting group i s close to eight times that of non-exporters. 133 Figure 6.3: Median Value added per worker in Figure 6.4: Median Value added per worker- Rwandaby industry (in USdollars) Rwanda and comparators (in US dollars) 10000 - 8000 - 6334 439; Source:RIMS, 2005; and InvestmentClimate Surveys, variousyears. 6.52 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 experience at the manageriallevel, and engagement in exporting all have a positive effect. The data indicate that strengtheningthe export orientation of firms may be an important means o f stimulating growth in the sector. In addition, regression analysis summarized in Annex 2 highlightsthe importance of policies to promote exports for strengthening the export orientation of firms. Figure 6.5: Capacity Utilization -Rwanda and NeighboringCountries 175%, 1 Source:RIMS, 2005; and InvestmentClimatesurveys., variousyears. 6.53 Despite their relatively better performance in terms value added per worker, Rwandanfirms reported being unableto use their full capacity. Duringthe fiscal year prior to the survey - 2004 - the average capacity utilization in Rwanda was about 62 percent; and the average for fiscal year 2003 was about 59 percent (Figure 6.5). These figures are comparable to those in neighboring African countries. Industrial establishments in African countries generally perform well below their capacity for a variety of reasons, including shortage of materials, shortage o f capital, low demand, and various regulations. 6.54 Growth performance can also be examined through employment trends. It is estimated that less than 5 percent o f the total labor force is employed in the industrial sector, although the level of industrial employment has been picking up recently. R I M S 2005 suggests that average employment growth between 2003 and 2004, for the whole industrial sector, averagedabout 6 percent, with a median growth rate of 5 percent. The textiles and garment sector dominated, with a 15 percent median growth rate of employment for the same period; followed by the foodbeverages and metal/furniture sub-sectors, which registered median growth rates of 134 13 and 10 percent, respectively. The median growth rates for the construction materials and mining sectors were close to 0 percent. 6.55 The survey data show that small-sized enterprises (10-49 workers) and large enterprises (100 or more employees) registered median employment growth rates o f about 6 and 5 percent, respectively; while for micro and medium-sized establishments, median employment growth was zero. That is, employment growth focused on the larger enterprises. 6.56 A look at growth performance by export status reveals that, despite the very highlevel of value added per worker created by exporters compared with the non-exporters, a relatively larger share o f employment growth i s generated by the non-exporting establishments. The median growth rate o f employment for the non-exporters is about 6 percent, compared with 3.5 percent for the exporters. 6.57 Firms in the manufacturing sector are already relatively export oriented - 42 percent of firms surveyed export 36 percent of their sales. These figures are high compared to figures for Uganda, where 19percent o f firms are exporters, with exports averaging 10 percent o ftheir sales. In the case of Tanzania, 26 percent of firms are exporters, with exports average 12 percent of their sales (Figure 6.6). InKenya, 57 percent ofthe firms are exporters and export 17 percent of their sales. These figures are encouraging, giventhat Rwandan firms face serious infrastructure- related obstacles. However, these figures also reflect the effects related to the size o f the domestic market and demand. 6.58 An examination of exports at the sector level in Rwanda reveals that food and beverage firms are more likely to be exporters than firms in other sectors; and firms in the food and beverage sector export a larger share of their output than do firms in other sectors. About 61 percent o f all food and beverage firms are exporters - well above the average of 42 percent for all firms in Rwanda(Figure 6.7). This i s followed by the miningand metal and furniture producers, of which about 38 percent are exporters. The average share o f export value to total sales inRwanda's food and beverage industryi s about 54 percent, compared with an average of 36 percent across all firms in Rwanda. About 20 percent of firms in the chemical industry have engaged inexport, and there i s room for growth. There are also prospects for growth inthe fruits and vegetables sector (provided the issues related to low agriculture productivity discussed inChapter4 are addressed), aswell as inwoodworking and artisanal activities. Figure6.6: RegionalParticipationin internationaltrade: Share o f exporters and average exports (In Percent) mfhnsapnting pacentofsales 57 Source: RIMS, 2005; InvestmentClimate surveys, variousyears. 6.59 An examination of the regional dimensions of exports and imports reveals that most exports go to the EU and most imports come from COMESA. The majority of imports in 2004 135 originated from neighboring COMESA countries, particularly Uganda and Kenya (13 percent each), followed by Asia, which accounted for 12 percent. A very small percentage o f imports originated from other neighboring African countries. Only around 1 percent of imports of industrial inputs originated from Tanzania, about 2 percent from the Democratic Republic of the Congo (DRC), and another 2 percent from other COMESA member countries. Outside COMESA, the major share of imported industrial inputsto Rwanda originated from Asia (at 12 percent in 2004), followed by the European Union (8 percent) and the Middle East (only 3 percent). This impliesthat COMESA dominates as the origin o f industrial inputsto Rwanda, but that close to three-fourths of the total imports for industrial use (72 percent) coming from COMESA originated from two countries: Uganda and Kenya. Figure 6.8 shows the comparison o f average share of exports to selected regiodcountries from Rwanda, with the corresponding average share of imports from the same regiondcountries into Rwanda for the year 2004. Figure6.7: Percentof exporters and averageexports by sector 6.60 The major destinations for exports are the European Union, North America, and Asia. These three regions account for, respectively, about 33, 15, and 11 percent o f Rwanda's industrial exports. Exports to neighboring African countries, on the other hand, were reported to have been relatively small. Only about 7 percent of exports were destined for Kenya, while another 2 percent were destinedfor Uganda, 3 percent to DRC, and 5 percent to other COMESA countries (for a total of 17 percent for COMESA) in2004. No industrial exports to Tanzania were reported duringthis period. 6.61 These trends suggest that measuresto increase the capacity utilized by firms would help to increase production and exports. Growth prospects are good for the food and beverage industry, as well as the metals/furniture and garment industrieswhich export to both international and regional markets. However, their exports are limitedby their level of production. 136 Figure6.8: Direction of Trade: Share of exports (by destination) and imports (by origin) (In Percent) - Source: RIMS, 2005. 6.62 There are particularlygood prospectsin the miningsector, 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. Much of the sector's growth has been due to the rising price of coltan, a mineral widely used in cellular phones and other electronic devices. Other ores are mined for use in construction. However, the industry is labor intensive and dominated by small-scale artisans. In 2004, an estimated 50,000 people were involved in mining. Growth in this sector could be a significant source of non-farm income for households inrural areas. 6.63 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 in Australia, the United States, and South Africa. However, it is important that Government not overlook the small-scale miner and the potential benefits of small-scale miningrelatedto job creation, off-farm employment, and poverty reduction. Table 6.6 shows the estimated numbers o f small-scale miners inkey comparator and competitive countries. 6.64 Through the privatization o f 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)helping small-scale miners obtain access to finance; (ii)improving health, safety, and working conditions, and reducing exposure to harmful chemicals; and (iii)providing training and transfer of skills to artisan miners, to improve productivity and reduce the price of production. Table 6.6: Employment in Small-ScaleMining Firms, RegionalComparison, 2005 Country Numberemployed Rwanda 50,000 Burkina Faso 100,000 - 200,000 Ghana 200,000 Malawi 40,000 Mozambique 60,000 Tanzania 550,000 Zambia 30,000 Zimbabwe 350,000 Source: InternationalInstitutefor Environmentand Development 137 6.4.3 Overviewof Constraints to Growthand Performanceof industries inRwanda 6.65 Electricity, finance, transportation, and skills are among the major business obstaclesinRwandathat limit growthand exports. The survey instrument used for the RIMS asked respondents whether a given type o f potential business impediment i s an obstacle to the growth and performance o f their establishment, and the level o f severity associated with each o f the obstacles. Respondents identified the five most important impediments to the growth and performance as: power (identified by 81 percent o f firms), cost o f financing (54 percent), tax rates (49 percent), transportation (41 percent), and shortage o f skilled workers (40 percent). Figure 6.9 shows business obstacles in Rwanda, along with the percentage o f respondents who rated them as major to severe. Figure 6.9: Ten Major BusinessObstaclesinRwanda (percent ofJrms that complained) 80% - 6oo/o - 49% 41% 4oo/o 40% - 34% 20% - I I I I 1 I I I I Source: RIMS,2005. 6.66 Firm perceptionsand rankingsof businessconstraintsinRwandavary significantly, according to the characteristics of the establishments surveyed. The following figures show such differences by region (between Kigali and outside Kigali), and by export status (between establishments that export at least 5 percent o f their sales and those with no exports). 6.67 Figure 6.10 shows that there i s a regional variation in the perceived investment climate constraints between establishments in Kigali and those located outside Kigali; with the availability o f skilled workers, access to finance, and telecom being o f greater concern to the latter. On the other hand, some o f the major business constraints seem to be common to both exporters and non-exporters: both groups identified power, cost o f financing, and tax rates as major to severe business obstacles. There are some differences between these groups, however. Electricity, skills o f workers, access to finance, macroeconomic stability, and business licensing were identified as impediments by more exporters than non-exporters; 85 percent of exporting establishments, for example, identified power as a serious obstacle, compared with 77 percent o f the non-exporting group. On the other hand, tax administration and anti-competitive practices were identified by a relatively larger share o f respondents in the non-exporting group than by those inthe exporting group. 138 Figure6.10: Major Obstacles,by Localand Export Status (percent identifiing them as major to severe) _. Bvhcation I Source: RIMS,2005 6.68 Business obstacles seem also to be different for the different size groups. For example, electricity, which i s the obstacle identified as major to severe by the largest majority o f establishments, is more of a concern to SMEs than to larger establishments (Table 6.7). The same i s true o f access to finance and issues o f macroeconomic instability, while issues such as anti- competitive practices, taxation, and cost o f finance are of relatively more concern for larger establishments. Table 6.7: BusinessObstacles inRwanda,by Sizeof Firm Percent of Firms evaluating Constraint as "Major" or "Very Severe" All Micro Small Medium Large Rwanda (empl(10) (empl 10-49) (empl50-99) (empl 100t) Electricity 81% 86% 82% 81% 74% Cost of Finance 54% 43% 53% 71% 47% Tax Rates 49% 43% 44% 62% 53% Transportation 41% 36% 40% 57% 26% Skills of workers 40% 36% 42% 48% 26% Access to Finance 34% 36% 37% 33% 26% Macroeconomic Instability 30% 36% 32% 33% 16% Tax Administration 27% 7% 30% 43% 16% Anti-competitive practices 27% 7% 21% 52% 32% Telecommunications 23% 21% 28% 10% 26% Access to Land 23% 21% 28% 14% 21% TradeICustoms Regulations 23% 14% 25% 38% 11% Source: RIMS.2005. 6.69 The discussion makes it clear that problems in power generation and distribution are the largest of all identified constraints to the development and competitiveness of the industrial sector. In the same survey instrument, establishments were asked to single out one obstacle that they think i s the biggest in terms of its effect on their growth and performance, and the aggregated results from this variable confirm that power is at the forefront: Overall, 48 percent o f respondents identified electricity or power shortages as their single biggest obstacle - 50 percent of non-exporting establishments, and 45 percent of exporters. This suggests that power-related problems may be the most important factor defining under-utilization o f capacity. 139 6.70 Objective indicators of the cost of energy also confirm the severity of power problems in Rwanda (Figures 6.11 and 6.12) During the fiscal year preceding the interview, Rwandan establishments reported having faced power outages an average o f 23 times, and being without power for an average o f 19 hours during each outage. The resulting loss o f output due to such power interruptions was also quite high; they reported that losses due to unreliable power supply were equal to an average o f 10percent o f their outputs (and some reported losses o f up to 85 percent). Due to its higher level o f connectivity, the problem i s most pronounced inKigali. Figure 6.11: Infrastructure indicators:Electricity Figure 6.12: Comparing Share of Energy Cost n (In unit indicated) as percent of sales (In percent) ..... ..... ..... .....:: 5.6 ::::: ....._... 20 4.0 10 2.0 0.0 0 Source:RIMS(2005) and Investment Climate Surveys, various years. 6.71 Cross-country comparative data between Rwanda and other countries also reveal that infiastructure-related problems, particularly electricity and transport, are o f greater concern to businesses inRwandacompared to those inother countries. Figure 6.13: Infrastructure: Electricity and Transport as major concerns (percent ofjrms complained) 58 45 Source: RIMS, 2005; and InvestmentClimate Surveys 6.72 I n the case of electricity, the data show that close to 75 percent of Rwandan businessesattempt to copewith unreliablesupply from the public grid by havingor sharing a generator (Figure 6.14). An estimated 45 percent of power used by Rwandan establishments comes from these generators, adding to the high cost o f production. However, smaller establishments, fully domestically owned establishments, those that do not participate in international markets (non-exporters) are less likely to own or share a generator, mainly due a lower level o f need, as well as the cost of upkeep, including the cost o f fuel. Estimates are that electricity shortages have increasedthe average costs to businesses by 16 percent. 140 Figure 6.14: Generator ownership by various establishment characteristics (Inpercent) Source:RIMS, 2005; and InvestmentClimate surveys, variousyears. 6.73 The cost of finance is another important obstacle to business performance and growth.The cost of finance was identifiedas a constraint by 54 percent of firms surveyed, and as a major or severe constraint by 34 percent of firms. The percentage o f firms identifying cost o f finance as an obstacle was larger (i)for firms in Kigali than for those outside o f Kigali; (ii) for non-exporters than for exporters; and (iii) for medium and small firms than for micro and large firms. 6.74 Smaller firms are more credit constrained than larger firms, given the need for collateralto obtaina loan. The data show that the majority o f loans to Rwandan establishments are provided against one or a combination of 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 of the owner/manager, such as hisher house (37 percent); and other types o f assets (19 percent). The approximate average value of collateralrequiredto get a loaninRwandais about 105 percent of the loanvalue. 6.75 Loan terms for small enterprises may also contribute to their credit constraints (Figure 6.15). In 2004, the average annual interest rate on loans was a reported 17 percent, and the average term o f 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 i s 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 17percent. 141 Figure 6.15: Interest rates paid (Znpercent)and terms of loans (months) by size interestrateperyear (averag) 51 errpl40 q l 1 0 - 4 9 anpl50-99 errpllW AUF&mch Source: RIMS,2005; and InvestmentClimateSurveys 6.76 Of the establishmentsthat did not have bank loans, about 80 percent reported that they did not apply to get a loan. Only 3 percent of firms that did not have bank loans reported that their application was rejected. These results reflect both the need for both collateral and the lack of knowledge of how to correctly complete a loan application. Figure 6.16: Establishments with overdraft facility or bank loans (In percent) 75% AmesstoFinana? 65% 55% 5W? - 41% 25% - W ? 1 I Source: RIMS,2005. 6.77 Overall, the majority of firms in the sector have access to some type of finance (Figure 6.17). The shares of firms that have access to overdraft facilities or bank loans are, respectively, 41 and 55 percent. Thus, the majority of establishments in Rwanda have access to some type of finance (either o f overdraft facilities or bank loan or both). The majority o f micro establishments also reported having access to some form o f finance from banks (either a loan, an overdraft facility or both). 142 Figure 6.17: Establishments with either or both overdraft facility andbank loans (Inpercent) arpl-40 arpl10-49 arplD!99 arpllw All- Withloans WahOwdaft ID WahM O Dor both Source:RIMS2005. 6.78 A largerproportionof overdraft facility or lines of credit owned by microand small establishments remain unused compared with those owned by larger firms (Figure 6.18). Just over 40 percent o f establishments have an Overdraft facility or line o f credit, and the figure is larger for smaller firms. However, about 21 percent o f the average value o f these facilities was not used inthe year prior to the interview period (2004). The corresponding figures for the micro and small size establishments were well above the averagesfor mediumand large establishments. Figure 6.18: Average unusedshare of overdraftsor lines of creditsby size of establishments(Inpercent) A 1 Q 35 30 25 21.0 20 15 10 5 0 Source: RIMS, 2005 6.79 In terms of working capital, the data show that retained earnings are the major financingsources, followed by local commercial banks. Generally, retained earnings account for 52 percent of total financing for both working capital and new investments, and commercial bank loans account for 20 percent of financing. Other sources o f finance include trade credits (2 percent), equity sales (8 percent), and informal sources (1 percent). For working capital, 53 percent of financing comes from retained earnings, compared to 80 percent o f firms in Uganda and 46 percent of firms in Kenya. For new investments, 51 percent o f financing comes from retained earnings, compared to 71 percent inUganda and 45 percent inKenya (Table 6.8). 143 Table 6. 8: SourcesofFinancingfor WorkingCapital andNew Investments (In percent) Sources o f fmancing Working Capital N e w Investments Uganda Kenya Rwanda Uganda Kenya Rwanda Retained Earnings 80.0 45.8 53.0 71.1 44.6 51.3 Domestic Commercial Banks 5.7 23.5 19.0 11.6 25.4 21.3 Foreign Commercial Banks 1.3 1.7 0.2 1.8 2.0 1.2 LeasingArrangements 0.1 0.4 2.2 2.4 0.2 1.3 Trade Credit 5.3 15.3 3.8 0.5 3.1 1.o Credit Cards 0.0 1.4 0.3 0.0 0.7 0.0 Equity, Sale o fstocks 1.8 1.1 8.5 2.0 0.3 7.6 FamilyFriends 1.4 1.2 3.O 2.0 0.8 4.6 Informal Sources 0.4 0.0 1.5 1.5 0.0 1.2 Other Sources 4.2 4.2 8.2 6.7 6.9 8.7 Source:RIMS, 2005; Investment Climatesurveys, various years. [ Figure 6.19: RelativeCost of the Main Constraintsto Doing Business (as a share of TotalSales) 7.4 4 3 2 1 n Source: RIMS2005 6.80 Firm perceptions indicate that the poor state of infrastructure should be the main area of focus. Although access to and cost of finance is also an issue, the constraints posed by infrastructure are more burdensome (Figure 6.19). The discussion above and in Chapter 3 shows that the cost o f finance is not exorbitant, and access i s improving through the growth of microfinance institutions. The Government is also focusing on identifying alternative financing mechanisms relevant to micro and small enterprises. 6.81 Firms identified the tax rate as an important constraint, yet the marginal effective tax rate (METR)compares favorably with that of other countries in the region; and it is not likely to be a significant barrier to investment. The FIAS study on METRs in Rwanda finds that these are generally lower than or roughly equivalent to those in the neighboring countries, except for taxes on inventories in countries with lower rates o f inflation, and taxes on some investments in Zimbabwe. METRs on investments in inventories in Rwanda would be reduced dramatically iffirms usedthe last-in-first-out inventory accounting system. 6.82 Along with the constraints posed by infrastructure, the lack of sufficiently skilled labor is an issue that will need to be addressed. A close look at the composition o f the industrial workers from the survey data confirms this. Indeed, 44 percent of industrial workers in 144 Rwanda did not complete primary education, compared with only 1 percent in Kenya and 4 percent in Uganda. The data also show that the share of workers who have completed secondary or vocational schooling is 38 percent (the same as in Tanzania), and less than the proportion for Kenya (67 percent), and Uganda (61 percent). Similarly, the share o f Rwandan industrial workers who have completed university level education is lower than that o f the neighboring countries. (Figure 6.20). Figure 6.20: Workers education: Share of workers by levelof educationcompleted (In percent) Rwanda Tanzania Kenya Uganda El None E Primary 0 Secondarynechnical 0 University Source: RIMS, 2005; Investment Climate surveys, various years. 6.83 A large share of the labor force in Rwanda (67 percent), have primary or lower level of education (compared to 59 percent in Tanzania, 21 percent inKenya and 24 percent in Uganda). Consequently, a larger share of Rwandan businesses identified a shortage o f skills as a major or severe business concern, compared to other countries inthe region (Figure 6.20). 6.84 A study by Temesgen and Ezemenari (2006) shows that the returns to education in the manufacturing sector increase with the level of education completed. Moreover, the study indicates that the completion of secondary schooling i s the most critical factor for providing a supply oftrainable labor for firms. 6.85 The availabilityof skilled laborwas reportedto be a major constraintby 40 percent of firms (Figure 6.21). The high returns to education, by level completed (e.g., 18 percent increase in income from completion of primary school), arising from employment in the manufacturing sector76indicates that educated labor is scarce in Rwanda. The increases by level are higher for women (48 percent for primary education) than for men (8 percent for primary education). Women have, on average, ten years o f education compared with nine years for men. They also have comparable years of experience inthe labor market and an equivalent number of hours worked per week across all education levels, but their average earnings are well below the 76RwandaIndustrialand ManufacturingSurvey(2005). 145 average for men. Women tend to be concentrated in low-paying occupational ~ategories,~'and very few have high-paying managerial positions. Figure 6.21: Availability of Skills and other Obstaclesto BusinessinComparator Countries (Inpercent) Source: RIMS, 2005; InvestmentClimatesurveys, various years. 6.86 Women are disproportionately under-represented in managerial positions in manufacturingfirms. Women's share inthe sample is about 30 percent, but they occupy only 17 percent of management positions (Table 6.9). On the other hand, there are relatively more women in the professional, services, health, and other non-production categories. Within the female sub-sample o f 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 o fwomen are inthe management category, compared with 5 percent for men. Table 6.9: Average Education, Experience, and Weekly Earnings inRwanda(In unit indicated) Average Average work Union Average Hours Average weekly education Experience membership worked weekly earnings (years) (years) (%I (number) (FRw) Men 9.2 20 23.00% 42.1 42,956 Women 10.3 16 24.10% 41.3 22,239 Total 9.6 21.8 23.30% 41.9 37,665 Source: RIMS, 2005. 6.87 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 77 The2001HouseholdLivingConditionsSurvey also foundthat more women inRwanda are engaged in low-payingoccupations. 146 assessment of the skills mix o f the labor force, in relation to the needs o f the industrial sector, is 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 important for the overall improvement in quality of, and access to, the broader educational system. 6.88 As shown above, the skill level of currently available workers in the Rwandan labor market i s identified by 40 percent o f respondents as a major o f severe constraint - more than macroeconomic instability (29 percent), anti-competitive practices (27 percent), and crime (5 percent) within Rwanda. By comparison, skills were identified as a major or severe by 24 percent of respondents in Tanzania and 30 percent inUganda. Clearly, issues related to crime, theft, and disorder are not significant business concerns compared with other problems, inbothRwandaand neighboring countries. It is also worth noting that compared with other countries, regulation issues are not identifiedas serious impediments inRwanda(Figure 6.22). Figure 6.22: Regulationissues as major concerns (Percent offirms complained) Source:RIMS, 2005. 6.4.4 Summary and Policy Recommendations 6.89 Productivity in Rwandan industry i s comparable to, and in some cases even better than, that of neighboring countries; however, productivity could be further improved by easing the constraints that firms face. Infrastructureproblems, particularly power and transportation, as well as lack o f skilled manpower, are among the major obstacles cited by the majority o f entrepreneurs and supported by objective measures. Thus, improving power generation and transport is clearly an investment and policy priority, to spur the development o f the industrial sector and improve the private sector's participationinthe country's development. The problemo f low access to and highcost of finance may require the development of financial instruments that are tailored to the specific needs of different types o f firms, includingthe large percentagethat are export firms. 6.90 It is important to raise the skills of workers to a level that meets the requirements of the industrial sector. An assessment i s needed of the skills mix outcomes o f the current curriculum and educational policy, with attention to on-the-job training. 147 6.5 THESERVICE SECTOR 6.5.1 Introduction 6.91 Within the service sector, both the PRSP and Vision 2020 identify tourism and ICT as key areas of focus. This section assesses the current situation inthe tourism and ICT sectors, and outlines the factors required to lay the ground work for developing those sectors. The results show that it will be necessary to have programs for skills development to target the market segment that the Government has in mind. To ensure that targets are achieved, there also needs to be increased focus on improving coordination and monitoring. As discussed above, infrastructure, particularly energy services, also needsto be improved. 6.5.2 Overviewof the TourismSector 6.92 While tourism contributes only marginallyto the economy, it hasthe potentialto be a key catalyst of economic growth and competitiveness over the next decade. In 2004, tourism accounted for only 0.9 percent of GDP. By comparison, Kenya's tourism receipts were 3.4 percent o f GDP; Uganda's were 4.0 percent; and Tanzania recorded the highest tourism receipts in the region at 9.6 percent o f GDP. Box 6.1 shows that Rwanda i s between the exploration and involvement stages inthe tourism sector. That is, recognizing the importance of the sector, the Government has taken the lead in investing in infrastructure and product development, although with limitedparticipation from the private sector. 6.93 Rwanda's main tourist attraction is gorilla tracking in the Virungas. The rare mountain gorillas are found only inthe Virunga Mountains shared by Rwanda and DRC, and in the BwindiImpenetrable ForestNational Park inUganda. There are no mountaingorillas inzoos around the world, and Rwanda's mountain gorillas are said to be the most accessible due to political instability in DRC, and the park's proximity to the capital (2 hours from Kigali, compared to 6 hours from Kampala). Rwanda is also endowed with a large number o f indigenous specieso f mammals, birds, reptiles, amphibians, and butterflies. 148 6.94 The main tourism sites in Rwanda include three national parks - VirungaNolcanoes National Park (PNV), Nyungwe National Park (PNN), and Akagera National Park (PNA) - as well as Lake Kivu and the surrounding towns (Gisenyi, Kibuye, and Cyangugu); the National Museum of Rwandain Butare; the King's Palace inNyanza; and the Gisozi Genocide Memorial Center inKigali. 6.95 As a latecomer in the region to the global tourism market, Rwanda's strategy is to market itself as an add-on destination for visitors to its East African neighbors. Under the EAC umbrella, Kenya, Tanzania, and Uganda are already moving toward joint marketing, and plans are underway to harmonize hotel standards, classification, and licensing in the region. Plans are also underway to liberalize the sector in order to market the EAC as a single tourist destination. Experience from Southeast Asia and Central America shows that well-designed regional tour packages can be beneficial, particularly for countries that offer a small niche. Rwanda's primate safari and cultural heritage sites can be a strong compliment to Kenya's Masai Mara Safari or Tanzania's Serengeti Safari, as well as the various popular beach destinations of East Africa. 6.96 The Government's approach is to focus on the high-end tourist market, while engaging local communities in an environmentally and socially sustainable manner, with the goal of generating US$lOO million in 2010. Rwanda's tourism development strategy, developed in 2002, aims to attract 70,000 visitors (from different parts o f the world, the region, and within Rwanda), each spending an average of US$200 per day and staying for one week. The strategy identifies three target groups - eco-tourists, explorers, and business travelers - and aims to create high-endtravel experience^.'^ 6.97 Along with eco-tourists and business travelers, Rwanda's tourism strategy paper has identified"explorers" as a main target segment. Rwanda's cultural product, however, is yet to be defined and developed. The United Nations World Tourism Organization (WTO) defines cultural tourism as "movements o f persons for essentially cultural motivations, such as study tours, performing arts and cultural tours, travel to festivals and other cultural events, visits to sites and monuments, travel to study nature, folklore or art, and pilgrimages." A growing number o f visitors are special interest travelers who rank the arts, heritage, and/or other cultural activities as among the top five reasons for traveling. According to the WTO, cultural tourism accounts for 37 percent o f all trips worldwide, and is growing by 15 percent every year, while tourism as a whole is growing at the rate o f 5 percent annually. 6.98 If developed well, Rwanda's fascinating cultural and historical heritage can be a major source of tourist attraction. The nation's cultural resources include heritage sites, museums, arts and crafts, music, and dance. Beyond the few existing sites such as the palace in Nyanza and the national museum in Butare, new sites should be identified and developed. In addition to introducing Rwanda's culture to the world and raising tourism receipts, this would helpto achieve a better distribution o ftourists and enable more Rwandans indifferentparts o fthe country to benefit economically. 6.99 The strategy document rightlyemphasizesthe needto avoid mass tourism, and aims to attract a low volume of high-spendingtourists. As a small country, where about 90 percent of the population i s engaged in agricultural activities and population growth i s 2.7 percent a year, Rwanda's natural resources are already overstretched. Poorly planned and managed tourism can '*RwandaNational Tourism Policy, February 2006, p. 14. 149 destroy ecological systems, raise the cost o f living for local people, and damage social and cultural traditions. Therefore, the need to carefully design a development plan and manage the country's environmental assets cannot be overemphasized. 6.100 In this regard, Rwanda has valuable lessons to learn from its neighbor, Kenya. In the 1970s, large investments were made along the coast of Kenya, with little control or regulation by the Government. After two decades of impressive growth in highvolume but low-yieldtourism, popular destinations such as Mombassa, Malindi, and Amboseli National Park started to suffer from environmental degradation. In the 1990s, with a large inventory o f accommodations and large package tour operators buying rooms in bulk at discounted rates, package prices fell drastically as demand declined. 6.101 In order to avoid such problems, Rwanda needsto develop a plan that provides the framework for development, management, and monitoring of investment in tourism infrastructure,as well as measures for quality assurance and environmentalsustainability. The strategy states that Rwanda will implement five tourism region development plans - for Kigali, Butare, Akagera, Kivu, Nyungwe, and Ruhengeri - within five years. As o f September 2006, only the Nyungwe Forest National ParkPlanhas beencompleted. 6.102 The strategy to share revenuefrom tourism earningswith communitiessurrounding the protected areas will help in the conservation of the nation's natural resources, and can make a significant contribution to poverty reduction. Since human-wildlife conflict and the shortage of natural resources (land, firewood, etc.) can impede local support for national parks, channeling tourism revenue to support development for local communities can improve attitudes toward conservation. Further, since tourism development in general may not have a substantial impact on reducingthe level of poverty, specific strategies and interventionsare neededto ensure that the poor also benefit from the overall growth o f the sector. For example, measures could be adopted to develop parallel activities or interventions linkedto the tourism industry, such as the production of handicrafts. 6.103 While not a core function inthe tourism industry,handicraftsplay a significant role in helpingto shape the tourist experience, and providea key channel for income generation amongthe poor. Currently, handicrafts constitutejust 0.1 percent o ftotal exports from Rwanda, but growth in tourism will also increase demand for Rwandan handicrafts and provide an opportunity for Rwandan producers to refine their work in preparation for growing exports. For example, countries such as Thailand, Indonesia, India, Vietnam, and China all had well- developed domestic handicrafts sectors prior to exporting globally. 6.5.3 Performanceof the Sector- Opportunitiesand Constraints 6.104 Since the re-launching of the tourism sector in 2002, the tourism industry has achieved steady growth in arrivals and receipts. The number o f tourist visits to the national parks increasedfrom 1,663 in 1995 to a peak of 27,000 in 1994, with a slight decline to 24,120 in 2005. However, the level of tourist visits is still below the 1984 higho f 39,000. 6.105 The end-of-year report prepared by the Rwandan Office o f Tourism and National Parks (ORTPN) in 2005 shows that the target for numbers o f visitors was exceeded in 2004 (27,000 compared to 20,000), and was almost met in2005 (24,120 compared to 25,000).79 These figures, ''Dataareavailable only for 2004 and 2005. 150 however, represent the number o f visitors to the three national parks, not the total number of tourist arrivals. In2004, the immigration department recorded35,000 international visitors. It is immigration records that are commonly usedto estimate tourist arrivals. Figure 6.23: Average Tourism Spending in Selected Countries in2004 (In USdollar) I I I I , 1 I I 1 Source: World Bank (2006): "The InvestmentClimate for the Hotel Industry inEastAfrica" 6.106 In terms of tourism receipts, the targets have been surpassed for both 2004 and 2005. In 2004, revenue from tourism amounted to US$17 million (compared to the target o f US$12 million), and 2005 earnings reached US$26 million (compared to the target o f US$22 million). A comparison of Rwanda's average spending per tourist with that o f its East African neighbors reveals that at US$485, in 2004, Rwanda's performance was better than the 2003 performance o f Kenya, Uganda and Tanzania. Rwanda's performance was higher than Kenya's at $365, but lower than Uganda's earningsat $868 and that o f Tanzania's at $1,280. 6.107 Due to lack of data, the actualeconomic impact of tourisminRwanda is not known. However, an indication of the potential benefits stems from the fact that tourism encompasses a number of sectors, including lodging, transportation, retail, food services, and recreation. Its overall impact on the economy includes the effects of direct, indirect, and induced spending. The direct impact represents the value added in those sectors that interact directly with the tourists, such as hotels and tour operators; the indirect impact represents the benefit to suppliers to the direct sectors, such as the farmers that supply the hotels; and the induced impact measures the impact o f tourism-generated wages as they are spent inthe economy (e.g., the portion o f wages of hotel employees spent on goods and services produced locally). There could also be beneficial externalities and spin-off effects through markets that develop around tourist activities, such as handicrafts. 6.108 This points to the need to have in place mechanisms and institutions to enable monitoring of key indicators in the sector, to ensure effective regulation and policy development. A first step inthis direction is to have well-defined indicators and targets for the sector. Monitoringthe appropriate indicators will also be instrumental in ensuring that the sector i s developing in line with the Government's strategic vision. 6.109 The Rwandan Office of Tourism and National Parks (ORTPN) has prepared a strategic plan coveringthe period2004-2008; however, the plan needsto be results oriented. 151 The strategic plan outlines ORTPN's objectives, along with indicators linked to specific goals; and identifies 11 strategic programs to support the mission o f the agency. The results orientation of the plan needs to be strengthened to increase the degree of prioritization and specificity in some key areas. This will make it easier to monitor and achieve results to achieve Government's targets. 6.110 Progressmade over the lasttwo years indicatesthat the goals set for 2010 -interms of tourist arrivals and receipts-will be attained(Figure 6.24). The national tourism strategy has set the goals of increasing tourist arrivals to 70,000, increasing the number o f days per visitor from 4 to 7, and raising the amount spent per day per tourist to US$200 by 2010. If achieved, these ambitious goals would increase overall earnings from the tourism sector to close to US$lOO million. This amount could be even exceeded, provided that the necessary investments are undertaken intime. Figure 6.24: Tourism Receipts in Rwanda in2004 and 2005, ActualandTarget (inmillionsofUSD) I 1P I 15 10 5 0 I 20% m I Source: ORTPN (Office Rwandaof TourismandNational Parks) 6.111 Judging from performance over the last two years, Rwanda seems on track for meeting the goals set for 2010, interms oftourist numbersand spending. However, consideringthat most of the planned investments to be undertaken between 2003 and 2010 have not yet materialized, the fact that the country was able to meet or surpass its goals with the current level of tourism infrastructure suggests that the targets may need to be revised upward, and that with targeted investment, Rwandacan attract more tourists and a higher level of spending, renderingthe sector among the top foreign exchangeearners. 6.112 Efforts to attract private investment and promote public-privatepartnershipsin a variety of tourism activities, including hotels and lodges, lake activities (rowboats, canoes, kayaks), and entertainmenthave been very limited. ORTPN and the Rwanda Investmentand Export Promotion Agency (RIEPA) have a key role to play in this regard. The continued improvement of the overall investment climate i s also critical. For example, primate discovery tours were designed in 2003, in line with the sector development objective o f promoting ecotourism. The circuit - consisting of Nyungwe Forest, Lake Kivu, and Parc Nationale des Volcans (PNV) - features a primate safari that includes the mountain gorillas of PNV, as well as 13 primate species in the Nyungwe Forest and other unique fauna and flora of the country. The primate safari i s designed as an 8-day tour. The plan, however, has not been operationalized, due mainly to lack of adequate accommodations in Nyungwe, lack of a boat on Lake Kivu, and the poor condition of the road from Cyangugu to Gisenyi (link between Nyungwe and PNV). 152 ORTPN recently secured funding from USAID to implement the Nyungwe development plan. It is crucial that at leastthe minimumlevel o f investment- accommodations inNyungwe and a boat on Lake Kivu- is inplace by the end of 2007. 6.5.4 Recommendations Issues and Constraints - 6.113 This assessmentofthe tourism sector highlightsthe needto address several key issues: A results-oriented plan is needed that identifies short, medium, and long-term results rather than activities, along with specific interventions and resources required to achieve those results. Inaddition to producing results, this approach improves transparency and accountability, promotes participation and empowerment, and builds capacity for evidence-based decision-making. Results orientation i s also particularly important to ensure that the Government's vision o f an environmentally sound tourism sector remains on track. In particular, an updated strategic plan could address issues such as zoning (the classification and conditions o f land use) and the carrying capacity o f the main tourist attractions (the number o ftourists that can be accommodated in a particular placewithout damaging the environment). Inthe case of national parks, for example, the presence o f large numbers of people can adversely impact natural resources as well as the quality o f visitors' experiences. The plan could also identify ways of strengthening the linkages between tourism and poverty reduction efforts. Beyond employment creation, tourism's impact on reducing poverty can be enhanced through, among other things, the provision of support for the development of arts and crafts, and micro and small enterprises. 0 Monitoring and evaluation would need to be an essential element of the strategy. Data on the various categories of visitors - leisure, business and conferences, visiting friends and relatives - who have different expenditure and length of stay characteristics, would be usefulinrefining the sector strategy and marketingefforts. Inadditiontobasictourismstatistics suchasarrivalsandexpenditures,the economic and social benefits and costs o f tourism need to be measured. Changes in the well-being of host communities (in terms o f employment, access to infrastructure, availability o f social services, number o f immigrants, etc.) and the level of tourist satisfaction (measured by sense o f value for money, number o f complaints received, etc.) are key indicators o f the sustainability of tourism. In particular for Rwanda, since ecotourism i s the chosen type of tourism, and because ecotourism takes place in a relatively undisturbednatural environment, it i s critical that key indicators are specified and monitored on a regular basis. Finally, certification programs could be another means (particularly for engaging the private sector) o f monitoring and regulating the sector, to ensure uniform standards. Worldwide, there are numerous types of certification programs, including self, second, and third-party certifications. There are more than 60 programs (two-thirds o f which are in Europe) that offer independent third-party certification, which guarantees that a business or an activity complies with certain established criteria. Certification programs 153 are available for accommodations, transportation and tour operations, national parks, etc. Inadditionto athird-party assessment, the certification processprovides a framework for improving business management. The establishment o f a certification program in Rwandawould help raise the quality o f service inthe tourism industry. Intensive (and extensive) skills development is needed to enable the industry to receive high-end tourists. The existing skills base is far from adequate to meet the standards o f an internationally competitive tourism industry. There i s no established training institute for the hospitality industry in Rwanda, although a private hotel and hospitality training institution is in the final stages o f the approval process. In the meantime, Rwanda must invest in short-term, medium-term, and on-the- job training programs for targeted groups, to address the critical need for skilled human resource inthe sector. In this regard, Government, donor, and other stakeholder interventions are needed to provide skills training, since most tourism enterprises inRwanda are too small to provide a cost-effective training program. Further, research shows that for investment intraining to pay off, there must already be a sufficient skills base. Studies have shown that the vast majority o f firms choose to train workers with at least secondary education, the completion rate of which is only 57 percent. However, the literacy rates are at 71 percent, which i s significantly above the SSA average o f 60 percent. While the introduction in 2006 of a fee-free, nine-year basic education program i s expected to improve the completion rate, the huge unfulfilleddemand for skilled labor in all sectors o f the economy poses a major challenge to improving the quality o f service in the tourism industry. It is, however, worth the effort, because if Rwanda manages to provide high-quality service, it can substantially improve its competitiveness v i s - h i s Kenya, Tanzania, and Uganda, where tourists' satisfaction with the quality satisfaction is generally low The main issue pertains to investing in the relevant infrastructure and building capacity among the various agencies and institutions engaged in the sector. Public-private partnerships are critically needed for infrastructure investment and training facilities. A pool of skilled and well-trained labor i s needed to support the high-end market that the Government is targeting. Given that tourism depends on a variety of sectors to thrive, coordinationwill be needed among a number of government and non-government industries. Lack o f coordination between the Government and stakeholders can result in poor infrastructure development, under-developed tourist sites, shortage of skilled workers, etc. The strength and efficacy ofthe institutional linksare key to the development ofthe sector. Coordination among the various stakeholders can be strengthened through an integrated planning and management framework. That is, the roles of the relevant agencies and organizations should be defined at the planning stage in order to coordinate and direct resources (both human and financial) to achieve specified short, medium and long-term 154 goals. Inthis effort, it would be crucial to ensure the early and continuous involvement of all relevant institutions. 0 Related to the issue o f coordination i s the importance o f developing a strategy for engaging the private sector and improving the business environment. The fact that more than 65 percent o f hotel rooms are concentrated inKigali, and that many o f the main tourist attractions in Rwanda lacking accommodations, i s an indicator o f larger supply constraints that the tourism industry is facing. Inaddition to accommodations, many o f the main tourist sites need upgrading - most parks in Rwanda compare poorly to other competing East African parks, roads to various attractions need to be improved, and airlift capacity for domestic flights needs to be expanded. 6.6 CURRENT SITUATION OFTHEICT SECTOR INRWANDA 6.6.1 Overviewof the ICT Sector 6.114 The Government aims to modernize the Rwandan economy and society using ICT. However, the penetration of ICT inRwanda is at a nascent stage, and the general population does not yet have exposure. The physical infrastructure needed to use ICT in an integrated manner i s not robust, and the penetration o f main telephone lines remains low. The challenge o f expanding the communications network i s amplified by Rwanda's hilly terrain. In the absence o f a reliable and fixed telephone network, the mobile phone is quickly acquiring popularity, as in other developing countries. 6.115 Rwandahas experienced rapid increasesincellular phone subscription, with a cumulative annual growth rate of 66 percent in 1999-2004. The network coverage o f MTN-Rwandacell, the largest mobile service provider in Rwanda, i s estimated at 70 percent. However, the number o f subscribers per 100 inhabitants still remains at 1.6, making Rwanda one o f the lowest cellular subscription countries in the region (Figure 6.25). Internet penetration i s also at a very early stage. The number of internet users i s estimated to be 0.45 per 100 inhabitants in 2004. Again, Rwanda i s trailing behindother countries inthe region. 6.116 Businesseshave not adopted computers on a largescale. While there are signs o f new ICT-based businesses (internet cafks, software development services, broadband services, IT consulting) being established, people's readiness to use ICT effectively remains low, and there are limited opportunities to acquire basic knowledge about how computers work, particularly in rural areas. 6.117 In order for ICT to have a broad impact throughout the economy, the following conditions have to be met: (i) highlyskilledstockofhumancapital; (ii) stronginformatiodtelecommunicationsinfrastructure;and (iii) conducivepolicy,legal,andregulatoryframeworks. 155 Figure 6.25: Cellular mobilesubscribers in2005 (Per IO0 inhabitants) Source: InternationalTelelcommunications Union, http://www.itu.int/home. 6.118 Two key public educational institutions, Kigali Institute o f Science and Technology (KIST) and the National University of Rwanda (NUR), are the most prominent players in developing core groups with IT skills. KIST i s offering advanced technical programs, including computer engineering and information technology. The Department o f Computer Engineering and Information Technology at KIST offers courses from basic computer literacy to diploma and degree-level training programs in computer studies. Student enrollment in KIST i s growing very rapidly - from 209 in 1998 to 3,304 in 2003. NUR is also active in this area. It is equipped with several computer laboratories and requires all students to take basic computer training. NUR offers training programs in network engineering (supported by the Cisco Networking Academy) and database management(supported by Oracle) among other computer-related courses. 6.119 The Government i s taking aggressive steps, too. The Ministry o f Education, Science, Technology and Scientific Research, in partnership with several donors, has launched a program aimed at providing computers and connectivity for all secondary schools in the country. The programaims for an average of 10 computers per school, and internet access for each school with a connection speed averaging 128 kbps. 6.120 Despite such efforts, the need i s so large that the effect will most probably remain limited for some time. For ICT to have a macroeconomicimpact,a large number of ICT usersand a highlysophisticated workforce will have to exist. Inbothdeveloped and developing countries, ICT usage tends to be higher among better-educated populations. Therefore, overall educational level is critically important for ICT to make a large impact. However, the adult literacy rate in Rwanda remains at 75 percent for males and 63 percent for females.*' ICT-specific training i s 80Data source: WorldDevelopment Indicators (2005). 156 important, but the larger challenge of improving the overall education outcome should not be overlooked. The challenge is clear, but it requires strong commitment and time to overcome.81 6.6.2 Current State of ICT inRwanda 6.121 Rwanda currently faces shortages in two key physical infrastructures for an ICT- driven economy: electricity and telecommunications infrastructure. Only 6 percent o f the population has access to the electricity grid (20 percent in urban areas and 2.5 percent in rural areas), and 1 percent of consumers use 56 percent of the available electricity. The current demand on the power grid is currently 45MW, butthe power supply capacity remains at 20MW, including imports from Uganda and the Democratic Republic of the Congo. The overall penetration level o f telecommunications service -main telephone line, mobile subscription, and internet usage - i s also very low. The number o f maintelephone lines per 100 inhabitants in 2004 is 0.27, which i s low compared to both neighboring countries and the average for SSA, which stands at about 2.5. Similarly, as notedabove, cellular phone and internet usage are amongthe lowest inthe region. 6.122 A survey conducted by the National University o f Rwanda (NUR)'* suggests, however, that Rwandans use telecommunications services more than these aggregated statistics may imply. They make use of such facilities as telephone kiosks, the usage o f which is quite high across varying parts o f the country. In rural areas, 55 percent o f respondents said that a household member has used a kiosks during the past three months. In major towns, kiosk usage is 67 percent, and 72 percent inother urban areas. 6.123 The same survey also suggeststhat the major internet access points for citizens are cyber cafes. O f all respondentswho use the internet, more than 50 percent say they use it at a cybercafk, while 30 percent use it at school or work. The price for an individual connection is very high-for a 256kbps line connected to fiber optics, the monthly fee was a reported US$1,250 at the end of 2005,83 while cybercafes and existing community telecenters charge, on average, FRw 400 (approximately $0.71) per There are three commercial ISPs, and KIST and the NUR also provide ISP services usingVSAT (very small aperture terminal satellite) technology. 6.124 The overall picture of ICT infrastructure clearly indicates that Rwanda needs to take aggressive measures to improve the quality, availability, and affordability of telecommunications infrastructure, as well as its electricity supply capability. To this end, the National Information and Communications Infrastructure (NICI) plan puts forward a comprehensive approach to improve infrastructure, access, affordability, and standards; and to strengthen private sector participationand local content. 6.125 There have been a number of positive developments, supported by favorable policies, in recent years. The old telephone and internet Government monopoly, Rwandatel, was sold to Terracorn Communications in 2005. Terracorn is implementing a fiber optic network India has a lower adult literacy rate, but the number of people capable of using ICT is far larger than in Rwanda. Accordingto NASSCOM (1999), Indiahad more than 1,832 educational institutions in 1999, which contributed more than 67,785 trainedcomputer softwareprofessionals per year. put inbibl Joseph(2002) estimatesthe share ofM-techs '*Thesurveytitle?was andPhDholdersto be 3.14and0.14 percent, respectively. conductedbetweenOctober andNovember 2004 as part ofthe ResearchICT Africa! (RIA) Network. 83 Government ofRwanda(year). - move rest to bibl The NICIPlan- The Government of Rwanda, Part 1: Situational Analysis and eReadiness-Release 3.2, pp.31 you never mentionedNICIplan 84 These services are most likely at lower speed. 157 backbone with wireless local loop access to reach out to users in remote areas. It has already installed fiber optics inside Kigali and is extending it outward.85TerraCom has now started to offer ADSL (asymmetric digital subscriber line) service in selected parts of Kigali, at a much lower rate than RwandaTel used to offer. The basic residential package is 128kbps, at a flat rate FRW 45,000 per month (or USD 81) and an initial installation fee FRw 100,000.s6 Further, the Rwanda Internet Exchange (RINEX), created in June 2004, allows internet service providers (ISPs) to exchange domestic internet traffic without having to send data across multipleEuropean and U S hopss7.This will help increase broadband speed inside the country and reduce overall internet costs. 6.126 Efforts must continue toward overall telecommunications sector reform and expansion of informationinfrastructurecoverage. Improvingthe service levels and capability of core infrastructure should be o f the highest priority, to establish a prerequisite foundation for the success ofNICIplans andthe transformationto aknowledge-based economy. 6.6.3 Priority Measures 6.127 The ICT sector, which includes both ICT production and IT-enabled services (such as call centers), can become an important component of Rwanda's development program, particularly in its effort to shift away from an agriculture-based economy. However, developing an ICT sector that can compete ina globalizedmarket i s challenging, and requires significant lead time to build core infrastructure and develop human capacity. Inaddition, supplyingcompetitive products in the globalized ICT market requires a high level of maturity in quality control and innovation. 6.128 For ICT to have a broad-based impact on Rwanda's economy, the highest priority need is to address deficiencies in infrastructureand humancapacity. Aggressive plans to do so are incorporated inthe NICI I1plan. 6.129 Improvements in the quality and availability of e-infrastructure, which poses the most bindingconstraint to growthinthis sector, will needtime. While investmentinICT will form a robust foundation for Rwanda's future, it is probably not realistic to expect, in the short term, a large-scale macroeconomic impact from ICT. However, steps can be initiated now to lay the foundation for higher ICT impact. 6.130 The entry point for Rwanda would be in segments lower in the value chain, such as product assembly for domestic or regional markets. There are some encouraging signs in this area. For example, it has been reported that a Chinese telecom firm specializing inthe production of low-cost handsets has agreed to invest US$2 million to set up the assembling plant at the proposed Kigali ICT Technology Park (techno-park)." The techno-park i s expected to house software development, call centers, hardware assembly, and other ICT-related firms. A number of domestic software service are also developing to meet local needs. 85 It now goes as far as Butare. 86 TerraCom Communications; httr,://mnv.terraeom.rw/. A hop is an intermediateconnectionina string of connectionslinking two network devices. A hop occurs eachtime a data packet is forwardedto the next router. See also webopedia definitionat httr,://www.webor,edia.com/TERM/h/hor,.html. Previously inRwanda, internet exchanges hadto go throughrouters locatedinthe US and/or Europe 88 "Rwanda to Make Mobiles," East African Business Week, May 22, 2006. 158 6.131 In the meantime, aside from the broad-based measures that will be needed in the areas of education and science and technology, Governmentcan begin to explore areas for immediately improving ICT use and literacy among the population. Small-scale local interventions with well-defined objectives can go a long way toward generating quick wins, which will help the population to prepare for projects with larger benefits. Many successful projects in developing countries are implemented in an environment with constraints similar to those in Rwanda. These projects succeed by maintaining a manageable level of technical sophistication, suitable for limited infrastructure and technical capacity. The integration of ICT is typically achieved by defining a clear set of goals and then using ICT as one o f the ways to achieve those goals. For example, applications in rural areas could include transmitting information on prices of agricultural products innearby markets; weather forecasts for fishermen; an interactive Q&A system on agricultural practices; and interactive tools for farmers to engage ine-commerce. Results from field surveys inrural China note the importance of integratingthese withtraditional means of communication. 6.132 ICT can also be used for e-government type projects, to improve transparency, publicservice quality,and accountability. ICT use should not become a goal in itself; rather, it should complement the sector development programs. It should be integrated into program activities as a tool for information dissemination, facilitation o f administrative processes, and other aspects o f development programs. Government would have a strong role to play in promoting these uses and approaches. 159 7. COMMODITY EXPORT DIVERSIFICATIONINRWANDA- MANY EXPORT DISCOVERIES-WITHLITTLESCALING-UP 7.1 INTRODUCTION 7.1 Increased diversification of commodity exports, and increased numbers o f high-value commodity exports, are neededto generate employment and meet Government targets for poverty reduction. This chapter presents evidence that increased diversification of exports is linked to increasedexport and GDP growth. Yet, commodity exports inRwanda are concentrated in a few commodities. Although a greater proportion of the population is engaged in sectors related to commodity exports as opposed to service-related exports, the service sector accounts for close to 50 percent of Rwanda's total exports of goods and services 7.2 Diversificationfrom traditionalto non-traditionalcommodity exports is imperative for Rwanda. The Government i s focused on generating economic growth to raise incomes above the current level of $250 per capita, but it faces some challenges. Exports, a key channel of economic growth, are caught in a low-growth trap. Export growth decelerated to only 9 percent nominal growth per year between 1996 and 2004, from its level between 1965 and 1990 of 20 percent per year.89Worse, excluding 2 or 3 primary commodity-based exports, very few other products have values of up to US$50,000 in any year. Given the limited prospects of simply scaling up Rwanda's leading exports, the country's policymakers have almost no option but to turn to export diversification in the hope of accelerating export growth and economic development. Export diversification i s indispensable for enabling Rwanda to graduate from a primary producer to a modern economy. 7.3 The relationshipbetween export diversificationand export growth in SSA confirms that for stable, sustainable, and higherexport growth, export diversificationis necessary.In addition to export promotion policies such as those which attract FDI, decision makers need to design policies to accelerate diversification o f the export mix. The effect of diversification on export growth will be maximized if diversification shifts the export mix in the direction of manufactured or resource-basedproducts (low or medium technology). This implies (i)sector- specific public strategiesthat can influence the export mix; and (ii) public investments in human capital and supporting infrastructure. 7.4 This chapter finds that in the past five years, East Asia's burgeoning demand has opened up new marketsthat are likely to sustain rapid growth for Rwanda's minerals and metalexports. But, while good for growth, such exports runthe risk o f reinforcing the trend in commodity exports basedon unprocessednatural resources, relegatingRwanda's export basket to products defined by its natural comparative advantage. A structural transformation o f the export basket, therefore, poses some hard questions. Inwhat direction should Rwanda diversify? Which sectors? Should the Government be indifferent to the types of products exported, or should it necessarily support higher-value products? Should the Government prefer sector-specific to general export development strategies? 7.5 In comparison to its neighbors and SSA, Rwanda's export mix is concentrated; the share of exports in the economytoo small; and export growthis low, declining, and volatile. *'Notethat the period 1991-1995 is excludeddue to the deleterious effect ofthe genocide. 160 Given that coffee and minerals (two types) account for 91 percent of its exports, these discouraging characteristics can only be altered if the Government implements a drastically differentpolicy aimed at export diversification. 7.6 The challenge of export diversification in Rwanda is not about the paucity of new export discoveries. Inlight o f its endowments, stage o f development, and geography, Rwanda's export discoveries are many but small. The problem o f export diversification is primarily a lack of scale inproducts other than coffee and minerals. Except for the latter, nearly all exports have miniscule values, and very few are larger than US$lOO,OOO in any given year. Each US$lOOO increment of a non-traditional export is a precious achievement. 7.7 This chapter provides insightsinto the export sector which policymakers may wishto use to inform an export diversification strategy for Rwanda. The chapter begins by examining how much Rwandahas diversified over the past decade. It then examines the degree of sophistication o f Rwanda's exports. Finally, it examines commodity-specific and general constraints to export growth. The chapter concludes with options for reducing the degree of concentration of exports in Rwanda, despite the constraint of limited natural and human capital endowments and the challenge of developingeconomies o f scale. 7.1.1 The Case for Export Diversification-To PromoteGrowth ofExports and GDP and Reduce Volatility 7.8 The relationshipbetweenexport diversificationand export growth for SSA confirms that for stable, sustainable, and higher-export growth, export diversification is necessary. Analysis of the links between diversification (measured by the Herfindahl Index, HI) and growth for SSA indicates that diversification helped SSA to transition from low to relatively higher export growth between the 1980s andthe ~ O O O S . ~ ~ 7.9 In the 1980s, Rwanda, Uganda and Burundihad some of the most concentrated export baskets in their sub-region and, indeed, SSA. Between 1980 and 1989, the share o f coffee, the primary export product inthese three economies, ranged between 70 and 95 percent, leadingto an extremely highHIS,inthe range o f 0.6 and 0.85 (Figure 7.1), and to negative export growth. In comparison, the HIfor SSA was just under 0.4 and the average export growth rate was about 3.4 to 4 percent per year. Between 1995 and 2004,91as the share o f coffee fell to about 50 percent, Rwanda and Uganda closed the gap with SSA - their HISand average export growth rates reachedthe averages for SSA inthe first period (1980-1989). However, by that time, the average for SSA had moved to the southeast quadrant -with even more diversification, although less than Rwanda and Uganda, SSA achieved significantly higher export growth rates o f 8 to 9 percent per year (southeast quadrant inFigure 7.1). 90 The Herfindahl index is used to measure the degree o f diversification. It is computed as the sum of the squared shares of exportsfrom each industry out oftotal exports.. 91 By 1995, most countries inSSA had liberalizedtheir trade, 1980-1989and 1995-2004are selectedas pointsof comparison. 161 Figure7.1: ExportDiversificationCausesExportGrowthin SSA Source: Chandraet al, based on UNCOMTRADE, SITC-23digit. 7.10 I t appears that the gains in export growth from export diversification, in the presenceof trade liberalization,are sensitive to the initial level of export diversification.In the first period, Tanzania and Kenya had far more diversified export basketsthan even SSA. Over the two decades, they enjoyed relatively high export growth in return for smaller changes in the HI. In contrast, weak diversifiers such as Rwanda and Uganda did not achieve higher export growth rates untilthey made significant progress inexport diversification. 7.11 These observationswere tested in a simple modelfor three time periodsfor Rwanda and its five neighbors - Uganda, Burundi, DRC, Tanzania, and Kenya (collectively called the sub-region) that share similar economic characteristics. Inthe fixed-effects model, export growth i s defined as a function of a lagged change inthe Herfindahl,used as a proxy for export diversification; the real exchange rate; and the two main commodity prices that drive exports - the price ofcoffee as a proxy for agricultural commodity prices, andthe price oftinas aproxy for mineral prices. Inaddition, the technology content of exports, proxiedby the share of primary and resource-based products (pprb), indicates the direction o f export diversification. Various measures of trade liberalization - years of trade liberalization experience, economic freedom, freedom to trade internationally, and the level o f trade (trade to GDP) - are included as controls to reflect structural change that affects growth over a longer term. 7.12 The hypothesistested is whether export growthwas caused by export diversification or was simply a manifestation of commodity price shocks. Export diversification did not occur in Rwanda and most of its regional neighbors until after structural adjustment in the mid- 1990s. Prior to this period, export growth was driven primarily by the prices of the primary commodities - coffee and minerals - that dominated exports. High commodity prices led to export concentration. To determine whether non-traditional products contributed to export growth, the effect o f commodity prices was netted out to obtain a non-price Herfindahl. The findings, indicatedby the negative sign on the lagged change in the HI, show that over the long term (1975-2004), non-commodity price-related export diversification caused export growth in the sub-region, including Rwanda.92 While small, growth in primary and resource-based products93raised export growth. Diversification toward more sophisticated products with a higher 92Note, anegativesign onthe laggedchange inthe HIdenotes an increaseinexport diversification, 93 A list of products definedby their technology content - primary, resource-based,low tech, medium tech andhigh tech is presentedinAnnex 3 andwill be discussedinalater section. 162 manufactured content was low and emerged slowly. For example, in Rwanda, diversification occurred from coffee to tin and other minerals, while in Uganda it was from coffee to fish. The effect o f trade liberalization was insignificant, perhaps because the challenges o f globalization have made it difficult for African exporters of primary commodities to compete in non-primary products. These findings are robust and indicate that in addition to the conventional export growth-promoting policies, special policies that spur export diversification are necessary. 7.13 Export diversificationand growth were unrelated in the pre-structural adjustment period, 1975-1995, when coffee was the dominant export in Rwanda and the sub-region. Closed trading regimes, ad hoc curbs on economic freedom, the genocide in Rwanda, and unstable socio-political conditions in several countries disrupted economic development and stymied diversification. Even though most countries had liberalized by the late 1980s, the implementation of reforms did not begin until the early 1990s. The post- structural adjustment period, 1996-2004, however, has witnessed a remarkable turnaround. Export diversification has contributed significantly to export growth, although the shift from primary toward more manufactured exports, coveted by governments for greater economic development, has been weak. These findings are robust across avariety o f specifications (Table 7.1). 7.14 A common beliefisthat special policies are unnecessaryand a stable macroeconomic and liberal trading environment that attracts FDIi s sufficient for export diversification. Undoubtedly, FDIhas playedan important role inthe region indeveloping exports based on the latter's natural comparative advantage, especially in the mining industry. For example, FDIflows facilitated the boom in Tanzania's gold exports, which accelerated from a share of nearly zero in 2000 to more than 15 percent in 2005, and were Tanzania's largest export product. Similarly, FDI has been critical for Rwanda's mineral exports. But the crucial point i s that without special export diversification-facilitating incentives, FDIhas not flowed into non-traditional, high-value export sectorsthat are necessary to reduce dependence on primary products. Special incentives to attract FDIto Uganda's non-traditional ex orts, such as uplandroses, fishfillets, organic cotton, textiles and garments, and animal product^!^ is also a good example. Table 7.1 shows the export growth model for Rwanda, Burundi,DCR, Kenya, Tanzania, andUganda. 94See Chandra and Boccardo (2006) on export diversification and competitivenessinUganda. 163 Table 7.1: Fixed Effects- Export Growth Model for the Sub-Region, 1975-2004 (Rwanda, Burundi, Dem.Rep.of Congo, Kenya, Tanzania andUganda) (Dependent variable: Export growth) 1975-04 1975-95 1996-04 1975-04 1975-95 1996-04 Years oftrade liberalizationexperience 0.010 0.035 0.010 Reergr -0.350*** -0.340*** -0.330* -0.340*** -0.330*** -0.260 Laggedchange inthe HIindexofnon-coffee price relatedexports -0.690*** -0,110 -0.990*** -0.670*** -0.020 -0.980*** shareofprimary andresourcebasedproducts intotalexports 0.000 o.ooo* 0.000 o*ooo** o.ooo** 0.000 Constant -0.020 -0.020 0.020 -0.020 -0.040 -0.060 Fraserfreedomto trade index for trade relatedvariablesonly - 0.030** 0.070** -0.110*** Fraserfreedomto trade indexfor all variablesassociated with aliveraltradingenvironment 0.030 0.190*** -0.190*** Reergr -0,340*** -0.350*** -0.620*** -0.340*** -0.350*** -0.360* Laggedchange inthe HIindexofnon-coffeepricerelatedexports -0.680*** -0.580 -0.950*** -0.670*** -0.027 -1.300*** Shareofprimary andresourcebasedproducts intotalexports o.ooo* o.ooo*** 0.000 o.ooo** o.ooo* -0.002 Constant -0.130 -0.880*** 1.100*** -0.150* -0.310** 0.650*** Ratioof importsto GDP 0.000 -0.004 0.007 Ratiooftradeto GDP 0.003 0.000 0.006 Reergr -0.310*** -0.290*** -0.350" -0.300*** -0.280** -0.340** Laggedchange inthe HIindexofnon-coffeeprice relatedexports -0.700*** -0.040 -1.130*** -0.690*** -0.010 -1.1lo*** Shareofprimary andresourcebasedproducts intotalexports O.OOO*** O.OOO* 0.000 0.000 0.000 0.000 Source: Chandraet. al. (2006), usingUNComtrade, SITC2-3 digit and World DevelopmentIndicatorsdatabases Observations: 134. Significance: ***1-5%;**5+ - 1O%;* IO+-1 5%. 7.15 Diversification from products that are more vulnerable to terms of trade shocks to products with relatively stable prices dampens the volatility of export growth. Rwanda's export growth was most volatile between the 1970s and 1990s when coffee accounted for nearly all of its exports. However, in the past 5 years, the emergence o f minerals has accelerated diversification. Booming minerals prices have increased their share to over 40 percent of total exports. It is unclear for how long these offsetting prices will sustain stable growth. Typically, export diversification from one primary commodity to another rarely dampens export price volatility. 7.16 The nexus between export diversificationand volatility of export growth is mapped in Figure 7.2. It illustrates that relative to the 1980s, in the past decade, Kenya, Tanzania and Uganda were the only countries that benefitedfrom a reduction inexport growth volatility as they diversified their exports. Most o f the benefits ensued from diversification into more high value horticultural, fishery products and low tech manufactures such as garments that have stable prices. While Rwanda also underwent impressive diversification in the past five years, over the past 10 year period or the longer term, the effect o f fluctuating coffee prices continued to worsen the volatility of overall export growth. 7.17 Looking ahead, the quality of diversificationwill matter for reducing the volatility of Rwanda's export growth. This will require scaling up exports o f more processed and high value agriculturalproducts, but especially manufactures. 164 Figure 7.2: Export Diversification Dampens the Volatility of Export Growth Source: Chandra et al, basedon UNCOMTRADE. 7.18 The set of six countries Rwanda, Burundi, Democratic Republic o f Congo, Kenya, Tanzania and Uganda - - is used to examine the pattern of export diversification. Over the years 1980 to 2004, two time periods are examined - 1980-1989 and 1995-2004. In the fixed- effects model, the pattern o f export diversification, again proxied by the Herfindah1 Index, is expressed as a function o f the technological sophistication o f the exported commodities. This is represented by primary products comprised of mostly unprocessed agricultural products (or primary products); resource-based products used as a proxy for good infrastructure and manufactured products denoted by low tech products (It) used as a proxy for superior human capital. The same measures of liberalization are used to delineate how the technological content o f exports changed. 7.19 The hypothesis tested is that an increase in primary or resource based products led to greater diversification(lower HT),as both indicatehorizontaland vertical diversification. An increase in low technology products would imply that an increase in skills accelerates diversification. Previously, it was noted that export growth in SSA has been driven by diversification and liberalization. The current model examines whether diversification was helped or hinderedby liberalization. Results are reported in(Table 7.2). 165 Table 7.2: FixedEffectModel-Pattern of Export DiversificationinAfrican Countries, 1975-2004 (Dependent variable is the log ofthe Herfindahl Index) 1975-2004 1975-1995 19962004 Recon SSA Region SSA Region SSA Denvariable: Loe HHI Log ofprimaryproductsas ashare o ftotal exports 0.302*** -0.077*** 0.316*** -0.034 0.074 -0.082*** Logofnaturalresource-basedproductsas ashare oftotal exports -0.116*** -0.157*** -0.076*** -0.117*** -0.163*** -0.148*** Log oflow-techproducts as ashare oftotal exports -0.056*** -0.019*** 0.007 0.005 -0.105*** -0.161*** Logof GDPper capita -0.860*** -0.111*** -0.579*** -0.138*** -1.790*** -0.278*** Constant 2.862*** -1.325*** 1.956*** -0.872*** 7.036*** -0.951 Fraserkeedomto trade index for trade-relatedvariablesonly - -0.058*** -0.070*** 0.006 -0.078*** -0.047 0.028 Logo fprimaryproductsas ashare oftotal exports 0.235*** -0.052** 0.324*** -0.029 0.08 -0.027 Log ofnaturalresource-basedproductsas ashare oftotal exports -0.097*** -0.144*** -0.076*** -0.105*** -0.161*** -0.137*** Log oflow-techproductsas a share oftotalexports -0.042** -0.015 0.006 0.005 -0.103*** -0.113*** Log ofGDPper capita -0.815*** -0.396*** -0.576*** -0.286*** -1.665*** -1.045*** Constant 3.007*** 0.756** 1.909** 0.421 6.659*** 3.480*** Fraserfieedomto tradeindex for all variablesassociatedwith aliberal -0.073*** -0.102*** 0.065 -0.116*** -0.176** -0.015 Log ofprimaryproductsas ashare oftotal exports 0.219*** -0.045* 0.361*** -0.022 0.012 -0.025 Log ofnaturalresource-basedproductsas ashareoftotalexports -O.lOO*** -0.134*** -0.085*** -0.091*** -0.161*** -0.132*** Log oflow-techproductsas ashare oftotal exports -0.058*** -0.025*** 0.008 -0.002 -0.101*** -0.112*** Log ofGDPper capita -0.714*** -0.363*** -0.687*** -0.164 -1.538*** -0.957*** Constant 2.438*** 0.697** 2.227"' -0.09 6.614*** 3.244*** Freedomheritage -0.481*** -0.240*** Log ofprimaryproducts as ashare oftotal exports 0.173* -O.IOl** Log o fnaturalresource-basedproductsas ashare oftotal exports -0.081 -0.129*** Log oflow-techproductsas a share oftotalexports -0.158*** -0.172*** Log of GDP per capita -2.807*** -0.823*** Constant 14.552*** 3.007*** Years oftrade liberalizationexperience -0.048*** -0.001 -0.013 -0.022* -0.054*** 0.01 Log ofprimaryproductsas ashare oftotal exports 0.069 -0.039 0.304*** 0.005 -0,052 -0.07 Log ofnaturalresource-basedproductsas a share oftotal exports -0.080*** -0.175*** -0.068*** -0.124*** -0.168*** -0.250*** Log oflow-techproductsas ashare oftotal exports -0.038** -0.021* 0.008 0.01 -0.042 -0.179*** Log ofGDP per capita -0.505*** -0.118* -0.547*** -0.076 -1.211*** -0.239 Constant 1.145*** -1.350*** 1.818** -1.191** 4.399111 -1.667 Ratio oftrade to GDP -0.001 -0.002*** -0,001 -0.003*** -0.005 -0.001 Log ofprimaryproductsas ashare oftotalexports 0.314*** -0.079*** 0.295*** -0.035 0.197*** -0.088* Log ofnaturalresource-basedproductsas ashare oftotal exports -0.115*** -0.154*** -0.077*** -0.115*** -0.131*** -0.148*** Log oflow-techproducts as ashare oftotal exports -0.053*** -0.018" 0.008 0.009 -0.087*** -0.151*** Log ofGDP per capita . . -0.942*** -0.128** -0.569*** -0.078 -2.118*** -0.504** lconstant 3.403*** -1.129*** 1.936** -1.021** 9.305*** 0.419 Source: Vandana, et. al. 2006,. usingUNComtradeandWorld DevelopmentIndicatorsDatabases. - Consistentwith FEmodelfor 6 countriesusingphonesper thousand ai proxy for infrastructureandratio of servicesto GDP as proxy o fdevelopment(see Annex). Observations: 134. Significance:***1-5%;**5+ -10%;*10+-15%. 7.20 The results indicatethat for all three periods, in SSA and the sub-region there was an unambiguous increasein natural resource-based products such as fish, wood, processed food, processed minerals, livestock and related animal products, and that this led to more export diversification.This may have been becausethere is greater scope for moving from raw to processed products than from primary to processed products; Le., it is easier to export frozen rather than fresh fish, as opposed to exporting roast coffee from unwashed beans. Exports that require a higher level of human capital, such as manufactures, supported diversification, though not during 1980-1989, when the economies were closed and vulnerable to domestic shocks. An increase in exports of pp supported diversification in SSA, but worsened it in Rwanda and the sub-region, where exports were already over-concentrated in coffee. 7.21 Unambiguously, liberalization affected export diversification positively. Over the longer term, the freedom to trade indices and trade liberalization contributedpositively (shown by 166 the decline inthe Herfindahl)to export diversification. Concentration inprimary products usually worsens diversification, but a shift towards low tech and resource-based products reinforced it. Much of the latter may have been a coincidence, due more to rapidly growing demand for minerals by China, which boomed after the mid-1990s. The effects o f emerging low-tech products, however small, occurred only in the post-structural period. A cross-section model for Rwanda and the sub-region also showed that land-lockedness was an important constraint to diversification. 7.22 The analysis suggests that there are at least two pointers that can be useful in developing manufactured (low technology) and resource-based exports. These are the development of skills necessary for manufacturing, and the development o f infrastructure to offset the land-lockedness as a disadvantage faced by Rwandan exporters. A further examination of the level of diversification in Rwanda, with a focus on specific commodities, demonstrates the importance ofthese factors. 7.2 HowMUCH RWANDA HAS DIVERSIFIED? 7.23 Rwanda is far less diversified compared to SSA and five Pigure 7.3) neighboring countries that are quite similar to it in topography, climate, soil, and natural resources. In fact, at least two - Uganda and Burundi - even share its geographic disadvantage o f being landlocked. Since the 1960s, the trends in Rwanda's pattern o f export diversification have tracked a course similar to that o f Uganda and Burundi. The Herfindahl Indices in these three countries movedtogether and ranged between 0.4 and 0.9, reflecting strongly concentrated export baskets that were dominated by coffee and ores (see Table 7.3 and Figure 7.5). In contrast, Tanzania, the only coastal country in the set and a coffee exporter, consistently enjoyed a well- diversifiedexport basket, withthe HIinthe range o f 0.3 to 0.1 (Table 7.3). 7.24 Typically, export diversification is a longer term process and Rwanda's case is a good example.A variety of export diversification measuresjointly show that it has taken almost 35 years for Rwanda's HIto decline from around 0.5 during 1970-75 to 0.35 in 2004 (Table 7.3). Between the 1980s and 2000, the number o f exported products valued at $10,000 or more in Rwandawas only 37 inthe 1970s; by 2000, it had risen to 73; and the share o f the top 5 products inRwanda's total exports declined from 93 to 89 percent. Inthe last five years (2000-2004), there was a slight reversal. The number of products exported declined and the share o f the top 5 exports increased, pointing toward a potentially rising trend in export concentration. In comparison, Rwanda's neighbors were more successful in diversifying their exports. Uganda and DRC were as diversified in 1970 as Rwandawas in2000. 167 Figure 7.3: Rwanda and Its LandlockedRegionalNeighbors Are Far Less Diversified than CoastalTanzania Source: Chandra et a1226; and UNComtrade. 7.25 The low levelof export diversificationinRwanda and its neighborscan be explained by several factors, including sociopolitical events that eroded the export base in several countries, leavinglittle else than natural vegetationon the ground. Examples include Uganda duringthe Idi Amin era (1971-1985), Rwandaduringthe genocide o f 1994, and Tanzania during the Nyerere's socialist regime. The annihilation o f the productive sectors and human capital duringthese conflicts limitedexported products to those which survived the shocks. Coffee that grew naturally or minerals that did not need new investments in Rwanda and DRC are good examples. Other reasons for weak diversification are economic policies and institutions that left the export mixg5to be determinedby natural comparative advantage. Most ofthe trends inthe HI also indicate a structural shift around 1995, a time when economic reforms in Rwanda and its neighbors initiatedexport diversification. 95InUgandafor instance, the IdiAmin era ledto massdestructionofthe economic base, causingexport concentration incoffee, the cropthat growsnaturally inUganda. 168 Table 7.3: DifferentMeasuresof Export Diversificationin Six Countries 1980-2004 (5 YearAverage) No. of products Avg Herfindahl Sh. Oftop Sh. Oftop Sh. Oftop Country Name Year exported Index 5 years - 5 products 10products20 products (> 10,000 U%) Burundi 1970 23 0.64 0.95 0.99 1.oo 1980 27 0.62 0.94 0.99 1.oo 1985 30 0.65 0.97 0.99 1.oo 1990 44 0.72 0.91 0.95 0.99 1995 42 0.62 0.96 0.98 1.oo 2000 40 0.63 0.96 0.97 0.99 2004 30 0.42 0.93 0.97 0.99 Congo, Dem. Rep. 1970 57 0.36 0.89 0.96 0.99 1980 79 0.27 0.90 0.95 0.99 1985 70 0.28 0.87 0.95 0.99 1990 86 0.32 0.90 0.97 0.99 1995 84 0.32 0.89 0.96 0.99 2000 82 0.49 0.94 0.98 0.99 2004 77 0.49 0.93 0.97 0.99 Kenya 1970 122 0.14 0.61 0.74 0.85 1980 125 0.23 0.73 0.85 0.94 1985 126 0.21 0.80 0.88 0.94 1990 128 0.20 0.69 0.83 0.91 1995 147 0.13 0.59 0.73 0.84 2000 143 0.11 0.62 0.74 0.84 2004 154 0.10 0.60 0.74 0.84 Rwanda 1970 11 0.50 0.97 0.99 1.oo 1975 21 0.43 0.97 0.99 1.oo 1980 27 0.57 0.98 1.oo 1.oo 1985 41 0.59 0.96 0.99 1.oo 1990 37 0.60 0.95 0.99 1.oo 1995 42 0.49 0.90 0.96 0.98 2000 51 0.45 0.91 0.95 0.97 2004 46 0.35 0.94 0.96 0.99 Tanzania 1970 81 0.10 0.68 0.82 0.96 1980 75 0.16 0.66 0.84 0.96 1985 69 0.18 0.74 0.87 0.95 1990 99 0.16 0.56 0.72 0.89 1995 115 0.10 0.64 0.78 0.91 2000 123 0.10 0.61 0.78 0.89 2004 129 0.08 0.49 0.72 0.87 Uganda 1970 57 0.38 0.95 0.98 0.99 1980 49 0.81 0.99 0.99 1.oo 1.oo 1985 38 0.85 0.99 0.99 1990 38 0.79 0.95 0.99 1.oo 1995 69 0.59 0.96 0.98 0.99 2000 84 0.52 0.88 0.94 0.98 2004 115 0.19 0.82 0.91 0.96 Source: Staff estimates ,COMTRADESITC 1. 7.26 The challenge for Rwanda lies intransforming its natural resource base, discovering high-value products, and scaling up existing (especially nascent) exports. Lake Victoria's vast fishery resources favored export diversification in Kenya, Tanzania and Uganda, but not in Rwanda. A sizeable fishing industry, good soil, and climate conditions suitable for floriculture and horticulture, along with policy interventions such as special support for Uganda's fishing 169 industry, and Kenya's floricultural exports, helpedto lower the shares o fthe top three products in total exports.96However, landlockedBurundi,DRC and Rwandahave not hadthe same fortune. 7.3 EXPORT SOPHISTICATIONINRWANDALOWER I S THANITSNEIGHBORS 7.27 Compared to exporters in its neighborhood and SSA, Rwanda's export sophistication ranking is low (Figure 7.2). The Rodrik-Hausman export sophi~tication~~ (EXPY) scores reflect the technological sophistication o f a country's exports when compared with its global competitors. A map of export sophistication scores shows that between 1990 and 2004, Rwanda's position was located mostly in the southeast quadrant of Figure 7.4 which indicates countries with the highest export concentration and lowest sophistication. By 2004, there was some diversification but almost no increase in its EXPY. This reflected the low sophistication o f its exports due to the concentration o f coffee and minerals. 7.28 There was some progressin export diversification between 1990 and 2004; however, in the last five years, Rwanda's export sophistication score slipped to below the SSA average. Within the region, even Uganda whose exports were once significantly more concentrated, has made impressive strides in diversifLing its exports and achieving a higher sophisticated (EXPY) content by shifting from the southeast quadrant to the northwest. In fact, during2000-2004, Uganda's export sophistication score was comparable to that o f Tanzania, its coastal neighbor. In comparison, in 2000-2004, Rwanda's score was about US$1,500 well below SSA's average o fUS$2,500 in 1990-1995. 96 InUganda's case, opennessto FDI, as well as collectiveeffortsby the donors, the World Bank, andthe Government of Ugandato establish and implementcompliance with phytosanitaq standards, facilitated scaling up of the country's world-class fish processing industry. In Kenya, the floriculture industry earned political favors through benign neglect when it was excluded from the regulationand Kenyanizationthat covered almost all other productive sectors. Flower exports were noticed by Kenya's political leaders because of their contribution to foreign exchange earnings; their profitability was prized by many well-connectedKenyanbureaucratswho had private businesses; and because of the difficulty in taxing non-durableexports. Initial political commitment and favoritism came disguised in the form of benignneglect of the sector, so that it could grow at arapidpace as opposedto becominganother stagnating economic sector. 97 The Rodrik-Hausmanexport diversificationindexcompares the level of sophisticationor the incomelevel implicit in a country's exports. InRwanda's case, its coffee exports, weighted by its GDP per capita and share of coffee in total exports and in world exports of coffee, are weighted by similar statistics from Brazil, Kenya, Tanzania, Ethiopia, Vietnam, and other coffee exporters. Its mineralexports are weightedby similar statisticsfrom mineralsexporters. 170 Figure 7.4: Rodrik-Hausman Score of Export Sophistication Rwanda - 7.29 Rwanda' export position is disappointing, but its vast natural endowments are largely underexploited, suggesting an abundant potential for catching up with its regional neighbors. One option is to follow in the footsteps o f its more prosperous neighbors; Le., gradually move away from over-dependence on minerals and coffee and toward agricultural exports. The share of such products in the more diversified export baskets o f Tanzania, Uganda, and Kenya ranges from 69 to 93 percent, compared to 54 percent in Rwanda (Figure 7.5). This may be a limited option, however, given that Rwandahas limitedarable land as a result of one of the highestpopulationdensities in SSA.The other option is to exploit more ofits naturalresource base, which impliesthat in addition to high-value crops, it needs to produce more mineral-based and chemical-based higher-value manufactured products. For either option, the crux of the diversification challenge is a larger number o f products, higher-value non-traditional products, and more manufactured and processed products, as opposed to the raw coffee and minerals that presentlydominate Rwanda's export basket. 171 Figure7.5: The Structure of Exports Rwanda and NeighboringCountries, - 1960sto 2000s (In unit as indicated) Burundi I Congo,Dem.Rep 10 LO LO LO 0.9 0.9 0.9 0.9 0.8 0.8 gg 0.8 0.8 0.7 0.7 8 0.7 0.7 g 0.6 0.6 5 g 0.6 0.6 fi 0.5 0.5 0.5 0.5 5 0.4 5g -d2 0.4 0.4 E 2 0.4 0.3 0.3 0.3 0.3 s 0.2 0.2 =b x 0.2 0.2 =b 0.1 0.1 0.1 0.1 0 .o 0.0 0.0 0.0 1960s 1970s 1980s 1990s 2000s 1960s 1970s 1980s 1990s 2000s I...............IAgricukure -Minerals r.........l Agriculture DMinerals Manufacturing +HerfmdahlIndex ----Hefmdahl Index Kenya Rwanda LO LO LO 1.0 0.9 0.9 0.8 0.8 0.8 0.8 VI E I-g 0 fi g 0.7 0.7 2 g wJ 0.6 0.6 x L ii0.6 0.6 0.5 0.5 5 0.4 0.4 u 6 - 0.4 V 0.4 & I- x 0.3 0.3 0.2 0.2 =z 25x 0.2 0.2 =b 0.1 0.1 0.0 0.0 0.0 0.o 1960s 1970s 1980s 1990s 2000s 1960s 1970s 1980s W90s 2000s Agriculture BMinerals -Agriculture BMinerals Manufacturing +Herfmdahl Index Manufacturing d H e r f m d a h l h d e x Tanzania 1.o 1.o Uganda I 1.0 1 .o - 0.8 0.8 E 0.8 0.8 0 e u $0.6 0.6 fi w9 w{ 8 0.6 0.6; L 4 0.4 0.4 g 0 0.4 0.48 c x 0.2 + 0.2 =b x 0.2 0.2* 0.0 0.0 0.0 0.0 1960s 1970s 1980s 1990s 2000s 1960s 1970s 1980s 1990s 2000s Agriculture -Minerals I.....T.:1Agricukure BMinerals Manufacturing +Herfmdahl Index Manufacturing +Herfindah1 Index Source: UNCOMTRADE. 7.4 EVOLUTIONOFEXPORTSTHEEMERGENCE MINERAL - OF EXPORTS 7.30 In the past five years, the emergence of mineral products in Rwanda has helped to increase export diversification in Rwanda. The predominance of coffee in its export mix has diminisheddramatically as the share of minerals, metals, and chemicals has increased - in 2003- 2004, the shares of these two main product lines in total exports were each about 40 percent. Global demand was a key driver o f the diversification, signaling a combination o f sensitivity to global forces and a diversification toward new export markets. 172 7.31 Export diversification from coffee to minerals mattered in two ways: it dampened the vulnerability of exports to coffee price shocks; and it introduced a new source of growth for at least the next decade, or certainly for as long as China continues to grow at double digit rates. Diversification toward minerals in Rwanda has occurred alongside diversification within the minerals category. Exports o f several nascent minerals have emerged in the past few years, mostly in response to the forces o f globalization, dominated by China. The rapid increase in demand from China's burgeoning industrial sectors boosted the demand for non-ferrous ores, which were previously imported in small quantities by the European Union and the US. Tin ores, imported almost exclusively by Malaysia in the past, got a boost with surging demand from Malaysia, Thailand, Singapore, and China. Demand for tungsten, which earlier came from only the US, increasedrapidly as Germany,the Netherlands, and Chinajoinedthe list of importers 7.32 Inadditionto coffee and minerals,at a lower98levelof disaggregationin2003-2004, Rwanda exported about 18 products that had a share of at least 0.05 percent in its export basket. Including miniscule products, the total share of manufactures was only 6 percent in 2003-2004. Albeit small, this development represents a 50 percent increase in manufactured exports compared to the 1990s. Excluding minerals, Rwanda's non-traditional exports reflect the diversityof its rich naturalresourcebase with potential for processing inseveral directions. 7.33 InRwanda's export basket, there were many low-volume, high-valueproductssuch as vegetables and fruits, flowers, tubers and foliage, and other materials of vegetable origin (seeds, oils, and so on). Examples o f crude or unprocessed commodity exports are various types of raw animal skins; minerals, metals and chemicals; and wood and forestry products such as gums and resins. Rwanda's export mix also contained miniscule amounts of processed or manufactured products, including chilled vegetables, dairy products(cream and preserved milk), beer, dried and salted animal hides and processed skins, mineral and chemical fertilizers, polyethylene, cement, leather, paper and products, furniture, yarn, alloys transformed into sheets and rolled metal plates, fabrics and so on. Unmistakably, the defining characteristic o f the non- coffee, tea, and minerals export mix was a lot of variety and numerous products o f minuscule weight. 7.5 EXPORTDISCOVERIES -MANY, WITHLITTLESCALINGUP 7.34 In2004, excludingcoffee, tea, and tin, Rwanda exported at least 180non-traditional commodities, which accounted for roughly less than 9 percent of its exports (Annex Table 12). Export discoveries were examined over two periods: 1976-1989 and 1995-2004. Products with near negligible values were excluded. A product was classified as an export discovery if its value doubled between the two periods.99 This methodology shows that there were only about 33-35 export discoveries of critical mass (Table 7.4). While no single product emerged with a big bang, all added miniscule but increasing values to non-traditional exports. As Rwanda's total exports were only US$97 million in 2004, every export discovery, even if as small as US$lOOO, was a significant inrelative terms. 98At SITC2-3 digits, the number ofproductsexportedwas 18. At afiner level, i.e., SITC2-4 digits, the number was 170. 99This definition of export discovery has been designed for small countries in Sub-Saharan Africa. Klinger and Ledermen(2006) use acut-off point of US$lO,OOO to identify export discoveries. 173 7.5.1 TechnologicalContent of ExportDiscoveries 7.35 A useful indicator of the sophistication or quality of an exported product is its technology content.'" Accordingly, products are labeled as follows: processed or unprocessed products with a high value-to-weight ratio are high value (HV);raw products are primary (PP); minerals and metals are resource based (RB); simple manufactures are low technology; and processed minerals and wood are medium technology. An application of this definition to Rwandan exports reveals that a variety of high value, resource based, low technology, and some medium technology products have recently emerged inRwanda's export basket (Figure 7.6). The hightechnology class of exports comprises fruits, vegetables, flowers, and dairy products. Low technology products comprise processed animal skins, leather, and animal oils; and medium technology products comprise pulp and related products, chemicals, fertilizers, and the like. Rwanda exports a large variety of primary products across various agricultural categories, as well as a fast growing range of resource based products such as processed and raw minerals (Table 7.4). 7.36 There have been many export discoveries, but at a low volume and scale."' 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 US$5000 to US$75,000 between 1976-1989 and 1995-2004. Inthe case o f hides and skins, which are locally produced for leather exports, the export value increased from US$23,000 in 1976-1989 to US$166,000 in 1995-2004, illustrating success in export diversification along the supply chain, from raw hides and skins to a manufactured export. The textiles chain represents the most sophisticated discovery, involving cross-sectoral synergies. An example i s the use o f chemicals to produce fabrics made from yarn and imported cotton, and exports of light technology-based garments and outerwear made from locally available materials. Some o f these product lines have exports inexcesso fUS$lOO,OOO. Figure7.6: Sector and TechnologicalComposition ofExports Sector Composition TechnoloeicalComDosition 100 80 60 40 20 0 1980-1984 19851989 1990-1994 19951999 2000-2004 19801984 19851989 199G1994 19951999 2wO-2004 Coffee 0 Non-coffeeagriculture R Primary Products (PP) High Value PP 0 Textiles E l Fuel 0 Minerals Manufactures R Resource Based LowTech Medium Tech FJ High Tech Note: PP =primary; RE3 =resource based; LT = low tech; M T =mediumtech; HT =hightech. Source:Chandra et al, 2006; and UNCOMTRADE (Rwanda mirror data). 7.37 The processingof mineral and metal-relatedcommodities into medium-technology manufactures (metal scrap, salts, anthracites, aluminum alloys) has contributed to the looLa11(2001). 101Export discoveries are defined here as products whose export value doubledbetween the two periods studied, 1976- 1989and 1995-2004. Productswith near negligible values were excluded, giventhe small size of Rwanda's economy. 174 increasein export values of non-ferrousores and concentrates, from US$2.5 million during the 1976-1989 periodto US$12.1millionin 1995-2004. The large variety o f export discoveries in resource-basedminerals and metals suggests that there is potential inthis area that has been largely unexploited. The emergence o f new minerals such as coltan, used in the manufacture of cellular phones, can be a good source of growth in the short to medium term. Other natural resourceswhose exports exceededUS$lOO,OOO are anthracite and aluminum and its alloys. 7.38 The small scale of discoveries highlights two crucial points: (i)the problem of commodity export diversification in Rwanda i s scaling up miniscule discoveries. In2004, even though 2 traditional products accounted for more than 51 percent of total exports, the large number of non-traditional export discoveries indicatedthat the constraint was not the lack of new and viable products, but their small values. Of the remaining 49 percent of exports which were non-traditional exports, tin ores and concentrates and non-ferrous ores accounted for 40 percent while all other products summed to only 9 percent. (ii) Every export discovery, even ifrelatively small inscale, was important. Table 7.4: Growing Exports and New Discoveries in Rwanda, 1976-2004 (In unit indicated) Techcode ProductsName Number of Transactions AverageTransactions (US$) 1976-85 1986-89 1995-99 2000-04 1976-89 1995-04 Manufacturing Products LT Tarpaulins,sails,awnings,sunblinds, tents, etc. 0 0 3 4 - 75,828 RB Portlandcement, cimentfondu, slagcement, etc. 1 0 5 5 5,571 134,445 LT Containers, of glass, usedfor conveyanceor packing 0 0 1 2 - 63,219 Wood and Paper RB Manufacturesof wood for domestic/decorativeuse 10 4 5 5 12,795 45,190 RB Manufacturedarticles of wood,n.e.s 0 2 2 3 952 29,848 LT Registers,exercisebooks, notebooks 1 1 1 3 3,520 34,232 LT Art. commonlyusedfor Dom.purposes, pot scourers 3 0 2 3 4,461 30,130 LT Chairs andother seats andparts 0 2 1 5 2,373 15,955 LT Otherfurniture and parts 5 2 3 3 17,924 96,738 LT Picturepostcards,greetingcards 5 2 5 4 2,647 10,454 Textiles LT Leather of other hidesor skins 3 1 2 2 23,311 166,516 LT Yam contain.85%by wgt.of synth.fi 2 0 4 4 5,484 17,264 LT Yam of regeneratedfibres, notfor retail sale 1 1 3 2 14,707 31,975 MT Fabrics, wovenof continuous synth. textile materials 1 0 2 3 4,911 11,101 MT Fabrics, wovencontain.85%ofdiscontin. synth.Fiber. 0 2 2 2 4,569 14,953 MT Fabrics, woven,of discontinuoussyntheticfibers 0 0 4 4 - 278,461 LT Blousesof textilefabrics 0 0 1 3 - 7,767 LT Other outer garments of textilefabrics 2 0 4 4 4,814 79,926 LT Otherouter garments & clothing, knitted 2 1 2 2 4,966 14,968 Source:Chandra et a1(2006) based on UNCOMTRADE. 7.5.2 MarketFailure is a Key FactorConstrainingScalingUp 7.39 In addition to the issue of the scale of new discoveries, some existing non-traditional exports seem to be on a declining trend in Rwanda's export basket (Table 7.5). Reversingthe definition that we used for discoveries shows that declining exports are concentrated in three 175 product categories: (i) agricultural high-value exports whose values plummetedfrom nearly US$1 million in the 1976-1994 period to only about US$667,000 during 1995-2004 in the case of floriculture exports; and from US$1.6 million in 1976-1994 to US$40,000'02 during 1995-2004 in the case of plant seeds; (ii) products such as goat skins andrelatedproducts; and(iii) animal low- tech manufactures of footwear and travel bags from leather. 7.40 The discussion in Chapter 3 on the constraintsto growth, coupled with analysis of growth prospectsfor the agriculture sector, suggests that the observed declining trends are relatedto the need for improvedtraining in handling, standards, and quality. For example, World Bank (2006a, p. 32) reports that "in the case o f hides and skins, poor post-slaughter treatment attributable to a lack of specialized equipment, low levels o f human capacity, and deficient coordination along the value chain, results in severe degradation o f hides and skins, which then often receive significant quality discounts in international markets. Hides and skins exports currently amount to approximately 2,000 tons per year, more than 90 percent o f which is low-value dry leather and less than 10 percent i s higher-valuewet blue leather.. .." Table 7.5: Diminishingand DisappearingExports in Rwanda, 1976-2004 (In unit indicated) Techcode ProductName NumberofTransactions AverageTransactions(US$) 1976-85 1986-89 1995-99 2000-04 1976-89 1995-04 AgriculturalProductsand Food HV Other fieshor chilledvegetables (excludingpotatoes,tomatoesandleguminous) 9 4 5 2 339,171 8,636 HV Fruit, fieshor dried. 7 4 5 5 38,062 18,244 PP Pepper ;pimento 10 4 1 1 70,582 16,503 PP Goatti kid skins, raw(fiesh, salted, dried,pickled 9 4 5 5 2,609,895 387,618 PP Bones,horns, ivory,hooves,claws, Cora 8 4 2 0 65,731 17,625 PP Plants,seeds,hit usedinperfumer 8 4 3 2 1,644,796 40,209 HV Bulbs, tubers& rhizomesofflowering or foliage/Cut flowersand foliage 9 4 5 5 1,000,770 666,831 Textiles LT Travelgoods, handbags, brief-cases,purses,sheath 6 3 2 3 28,551 2,181 LT Footwear 4 3 2 5 58,824 7,123 Mineralsand Metals LT Jewelry ofgold, silver or platinum 5 2 1 2 3,219 1,429 Source:Chandra, et al, based on UNCOMTRADE 7.41 In the case of horticulture products, the need for organic certification is a major requirement that is difficult to fulfill without quality certification services. Given the high standards for fresh vegetables and fruits for the international export market (primarily Europe), and the susceptibility of these crops to diseases, only commercial farms or very well-structured farmers' associations would be able to supply the required standardized and consistent export quality. The European market imposes traceability requirements that are difficult to fulfill by infrastructure. World Bank (2006a, p.52) reports that technical constraints include: ". ..(i) informal small-scale producers. In addition, there are constraints posed by the lack of adequate 102 This decline seems exaggeratedand is probably due to are-classification. It will be validated duringthe field visit. 176 infrequent cargo flights with limited capacity and high cost o f transport'03; (ii)lack o f cold storage and handling facilities at all stages o f the chain; (iii) lack of technical knowledge at all levels; (iv) absence of certification and quality control services and laboratories; and, (v) absence of specific inputsand cheappackaging material...." 7.42 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 o f property and assets in 1994 put the country at a lower starting base, which was compounded by years o f neglect, lack of investment, and little or no maintenance o f economic services. For example, many coffee trees were destroyed, but investments to replace destroyed and low-producing trees, and to develop a stock of improved varieties, were not made until2004. The destruction also led to similar declines in exports from other sectors, such as tea and livestock. 7.43 The main barriers to trade have been energy, transport, and ICT. Energy costs have beena significant deterrent to the expansion ofmanufacturing and processingactivitie~.''~ 7.44 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 inChapter 3 of Volume 2 of this report also show that for the case of coffee, improved access to rural roads would greatly increasefarmers' income. 7.45 The high cost and unreliability of using corridor routes to reach gateway ports are severe impediments to Rwanda's capacity to access world markets. The cost per ton of transport 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 inthe cost of trade. In addition, there are extreme delays in the transit o f goods fiom Rwanda along the corridors. The average transit time between Kigali and Mombasa is four weeks. The long delays arise from congestion inthe ports as well as procedural obstacles, which account for up to halfo f the time intransit. 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 o f a transportation network with neighboring countries, which would lower cost and increase efficiency by reducing border transactions. 7.46 For the transport of perishableand high-valuegoods such as horticultural products, the lack of regular flights to major markets and 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 o f cargo, which can take a substantial number o f days and additional delays for a large container. There are also insufficient cold storage facilities at the airport. Airfreight rates are about 2.30/kg for the European market and 1.60/kg for the Middle East and South Africa markets. This rate is relatively high compared to that paid by Rwanda's competitors, mainly South Africa and Zambia, where output is increasing. Kenyan and Ugandan exporters have direct airline connections. Airfreight rates to Europe are estimated at US$1.57/kg from South Africa and US$1.30-1.50/kg from Zambia. At these rates, Rwandan exporters will be starting out at a competitive disadvantage, and the negotiation of a more beneficial rate will be critical to success (World Bank 2006a). lo4Diagnostic Trade Integration Investment Survey (DTIS), 2005. See also Chapter 5 of this volume. 177 7.47 ICT will be an important factor in providing households and firms with access to market information, direct contact with buyers, and linkages to financial institutionsand government. These links are particularly important for export diversification. Measures to improve access to low-cost ICT services will be important indeterminingthe flow o f information into and out o f rural areas, as well as the extent to which rural households can effectively participate in trade 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. 7.48 Finally, the lack o f research and extension poses a significant barrier to the adoption o f new methods and approaches to increase agriculture production. Farmerswill needadvice, basedon sound research, onways to move into new initiativesthat require new techniques and skills. Inthe case o f coffee, for example, roughly half a million farmers will require extensive training in all aspects of production, from planting to care and maintenance to harvesting, handling, and transportation. 7.6 SECTOR-SPECIFIC STRATEGIC FocusToPROMOTE DIVERSIFICATION 7.49 As noted inthe Rwandan Development Gateway Projectreport, " after independence, the export sector covered agro-industrial companies based on coffee, tea, sugar, processed h i t s and pyrethrumas well as a tannery. Import substitutingindustrieswere also introduced. The current generation of Rwandan industries includes chemical companies, building materials, printing offices, agro-industries, wooden products, textiles and service industries." 7.50 The abundanceof export discoveries inRwandasuggests that entrepreneurialfirms and farmers are able to transformcreativeideasinto exports.Infact, even inthe aftermathof the genocide, there were many discoveries across sectors (Table 7.4). However, the failure to scale up indicates that the factors driving discovery are different from those necessary for reducing the dominance of commodities inthe export mix. 7.51 Export discoveries are the result o f good fortune; but export discoveries transformed into export product lines are the result o f strategic nurturing, driven by a variety o f factors, including production technologies, technological standards, type o f land or soil, climate, production facilities, transport logistics, and market information. The factors needed to nurture an export discovery are quite distinct and often sector specific, even for seemingly similar products, such as those inthe class of agricultural products. The land, water, new farm technologies, phytosanitary regulations, and air transport that enable high-value horticultural products to be airlifted to the quality-conscious European markets are of little use in nurturingthe development of the animal products chain. The latter needs fixed foreign investment in tanneries, as well as electricity, new production technologies, and roadhail transport to transform skins and hides into leather for overseas markets. Diversifying away from wood into higher-value manufactured products, such as pulp, requires factors that are quite different from the skills requiredby the furniture sector, which sourceswood. 7.52 Each sector has its special logistical and marketing requirements. Exploringnew markets for coffee and tea with international coffee marketing firms is very different from, for example, doing business with Asian firms in the market for raw minerals. Skills requirements also vary across sectors. The skills of agronomists needed for horticultural production are non-transferable to the leather, minerals, or chemicals export sectors. Such realities suggest that there may be 178 larger gains from focusing on a few sectors with the greatest export potential, given limited resourcesand capacity. 7.6.1 Which SectorsHavethe GreatestExportPotential? 7.53 An examination of trends in export destination and world demand for strategic exports from Rwanda, and their destination, provides an indication o f which exports have the greatest potential. For example, exports to East Asia (China, Hong Kong, Malaysia) have increased from a nearly negligible share of total exports until 1999, to about 27 percent o f exports over the period 2000 to 2004 (Table 7.6). Table 7.6: Rwanda's Export Destinations, 1978-2004 (Percent of total exports) Partner Country 1978-1989 1990-1993 1995-1999 2000-2004 Germany 28.68% 28.10% 30.60% 13.60% China 0.01% 0.50% 1.60% 11.90% Hong Kong, China 0.00% 0.00% 0.10% 10.90% Netherlands 4.10% 13.go% 8.30% 7.80% United States 20.80% 8.60% 7.30% 7.70% Belgium 0.00% 15.80% 6.50% 5.90% Pakistan 1.70% 8.90% 9.80% 5.70% South Aftica 0.00% 0.04% 0.50% 3.80% Malaysia 0.00% 2.50% 0.90% 3.80% UnitedKingdom 4.30% 3.30% 3.60% 2.80% 59.50% 81.60% 69.10% 73.90% Source: Staffs calculations 7.54 The emergence of East Asia opened up large opportunitiesfor Rwandan exporters, allowingthem to export some non-traditional products that were, until recently, limited in their traditionalexport markets.However, export market diversification has beenvery product specific, and has benefitedmostly exporters of minerals and metals. In terms o f traditional trading partners, these continue to be important for its export sector, even as demand for Rwanda's non- traditional products has picked up in the EU and the US. New demand for Rwanda's traditional exports has also emerged inEuropean countries with which Rwanda has not previously traded. A good example is the demand for Rwandancoffee inPoland, Ireland, Hungary, and Greece. 7.55 Driven by growthin China, non-ferrous minerals-which were, until the late 1990s, an under-unexploited natural resource worth less than 2 percent of total exports - have surged dramatically, to become Rwanda's second largest export product. China's growth raised the demand for Rwanda's minerals and ores, comprising mainly non-ferrous ores such as titanium, vanadium, and molybdum, by 116 percent per year between 1997 and 2004. Exports of these minerals and metals increased from US$ 635,000 in 1997 to US$16 million in 2004. China presently accounts for 82 percent o f Rwandan exports of ores, and its demand almost single- handedly altered Rwanda's export mix. Import demand for the same products, as well as other minerals in the region, from Hong Kong (China), Japan, Singapore, and Thailand - has also surged since 1997, reflecting the shift toward a certain pattern o f industrial production and trade in East Asia's middle-income economies. In Hong Kong, for example, imports of Rwandan minerals averagedabout US$6 million and grew at 177 percent per year during2000-2004. 7.56 The demand for Rwanda's other non-traditionalexports tin and tin alloys, metal - sheets, chemical and other metallic alloys - increased in Hong Kong, Thailand, Singapore 179 (mostly tantalum), and Malaysia long before the surge in the demand for ores. A leading reason for exports of tin and its alloys to Malaysia, at one time a leading producer o f tin, is due to the increase in wage levels, which have made Malaysia's tin mining industry uncompetitive. Apparently, instead o f sourcing tin from its own mines, it i s more cost effective for Malaysia to import the metal from Rwanda and process it in its refineries. East Asia has also opened the doors to Rwanda's newly discovered manufactured products such as synthetic fibers and hides and skins (leather). Some animal skin product discoveries were, until 2000, exported predominantly to Italy. In 1999, Hong Kong began importing bovine products from Rwanda; by 2001, its share equaled Italy's, and by 2004, it had crowded out Italy. 7.57 About 37 percent of Rwanda's exports are directed to its traditional European partners Germany, Belgium, the Netherlands,UK, Germany, and France. In addition to - coffee and tea, many export discoveries - fresh fruits and vegetables, plants, saps, extracts, cut flowers, wicker articles, mineral ores and metals (tungsten, anthracite, metals, steel plates) are facing steadily rising demand. Even demand for one o f Rwanda's old exports - animal products, especially hidesand skins -is still strong inItaly, its oldest trading partner inthis product line. 7.58 Beyond its old markets, Rwandan exporters have also discovered new markets in Europe. While the share of its new Europeantrading partners -Ireland, Poland, Hungary, Spain, Greece, the Czeck Republic, and Turkey - is presently minuscule, their demand for Rwandan exports i s in the rise. Turkish demand for Rwandan yam made o f regenerated fiber, as well as pulp and chemical products and re-rolled iron and steel coils is typically below close to us$100,000. 7.59 However, Rwanda's exports to the U.S. have declined from 21 percent of total exports in the 1980s to about 8 percent over the period 2000-04. The bulk of Rwanda's exports to the American market have historically comprised coffee, tea, minerals, and metals. Other exports, such as textile-related products (regenerated fibers, textiles, garments, bags, etc.), have rarely been larger than US$4000. Yet, starting in 2005, special USAID efforts to develop Rwanda's exports boosted exports o f non-traditional products such as basketwork and wickerwork to stores such as Macy's. A decimal point worth of Macy's imports translated into a huge opportunity for Rwanda's exporters. Thejump inthe value o f handicrafts from almost zero to US$ 90,000 is a sign o f the enormous untapped opportunities that AGOA offers to Rwanda's non-traditional products exporters. AGOA has enabled Rwandan exporters to diversify vertically within the coffee product chain. The recent partnership between Starbucks and Rwandan coffee producers, again with the help o f USAID, has enabled exports o f washed coffee beans, a higher- value product. There is large scope for further exploring the development o f specialty coffee, the demand for which is rapidly growing, or the long list of other products that are AGOA eligible and therefore duty free and quota free. 7.60 Regionalmarkets provide another opportunity for the growth of Rwanda exports; however, it seems far more challenging for Rwanda to export within the region than to export out of Africa. In the past five years, South Africa has emerged as the new African destination for Rwandan minerals, metals, and chemicals-related exports. Within the region, the market is much smaller, and the natural endowments o f the neighboring economies quite similar. Exporters in Uganda, Tanzania and Kenya produce many of the same product^,"^ but enjoy far more developed and diversified domestic markets, which enables economies of scale in production. This creates tough competition for Rwandan exporters, who are also handicapped 105Examplesare freshfruits andvegetables, flowers, fabrics, garments, wood andrelatedproducts, metalproductsetc.. 180 because of higher transport costs associated with land-lockedness. To all three markets, Rwanda exports fabrics (synthetics), beer, chemical products (insecticides, fertilizers), and metal products. InBurundi, Rwandanexporters enjoy a larger market and are able to sell cement. All recorded exports to the immediate region were only US$1.8 million in 2004, and accounted for percent 2 o f total exports. However, there appearsto be a large volume of unrecorded trade acrossborders. 7.6.2 Which Sectors Should GovernmentSupport? 7.61 The trends in demand for Rwandaexports suggest that the greatest potentiallies in horticulture/floriculture products, leather and textiles, processed fruits and vegetables, mining, tea and processed coffee.lo6 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 explored for the main export discoveries - minerals, chemicals, textiles, pyrethrum, wood, livestock, hides and skins, flowers, and horticultural products. 7.62 Minerals and metals. This sector is the least diversified vertically - its first two tiers comprise resource basedproducts, the demand for which has boomed inthe past five years due to rising demand for minerals from China and the electronics industry.Thefirst tier's five or six raw comm~dities'~~have beenthe main drivers o f overall export diversification and export growth in the past 5 years.'" The diversification story is dominated by tin and non-ferrous ores, which account for 40 percent of total export^.'^' No public intervention i s required to develop this tier, except for regulationto address environmental issues. 7.63 The secondtier ofthe product chain accounts for several export discoveries with promise o f rapid scaling up - alloys, concentrates, scrap and pulverized products - that have potential for heavy manufacturing, though perhaps not in the medium term, and not in all resource base products. REDEMI,the state-owned miningcompany, intendsto offer its undeveloped beryllium, kaolin, and peat deposits, as well as its present facilities, for sale in 2007. REDEMI, COPIMAR and other companies operate processing facilities for Columbium (Niobium) and Tantalum at Gatumba; for cassiterite at Rutongo; and for wolframite at Nyakabingo."' Portland cement, a resource base export, is operatedby Cimenva, Rwanda's only producer of cement. Nearly all the firms operating in this tier, even if only one in each mineral sub-sector, are constrained from scaling up as long as the Government does not take measuresto increasethe supply o f electricity. Except for power, the Government's other plans to privatize REDEMI's activity in this tier appear to be adequatefor the industry's development. Preserving the status quo or improving the situation - Le., scaling up tiers one and two o f the minerals product chain -simply requires that Government ensure a steady supply of power to the minerals industry. This would preserve the concentration ofthe export mix. lo6 coffee, horticulture, and handicraftswere also noted as candidates for sector-specifictargeting in the Tea, DiagnosticTradeIntegrationStudy(World Bank, 2005). lo7 Rwanda's mainmineralexportsare columbium,tantalum, tin andtungsten lo* increaseinproductionwasmainlyduetogreaterdemandfor cassiterite(tinproduct), coltan,andwolframite. The Cassiterite accounted for 16% of Rwanda's exports in 2004. The demand for tin has been increasingin recent years, especially in China, due to its use in making consumer electronics and the ban imposed by EU on lead solders in electronicdevices(Thomas R. Yager, 2004). log Rwanda's exports also include re-exportsfrom sources in DRC, but there is no clear record o f how much comes from DRC. 'lo (2004). Yager 181 7.64 Unlikethe first two tiers, scaling upthe third tier ofmanufactured products -metalscrap, waste, plates, sheets, and strip from wrought metal (all medium technology products) - needs special government efSort, as does the rehabilitation o f the Metal Processing Association's tin smelter at Gisenyi. 7.65 Several entry points are possible to provide an enabling environment for the sector: 0 Outdated technology which in some sectors (i.e. tin smelting in 2004), led to pollution and excessive consumption of electricity. Giventhe potential of tin and its products, this disappearing manufactured export provides a perfect example of a tier that has scale economies and needs large scale investments for technological upgrading. Shortage o f adequately trained workers. Presence of scale economies that deter private firms as the profitability o f the manufactured metals products is unknown at this point - ie. information externality. 0 Coordination failure associatedwith mine-specific dilapidated infrastructureetc. Land-lockedness that raisesthe costs oftransport to the coast. 7.66 The mining sector is characterized by small producers; and the Government can take advantage of this aspect of the industry to attract investment. First, organizing producers into cooperatives would enable them to take advantage o f economies o f scale. Itwould also promote investment in the sector and facilitate the provision o f 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 o f adequately trained workers. Forming public-private partnerships to support prospecting would improve the information base on the stock o f mineral resources. Partnerships would also be o f benefit in establishing the profitability o f manufactured metal products, and would mobilize private investment to rehabilitate dilapidated and outdated machinery. As noted throughout this chapter, improving the transport and electricity infrastructure is also important for attracting investment. Box 7.1: The Case for RefinedTin Exports 7.67 Chemicals. The first tier o f Rwanda's chemicals product chain is based on diverse, naturally available organic and inorganic chemicals which are presently not available at adequate levels for production. In 2004, chemicals had a share o f only 0.06 percent in Rwanda's total exports of US$97 million. Although they have links to higher-value second and third-tier products, each link is miniscule. In 2003-2004, cyclic hydrocarbons, polyethylene, metallic oxides, and chemical fertilizers were some of the leading manufacturedprocessed chemical products that were exported directly. There is also potential for scaling up the production o f carbon black, a specialty product added to rubber tires, derived almost exclusively from the burningofnatural gas. 182 7.68 Scaling up of this sector is important for scaling up the overall export sector. Chemical products provide useful inputs for other export sectors - agricultural food, forestry/wood and paper, animal skins and hides, and textiles. They are also used to process products from other sectors for export. For example, the sector processes flowers to produce extracts that are usedto manufacture perfumes, pesticides, and fertilizers. 7.69 The most important chemical exports are derived from Rwandan pyrethrum flowers, which have a high concentration o f pyrethrums. Rwanda i s one o f the largest producers o f pyrethrum in the world, and its pyrethrum concentrate is exported mainly to the U S under AGOA, and to Kenya and South Africa. Presently, Sopyrwa, the pyrethrumproducing company, exports crude pyrethrum extract to be refined abroad. Rwanda can also use its flowers for oil distillation or solvent extraction of other products, such as paprika. A pyrethrum spray formulation, which adds value to the product, is available from the PyrethrumBoard of Kenya."' 7.70 Expansion of the Pyrethrumsub-sector is constrained by a dormant refiningunit - due partly to electricitysupply -which, ifoperational,could allow Sopyrwa to complete the refining process and capture significant additional value added. To partly address the electricity issue, a critical sector-specific action would be to introduce solar''2 dryers to dry f l 0 ~ e r s . l ' ~Below-capacity production o f pyrethrummeans that Rwanda i s not taking advantage of high demand in the world market and existingpreferential trade arrangements with the United States andthe European Union. 7.71 Textiles: Exports from this sector comprise one o f the larger product chains with newly discovered exports o f nearly US$ 500,000 in 2000-04. The potential o f this sector for export diversification far exceeds that of others, especially because it produces manufactured goods which have a stable and growing demand, it enjoys special preferences in the US market, and its growth i s not constrained by Rwanda's relatively less skilled labor force. 7.72 Rwanda is one o f the few AGOA eligible countries in SSA that can export textiles, yarn and fabric without cultivating cotton. This should flag it as an attractive location for textile and garments producing multi-national corporations, similar to those in Lesotho, SSA's largest garment exporter to the U.S. However, currently, there is only one major privately owned textile firm, UTEXRWA that operates in Kigali. In operation since 1985, the firm employs 1000 workers, in cut make and trim operation to produce up to 6,000 garments each day from fabric made of raw cotton from Uganda and Tanzania, and polyester fabric made in South Africa. The firm has rapidly expanded in the last few years to become an integrated textile company with spinning, weaving, dyeing, printing, and apparel-making capabilities. It maintains a designs library o f more than 2,000 products and can easily add to current production and shift manufacturing lines within days to meet new demand.'14 Other small textile weaving companies are also expanding, including a firm that specializes inhand-loomed textile products. ''"Agribusiness in SustainableNaturalAfricanPlantProducts(ASNAPP)NaturalProducts Assessment: Potentialfor Economic Growthand Trade inRwanda," Chemonics for USAID (2002). preparedby Chemonics Internationalunder ADAR,July 2002. This would eliminatethe needfor harvestingfuel woodto fire the ovens, thus contributingto improvedand sustainable environmental management. `I3AgribusinessDevelopmentAssistanceProjectinRwanda(ADAR),SecondQuarterProgressReport, Chemonicsfor USAID(2004). "Investment Climate Statement-Rwanda," US Departmentof State (2005). mnr.state.~ov/e/eb/ifd/2005/42425.htm. 183 7.73 Exports from the textiles sector comprise one of the larger product chains, with newly discovered exports of nearly US$500,000 in 2000-2004. The potential of this sector for export diversification far exceeds that o f other sectors, because (i) it produces manufactured goods for which there is a stable and growing demand; (ii) it enjoys special preferences inthe US market; and (iii)its growth i s not constrained by Rwanda's relatively less skilled labor force. However, high utility costs have been a constraint to production. While export discoveries of almost US$500,000 in 2000-2004 have triggered rapid growth in the sector, Rwanda is not yet fully exploiting its textiles and garments exports potential, which would enable it to take full advantage o f AGOA and scale up."' 7.74 One of the potentialexports being considered in Rwandais silk textiles. The case o f this nascent industrymay have lessons for other sectors. Silk is not native to Rwanda, but with help from the UN's Food and Agricultural Organization (a), and in cooperation with the Institut des sciences agronomiques du Rwanda (ISAR), under the Ministry o f Agriculture,'16 mulberry trees have been planted in the Butare region, as a first step toward silk production. A similar approach could be adopted for other export sectors. 7.75 Wood. In 2000-2004, Rwanda exported about US$270,000 worth of newly discoveredsecond and third-tier manufacturedwood products, such as pulp, cellulose, and paper, most of whichwere either resourcebased or light technology manufactures. Exports o f furniture, a second-tier, low-tech product, have recently exceeded US$lOO,OOO 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 o f electricity is one important factor affecting the sector; and an equally important factor i s the availability o f skilled labor. 7.76 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. Until2004, wicker basket and hat exports averaged US$24,000 per year. Through support from a USAID project, one enterprising woman's contact with the Macy's department store causedthe industry's exports to increaseto US$25,000 in 2005. The challenge will be to continue to meet increasing demand while maintaining quality. 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 handicraftproduction training program should helpto further expand this promising export industry. 7.77 In the case of handicraftproduction,there has already beena move to explore more opportunities inthe US marketunder the umbrellaof African Growthand OpportunityAct (AGOA). It should also be possible to explore other markets in Europe. To increase sales, a handicraftmarketingprogramwill be needed, as will handicraft production training. `15 "The African Growth and OpportunityAct: AchievingSuccessthroughthe African GlobalCompetitive Initiative," June2006, www.agoa.vov l6"Rwanda: Diagnostic Trade IntegratedStudy." World Bank (2005). 184 ' Box 7.2: RwandanHandicraftsExports The handicrafts industry in Rwanda is largely informal; artisans use local materialto producehandicrafts and sell them locally. Some of the crafts produced are metal works, stone cuttings, basketwork, pearl work, and some leather products. Women in Rwanda have been weaving baskets for years, mostly for decorative purposes at home or to sell locally. After the genocide, a local woman started organizingthe weavers and selling their products incraft fairs inthe US. A woman's rights activists in the US presented a proposal to Macy's, and the baskets were sold for the first time during the 2005 holiday season. Currently, there are 1,500 women weavers employedinthisproject.They sell the larger, foot-high baskets for about $20 each, and smaller ones for about $5. Since many weavers produce halfa dozen of the large baskets a month, that means they can earn $120, better than the average per capita monthly income of ~ ~ $ 1 0 0 . ~ ~ ~ 7.78 Livestock.Instead of starting with exports of live animals and meat, Rwanda's livestock product chain, which accounted for less than 2 percent of total exports, begins at the second tier with high-value animal products (hooves, claws, bones, ivory) and light and medium technlogy manufactures. The five tiers in the livestock chain offer increasing diversification potential, moving up from a large variety o f animal skins and hides to leather and related products. Inspite o f the export potential, however, scaling up in this industry i s stymied by two mutually reinforcing forces. First, exports o f higher-value skinproducts, such as footwear and travel bags, are rapidly slowing down, indicating a loss o f global competitiveness. Further progression inthe same direction is likely to lead to a disappearance o f these two links in the product ladder. Second, raw animal skins, the most lucrative traditional export product, are also disappearing. Rwanda traditionally exported 80 to 90 percent o f skins and hides and leather to Italy, and more recently to some East Asian countries, but shrinking exports suggest that domestic supply constraints are probably to blame. Ifso, the industry i s threatened from the first as well as the top tiers and i s unlikely to survive. Leather export discoveries are a rare exception to other export products inthe livestock sector. 7.79 What explains the diminishing size o f animal skin-related exports? Market analysis shows that globalization is pushing the tanning industry from Italy and Spain, the traditional importers to China and Thailand. While some o f this adjustment is visible in the direction o f Rwandan exports, the diminishing size o f the market remains unexplained. Rwanda and its five regional neighbors exported half o f one percent o f total hides and skin exportedto Italy in 2004, representing sharp declines in exports, and signaling a regional loss of competitiveness. Since 2001, China has become the largest importer of hides and skins. Some analysts believe that China i s trying to cut down its import of raw products due to environmental concerns, and will increase the import of semi-processed (wet-blue) products."' This should open the door to processedskins exports for Rwanda. 7.80 To take advantage of these global trends in the hides and skins sector, a national strategy is needed to support production, and to invest in upgradingslaughterhouses and collectionsystems to ensure consistentquality. A key component ofthe strategy would include training in handling, to ensure consistent quality. Organizing small leather producers into `17 Poolos (2006), "Rwandan Women Weave New Futurefrom Baskets," Awakened Woman e-magazine, (March 1). www.awakenedwoman.com. `18 FoodandAgricultural Organization(2006), "Consultation on Hidesand Skins," Arusha, Tanzania, (February lPut bothinbib1 185 production units would also help in establishing a value chain and give them access to market information and finance. 7.8 1 Edible agricultural products. This sector is well diversified horizontally, to support several product chains that produce high-value fresh fruits, vegetables, legumes, malt, and coffee and tea for export. While scaling horizontally can be sufficient inthe short term, there is ample scope for vertical diversification toward processed foods in the longer term. Presently, Rwanda exports frozen and chilled products, but potential exists for further transformation into manufactured foods such asjams, canned and bottled products. 7.82 Rwanda's favorable climate, relatively cheap labor force, and proximity to European markets give it some comparative advantage in producing horticulture products. The demand for horticulture products inthe European market has grown very strongly. Compared with year 2000, the demand for fresh vegetables had increased by 61 percent and fresh fruit by 66 percent in 2004. 7.83 Scaling up has several obstacles, however. Perishable horticulture products require post- harvest technologies, and quick and regular transportation. It also needs economies o f scale, which are difficult to achieve inRwandabecause of its small size. Annual exports o f horticulture products in 2002-2004 averaged US$643,000, accounting for one percent o f total ex orts. Further, production patterns have been unstable, usually driven by exports of one product. A USAlDreport in2001120statedthat there was sufficient air capacity out of Kigali to Europe and South Africa. However, there were no bindingcontractual arrangements between air carriers and exporters; and because of low production o f exporting goods, exporters had no bargaining power to negotiate a favorable schedule. 7.84 These externalities provide a case for special support for the horticulture industry. Examples121of how Uganda, Kenya, India, and even Chile scaled up in horticultural exports may have some relevance for the Government. Specific interventions in those cases included public provision of post-harvest technologies for grape growers' cooperatives (India); refrigeration facilities (Kenya); foreign consultants who brought superior farm technologies (Chile); and government-donor partnership (Uganda). 7.85 There are significant prospects for exports in a number of key products where global market trends are positive, if yields and quality can be increased. Rwanda has the opportunity to benefit from preferential access to the EUmarket under the Everythingbut Arms (EBA) and the US market under AGOA. As an example, consider the case of scaling up banana, the top cash crop in Rwanda, which has been problematic. Banana growers face declining yields due to poor soil fertility, as well as pests and diseases. Currently, ISAR (Institut des science agronomique du Rwanda, or the National Agricultural Research Institute)122is conducting research to improve productivity and yields for small producers, and i s also promoting cooking and dessert bananas as ` I 9 As an example: the United States was the principal importer of some unspecified vegetable materials (recorded as "other material ofvegetable origin") from Rwanda.A quick rise ofthe USmarket inlate 1980sto early 1990s boosted the sector. But suddenstop of export to US since 1993dragged it downto the bottom. With very limited capacity, it is a more easible strategy for exporters to focus on one sector and one market. `'O "ATDT BananaProject-Rwanda" mrw.isar.cgiar.org/atdt/Banana/banana.htm. Chandra, V (ed.). 2006. Technology,Adaptation andfiports: How some developingcounties got it right. World Bank ISAR stands for Znstitut des science agronomique du Rwanda, or the National Agricultural Research Institute ( N M ) . 186 exports ATDT Banana Project). The experience o f developing countries in arresting declining yields as inthe case o f maize on small farms inIndiamay be useful for Rwanda 7.86 In Rwanda, sector-specific support for horticulture implies crop-specific support for special crops such as passion fruit, the most highly prized fresh fruit by European consumers. Demand for its juice and puree i s also high. But scaling up i s obstructed by crop diseases; the producers are finding it hard to obtain disease-free seeds for cultivation. USAID'SADAR'23 project has worked with passion fruit producers and Rwandan exporters to increase the production of export-quality fruit and develop appropriate post-harvest handling and export strategie~'~~; and a local company called SHEMA FRUITSplans to export passion fruit puree to Europe. There are also plans to build a factory to manufacture products such as passion fruit pulp, for export to makers of ice cream, sherbet, and pastry in Europe and the US. ADAR i s also helpingfarmers to managedisease with minimumuse ofpesticides. These examples illustratethe kindsofmeasuresthat are neededto scale uphorticultural exports inRwanda. 7.87 To some extent, scaling up of the horticultural exports will depend on the alleviation of the key constraints to improve access to transport, finance, and energy, and achieve more effective organization o f the rural sector. But most o f the scaling up will depend on the Government's efforts to identify the sub-sector specific and crop-specific obstacles to exports. Currently, there is no organized strategy to developRwanda's horticulture sector. 7.88 In the case of tea and coffee, priority measures for improving productivity include improving access to finance and inputs; disseminating information; and establishing and enforcing standards. For the coffee sector, measures are also neededto improve the capacity o f 123ADAR standsfor AgribusinessDevelopmentAssistanceProjectinRwanda,fundedby USAID. lZ4USAID(2004). RwandaIntegratedStrategic Plan2004-2009, Volume 1(January). lZ5"Considerations for Initiating a Specialty Coffee Industry in Rwanda,"ADAR Project, Chemonics for USAID (2002). "The African Growth and OpportunityAct: Achieving Success through the African Global Competitive Initiative," (2006). www.agoa.gov. 187 washing stations, which often operate below capacity due to water and electricity shortages. Further, the current installed capacity of washing stations can process only half o f the beans produced, and more washing stations will needto be built. 7.89 Increased productivityin the tea sector will require training on care of the bushes, and a pricing and 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 investmentinfactories currently operating below capacity. 7.6.3 Sector-SpecificPolicyObstacles May Be the Critical Constraint 7.90 In sectors where there are many small exporters, scaling up requires addressing sector-specific obstacles such as pests, diseases, and lack of access to modern technologies. Further, small exporters need assistance in negotiating air transport. Insome cases, Government will needto intervene, probably with the help o f donors, to reach foreign buyers. Often one buyer is sufficient - Macy's and Starbucks are examples. Wooing one large firm to start processing fruits and vegetables may be necessary, as this i s unlikelyto happen on its own. The Government can also enable scaling up by promoting the Rwandan brand name and enacting regulations to maintain quality control, especially inagricultural export^.'^' 7.91 Inmanycases, Rwandanexporters canbenefitfrompreferentialtrade arrangementswith the EU and the US. Special efforts to unravel what is neededto scale up imports in each sector may help Rwandan exporters to move faster into more manufactured products. Collaboration with foreign buyers and firms may facilitate this effort. 7.92 Access to investment and working capital for commercial agriculture i s extremely limited. As noted in Chapter 5, currently only 2.3 percent o f bank credit is used to finance activities in agriculture (which employs more than 90 percent o f the population and produces 40 percent o f GDP).'*' Microcredit by itself will not suffice; it is necessaryto extend bank finance to the rural sector to commercialize farming. This issue needs to be addressed in the context o f the strategy to develop the financial sector, which i s currently beingdeveloped and implemented. 7.93 Excessive fragmentation of farming due to the small size o f farms i s a major constraint for horticulture and floriculture. As land i s consolidated, bank finance can play an important role in helping large-scale farming to become viable. There is also a need to link rural producers to local, national, regional, and international markets. Cooperatives have the potential to play a strong role, provided they have the legal status to support farmers inmarketing their products, as inthe case ofMAHAgrapes inIndia.I2' 127A goodexample is Chileanwine, promoted aggressivelyby ProChile andChile Vid (Chandraed., 2006). lZ8DTIS: DiagnosticandTrade IntegrationStudy, World Bank (2006) lZ9Bridging the KnowledgeGap in CompetitiveAgriculture: Grapes inIndia (V. Chandraed.), 2006. 188 Box 7.4: The African Growth and Opportunity Act Rwanda is one of beneficiary countries of the African Growth and Opportunity Act (AGOA) of 2000. AGOA promotes two-way trade between US and Sub-SaharaAfrican countries by providing beneficiary countries with the most liberal access to the US market availableto any country or region with which the UShasnoFreeTradeAgreement. AGOA offers Rwanda several special opportunities for export diversification, but these cannot be accomplished with a single stroke of the policy pen. Diversification within each product chain requires concertedeffort. Rwandancoffeeandwicker basketexports are two examples. By investing in local infrastructure and coffee washing stations, and by supplyingtechnical training and support, USAID fosteredthe development of the Rwandan specialty coffee industry. The Rwandan coffee industry went fiom producing no specialty coffees in 2001 to producing 1,200 tons in 2005, yielding approximatelyUS$3 million insales, mainly to the US. Exportsofspecialtycoffeewere projectedto reach US$6million in2006. Source:www.agoa.org. 7.6.4 GeneralConstraintsto ExportDiversification 7.94 In almost all sectors, landlockedness and lack of sufficient power are problems. Provision of moreenergy, probablyfrom Lake Kiwu's natural gas project, as well as better rail, road,and air transport, seem indispensablefor faster export diversification.One way to achieve this may be regional integration, which is important for Rwanda, not only to expand its market, but also to improve the business environment and enhance competition. 7.95 A transportation network with neighboring countries will reduce border transactions and delays, leading to lower cost and greater efficiency. In 2004, the Governments o f Burundi, Congo (Kinshasa), and Rwanda discussed the rehabilitation o f the Ruzizi Ihydropower station in Burundi. Repairs to Ruzizi Iwould increase capacity from 28.2 MW to 39.6 MW, and allow the plant to export power to Rwanda. There have also been discussions to rehabilitate Ruzizi I1and the possibly construct Ruzizi 111. 7.96 Currently, transportationcosts are pushing up the cost of trade by an estimated 30 to 50 percent. Targeted support for infrastructure will probably help, as the Government cannot improve it economy-wide at once. Inparticular: 7.97 Better air transport will help high-value, low-volume goods to be flown directly to external markets. Ideally, the investment would come from the private sector; however, Government may needto intervene. 7.98 The cost oftransport from Kigali to Mombasa is 70 percent higherthan from Kampala to Mombasa. There are extreme delays along transport routes, especially main transit corridors. The average time from Mombasa to Kigali i s 4 weeks. Rural transport costs from the farm gate in Rwanda to Mombasa are 80 percent of the farm gate price; while from the farm gate to Kigali is about 40 percent. This amounts to an implicit tax on coffee and other exportables (DTIS, 2006). 7.99 Two straightforward solutions: build corridor routes that provide access to gateway ports and world markets; and provide rural roads that link agricultural areas to local markets. 189 7.100 The regionaldevelopment of power systemswill increasethe availability and reduce the cost of electricitywhich is one of the major deterrentsto doing business in the country. Regional capital markets will enable government to borrow at lower costs and encourage competition. Rwanda can also benefit by becoming the distribution center for the sub-region. Rwanda has the advantage o f greater security for people and property and a lower level of corruption compared to its neighbors. The country is already re-exporting to DRC, Tanzania, Uganda, and Burundi. The development o f ICT structure will help Rwanda become a regional distribution center for exports. Another benefit may be better employment opportunities for semi- skilledand unskilledlabor, due to greater mobility. The country's agriculture sector may benefit from greater demand. However, since most agricultural trade is informal, it will be hardto assess the benefits to that ~ect0r.l~' 7.101 Finally, capacity issues are a constraint faced across all sectors. Most entrepreneurs in the private sector do not have the training, experience, and skills needed to manage their business competitively. There is a need to provide technical, managerial, and international marketing training programs for private entrepreneurs and managers; and to develop sector- specific skills. Examples include skills for managing coffee and tea plantations; applying new agricultural technologies on farms, operating minerals and metals factories, linking to a supply chain, and marketing. 7.102 The Government has established several public sector institutions to support the private sector. However, the private sector needs help in adopting and implementing management, quality, safety, environmental, and other standards and certifications. Currently, Rwanda has no quality certification or management systems, and both have to be in place before the country can enter high-valuemarkets. 7.7 CONCLUSION 7.103 First, eliminate the barriers to trade. Foremost among these is the electricity problem, which constrains the capacity of processing plants. The recent investment in Lake Kivu methane gas development should help to reduce the cost and increase the reliability of electricity in the long term. Second, invest in rural roads to link rural households and small producers to markets. Third, identify sector-specific constraints to productivity, to help develop the value chain and increase producers' access to training and finance. Fourth, facilitate and promote engagement of the private sector. Fifth, provide assistance to producers' organizations, including training in the use o f 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 MINAGRI; and for other producers coming from MINICOM. RIEPA would have the role of trade promotion outside the country. 7.104 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. 130ImaniDevelopmentGroup(2006) 190 8. PRIORITIZATIONOFCONSTRAINTS: WHAT I S THE MOSTEFFECTIVEUSEOFAID FUNDS? 8.1 This report has examined the sources and constraints to growth, as well as the sector- specific issues and constraints that influence growth at the sub-sector level. It has also discussed several challenges that the Government faces as it works to meet its Vision 2020 targets. 8.2 The country's narrow focus on rain-fed subsistence agriculture has made it vulnerable to both climatic shocks and terms-of-trade shocks. Inaddition, the country relies heavily on foreign assistance for its financing needs. In the case of private sector development, although the main reforms to minimize price distortions and liberalize trade have been implemented, the payoff in terms o f increased growth has been limited because o f the poor condition of infrastructure - the result of lack of investment and poor maintenance. Electricity i s viewed as more o f a constraint by large and manufacturing firms, followed by the cost o f finance, the tax rate, and transport costs. This contrasts with microenterprises, which identify their main constraints as low level of purchasing power and demand, limited access to markets (due to poor roads), lack o f access to finance, lack o f support for business services, and lack o f skills. A poor business environment and low skills level are also constraints for the service sector. In the case o f exports, the challenge i s to implementmeasuresto scale upthe minute but diverse areas o f discovery. 8.3 The analysis in this Country EconomicMemorandumsuggests a number of priority actions for transforming the economy and sustaining growth. Given the country's high dependence on aid, there needs to be a focus on improving the efficiency of spending to yield the greatest growth benefits. 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, the substantial progress in the social indicators has not been accompanied by progress in the delivery o f economic and infrastructure services. To generate growth and reduce poverty, there will need to be increased focus on private investment. Spending should go to areas where absorption can take place quickly. In the short to medium term, this means a focus on improving productivity in the agriculture sector. In the medium to long term, it means increased capital investment to reduce the costs of energy, water, and transport, and to increasethe supply of skilled labor. Four mainPriorityMeasuresare identifiedto support growth(Table 8.1). These are: Priority Measure 1: Invest to ease the inJFastructure constraints to growth and export diversijcation; Priority Measure 2: Put inplace supportivepoliciesfor improved spending eflciency; Priority Measure 3:Transform the agriculture sector to be more market oriented; Priority Measure 4: Provide support and incentives for private sector development. PriorityMeasure 1:Investto ease the infrastructureconstraintsto growthand export diversification. 8.4 In the medium to long term, investments in energy supply, water supply and irrigation, and transport will be the most crucial for easing infrastructure constraints to 191 sustained growth. The high transaction costs resulting from poor infrastructure have seriously undermined the productive capacity of 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 i s also needed to elaborate a comprehensive strategy for developing alternate sources o f 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 increasedproductivity, 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 supply constraints. In particular, there needs to be a systematic, planned approach to rehabilitation and maintenance o f roads, with a focus on communal and non-classified roads and options for cost recovery. Inaddition, investments to improve standards and quality, and the construction of cold storage facilities at airports, would facilitate the development o f perishable high-value exports such as horticulture products. PriorityMeasure2: Put in placesupportive policiesto improveabsorptivecapacity. 8.5 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 of expenditure needsto 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 help to increase private investment and the absorption o f aid. Inthe case o f 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. 8.6 Additional measures to support increased absorption include wide-scale skills development and formulationof a populationpolicy. The analysis inthe CEM highlightsthe shortage o f technical and managerial skills to support business development. To address this issue, the Government could play a facilitating role in strengthening partnerships between higher education and the private sector, to ensure that the education system is producing graduates that can meet the demand of 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 the policy measures to reduce fertility (including family planning) would helpto reduce the populationgrowth rate. Priority Measure3: Transformthe agriculturesector to be moremarketoriented. 8.7 In the short to medium term, the focus will 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. Constraints to production in 192 agriculture have been found to be the most binding constraint to exports, particularly for the traditional exports of coffee and tea. Moreover, the low levels of returns to agricultural labor are a result o f the high degree o f labor surplus, and not due to the low returns to investment in land. The yield gap is very low, and measures to improve inputuse, water management and irrigation could help to improve labor productivity inthe sector. 8.8 It is important to increase the market orientation of the sector, and the degree of value addition, in order to raise incomes and reduce poverty. Measures will be needed to raise awareness about the appropriate use o f improved inputs. One approach would be to support the organization of farmer associations and cooperatives around clearly defined value chains. This would require investments in developing viable commodity chains (in horticulture, leather, or maize for domestic market) or strengthening existing ones (coffee, tea, pyrethrum). Training would be needed in the proper use and application of inputs. To improve production for the market, measures are also neededto improve the state o f rural roads that link households to local markets. The Government could play a facilitating role interms o f providing information, setting standards, and regulation. Strategies would also be needed to increase farmers' access to micro- finance. Investments in strengthening the extension system are important, and particularly increasing the number of agronomists and level o f veterinary services. Needed investments in rural infrastructure would be closely aligned with measures already outlined in Priority Measure 1. Inthe case of rural development, the focus would be on rehabilitation and maintenance o fthe non-classified gravel and unpaved roads. Priority Measure4: Providesupport and incentives for private sector development. 8.9 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 strengthenthe links between the agriculture sector and those non-agriculture sectors that rely on agricultural inputs. Establishing such links will require the relevant investments ininfrastructure to support private investment.Itwill also require sector- relevant skills development to support the integration of farmers and MSSE owners into market and export-oriented value chains (basket weavers, hrniture makers). The tourism sector also provides an opportunity to develop non-farm sources of income. However, to ensure the development of a high-end niche tourism market that is environmentally sound and supports community development, a comprehensive development plan i s needed that specifies standards, regulations, and zoning. 8.10 The problem of low availability of skilled labor cuts across all types and levels of firms. Inthe case of micro and small enterprises, the analysis inthis CEM indicate that 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 Survey and the assessment o f the microenterprise sector, discussed in Chapter 4. Training support for business development and to facilitate access to information should helpto reduce some of the constraints faced bythese firms. Training could also be used as an incentive for firms to move to the formal sector. Many firms remain informal due to lack o f knowledge or information about the requirements for formal registration o f businessesand property, or exporting, or taxes. Therefore, a first step would be to streamline registration procedures and make information easily accessibleto firms. 8.11 Private sector firms, 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 o f 193 Rwandan businesses. This should involve the combination o f full-time training; continuing and part-time education in tertiary academic institutions and vocational training centers; and occasional management and technical training workshops in specialized private and public institutions. The Kigali Institute of Science and Technology has initiated various levels o f full and part-time technical and management training programs that can serve as models 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. 8.12 SME operators are willing to pay for businessservices, particularly training, if the cost is reasonable, but currently there are a limited number of 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. About 50-70 percent o f SMEs in the rural areas o f Rwanda are organized into associations or economic interest groups, yet there still appear to be gaps in the supporting industriesand services for many businesses. O fthe associations that are formed, the majority are not effective in serving their members. Inorder to fully integrate SMEs into export value chains, there need to be improved institutional support and business services. Supplier inputs, quality control regulatory bodies, efficient transport and logistics services that can be accessed by SMEs, and networks to facilitate marketingactivities are critical. 8.13 I t will also be criticalto ensure equal access of both women and men entrepreneurs to the training programs, to ensure that both groups benefit from growth. The results presented here indicate that education is an important tool for reducing gender inequality in earnings (as well as inequality across socioeconomic groups). The results also show that informality is characterized by low education levels o f firm operators, and that women are over- represented in the sector. Many firms have remained inthe informal sector mainly because of a lack of knowledge of what is required to formally register their enterprises, or pay taxes; or because of the high costs (in both resources and time) o f starting or registering a new business. Therefore, increased training and education for men and women entrepreneurs could be a powerful means of supporting the growth of small andmicro businesses. 8.14 Skills development is particularly important for the competitiveness of the manufacturing, ICT, and tourism sectors. In the immediate term, the 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, and to facilitate increased enrollment and completion rates at the secondary school level. These measures will be important for reducing the difference in earnings between men and women. 8.15 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 o f financing instruments that are more suited to smaller firms, perhaps in collaborationwith the Union banque populaire du Rwanda, would help to ease these constraints. Support for integratingfirms into export value chains would be another means of increasing their accessto finance. 194 Y 0 .* a h B . e . . . 6 e, m *0 0 Y 9. Annex 1: New Estimatesof Gdp Rebasedto 2001- Overviewof Methodology 1.1 In 2006, the newly established National Institute of Statistics Rwanda released its first official national accounts estimates.131 A benchmarking exercise was undertaken to establish the best possible estimates of GDP for 2001, given the data available. The approach usedto derive the GDP estimateswas basedon the UnitedNations System o f National Accounts 1993 (UN93). Previous to this, estimates of GDP had beenbased on the UnitedNations Systems o f National Accounts 1968 (UN68). The figures for 2001 were compiled using the commodity flow approach, which involves balancingestimates o fthe supply and demand for a detailed list o f goods and services, resulting inmore robust estimates of GDP. 1.2 Estimates of the value of household consumption were available from Household Survey (IntegratedSurvey of LivingConditions, EICV) data 2001. In estimatingsupply, four modes of production were examined - private formal sector (businesses, excluding agricultural activities, registered for the value-added tax, VAT, at the Rwanda Revenue Authority); the informal sector (marketed production not subject to VAT); the non-monetary sector (goods and services produced by the user); and government and NGOs (mainly public administration, education, and health). 1.3 The gross value added of each activity is measured at basic prices; i.e., excluding VAT and other taxes on products. VAT and other taxes on products (less subsidies) are added separately inorder to evaluate GDP at market prices. Previously, only import duties were added separatelybecauseother taxes were, inprinciple, included with the corresponding activity. 1.4 As a result of the benchmarking exercise, the base year for national accounts estimateswas changedfrom 1995to 2001. Estimates o fthe overall level of GDP and its growth since 2001 did not change significantly, despite several changes at a more detailed level. However, estimates of levels of GDP for more recent years, including 2005, are higher than previously projected. This i s due mainly to the fact that the revised methodology took into account the significant increase inthe price of basic food crops since 2002. Manufacturing output i s lower, mainly because excise taxes have been excluded; the estimated level o f construction activity in rural areas was considered to be high relative to the results coming from the 2005 household survey; and finally, the estimates for the education and health sectors cover both government and non-government activities and are significantly higher than in2001. Implicationsfor Sourcesof GrowthEstimation 1.5 As mentioned in Chapter 3, the broad results on the sources of growth do not change, whether GDP is based on either UN93 in constant real 2001 prices, or on UN68 in constant 1995 real prices. In particular, food crop production accounts for the single highest 13' See Governmentof Rwanda. 2006. The NationalAccounts of Rwanda- Sources andMethods(base year 2001), NationalInstituteof Statistics, December. The document details the methodologyusedto derivethe new GDP estimates. 198 contribution to GDP growth. However, across the broad categorization of sectors (primary, secondary, tertiary, or services), the service sector displays a marginally higher sectoral contribution to GDP compared to the agriculture sector. Annex Table 1 and Annex Table 2 show the corresponding supply and demand-side decompositions based on the 2001 constant real GDP estimates. Annex Table 1: Sectoral Contribution to GDP Growth, 1998-2004 (Period average, base 2001=I0O)d 1998-2004 1998-2001 2002-04 Growth Share Contribution Growth Share Contribution Growth Share Contribution Agriculture 6.5 37.1 2.4 8.8 37.3 3.3 3.5 36.8 1.3 Foodcrop 6.6 31.8 2.1 9.0 31.8 2.9 3.4 31.9 1.1 Export Crop 25.8 1.1 0.3 34.5 1.1 0.4 14.2 1.1 0.2 Livestock 1.2 2.4 0.0 0.1 2.5 0.0 2.7 2.2 0.1 Forestry 5.8 1.4 0.1 8.3 1.4 0.1 2.6 1.3 0.0 Fisheries 11.7 0.4 0.0 18.5 0.4 0.1 2.6 0.4 0.0 Industry 5.9 14.2 0.8 5.1 14.4 0.7 6.8 14.0 1.0 Miningandqumying 32.4 0.4 0.1 55.8 0.4 0.2 1.3 0.5 0.0 Manufacturing 2.8 7.2 0.2 0.8 7.4 0.1 5.5 6.9 0.4 Ofwhich : Food 9.6 1.6 0.2 12.1 1.6 0.2 6.2 1.6 0.1 Beverages,&tobacco -3.9 2.9 -0.1 -9.3 3.2 -0.3 3.3 2.5 0.1 Others 9.2 2.6 0.2 10.1 2.6 0.3 7.9 2.7 0.2 Electricity, gas, &water 4.7 0.5 0.0 6.3 0.6 0.0 2.4 0.5 0.0 Construction 9.3 6.1 0.6 8.8 6.0 0.5 10.0 6.1 0.6 Semces 7.4 43.0 3.2 7.7 42.6 3.3 7.1 43.5 3.1 Wholesale &retail trade 3.8 9.9 0.4 3.1 10.1 0.3 4.8 9.6 0.5 Hotelsandrestaurants 11.6 0.9 0.1 16.7 0.9 0.2 4.8 0.9 0.0 Transport, storage, communication 12.1 5.5 0.7 15.7 5.3 0.8 7.3 5.9 0.4 Finance,insurance 15.0 3.1 0.5 13.3 2.8 0.4 17.1 3.5 0.6 Realestate, businessservices 5.7 9.9 0.6 6.6 10.3 0.7 4.5 9.5 0.4 Publicadministration,education,health 7.3 12.9 0.9 6.6 12.7 0.8 8.3 13.2 1.1 other personalservices 73.1 0.7 0.5 120.1 0.5 0.6 10.4 0.9 0.1 Adjustments 6.1 5.1 0.3 7.3 5.8 0.4 4.4 5.7 0.3 Less: lmputedbankservicecharge 10.7 -1.7 -0.2 11.0 -1.7 -0.2 10.2 -1.7 4.2 Plus:VAT andother taxesonproducts 7.0 7.4 0.5 8.1 7.4 0.6 5.5 7.4 0.4 Gross DomesticProductbl 6.7 99.9 6.7 7.6 100.4 7.1 5.5 99.2 5.5 Source: Staff calculations using data from Ministry of Finance and Economic Planning, Department o f Statistics. BasedonUN93 GDP estimates Note: Real values of GDP and its components are obtained by dividing nominal values by their own GDP deflator, usingreviseddata; andtotal maynot sumperfectlydue to roundingup. 199 Annex Table 2: Sectoral Shares in GDP, 1995-2005 (In Percentage, based onyear 2001 relativeprices) a/ 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Agriculture 39.7 41.1 35.8 36.9 31.1 37.2 37.3 38.1 36.8 35.0 34.2 Foodcrop 33.2 33.9 30.4 31.3 32.1 31.9 32.0 33.8 32.0 29.8 29.6 Export Crop 1.7 1.2 0.5 1.0 1.2 1.1 1.1 1.1 0.9 1.3 0.9 Livestock 3.0 3.9 3.2 2.8 2.5 2.4 2.4 2.2 2.3 2.2 2.1 Forestry 1.5 1.8 1.4 1.4 1.5 1.4 1.4 1.3 1.3 1.3 1.2 Fisheries 0.2 0.2 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.4 0.3 Industry 15.9 16.3 15.6 15.3 14.5 13.6 14.2 13.4 13.8 14.7 14.7 Miningandquarrying 0.2 0.2 0.3 0.3 0.2 0.3 0.8 0.5 0.4 0.6 0.7 Manufacturing 9.4 9.4 9.0 8.4 7.4 6.8 6.9 6.8 6.8 6.9 6.7 Ofwhich: Food 1.8 1.6 1.5 1.5 1.6 1.6 1.8 1.6 1.5 1.8 1.7 Beverages,&tobacco 5.9 5.7 5.1 4.4 3.2 2.7 2.5 2.7 2.4 2.3 2.4 Others 1.8 2.1 2.4 2.5 2.6 2.5 2.6 2.6 2.9 2.8 2.5 Electricity,gas, & water 0.6 0.5 0.5 0.7 0.7 0.5 0.4 0.5 0.5 0.4 0.5 Construction 5.7 6.3 5.8 5.8 6.2 6.0 6.0 5.6 6.0 6.8 6.9 Services 42.2 38.9 42.1 41.9 42.2 43.5 42.8 41.9 43.7 44.8 45.6 Wholesale& retailtrade 16.2 10.9 11.9 10.8 9.7 10.1 9.9 9.7 9.5 9.7 9.9 Hotelsandrestaurants 0.5 0.7 0.7 0.8 1.0 1.0 0.9 0.8 0.9 0.9 1.0 Transport, storage,communication 4.5 4.3 4.4 4.6 5.1 5.5 5.8 5.8 5.8 6.1 6.3 Finance,insurance 2.3 2.1 2.4 2.5 2.7 3.1 3.0 2.8 3.6 4.0 4.2 Realestate, businessservices 10.0 10.0 10.1 10.5 10.7 10.2 9.7 9.2 9.7 9.5 9.6 Publicadministration,education, health 8.6 10.8 13.1 12.6 12.7 12.7 12.6 12.8 13.3 13.6 13.6 Other personalservices 0.0 0.1 0.1 0.2 0.2 0.8 0.8 0.8 1.0 1.0 1.1 Adjustments 2.3 3.7 5.8 5.9 5.7 5.7 5.7 6.0 5.1 5.5 5.5 Less: Imputedbank servicecharge -2.3 -1.8 -1.5 -1.5 -1.7 -1.7 -1.7 -1.5 -1.7 -1.9 -2.0 Plus:VAT andother taxes onproducts 4.6 5.5 7.3 7.4 7.4 7.4 7.4 7.4 7.4 7.4 7.4 GrossDomesticProduct bl 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: RwandaNational Institute of Statistics Note: Real values of GDP and its components are obtained by dividing nominal values by their own GDP deflator using revised data; and total may not sum perfectly due to rounding up. 200 10. Annex 2: Overview Of The DeterminantsOf Exports 2.1 The volatility of Rwanda's export receipts has been influenced by both commodity prices and export volumes. However, the volatility o f export volumes has weighed more heavily in the overall trend in export receipts. Although Rwanda's export growth has been relatively strong, averaging 23.3 percent during 1996-99 and 15.6 percent in 2000-05, the pattern o f growth has been erratic, with periods of strength interspersed with periods o f weakness (Annex Figure 1). A mix of domestic and foreign factors hasdetermined these trends. Annex Figure 1: Changes inAnnual RealExports, 1984-2005 (`percentagechange) I100.0 80.0 60.0 40.0 20.0 0.0 -2q` 1 -40.0 1-60.0 -80.0 I I Source: World DevelopmentIndicators,2006; and World Bank staff estimates 2.2 These trends have been influenced,on the domestic side, by the structure of exports, which exposes the agriculture sector to shocks. More than 70 percent o f exports are agricultural commodities, which leads to high exposure to supply-side shocks. Shocks have also arisen from drought and civil unrest. The low level of modern inputs and techniques has also significantly and negatively affected productivity inthe agriculture sector, and therefore impeded exports. 2.3 The level o f volatility and exposure to supply-sideshocks is evident in Annex Figure 2, which illustrates the annual real agriculture GDP and export GDP growth rates over the past 35 years. This is only a crude measure of the fluctuations in agricultural production, but it clearly shows the positive and negative supply-side shocks originating from this sector. The trend in Rwanda's total exports broadly resembles fluctuations in agricultural output over the past nearly four decades, largely due to the concentration of Rwanda's exports inthis sector.132 13' Bond (1987) shows that supply-side shocks havea significantimpacton exports, particularly in developing countries. 201 Annex Figure2: Agriculture supply-sideshock and real exports growth, 1966-2005 (percentage change) 80.0 60.0 40.0 20.0 0.0 -29&b -40.0 ' -60.0 -Agricultural -.-.- -. -80.0 Source: World DevelopmentIndicators,2006; andWorld Bank staff estimates. 2.4 Foreignfactors influencingthe volatility of exports are relatedto the small size and relativeopenness of the economy and its dependenceforeign demand for its products. The price and quantity o f Rwanda's major exports (tea and coffee) are determined by the world market, in which Rwanda (like other developing countries) plays no significant role. Therefore, the demand for Rwanda's exports is affected by fluctuations inforeign income. Rwanda's recent trade policy reforms, which were intended to boost the country's trade relations with its major trading partners, might also have subjected the economy to increased external shocks. The volatility of foreign income could be directly transmitted to Rwanda's exports as a result of trade and financial linkages. A simple correlation between Rwanda's export growth and major trading partners' GDP growth shows that there is a significant and positive relationship o f (0.3). 2.5 While broad movements in export volumes seem to be largely underpinned by supply-side considerations, and to some extent by foreign demand, there is also a correlation with relative prices. In theory, depreciation o f the exchange rate is positively correlated with a growth in exports. A decrease inthe exchange rate means a depreciation o f the domestic currency, which makes the exportable goods less expensive. If the exchange rate depreciates, the demand for exports is likely to increase, and the reverse will occur if the exchangerate appreciates. 2.6 Rwanda's exchange rate has been more or less stable over the past 35 years, although there were some outliers in the late 1960s and mid-1990s that coincided with civil unrest. The data indicate that, to varying degrees, each appreciation of the exchange rate is accompanied by a decline in exports (Annex Figure 3). Correspondingly, depreciation tends to correspond with higherrates of export growth.'33 133 Bayoumi (1990), Khan (1985); andKatao and Falcetti (1999) have also shown the inverserelationshipbetweenthe exchange rate andexports. 202 Annex Figure 3: Exchange rate volatility and real exports growth, , 1962-2005 (Inpercent) 100.0 I Source: World Development Indicators,2006 2.7. An error correction model (ECM) estimation is used to further examine the potential relationshipamong Rwanda's exports, trading partners' income, exchange rate, and agriculturalproductivity.For simplicity, we use the weighted'34average income o f major European trading partners (Austria, Belgium, Denmark, France, Germany, Italy, Netherlands, Spain, Sweden, and the UnitedKingdom). These European countries were selected because they account for a larger share o f Rwanda's exports for more than 30 years. Changes in the real effective exchangerate could be usedas a measure o f price fluctuations between the exporter and importer, but due to lack of sufficient data, changes inthe official exchange rate are used. Finally, the crop production index is used as a proxy for agricultural productivity and other supply-side factors. 2.8 The outcome of the ECM regressionis summarized in Annex Table 3. Inthe short run, the results highlightthe significant impact of volatility inthe agricultural supply-side shock variable on real exports. The previous year's agricultural supply-side shock variable (measured by the crop production index) is the only variable that explains the sharp swings in real exports, and is statistically significant. Other factors that seem to have some influence on the volume o f Rwanda's exports are its competitiveness (reflected by its exchange rate variation) and the economic conditions of its trading partners. However, neithero fthese variables yields statistically significant relationships. 2.9 One of the strongest results to emerge from the long-run analysis is the strong empirical support for a positive relationshipbetween Rwanda's exports and the income of its major trading partners. The estimated equation provides strong evidence o f a co-integrating vector, with the relevant ECM coefficient being very significant. The long-run income elasticity is around 0.53, indicating that over the past three decades, a 10 percent increase in demand from Rwanda's trading partners (predominantly the EU), yields a 5.3 percent increase in exports. It i s also important to note that although the exchange rate did not seem to have a significant short- term impact on Rwanda's exports, the long-run results show the presence o f a strong negative 134 Rwanda's exportsto the respectivecountries, as outlined inthe text. 203 relationship. Specifically, over the long term, a one percent appreciation o f the exchange rate yields a 1.2 percent reduction in exports. These results highlight the importance of having an exchange rate equilibrium that is appropriately valued. Annex Table 3: Error CorrectionModelofthe Determinantsof Exports, 1968-2005 Cointegrating Eq: CointEql CPI -0.07 (-8.42) InEurope GDP -0.58 (-2.65) InEx.rate 1.23 (3.45) Error Correction: In Exports CPI In Europe GDP In Ex. Rate CointEq1 0.43 50.61 -0.14 0.12 (1.64) (6.24) (-0.56) (0.92) InExports(-1) -1.25 -34.83 -0.27 -0.2 1 (-4.63) (-4.15) (-1.06) (-1.54) InExports(-2) -0.67 -12.47 -0.61 -0.11 (-3.62) (-2.18) (-3.56) (-1.24) CPI( -1) 0.03 1.32 0.01 0.004 (2.36) (3.54) (0.47) (0.70) CPI(-2) 0.01 0.06 0.01 -0.002 (1.57) (0.30) (1.37) (-0.5 9) InEurope GDP(-1) 0.32 27.70 -0.62 0.13 (1.38) (3.82) (-2.82) (1.15) InEurope GDP(-2) 0.17 17.22 -0.2 1 0.12 (0.94) (3.02) (-1.22) (1.32) InEx. Rate(-1) -0.52 -3 1.37 -0.43 -0.52 (-1.OS) (-2.08) (-0.94) (-2.16) InEx. Rate(-2) 0.09 -8.52 -0.18 -0.15 (0.24) (-0.77) (-0.55) (-0.86) Constant 0.03 3.91 -0.02 0.002 (0.59) (2.18) (-0.42) (0.07) D94-95 -0.57 -72.35 0.3 -0.11 (-1.12) (-4.58) (0.62) (-0.44) R-squared 0.59 0.77 0.64 0.45 Adj .R-squared 0.44 0.68 0.5 0.23 Source: Datafrom World DevelopmentIndcators2005. Note:Exports= export inconstant US$; CPI=Crop ProductionIndex; Europe GDP= Weighted average realGDP often Europeancountries; Ex. Rate= Official exchange rate; All variables, with the exceptionof crop productionindex, are in logarithms. 204 11. Annex 3: EconomicReforms and TotalFactorProductivity 3.1 During the period 1961-1990, Rwanda hadan administeredeconomy, characterized by the imposition of severe restrictions on trade and foreign exchange transactions, as well as a fixed exchange rate. Inthe early 1 9 9 0 the average tariff rate was 34.8 percent, with five ~ ~ tariff bands ranging from 0-60 percent. Each imported product was subject to a quota, as was each importer. In addition, all import operations were subject to a license authorizing external currency disbursement. Licenses to export were authorized only by the Banque Nationale du Rwanda (BNR). All exports were subject to a preliminary declaration, which obliged exporters to transfer export earnings (expressed in foreign currency) to Rwanda. Export earnings were then transferred to the BNR, which managed them. Likewise, the BNR had to give its prior approval for certain transactions, including medical care, tourist trips, and study abroad. The purchase of currencies to finance these transactions was subject to ceilings. The over-protection o f the economy did not give any incentive to businesses to compete with foreign firms. This lack of competition prevented local companies from innovating and adopting new technologies. As a result, TFP growth fell by an average of 0.21 percent per year during the 1961-1990 period, as shown inTable 3.15, Volume 2 inthe text. 3.2 The period 1991through 1994 marks the beginningof the removalof restrictionson trade and foreign exchange transactions, and the slow establishment of a market economy. While this period covered the first phase of trade reforms, it was also characterized by macroeconomic and political crisis inRwanda. The genocide that occurred in 1994 led inevitably to the destruction of rnanpo~er'~'and capital stock, and a total absence ofthe state, which created uncertainty to trade reforms and unpredictability inthe incentive structure. With uncertainty, the few businesses that survived were reluctant to incur the sunk cost o f adjustment caused by the first wave of trade reforms. The inability o f businesses to adjust, and the uncertainty o f the macroeconomic and political environment, contributed to the poor growth in TFP. Over the period 1991-1994, TFP gowth grew an average o f -11.98 percent per year, or by -4.5 percent when 1994 is excluded (Table 3.15, Volume 2). 3.3 During the period 1995 through 2003, Rwanda embraced the establishment of a market economy, characterizedby a continuationof trade reforms,and liberalizationof the monetary and financial sectors. Trade reforms reduced tariff rates considerably; the average tariff rate was lowered from 34. to 18 percent, and the tariff band was narrowed to four kinds of tariffs, ranging from 0-30 percent. 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 state-owned banks.'36Current account operations (imports, exports, services) were liberalized, and certain restrictions on capital flows were eliminated. For instance, the transfer of capital and revenues related to FDI, and free withdrawal on foreign exchange accounts in 135It is estimatedthat 800,000 people lost their lives betweenApril andJune 1994. 136Rwandahadthree commercialbanksprior to 1995. Duringthe period 1995-2003,three more commercial banks were opened, for atotal of six. 205 commercial banks, were all~wed.'~'The financial reforms included the introduction of a flexible exchangerate inMarch 1995. 3.4 During the same period, the Government further consolidated its commitment to trade, financial, and exchange reforms, and became increasinglycredible and stable. The reforms therefore helped remove some of the distortions, especially in terms o f price subsidies. The credibility and stability of the economic reforms created an environment inwhich businesses adjusted their productive capacities, which improved the competitiveness of the Rwandese economy. As a result, productivity grew, on average, by 7.76 percent per year, as shown inTable 3.15 of Volume 2. 3.5. This change clearly shows an increasing degree of openness of the economy. In particular, imports of goods and non-factor services over GDP, which stood at 14 percent during 1980-1989, rose to 27 percent during 1995-2003. In addition, the share o f capital goods in GDP increasedfrom 3.9 percent during 1980-1989 to 6.5 percent in 1995-2003. 3.6 The share of taxes collected from international trade as a share of total revenues, which stood at almost 47 percent during 1980-89, declined to 23 percent during 1995-2003, due to a reductionand harmonizationof trade tariffs. The Government instead implemented a VAT to replace taxes on international trade and improve tax collection. Aside from exports of goods and non-factor services as a share of GDP,13* all the indicators of trade reforms improved significantly during 1980-1989 and 1995-2003. Exports o f goods and services did not increase following trade liberalization, due mainly to low investment inthe coffee andtea sectors, coupled with a fall inworld prices for coltan. However, total factor productivity, which was negative in the 1980s, becamepositive between 1995and 2003 (AnnexTable 4). Annex Table 4: Indices of Trade Reform Period TFP Imports/ Imports Trade Trade- Exports/ GDP (cap)/GDP ratio tax ratio GDP 1980-1989 -2.4 14.7 3.9 26.7 46.6 12.0 1995-2003 7.8 27.0 6.5 35.2 23.2 8.2 Source: Calculatedby the author, basedon World Bank Database.Note: Trade ratios: Exports and imports of goods and services as a % of GDP; Trade tax ratio: Trade tax over total tax 3.7. The second set of quantitative measures used to examine the relationshipbetween Government policy and growth are the financial indices - broad money (M2)as a share of GDP; credit to the private sector over GDP; and the real interest rate (Annex Table 3.3). Broad money as a share of GDP, which stood at 14 percent during 1980-1989, rose to 16 percent 13' Kanimba(2004). 13*Exportsof goods and non-factor services over GDP, which was expected to increase as trade liberalization progressed, instead declined over the two periods. This counter-intuitive result is due to the fact that the productionof coffee, Rwanda's main export, decreased significantly invalue and volume. There were three main reasons: lack of maintenance of coffee trees; ageing of coffee trees; andrejection by farmers of coffee production, which hadbeen substitutedwith food crop production between 1990 and 1994. 206 during 1995-2003. Likewise, credit to the private sector as a share of GDP rose from 7 percent during 1980-1989 to 9 percent during 1995-2003. However, as would be expected, the real interest rate dropped from 9 percent during 1980-1989 to 8.5 percent during 1995-2003. Financial reforms, particularly a reduction inthe number of non-performing loans, helpedto reduce the cost o f financing compared to levels in the 1980s. Annex Table 5 shows that there was a positive association between the trend inproductivity and reforms inthe financial sector. Annex Table 5: Indicesof Financial Reforms Private Real Period TFP M2/GDP CredWGDP Interest 1980-89 -2.37 14.36 7.2 9.13 1995-03 7.78 16.32 9.2 8.5* Source: World Bank staff estimates. Note: * 2003 3.8 The third set of quantitative measures comprises exchange rate reform indices, which include the real effective exchange rate'39and the exchange rate regime. Rwanda established a fully liberalized and market-determined, (managed float), exchange rate system in 1995. The change from a fixed to a flexible exchange rate coincided with the period when productivity became positive. Improvement in productivity also corresponded to the period of real exchange rate depreciation (Annex Table 6). Indeed, Rwanda adopted a policy aimed at stabilizing the exchange rate through foreign exchange auctions, to maintain the external competitiveness ofthe country's exports. The stability ofthe exchange rate was also supported by a continuous flow o f foreign financial assistance. Official development assistance (ODA) and grants or official aid'40as a share of GDP rose sharply, from an average o f 10.83 percent during the years 1980-1989 to 23.65 percent between 1995 and 2003. These findings also confirm the existence of an association between a change inproductivity and exchange rate reforms. Annex Table 6: Indicesof Exchange Rate Reforms Real Effective Exchange Period TFP Exchange Rate Rate Regime 1980-89 -2.4 102.3 Fixed rate 1995-03 7.8 93.9 Flexible rate Source: World Bank staff estimates. Note: * 2003 13' The real effective exchange rate i s defined as E*(PD/P), where E is the nominal effective exchange rate (the number o f unitso f $ to get one unit o f Franc Rwanda), P D is the domestic price, and PF is the foreign price. Here an increase inthe real effective exchange rate is anappreciation. 140 Official development assistance and official aid over GDP consists o f disbursements of loans made on concessional terms (net o f repayments o f principal) and grants by official agencies o f the members o f the Development Assistance Committee (DAC), by multilateral institutions, and by non-DAC institutionsto Rwanda. 207 12. Annex 4: DeterminantsofProductivityand Employment in the Manufacturing Sector Annex Table 7 : DeterminantsofProductivityacross ManufacturingFirms Independentvariable Dependentvariable=yovern Dependentvariablwoovem Dependentvariableyovem Dependentvariable=voovern (1) (2) (3) (4) (5) (6) (1) (2) (3) (4) (5) (6) Ratiooftotalannual cost of -0.1 2.5 2.0 -0.8 2.2 2.0 energyto totalannual sales 0.0 -0.6 -0.5 -0.2 -0.5 -0.5 Ratio oftotal annualcost of -2.8 -2.4 -2.2 -4.5 -4.0 -3.8 transport to totalannual Sales -0.8 -0.7 -0.7 -1.2 -1.2 -1.1 Ratio oftotal annualcost ofinterest 1.0 0.7 0.8 1.3 1.0 1.0 paymentto total annualSales (2.15)" (1.80)* (1.89)* (2Sl)** (2.22)** (2.13)** Log oftotal annualcost ofenergy -0.1 0.0 0.0 -0.1 0.0 0.0 -1.2 -0.4 -0.5 -1.3 -0.6 -0.6 Log oftotal annual cost oftransport 0.0 0.0 0.0 0.0 0.0 0.0 -0.5 -0.5 -0.5 -0.3 -0.1 -0.1 Log oftotal annual cost of interest 0.5 0.6 0.6 0.6 0.7 0.7 payment (4.88)*** (6.17)*** (5.90)"' (4.60)*** (6.29)*** (6.08)*** Logofcapital(book value of 0.3 0.2 0.2 0.2 0.1 0.1 0.1 0.1 machineryandequipment) perworker (3.93)*** (3.66)*" (3.62)*** (3.37)*** -1.1 -0.9 -1.1 -1.1 Logoftotalnumberof employees -0.2 -0.2 -0.3 -0.3 -0.8 -0.8 -1.0 -1.0 -0.6 -0.6 -1.1 -1.1 (3.64)*** (3.56)*** (4.13)*** (3.99)*** Dummyequals 1ifestablishment -0.1 -0.3 0.1 -0.3 is inthe food andbeveragesindustry; -0.1 -0.4 -0.1 -0.4 zero otherwise Dummy equals 1ifestablishment 0.5 0.3 0.2 -0.1 is locatedin Kigali;zero otherwise -0.7 -0.3 -0.3 -0.2 Constant 9.4 8.8 8.6 8.9 9.0 9.0 4.7 6.7 6.5 3.9 6.3 6.5 (22.13)*** (8.53)*** (7.43)*** (18.72)*** (7.90)*** (6.91)*** (5.16)*** (6.23)*** (5.50)*** (3.91)*** (5.71)*** (5.20)*** Observations 69 69 69 60 60 60 48 48 48 4 4 4 4 4 4 R-squared 0.1 0.3 0.3 0.1 0.3 0.3 0.4 0.6 0.6 0.4 0.7 0.7 Source: RIMS2005 Absolute value o f t-statistics inparentheses * significant at 10%; ** significant at 5%; *** significant at 1% 208 Annex Table 8: Determinants of Employment Growth Dep. Var=employment growth Ratio o f Book value of equipment to 0.04 0.04 total number of employees (in log) -1.66 -1.63 Logof total number of employees 0.18 (3.29)* ** Durnrnyequals 1ifestablishment provides 0.06 0.05 internal training (within the organization) to its -0.45 -0.36 employees; zero otherwise Durnrny equals 1ifestablishment provides extema -0.29 -0.23 (outside the the organization) training to (1.82)* -1.34 its employees; zero otherwise Share of the workforce inthe establishment 0.01 0.01 with only primarylevel education (1.96)* (1.83)' Share o f the workforce inthe establishment with 0.01 0.01 secondary or vocational level education -1.62 -1.32 Share o f the workforce inthe establishment with 0.00 0.00 university or higher1level education -0.33 -0.49 Establishment's manager has a university -0.22 -0.19 level education -1.67 -1.34 Manager's experience inthe industry before -0.10 -0.08 joining the current establishment (inyears) (2.05)** -1.67 Establishment located inKigali 0.38 0.36 (2.16)** (1.92)* Establishment inthe food and beverage industry 0.19 0.20 -1.65 -1.52 Establishment exports at least 5% of its output. -0.18 -0.16 -1.38 -1.16 At least 10% ofthe establishment is owned by f o r e -0.27 -0.22 (1.92)* -1.5 1 Firmage (inyears) -0.13 -0.09 (1.73)* -1.14 Du-y equals 1ifthe establishment had a loan; -0.08 -0.12 zero otherwise -0.59 -0.89 Durnrny equals 1ifthe establishment 0.2s 0.18 hadan overdraft; zero otherwise (1.91)* -1.18 micro (Durnrny equals 1iffirm has less -0.63 than 10 employees; zero otherwise) (2.42)** small (Dumy equals 1if firm has from 10 -0.33 to 49 employees; zero otherwise) (1.98)- m e d i u m (Durnrny equals 1if firm has from SO to -0.38 99 employees; zero otherwise) (2.02)* Constant -1.10 -0.19 (2.70)** -0.45 Observations 54.00 54.00 R-squared 0.50 0.46 Source: Rwanda Industrial andMining Survey, 2005. Note: Absolute value o f t statistics inparentheses; * significantat 10%;**significant at 5%;***significant at 1% 209 13. Annex 5: Export Diversification Annex Figure 4: Over Time, Greater Diversification I s Associated with Lower Export Growth Volatility Source: Chandra et al, 2006. 210 Annex Table 9: Export Diversification, Growth, and Share of Exports in GDP, 1970-2004 1970 1975 1980 1985 1990 1995 2000 2004 Burundi Avg Herfindah1Index 5 years 0.64 0.59 0.62 0.65 0.72 0.62 0.63 0.42 Avg ExportsI GDP 5 years 0.10 0.10 0.13 0.10 0.10 0.10 0.08 0.07 Avg ExportGrowth 5 years -- - 0.01 0.14 0.22 0.02 0.00 0.20 -0.16 -0.07 Share of top 5 exports 0.95 0.93 0.94 0.97 0.91 0.96 0.96 0.93 Correlation- export gr. & HI 1970-04 0.16 Congo, Dem. Rep. Avg Herfiidahl Index - 5 years 0.36 0.35 0.27 0.28 0.32 0.32 0.49 0.49 Avg ExportsI GDP 0.18 0.13 0.13 0.19 0.26 0.20 0.25 0.19 Avg Export Growth 5 years -- 5 years 0.16 0.10 0.13 -0.04 0.07 -0.02 -0.06 0.04 Share of top 5 exports 0.89 0.83 0.90 0.87 0.90 0.89 0.94 0.93 Correlation- export & HI 1970-04 gr. -0.30 Kenya Avg HerfmdahlIndex 5 years 0.14 0.12 0.23 0.21 0.20 0.13 0.11 0.10 Avg Exports/ GDP 5 years 0.30 0.29 0.30 0.25 0.24 0.34 0.27 0.26 Avg Export Growth 5 years -- - 0.07 0.19 0.18 0.00 0.04 0.11 -0.01 0.09 Share oftop 5 exports 0.61 0.62 0.73 0.80 0.69 0.59 0.62 0.60 Correlation- exportgr. & HI1970-04 0.08 Rwanda Avg Herfiidahl Index 0.50 0.43 0.57 0.59 0.60 0.49 0.45 0.35 Avg Exports/ GDP 0.10 0.10 0.16 0.11 0.08 0.06 0.07 0.09 Avg Export Growth 5 years -- - 5 years 5 years 0.14 0.52 0.23 0.02 0.10 -0.17 0.08 0.10 Share of top 5 exports 0.97 0.97 0.98 0.96 0.95 0.90 0.91 0.94 Correlation- exportgr. & HI 1970-04 -0.19 Tanzania Avg Herfidahl Index 5 years 0.10 0.10 0.16 0.18 0.16 0.10 0.10 0.08 Avg ExportsI GDP 0.13 0.16 0.15 0.17 Avg Export Growth 5 years -- - 5 years 0.06 0.09 0.10 -0.07 0.06 0.10 -0.01 0.12 Share of top 5 exports 0.68 0.64 0.66 0.74 0.56 0.64 0.61 0.49 Correlation- exportgr. & HI 1970-04 -0.49 Uganda Avg Herfidahl Index 5 years 0.38 0.50 0.81 0.85 0.79 0.59 0.52 0.19 Avg ExportsI GDP 0.24 0.16 0.15 0.12 0.09 0.09 0.12 0.13 Avg ExportGrowth 5 years -- - 5 years 0.10 0.04 0.18 0.02 -0.10 0.33 -0.11 0.07 Share oftop 5 exports 0.82 Correlation- export 0.95 0.95 0.99 0.99 0.95 0.96 0.88 gr. & HI 1970-04 -0.06 Source: COMTRADE data, basedon Staff estimates. 211 Annex Table 10: BesidesCoffee and Minerals,What ElseDoes Rwanda Export? (Products in total exports with a share of at least one halfpercent in total exports-SITC2 -4 digit) SITC2 ProductName 1976-77 1980-81 1984-85 1990-91 1998-99 2000-01 2002-03 2004 00. Food and live animals 224 Milk & cream,preserved, concentrated __ __ 0.1 __ 0.1 545 Other fresh or chilledvegetables 0.1 0.3 0.0 0.2 0.0 I 0.0 0.0 711 Coffee,whether or not roasted 81.4 69.6 75.2 64.5 50.9 38.0 37.2 42.2 723 Cocoabutter and cocoapaste 741 Tea 6.9 8.3 11.8 11.8 12.5 9.8 4.7 8.5 Subtotal 88.4 78.3 87.1 76.5 63.5 47.9 42.0 50.8 01. Beveragesand tobacco 1110 Non alcoholic beverages,n.e.s. 0.0 0.0 - __ 0.5 1123 Beer made from malt ____ _- 0.0 0.4 0.0 0.0 Subtotal 0.0 0.0 0.0 0.0 0.0 0.4 0.0 0.5 02. Crude materials, inedible, except fuels 2111 Bovine & equinehides (other than calf),raw 3.1 0.3 1.2 0.3 0.8 0.5 0.3 0.1 2112 Calf skins,raw (fresh,salted,dried,pickled) 0.0 0.0 0.6 0.1 0.1 0.5 0.2 2114 Goat & kidskins,raw (fiesh,salted,dried pickled) 6.7 2.3 2.4 2.0 0.4 0.2 0.6 0.5 2116 Sheep & lamb skins with wool oqraw 0.7 0.2 0.3 0.4 0.4 0.1 0.3 0.4 2117 Sheep & lamb skins without the wool,raw 0.0 _- 0.3 0.0 0.0 __ 2224 Sunflower seeds _- 0.0 ____ 2516 Chemical wood pulp,dissolving grade - 0.5 2784 Asbestos - 0.5 0.0 ____ 0.0 2814 Roastediron pyrites __ ______ 0.6 --_- ________ 2876 Tin ores and concentrates 9.5 0.1 2.0 1.4 4.8 3.0 12.6 2879 Ores & concentrat.of non-ferrous metals 3.8 2.3 0.9 6.4 35.8 37.3 28.3 2881 Ash & residues,contain.metals/metalliccompound - 0.1 0.0 2922 Shellac,seed lac,stick lac,gum-resins,etc. 0.3 __ __-- _- 0.0 - __ -- 2924 Plants,seeds,fiuit used inperfumer - 1.8 1.3 0.5 0.0 0.0 0.0 2926 Bulbs,tubers & rhizomes o f flowering or foliage 0.1 0.5 0.9 1.2 0.4 0.4 0.7 0.4 2927 Cut flowers and foliage 0.0 0.0 0.0 0.4 2.3 0.3 0.3 2929 Other materials o fvegetableorigin 0.0 0.6 0.3 1.6 0.6 0.1 0.0 Subtotal 10.9 19.5 8.7 10.2 10.9 44.3 43.6 42.7 03. Mineral fuels, lubricantsetc. 3221 Anthracite,whether/not pulverized __ 0.2 0.6 0.0 Subtotal 0.0 0.0 0.0 0.0 0.0 0.2 0.6 0.0 05. Chemical products 5112 Cyclic hydrocarbons -- -- 0.5 5 156 Heterocyclic compounds/acids 0.0 __ -_ 5221 Chemical elements - ____ 0.3 __---_ 5225 0th.inorg.bases & metallic oxides 0.0 0.1 0.1 5621 MineraVchemical fertilizers 1.o 0.1 0.5 __ 5629 Fertilizers,n.e.s. 0.4 0.0 __ 5831 Polyethylene 0.1 __----____-_ ____ 0.0 0.0 0.0 Subtotal 0.0 0.I 0.0 1.0 0.5 0.0 1.5 0.1 06. Manufacturedgoodsby material 6116 Leather of other hides or skins 0.0 _- 0.1 0.3 6413 Kraft paper and paperboard - __ 0.0 0.3 6418 Paper& paperboard ---- 0.3 6513 Cotton yarn 0.3 ______ 6534 Fabrics,woven,of synthetic fibres ___- ______ 0.4 0.0 1.0 0.7 8439 Other outer garments oftextile fabrics 0.0 0.0 0.0 0.0 0.6 0.0 6612 Portlandcement,ciment,slag cement 0.0 0.0 0.1 0.8 0.4 6672 Diamonds,unwork.cutiotheMise ___- I 0.1 0.2 1.0 6673 0th.precious & semi-precious stones 0.0 0.0 __ 0.0 0.2 0.4 6724 Puddled bars andpilings; ingots etc. __ 0.8 6746 Sheets & plates,rolled ___- 0.1 0.3 6842 Aluminium andaluminium alloys 0.7 0.2 _- 0.0 6871 Tin and tin alloys,unwought 3.1 0.5 0.7 0.2 6891 Tungsten,molybdenum,tantalum & magnesium - 0.0 1.0 6999 Semi-manufactureso ftungsten ____ -- 0.3 ____ Subtotal 0.0 0.1 3.1 0.6 1.4 1.7 6.5 1.9 08. Miscellaneousmanufactured articles 8931 Art.for transportatiodpacking of goods __ 0.1 0.1 0.0 0.0 8947 Other sporting goods and fairground amusements 0.0 0.1 0.0 0.0 0.1 0.7 8960 Works o f art & antiques 0.0 0.0 0.0 0.0 0.1 0.0 Subtotal 0.0 0.0 0.I 0.0 0.I 0.I 0.1 0.7 TOTAL 99.2 98.1 99.0 88.3 76.5 94.5 94.3 96.7 Source: UNCOMTRADE 212 Annex Table 11: Price Trends for SelectedCommodities. 1975-2004 (In unit as indicated) 1976-1989 1995-2004 Bananas, CentralAmerica andEcuador,U.S.importer'sprice, FOB US.ports(#Ab.) 17 21 Aluminiumhighgrade, LME, cash 1403 1512 Tin, LME, cash 11279 5790 Tin, KualaLumpurTin Market, ex-smelter(Mal$/kg) 26 19 Tin, KualaLumpurTinMarket, ex-smelter, 11030 5560 Tungstenore, minimumcontentofW03: 65%, CIF Europe($/mtuWo3) 100 50 Gold, 99.5% fine, afternoonfixing London($/troyounce) 354 331 Source: Annual averages of free- marketprices(1960 2005), UNCTADCommodityPriceBulletin. - 213 Annex Table 12: Shares of Top 3 Products in Total Exports, 1970-2004 (SITCZ -4 digit) 1970 1980 1990 2000 2004 Burundi Coffee 0.83 0.81 0.78 0.79 0.63 Cotton 0.06 -- -- -- -- Hides & skins,-exc.fur skins- undre 0.04 -- 0.05 -- Pearls and precious and semi-precious 0.04 -- -- 0.15 -- Natural abrasives-incl.industria1 d -- 0.03 -- -- -- Tobacco, unmanufactured -- -- 0.03 -- -- Tea and mate -- -- -- 0.11 0.13 Ores & concentrates o f non-ferrous -- -- -- 0.03 -- Share intotal exports 0.93 0.88 0.86 0.93 0.90 DRC Copper 0.62 0.54 0.53 -- -- Miscell.non-ferrous base metals 0.11 0.13 -- 0.05 0.11 Coffee 0.06 0.15 0.11 -- -- Pearls and precious and semi-precious -- 0.16 0.75 0.48 -- Ores & concentrates o f non-ferrous -- -- -- 0.10 0.27 Share intotal exports 0.80 0.82 0.80 0.89 0.86 Rwanda Ores & concentrates o f non-ferrous 0.59 0.10 -- 0.23 0.41 Coffee 0.32 0.75 0.71 0.50 0.42 Tea and mate 0.04 0.07 0.12 0.15 0.09 Hides & skins,-exc. fur skins-ndre -- -- 0.05 -- -- Share intotal exports 0.94 0.92 0.88 0.88 0.92 Kenya Coffee 0.32 0.39 0.21 0.11 -- Tea and mate 0.20 0.20 0.30 0.28 0.16 Crude vegetable materials,nes 0.03 0.06 0.07 0.12 0.17 Clothing except fur clothing -- -- -- -- 0.14 Share intotal exports 0.56 0.64 0.58 0.50 0.46 Tanzania Coffee 0.17 0.3 1 0.22 0.13 -- Spices 0.16 0.13 -- -- -- Vegetable fibres,except cotton and 0.14 -- -- -- -- Cotton -- 0.13 0.18 -- -- Copper -- -- 0.06 -- -- Fish,fresh & simply preserved -- -- 0.24 0.16 Pearls and precious and semi-precious -- -- 0.10 -- -- Gas,natural and manufactured -- -- -- -- 0.09 Tobacco,unmanufactured -- -- -- -- 0.09 Share intotal exports 0.47 0.57 0.45 0.47 0.34 Uganda Coffee 0.55 0.95 0.76 0.56 0.29 Cotton 0.22 0.02 0.05 -- -- Copper 0.10 _- -- -- -- Hides & skins,-exc.fur skins- undre -- -- 0.08 -- -- Fish,fresh & simply preserved -- -- -- 0.14 0.26 Tobacco, unmanufactured -- -- -- 0.09 0.13 Share intotal exports 0.87 0.97 0.89 0.79 0.67 Source: UNComtrade. 214 Annex Figure 5: Trends in Hides, Skins and Leather Exports (In unit as indicated) Figure 5.a :Trends inHides, Skins and Leather Figure 5.b :RegionalExporters of Hides and EXDOrtS. 1976-2004 1'000 us$) Skin to Italy, 1978-2004 ('000 US$) 0000 io000 8000 A 15000 6000 10000 I5000 4000 10000 5000 - 2000 0 - C E l q r o r l v i l u e o f n w r k m andhde COO0 US$) 1978 1981 1984 1987 1990 1993 1996 1999 2002 --Share oftotalsxprts ofrawrkhandkathsr 0 -UgandaRvmda -Kenya .........x Tanzania 1976 1980 1983 1986 1989 1992 1995 1998 2001 2004 +Conw. Dem.Reo. --BIUm& Source: UNCOMTRADE Annex Table 13: SubstantialGrowth inRwanda's Hides and Skins Export Products (In unit as indicated) 1999-2001 2002-2004 Period Change (US$ 000) (US$ 000) (Percent) Bovine & equine hides 375.0 102.0 -72.9 Calf skins, raw 54.0 196.0 260.6 Goat & kid skins, raw 105.0 291.O 176.7 Sheep & lamb skins with wool on, raw 134.0 176.0 31.9 Leather of other bovine cattle and 0.0 15.0 Sheep and lamb skin leather 8.0 41.0 419.0 Leather of other hides or skins 0.0 125.0 Total (hides, skin and leather) 694.0 946.0 36.3 Source: UNComtrade 215 Annex Table 14: IncreasedTrends in Exportsfrom Rwanda to Major World Markets (US$ 000) Hides and Skin Leather 1999-2001 2002-2004 1999-2001 2002-2004 Italy 398 239 8 40 Spain 194 59 0 85 Kenya 14 95 Pakistan 66 88 Greece 64 143 Turkey 0 225 Hong Kong, China 142 79 Total o f all markets 681 764 13 181 Source: UNCOMTRADE. Annex Figure 6: Trends in Horticulture Exports Exports, 1976-2004 ('000 US$) 3000 2500 2000 1500 1000 500 0 - 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 cVegetables ... ...Fruits . -Bulbs,tubers & rhizomeso fflowrin -Cut flowrs and foliage -Other materialso f vegetable origin Source:UNCOMTRADE. 216 Annex Table 15: SubstantialUpward Trend and Growth inRwanda's Horticulture Exports, 1995-2004 ('000 USDollar) 1995-98 1999-01 2002-04 2004 Vegetables 34.6 55.5 4.5 9.0 Fruits 46.9 232.5 121.0 132.3 Fresh or dried h i t s 46.8 232.3 119.1 126.9 of which: Banana 28.7 228.6 93.4 94.8 Prepared fruits (juice, jelly etc.) 0.1 0.2 1.8 5.5 Bulbs, tubers and rhizomes 104.3 282.8 305.9 266.9 Cut flower and foliage 0.4 1252.2 177.0 221.o Other materials o f vegetable origin 990.0 180.2 23.8 28.0 Source: UNCOMTRADE statistics Annex Figure 7: Demand for Horticultural Products in the EUand US Markets I s Growing Figure 7.a: European Market Dermand for Figure 7.b: US Market Demand for Horticulture HorticultureProducts, 1978-2004 ('000 USDollars) Products 1978-2005 ('000 USDollars) 3.00EH7 6.00Et06 2.50EH7 S.OOEffl6 2.0OEtO7 4.00Effl6 1.50Effl7 3.00Effl6 1.00Et07 2 00E+06 1.00Et06 0 00EH0 1978 1981 1984 1987 1990 1993 1996 1999 2002 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 - -Vcgetab.&shchikqfrordpm - - +- -Fresh -1- -Preparedvegetables -a - --r-F~~&nuts(notmcLdoPnutr), ,FNt,p~scrre~and~ipnarati +Vegetab.mots &tukn,prsparedip --c-FreshVegetables --icBulk,tubersandrhizomes fndtr +Prepared -Cut flowcrfrrnts andfoliage -x-Cuthxen atdfohage .(, ,,,,Bubs,tukn Blhaomei o fh n e h ,*, Source: UNCOMTRADE. Annex Table 16: Kenya DominatesRegionalExports to EUHorticulture MarketJ995 and 2004 (in unit as indicated) 1995 Value Percentage of Value Percentage of ('000 U S $) European market ('000 U S $) European market Fresh Vegetables 79314 0.56 209839 1.13 of which: Kenya 70582 0.50 190085 1.02 Fresh Fruits 22866 0.13 35658 0.14 of which: Kenya 20729 0.12 33429 0.13 Bulbs, Tubers and Rhizomes 5795 0.20 54354 1.16 of which: Kenya 5240 0.18 37388 0.80 Cut Flowers and Foliage 111017 3.26 336968 7.21 of which: Kenya I03963 3.06 296789 6.35 Source: UNCOMTRADE. 217 REFERENCES Aersten, George, Augustin Mutijima, Sylvain Mbarubugeye, March 2006. Rwanda Agricultural Policy Note, Background study # 1: Prospects for achieving sustainable agricultural growth inRwanda, Ministryof Finance and Economic Planning, Republic o f Rwanda. Afiica Mining Intelligence. 2004, As quoted in Thomas R. Yager "The Mineral Industry of Rwanda" inU.S. Geological Surveys Minerals Yearbook -2004, pp33.1. Akitoby, Bernardin & Matthias Cinyabuguma. 2004. Sources of Growth in the Democratic Republic of the Congo: A Co-integration Approach. International Monetary Fund, WP/O4/114. ATDT BananaProject -Rwanda" www.isar.ceiar.ordatdtiBanana/banana.htm. Aureille, Yves G. & Serge Rwamasirabo. Spring 2006. Patterns of Competitiveness and Comparative Advantage inRwandaAgriculture. Byakweli. 2004 (incomplete). Chandra, Vandana, I.Osorio, M Desai, and Y. Li. 2006, Export diversification Rwanda too many export discoveries, too little scaling-up (Background paper for the Rwanda Country Economic Memorandum). Chandra, V. (ed.). 2006, Technology, Adaptation and Exports - how some developing countries got it right, The World Bank. Coulibab, Kalarnogo and Tembo R.B. Maburuki. 2006, Cross-border trade in Rwanda (Backgroundpaper for the RwandaCountry Economic Memorandum). Dabalen, Andrew, Stefan0 Paternostro and Gaelle Pierre. December 2004, The Returns to Participation in the Non-Farm Sector in Rural Rwanda, World Bank Policy Research Working Paper 3462. Diao, Xinshen and Yu Bingxin, May 2006. Agricultural Growth and Poverty Reduction Options in Rwanda - Background Study #4 for Rwanda Agricultural Policy Note, International FoodPolicy ResearchInstitute. Diao, Xinshen and YuBingxin. August 2005, Growthand Poverty Reduction Options inRwanda: An Economy wide, Multi-Market Model Analysis for 2004-2015, Second Draft, Development Strategy and Governance Division, International Food Policy Research Institute. Diamond, Jack and Poka Khemani, 2005. `Introducing Financial Management Information Systems in Developing Countries'. IMF Working Paper. International Monetary Fund, October. Diop, Ndiame, Paul Brenton and YakupAsarkaya. 2005, Trade Costs, Export Development and Poverty inRwandaWorld Bank Policy Working Paper 3784. The World Bank. 218 EconomistIntelligent Unit.2006, Country Report, Rwanda, 2006. Elbadawi, Ibrahim. 2002. "Real Exchange Rate Policy and Non-traditional Exports in Developing Countries," in G.K. Helleiner (ed.), Non-Traditional Export Promotion in Africa: Experiences and Issues, Palgrave: New York. Elbadawi, Ibrahim, L. Kaltani and R. Soto. 2006, "Real Exchange Rate Misalignment in SSA: How Serious, How Dangerous,'' UnpublishedWorld Bank Memo (forthcoming). Electrogaz. 2004: Annual Report. Ezemenari, Kene and Kalamogo Coulibaly. 2006A, Total factor productivity growth and economic reforms: Evidence from Rwanda (Background paper for the Rwanda Country Economic Memorandum), World Bank. Ezemenari,Kene and Kalamogo Coulibaly. 2006B, Sources o f growth inRwanda: Four different complementary approaches (Background paper for the Rwanda Country Economic Memorandum), World Bank. Ezemenari, Kene and Ephraim Kebede. 2006, Rwanda's growth Path and Binding Constraints (Background paper for the RwandaCountry Economic Memorandum). Ezemenari, Kene, Kebede Ephraim and Rutayisire Musoni. 2006, Sources o f Inflation in Rwanda, (Background paper for the Rwanda Country Economic Memorandum) Unpublished, World Bank. Ezemenari, Kene, Kebede Ephraim. 2006, The Impact o f Aid on Fiscal Policy in Rwanda Implications for Growth and Poverty Reduction. (Background paper for the Rwanda Country Economic Memorandum. Ezemenari, Kene and Rui Wu. January 2005, Earnings Differences Between Menand Women in Rwanda, Africa RegionWorking Paper SeriesNumber 81,32053, The World Bank. FAO. 2006, "Consultation on Hides and Skins", Arusha, Tanzania, February 1,2006. FAOSTAT 2005 FIAS. 2006a. Government o fRwanda- Sourceso fInformalEconomic Activity. International Finance Company. November. FIAS. 2006. Sector Study ofthe Effective Tax Burden, Rwanda. ForeignInvestmentAdvisory Service. A joint service o fthe International Finance Corporation and the World Bank. January. GoessensandRwamasirabo2006. (incomplete). Agriculture Policy Note BackgroundPaper. Ghura, Dhaneshwar and Benoit Mercereau. 2004. Political Instability and Growth: The Central African Republic, InternationalMonetary Fund, WP/04/80 Governmentof Rwanda. 2006a. The National Accounts for Rwanda-Sources and Methods for GDP Estimates (base year 2001), National Institute of Statistics. December. 219 Governmentof Rwanda. 2006b. RwandaDemographic and Health Survey. National Institute of Statistics, July. Government of Rwanda. 2005. Rwanda's Coffee Industry-Current Situations and Proposal for Action, Report ofthe Coffee Task Force, December. Governmentof Rwanda. 2001. RwandaHouseholdLiving Condition Survey (EICV). Governmentof Rwanda. 2002. Population Census. Haussman, Rodrick and Roberto Rigobon. 2002. An alternative interpretation o f the `Resources Curse': Theory and policy implications, NBERWorking Paper 9424 Haussman,Ricardo, Dani, Rodrik, andAndres Velasco,2005, "Growth Diagnostics," John F. Kennedy School of Government, HarvardUniversity. Imani Development Group. 2006. Support to Rwanda's Accession to the East African Community: Volume 1, Economic Impact Report. May. Imbs and Wacziarg.2003 (Incomplete). International Monetary Fund (IMF).2006. http://www.friendscentre.net/home.html.International Financial Statistics database. IMF,2005. "The Macroeconomics ofManagingIncreasedAid Inflows: Experiences ofLow- Income Countries andPolicy Implications" Available at http://www.imf.org/external/np/pp/eng/2005/080805a.pdf.IMF, Staff reports, various years InternationalTelecommunication Union. Joseph, K.J. 2002. `Growth o f ICT and ICT for Development, Realities o f the Myths o f the Indian Experience'. Discussion Paper No. 2002178. World Institute for Development Economics Research(WIDER).UnitedNations University.August. J.E.AustinAssociates,Inc.2006, Rwanda: Export CompetitivenessDesk Study (Background paper for the Rwanda Country Economic Memorandum). Kanimba, Francois, Deepening the financial market in Rwanda. Power point presentation, September 2004. Kassahun,Rahel.2006, Developmentof Tourism inRwanda: Unleashingthe Sector's Potential to Drive Export Growth (Background paper for the RwandaCountry Economic Memorandum). Kawai, Masahiro and Shinji Takagi, 2001. "Proposed strategy for a regional exchange rate arrangement inpost-crisis East Asia," manuscript, World Bank. Kelly andMurekezi. 2000 (Incomplete.) Klinger, B. andD.Ledermen.2006. Diversification inthe Search for Innovation and Export Booms. Draft Working Paper, World Bank. Lall, S. 2001, "Technology, Competitiveness and Skills," Cheltenham, EdwardElgar. 220 Loveridge,S. 1989. (incomplete). Lull, S. 2001. "Technology, Competitivenessand Skills," Cheltenham,Edward Edgar. Lasbille, Gerard and Jee-Pan Tan.2004, The returns to Education inRwanda, Journal o fAfrican Economies, Vo1.14, 92-116. Linjouom, Mireille. 2006. The impact of the higher oil prices on Rwanda economic (Background paper for the RwandaCountry Economic Memorandum). Madon S. and G. Kiran. 2002. Information Technology for Citizen-Government Interface: A study o f FRIENDS project in Kerala. World Bank Global Knowledge Sharing Program (GKSP). Ministry of Finance and Economic Planning, NationalAccounts, Direction o f Statistics, Republic of Rwanda. Ministry of Finance and Economic Planning, Government Operations 1995-2005, Republic of Rwanda Ministry of Agriculture andForestry. February 2006, Statistics series, Republic o fRwanda. Ministry of Agriculture andForestry. 2004, Statistics series, Republic o f Rwanda. Ministry of Agriculture, and Forestry and BM. May 2004. Utilisation et AccessibilitB des Engrais Inorganiques au Rwanda. Minist2re des Finances et de la Plani3cation Economique.Novembre 2005. Impact de la crise de 1'6nergie sur la croissanceBconomique au Rwanda, RBpublique du Rwanda National Bank of Rwanda(BNR),Republic o fRwandaOCIR CafB. 2005, Republic o fRwanda. Nugawela, P, V. Bicali, and V. Ndagijimana. 2004 "Review and Assessment o f Micro- and Small-scale Enterprises in Rwanda, Background paper for the Country Economic Memorandum Nkama, Arsene Honore Gideon.April 2005, Sources o f Growth and competitiveness in Rwanda, (Background paper for the RwandaCountry Economic Memorandum), World Bank. Obidegwu, Chukwuma. 2003, "Rwanda: The Search for Post-Conflict Socio-Economic Change, 1995-2001." Africa RegionWorking Papers Series, Number 59, World Bank. OCIR Tea. 2005, Republic of Rwanda. OECD. 2003a. `ICT and Economic Growth: Evidence from OECD Countries, Industries and Firms'. Organizationfor Economic Co-operation andDevelopment. OECD Development Centre (?).Providing Low-Cost Information Technology Access to Rural Communities in Developing Countries: What Works? What Pays? OECD Development Centre Working PaperNo. 229. On TheFRONTIER, RwandaNational ICT Usage Survey, August 2003. 221 Peters, Marjolein and Huib Poot. November 2005, The role o f employment inRwanda's Poverty reduction strategy. Labour market, employment considerations, employment policy, ECORYS-NEI, Labour & Social Policy (Background paper for the Rwanda Country Economic Memorandum). Recensementdes activitks lucratives de la ville de Kigali, 2002 (incomplete). Rutayisire, J. Musoni. Monetary Policy and demand for money in Rwanda: application of co- integrationtechniques and the error-correctionmodel (1980-2000), May 2004. S. Eken, Abdourahmane Sarr, Jacques Bouhga-Hagbe and Jerome Vandenbussche, 2005. InternationalMonetary Fund, Morocco SelectedIssues, Country Report No. 05/419. Scetauroute. 2002. Report on Roads. Prepared for the European Union. Paris. Sekkat, Khalid and Aristomene Varoudakis. 1998. Exchange-rate management and manufactured exports inSub-SaharanAfrica, Working PaperNo. 134, OECD Development Centre. US. CensusBureau, Global PopulationProfile. 2002. USAID. 2006, USAIDlRwanda Integrated Strategic Plan 2004-2009, Volume 1, January. Tahari, Amor; Dhaneshwar Ghura; Bernardin Akitoby and Emmanuel Brou Aka. 2004 Sources of Growth inSub-SaharanAfrica. InternationalMonetary FundWP/04/176. Temesgen, Tilahun, Kene Ezemenari, Louis Munyakazi and Emmanuel Gatera "The Rwandan Industrial and Mining Survey (RIMS), 2005: Survey Report and Major Findings" The World Bank and National Institute o f Statistics of Rwanda(NISR); October 2006. TerraCom Communications: httdlwww.terracom.rwl. TheNew Times,"Axis of ICT Success'' May 6,2006. TheRwandanIndustrial andMining Survey (RIMS), 2005. World Bank, 2002. World Bank Development Indicators VDI)2002, World Bank, Washington DC. World Bank, 2003. Nigeria, Policy Options for Growth and Stability, By Douglas Addison, ReportN0.26215-NGA, WashingtonDC: The World Bank. World Bank, 2004. Micro and Rural Finance in Ghana: Evolving Industry and Approaches to Regulation.Findings Newsletter number 234, January. WorldBank. 2005a.RwandaDiagnostic TradeIntegration Study,November WorldBank. 2005b, WorldEconomicIndicators, World Bank: Washington, DC. World Bank. 200.5~.Private Solutions for Inj?astructure in Rwanda, A country Framework Report, the Public-PrivateInfrastructure Advisory Facility and the World Bank Group. WorldBank. 2005. Study on Labor Market Policy in RwandaFinalreport (Background paper for the RwandaCountry Economic Memorandum), November. 222 World Bank. 2006a. Rwanda Agricultural Policy Note, Rwanda Agricultural Markets Overview BackgroundStudy #2 March3 1. World Bank. 2006b. World Bank Development Indicators (WDg database, World Bank, Washington DC. World Bank, 2007. Promoting Pro-Poor Agricultural Growth in Rwanda: Challenges and Opportunities. Report No. 39881-RW. June. Sustainable Development (AFTS3), Country Department 2 (C2), Africa Region. WorldBank, Several Years.DoingBusiness Report,httd/www.doingbusiness.org. WTO:TradePolicy Review Rwanda.Reportby the SecretariatWT/TPR/S/129 13 April 2004. Yongling, Zhong.2004. `Information Services in Rural China - Field Surveys and Findings'. Information Centre, Ministry of Agriculture, People's Republic of China. RAP Publication 2004/03.Food and Agriculture Organization of the UnitedNations Regional Office for Asia andthe Pacific. Bangkok. TheAfrican Growth and Opportunity Act: Achieving Success through the African Global Competitive Initiative", June 2006,www.apoa.pov. 223 224