Report No. 39021-TZ Tanzania Sustaining and Sharing Economic Growth Country Economic Memorandum and Poverty Assessment (In Two Volumes) Volume II: Main Report March 1, 2007 The World Bank The United Republic of Tanzania Poverty Reduction and Economic Management Unit 2 Ministry of Planning, Economy, Africa Region and Empowerment Document of the World Bank Contents Acknowledgments xv Summary xix Acronyms andAbbreviations xxxiii Part 1 Poverty Reduction and Growth: Recent Performance and Prospects 1 Chapter 1 A DecadeofReforms,MacroeconomicStability, andEconomic Growth 3 Economic Growth Since 1990 3 ImprovedMacroeconomic Fundamentals 8 Determinants of Economic Growth inTanzania 14 Conclusions 20 Notes 21 iii Contents Chapter2 The ChallengeofReducingPovertyinTanzania 23 Monetary Poverty: Levels and Changes 23 Economic Inequality, Poverty, and Growth 25 Nonmonetary Poverty Measures 31 Economic Characteristics of the Poor 32 Explaining HouseholdConsumption 40 Conclusions 41 Notes 43 Chapter3 SpatialDimensionsof GrowthandPovertyReduction 45 Overall RegionalIncome Patterns 46 Implications for Regional Policy 54 Notes 55 Chapter4 Outlook on Growthand Poverty Reduction 57 Growth Scenarios 58 ReviewofTanzania's GrowthProspectsinHistorical and InternationalContexts 59 Policy-BasedProjections 60 Input-Based Projections 61 Sectoral Projections 65 Reaching the MDGandNSGRP Targets 66 Conclusions 74 Notes 75 Part 2 Sectoral Perspectives on Growth 77 Chapter5 AgriculturalProductivityand SharedGrowth 79 Removing Constraints onAgricultural Growth 87 Public Expenditures to Support Agricultural Growth 110 Notes 121 Chapter 6 FosteringGrowth,ExportCompetitiveness,andEmployment inTanzania's ManufacturingSector 123 Background 123 Determinants of Manufacturing Sector Growth 125 Enhancing the Export Performance o fthe Manufacturing Sector 132 Conclusions andRecommendations for aManufacturingSector Growth Strategy 135 Chapter 7 TourisminTanzania 139 Economic Contribution of Tourism 140 The Tanzanian Tourism IndustryComparedwith That of Other Countries 143 Growth Potential 145 iv Contents Recommendations 147 Notes 148 Chapter 8 The InformalEconomyin Tanzania 149 Constraints to Growth of Enterprises inthe Informal Sector and Formalization 152 Benefits o f Increasing Formalization 155 Implications for Policy 156 Notes 158 Part 3 Elements of a Strategy for Shared Growth 159 Chapter 9 FosteringInnovation,Productivity,andTechnological Change: DriversofInnovation,Productivity,and TechnologicalChange 161 Education 162 Innovation 171 Informationand CommunicationTechnologies 177 Summary o f Issues and Recommendations 181 Notes 183 Chapter 10 Enhancingthe BusinessEnvironment 185 Scaling Up Access to Infrastructure 187 Policy Recommendations for Scaling Up Investment inInfrastructure 193 Scaling Up Access to Capital andFinance 195 Enhancing the Public-Private Interface 200 Notes 215 Chapter 11 HarnessingNaturalResourcesfor SustainableGrowth: The Contributionof NaturalResourcesto Growth 217 Public Investment inNaturalResource-Based Growth 220 UntappedGrowth Potential 221 Potentialfor Local Spinoff Effects 222 Potentialfor PovertyReduction 223 Sustainability of Growth 225 Externalities 227 Recommendations 228 Chapter12 Enhancingthe Capacityofthe Poorto ParticipateinGrowth 231 Improving HumanCapital of the Poor 231 Building Physical Capital ofthe Poor 248 Dealingwith Vulnerability 255 Conclusions 259 Notes 260 V Contents Part 4 Managing Policies and Expenditures for Shared Growth 261 Chapter 13 ScalingUpPublicExpenditure:NationalStrategy for Growth andReductioninPovertyandMillenniumDevelopmentGoal FinancingRequirements 263 Domestic Resources 265 Scaling UpForeignAid 266 Debt Sustainability 269 Reducing Aid Dependency 271 Conclusions 272 Notes 272 Chapter 14 CoordinationofEconomicPolicyFormulationand ImplementationinTanzania 273 Review of Institutions for Economic Policy 274 Challenges 276 Implementationof the Growth Agenda of the NSGRP 280 Note 281 Part 5 Appendixes: Statistical Tables 283 Appendix A PopulationandDemographics 285 Appendix B The Economy 289 Appendix C Exportsand Imports 311 Appendix D ExternalDebt 327 Appendix E CentralGovernmentRevenueandExpenditure 337 Appendix F MonetarySituation 351 Appendix G AgriculturalProduction 361 Appendix H Employment,Labor,andProductioninthe Manufacturing Sector 369 Appendix I ConsumerPricesand CostofLiving 377 References 381 List ofBackgroundStudies 389 Boxes 1.1 Overview of Structural Reforms inTanzania 4 1.2 Government Spending andEconomic Growth 19 2.1 I s Tanzania's Poverty Line Too Low? 24 3.1 RegionalDifferences inCoping with External Shocks 48 5.1 Organization of the Marketing Chain for Oranges and Onions 109 8.1 Examples of Voluntary Formalization 156 vi Contents 9.1 BenchmarkingTanzania inthe Global Context 162 9.2 The United Nations DevelopmentProgramme's Technology Achievement Index 173 9.3 Constraints to Technology Access inthe Horticulture Sector inTanzania 175 10.1 Aspects o f Governance 201 12.1 GenderDifferencesinEducation 235 12.2 Marketing Opportunities and Crop Adoption 250 12.3 Positive Consequences o f a Shock 252 12.4 A PovertyTrap inShinyanga 256 12.5 Analysis Helps Clarify Whom to Target 258 14.1 Governance Arrangements to Strengthenthe Effectiveness o f Growth-EnhancingInterventions 278 Figures 1.1 Annual Growth o fReal GDP at Factor Cost, 1960-2005 4 1.2 Merchandise Exports: Traditional and Nontraditional, 1990-2004 9 1.3 Government Finance, 1991/92-2004/05 10 1.4 Money and Inflation, 1990-2004 10 1.5 Domestic Credit andInterest Rates 11 1.6 CapitalFormation, 1995-2005 12 1.7 Saving andInvestment, 1990-2005 13 1.8 Exchange Rate andBalance of Payments, 1990-2004 13 1.9 Public andPublicly GuaranteedDebt and Debt Service, 1990-2004 14 1.10 Decomposition o f Economic Growth per Worker into Contribution o fHuman andPhysical Capital Accumulation and Total Factor Productivity (TFP), 1985-2005 16 1.11 Growth Rates o f GDP, Inclusive and Exclusive o f Government Spending, 1990-2005 19 1.12 Contribution o f Public and Private Expenditure to Economic Growth, 1990-2005 20 2.1 Simulated Changes inPoverty, 1992-2002 26 2.2 Growth Incidence Curve: Nationas a Whole 28 3.1 Average GDP by Region, 1996-99 and2000-03 47 3.2 Average Maize andPaddy Yield per Hectare, Fiscal Years 1995-2001 51 3.3 InterregionalOutput DisparitiesinTanzania, 1980-2003 54 4.1 Average Annual Per Capita GDP Growth for Five-Year Periods, 1961-2005 58 4.2 Projections o f GDP Per Capita and Poverty, 2003-25 59 4.3 Average Per Capita Real GDP Growth in 185 Countries, 1994-2003 60 4.4 Average Years o f Schooling in Seven Afiican Countries, 1960-2000 62 4.5 Contribution o f Capital Accumulation to Growth, 2005-25 63 vii Contents 4.6 ProjectedReduction inConsumptionPoverty inTanzania, 2001-15 68 4.7 Projected ReductioninConsumption Poverty inTanzania under Alternative Compositions of Growth, 2001-1 5 69 4.8 Projected ReductioninMalnutrition (Underweight) inTanzania, 2004-1 5 71 5.1 Average Annual Agricultural Growth, 1990-2003 80 5.2 Labor Productivity Levels inTanzania and Comparators, 1990-2002 81 5.3 Labor Productivity Trends inTanzania and Region, 1990-2003 82 5.4 Producer andIntermediaryReturns as a Percentage of the Border Price 100 5.5 Cost Components of Marketing Margins 101 5.6 Cost Components ofProducer Margins 102 6.1 Growth o f Manufacturing Sector Output and Exports 124 7.1 Tourism Receipts inTanzania, 1991-2005 140 7.2 Total Visitor Arrivals inKenya andTanzania, 1996-2004 143 7.3 World Travel and Tourism Council Competitiveness Index, 2004 144 8.1 Size of the Informal Economy for SelectedCountries, as a Percentageo f GrossNational Income 150 8.2 Estimated UnreportedRevenuefor Tax Purposes 154 8.3 Median Value Added per Worker 155 9.1 Adult Literacy Rates, 1970-2002 164 9.2 Average Years of Schooling, 196&2000 165 9.3 Reading Scores and Mathematics Scores 167 9.4 Predicted EarningsinManufacturing Sector Based on ManufacturingFirmSurveys 168 9.5 Marginal Social Returns per Year of EducationBased on Integrated Labor Force Survey 169 9.6 Projected Shortfall of Health Care Workers inTanzania 171 9.7 ICT Infrastructure: Telephones, Personal Computers, andInternet 179 10.1 Cost o f Inefficienciesinthe Business Environment as aPercentage o f Sales, Various Countries 186 10.2 Cost Structure of Firmsby Average Percentageo f Total Costs 186 10.3 Percentageo fEnterprises Rating Problems as Major or Very Severe Constraints on Enterprise Operations and Growth inTanzania, 2003 187 10.4 Effect o f Low Levels of Infrastructure on Economic Growth 188 10.5 Proportion of Rural PopulationLiving within Two Kilometers of an All-Season Road 190 10.6 RealInterest Rates for,T-Bills, Lending,and Saving, 1993-2005 198 10.7 Governance Indicators for Tanzania, 1996-2005 202 10.8 Informality inTanzania andinComparator Countries 204 10.9 Percentageof FirmsInspectedandNumber o f Inspections perYear, by Government Agency 207 viii Confenfs 10.10 Ratingo f Problems by Enterprises inTanzania and inComparator Countries 210 10.11 Enterprises Reporting That Bribes to the Government Affect Their Businesses in Tanzania and Comuarator Countries 212 10.12 Requests for Bribes during Tax Inspections 214 10.13 Interactionwith InstitutionsThat DemandBribes: Micro Enterprises versus Small, Medium, and Large Enterprises 215 11.1 Annual License Revenue and Number o f Foreign Vessels inEEZ, 1998-2004 222 11.2 Income to Ololosokwan Village, Ngorongoro District Council, 1999-2003 223 11.3 Village Incomes from HuntinginLunda-Mkwambi (Game- Controlled Area), Idodi, and Pawaga Divisions, 1996-99 224 12.1 Primary EducationPerformance, 1995-2004 233 12.2 Changes inthe Distribution o f Access to EducationinRural Kilimanjaro, 2001 and 2003 234 12.3 Enrollment in Secondary Schools, 1990-2004 236 12.4 Percentage o fUndernourished Children under Age Five, 1991-2004 237 12.5 Fraction Stunted by Region (1992-99) and Under-Five Mortality Rate per 1,000 LiveBirths (2002) 238 12.6 Current Height o f Childrenby Stunting Status 10 Years Ago inKagera 239 12.7 Nutritional Status o f Childrenby Age, 1999 240 12.8 Effect o f Community Interventions on Average Nutrition Scores inKagera 241 12.9 Infant Mortality, 1988 and 2002 242 12.10 Concentration Curves for Different Health Care Consultations 244 12.11 Birthper WomanAge 15-49, byWealth Quintile 247 12.12 Fractiono fPeople Working as Own-Account Laborers inAgriculture 249 12.13 Disability andOrphanhood Relative to Primary School Attendance 257 13.1 Effective Exchange Rates, 1990-2004 267 13.2 Exports o f Manufactures as a Percentage o f GDP andExports o f Goods and Services, 1990-2002 268 13.3 Multilateral Credit Disbursements andDebt Sustainability: Net Present Value o fDebt-to-Export Ratio, 2006-26 270 Tables 1.1 Real GDP Growth Rates, 1988-2003 6 1.2 Sources o f Growth andProduction, 1990-2005 7 1.3 Structural Change o fthe Tanzanian Economy, 1990-2005 8 1.4 Decompositiono f Tanzania's Growth, 1995-2005: Depreciation of Initial Capital Stock by 0,25, and 50 Percent 15 1.5 Sources o f Growth by Region, 1990-2000 17 ix Contents 1.6 Sources o f Growth: Expenditure, 1990-2005 18 2.1 PovertyIndices, 1991/92-2000/0 1 24 2.2 Poverty Status inTanzania, 1991/92-2000/01 25 2.3 Distribution o f the Populationby Strataby National Quintile, 1991/92-2000/01 27 2.4 Increase inConsumer Prices between 1991and 2001 27 2.5 GiniCoefficient andTheil Index, 1991/92-2000/01 28 2.6 Share o f Inequality Created by Between-Group Differences inTanzania, 1991/92-2000/01 29 2.7 Decomposition o f Change inPoverty inTanzania, 1991/92-2000/01 30 2.8 Food Share by Strata andPoverty Status, 1991/92-2001/01 31 2.9 Households' Perceptions o fProblems with SatisfyingFood Needs inRelation to Actual Poverty Status, 2000/01 32 2.10 HouseholdAsset Holdings by Quintiles inTanzania, 1991/92-2000/01 33 2.11 Housing Quality inTanzania, 1991/92-2000/01 34 2.12 Poverty by Number o f ChildrenAge Five or Younger, 1991/92-2001/01 35 2.13 Poverty by Civil Status o f Heado f Household, 1991/92-2001/01 35 2.14 Level o f Completed Schooling o f Heado fHouseholdby Quintiles, 1991/92-200 1/01 36 2.15 Employment o f Heado f Householdby Quintile and Strata inTanzania, 1991/92-2000/01 36 2.16 Changes inAverage Consumptionper Adult Equivalentby Employment o f Head o fHouseholdandby Strata, 1991/92 and2000/01 37 2.17 Average Consumptionper Adult Equivalent by Employment o f Head o fHousehold inDar es Salaam, Other UrbanAreas, andRuralAreas, 1990/91-2000/01 37 2.18 Index Number o fAverage Consumptionper Adult Equivalent by Employment o fHeado fHousehold, 1991Tanzania Basis 38 2.19 Share of Labor Forceby Employment Sector andby Gender, 1991/92-2000/01 39 2.20 Sector o f Employment o f Spouses Compared with That o f Heads o fHousehold, 1991/92-2000/01 39 2.21 Main Type o f Business by Quintiles, 1991/92-2000/01 40 2.22 Average Amount o f Educationo fHeado f Household, 1991/92-2000/01 40 2.23 RegressionResults, Determinants o f Consumptionfor Households, and CoefficientsinLevels (Regional Dummies Included), 1991/92-2000/01 42 3.1 Distribution o f IndustrialEstablishments, Workers, and Value Added, 2000 and 2004 49 3.2 Regional PopulationDynamics 52 3.3 Selected Indices o fHuman Capital Developmentby Region 53 X Contents 4.1 Projections ofPer Capita GNI and Share of Population below PovertyLine, 2010-25 59 4.2 Policy-Based GrowthProjections 61 4.3 Effect o f Additional Years of Schoolingon Economic Growth 62 4.4 Contribution of Investmentto Growth: Average over 10Years 63 4.5 Growth and Total Factor Productivity inSelectedEast Asian Countries, 1960-94 64 4.6 Overall Input-Based Projections 65 4.7 Structural Transformation, Selected Countries, 1980-98 65 4.8 Scenarios for Economic Growth and Structural Transformation 66 4.9 MDGBaseline, MostRecentEstimate, andTarget 67 4.10 Scenarios for Economic Growth 68 4.11 Reduction inMalnutrition inKagera 73 5.1 Sectoraland SubsectoralContributions to GDP Growth, 1995-2003 80 5.2 Labor Productivity Growth: Contributing Factors 84 5.3 Contributionsof Subsectorsto 5.3 Percent Growth inAgricultural Gross Value of Production, 1995/96-2002/03 84 5.4 Land Use and Potential for Agricultural LandExpansion, Mid-1990s 88 5.5 Inventory of Technologies Coming Out o fthe ResearchSystem inthe 1990s 92 5.6 InstitutionalFramework for SustainableDevelopment of Smallholder Irrigation Systems 97 5.7 Proposed Growth inDevelopment Expenditures for ASDP, 2002103 114 5.8 2005/06Budget Proposals 115 5.9 Budget Ceilings for RuralDevelopment andAgriculture inLine Ministries, 2005/06 116 7.1 Key Tourism Statistics, 1991-2005 141 7.2 Hotel Occupancy Rate 144 7.3 Perception o f Infrastructure Services 145 7.4 Obstacles Encountered inthe Business Environment 146 7.5 Electricity Provision Indicators 148 8.1 Typology of Forms of Enterprise inTanzania 150 9.1 Role o fthe Public Sector inFostering Innovation 172 9.2 ICT Indicators for Tanzania and Comparators 180 10.1 RoadNetwork, February 2004 190 10.2 FinancialVariables as aPercentageo f GDP, 1997-2005 197 10.3 Commercial Bank Lending to Some Sectors, 1997-2005 198 10.4 Doing Business Indicators 205 10.5 Percentageof FirmsInspectedandMedianNumber of Inspections, by Sector 208 10.6 Likelihoodof ReportingThat Bribes Were Neededto Get Things Done, by Size o fEnterprise 213 11.1 Ministry o fNaturalRevenues and Tourism Annual Revenue, 2002/03 and 2003/04 219 xi Contents 11.2 Budgeto fMinistry ofNatural Revenuesand Tourism, as Distributedby Subsector, 2002/03-2003/04 220 12.1 IncreaseinPer Capita Consumption Relative to Households Headedby Individualswith No Education 232 12.2 Benefit-Cost Ratios ofNutritionInterventions 239 12.3 Number ofDays of School or Work MissedBecauseofIllness, by Consumption Quintile 242 12.4 Distance to HealthFacilities, 1991 and2000 243 12.5 Differences inHealth Outcomes, by Quintile 244 12.6 Access to Preventive Health Services, by Quintile 245 12.7 Socioeconomic Aspects ofHealth, by Quintile 245 12.8 Ordinary Least Squares Regression of EducationGap of ChildrenAge 8-1 1 247 12.9 Consumption TransitionMatrix inKagera, 1994 and 2004 250 12.10 Shocks with Major Consequencesfor Well-BeinginKagera, by Quintile 1994-2004 252 12.11 Five Main Causes ofMortalityby Age Group 253 12.12 Access to Savings Services inRural Areas, 1991 and2000 254 13.1 Per Capita MDGInvestment Needs andFinancing Sources, 2006-1 5 264 13.2 PotentialContribution ofDomestic Revenueto Finance MDGs, 2006-25 265 A.1 SummaryofPopulationinTanzania, 1988and2002 286 A.2 PopulationSize by Region, 1967-2002 287 A.3 Dependency Ratios 288 B.1 GrossDomestic Product by Sector at Current Prices, 1995-2005 290 B.2 Share of Gross Domestic Product at Market Prices by Sector, 1995-2005 292 B.3 Share of Gross Domestic Product at Factor Cost by Sector, 1995-2005 293 B.4 Gross Domestic Product by Sector at Constant 1992 Prices, 1995-2005 294 B.5 Annual Growth o f Gross Domestic Product at Constant 1992 Prices, 1995-2005 296 B.6 GrossNational Product by Expenditure at Current Prices, 1995-2004 298 B.7 Share of GrossNationalProduct by Expenditure at Current Prices, 1995-2004 300 B.8 Gross Domestic Product Deflators by Sector, 1995-2005 302 B.9 Annual ChangesinGross Domestic Product Deflators by Sector, 1995-2005 303 B.10 Investment at Current and at 1992 ConstantPrices, 1995-2005 304 B.11 Fixed Capital FormationbyEconomic Activities at Current Prices, 1995-2004 306 B.12 Fixed CapitalFormationby Types o f Assets at Current Prices, 1995-2004 308 xii Contents B.13 Incremental Gross Capital to Output Ratio, 1987-2005 309 B.14 Incremental Fixed Capital to Output Ratio, 1995-2005 310 c.1 Balance of Payments, 1995-2005 312 c.2 Merchandise Exports: Value, 1995-2005 314 c.3 Volume andUnit Price of SelectedMerchandise Exports, 1995-2005 316 c.4 IndicesofValue, Volume, andUnit Price of SelectedExports, 1995-2005 317 c.5 Merchandise Imports: Value, 1995-2005 319 (2.6 Indices of Terms of Trade, 1995-2005 320 c.7 Direction of Trade: Value, 1995-2005 321 C.8 Directionof Trade: Share of Total, 1995-2005 323 c.9 ExternalReserves, 1995-2005 325 c.10 ScheduledDebt Service Payments, 1995-2005 326 D.1 Summary of ExternalDebt, 1995-2005 328 D.2 ExternalPublic Debt, 1995-2005 330 D.3 ExternalPublic Debt Outstanding, Commitment, andDisbursementby Creditor Type, 1995-2005 331 D.4 ExternalPublic Debt Service Paymentsby Creditor Type, 1995-2005 332 D.5 Share of ExternalPublic Debt Service Paymentsby Creditor Type, 1995-2005 334 D.6 Obligationsto the InternationalMonetary Fund, 1995-2005 335 E.1 Summary of Central Government Operations, 1995-2006 338 E.2 Central Government Revenue, 1995-2006 341 E.3 Share of Central Government Revenue, 1995-2006 343 E.4 Central Government Expenditure, 1995-2006 345 E.5 Share of Central Government Expenditure, 1995-2006 347 E.6 Central Government Expenditure by Economic Function, 1995-2005 348 F.1 Monetary Survey, 1995-2006 352 F.2 Change inMoney Supply and Sources of Change, 1997-2005 353 F.3 Balance Sheet of the Bank of Tanzania, 1995-2005 354 F.4 Balance Sheet of Commercial Banks, 1995-2005 356 F.5 Domestic Lendingof Commercial Banks by Economic Sector, 1995-2005 357 F.6 Interest Rate Structure, 1995-2005 359 F.7 Exchange Rate Movement, 1995-2005 360 G.1 Productionof Principal Cash Crops andFoodCrops, 1994195-2004105 362 G.2 SelectedAgricultural ProductionIndices, 1995196-2004lO5 363 G.3 Producer Price of SelectedCash Crops andFoodCrops, 1994195-2004105 364 G.4 Producer Price Index o f SelectedCashCrops andFood Crops, 1994195-2004105 365 G.5 Average Yield per Hectare for Tea and Sisal, 1994195-2004105 366 xiii Contents G.6 Coffee Production, 1994195-2004lO5 367 H.1 Estimated Employment inManufacturingEnterprises, 2000-05 370 H.2 EstimatedLabor Costs inManufacturing Enterprises, 2000-05 372 H.3 Productiono f Selected ManufacturedCommodities, 1995-2005 374 H.4 MineralProduction, 1995-2005 375 I.1 National Consumer Price Index, 1996-2005 378 1.2 Dar es Salaam Cost o fLivingIndex: Middle-Income Group, 1996-2005 379 1.3 Dares SalaamRetailPriceIndex, 1996-2005 380 xiv Acknowledgments This report was prepared in collaboration East Africa), who provided substantive by the Tanzanian Ministry o f Planning, inputs, comments, and support at all stages Economy, and Empowerment (MPEE) and o fthe preparation process. the World Bank, with support from the African Development Bank and various The task manager and principal author o f local and international researchers. It was the report is Robert Utz (Senior Econo- prepared under the overall supervision o f mist). The Government team was initially Dr. Enos Bukuku (Permanent Secretary, led by Arthur Mwakapugi (Permanent Sec- Ministry o f Infrastructure Development, retary, Ministry o f Energy and Mining, previously Permanent Secretary, MPEE), previously Director for Macroeconomics, I Charles Mutalemwa (Permanent Secretary, MPEE), followed by LastonMsongole (Di- MPEE), Judy O'Connor (Country Director rector of Macroeconomics, MPEE). Sub- for Uganda and Tanzania), and Kathie stantive inputs andbackground papers were Krumm(Sector Manager, Poverty Reduc- prepared by Johannes Hoogeveen (poverty, tion and Economic Management Unit for nutrition, and strengthening the capacity o f xv Acknowledgments the poor to contribute to growth); Louise tion [ESRF]); Erik Thorbecke, Steven Fox and Marianne Simonsen (poverty as- Younger, and David Sahn (Cornel1 Uni- sessment); Meera Shekar (nutrition); Henry versity); and Dani Rodrik (Harvard Uni- Gordon (agriculture); Anuja Utz and Jean- versity). During the consultations held in Eric Aubert (education, innovation, and Dar es Salaam, Haidari Amani (ESRF), information and communications technol- Brian van Arkadie (ESRF), Amon Mbelle ogy); Michael Wong (business environ- (ERB), Peter Noni (Bank o f Tanzania), ment); Ravi Ruparel (financial sector Adolf Mkenda (University o f Dar es Sa- development); Vandana Chandra (manu- laam), and Robert Mbelle (University o f facturing sector analysis); Philip Mpango Dar es Salaam) served as discussants for (spatial dimensions o f growth); Allister the draft background papers. Detailed Moon (institutional reform); and Em- comments were also provided by mem- manuelMungunasi (data and statistical an- bers of the World Bank country team, in- nexes). Peter Mwanakatwe (of the African cluding Keith Hinchliffe (education); Di- Development Bank) contributed the analy- eter Schelling (transport); Karen Rasmus- sis o f infrastructure. Mary-Anne Mwakan- sen, Arun Sanghvi, and Duncan Reynolds gale and Arlette Sourou provided dedicated (energy); Mavis Ampah (information and logistical support. Arlette Sourou was also communications technology); Mathew responsible for word processing and physi- Glasser (decentralization and local gov- cal productiono fthe report. ernment); and Denyse Morin (institutional reforms). The governments o fAustria, Denmark, the Netherlands, and Sweden financed the The Country Economic Memorandum preparation o f background papers by An- benefited also from participation in the nabella Skof (consultant on tourism); "growth path" project led by Roberto Kirsten Pflieger (COW1 consultants on Zagha, in cooperation with Harvard Uni- natural resourcebased growth); Meine versity, which helped to sharpen the Pieter van Dijk and Klaas Schwarz growth diagnostic. (United Nations Educational, Scientific, and Cultural Organisation Institute for Preparation o f the report included several Water Education, for analysis o f the role rounds o f consultations in Tanzania or- o f urban areas); and Jorgen Levin (con- ganized by the President's Office- sultant on computable general equilibrium Planning and Privatization. In initial con- analysis). InfoDev financed a study on sultations in September 2003 and July growth, competitiveness, and information 2004 the team defined and agreed on the and communications technology carried scope and focus o f the study as well as on out by OTF Group consultants. Appendix collaborative arrangements. The main A presents the list of background studies mission took place in November 2004 prepared for this report by local and inter- and included field visits to Kigoma, national researchers and by World Bank Lindi, and Mtwara. InMarch 2005, a se- staffmembers. ries o f workshops in Dodoma, Morogoro, Moshi, and Dar es Salaam were organized Insightfuland challenging comments were to obtain feedback and input on the draft provided at various stages o f the prepara- background papers before the drafting o f tion process by peer reviewers Benno the main report. At that stage, the Presi- Ndulu (World Bank); Josephat Kweka dent's Office-Planning and Privatization (Economic and Social Research Founda- also organized a review meeting with xvi Acknowledgments permanent secretaries and senior officials and recommendations. The team would from a large number o f ministries for a like to express its sincere gratitude to all briefing on the consultations and a dis- who have provided valuable comments cussion o f the emerging main messages andinputduringthe consultations. xvii Summary Tanzania's National Strategy for Growth What must be done to sustain eco- and Reduction o f Poverty (NSGRP) em- nomic growth that i s pro-poor? phasizes the importance o f fostering eco- nomic growth for poverty reduction. It sets The report presents evidence from the mac- an ambitious target o f 6 to 8 percent an- roeconomic, sectoral, and firm and house- nual economic growth to achieve rapid hold levels that sheds light on these ques- reduction in poverty. This report focuses tions. This summary provides an overview on three issues that are central to the suc- o fthe main findings andrecommendations. cess o f Tanzania's poverty reduction efforts: What Factors Explain 0 What factors explain Tanzania's re- Tanzania's RecentAcceleration cent acceleration ineconomic growth? in Economic Growth? 0 Has the accelerated economic growth The average annual growth o f Tanzania's translated into reduced poverty? gross domestic product (GDP) of 6.0 per- xix Summary cent during 2000 to 2005 has been high, The transition o f Tanzania to a market not only compared with its own historical economy began in the mid-1980s with an growth performance but also compared initial focus on the liberalization o f the with international growth rates. Growth economy through the removal o f con- rates increased across all sectors, with in- straints on private sector activities and the dustry growing by 8.7 percent, services by abolition o f controls on prices and ex- 5.9 percent, and agriculture by 4.8 percent change and interest rates. The reforms also during the same period. Mining (growth included a restructuring o f the public sec- rate o f 15.2 percent), construction (10 per- tor and an ambitious privatization pro- cent), manufacturing (7.0 percent), and gram. Inthe mid-1990s, the reform agenda trade hotels and restaurants (6.9 percent) was augmented by a strong focus on mac- were the fastest-growing sub-sectors. The roeconomic stability and the quality o f contribution o f the various sectors to public financial management. Initially, this growth, which depends on both the growth effort involved sharp cuts in government rate o f the sector and its share in the econ- expenditures to minimize the govern- omy, shows that agriculture contributed ment's domestic and nonconcessional bor- 2.3 percentage points, services 2.1 per- rowing. These cuts served as the basis for centage points, and industry 1.6 percent- a prudent monetary policy that reduced the age points o f the average annual growth o f rate of inflation to well below 10 percent. 6.0 percent during 2000 to 2005. The Subsequently, reform efforts focused on analysis o f the sectoral contributions to the improving Tanzania's tax system and pub- increase in the average GDP growth rate lic financial management to improve allo- from 2.5 percent during 1990 to 1994 to cative and operational efficiency o f public 6.0 percent during 2000 to 2005 confirms expenditures and to minimize resource that growth accelerated in all sectors. leakages. An important result o f prudent Growth in the service sector contributed monetary and fiscal policy, combined with 1.4 percentage points to the increase, in- financial sector reforms, i s the recovery o f dustry 1.3 percentage points, and agricul- credit to the private sector which grew by ture 0.8 percentage point. more than 30 percent annually in recent years. The environment for economic The implementation o f a comprehensive set growth is thus vastly improved, and cur- o f macroeconomic and structural reforms rent government efforts are targeting laid the foundation for the recent growth higher levels o f investment inhuman capi- acceleration. These reforms enhanced the tal and physical infrastructure, improve- incentives for private sector activities and ments in the business environment, and ledto improved efficiency o fresource allo- strengthening o f government capacity. cation anduse inthe economy. The domes- tic and foreign private sectors as well as The intensification of reforms since 1995 Tanzania's development partners reacted to and improvements in the business environ- the improvements in the economic and in- ment, as well as sector-specific reforms- centive regime in a variety o f ways that ex- especially in the mining sector-have trig- plain the increase in economic growth. A gered an increase in foreign direct invest- central element o f Tanzania's recent ment (FDI) and aid inflows. FDI has in- growth performance is large inflows o f pri- creased rapidly since the mid-1990s and vate and public capital that were triggered reached about US$542 million or 5 percent bythe reformsundertakenbygovernment. o f GDP by 1999, partly driven by large Summary investments in mining and privatization- inDar es Salaam-as the result o f various related investments. Following the comple- factors. These include the liberalization o f tion o fmajor investments inthe miningsec- the economy, the tolerance o f many infor- tor and the major privatizations, FDI de- mal sector activities that were previously clined to US$375 million or 2.5 percent o f illegal, the need for laid-off government GDP by 2005, a level that i s still high in workers and migrants to generate new in- comparison with that o f most other Afkican come-earning opportunities, and the in- countries. The sectors that received the bulk creased demand for informal sector prod- o f the FDIshowed the highest growth rates, ucts and services as a trickle-down effect including mining, manufacturing, and trade from growth inthe formal economy. and tourism, which together attracted about 75 percent ofFDIduring 1999to 2001. Another significant economic develop- ment during the past decade has been the The reforms implemented by the govem- rapid expansion o f mining and gold ex- ment also triggered a continuous increase in ports, whose share in total exports in- aid inflows that, together with improved creased from 4 percent in 1998 to 56 per- domestic revenue collection, supported the cent in2005. However, the contribution o f increase in government spending from 16 mining to overall growth was only 0.4 percent o f GDP in 1999/2000 to 26 percent percentage points, reflecting the relatively in 2005/06. National accounting statistics small size o f the mining sector, as well as suggest that this increase in government the high import dependence o f the sector spending contributed significantly to the ac- for machinery and its very limited domes- celeration in economic growth. Inthe short tic backward and forward links. Aside term, the increased demand for goods and from gold, fish, and tourism, the value o f services by the government led to increased exports remains low and volatile. Between use o f available capacity. For example, the 1995 and 2001, the real effective exchange rehabilitation and expansion o f administra- rate appreciated by almost 50 percent and tive, economic, and social infrastructure are then returned to its 1995 level. The real reflected in the fast growth o f the construc- appreciation has had a significant influ- tion sector by about 10percent annually dur- ence on the competitiveness o f Tanzania's ing2000 to 2005. Fast growth inthe service tradables sector. Although merchandise sector i s also partlyrelatedto increased gov- exports declined during 1995 to 2001, they ernment expenditures. In addition to these started a recovery in parallel to the recent direct effects o f increased government real depreciation. To date, exports other spending, traditional multiplier effects trans- than gold and fish have played a relatively late increases in government spending into small role as a dynamic source o f growth increased demand for goods and services in and learning and have seen little diversifi- all sectors. In the medium to long term, if cation. Thus, a key challenge for the Tan- government spending contributes effectively zanian economy i s to strengthen its export to the buildingo f humancapital and the ex- competitiveness. Doing so would ensure pansion o f economic infrastructure, then that, aside from the dynamic growth ef- sustained levels o f increased government fects o f a strong export sector, exports will spending have the potential to expand the provide an important demand stimulus for productivecapacity o fthe economy. the economy, especially because the scope for continued increases o f government A noteworthy development is the rapid spending as the primary demand stimulus growth o f the informal sector-particularly i s clearly limited. xxi Summary The analysis o f factor inputs suggests that to 14 percent by 2005, reflecting increased the acceleration ineconomic growth i s not investor confidence in response to sus- so much grounded in a rapid expansion o f tained implementation o f investor friendly human and physical capital but is primar- reforms andincreased demand. ily due to an increase incultivated land in the agriculture sector and increased factor Drawing on the review o f Tanzania's recent productivity for the other sectors. The in- growth performance, the report assesses the crease in total factor productivity reflects prospects for sustained highgrowth and the both increased capacity use in response to key challenges that need to be addressed. increased aggregate demand and economic Policy-based growthprojections suggest that efficiency gains in the wake o f the re- growth of 6 to 8 percent per year i s feasible. moval o f economic distortions. Innovation However, some o f the factors behind the and technological change have so far recent growth acceleration are unlikely to be played only small roles in improving Tan- sustainable inthe medium to long term. The zania's total factor productivity, mainly in demand-side impulses o f foreign aid and the form o f FDI but also as some encour- government spending depend on ever- aging innovations emerging from the agri- increasing amounts o f aid and government cultural research system. At the firm level, spending. There is also a clear limit to the there i s some evidence that the structural extent that agricultural production can be reforms have resulted in a more dynamic increased solely by increasingthe landunder and competitive private sector. Increased cultivation. Signs o f environmental and so- competition in the private sector is evi- cial stress (especially between pastoralists denced by an increasing number o f firms and agriculturalists) o f increased land use exiting and entering the market. The fact already exist in some areas o f Tanzania. that firms entering the market are typically Similarly, the effect o f reform-induced effi- more competitive than those that exit i s an ciency gains on economic growth will di- important driver o f the increase in total minishwhen the higher level of efficiency factor productivity registered at the aggre- has beenreached. gate level. Thus, for Tanzania to achieve sustained Although the contribution o f human and high growth, increases in government physical capital accumulation to economic spending and expansion of landunder cul- growth has been relatively small, the recent tivation need to be gradually replaced by increases in school enrollment can be ex- increased productivity, saving, and in- pected to be reflected in higher economic vestment by the private sector as primary growth inthe future. Public investment has drivers of growth. Sustained economic recovered from an average o f about 3 per- growth will depend on the ability o f the cent o f GDP during the late 1990s to about economy to diversify and to increase its 8 percent o f GDP in recent years. The international competitiveness. Diversifica- analysis of public investment suggests, tion requires efforts both to enhance the however, that only about one-third o f it was capacity to innovate and to find new areas used on public infrastructure such as roads o f economic activity where Tanzanian en- or electricity, while the remainder was de- terprises can successfully compete. En- voted to the rehabilitation and expansion o f hancing international competitiveness re- administrative and social infrastructure. quires measures that enhance productivity Private sector investment hadbeen stagnant and reduce the cost o f doing business at at about 11percent until2002 but increased the microeconomic leveI and macroeco- xxii Summary nomic policies that ensure a competitive tures o f all income groups grew at about the exchange rate as well as interest rates and same pace, probably because growth in ag- access to capital that are not distorted by riculture, which i s the source o f income for highpublic demand for funds. most o f the poor, was similar to growth in other sectors during 1991 to 2000. Since Has the Accelerated Economic 2000, growth in the industry and service Growth Translated into sectors has been higher than in the agricul- Reduced Poverty? ture sector, which may have caused an in- crease ininequality. Sustained economic growth i s critical to achieving progress in poverty reduction. The Household Budget Survey data show The mechanisms through which the poor large regional differences in poverty reduc- contribute to and participate in economic tion. Although poverty dropped fiom 28.1 growth include the following: percent to 17.6 percent inDar es Salaam, in other urban areas poverty declined only 0 Increased incomes from the main from 28.7 percent to 26 percent and inrural sources o f livelihood o fthe poor areas from 40.8 percent to 38.7 percent. The faster pace o f poverty reduction in Dar es 0 New income-generating opportunities for the poor Salaam reflects Tanzania's pattern o f growth. In particular, Dar es Salaam ac- 0 Reduced vulnerability to shocks that counts for about 50 percent o f the FDIstock affect the incomes o f the poor and flows, and as the seat o f central gov- ernment and most donor agencies, it also 0 Increasedgovernment revenue for pro- poor expenditures benefits disproportionately fiom the increase inaidinflows. Althoughcentral government 0 Increased private transfers and strength- expenditures increased fiom 18 percent to ened social safety nets. 25.6 percent o f GDP between 2000 and 2005, transfers to local governments in- Inaddition, the report examines the effec- creased only fiom 2.9 percent to 3.3 percent tiveness o fmeasuresthat support the poor in o f GDP during that period. Growth in the efforts to accumulate human and physical formal sector in Dar es Salaam also sup- capital, which would enhance their pros- ported an increase inthe size andincomes in pects o fcontributing to economic growth. the informal sector, which contributed sig- nificantly to poverty reduction during the Modest per capita GDP growth rates during periodfrom 1991/92 to 2000/01. the early and mid-1990s resulted inequally modest poverty reduction. In 2001, gov- More than 80 percent o f Tanzania's poor ernment estimates show 35 percent o f the derive their livelihoods from agriculture. population living in poverty. The potential Between 1991 and 2000, the agriculture effect o f the recent GDP growth accelera- sector grew by an average o f 3.5 percent, tion has not yet been captured in available which suggests per capita growth o f less poverty data. Ownership o f assets such as than 1percent. The increase inper capita improved housing, radios, and bicycles by expenditure by farm households i s equally the poor has also increased. The expansion modest at 7.3 percent during the period o f access to free primaryeducation has also from 1991/92 to 2000/01. Nonetheless, clearly benefited the poor. The analysis o f because most o f the poor derive their live- growth incidence suggests that expendi- lihood from agriculture, this modest in- xxiii Summary crease explains more than half o f the total Dar es Salaam, funding credit to the pri- decline in poverty observed during that vate sector and government in Dar es Sa- period. Between 2000 and 2005, growth in laam, as well as overseas investments by the agriculture sector accelerated to an av- the banking sector. Although the mobility erage o f 4.8 percent annually, which ac- o f human and financial resources toward cording to poverty simulations i s likely to opportunities where the returns are highest have generated a further drop inrural pov- i s supportive o f high economic growth in erty. The study argues that given Tanza- Tanzania, measures that counteract an in- nia's agricultural potential, there i s signifi- creasing marginalization o f the rural poor cant scope for reducing poverty by in Tanzania's growth process are needed. measures that would foster growth in agri- Such measures would include enhanced culture andthus the incomes o ffarmers. rural access to quality education and poli- cies that support agriculture. Another path out o f poverty i s the move- ment from agriculture to other sources o f The report highlights instances in which income, possibly combined with migration the lack o f integration o f rural areas in the from rural to urban areas. Data suggest economy significantly reduced rural that the shift from agriculture to nonagri- growth. Key among these instances i s the cultural activities in rural areas has been access o f rural areas to markets as well as an important contributor to poverty reduc- to agricultural inputs. For example, sur- tion. Informal sector activities have been veys carried out in the Kilimanjaro and an important entry point for the poor to Ruvuma regions suggest that lack o f ac- engage innonagricultural activities. Rural- cess to agricultural inputs results in low urban migration has also contributed to agricultural productivity and, conse- poverty reduction. However its quantita- quently, limited progress in rural poverty tive significance was less than that o f the reduction. This limited access to agricul- other channels, probably because most of tural inputsi s the result o f two equally im- the migrants are from households above portant problems: (a) limited access to in- the poverty line. However, migration i s put credit and @) lack o f a rural input only one path in which fast urban growth supply infrastructure that would allow can benefit the poor inrural areas. Indirect farmers to purchase these inputs. channels include higher demand for rural products, wage effects, and transfers. These results suggest that rural develop- However, the fact that rural growth and ment and informal sector activities are the poverty reduction lag significantly behind primary direct drivers o fpoverty reduction urban growth and poverty reduction sug- inTanzania, where the informal sector has gests that these links are still weak. been an important transmission mecha- nismthat allowed the poor to participate in Nevertheless, urban-rural links can also economic growth opportunities originating result in a deepening o f urban-rural differ- inthe formal andpubic sectors. This inter- ences. There i s evidence that rural mi- pretation i s reinforced by the fact that al- grants are typically better educated than though economic growth was significantly the average rural population, leading to a higher inurban areas than inrural areas in widening o f the education gap as these the period from 1990/01to 2000/01, mod- migrants move from rural to urban areas. est rural growth has clearly dominated the A large share o f financial savings col- faster urban growth with respect to its ef- lected by banks inrural areas flows toward fect on poverty reduction. Furthermore, xxiv Summary even in an environment o f relatively high ditures been pro-poor. In other sectors, growth differences between rural and ur- such as water, primarily nonpoor house- ban areas, the contribution o f migration holds benefited from improved quality and and other urban-rural links to poverty access to services. The focus on social ex- reductionhas beenrelatively modest. penditures also limited the availability o f funds for growth-enhancing expenditures. Appropriate tax and public expenditure policies play an important role infostering What Must Be Done to shared growth. As the report highlights, Sustain Economic Growth enhancement o f the domestic revenue base through sustained economic growth i s cen- That Is Pro-Poor? tral to the sustainable financing o f public The review o f Tanzania's recent growth expenditures and reduction o f aid depend- performance suggests that enhancing the ence in the medium to long term. In turn, pace of structural change and diversifica- tax policies have a direct influence on the tion and increasing the international com- level o f investment and economic activi- petitiveness of the economy remain the key ties. Similarly, public expenditures play an challenges for sustaining growth. The pov- important role not only in improving the erty analysis highlights the importance o f a environment for economic growth, but productive agriculture sector and o f a con- also in enhancing the access and quality o f ducive environment for the activities o f public services for the poor. small and medium enterprises (SMEs) as key elements o f a shared-growth strategy. It Overall, the Tanzanian tax policy i s as- also emphasizes that participation o f the sessed as being sound and not inimical to poor in the growth process requires that growth. However, several measures could policies support the accumulation by the enhance the contribution o f the tax system poor o f primarily human capital and physi- to fostering shared growth. First, there re- cal and financial capital. Finally, the report mains an urban bias in the tax system: ef- underscores the importance o f appropriate fective tax rates are higher for farmers policies and institutions to manage the de- than for businesses, which are mostly ur- sign and implementation o f a shared- ban. In particular, the crop cess collected growth strategy, as well as resources for by local authorities imposes a relatively their implementation. Here the report dis- heavy tax burden on agriculture. Second, cusses not only the management o f public the presumptive tax regime for small busi- finances, but also the equally important nesses i s one of the most sophisticated in management o f natural resources. Im- the region. Nonetheless, it i s regressive for proved governance o f Tanzania's natural small businesses that do not keep records. resources, strengthened capacity to ensure Third, there are significant weaknesses in that tax and expenditure policies are sup- the taxation o f natural resources, which portive o f a shared-growth agenda, andbet- result inboth distortions to the sustainable ter institutional coordination framework for exploitation o f natural resources and development and implementation of a suboptimal collection o frevenue. growth strategy require attention in Tanza- nia's quest for sustainable shared growth. Social sector expenditures have seen sig- nificant increases in recent years. How- To sustain and share economic growth ever, incidence analysis suggests that only across all income groups o f society, Tanza- inthe education sector have public expen- niawill need to preserve achievements, con- XXV Summary solidate ongoing reforms, and strengthen potentially harmthe efficiency o f markets institutional capacity for both policy advice and reduce farm incomes. andprogram implementation. It will also be important to guardagainst backsliding inthe Scaled-up investment in agricultural re- face o f pressures fiom vested interests search and a reform o f Tanzania's exten- or impatience with the pace o f poverty sion service have important roles to play in reduction. supporting farmers in the move to raising crops that yield higher returns. The study The sectoral distribution o f growth has a highlights the large productivity losses at- significant effect on the pace o f poverty tributable to human diseases that make reduction and inequality. Tanzania's com- health intervention an important element parative advantage in agricultural produc- o f efforts to increase agricultural produc- tion and its large potential to enhance agri- tivity. Inaddition to improving farm-level cultural productivity provide a good basis productivity, the focus needs to be on for a focus on agriculture and agriculture- promoting upstream activities such as ag- related activities as the central element o f roprocessing and enhancing links to do- its efforts to reduce poverty. Agricultural mestic and foreign markets. Efforts toward activities are the source o f livelihood for 75 raising agricultural productivity must also percent o f the population, more than 40 take into account the general question o f percent o fwhom are poor. Efforts to reduce the impact that risks have on agricultural poverty must focus on measures that will activity in general and the more specific help the poor to (a) generate more income consideration o f the capacity o f the poor to from their current agricultural products, (b) grow out o fpoverty. shift their production to more profitable agricultural products, and (c) shift to in- Increased income-generating opportunities come-generating opportunities outside o f in nonagricultural activities, especially in agriculture inboth rural andurban areas. In rural areas, are also important for poverty addition, a decline in prices attributable to reduction and for the medium- to long-term productivity increases benefits poor people structural transformation o f the economy. who are net buyers o f agricultural products. Providing opportunities for Tanzanians to move out o f the agriculture sector can be Increasing agricultural incomes requires expected to improve the labor productivity policies that target both improvements in inthe sector and to provide higher incomes market access and increases in agricultural for those moving out o f the sector. SMEs, productivity. Enhancing rural infiastruc- often in the informal economy, provide an ture remains critical to ensure market ac- important entry point for the poor to engage cess for farmers. It i s equally important to in industrial and service sector activities. ensure that institutional arrangements are Measures that support SMEs are thus im- inplace that link farmers to domestic and portant. These measures include facilitating international markets. An example o f re- informal sector activities such as easier ac- forms in this area would be to ensure that cess to credit and public recognition and the crop boards function efficiently, with a support for informal sector activities, in- clear separation o fpublic andprivate func- stead o fthe frequently observed harassment tions, and that they are accountable to o f informal sector operators. Formalization farmers. Regulations such as mandatory should be primarily incentive based in the auctions and single license rules for coffee case o fmicro enterprises. xxvi Summary The report highlightsthe importance o f the private sector can readily support. This ad- manufacturing sector as a potential dy- vice may seem tantamount to recommend- namic driver o f diversification and growth. ing everything-that is, redressing all bani- Analysis o f firm-level data from the enter- ers to production presently facing all prise survey suggeststhat, inorder o fprior- manufacturing firms in Tanzania. But it is ity, the leading five factors that affect firm not. To circumvent the high financial and growth andthat deserve special attention by time costs and the government's weak im- policy makers are (a) access to and cost o f plementation capacity, the report recom- financial capital; (b) access to technology mends focus and pragmatism in catering to to improve productivity; (c) infrastructure, existing and potential exporters. A sound especially energy; (d) skilled labor; and (e) strategy for delivering physical inputs (such the regulatory environment for business as infi-astructure) andfinancial inputs (which activities. will make bank finance more accessible) includes the identification o f spatial loca- Because firm growth i s intricately tied to tions where export activity is most prevalent growth in exports, an aggressive and pro- and where exporters most likely to locate. active policy stance promoting manufac- The government i s pursuing this strategy tured exports is likely to have the greatest throughexport processing zones andspecial effect on manufacturing growth in Tanza- economic zones. However, in rolling out nia and i s recommended. The rationale for this strategy, it will be important to sequence this selective approach is motivated by these activities by initially addressing prob- today's global reality: if a firm cannot lems o f existing zones before new ones are compete in the global market (that is, if it created. A review o fTanzania's exportproc- cannot export), it i s unlikely to survive too essing zones highlights infiastructure long in Tanzania's domestic or Africa's weaknesses, especially reliable access to regional markets, which are flooded with electricity and water, as a main constraint cheaper imports from low-cost and high- for firms located in those zones (World skills producers such as those from East Bank 20050. The targeted improvement o f and SouthAsia. The policy implications o f infiastructure services to the manufacturing an export-oriented stance have several sector thus needs to be a priority in Tanza- overlaps with factors that promote growth nia's efforts to spur growth and structural in nonexporting firms, but an aggressive transformation. Spatial targeting helps in focus on incentives that facilitate the ex- targeting exports. This approach would ren- pansion o f existing firms and promote new der public support in a financially feasible entrants in the export sector i s likely to and timely manner for fast-growing export- yieldthe most benefits. ers and potential new entrants into the ex- port business. Priority areas that require improved policies or scaling up o f expenditures include invest- Sustained economic growth will increas- ing in infrastructure; enhancing access to ingly depend on the capacity o f economic finance, notably for the rural and SME sec- actors to innovate, to produce an increased tors; and building an effective interface be- array o f goods and services, and to accel- tween the public and private sectors, be- erate the pace o f technological change. It cause the economy currently suffers from will require foremost a greater focus on ineffective regulation, bureaucracy, and cor- investment in human resource develop- ruption. Focus on these issues promises in- ment (and vast improvement in secondary, creased economic activities inareas that the technical, and tertiary education), as well xxvii Summary as strengthening o f the innovation envi- convergence in per capita incomes across ronment and Tanzania's fledgling infor- regions. However, in most regions, the mation and communication technology growth performance remains intimately (ET) infrastructure. linked to the fortunes o f individual crops. Several policy lessons emerge from the Tanzania's natural resource endowment analysis o f subnational growth patterns. could be an important source o f growth and First is the importance o f capturing local poverty reduction. Strengthening the gov- knowledge to l l l y exploit growth opportu- ernance of natural resource use and the nities. This lesson suggests that decentrali- backward and forward links to other activi- zation not only i s a means for improved ties is critical to ensure that Tanzania bene- service delivery, but also has an important fits fiom the exploitation o f its natural re- role to play in the implementation o f Tan- sources. In addition, improved governance zania's growth strategy. Local knowledge arrangements are important for the sustain- about growth opportunities i s critical for a able exploitation o f renewable resources series o f public interventions to foster such as fish or forests and are necessary to growth, ranging from infrastructure in- minimize the impact o f negative external- vestments to targeted, crop-specific inter- ities such as that o f commercial fishing and ventions in the agriculture sector. Second, miningontheir artisanal counterparts. local governments have a significant effect A shared-growth strategy requires also a on the business environment, ranging fiom their attitude toward the informal sector to focus on the capacity o f the poor to con- local tax policy and administration, licens- tribute to-and participate in-economic ing, land management, and so forth. Fi- growth. This strategy includes opportuni- nally, attention to subnational growth i s ties to build human capital through meas- important to identify successful strategies ures that center on equitable access to, and that can be scaled up andreplicated inother improved quality of, education (primary, secondary, and technical), nutrition, and parts o f the country. The study argues that the regional distribution o f public invest- health services; to reduce the burden of ment should be determined by the growth communicable diseases; to improve child nutritional status; and to reduce maternal opportunities, whereas distributional objec- tives should be primarily pursued through mortality by helping women achieve their targeting o f access to education. desired family size. Other measures in- clude supporting household savings and As the report highlights, Tanzania has investment through development o f ap- been successful in establishing a sound propriate finance and ancillary institutions, basis for sustained and shared economic especially in rural areas. Limited social growth through the implementation o f a protection measures can also play a role in broad reform agenda. Tanzania's strategic mitigating the impact o f large shocks that frameworks for growth and poverty reduc- may create poverty traps for the poor. tion, including the National Development Vision 2025, the Medium-Term Plan for Regional differences in economic perform- Growth and Poverty Reduction, and the ance reflect not only inherent differences in economic potential, but also factors such as NSGRP, adequately identify core interven- past investments in infrastructure and hu- tions that are needed to sustain economic growth. These interventions include the man resources, local governance, and con- strengthening o f economic infrastructure, nectivity. Data suggest some degree o f scaling up of human resource development xxviii Summary from an initial focus on primary education priate policy and regulatory frame- to secondary and higher education, and the works in place, which would support implementation o f reforms to strengthen a greater participation by the private the business environment. In many key sector in selected areas o f infrastruc- areas, specific reform programs are in ture development, operation, and place, including the Business Environment maintenance. Strengthening Program in Tanzania, the Agricultural Sector Development Pro- 0 Devote greater attention to fostering gram, the Primary and Secondary Educa- structural transformation. To sustain its tion Development Program, and the economic growth, Tanzania will in- Second-Generation Financial Sector Re- creasingly need to broadenthe range o f form Program. The report broadly en- goods andservices it produces. The role dorses this reform program. The report o f government in the process o f struc- highlights three elements o f the reform tural transformation i s not to identify agenda that deserve increased attention: new growth opportunities, but rather to support the identification and exploita- tion o f opportunities by the private sec- 0 Enhancing international competitive- ness and accelerating diversification tor. This effort will require scaling upo f investment inhigher education, with the 0 Makinggrowth pro-poor availability o f skilled labor already be- ing a constraint to economic growth. 0 Managing policies and resources for shared growth. FDI and the import of technology are the primary sources o f new technology Enhancing International for Tanzania. However, a strengthening Competitivenessand Accelerating o f Tanzania's research and develop- Diversification ment systems, especially in agriculture, also has an important role to play inthe Enhancing international competitiveness adaptation and dissemination o f new and accelerating diversification with a fo- technologies. Finally, greater access to cus on macroeconomic management, in- ICT is an important tool to accelerate frastructure, access and cost o f credit, the the acquisition o f technology and regulatory environment for private sector knowledge. Inaddition, microeconomic activities, human resource development, evidence supports a direct link between the innovation environment, and the roll- greater use o f ICT, especially cell out anduse o fI C T involves the following: phones, and the productivity o f farms and rural enterprises. In addition to 0 Resolve infrastructure bottlenecks. Recurrent energy shortages are the economywide support, there is also some scope for more direct support at most visible constraint to economic growth. However, general underin- the sector and fm levels. An example vestment in the development and o f such support would be the estab- maintenance o f transport infrastruc- lishment o f a matching grant scheme for the introduction o f business activi- ture, especially in the rail sector and for rural roads, also holds back ties that are new to Tanzania. growth, particularly in rural areas. In 0 Develop an urban strategy. Urban ar- addition to scaled-up investment, eas play an important role in the eco- there i s an urgent need to get appro- nomic growth process. Thus, an impor- xxix Summary tant challenge i s not only to sustain the source envelope have accrued primar- good performance o f Dar es Salaam, ily to central government agencies but also to enhance the role o fregional and to a much lesser extent to local urban centers as hubs o f economic ac- governments. Progress in the devel- tivity. Such an urban strategy would opment o f a sound intergovernmental define the role o f urban areas in Tan- fiscal framework i s a necessary pre- zania's economy, as well as the neces- condition to strengthening the devel- sary investments and an appropriate opment, operation, and maintenance institutional and fiscal framework that o f local infrastructure and delivery o f would allow the implementation o f local services, which, in turn, directly such an urbanstrategy. affect rural growth andpoverty reduc- tion. Fourth are government policies Making GrowthPro-Poor and expenditures to support agricul- Makinggrowthpro-poor highlightsreforms ture. Here the report suggests that that would accelerate growth o f the agricul- public expenditures on research and ture sector; expand income-generating op- extension or irrigation can play an portunities in micro-, small-, and medium- important role in supporting the sec- size enterprises; and strengthen the capacity tor. However, scaling up the public o f the poor to participate and contribute to support for agriculture needs to be economic growth: grounded in a careful analysis o f the effectiveness o f expenditure pro- grams. Finally, a better integration o f 0 Ensure that reforms beneJit rural ar- eas. To date, the effect o f reforms un- rural areas in the growth process will dertaken has been visible primarily in require an effort to draw effectively urban areas, especially Dar es Salaam, on local knowledge in the develop- while rural areas have seen much ment and implementation o f regional less systemic improvements in growth and district growth strategies. This ef- performance and poverty reduction. fort will require an increased focus o f The report highlights five areas in Tanzania's decentralization program which reform programs should scale on economic growth, in addition to up the focus on rural areas. First, im- the provision o f social services. proved access to markets i s central to 0 Encourage micro-, small-, and me- increasing real incomes o f the rural dium-size enterprise activities. Micro population. Such access requires in- enterprises, mostly in the informal vestments inroad infrastructure and in economy, not only are an important storage and market facilitation (stan- source o f income for many Tanzani- dards and business climate).The sec- ans, but also are an important entry ond area i s financial sector reform. point for Tanzanians into entrepreneu- Despite the overall progress made, ru- rial activities. The primary focus ral areas remain largely cut o f f from should be on facilitating and support- access to financial services, which, in ing such activities. Facilitating the turn, has been identifiedas a key con- transition to the formal sector i s im- straint to both farm and off-farm ac- portant, but the move should be vol- tivities in rural areas. The third area i s untary unless the primary motive for public expenditure reform. In this being in the informal sector i s clearly area, the increases in the overall re- tax evasion. Summary Devote greater attention to the effect to coordinate the formulation and im- of social expenditures. The report plementation o f a shared-growth (pro- highlights that the substantial increase poor) strategy in Tanzania. This effort in public expenditures since the late includes the ongoing strengthening o f 1990s has directly benefited the poor the budget process to examine and pri- to only a limitedextent. This situation oritize investments in infrastructure. calls for a revision o f spending priori- Given the large regional differences in ties and targeting approaches to en- productive potential, strengthening in- sure that public expenditures do bene- stitutions that are able to respond to fit the poor. these differences will be critical. Managing Policies and Resources Strengthen analytic underpinnings for Shared Growth and participation in the design of growthpolicies. The NSGRP suggests Managing policies and resources for a very welcome greater focus on eco- shared growth means strengthening insti- nomic growth, supported by a scaling tutions to develop and implement a growth up o f expenditures that would support strategy and to harness both public fi- accelerated growth, ranging from in- nances and natural resources toward the creased infrastructure investments to objective o f sustained shared growth: subsidies for agricultural inputs or credits to targeted segments o f the Strengthen the capacity for the im- economy. Inaddition, the government plementation of Tanzania's reform is implementing a range o f specific agenda. As highlighted throughout the programs, such as Tanzania's Mini report, Tanzania has made great strides Tiger Plan 2020, the Property and in establishing a vastly improved eco- Business Formalization Programme, nomic and incentive regime for eco- and the National Economic Empow- nomic activities. However, inmany ar- erment Policy. For these measures to eas where appropriate policies and be effective, they must be grounded in regulations are in place, their effec- a solid analytic basis that supports the tiveness i s hindered by limited imple- choice and prioritization of specific mentation capacity. Enhancing imple- strategies and expenditure policies. mentation capacity will rely on the Strong processes for stakeholder in- continued implementation o f public puts, adequate governance arrange- sector reforms, but the increased use o f ments for expenditure programs at the the private sector, when appropriate, i s sectoral level, and a strong monitoring an important way to realize reform ob- and evaluation system are o f central jectives. Examples range from an in- importance. Although these issues creased role o f the private sector inthe have received significant attention in provision of agricultural support ser- the social sectors, they are yet to be vices to the contracting in o f capacity fully developed for growth-enhancing at the National Audit Office. expenditure programs. Strengthen the management of the Ensure the egectiveness of govern- growth process. The report highlights ment interventions through appropri- the importance o f developing and ate governance arrangements. The strengthening institutions that are able report highlights a range o f govem- xxxi Summary ment interventions and supports the as fish or forests. Strengthening gov- government's proactive efforts that ernance can minimize the impact o f complement the focus on creating an negative externalities such as com- enabling environment for private sec- mercial fishing and mining on their tor activities. Recent government in- artisanal counterparts. Finally, im- terventions-such as targeted credit proved governance should aim at an guarantee schemes, export processing equitable sharing o f natural resource zones, and targeted agricultural subsi- rents. dies-are innovative efforts with the potential to foster growth, but they also carry significant risks. These Issuesfor Further Study risks include the potential for govern- Finally, it is important to highlight a set o f ance problems that could make such issues that are raised in the report, but that interventions ineffective or even need further analytic work to informpolicy: counterproductive. The report sug- gests a range o f measures that would 0 A better understanding o f private sec- improve the likelihood o f success o f tor consumption and saving behavior, government interventions. because private saving has remained relatively low and has thus limited the 0 Pay greater attention to the manage- ment of natural resources. Tanzania's domestic accumulation of assets natural resource endowment could be 0 A better understanding o f population an important source o f growth and dynamics and their effect on eco- poverty reduction. Governance o f nomic growth natural resource use, as well as the backward and forward links to other 0 A better understanding o f the rela- activities, must be strengthened if tively large movements inthe real ex- Tanzania i s to benefit from the exploi- change rate and their effect on eco- tation o f its natural resources. More- nomic performance and trade over, governance arrangements must 0 A better understanding o f economic be improved for the sustainable ex- developments at the regional and dis- ploitation o f renewable resources such trict levels xxxii Acronyms and Abbreviations AIDS Acquired immune deficiency DADG District Agricultural Develop- syndrome ment Grant ASDP Agricultural Sector Develop- DADP District Agricultural Develop- ment Programme ment Plan ASDS Agricultural Sector Develop- DANIDA Danish InternationalDevelop- ment Strategy mentAgency BCI Business Competitiveness DAWASA Dar es Salaam Water and Sew- Index erage Authority CBO Community-basedorganization DHS Demographic and Health CDTT Centre for the Development and Survey Transfer o f Technology DRD Department o f Research and CET Commonexternal tariff Development CPIA Country Policy and Institutional EAC East African Community Assessment EEZ Exclusive Economic Zone xxxiii Acronyms and Abbreviations EWURA Energy and Water Utilities NEER Nominal effective exchange Regulatory Authority rate FDI Foreign direct investment NGO Nongovernmental organization GCI Global Competitiveness Index NPV Net present value GDP Gross domestic product NSGRP National Strategy for Growth GLS Gray leaf spot andPoverty Reduction GNI Gross national income ODA Official development assistance HBS Household Budget Survey OECD Organisation for Economic Co- HIPC Heavily IndebtedPoor Coun- operation and Development tries (Initiative) PASS Private Agricultural Sector HIV Humanimmunodeficiency virus support ICOR Incremental capital output ratio PBRs Plant Breeder's Rights ICRG International Country Risk PC " Personal computer Guide PEDP Primary Education Develop- ICT Informationand communica- mentProgram tion technology PORALG President's Office-Regional ILD Instituto Libertady Democracia Administration and Local Government ILO International Labour Organization PRSP Poverty Reduction Strategy Paper IMF International Monetary Fund R&D Research and development I S 0 International Standards Organization REER Real effective exchange rate I T Informationtechnology ROSCAs Rotating saving and credit associations KAM Knowledge Assessment Methodology S&T Science and technology KEI Knowledge Economy Index SACCO Savings and credit cooperative MAC MinistryofAgriculture and SACMEQ Southern and Eastern Africa Cooperatives Consortium for Monitoring Educational Quality MAFS Ministry ofAgriculture and Food Security SADC South African Development Community MBICU Mbinga Cooperative Union SEDP Secondary Education Devel- MCM MinistryofCooperatives and opment Program Marketing SMEs Small and mediumenterprises MDG MillenniumDevelopment Goal SPILL Strategic Plan for Implementa- MNRT MinistryofNaturalResources tion o fLandLegislation and Tourism TAFOPA Tanzania FoodProcessors MTEF Medium-Term Expenditure Association Framework Technology Achievement Index MVIWATA Mtandao wa Vikundivya Wa- TAI kulima Tanzania TANAPA Tanzania National Parks MWLD MinistryofWater andLive- TANESCO Tanzania Electricity Supply stock Development Corporation NARS National Agricultural Research TANROADS Tanzania National Roads System Agency xxxiv Acronyms and Abbreviations TASISO Tanzania Small Industrialists TVET Technical and vocational edu- Society cation andtraining T A Z A U Tanzania-Zambia Railway UNCTAD UnitedNations Conference on TBS Tanzania Bureauof Standards Trade and Development TCR4 Tanzania Communications UNDP UnitedNations Development Regulatory Authority Programme TFP Total factor productivity UNIDO UnitedNations Industrial De- TICTS Tanzania International Con- velopment Organization tainer Terminal Services USAID U.S. Agency for International TNBC Tanzania NationalBusiness Development Council VAT Value added tax TPA Tanzania Port Authority VIBINDO Vikundivya Biashara Ndogon- TRC Tanzania Railways Corporation dogo (Small Industries and Petty Traders Association) TRCHS Tanzania Reproductive and Child Health Survey WUA Water user association TTCL Tanzania Telecommunications ZRDC Zonal research and develop- Company Limited ment centers PART 1 Poverty Reduction and Growth Recent Performance and Prospects CHAPTER 1 A Decade of Reforms, Macroeconomic Stability, and Economic Growth Economic Growth Since 1990 cessive governments. The reversal of the negative long-term trend and the gradual Per capita gross domestic product (GDP) acceleration in growth are consistent with growth was negative inTanzania duringthe the gradual broadening and deepening of first halfof the 199Os, but it has accelerated reforms that has taken place and the cau- subsequently and reached 4 percent in re- tious but steady private sector response to cent years. Looking at long-term trends those reforms. The temporary slump in (figure 1.1)' we see that 1985 was a turning economic growth during the early 1990s point in the secular trend in Tanzania's was caused by a weakening o f macroeco- economic performance, when Tanzania nomic policies. This slowdown highlights embarked on a market-oriented reform pro- the fragility of growth in poor countries gram. Tanzania's economic performance such as Tanzania and the importance o f since 1985 provides a strong endorsement sustaining macroeconomic stability, a cor- for the policies pursued since then by suc- nerstoneo fTanzania's growth strategy. 3 PovertyReduction and Growth Figure 1.1. Annual Growth of Real GDP at Factor Cost, 1960-2005 (a) Real GDP (b) Real GDP per capita 14 0 7 30 1 _____I 12 0 25 10.0 2 0 ~ 8.0 I a.oJ I Source:Authors calculations, based on United Republic of Tanzania various years. In the mid-l990s, Tanzania resumed its including privatization o f state-owned en- reform course with a clear and sustained terprises, liberalization o f the agriculture commitment to macroeconomic stability sector, efforts to improve the business en- through sound fiscal and monetary policies vironment, and strengthening o f public as the foundation for economic growth. expenditure management (box 1.1). Those Macroeconomic stabilization was accom- reforms have resulted in sustained high paniedby wide-ranging structuralreforms, growth. Box 1.I. Overview of Structural Reforms in Tanzania Financial Sector During the past decade, the financial sector has seen significant change. Frombeingthe sole pre- serve o f state-owned financial institutions, it has gone through a process o fprivatizationand has been opened up to new entrants. The two largest state-owned banks have been successfully pri- vatized: a 70 percent stake o fNational Bank o f Commerce and a 49 percent stake o fNational Mi- crofinance Bank. Those sales have contributed to increased efficiency and competition inthe banking sector and to narrowed interest spreads and a fairly rapid increase incredit to the private sector. Financial sector reforms also involved the strengthening o fthe legal and regulatory framework, including for microfinance, as well as a strengthening o f supervision by the Bank o f Tanzania. The recent establishment o f a credit rating agency i s a further step inenhancing the efficiency o f financial intermediation inTanzania. The mainchallenges for the sector include fur- ther reduction ininterest spreads and enhanced access to credit by the private sector, especially in rural areas. The government has prepared a comprehensive plan-the Action Plan for Second- GenerationFinancial Sector Reforms-and has begun implementing it. Parastatal Sector Tanzania has been aggressively implementing its privatization agenda. The privatization o f manufacturing and commercial parastatal entities was virtually complete by 2000 and a solid success. continued 4 A Decade of Reforms, MacroeconomicStability, and Economic Growth Box 1.I continued The reform and development o f public private partnerships inthe infrastructure sector proved much more difficult. At the end o f the 1990s, Tanzania launched the privatization o f its infra- structure enterprises. By 2003, five key infrastructure enterprises had some form o f private participation: Tanzania Electricity Supply Corporation (TANESCO) Tanzania International Container Terminal Services (TICTS) Dar es Salaam Water and Sewerage Authority (DAWASA) Tanzania Telecommunications Company Limited (TTCL) Air Tanzania, the national airline However, by early 2007, only the private sector participation inTICTS was still inplace and con- sidered to be successful, whereas private sector participation inDAWASA and inAir Tanzania hadbeen dissolved and that o f TANESCO was not renewed as a result o fperformance. As for the two railroads, Tanzania Railways Corporation (TRC) was undernegotiation, andnegotiations regarding TAZARA had not yet started, as was the case for the harbor authority (Tanzania Port Authority, or TPA). Trade Policies and Institutions Reforms o f trade policies have taken place mainly inthe context o fregional agreements, includ- ingthose o f the South African Development Community (SADC) and the East African Commu- nity(EAC). Tanzania adoptedthe commonexternal tariff (CET) ofthe EAC inJanuary 2005 and loweredits average tariff from 13.8 to 12.3 percent; however, it further raisedthe dispersion o f protection. The lowering o fthe maximumrate o f the CET from the current 25 percent to 20 per- cent, as is expected to happen by 2010 in accordance with the Customs Union Protocol, should helpcorrect some o fthe dispersionofprotection. Onthe export side, the mainissue pertains to export taxes. International experience has shown that export taxes and bans have generally failed to achieve industrial development objectives, have ledto informal trade, and have frequently hurt smallholders, who receive lower prices as a result. Factor Markets: Labor and Land The revisionof land and labor legislation is complete, with most o f the emphasis onthe reform of institutions. The main challenge now is to implement the new legislation. Inthe case o f landre- form, the implementation process is expected to be lengthy and costly. Infrastructure: Power Sector and Transportation The establishment o fthe executive agency, Tanzania National Roads Agency (TANROADS), with responsibility for the trunk roadnetwork has been a major step forward for the transportation sector. However, a clear separation o f the responsibilities o f the Ministry o f Works, TANROADS, and the districts is needed. The overlap o fresponsibilities hampers effective road maintenance and development activities. The current formulation o f a new RoadAct offers a real opportunity to establish a more appropriate policy and institutional framework and to provide the basis for accelerated infrastructure development. continued 5 Povetty Reduction and Growth Box 1.1continued Detailedwork on the restructuringo fthe power sector hasbeen carried out, but the implementation o f the restructuringhas beendelayed, partly as a consequence o f developments inthe international energy market. The performance o f TANESCO has improved significantly under a performance- based management contract. Nonetheless, reform o f the policy andinstitutional framework for the power sector i s essentialto ensure the effectiveness o f future investments inthe sector. Public Institutions Interfacing with the Private Sector Redtape, corruption, and overly burdensome regulatory and licensingrequirements are among the mainconstraints to private sector development inTanzania. The government has started reviewing regulations, focusing on removing obstacles and reorganizing the most important tasks o f government. Inpractical terms, this effort requires (a) harmonizationo f local govern- ment taxation to remove excessive tax burden on private enterprise, (b) streamlining o f work permit procedures, (c) review and amendment o f licensinglegislation to reduce the cost o f business establishment and continuation, (d) review and revision o f export-import procedures to reduce time costs and corruption-related costs, and (e) design and implementation o f a pro- gram for enhancing access to commercial courts by small and mediumenterprises. Tanzania has reformed the legal framework for regulatory institutions. The effectiveness o f those newly established regulatory institutions, especially given the current oversight arrangements, needs to be closely monitored. With an average growth rate o f 5.2 percent economic growth, a relatively wide gap between 1998 and 2003, Tanzania's per- remains when growth i s measured on a per formance was close to that of South Asia capita basis. That gap reflects different (5.4 percent) and South East Asia (5.6 per- rates o f population growth. On a per capita cent) (table 1.1). However, though Tanza- basis, Tanzania grew by a respectable 2.7 nia was able to catch up to the South and percent. However, South and East Asia East Asian nations with respect to real grew by 3.6 and 4.6 percent, respectively. Table 1.I.Real GDP Growth Rates, 1988-2003 Annual real GDP growth - Annual real per capita GDP (%.) growth(%) Countriesand regions 1988-93 1993-98 1998-2003 1988-93 1993-98 1998-2003 Countries CBte d'lvoire 0.3 13.0 -0.3 -3.1 9.7 -1.8 Ghana 4.6 4.3 4.4 2.0 1.5 2.4 Kenya 3.1 2.5 1.o 0.1 0.0 -1.2 Tanzania 3.6 3.0 5.2 0.4 0.1 2.7 Uganda 6.0 7.6 5.8 2.1 4.7 3.1 Regions East Asia and Pacific 8.6 7.7 5.6 6.9 6.4 4.6 Latin America and the Caribbean 2.1 3.6 1.2 0.2 1.9 -0.3 South Asia 5.3 5.7 5.4 3.2 3.7 3.6 Sub-SaharanAfrica 1.3 2.9 3.0 -1.4 0.3 0.6 Source: World Bank, World Development Indicators database. 6 A Decade of Reforms, Macroeconomic Stability, and Economic Growth Table 1.2. Sources of Growth and Production, 1990-2005 Average annualgrowth rate Average contributionto growth (%I (%I Type of economic activity 1990-94 1995-99 2000-05 1990-94 199599 2000-05 Agriculture 3.I 3.6 4.8 1.5 1.8 2.3 Crops 3.2 3.9 4.8 1.I 1.4 1.7 Livestock 2.5 2.7 4.1 0.2 0.2 0.3 Forestry and hunting 2.8 2.4 4.0 0.1 0.1 0.1 Fishing 3.4 3.7 6.7 0.1 0.1 0.2 Industry 2.0 5.4 9.0 0.3 0.9 1.6 Mining and quarrying 11.8 14.8 15.2 0.1 0.2 0.4 Manufacturing 0.4 4.6 7.3 0.0 0.4 0.6 Electricity and water 4.0 5.7 4.4 0.1 0.1 0.1 Electricity 4.5 6.3 4.5 0.1 0.1 0.1 Water 0.8 1.9 3.5 0.0 0.0 0.0 Construction 2.2 3.5 10.3 0.1 0.2 0.5 Services 1.9 3.8 6.I 0.7 1.3 2.1 Trade, hotels, and restaurants 2.0 4.5 7.1 0.3 0.7 1.2 Transportation and communica- tion 3.6 4.8 6.0 0.2 0.2 0.3 Financial and business 2.9 3.6 4.5 0.3 0.4 0.4 Finance and insurance 2.6 3.5 3.9 0.1 0.1 0.1 Real estate 3.0 3.7 4.8 0.2 0.2 0.3 Business 3.6 4.5 5.6 0.0 0.0 0.0 Public administration and other 1.9 1.6 4.1 0.2 0.1 0.3 Public administration 0.6 -0.2 2.8 0.0 0.0 0.1 Education 4.9 4.2 6.6 0.1 0.0 0.1 Health 3.9 3.6 5.8 0.0 0.0 0.0 Other 4.7 6.0 5.3 0.1 0.1 0.1 Less financial services (index measured) 5.7 3.4 3.2 -0.3 -0.2 -0.2 Total GDP (factor cost) 2.5 4.0 6.0 2.5 4.0 6.0 Source: Authors calculations, based on United Republic of Tanzania various years. Sectoral growth rates have accelerated contribution to economic growth remains across the board from 2000 to 2005 (table small. The manufacturing sector has 1.2). Industry has been the most dynamic started to recover, growing at an average sector. The construction sector grew by an o f 7 percent per year from 2000 to 2005. average of 10 percent during that period. This strong performance i s partly attribut- Growth o f agriculture averaged 4.8 per- able to public investment ininfrastructure, cent from 2000 to 2005-which i s more but investment in residential and business than 1percentage point higher than during structures has also increased. Gold mining the period from 1995 to 1999. Within ag- expanded rapidly as several gold mines riculture, fishing was the most dynamic started production. However, its overall sector. However, crops remain the main- 7 Poverty Reduction and Growth Table 1.3. Structural Change of the Tanzanian Economy, 1990-2005 Average annual growth rate (%) Share of GDP (%) Sector 1990-94 1995-99 2000-05 1990 1995 2000 2005 AgricuIture 3.1 3.6 4.8 50 51 48 46 Industry 2.0 5.4 9.0 16 15 17 20 Services 1.9 3.8 6.1 35 35 35 35 Total GDP (factor cost) 2.5 4.0 6.0 100 100 100 100 Source: Authors calculations, based on United Republic of Tanzania various years. Note: Total may not equal 100 because of rounding. stay of the Tanzanian economy and output age points. There i s concern that both the grew by 4.8 percent during 2000 to 2005, gold and fishing industries are reaching thereby contributing 1.7 percentage points the limits o f expansion o f natural resource to Tanzania's overall growth. extraction, with only limited prospects of future growth. The environmental impact The service sector grew by 6.1 percent, rep- o f these industries i s also a concern. Ex- resenting a significant improvement as ports of agricultural products declined comparedwith the growth o f the sector dur- since the mid-1990s and started to recover ing the previous five-year period. Growth only recently. Exports o f manufactured was particularly strong inthe areas of trade, goods have recovered inrecent years, from transportation, andcommunication. about US$30 million in 1999 to US$156 million in 2005, which i s only 23 percent Sectoral growth patterns have resulted in higher than the value o f export o f manu- modest structural change o f the Tanzanian factured goods that had already been economy (table 1.3). The relatively fast achieved in 1996. A key challenge for the growth of industry has led to an increase Tanzanian economy is thus to strengthen by 2 percentage points in its contribution and diversify its export base. to GDP, and as o f 2005 it accounts for 20 percent of GDP. Most o f the increased contribution o f industry to GDP i s attrib- Improved Macroeconomic utable to the expansion o f the mining and Fundamentals construction sectors. Gains in manufactur- ing were more modest. Conversely, the Since 1995, Tanzania has successfully share o f agriculture has fallen by 2 per- pursued a policy o f macroeconomic stabi- centage points from 48 percent in 2000 to lization. This policy has resulted in low 46 percent in2005. inflation and accelerated economic growth. At the core o f Tanzania's stabili- Rapid expansion o f nontraditional exports, zation efforts i s fiscal consolidation. In especially o f gold and fish, plus a recovery 1996/97, Tanzania adopted a cash budget o f exports o f manufactured goods, has led system, under which expenditures are to rapid export growth (figure 1.2). How- strictly limited to available resources from ever, even though exports o f gold rose domestic revenue and foreign aid. This from virtually nothing to about 5 percent system virtually eliminated net domestic of GDP, their contribution to economic borrowing. In parallel to the country's re- growth has been only around 0.4 percent- gaining fiscal control, donor assistance in a Ia) Traditionaland ~ o n t ~ aExports~ ~ ~ ~ a l ~ ~ Source:Bank of Tanzania vari~usyears. the fomi o f grants and concessionat Eend- ~ n ~ e r e s ~to~ nnote that higher monetary g ing increased ~~ibst~~i~ially.assis- 5ucli growth not result in increased i ~ r ~ ~ t i o n ~ tance firtanced the increase in g o ~ e m ~ i e nr~e ~ e ~ both n ~ ~ thegfaster economic growth e~pend~~ures from about 16 percent of and the increased ~ i o ~ e ~ andzdeep- ~ ~ n ~ a ~ CDP in 1997J98 to more than 24 percent ening of financial markets in recent years in2 ~ ~ (figure 1,0 ~I 4 ~ ffigLJre 1.4). ~one~he~ess,the n1Otley- to-GDP ratio is still tow in Tanzania Tanzania's tax system has undergone sig- compared with that in other African coun- nificant r e f o ~ i~n~c,~ ~ d i nthe r ~ ~ l ~ c e ~ r rtries,~M2, which is a measure of the g e n with a vafuc added tax in ~iio~iey supply, stood at 21.6 percexie of n a t i ~ no f ~uisa~icetaxes, GDP in ~ a n ~ a For~Kenya it was 40.0 n ~ . the remotal of tax e ~ e ~ ~ t the n s ~ oadop-~ percent, ~ ~ ~ ~ a ~29.6i percent,eand b ~ L i tion o f a new income tax act in 2004, and Uganda 19.0 percent. the ~ a ~ ~ o no f~ local ~ o~v ~ o ~ i l ~ e ~ ~ i e n t taxes. A program to s t r e n ~ tax~adrnini-~ ~ ~ e ~ Credit to the private seetor has started to stration ~cco~?ip~niedrefom and re- the recover, hut it is still XQW. Credit a5 a share sulted in a gradual increase in doniestie o f GDP declined d r ~ m a ~ ~ cfroml yabout a ~ revenue from 11.3 percent of CDP in 35 percent to GDP in 1993 to 8 percent in ~ 9 ~ 9 ~ 2to013.3 percent in 2 0 ~ 4 ~ ~ 5 . 2003, bur it recovered to 15 percent by 0 0 2005, as shown in panel (a) of figure 1'5. Fiscal co~isol~da~~on provided the has Most of this decline is due to the fiscal con- space for sound n ~ o policy,~aimed at ~ ~ a ~so~ida~ion,which resulted in a reduction in price stability. A I ~ ~ ~ oaniirtal~~~ u g ~ credit to the public sector from~ 23 percent 0 1 ~ ~ ~ ~ ~ r O ~was 30 to 40 percent during the t h of GDP in 1993 to 0.2 percent in 2004. first he1990~~droppedtobclou it Credit to the private sector cO~i~ra~ted froni I O perccnt at the start of ~ a n z a ~ ~stabi- a ' s about 15 percent of GDP to only 3 percent 1. In parallel, the rate of in 1996.However, since then it has stcadily c o n ~ ~ ~ i u ofromsabout ~ ~ l y recovered and stood at 11 percent of GDP 40 percent in 1995 to around 5 percent in in 2005, The ~ ~ ~ i t e led~ton ag s ~ ~ n ~ ~ c a n ~ n onetary growth started to increase in real interest rates. Panel (b) of n rcccirt years. mainly figure 1.5 slows that lending rates in- age itiflows o f forcign aid, creased from 5 percent in 1993 to X 5 per- only partially sterilized. It is cent by 2000, but they decfined subsc- Figure 1.3. Government Finance, ~ 9 9 ~ ~ 9 2 - 2 ~ 0 4 ~ ~ 5 (a) Xotaf revenue and expenditure (b) Grantsand fiscal deficit beforegrants ," " ....... ................... ,.. ......... ._ _ . _" ............................................... (e) Foreignand domestic financingof the fiscal deficit -,lx _,I ........... ...... ..,-,_,-__I go ~, ....................................................................................... a * ! - __.-____ ' ' ' ......................... ...x.. ......~,.~,. x_,,_x .... ....... .." Source:Tanzanianauthori~i~s l ~ t ~ ~ n aMonetaryaFundstaff sstrrnates. and t ~ ~ n ~ Figure 1A. Money and Inflation, 1990-2004 {a) Monetary growth and the rate of inflation _---_-_- (b) Moneyand quasi money (7%of GDP) A Decade of Reforms, MacroeconomicStability, and Economic Growth Figure 1.5. DomesticCredit and Interest Rates (a) Domestic credit (b) Real interest rates 40 7 20 35 n II 30 g l5 B 10 P 25 ; e 1 20 $ 5 .E B S 15 0 10 I K -5 5 0 1990 1992 1994 1996 1998 2000 2002 2004 /.CsntraiGiernmnt Rivate sector 0 Otherj ~ T - b l i l s---e-Lendlng --b- Saving Source: Based on Bank of Tanzania various years. quently to 10 percent as the money supply investment (FDI). A survey o f direct for- was allowed to grow faster. eign investment (Bank o f Tanzania 2004b) provides more detailed information on Large foreign aid inflows have financed an FDI. The value of the foreign investment increase inpublic sector investment during stock was estimated to have risen from the past five years, but the ratio of private US$1.7 billion in 1998 to US$2.6 billion investment to GDP has beenfairly constant in 2001. Manufacturing is the biggest re- since 1995. Following successful fiscal sta- cipient o f FDI and accounts for about 34 bilization efforts in the mid-1990s, which percent o f the stock o f FDI. Mining ac- included significant cuts in public invest- counted for 28 percent and tourism for 8.1 ment, capital formation declined to 16.3 percent o f the stock o f FDI at the end of percent o f GDP in1997. As shown inpanel 1999. Agriculture, despite its importance (a) of figure 1.6, it has recovered since and potential, accounts only for 6.7 per- then, however, reaching 22 percent of GDP cent o f the FDIstock inTanzania. in 2005. This recovery is primarily due to an increase in public sector investment Public investment has witnessed a dra- from 3.2 percent of GDP in 1997 to 7.5 matic shifi from low return investment by percent in 2004. While public sector in- parastatal enterprises to investment by the vestment increased, private sector invest- central government in public infiastruc- ment fluctuated between 12 and 13 percent ture. During the first half o f the 199Os, o f GDP and only in2005 increasedto 14.4 parastatal enterprises accounted for most percent. The increase in public sector in- o f public sector investment. The privatiza- vestment affected primarily investment in tion o f most parastatals has ledto a virtual buildings and other works. As panel (b) o f disappearance o f this form o f low-return figure 1.6 shows, investment in equipment investments. As panel (c) o f figure 1.6 il- fluctuated between 7 and 9 percent of GDP lustrates, public sector investment i s now since 1996, showing, however, a sustained almost exclusively investment by the cen- upward trend inrecent years. tral government. Such investment includes the rehabilitation o f government facilities Since 1995, Tanzania has also benefited inall areas and the expansion ofthe infra- from significant inflows of foreign direct structure for social service provision (in 11 Poverty Reduction and Growth Figure 1.6. Capital Formation, 1995-2005 - - 1 12 20 - ! 10 y, -4 ~~~ 5 % " - J 2 04 , , , , , , , 0 , . , I 1995 1885 1997 4998 1999 ZOW 2001 ZW2 2W3 2w4 20m 10% lose ioe7 iws low zwo 2001 zoo2 zoo3 zoo4 zoo5 +Buildings -Other Wwks +Equipment (c) Publicsector investment (d) Privatesector investmentand FDI I I Source: United Republicof Tanzania various years. particular the construction o f classrooms, ample, Rodrik 2000) suggests that in- as well as investments inroads, water, and creases in the saving rate are typically the power). result o f accelerated economic growth rather than being a driver of economic Domestic saving has shown a dramatic in- growth. Consequently, measures to stimu- crease from -5 percent of GDP in 1993 to late saving have a lesser priority thanthose almost 15 percent in 2005 (see figure 1.7). aimed directly at stimulating growth and The increase insaving i s almost entirely the investment. result of increasedpublic sector saving. Be- tween 1990 and 2005, public sector con- Macroeconomic stabilization has also re- sumption declined from 18 percent o f GDP sulted in a slowdown o f the depreciation to 7 percent. During that period, private o f the nominal effective exchange rate sector consumption increasedinitially from (NEER) and an appreciation of the real 83 percent o f GDP in 1990 to 88 percent in effective exchange rate (REER) between 1999, but it dropped subsequently to 80 1996 and 2001. Since 2001, both the percent by 2005. The difference between NEER and the REERhave been depreciat- savings and investment, which corresponds ing, as panel (a) o f figure 1.8 demon- to the current account deficit, declined strates, and the 2004 level o f the REER i s dramatically since the early 1990s. considered to be consistent with equilib- rium inthe external accounts (IMF 2004). Inan internationalcomparison, Tanzania's The current account (exclusive and inclu- savingrate i s still low. The average saving sive of current transfers) and the overall rate for Sub-Saharan Africa is 18 percent. balance o f payments have improved sig- However, the economic literature (for ex- nificantly since 1993, as shown in panel 12 A Decade of Reforms, Macroeconomic Stability, and Economic Growth Figure 1.7. Saving and Investment, 1990-2005 25 ? 20 15 I ?---=,' I Figure 1.8. Exchange Rate and Balance of Payments, 1990-2004 (a) Realand nominal effective exchange rates (b) Balance of payments 2004 20, 1 10 250% o 230% B!? 210% -10 190% -20 170% -30 -40' I 110% - 7- 9 0 % - = - - - - / - - + - ~:- \'n 70% - -Current account (incl. transfers) &NEER&REER -A- Overallbalance Source:Tanzanian authorities and InternationalMonetary Fund staff estimates. (b) of figure 1.8. The current account defi- ternal accounts coincides with an extended cit (excluding grants) has declined from 35 period o f strong appreciation o f the REER, percent o f GDP in 1993 to only 4 percent suggesting that the positive effect o f over- in2002. The current account balance after all macroeconomic stabilization and struc- grants improved from a deficit o f 26 per- tural reforms outweighed the exchange cent in 1993 to 0 percent in 2002. Subse- rate appreciation. quently, partly on account o f the absorp- tion of increasing aid flows, the current Panel (a) of figure 1.9 shows that external account deficit increased again to 12 per- debt declined dramatically during the cent by 2005. The improvement in the ex- 199Os, from more than 130 percent of GDP 13 Poverty Reduction and Growth Figure 1.9. Public and PubliclyGuaranteed Debt and Debt Service, 1990-2004 (a) Debt and debt service (b) Gross international reserves Source: Tanzanianauthorities and InternationalMonetary Fund staff estimates. in 1990 to 60 percent in 2004. Inparallel, factor productivity to economic growth. debt service also declined from a peak of The analysis requires estimates o f the almost 40 percent of exports to 4 percent in physical andthe humancapital stock:2 2004. Followingimplementation of the En- hanced Highly Indebted Poor Countries Human capital. Humancapital is typi- Initiative and the Multilateral Debt Relief cally estimated on the basis o f average Initiatives, Tanzania's debt sustainability years o f schooling. Between 1992 and indicators are well below the debt sustain- 2001, average years o f education ability thresholds. The net present value o f among Tanzania's adult population in- debt-to-export ratio is estimated to be 64 creased from 3.8 years to 4.2 years. percent in2006. At the same time, as panel According to Cohen and Soto (2001), (b) of figure 1.9 shows, gross international Tanzania's average number o f years of reserves increased more than threefold be- schooling in2000 was higher than that tween 1997 and 2005 and stand now at o f Uganda (3.22), but lower than that US$2.0 billion or more than five months o f of Kenya (5.80) and far from that o f import coverage. South Africa (7.22). e Physical capital. The capital stock i s Determinants of Economic estimated by using the perpetual in- Growth in Tanzania ventory method. After the capital stock is estimated for an initial pe- This section explores two questions. First, riodY3estimates for subsequent peri- what has been the contribution of the main ods are obtained by assuming a cer- factors o f production-that is, capital per tain rate o f annual depreciation4 and worker and education per worker-to the by adding new investment.However, recent growth performance? Second, what inmany instances, the value of an in- have been the demand-side factors that vestment will not automatically add an triggered Tanzania's growth acceleration? equivalent amount to the capital stock.5 Growth accounting i s a widely used Among the reasons for this discrepancy method to estimate the contribution o f are poor investment decisions and physical and humancapital as well as total overpriced investment cost. For Tanza- 14 A Decade of Reforms, Macroeconomic Stability, and Economic Growth nia, it i s likely that in the past govern- education remains unaffected b changes in ment investment did not add commen- the capital stock (see table 1.4).K surately to the productive capital stock. Examples abound, such as the The growth accounting calculations sug- Morogoro shoe factory, which never gest that the recent acceleration of eco- produced at more than 4 percent of its nomic growth is primarily due to more production capacity. Once the govem- rapid accumulation of physical capital and ment decided to privatize such enter- increased factor productivity, as shown in prises, their sales value, ifthey could be figure 1.10. From 1995 to 1999, the ratio sold at all, was typically low and in of investment to GDP declined from 19.9 many cases negative. Inthe case of in- to 15.5 percent, and the contribution o f vestments in infiastructure, the rate of capital accumulation was negative. After depreciation quite likely exceeded the 1999, both public and private sector in- commonly assumedrate o f 5 percenton vestment increased significantly, reaching account of insufficient investments in 22.2 percent o f GDP by 2005. Duringthat maintenanceandrehabilitation. period, private and public sector invest- ment increased by 2.1 and 4.4 percentage Thus, 1985, when Tanzania abandoned its points, respectively. However, national socialist policies and started to introduce accounts data may not fully reflect the sig- market reforms, might be an appropriate nificant increase FDI that Tanzania has year for the reestimation of the capital witnessed since the mid-1990s. Although stock. Given the paucity o f data, we exam- there was virtually no FDI until the early ine two scenarios. Under the first scenario, 199Os, by 2000 FDI was more than the actual capital stock in 1985 i s assumed US$500 million, or about 5 percent o f to be only 50 percent o f the estimated capi- GDP. An investment report prepared by tal stock. Under the second scenario, it is the Bank o f Tanzania (2004b) shows that assumed to be 75 percent. With respect to most of the FDI flows go into mining (30 the growth accounting exercise, a lower percent); manufacturing (3 1 percent); and initial capital stock implies reduced abso- wholesale and retail trade, including tour- lute amounts of depreciation and, thus, ism (14 percent), with agriculture receiv- higher levels of net investment7 and addi- ing only a small share (7 percent) o f total tions to the capital stock. Intum, the share FDI.9 The sectors that were the main o f growth attributed to capital will be beneficiaries o f FDI flows are also those higher, and the share attributed to factor that showed the highest growth rates inthe productivity will be lower, with a lower past five years. The FDIsurvey also shows initial capital stock. The contribution of the increasing importance o f regional FDI Table 1.4. Decompositionof Tanzania's Growth, 1995-2005: Depreciation of Initial Capital Stock by 0, 25, and 50 Percent Contribution(%) ~~ Adjustment in capital Output per Physical Factor stock (%) worker capital Education productivity 0 2.31 0.34 0.73 1.oo 1.24 25 2.31 0.58 0.73 50 2.31 0.88 0.73 0.70 Source: Author's calculations. 15 Poverty Reducfion and Growth Figure 1.IO. Decomposition of Economic Growth per Worker into Contribution of Human and Physical Capital Accumulation and Total Factor Productivity (TFP), 1985-2005 2.5% 1 2.0% 1.5% 1.O% 0.5% capital Education 0.0% -2.0% ` I985-89 1990-94 1995-99 2000-05 Source: Author's calculations. flows from member countries o f the South cumulation i s labeled totalfactor produc- African Development Community and the tivity, but it contains a variety o f different East African Community (primarily South factors that contributed to growth. First, it Africa, but also Kenya). Such flows ac- covers the increase in land under cultiva- count for 42.2 percent o f total FDI flows tion inresponseto improvedincentives for in2001, compared with 52.2 percent from agricultural production, which has been a Organisation for Economic Co-operation major contributor to growth inthe agricul- and Development countries. In2004, FDI ture sector. It also includes increases in had declined to about 2.4 percent of GDP, capacity utilization as a result o f increases because the major investments inthe min- inaggregatedemand. Enhanced efficiency ing sector hadbeen completed and privati- o f resource allocation and use i s another zation-related FDIhad declined. important component o f the observed in- crease in factor productivity in Tanzania. Recent investments in education have a Analysis of the manufacturing sector sug- relatively long gestation periodbefore they gests that increased productivity in the will lead to an effective increase in the sector is the result of accelerated exit and human capital o f Tanzania; hence, the con- entry o f firms. As older, inefficient firms tribution o fhumancapital to growth i s still leave, new, more productive firms enter small. However ifrecent trends inexpand- the sector. All those factors contribute to ing primary and secondary education are bringingTanzania's economy closer to its sustained, the contribution o f human capi- production frontier. Productivity increases tal to growth i s likely to increase. that lead to an expansion o f the production frontier, such as the adoption o f improved The share o f growth that cannot be ex- technologies, have been o f lesser impor- plained by human and physical capital ac- tance, but a number o f cases inwhich new 16 A Decade of Reforms, Macroeconomic Stability, and Economic Growth Table 1.5. Sources of Growth by Region, 1990-2000 Contribution(%) OutDut per Phvsical Factor Country or region woiker i%) cipitai Education productivity Tanzaniaa 2.4 0.4 to 0.9 0.7 0.8 to 1.3 Africa -0.2 -0.1 0.4 -0.5 China 8.8 3.2 0.3 5.1 East Asia (except 3.4 2.3 0.5 0.5 China) Latin America 0.9 0.2 0.3 0.4 Middle East 0.8 0.3 0.5 0.0 South Asia 2.8 1.2 0.4 1.2 World 3.5 1.2 0.3 1.9 Industrialcountries 1.5 0.8 0.2 0.5 Source: Data for Tanzania, author's calculations; other data, Bosworth and Collins 2003. a. Data for Tanzania are for 1995-2005. technologies have been successfully intro- the contribution of factor productivity duced demonstrate the scope for future seems to be less clear cut. For example, in growth impulses. Such technologies in- the case o f China, almost 60 percent of clude innovations introduced through FDI the country's growth i s explained by en- but also the results of Tanzania's agricul- hanced factor productivity, whereas in tural research system. other East Asian economies, only about 20 percent o f growth i s explained by in- To interpret the findings o f the growth creases in factor productivity. The contri- accounting exercise for Tanzania, one bution o f education to growth shows less may find comparisons to other countries variability across regions. The contribu- useful. Table 1.5 compares the contribu- tion ranges in a rather narrow band be- tion o f capital, education, and factor pro- tween 0.2 percent and 0.5 percent. ductivity to growth in Tanzania with re- sults obtained for other regions. Several The increase in average years o f education interesting facts are apparent from this during the past decade is reflected in a table. First, average growth rates varied relative large contribution o f human capi- significantly during the 1990s. The Asian tal to growth. However, compared with the economies performed strongest, led by performanceo f other countries, Tanzania's China, where output per worker grew by performance during the 1990s appears to 8.8 percent. On the other end o f the spec- be plagued by several weaknesses. First, trum, the African countries in the sample the contribution of capital formation seems saw their output per worker decline by to be low, evenwhen we allow for the fact 0.2 percent. The wide variation in growth that the initial capital stock may have been rates is reflected in similarly large varia- lower than suggested by the commonly tions in the contributions o f capital and used estimates. In addition, the data also factor productivity. For the Asian econo- suggest scope for greater increases in fac- mies, capital formation clearly played an tor productivity, although estimates for important role, and it appears that, in more recent years actually suggest that general, faster growth o f output per factor productivity has indeed increased in worker i s associated with higher growth Tanzania to levels achieved by other re- incapital per worker. On the other hand, gions. However, it is likely that this recent 17 Poverty Reduction and Growth Table 1.6. Sources of Growth: Expenditure, 1990-2005 Average annual growth rate Average contribution to growth ("/I ("/I EconomicActivity 1990-94 1995-99 2000-05 1990-94 1995-99 2000-05 GDP (market prices) 2.5 3.8 7.0 Consumption 2.1 6.0 5.4 2.1 6.2 5.6 Privateconsumption 2.6 5.7 4.0 2.1 5.0 3.4 Governmentconsump- tion -0.3 7.3 10.9 0.0 1.2 2.2 Grossdomestic invest- ment 5.8 -1.3 15.3 1.2 -0.3 3.5 Net exports -1.4 -2.1 9.1 0.4 0.4 -1.7 Exportsof goods and nonfactorservices 12.1 5.4 19.4 1.6 1.o 5.0 Importsof goods and nonfactorservices 3.2 1.4 15.0 1.3 0.6 6.7 Statisticaldiscrepancy -37.4 -285.0 4.3 -1.2 -2.6 -0.4 Investment by type of as- set Construction 2.1 -4.4 18.9 0.2 -0.4 1.8 Machineryand vehicles 9.9 1.o 12.5 1.o 0.1 1.6 Investment by public and private sector Privateinvestment 20.5 -0.3 11.6 2.3 0.0 1.9 Public investment -1 1.5 -5.0 25.9 -1 .o -0.2 1.5 Source: Authorscalculations, basedon UnitedRepublicof Tanzaniavarious years. upturn in factor productivity is the result points to the overall growth o f 6.8 percent o f demand-side factors and efficiency o f GDP at market prices during the period gains from reforms that have moved Tan- from 2000 to 2005. Figure 1.11 shows zania closer to its production frontier; dramatically different trajectories for the there is little evidence that the increased growth rates o f GDP at market prices, in- factor productivity indeed represents tech- cluding and excluding government spend- nological change that would increase Tan- ing. Overall GDP at market prices shows zania's productive capacity on a sustain- the familiar acceleration o f economic able basis. growth inTanzania over the past 15 years. However, GDP net o f public sector expen- On the demand side, aid-financed public diture shows a markedly different growth sector investment has provided an impor- path. There, growth duringthe period from tant stimulus for economic growth. These 1995 to 2005 was 4.5 percent, compared growth impulses have translated into with 4.8 percent during the first halfo f the growth in consumption and also higher 1990s. Most o f the increase ingovernment demand for imports (table 1.6). spending i s financed from increased aid inflows to Tanzania. In the national ac- Increased government spending has been counts data, this situation i s reflected in a an important engine o f economic growth widening o f net exports during the period (see box 1.2), contributing 3.8 percentage from 2000 to 2005. 18 ~ ~ ~ spending affects~e e o~n v ~ activity boththrough tts ~ e e~ c n ~ nd-side eff'ecls and its supply- side effects, Tne ~ a ~ i t u anderime pattern o f de~and-s~de d effe pends on wheehcr there arc pacities inthe econon2y and whether changes nng o ~ e ~ m spending are ~ e n ~ ~ ~or p ~ r a ~ t. Inrhc cabc of CXCCSS capaaty rcases rn~ v ~ ~exp~ndi~ureadd to e~c o n o ~ ~ c e ~ ~ e n ctiy through the added denim the fomi of~ o ~ ~ purchasesasnwell as kough ~ m e ~ ~ u ~ t i p leffects onpriyate sector consi~mp~~on.~ ~ p ~estimatesaoflthis d e ~ ~ n d - s ~ ~ e ~ e r ~ c d e ~ e l u c~a~~~ng~ ~~ e~s ~ m Kell,m ~andnhfahfauz1002;Schclarelc20033 suggestthat t ~ , mandeffects arc ~ ~ ~ i clargerliny~ c ~ e ~than ~in~i n~ ~d ucountnes.aIf~the increase in gvv- a ~ o s ~ ~ emment e~pen~itureexceeds unusedcapacity, it ts likdy to exen inflatz pressures inthe short runand-if increases in g o ~ ~ e ~e~xe~n~ t~ d are~considered to be i t ~ r e ent-to induce an ex- pansioninthe produG~i~~~ ofthe e c o ~ a ~ y . capacity Public e ~ ~ e ~ ~ dalsouaffcct aggregate supply, Go i t r ~ s ments ~ e n d j ~ g ~ sa1piyegc ~a ~ ~ ~ ~ e ~ t ~ n ~ ' e S ~ in~economic ~ n f r a s ~ r ~ chuman e , n e n ~ ~ ~ ~ r capita research and ~ ~ ~ ~ l ~ ~ncrcase ~ ~ - c a n ~ ~ c rhcp r o ~ ~ c ~capacity ufthe economy, Whereas the de~and"sieffects typically occur inthc i ~ t e d ~ short term, s ~ p p ~ ~ -effectsehave a ~ l d d. Inaddition, though d e ~ a n d - S ~ ~ e efSerts affect only the level o f ~ e o n o ~ai ~ c e cffccts have the potent~alto lead to a sustained increase in the gro%\thrate. Separate from the effect of rnment spending oneconomic activity is the ~ o v e r n ~ spending e n ~ 1s fin .In general, higher taxation and domestic b negative ~ ~ pon ecc o~n ~ ~activity. The effect of donor ~ n a ~ --separate~from that ofthe a ~ ? ~ e c ~ n associated Increase in ~ o ~ e r spending-isn ~p r i ~ a~~ l ~~ ~ t cffcct onothe u n ~ ~ s ~ ~rate. ~ ~~~~~~~ Figure 1,11 Growth Rates of GDP, l ~ ~ land Exclusive~of G ~ ~ i v ~ ~ ~ i n ~ Spending, 1 9 9 ~ - 2 ~ ~ 5 I 1995-99 2000-05 1 Source:Author's ~ a l ~ ~ ~ abasedoon ~ , ~ i n United R e ~ u baf~Tanzaniavarious years. i ~ Figure 1.12. of Public and Private ~ ~ pto Ecanomic ~ ~ d ~ t ~ ~ ~~~~~~~~~~~~ Growth, 1 9 9 ~ - 2 ~ ~ ~ ~ ." 6.0% 5 0% 4.00,b 3 0% 2 0% 1*0% 0 0% -1 0% -2 0% -3.0% 2 990-94 1995.93 2000-05 __ - Source:Author's ~ a l c ~ ~ ~based onsUnited Republicof Tanzania variousyears. ~ i o n ~ Box 1.2 presents the ~ o n ~ ~ i ~ouf pri-o n ~ j that the ~emand-sj~eimpulses vate and public e x ~ e n d ~ ~tor eecononiic u sed ~ ~ ~ e ~ ~ i e n ~ grawtlt and shows a similar picture as s p ~ n have only partly tra~sla~ed d ~ ~ ~ into figurc 1.11, The c o n ~ r ~ b o~ftprivate ~ ~ n higher d o ~ ~ eps~ o~di u~c ~ ~butnbavc also o e~pend~ture to growth i n ~ r e a sonly ~ ~ c ~ n t ~ ta ~increased dimports and a ~ u ~ ~ from 4.5 percentage points during the pe- \ ~ ~ d e ~ icurrent account deficit. j n g II 1990 and 2994 to 5.3 per- centage points during period bctwcen , ~ ~ ~exportsuhave h ~ o ~ been rising rapidly, 2000 and 2005 (figure 1.12). However, this effect has been more than o the direct c o I i ~ r i b ~ t ~oofn ~ o ~ r e ~ n n ~ eiitcrease n t in imports (table lh}. The in- spending to G ~ o n o ~ growth increased ~ i c crease in exports is ~ ~due to ~thc a ~ l d ~ ~ n ~ a t ~ cduring that period, Public a l l y rapid Gxpans~onof gold p r ~ d ~which ~ ~ ~ ~ ,~ c sector refonns during the first half of the fostered large imports o f c a p ~ ~goods and a ~ 1990s involved a s i g n ~ ~ reductjon in c a ~ ~ ~ ~ ~ ~ pThe netneffect o f the increase ~ c t . the size of ~ ~ ~ ~ cand thusi slotvcdn ~r ~ ~ c in gold p r ~ d ~ c ~asj oseen~inlhc contribu- n down growth of GDP. The period from tion of mining lo growth, was oslly 0.4 1995 to 1999 sau' a slight recovery o f percentage points. ~ ~ ~ ~ ~ ~ spending,i iand the contribu- n i i i e t tion to GDP growth was 1 ~ e r c e ~ i ~ aConclusions ~ e point, The period from 2000 to 2005 saw a sharp increase in ~ o ~ espending, e ~ ~ ~ t I ~ a c r ~ e ~s toa ~b ~o~~~~~ a~ ~ ~ o ~ atid its direct ~ o ~ i ~t ~ i b ~ ~ ~and o ~ ~ the i ~ p ~ ~ n ~ Ganf~a ~broad ~ i t j a range growth was 3.8 percent, o i structural reforms have resulted in a tion o f the sum of net exports and unre- steady a c c e l e ~ ~ ~ini oeconomic growth ~ corded trade and the s ~ a ~ ~discrep-~ adl s ~ ~ L ~ rthe~past decade. This a c h ~ ~ ~ ~ e ~ ~ e ~ ~ i ~ ancy has been negative and ~ n ~ i sr a strong~e ~ tsd o r~s efor~c~~~~i it~ ~ n ~ aoft j o n during the ttirtc pc ods. This findi1~~the reforni course with~ n~a~roec~noIit~c ~ 20 A Decade of Reforms, Macroeconomic Stability, and Economic Growth stability, clear definition o f the roles o f the tions inaid flows andunderpins the impor- public and private sector, market deter- tance o f using aid to strengthen the com- mined prices, and openness o f the econ- petitiveness o f the economy through omy as the foundation for private sector- investments in infrastructure and the im- driven growth. Increasing inflows o f for- plementation o f reforms that would in- eign aid have supported Tanzania's reform crease competitiveness. efforts and have contributed significantly to the recent growth acceleration by stimu- In addition, the growth in exports is pri- lating domestic demand. However, the re- marily due to increased natural resource view o f growth and macroeconomic per- exploitation-namely, exploitation o f gold formance also highlights weaknesses that and fish-while exports o f agricultural need to be addressed if growth i s to be products and manufactured goods account sustained. for a declining share o f exports. Diversifi- cation o f exports is thus critical both with Most o f the growth acceleration can be respect to the dynamic effect o f greater explained by demand-side effects o f for- integration into international markets as a eign aid as well as by greater efficiency o f driver o f innovation and technological the economy. However, the growth effect change and as an important source for ef- o f efficiency gains i s likely to diminish ficiency gains and scale effects achieved over time, and aid inflows cannot be ex- byproducing for a larger market. pected to increase indefinitely. Thus, an important element focus o f future reforms Finally, reforms and the acceleration in needs to be on strengthening the invest- economic growth have so far shown only a ment climate at the firm and farm levels. limited effect on the lives o f the poor. In the next chapter, we review the link be- Also, the fact that increasing aid inflows tween growth and poverty and analyze have provided an important demand-side Tanzania's poverty profile with the objec- stimulus to the economy highlights the tive o f identifying reform strategies that vulnerability o f the economy to fluctua- would ensure pro-poor growth. For a detailed discussion see Pritchett (2000). ` Developments in Even after 1985 untilthe mid-l990s, public sector government finance are moni- investment and investment by parastatal entities tored and reviewed under the annual Public Expen- was relatively large. Therefore, it might indeed be diture Review process. appropriate to discount up to 50 percent of public Estimates o f humancapital are based on Bosworth 'sectorinvestmentisgross investment duringthat period. andCollins (2003), who provide data up to 2000 by Net investment minus depre- averaging time series for years o f schooling assem- bled by Barro and Lee (2000) and Cohen and Soto *ciation.effect This is purely a consequence o f the under- (2001) and applying a rate o f return o f 7 percent to lying growth accounting methodology. It is likely years o f schooling. Estimates o f the capital stock are that a lower ratio o f capital to qualified labor would also drawn f?om Bosworth andCollins (2003). implyhigher returns to capital and lower returns to Nehru and Dhareshwar (1993) provide estimates qualified labor. of the capital stock for a large number o f countries, Shares represent the average distribution of FDI including Tanzania. by sector, 1999-2001 (BankofTanzania 2004b). W e assume an annual rate o f depreciation o f 5 Io To the extent that increased government spend- percent. ingfell on imports, the effect on growth is reduced. 21 CHAPTER2 The Challenge of Reducing Poverty in Tanzania Monetary Poverty: error for the two surveys and, therefore, we Levels and Changes cannot conclude that poverty declined. The reduction in the severity of poverty at the Duringthe 199Os, the poverty headcount in national level shows a 10 percent decline in Tanzania remained roughly the same as a the poverty gap (average distance o f the percentage o f the population, but the sever- poor to the poverty line) and a 20 percent ity o f poverty seems to have been reduced decline in the poverty gap squared (sever- (box 2.1). According to the data fkom two ity), butthat difference is not significant for household surveys (National Bureauof Sta- the same reason(see table 2.1). tistics 1993, 2002), the national poverty headcount declined from about 38 percent The modest decline inthe national poverty to 35 percent. However, because the levels conceals large regional differences 1991/92 survey had a small sample size and inlevelso fpoverty andchanges duringthe the 2000/01 survey had some sampling is- decade, as can be seen in table 2.2. At the sues, that difference is within the marginof strata level (rural areas, other urban areas, 23 Poverty Reduction and Growth The Tanzania poverty line applied in2001was T Sh 7,253 per adult equivalent per 28 days (De- cember 2000 prices), which is equivalent to US$0.79 per capita per day, usingthe international purchasing power parity conversion. That figure i s considerably less than the US$l.O8, or "dollar- a-day," poverty line often usedininternational poverty comparisons. Ifone were to calculate poverty inTanzania usingthe dollar-a-day poverty line, then the poverty line expressed inlocal currency would be T Sh 9,900 and the poverty incidence would be around 57.5 percent, consid- erably more than the 35.6 percent usedinTanzania. One can compare Tanzania's poverty line with that used elsewhere inthe region. Uganda's pov- erty line, for instance, i s higher and equivalent to US$l.12per capita per day. Application ofthat poverty line to Tanzania gives a poverty incidence o f 59.8 percent. One can safely conclude that Tanzania's poverty line i s low from an international and a regional perspective. Table 2.1. Poverty Indices, 1991/92-2000101 Poverty Poverty gap Population share headcount Poverty gap squared Region 1991192 2000101 1991192 2000101 1991192 2000101 1991192 2000101 Tanzania 100.0 100.0 38.6 35.3 11.8 10.4 5.3 4.4 Rural 82.1 79.0 40.8 38.6 12.7 11.5 5.8 4.9 Other urban 12.6 13.6 28.7 25.9 8.0 7.7 3.2 3.4 Dar es Salaam 5.3 7.4 28.1 17.6 7.5 4.1 3.0 1.6 Coastal 12.8 12.8 40 34.7 11.8 8.7 4.8 3.2 Rural 10.8 10.7 44.1 36.7 13.1 9.2 5.4 3.4 Urban 2 2.1 17.9 24.9 4.8 6.2 1.7 2.4 Northern Highlands 10.1 11.0 20.2 36.1 3.0 10.9 0.9 4.7 Rural 9.3 9.4 20.4 38.7 3.0 11.9 0.8 5.2 Urban 0.8 1.6 18.4 20.8 3.7 5.3 1.3 2.0 Lake 35.1 37.4 37.0 39.0 12.3 12.9 5.7 5.8 Rural 31.3 32.9 39.6 39.8 13.3 13.0 6.2 5.9 Urban 3.8 4.4 15.2 32.7 4.4 11.5 2.0 5.5 Central 9.4 8.3 48.8 42.4 16.1 11.8 7.6 5.0 Rural 9.0 7.3 49.5 44.8 16.5 12.5 7.8 5.4 Urban 0.4 I.o 34.5 24.1 9.5 6.2 3.6 2.3 Southern Highlands 15.3 14.0 46.6 25.8 15.9 6.0 7.8 2.1 Rural 11.0 11.2 48.2 28.0 16.9 6.6 8.8 2.3 Urban 4.3 2.8 42.7 16.9 13.2 3.7 5.4 1.3 South 11.9 9.1 43.9 43.2 11.2 13.0 4.2 5.4 Rural 10.3 7.5 44.0 45.7 11.6 13.6 4.3 5.6 Urban 1.2 1.6 42.6 31.2 8.3 9.9 2.9 4.4 Source: Based on Household Budget Survey 1991/92(National Bureau of Statistics 1993)and HouseholdBudget Survey 2000/01 (National Bureau of Statistics2002). Note: Bold typeface indicatesthat the difference across years is significant at the 5 percent level. 24 The Challenge of Reducing Povertyin Tanzania and Dar es Salaam), there was an impor- sources o f those differences and analyzing tant decline in poverty in the Dar es Sa- them could yield insights into the effects on laam region even as the population grew. households o f the economic roller coaster Disaggregating the data further into seven o f the 1990s. Unfortunately, the data qual- regional zones, other urban areas, and ru- ity is not adequate for that type o f analysis. ral areas,' the picture becomes even more A new survey will be needed to provide a heterogeneous. In addition to that in Dar basis for such analysis. In the rest o f this es Salaam, a significant decline in poverty chapter, we usually show data disaggre- occurred in the Southern Highlands, in gated by the three strata (rural areas, other both urban and rural areas (see table 2.2). urbanareas, andDar es Salaam) only. In the Northern Highlands, poverty in- creased in rural areas but remained con- Economic Inequality, Poverty, stant in urban areas, whereas the opposite and Growth occurred in the Lake districts. In the case o f the Southern Highlands, the decline in In the simplest terms, poverty reduction poverty was quite a reversal in fortune. results from (a) an increase in per capita However, because o f the small sample size household income and (b) the share o f this o f the 1991192 survey, estimates of increase that goes to the poor. That simple changes in poverty at the zone level are identity can be further disaggregated by imprecise and should be interpreted with region o f the country or other relevant great care. The more precise changes in household group. The change in income table 2.2 are in bold (that i s they are sig- that goes to the poor i s measured by the nificant at the 5 percent level). change in share o f the poor in total con- sumption, but it i s also helpful to look at Insum, the extent o fpovertyreduction dur- the change in common measures o f ine- ingthe 1990sseems to be uneven, with ma- quality, including the Gini coefficient and jor gains in some areas and an overall the Theil index, the latter o f which is de- worsening in others. Understanding the composable and so i s often the measure Table 2.2. Poverty Status in Tanzania, 1991/92-2000/01 Populationshare (%) Povertyheadcount(%) Location 1991192 2000101 1991192 2000101 Tanzania 100.0 100.0 38.6 35.3 Ruralareas 82.1 79.0 40.8 38.6 Other urban areas 12.6 13.6 28.7 25.9 Dares Salaam 5.3 7.4 28.1 17.6 Zones Coastal 12.8 12.8 40.0 34.7 Northern Highlands 10.1 11.0 20.2 36.1 Lake 35.1 37.4 37.0 39.0 Central 9.4 8.3 48.8 42.4 Southern Highlands 15.3 14.0 46.6 25.8 South 11.9 9.1 43.9 43.2 Source: Based on Household Budget Survey 1991/92(National Bureau of Statistics 1993)and HouseholdBudget Survey 2000/01(National Bureau of Statistics 2002). Note: Bold typeface indicates that the difference between the two surveys is significant at the 5 percentlevel. 25 Poverty Reduction and Growth Figure 2.1. SimulatedChanges in Poverty, 1992-2002 "4 Source: Demombynes and Hoogeveen 2004. o f choice. In this section, we look at the tor-then poverty most likely rose between role those two components o f poverty re- 1992 and 1994 and then fell in the final duction played inthe observed results. years o fthe decade. As with other analyses, the pattern shown in figure 2.1 suggests The small growth in per capita consump- that national poverty incidence i s mostly tion recorded inthe household survey data determinedbyruralpoverty incidence. during the period from 1992 to 2001 is caused by a poor economic growth per- Table 2.3 suggests a widening poverty gap formance during that period (see figure between urban areas-specially Dar es 2.1). Real GDP grew at an average o f 3.6 Salaam-and rural areas. Between percent, which resulted in annual per cap- 1991/92 and 2000/01, the share o f Dar es ita growth o f only 0.7 percent when com- Salaam's population in the lower (na- binedwith ahighannual rate o fpopulation tional) quintiles declined, while the share increase o f 2.9 percent. We must note that o f Dar es Salaam's population inthe high- the household survey data provide only est national expenditure quintile jumped two snapshots in time and fail to represent from 23.5 percent to 43 percent. At the the full evolution o f poverty between 1992 same time, in rural areas the share o f the and 2001. What actually happened to av- population in the lowest quintiles in- erage household consumption during that creased marginally, while that in the high- period? We do not know the answer, but est quintile decreased. In other urban ar- one way to estimate it would be to assume eas, inequality seems to have increased, that it tracked economic growth duringthat becauseboththe share o f the population in period. By applying macroeconomic the lowest and that in the highest national growth data to the microlevel household quintile increased, while the share of the survey data, we can simulate changes in population in the middle quintiles de- consumption year by year.2 As noted ear- creased. However, the extent to which the lier, income growth varied substantially cost o f living in urban areas and that in during that period. If poverty tracks that rural areas have divergedmay be less than growth-especially if poverty in the rural these numbers suggest. The same holds for areas tracks growth in the agricultural sec- changes across quintiles, to the extent that 26 The Challengeof Reducing Poverty in Tanzania Table 2.3. Distributionof the Population by Strata by National Quintile, I991192-2000/01 percentage of population Dar 8s Salaam Other urban areas Rural areas Quintile 1991/92 2000/01 1991192 2000101 1991/92 2000/01 Poorest 12.7 8.2 13.1 14.4 21.5 22.1 2nd 16.5 12.2 16.6 15.6 20.8 21.5 3rd 22.8 17.3 19.1 14.9 20.0 21.I 4th 24.4 19.2 19.9 22.7 19.7 19.6 Richest 23.5 43.0 31.3 32.4 18.0 15.7 Source: Based on Household Budget Survey 1991/92(National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureau of Statistics 2002). Note: Boldtypeface signifies that the difference across years is significant at the 5 percent level. Table 2.4. Increase in Consumer Prices between 1991 and 2001 Product Increase (%) Food 221 Drinks and tobacco 363 Rent 614 Fuel and light 723 Clothing and footwear 63 Furniture and utensils 204 Household operation 87 Personal care and health 103 Recreation and entertainment 145 Transportation 340 Source: Bank of Tanzania 2004a. the consumption baskets by quintile differ Figure 2.2 shows the growth rates o f con- significantly. Table 2.4 shows the increase sumption for the entire distribution. The in the components o f the consumer price vertical line represents the poverty line. index. Rent, which has increased by more The curve shows that growth was positive than 600 percent, is an important compo- for poor and nonpoor households alike, nent o f urban consumption baskets, which i s indicated by the fact that the whereas clothing and footwear, furniture growth incidence curve i s above zero at all and utensils, or personal care and health, points. The flatness o f the curve indicates which showed more modest price in- that growth was evenly distributed and not creases, are likely to claim a larger share highly concentrated among either high- in the expenditure baskets o f the rural income or low-income households. The population. Similarly, these items are slight rise at the far right o f the curve i s likely to claim a larger share o f the expen- attributable to the larger gains in Dar es diture baskets o f the poor population than Salaam, in particular among high-income of those o f the better-offpopulation. households inthe capital. The growth incidence curve i s another Reflecting the consumption changes noted way of representing that story (figure 2.2). above, the national measures of in- 27 Poverty Reduction and Growth Figure 2.2. Growth Incidence Curve: Nation as a Whole 6 1 I - Q M h noom. O Y H ~ 3 g- sp 2;'- zz Source: Demombynes and Hoogeveen 2004. Table 2.5. Gini Coefficient and Theil Index, 1991/92-2000101 Gini Theil Location 1991192 2000101 1991192 2000101 Tanzania 0.33 0.34 0.185 0.199 Rural areas 0.33 0.32 0.184 0.177 Other urban areas 0.34 0.35 0.201 0.214 Dares Salaam 0.30 0.34 0.152 0.208 Zones Coastal 0.34 0.30 0.204 0.165 Northern Highlands 0.26 0.32 0.120 0.170 Lake 0.33 0.34 0.191 0.207 Central 0.30 0.34 0.161 0.185 Southern Highlands 0.37 0.32 0.244 0.169 South 0.29 0.33 0.157 0.197 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureau of Statistics 2002). Note: Bold typeface signifies that the difference across years is significant at the 5 percent level. equality-the Gini coefficient and the The Theil index can be decomposed into Theil index (see table 2.5)-changed within- and between-group inequality (see modestly. We can calculate those indices Shorrocks 1984).4 Table 2.6 shows the by region (with the same caveats as be- share o f the between-group inequality o f fore). Not surprisingly, in places such as total inequality for different groups o f the Northern Highlands, where the rural areas population. Although the within-group es- suffered relative to the urban areas, ine- timate contributes the bulk of the inequal- quality increased. For the country in gen- ity for all groupings, the share o f inequal- eral, which in 2000/01 had a Gini coeffi- ity between groups has also ~ i d e n e d . ~ cient o f 0.34, the level o f inequality i s low Inequality between households classified compared to other African countrie~.~ by the education level of the head of 28 The Challenge of Reducing Poverty in Tanzania Table 2.6. Share of InequalityCreated by Between-Group Differences in Tanzania, 1991192-2000101 percentage of total inequality Theil index (%) Groups 1991192 2000101 Dares Salaam, other urban areas, and rural areas 2.0 6.5 Geographic zones 2.8 5.2 Education levels 4.7 12.0 Household sizes 18.2 15.8 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureauof Statistics 2002). household increased the most, with geo- contribution to poverty reduction (a reduc- graphic inequality also increasing. Ine- tion inthe headcount). quality between geographic zones i s higher than in Mozambique but much Growth has clearly decreasedpoverty, but lower than in Uganda. The widening ine- at the same time, higher inequality has quality by education level o f head o f worked in the opposite direction. Outside households may reflect increased returns of Dar es Salaam, both the consumption to education at the upper levels relative to growth andthe inequality change are small, the lower levels, or it may be caused by such that they almost cancel out one an- factors correlated with education (for ex- other. In Dar es Salaam, inequality in- ample, living inanurbanarea). creased, but household consumption growth occurred as well, andthe poor bene- What has been the effect on poverty of the fited substantially from this growth. Thus, increase in inequality within and between despite the much larger increase ininequal- regions? The change inpoverty can be de- ity in Dar es Salaam, poor households composed into a growth component, an gained more than in other areas where the inequality component, and a residual com- inequality increase was more modest. In ponent (Datt and Ravallion 1992). Table other words, although the rising inequality 2.7 shows such a decomposition for in Tanzania is somewhat of an issue, in- mainland Tanzania, Dar es Salaam, other creasing overall income growth across the urban areas, and rural areas. Inthat table, country in the poor households i s an even the growth impact refers to the change in greater issue. Income inequality does not the poverty headcount brought by growth seem to have been an impediment to in household consumption per capita as growth inDar es Salaam. reported in the survey (in Tanzania, a re- duction o f 8.4 percentage points in the Alternatively, the change inpoverty canbe poverty headcount). The inequality impact decomposed into regional composition i s the change in the poverty headcount effects. There are intraregional effects, caused by the change inthe distribution of population-shift effects, and interaction per capita consumptionas measuredby the effects (see Huppi and Ravallion 1991). household survey (an increase of 5.5 per- Those effects are depicted in the second centage points in the poverty headcount). part o f table 2.7. During the period from If the change is negative (for example, 1991/92 to 2000/01, poverty declined sub- -8.4 percentage points), it is a positive stantially inDar es Salaam and other urban 29 PovertyReductionand Growth Table 2.7. Decompositionof Change in Poverty in Tanzania, 1991/92- 2000/01 percentage points Dares Other Rural Aspect of change Tanzania Salaam urban areas areas Poverty 1991 38.6 28.1 28.7 40.8 Poverj/ 2001 35.4 17.6 26.0 38.7 Total change in poverty -3.2 -1 0.5 -2.7 -2.1 Growth impact -8.4 -1 8.4 -6.6 -5.3 Inequalityimpact 5.5 12.4 4.0 2.7 Residual -0.2 -4.5 -0.2 0.6 Urban-ruraldecomposition Change in poverty in Dares Salaam -0.6 Changein poverty in other urbanareas -0.3 Change in poverty in rural areas -1.7 Total intraregionaleffect -2.6 Population-shifteffect -0.4 Interactioneffect -0.2 Types of employment Change in povertyin farming and fishing -2.0 1.1 -0.2 -2.5 Changein poverty in paid employment -0.4 -7.9 -0.9 0.2 Changein poverty in self-employment -0.6 -5.8 -0.9 -0.2 Change in poverty in family employment 0.1 -0.2 0.2 0.0 Changein poverty in noneconomicactivity 0.0 0.1 -0.1 0.1 Total intrasectoraleffect -2.9 -12.6 -1.8 -2.5 Population-shifteffect -0.6 0.6 -0.8 -0.7 Interactioneffect 0.4 1.5 -0.2 1.2 Source: Basedon HouseholdBudget Survey 1991/92(NationalBureauof Statistics 1993)and HouseholdBudgetSurvey2000/01 (National Bureauof Statistics2002). areas. That decline attracted people from widens. Nonetheless, rural-urban migra- rural areas, resulting ina 4.0 percent popu- tion may indirectly promote poverty re- lation growth rate in other urban areas, duction, as in facilitating stronger rural- compared with 2.7 percent in rural areas. urban links. However, migration itself is Despite those population inflows and the not a major contributing factor. Most o f substantial urban poverty reduction, only the poverty reduction during the decade 0.4 percentage points o f the total drop in has happenedwithin those areas. national poverty of 3.2 percentage points (12 percent) can be attributed to a shift in Finally, table 2.7 also shows the power o f the population from poorer rural areas to a rural poverty reduction strategy for Tan- wealthier urban sectors. This finding sug- zania. Declines in poverty in Dar es Sa- gests that the majority of migrants are laam have had only a minor effect on na- nonpoor. As nonpoor move from rural to tional poverty rates because only 7.5 urban areas, the urban-rural poverty gap percent o f the population lives there. The 30 The Challenge of Reducing Poverty in Tanzania slight income growth for the rural poor vulnerability. Therefore, analysis of non- had a large effect on overall poverty in monetary indicators i s helpful. Two types Tanzania, accounting for one-half o f the o f analysis were performed: drop inthe nationalpoverty rate. The same circumstances apply to households en- 0 Analysis o f food share and food gaged infarming and fishing, becausethey security are a large share o f the population and their income i s low. However, the last sec- 0 Analysis o f change in reported assets tion of table 2.7 shows that the movement andhousingconditions of households out o f agriculture has also Normally, if income i s increasing, then played a strong role in poverty reduction and i s likely to remain important in the households would purchase more nonfood future. Nevertheless, acceleration in na- items with the marginal income, whereas if poverty is going up, the food share tional poverty reduction can be achieved would increase. Food share did go down only through an accelerated decline in among both the poor and the nonpoor dur- poverty inrural areas. ing the period between the surveys (see The analyses in this section suggest that a table 2.8), suggesting increasing welfare, successful poverty reduction strategy and the food share declined the most in would include increases in rural income Dar es Salaam, where poverty fell the levels, an urban growth strategy, and fa- most. Another important dimension i s per- cilitation of rural-urban migration. ceived food security. Unfortunately, such information i s available only for 2000/01. Table 2.9 shows that perceived problems Nonmonetary Poverty with satisfying food needs are highly cor- Measures relatedwith actual poverty status. The measurement o f poverty using con- Table 2.10 presents household asset hold- sumption data can be imprecise, and it ings by quintiles in Tanzania across the does not adequately capture other dimen- years. Inall areas and inall quintiles, data sions o f poverty, such as insecurity and show some increase in assets, another Table 2.8. Food Share by Strata and Poverty Status, 1991192-2001101 1991192 2000101 Foodshare Foodshare Tanzania 74.7 72.4 Dares Salaam 72.7 58.7 Other urban 70.3 65.0 Rural 75.5 70.5 Non-poor 74.7 67.9 Poor 74.6 70.8 Source: Based on Household Budget Survey 1991/92 (National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureau of Statistics 2002). Nofe: Bold typeface indicates that the difference across years is significant at the 5 percent level. 31 Poverty Reduction and Growth Table 2.9. Households'Perceptionsof Problemswith Satisfying Food Needs in Relationto Actual Poverty Status, 2000/01 Problems with satisfying food needs Share below the poverty line (%) Never 26.3 Seldom 36.2 Sometimes 39.1 Often 47.2 Always 46.7 Source: Based on Household Budget Survey 2000/01 (National Bureau of Statistics 2002). indicator o f increasing welfare. Only one tio o f dependents to earners? What are the asset (chair) in the poorest quintiles had a human capital assets of earners in poor significant decrease. However, because households? How do earners fare in the some assets are held by very few house- labor market? Beginning with the demo- holds (for example. fishing nets), the in- graphic analysis, earners in poor house- crease i s hard to interpret. But indices of holds clearly support more people than in consumer durables did go up, especially higher-income households. Poor house- ownership o f bicycles, radios, and stoves holds have more children, and large inall quintiles and ownership of motorcy- households are a disproportionate share o f cles and televisions inthe highest quintile. the poor. Households that have four or The quality o f housing also improved (see more children have a 50 percent poverty table 2.11). When measured at the national rate, which is much higher than the na- level, the share of households without a tional rate o f 35 percent (table 2.12). The foundation or without a durable roof went number of widowed heads of household down substantially. Quintile analysis has more than tripled, rising from 2.5 per- shows improvements across the spectrum. cent to 7.8 percent, and those households The disaggregation by regional strata sug- are also more likely to be poor (table gests improvements, but outside of Dar es 2.13). Female hardship has also increased Salaam, the results are not significant. (table 2.13), but it does not seem to be linked to poverty. Heads o f households In sum, the monetary and nonmonetary remain primarily uneducated: 37 percent data are in broad agreement: welfare ap- have no education, and another 6 percent pears to have improved, but by a larger have adult education only. Less than 3 margin inDar es Salaam than inthe rest of percent o f the poor have any postprimary Tanzania. We now look at the economic education, but the share with some or characteristics o f households to learn more complete primary education has risen dur- about how the growth process affected ingthe past decade (table 2.14). households. The economic activities o f the very poor Economic Characteristics have remained virtually the same during of the Poor the period, while those o f the nonpoor have undergone a structural change. For Developing effective antipovertyprograms approximately 80 percent o f households in requires going deeper into the economic the lowest two quintiles, the head o f characteristics of the poor. What i s the ra- household works in farming or fishing, as 32 33 34 The Challengeof Reducing Poverty in Tanzania Table 2.12. Poverty by Number of ChildrenAge Five or Younger, 1991/92-2001 IO1 percentage 1991/92 2000/01 1991/92 2000101 Poverty Poverty Populationshare Populationshare 0 32.0 30.8 36.8 32.9 1 39.8 32.3 35.6 32.7 2 45.3 40.5 18.0 23.2 3 42.3 44.5 5.4 6.8 4 or more 53.4 50.1 4.2 4.4 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household Budget Survey 2000101(National Bureau of Statistics 2002). Note: Bold typeface signifies that the difference across years is significant at the 5 percent level. Table 2.13. Poverty by Civil Status of Head of Household, 1991/92-2001/01 percentage 1991192 2000101 1991192 2000/01 Poverty Poverty Populationshare Populationshare Never married 13.3 20.7 3.8 5.5 Married 40.2 36.3 87.6 81.6 Divorced 29.0 28.8 3.7 5.1 Widowed 37.3 40.0 2.5 7.8 Female headed household 35.3 34.8 15.1 24.4 Source: Based on Household Budget Survey 1991/92(National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureau of Statistics 2002). Note: Bold typeface signifies that the difference across years is significant at the 5 percent level. they did in 1991 (see table 2.15).6 Mean- most likely absorbed much o f the govern- while, the highest quintile showed a large ment and parastatal sector layoffs, as well shift into paid employment and self- as new entrants and those who left farming employment and a shift away from unpaid (10 percent o f the urban heads of house- nonagricultural family labor and agricul- holds outside o f Dar es Salaam). However, tural labor. The middle and top quintiles without panel data drawn ftom surveying register an increase in self-employment, the same households, it is difficult to know and-especially in the top quintile- how this transitionreally happened. movements into paid employment. In Dar es Salaam, the heads o f households have Because we have no data on fanning or mainly moved out o f paid employment and self-employment income, we use the con- into self-employment, probably as a result sumption per capita o f the household as a of the government sector restructuring. In proxy. Table 2.16 shows the change across other urban areas, net paid employment7 the years in average household per capita was constant, but the self-employed sec- consumption according to the sector in tor-most of such businesses are without which the head o f household i s employed employees-grew to be almost as large as and the location. The differences are strik- the farming and fishing sector. That sector ing; the regional differences in poverty 35 r g 0 0 N N e r sm 36 The Challenge of Reducing Poverty in Tanzania Table 2.16. Changes in Average Consumption per Adult Equivalent by Employment of Head of Household and by Strata, 1991/92 and 2000/01 percentage change Other urban Employment Tanzania Dares Salaam areas Rural areas Farming and fishing 7.1 -6.3 7.8 7.3 Paid employment 28.2 60.0 17.8 19.2 Self-employment 18.3 64.5 7.2 5.7 Family employment -1 9.2 17.6 -5.7 n.a. No economic activity 15.9 25.5 44.4 7.3 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993) and Household Budget Survey 2000/01 (National Bureau of Statistics 2002). Note: n.a. = not applicable, because family employment is mainly an urban category. Family em- ployment comprises unpaid family helpers whose work is in nonagriculture and, for 1991192, houseworkers in an urban area. Table 2.17. Average Consumption perAdult Equivalent by Employment of Head of Household in Dar es Salaam, Other UrbanAreas, and Rural Areas, 1990/91-2000/01 constant Tanzaniashillings I990191 2000101 Other Other Dares urban Rural Dares urban Rural Tanzania Salaam areas areas Tanzania Salaam areas areas Farming and fishing 3,681 3,820 3,895 3,663 3,954 3,581 4,199 3,928 Paid em- ployment 4,971 4,404 5,735 4,810 6,373 7,045 6,754 5,733 Self- employ- ment 4,857 4,046 5,665 4,608 5,745 6,655 6,074 4,871 Family employ- rnent 4,964 4,420 5,220 ma. 4,011 5,201 4,921 3,162 No economic activity 3,292 3,976 2,912 3,331 3,815 4,989 4,206 3,573 Source: Based on Household Budget Survey 1991/92 (National Bureau of Statistics 1993) and Household Budget Survey 2000101 (National Bureau of Statistics 2002). Note: n.a. = not applicable, because family employment is mainly an urban category. Family em- ployment comprises unpaid family helpers whose work is in nonagriculture and, for 1991 house- workers in an urban area. performance emerge again as a notable cor- more than 80 percent of all Dar es Salaam relation with poverty reduction. In Dar es households. Inthose sectors, Dar es Salaam Salaam, where poverty fell, all types of households began on average behind other households, except farming and fishing, had urban areas, but duringthe decade, they not real consumption growthper capita, with the only caught up but passedtheir counterparts most growth occurring in the paid employ- elsewhere (tables 2.17 and 2.18). In Dar es ment and self-employment households- Salaamduringthat period, the informal self- 37 Poverty Reduction and Growth Table 2.18. Index Number of Average Consumption per Adult Equivalent by Employment of Head of Household, 1991 Tanzania Basis constant Tanzaniashillings 1990191 2000101 Dares Other Dares Other Tanzania Salaam urban Rural Tanzania Salaam urban Rural Farming and fishing 1.O 1.o 1.1 1.0 1.1 1.o 1.1 1.1 Paid employment 1.o 0.9 1.2 1.0 1.3 1.4 1.4 1.2 Self-employment 1.o 0.8 1.2 0.9 1.2 1.4 1.3 1.0 Family employment 1.O 0.9 1.1 n.a. 0.8 1.o 1.0 0.6 No economic activity 1.o 1.2 0.9 1.0 1.2 1.5 1.3 1.1 Source: Based on Household Budget Survey 1991/92(National Bureau of Statistics 1993)and Household Budget Survey 2000/01(National Bureau of Statistics 2002). Note: n.a. = not applicable, because family employment is mainly an urban category. Family em- ployment comprises unpaid family helpers whose work is in nonagriculture and, for 1991house- workers in an urban area. employment sector was not a stagnant pov- employment and in agriculture, whereas erty trap for most households. consumption in households headed by a self-employed person grew slower than in Inother urbanareas, thepictureisdifferent. other urban areas. But because they started Households headed by someone in paid out behind the other sectors, even house- employment (about one-fourth o f house- holds in paid employment remain vulner- holds) realized substantial gains, but these able to poverty, and those inthe other sec- gains were only one-third o f those received tors are highly vulnerable. Within all areas, by similar households in Dar es Salaam. the ratio o f average household consumption The fastest-growing and largest sector, self- in agriculture households compared with employment households had a very small nonagricultural households widened, again gain (on an annual basis, almost nothing). highlighting the need for a rural poverty We are not surprised that the poverty re- strategy. duction performance was much better in Dar es Salaam than inother urbanareas. In Ifwe consider the labor force as a whole, other urban areas, the high inflow into the the shifts in employment patterns mirror self-employment sector from agriculture those o f heads o f household, with some and the low average consumption gains for differences by gender (see table 2.19). that sector during the period suggest that Both genders moved out o f agriculture, but the labor force shift was more likely caused women moved proportionately more. by a pushfrom other sectors thana pulland Nonetheless, women are still more likely that the sector has a substantial low produc- to be employed in agriculture than men. tivity and subsistence component. InDar es M e n still dominate the paid employment Salaam, however, some pull elements into and self-employment sectors, but women the self-employment sector may be present, were able to move into those sectors. The given the income gains. main difference with heads o f households i s that the whole labor force has a higher In rural areas, households did equal or share working as family employment, better than other urban areas in self- which i s to be expected. 38 The Challenge of Reducing Poverty in Tanzania Table 2.19. Share of Labor Force by Employment Sector and by Gender, 1991192-2000101 Share of females Share of all Share of males Share of females (as a % of (as a % of total (as a % of total (as a % of total sectoral labor Sector of laborforce) laborforce) laborforce) force) employment 1991192 2000101 1991/92 1991192 2000101 2000/01 1991/92 2000101 Farming and fishing 81.3 71.5 77.5 70.0 84.8 72.8 55.0 55.1 Paid employment 7.9 6.8 12.7 10.4 3.6 3.8 24.2 55.1 Self- employment 5.5 8.9 7.4 11.9 3.7 6.4 36.0 38.7 Family employment 5.2 12.8 2.3 7.9 7.7 17.1 78.9 71.9 Total 100.0 100.0 100.0 100.0 100.0 100.0 n.a. n.a. Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household BudgetSurvey 2000/01 (National Bureauof Statistics 2002). Note: n.a. = not applicable. Bold typeface signifies that the differenceacross years is significant at the 5 percent level. Totals may not equal 100.0 because of rounding. Table 2.20. Sector of Employment of Spouses Compared with That of Heads of Household, 1991192-2000/01 percentage Sector of Sector of employment of head of household employment Farmingand fishing Paid employment Self-employment of spouse 1991/92 2000101 1991192 2000/01 1991192 2000101 Farming and fishing 97.8 94.1 52.1 41.7 47.3 39.3 Paid employment 0.5 0.6 13.2 12.3 6.3 3.3 Self- employment 1.o 2.3 8.4 10.9 8.3 18.4 Family employment 0.5 3.0 26.3 35.1 37.9 39.0 Source: Basedon Household Budget Survey 1991/92(National Bureau of Statistics 1993)and Household Budget Survey 2000/01 (National Bureau of Statistics2002). Note: Boldtypeface signifies that the difference across years is significant at the 5 percent level. Households normally have several earners spouse to work inagriculture while the head and several sources of income. One way to of household worked in a nonagricultural get a snapshot of this is intable 2.20, which occupation. That situation has shifted as shows the sector of employment of the spouses have moved out of agriculture and spouse comparedwith the sector of the head into mostly nonagricultural family work- of household. Ingeneral, agricultural house- but also into self-employment ifthe headof holds have diversified, whereas nonagricul- household is in self-employed. That shift tural households have movedthe other way. may help explain the consumption gains o f The dominant trend in 1991 was for the those households. Consistent with those 39 Poverty Reduction and Growth trends, the number o f households with a and that policy needs to focus on those business has risen (table 2.21). But this areas. trend, too, has left the bottom two quintiles behind. From the cross-section data, we cannot tell whether a household that added a Explaining Household business tended to move out o f poverty or Consumption whether only those households not in pov- Up to this point, we have relied primarily erty could affordto add a business. on bivariate analysis. We now turnto mul- tivariate analysis to isolate the separate In sum, where economic growth has oc- effects o f independent variables on house- curred, the labor market has responded. In hold consumption. The main advantage o f the self-employment sector, where supply such analysis i s that it allows us to isolate equals demand, activity has increased, as the effect of a specific variable while hold- well as incomes. Does the income growth ing all other (observable) factors constant. represent simply an expansion because the For example, households inDar es Salaam demand for intermediate and final goods are less likely to be poor than households and services has changed, or is it ex- inrural areas. Simultaneously, the level of plained by productivity increases (more education is also higher in the Dar es Sa- capital, for example)? This analysis cannot laam thaninrural areas (see table 2.22). answer those questions. However, we see clearly that earnings do not seem to Regression analysis allows us to separate have increased in areas with low growth those two effects. Our regression analysis Table 2.21. Main Type of Business by Quintiles, 1991192-2000/01 Type of business Year Poorest 2nd 3rd 4th Richest No business 1991I92 67.3 60.7 55.4 53.8 55.4 2000101 66.2 56.8 55.6 54.1 49.3 Agriculture 1991I92 20.9 21.7 30.4 27.2 25.5 2000101 15.4 22.2 22.2 20.9 22.3 Wholesale and I991192 9.6 13.7 11.2 16.1 13.8 retail 2000101 6.4 8.5 8.9 13.3 16.8 Other 1991192 2.3 3.9 3 2.9 5.3 2000101 12.0 12.6 13.3 11.7 11.7 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household Budget Survey 2000101 (National Bureau of Statistics 2002). Note: Bold typeface signifies that the difference across years is significant at the 5 percent level. Table 2.22. Average Amount of Education of Head of Household, 1991/92-2000/01 Location 1991192 2000/01 Tanzania 4.0 4.9 Dares Salaam 7.4 7.8 Other urban areas 5.8 6.5 Rural areas 3.5 4.3 Source: Based on Household Budget Survey 1991192(National Bureau of Statistics 1993)and Household Budget Survey 2000101 (National Bureau of Statistics 2002). Note: Bold typeface signifies that the difference across years is significant at the 5 percent level. 40 The Challenge of Reducing Poverty in Tanzania compares the determinants of the log o f parable across areas and even higher in consumption per capita in rural areas and Dar es Salaam for primary schooling and other urban areas. We regress log per postsecondary schooling. l1Inall areas, the adult-equivalent real consumption on a set marginal effects o f schooling have risen o f explanatory variables (compare with duringthe decade, which is consistent with Datt and Joliffe 1999). The coefficients our findingson inequality. should be interpreted as the percentage change in per adult-equivalent real con- The 1991/92 results on the influence of sumption resultingfrom amarginal change type o f employment are puzzling because in the explanatory variable. We did three there are no significant differences in re- separate regressions by regional strata be- turns to different types of employment in cause, although most of the explanatory Dar es Salaam, whereas self employment variables are significant in both regres- leads to significantly higher consumption sions, the coefficients are different, and, inother urbanareas and inruralareas. Not overall, tests o f significance confirmed a surprisingly, paid employment increases different structure at the 99 percent confi- consumption in rural areas. That pattern dence level. The regression results are pre- changes in 2000/01. Now, both paid em- sentedintable 2.23. ployment and self-employment lead to significantly higher consumption, and the With respect to demographics, household returns are almost twice as high in Dar es composition affects consumption even Salaam compared with other urban areas though consumption i s measured in per and rural areas. Furthermore, no employ- adult equivalents. The net effect o f more ment reduces consumption in rural areas, household members i s negative independ- and family employment increases con- ent o f age, but it i s generally less negative sumption in other urban areas. No em- in rural areas.8 Furthermore, the negative ployment increases consumption inDar es effect o f more adult male members i s lar- Salaam, but those households may depend ger than the effect o f more adult female on transfers. In general, this set of vari- membersin all three areas in all years, al- ables usually measures a range of nonob- though it seems stronger inDar es Salaam. servable household characteristics that are The age of the head o f household has no related to income earningpower, and thus effect, either in 1991/92 or in2000/Ol.9 they can be hard to interpret and should be viewedwith some caution. Education in general leads to significantly higher consumption in all areas compared Overall, Dar es Salaam seems to havebene- with no schooling, which is the excluded fited &om substantial increases inreturns to category. loFurthermore, more education education. The combination of higher levels leads to even higher consumption, which of schooling in general and a much larger i s extremely positive given the recent pol- concentration o f paid employees and self- icy o f providing free primary schooling. employed workers in the capital area helps However, returns may change if the level explain why Dar es Salaam fared muchbet- of education changes for the population in ter in2000/01 comparedwith other areas. general. In 1991/92, the effects of school- ing were higher in other urban areas and Conclusions rural areas as compared with Dar es Sa- The household budget survey o f 2000/01 laam. That relation changed in 2000/01, showed only a modest reduction in pov- though, when the effects were more com- erty during the 1990s, reflecting the 41 2- 42 The Challengeof Reducing Poverty in Tanzania relatively poor growth performance during nonwage sectors. In the language o f the that period. Those poverty estimates do 1970~~ growth did trickle down in Dar es not yet capture the effect o f higheconomic Salaam, and it seems to have happened growth recorded since 2001, but poverty because o f the economic links betweenthe simulations suggest that they are likely to formal and informal sectors interms o f the have had a significant effect on poverty. demand for goods and services provided However, that effect can be confirmed by the informal sector. Even though the only with the new household budget sur- growth process brought more inequality to vey that i s plannedfor 2006/07. Dar es Salaam, poverty reduction was not hampered. On the contrary, in other urban Nonetheless, even during the 1 9 9 0 ~pov- ~ areas, where inequality was reduced only erty reduction occurred in some regions o f slightly, growth did not take place. So Tanzania. The experience o f those regions, growth was more pro-poor in Dar es Sa- especially Dar es Salaam, suggests that laam, despite the inequality increase. We significant poverty reduction i s possible. It do not to argue that growth needs inequal- also suggests that the reforms o f the 1990s ity or even that inequality is good. We and the flow o f foreign aid they triggered simplypoint out that some increase inine- have, for the most part, benefited the bet- quality can be tolerated along with growth. ter-educated, better-capitalized areas such Tanzania's challenge will be to maintain as Dar es Salaam. the pace o f growth and poverty reduction in Dar es Salaam and similar areas (for At the same time, that experience has also example, Southern Highlands), while shown that when the private sector i s able adopting new strategies and measures to to create jobs, there i s spillover to the non- reach the rest o f the country. Notes The within-group estimate is calculated as the 'The difference o f 100 percent and the between-group seven regional zones are Coastal (Morogoro, estimate reported intable 2.4. Pwani, and Tanga), Northern Highlands (Arusha The survey records a decrease o f farm activities and Kilimanjaro), Lake (Kagera, Kigoma, Mara, inthe middle quintile andanincrease in"no activ- Mwanza, Rukwa, Shinyanga, and Tabora), Central ity," However, we suspect that those differences (Dodoma and Singida), Southern Highlands (Ir- reflect a change in coding. inga, Mbeya, and Rukwa), South (Lindi, Mtwara, Net paid employment is government-private andRuvuma), andDar es Salaam. employment. The approach followed i s the one developed by * As consumption is measured as household con- Datt and Walker (2002) and Datt and others sumption divided by household sue and household (2003). That approach was used by Demombynes size also appears on the right-hand side o f the re- and Hoogeveen (2004) to investigate the pattern o f gression, the negative sign could very well be the growth inTanzania during the 1990s. result o f a measurement error inhousehold size. In 1997, the Gini coefficient for Mozambique A different model specification with age dummies was 0.40. In 1999, the Gini coefficient for Uganda didnot give any economically and statistically sig- was 0.43. In1997, the Ginicoefficient for Kenyawas nificant effects either. We included in earlier esti- 0.45. (World Bank 2005a). mated models a dummy for female heads of house- The decompositions are implemented using holds, for civil status variables, and for raising o f a Stephen P. Jenkins's Stata program, ineqdeco. foster child, along with interaction terms between Given the survey comparability problems discussed those dummies. The dummy for female heads o f earlier, this analysis should be considered as only household was not significant despite the specifica- indicative. tion, and the same was true for civil status. But 43 Poverty Reducfion and Growth prevalence o f a foster child was significantly posi- tive, indicating that households with foster children are those that have the economic means to do so. However, given the endogeneity o f that variable, we didnot use it inthe final results. lo W e include the type o f schooling instead o fjust years o f schooling to allow for nonlinear effects of education. I' W e reject the simple model with no variation in returns to education across areas and years. 44 CHAPTER3 Spatial Dimensions of Growth and Poverty Reduction This chapter summarizes regional patterns nomic growth and the skewed income dis- o f economic growth and poverty reduction tribution is vital for at least two reasons. inTanzania, usingregional gross domestic First, such understanding i s considered to product (GDP) data, poverty measures, be the key for unlocking secrets o f how to and other socioeconomic indicators for the kindlegrowth o fregions lagging economi- period from 1992 to 2003. The main ob- cally and sustain growth inregions that are jectives are to highlight cross-regional' better off economically. As Krugman variations in incomes, to identify the most (1991) stated, if we want to understand common barriers to economic growth, and differences in national growth rates, a to suggest regional policy options for im- good place to start i s by examining differ- proving economic growth performance.* ences inregional growth. Second, extreme Understanding the causes o f the geo- economic gaps among different regions graphically uneven distribution o f eco- are potential flash points for social and 45 Source:~ ~ a2005.g o ~ political ins~abil~~y.~ n s e q ~ ~the~ ~ which is homc to less than 8 percent of ~ e l y * ~ ~ a of~a ~ ~ a ~ei cthegkiiid o~f growth inn ~ Tanzania"~~ ~ p u ~ a ~c ~~onn t, ~ aboiit~ ~ e s ~ ~ that creaks benefits ~ ~ i rsociety has ~18opercent of Tanzania's GDP, equal to o ~ ~ ~ ~ figured ~ r o ~ ~oni theed ~ e~ ~~ ~ ~~theo c o~~ b~i neecontribu~~or~ n ~~ ~ to agenda in nzariy co~~Ii~riesand ~ o ~to ~ GDP by the bottom six~ regions ~ i n ~ ~ do so. ~ o d o ~ aEindi, Kigorria, Itlara, and , Mtsvara).4 Overall Regionalincome Patterns In ternis of r a ~ k ~ nper capita CDP over g time, the top five regions for the period ~ ~ a i ~ Tanzania ~i s c h a ~ a c ~ ~ rbyc d from 1996 to 1999 were Amsha, Dar es i ~ a n j ~ iiighly uneven d ~ s t ~ b ~o f~ econoniic ~ o n Salaam, Iringa, Rukwa, and Ruvuma, activity and incomesacross its 21 adminis- Those regions with the Iowest GDP per 0n2~ with ~ azsrSalaam domi- capita i ~ c ~ u d~~odd ~ ~Kigorna, Kili- i a , other rcgions. From 1992 to a, and Tanga. However, the 2003, ribour 52 pcrccn ranking changed during 2000 to 2003 (scc ttunat GDP was produ figitrc 3.1). The ,IJfttvara and Mwanza re- gions: Arusha, Dar gions r ~ ~~ ~ ~ a ~andi Rukwa~among ~ ~u d a 9 and ~ l ~ ~ (mapy ~thentop five regions for per capita GDP. n ~ ~ 3.1). The Dar es Salaam region alone, ~ ~ ~ ~the~Coast,a Kagera,~and Tabora I ~ l , 46 Spatial Dimensions of Growth and Poverfy Reduction Figure 3.1. Average GDP by Region, 1996-99 and 2000-03 (a) Ranking in 1996-99 (b) Ranking in 2000-03 RankingofAwrageGDP Per CapitaByRegion Rnnldngof Average GDP Per Caplts ByRegion 1996-99 1000-03 I T; 0 I00000 200000 100000 400000 500000 TShs Source: Nationalincomeaccounts. regions joined the ranks of regions with city i s Tanzania's major port and the com- the lowest GDP per capita while the Mara, mercial and financial capital, and therefore Kigoma, and Tanga regions made mar- it is more connectedto the global economy. ginal gains. The reversals in the regional As a port, it is the main conduit of export ranking of average GDP per capita reflects and import trade not only for Tanzania but inpart the new investment inmining, fish- also for the neighboring countries o f Bu- ing, and relatedservices around Lake Vic- rundi, Democratic Republic of Congo, Ma- toria; the improvement o f the cashew nut lawi, Rwanda, Uganda, and Zambia. Third, industryinMtwara; andthe collapse ofthe Dar es Salaam has the country's highest coffee industry in Kagera, Kilimanjaro, concentration o f manufacturing and service andRuvuma. See also box 3.1. industries and very little o f traditional low- productivity agriculture. The dominance o f Dar es Salaam i s attrib- utable to three major factors. First, Dar es Ingeneral, there are many reasons behind Salaam i s the de facto seat of government the observed spatial inequalities in Tanza- andtherefore has the highest concentration nia, including economic and noneconomic o f political power, resources, and related factors. Noneconomic factors relate mainly support and ancillary activities. Second, the to historical reasons, especially colonial 47 Poverty Reduction and Growth Box 3.1. Regional Differencesin Copingwith External Shocks Inrecent years, Tanzania sufferedfrom declining internationalcoffee prices. The pricepaidto growers o f Arabica coffee inTanzania declined from a peak o f US$1.36 in 1997 to less than US$0.24 per pound in2004 (see International Coffee Organization Website, http://www.ico.org). However, the impact o fthis external shock on growth performance differed significantly across regions. A study by Gresser and Tickell (2002) indicates that declining international coffee prices have had adverse consequences on smallholder coffee farmers inKilimanjaro. These consequences include farmers exiting from growing coffee or turningto food crops andhorticulture, as well as farmers migrating to towns and miningcenters. Likewise, coffee traders gave up or closedtheir business. However, for some regions (such as Mwanza, Shinyanga, and Kilimanjaro), the decline inreal incomes, causedbyproductivity decline for the major cashcrops, was partly mitigated by substitution from coffee to other crops andnew opportunities, by business and trading inother regions, andby disposal o f assets. Inaddition, there was growth o f other sources o f income, par- ticularly from new investmentsinmining, infish processing, and inrelated services around Lake Victoria. Inthis regard, it seems plausible that the coffee-producing regions inthe southern high- lands zone did better than the northern highlands because the agricultural economy o f the former i s more diversified, with a wide range o f agricultural sources o f income (tea, pyrethrum, coffee, maize, potatoes, beans, paddy). Incontrast, the northern highlands zone i s dominated by coffee and has relatively fewer alternative crops. Therefore, a key messagehere is that the composition o f productionand the responseso fthe population to new challenges (including changes inrela- tive prices) and opportunities have had a strong bearing onthe relative growth performance o f different regions. legacy as well as culture. For example, the business services. For instance, in addition colonial administration designated some to Dar es Salaam, other regions that are regions, such as Kigoma, Mtwara, and among the top five both interms o f contri- Rukwa, as labor reserves. Following inde- bution to national GDP and in terms o f pendence, some investments in human shares in manufacturing value added, capital development, such as schools and number of industrial establishments, and hospitals, were directed to such regions. employment are Mwanza and Shinyanga. Also, because of differences in opportuni- The opposite is true of the poorer regions, ties, some regions such as Kilimanjaro have where more than two-thirds of employed developed a stronger entrepreneurial cul- people earn a living from traditional low- ture than others. productivity agriculture, livestock rearing, and fishing (see table 3.1). Regarding economic factors, variations in regional incomes inTanzania are driven in The importance o f nonagricultural activi- part by the concentration of nonagricul- ties in regional economies is also clearly tural economic activities. The major con- reflected in many other socioeconomic tributors to national GDP by region also indicators. For example, 46 percent o f have the highest concentration o f manu- electricity sold in mainland Tanzania i s facturing; mining and quarrying; produc- consumed in Dar es Salaam. The other tion and distribution of electricity, gas, and major consumers of electricity-Arusha, water; trade; tourism; and financial and Tanga, Kilimanjaro, Morogoro, Iringa, 48 Spatial Dimensions of Growth and PovertyReduction Table 3.1, Distributionof Industrial Establishments,Workers, and Value Added, 2000 and 2004 percenfage Workers Workers Value added Value added Region Establishmentsa (2000)b (2004)' (2000)d (2004)e Arusha 9.3 6.8 6.6 4.3 7.3 Coast 0.2 0.0 0.0 0.0 0.0 Dares Salaam 43.4 25.9 33.5 41.6 59.0 Dodoma 0.2 0.2 0.2 0.6 0.3 lringa 2.9 15.4 13.2 4.2 2.8 Kagera 4.8 1.8 1.8 2.6 1.8 Kigoma 0.4 0.2 0.2 0.3 0.1 Kilimanjaro 4.4 8.1 7.9 6.4 3.4 Lindi 1.o 0.1 0.1 0.6 0.1 Mara 2.7 0.8 0.1 4.4 0.0 Mbeya 3.2 3.6 3.4 3.3 6.1 Morogoro 2.9 14.3 11.6 4.3 2.9 Mtwara 0.6 0.1 0.1 0.1 0.0 Mwanza 7.2 5.0 6.8 9.2 2.2 Rukwa 0.2 0.1 0.1 0.1 0.0 Ruvuma 1.9 2.6 0.0 0.8 0.4 Shinyanga 2.9 2.9 2.9 6.0 0.9 Singida 0.4 0.1 0.1 0.2 0.6 Tabora 1.o 1.2 1.9 2.8 2.1 Tanga 10.7 10.2 9.5 8.3 9.9 Mainland Tanzania 100.0 100.0 100.0 100.0 100.0 Source: United Republic of Tanzania 2003a,2004a. Note: Totals may not equal 100.0 because of rounding. a. The number of establishments was 525. b. The number of workers in 2000 was 84,589. c. The number of workers in 2004 was 89,826. d. Value added in 2000 equaled T Sh 441,482 million. e. Value added in 2004 equaled T Sh 701,057 million. Mwanza, Mbeya, and Shinyanga) are also of the projects were in manufacturing (43 the major centers o f industrial activity. By percent), agriculture and livestock devel- contrast, the poorer regions (Coast, Lindi, opment (7 percent), construction (7 per- Rukwa, and Kigoma) together consume cent), services (6 percent), and tourism (5 only 1.5 percent of the total quantity o f percent). Similarly, the revenue collection electricity sold in the mainland. Analo- pattern also reveals that about 85 percent gously, the regional distribution of pro- o f total tax revenue comes from Dar es jects registered by the Tanzania Invest- Salaam, although that figure also reflects ment Centre from1990 to June 2002 Dar es Salaam's role as the main entry indicates that 58 percent o f the total num- point for imports and the related customs ber of investment projects was for the Dar payments. Other regions that are the main es Salaam region alone. Other regions that contributors to total tax revenue include attracted a significant number o f new pro- Arusha (3.2 percent), Tanga (2.8 percent), jects were Arusha (11 percent), Mwanza Mwanza (2.4 percent), and Kilimanjaro (7 percent), and Tanga (4 percent). Most (2.1 percent). Analogously, more than 50 49 Poverty Reduction and Growth percent o f local government authorities' where smallholder agriculture and live- own revenue i s collected by local govern- stock keeping are the dominant economic ment authorities inArusha, Dar es Salaam, activities, where the cultivated area for Kilimanjaro, Mbeya, and Mwanza. The each crop and the average farm size are pattern described above reflects the fact small, and where productivity i s very low. that concentration o f secondary activities Most of the regions that contribute least to (manufacturing and so forth) tends to gen- national GDP also have very low produc- erate a wide range o f supporting services tivity for all the major food crops com- as well as forward and backward links. pared with the regions that contribute more. Evidence also exists that major pro- Climate and uneven natural resource en- ducers o f traditional export crops have suf- dowments have also had a strong bearing fered extreme drops in real income inpart on economic growth of different regions. because o f a decline in world market That fact i s perhaps most borne out by the prices, average farm size, and farm pro- recent growth o f mining, concentrated ductivity or population pressure. However, around Arusha, Mbeya, Mwanza, and for some regions, such as Mwanza, Shin- Shinyanga, and tourism, centered in yanga, and recently Mtwara, the decline in Arusha, Manyara, Kilimanjaro, and real incomes (mainly from export crops) Morogoro. Similarly, differences in agri- was mitigated in part by the growth of cultural production (crops grown, volume, other sources o f income, particularly from farm productivity, and relative unit prices) new investments in mining, fish process- have played an important role in shaping ing and related services, and the cashew regional growth patterns in Tanzania. Ba- nutindustry. sic agricultural statistics show marked re- gional differences in the crops cultivated, The unequal contribution by regions to largely dictated by climatic conditions. For national income i s also reflected in the example, only three regions-Mara, poverty headcount ratios in the 1991/92 Mwanza, and Shinyanga-produce about and 2000/01 household budget surveys. 88.6 percent o f Tanzania's total annual The ratios indicate that poverty is more yield o f cotton. Analogously, Arusha, severe in rural Tanzania than in urban ar- Kagera, Kilimanjaro, Mbeya, and Ruvuma eas. The data also suggest that during the together produce more than 90 percent of 1990s, poverty fell in the southern high- Tanzania's total yield o f coffee. The lands but increased in the northern high- Mtwara region alone accounts for two- lands (see chapter 2 and table 3.2). thirds of the total national cashew nut pro- duction. Similarly, Arusha, Iringa, Mbeya, The observed regional income diversity in Rukwa, and Ruvuma are the mainproduc- Tanzania also involves, at least in part, ers o f legumes, maize, paddy, and wheat, demographics. Generally, regions that lie just as Arusha, Dodoma, Mara, Mwanza, at the bottom or top o f the scale for GDP Shinyanga, Singida, and Tabora account per capita also have the lowest or highest for more than three-fourths o f the total working age group, respectively. No clear number o f cattle, goats, and sheep in the relationship exists between the level o f country (figure 3.2). Significant variations poverty in a region and the net migration also exist inindividual crop-yield per hec- flows betweenregions. tare across regions and even within a par- ticular region. In general, income levels Regional growth patterns in Tanzania are are found to be extremely low in regions influenced by changes in government 50 Spatial Dimensions of Growth and Poverty Reduction Figure 3.2. Average Maizeand PaddyYield per Hectare, Fiscal Years 1995-2001 (a) Maize I Chart 2a: Awrage Maize YieldPer Hectare FY95-FWl 2500 2000 pn1500 3 1000 500 0 Regions ~ (b) Paddy I Chart Zb: AwragePaddyMeldPer Hectare FY95-FM)I 3500 2 3000 5 2500 2000 2 I500 A 1000 500 0 Regions ~ Source: Data from Statistical Unit of the Ministry of Agriculture and Food Security. policies, particularly those related to fiscal less elsewhere. Data on actual expenditure and trade regimes. A recent study using by region for 1999/2000 to 2003/04 indi- evidence from the 2000/01 household cate that although recurrent expenditure budget survey (Fan, Nyange, and Rao (dominated by expenditure on social sec- 2005) indicates that Tanzania can improve tors) has been relatively more evenly dis- the effects o f public expenditure on tributed across the regions (rightly so), de- growth andpoverty through better regional velopment expenditure has been skewed targeting. Specifically, investment in rural and has been mainly in line with donor roads i s found to have a larger effect on support preferences for the various re- per capita incomes in the western, central, gions. For example, although 11 regions and southern areas o f Tanzania and much out o f 20 together claimed more than two- 51 Poverty Reducfion and Growth Table 3.2. Regional Population Dynamics Average Population Populationgrowth Net Population household distribution rate migration density size Region 1988 2002 1978-88 1988-2002 1988 census 1988 2002 2002 Dodoma 5.5 5.1 2.4 2.3 -101.085 30 41 4.5 Arusha 6.0 3.8 3.8 4.0 141;724 20 35 4.5 Kilimanjaro 4.9 4.1 2.1 1.6 -1 24,383 83 104 4.6 Tanga 5.8 4.9 2.1 1.8 -52,168 48 61 4.6 Morogoro 5.6 5.2 2.6 2.6 30,437 17 25 4.6 Pwani 2.8 2.6 2.1 2.4 -103,912 20 27 4.4 Dar 6.0 7.4 4.8 4.3 500,621 977 1793 4.2 Lindi 2.9 2.4 2 1.4 -49,831 10 12 4.1 Mtwara 3.9 3.4 1.4 1.7 -98,689 53 68 3.8 Ruvuma 3.5 3.3 3.4 2.5 -1 5,219 12 18 4.8 Iringa 5.3 4.5 2.7 1.5 -120,198 21 26 4.3 Mbeya 6.5 6.2 3.1 2.4 46,999 25 34 4.2 Singida 3.5 3.2 2.5 2.3 -63,880 16 22 5.0 Tabora 4.6 5.1 2.4 3.6 66,370 14 23 5.9 Rukwa 3.1 3.4 4.3 3.6 38,305 10 17 5.1 Kigoma 3.8 5.0 2.8 4.8 -1 02,923 23 45 6.9 Shinyanga 7.8 8.4 2.9 3.3 6,763 35 55 6.3 Kagera 6.0 6.1 2.7 3.1 -5,980 47 72 5.2 Mwanza 8.3 8.8 2.6 3.2 -33,504 96 150 5.9 Mara 4.3 4.1 2.9 2.5 -39,878 50 70 5.5 Manyara ma. n.a. ma. 3.8 n.a. 13 23 5.2 Mainland Tanzania 100.0 100.0 3 3.1 n.a. 26 38 4.9 Source: United Republic of Tanzania 2002. thirds of actual annual recurrent expendi- pared with peripheral regions with ture by region, more than 70 percent of smaller branches. Restrictions also most annual regional development expenditure likely forced private firms to locate in was spent in only 7 out o f the 20 regions. Dar es Salaam, where they could easily The biggest beneficiary (Kagera) received maneuver with the controls. By contrast, about 27 percent annually, while the trade liberalization and removal o f trade smallest beneficiary (Rukwa) received monopolies did create a window o f op- only 1 percent. Furthermore, with only a portunity for growth o f other regions, par- few exceptions (Dar es Salaam and Kili- ticularly in the southern highlands. That manjaro), regions that received less than 2 window is partly supported by the re- percent each o f total annual development emergence o f an active private sector in expenditure also had the highest basic- crop and food grain marketing and distri- needs poverty rates. bution (for example, cashew nuts in the coastal regions), as well as new private With regard to trade policy, one could investment in the tea sector, which i s argue that during the era o f the control dominant inthe Southern highlands and i s regime (1967-86), Dar es Salaam (and a much less affected by a decline in world few urban centers), as the headquarters o f market prices compared with coffee, most o f the parastatals, did benefit more which i s the main export crop o f the from parastatal sector operations com- northern zone. 52 Spatial Dimensions of Growth and Poverty Reduction Table 3.3. Selected Indices of Human Capital Development by Region Meanmonthly consumption Life Adult literacy Total net Numberof perthousand, expectancy capita at birth Region rate (% age 15 enrollmentrateschools(2004) (T Sh2000) secondary and above) (2004) (years, 1988) Dares Salaam 91 93.1 87 21.9 50 Arusha 78 91.9 107 10.3 57 Rukwa 68 87.9 33 6.7 45 Ruvuma 84 99.3 48 9.6 49 lringa 81 99.1 82 11.2 45 Shinyanga 55 86.3 51 8.0 50 Mwanza 65 99.5 75 8.1 48 Singida 71 85.0 35 6.9 55 Mbeya 79 99.3 104 12.6 47 Tabora 65 68.2 41 10.4 53 Morogoro 72 81.9 58 10.0 46 Lindi 58 84.1 19 9.5 47 Coast 61 94.5 40 10.5 48 Mtwara 68 94.2 41 12.4 46 Mara 76 100.0 59 8.0 47 Kilimanjaro 85 100.0 160 11.2 59 Tanga 67 97.9 83 9.3 49 Dodoma 66 76.3 50 8.5 46 Kigoma 71 77.2 38 7.3 48 Kagera 64 86.8 70 9.0 45 Sources: Ministry of Educationand Culture 2004; United Republic of Tanzania 2003b. The pattern o f economic growth inTanza- civil servants (permanent secretaries, nia also reflects in part the differences in teachers, and so forth) in the relatively the level o f human capital development well-to-do regions, finding them in the (see table 3.3). The average levels o f edu- poorer regions i s a quite rare occurrence. It cation and o f general skills available in a i s also interesting to note that the forma- region are critical because they are funda- tion o f effective and operational civic de- mental to private sector development. In velopment forums has begun taking root in particular, the level o f human capital de- regions with a better human capital base velopment dictates the capacity o f individ- (Kilimanjaro, Mheya, Iringa, and Ru- ual regions to learn and adopt better land- vuma). These development forums are use systems and new farming practices, to aimed at building consensus among stake- introduce new high-value crops, or to ven- holders o f a particular region on binding ture into other business opportunities as constraints and articulation o f a shared they emerge. strategy for faster progress. Similarly, the current drive in some regions to form The general level o f education i s also community banks to deal with the problem paramount to the extent that it affects the o f lack o f financial capital is seen to be quality of leadership in a region down to much stronger in regions that have, among the village level. Indeed, though it is rather other things, a better education base (Dar es common to find village and district coun- Salaam, Mheya, and Kilimanjaro). Basic cil chairpersons who are retired senior education statistics and poverty indicators 53 Poverty Reduction and Growth Figure 3.3. InterregionalOutput Disparities in Tanzania, 1980-2003 Fig. 1: Inter-regional Output Disparities inTanzania 1980-2003 (Coefficients of Variation of GDP per Capita) 1.oo 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 1915 1980 1985 1990 1995 2000 2005 Sources:World Bank 2000b and World Bank staff computations. for Tanzania largely conform to the re- 3.3). On average, although the per capita gional variations in incomes. Overall, Dar income o f the top five regions doubled be- es Salaam, Mwanza, Kilimanjaro, Mbeya, tween 1992 and 2003, that o f the bottom and Ruvuma still rank as the top five re- five regions increased slightly faster (two gions in terms of human capital develop- and one-half times). The apparent conver- ment, whereas Lindi, Mtwara, Kigoma, gence o f regional per capita incomes in Dodoma, and Coast generally lie at the bot- Tanzania is driven by a combination o f tom of the spectrum. factors including sustained macroeco- nomic reforms, rural-urban migration, re- Evidence shows the emergence o f regional mittances, and emergence of other re- growth poles other than Dar es Salaam. gional growth poles-Arusha, Mbeya, Regional per capita incomes in Tanzania Mwanza, and Shinyanga-because o f new have tended to vary significantly over investments, particularly in mining, fish time. That variation i s revealed in part by processing, transportation, tourism and gaps between the top five and bottom five related services, manufacturing, and high- regions in terms o f the average GDP per value crops. Those new growth poles have capita between and the coefficients o f generally recorded positive net migration variation of GDP per capita for the 20 re- inrecent years. gions o f mainland Tanzania. For the pe- riod from 1980 to 2003, the coefficients o f Implications for Regional variation constructed for the 20 regions o f Tanzania suggest that income differential Policy relative to the national average has de- This chapter has highlightedsignificant in- clined over the past two decades (figure come disparities among the 21 administra- 54 Spatial Dimensions of Growthand Poverty Reduction tive regions o f Tanzania. It has also pointed quate capacity at the central level to to the most formidable obstacles to growth, evaluate projects and proposals from some that are region specific and others the various regions inorder to allocate that cut across regions. It i s also quite ap- resources to those opportunities that parent that the regions have different have the highest expected economic opportunities, some dictated in part by lo- returns or the largest effect onpoverty. cation and climate, natural resource en- dowments, and even economic history. An Supporting measures by the local au- important question then follows: does re- thorities. Local revenue collection and gional policy have arole? business licensing requirements affect the local business environment. Given the substantial disparities in eco- Strengthening the capacity and incen- nomic activities and incomes across the tives o f the local authorities to carry administrative regions of Tanzania and the out their functions in a manner that fact that regional income convergencei s far supports economic growth would be from inevitable, region-specific interven- important. In particular, those aspects tions that are based on geographically dif- should receive consideration inTanza- ferentiated growth strategiesmust evolve: nia's decentralization program and ca- pacity-building efforts through the lo- e Identi3ing and supporting growth op- cal government reformprogram, which portunities that promise the greatest to date primarily focuses on service effect on national growth and poverty delivery. In addition, local authorities reduction. The main policy implica- are also in the best position to allocate tions of providing such opportunities locally available resources to expendi- are that an effective growth strategy tures that can contribute most to a dis- must identify regional growth oppor- trict's economic development. tunities in order to provide adequate support through the provision of ade- Sharing growth through social service quate infrastructure. That identification provision. Although different growth implies a process inwhich the various potentials o f regions may lead to regions and districts compete for cen- greater income disparities, public pol- tral government-funded infrastructure icy must reduce or prevent the emer- investments and other public expendi- gence o f disparities with regard to ac- ture measures that would facilitate the cess to social services, such as exploitation of growth opportunities. education, health, and nutrition, or to That identification also requires ade- water and sanitation. Salaam, Dodoma, Iringa, Kagera, Kigoma, Kili- manjaro, Lindi, Mara, Mbeya, Morogoro, Mtwara, ' Economic activity is unevenly distributed both Mwanzan, Rukwa, Ruvuma, Shinyanga, Singida, across andwithin individual regions. Tabora, and Tanga. In2003, Arusha was split into The concentration o f economic activities around two regions: h s h a and Manyara. Zanzibar is di- vided into five regions: North Pemba, South certain centers i s also often associated with other Pemba, North Unguja, South Unguja, and Urban costs, such as traffic congestionandpollution. West. Most of the data used inthis sectionrefer to Until 2003, Mainland Tanzania was divided into the 20 regions o f Mainland Tanzania, combining 20 administrative regions: Arusha, Coast, Dar es Arusha andManyara. 55 Poverty Reduction and Growth Regional GDP data for Tanzania could be over- stating regionalincome disparities, inpart because production in the regions that contribute little to nationalGDP is largely for subsistence and is not fully capturedinmarket-based GDP numbers. The purchasing power of the shilling also tends to be higher inthe poorerregions. 56 CHAPTER4 Outlook on Growth and Poverty Reduction This chapter assesses the likelihood that also make a global comparison o f growth Tanzania will achieve and sustain high performance. Policy-based growth projec- growth rates and reduce income poverty tions indicate what growth rates seem and other dimensions of poverty. We start feasible on the basis o f the quality o f by reviewing various scenarios for Tanzania's policies and institutions. growth and poverty reduction to illustrate Growth accounting is usedto estimate the the relationship between economic input requirements of sustained high growth and the achievement o f Tanza- growth in terms o f investment in both nia's objective o f reaching middle- human and physical capital. This chapter income status by 2025 and halving in- also examines the implications o f sus- come poverty by 2010. We then assess tained high growth for structural trans- Tanzania's growth targets against Tanza- formation. The discussion o f sectoral nia's historical growth performance and growth rates provides the basis for an 57 Poverty Reduction and Growth Figure4.1. Average Annual Per Capita GDP Growth for Five-Year Periods, 1961-2005 E n 5 0 E 4 0 3 0 kn 2 0 c 1 0 $. 0 0 -m2-10 -20 m -30 " I 1961- 1966- 1971- 1976- 1981- 1986- 1991- 1996- 2001- 1965 1970 1975 1980 1985 1990 1995 2000 2005 Source: World Bank,World Development Indicatorsdatabase. assessment o f whether macroeconomic capita GDP as well as associated poverty growth projections are consistent with rates for these growth scenarios. Tanza- sectoral prospects for continued high nia's Development Vision (Vision 2025) economic growth. aims for the country to reach middle- income status by 2025.' To reach the Disaggregated poverty simulations are lower threshold for middle-income coun- then used to project the effect o f various tries per capita-gross national income growth scenarios on income poverty and (GNI) o f US$765-Tanzania's per capita inequality. Although the main focus i s on income must grow by at least 4.1 percent income poverty, the chapter also reviews annually. If per capita GNI were to grow the prospects o f reaching other National at only 2 percent annually, it would in- Strategy for Growth and Poverty Reduc- crease from US$330 to US$500 by 2025. tion (NSGRP) and Millennium Develop- If it grew by 8 percent annually, GNI ment Goal (MDG) targets, such as reduc- would reach US$1,661 (see table 4.1 and tion in hunger and improvements in figure 4.2). The results of these calcula- education and health. tions are quite sobering: they imply that even with growth rates significantly higher Growth Scenarios than Tanzania has achieved inthe past, the country will still be relatively poor in 20 We start the discussion with some simple years. illustrations o f the effect o f various growth rates on per capita income and poverty We now turnto the question o fwhat effect levels in Tanzania. The scenarios we con- various growth scenarios have on poverty. sider are annual growth rates of real per As shown in panel (b) o f figure 4.2, with capita gross domestic product (GDP) o f 2 per capita GDP growth o f 2 percent annu- percent, 4 percent, 6 percent, and 8 per- ally, poverty would decline to 25 percent cent. Figure 4.1 shows projections o f per by 2010 and to 21 percent by 2015. This 58 Outlook on Growth and Poverty Reduction Table 4.1. Projections of Per Capita GNI and Share of Population below Poverty Line, 2010-25 Average annual real GNI growth rate Year 2% 4% 6% 0% 2010 Percentagebelow poverty line 25 19 13 10 GDP per capita (US$) 372 418 468 524 2015 Percentagebelow poverty line 21 11 24.0 21.2 20.0 18.2 -Medium growth scenario 15.0 +Medium growh scenario, 0.5% annual increase in inequality &Mediumgrowthscenario,1%annualincreaseininequality .......Lowgrowth scenario, unchangedinequality 10.0 Year Source: Author'scalculations. instance, in the medium- and high-growth suses, drawn from the respective sector scenarios, the agricultural growth rates o f GDP growth rates. 3.7 percent and 5.0 percent with the rural growth rates of 5.0 percent and 7.4 per- Figure4.6 presents poverty simulations for cent. Per capita growth rates are calculated the medium- and low-growth ~ ~ ~ a r i o s . by deducting the populationgrowth rate of For the medium-growth case, the f i W e 2.9 percent-fiis i s the population growth also presents scenarios inwhich inequality observed between the 1988 and 2002 ten- increases over and above any increases 68 Outlook on Growth and Poverty Reduction associated with the differential rural and 2015. This result i s high-at least from the urban growth rates. The horizontal line in current perspective-but not unimagin- figure 4.6 reflects the MDG and NSGRP able. For instance, in Uganda inequality objectives. was 0.429 in2000/01. According to the medium-growth scenario Even inthe absence o f additional increases (the line inbold), the NSRGP objective o f in inequality, the NSGRP and MDG pov- 19.8 percent poverty incidence by 2010 erty reduction targets will be missed under will not be met. Yet the MDGobjective o f the low-growth scenario (represented by attaining the same level o f poverty five the top line), underscoring the need for at years later, by 2015, i s attainable. least medium growth if the objective i s to attain at least the MDG poverty reduction The baseline scenario incorporates differ- target by 2015. The NSGRP target o f re- ential rural-urban growth rates, leading to ducing poverty by half by 2010 can be at- a worsening o f inequality. The Gini coef- tained only under a high-growth scenario. ficient increases from 0.337 in 2001 to Under that scenario, poverty would be re- 0.352 in 2015. If one assumes in addition duced to 16 percent by 2010. that inequality increases further, not only the NSGRP target but also the MDGtarget The projections are sensitive to changes in will be missed. This is illustrated by two rural growth. Because the majority o f the lines in the middle o f figure 4.6; in them poor live in rural areas, the national pov- an autonomous increase ininequality ofo.5 erty incidence i s driven by rural poverty. percent and 1.0 percent per year, respec- A small change in rural growth therefore tively, i s added to the medium-growth leads to a substantial change in national scenario. Under these scenarios, inequality poverty. Figure 4.7 illustrates what would rises to 0.377 and 0.402, respectively, by happen to the medium-growth scenario if Figure 4.7. Projected Reduction in Consumption Poverty in Tanzania under Alternative Compositions of Growth, 2001-15 1 40'0 35.0 30.0 - 5 h 2 25.0 - 8? 22.0 20.0 -119.8 18.2 -Medium 15.0 - - - - growth scenario -Mediumgrowth scenario, with reducedrural growth 1 15.6 &Medium growth scenario, with increasedrural growth - o ~ om -o0= ro m ow ~ o 1 c o a m o - ~ ~ ~ " 0 o0 o ~0 N N 0 N 0 N N 0 N 0 N 0 N0 N 0 N 0 N N 0 N 0 N 0 N Year Source:Author's calculations. 69 Povetfy Reduction and Growth the composition o f growth were different. spite these various uncertainties, what i s It presents two scenarios: one in which very clear from these simulations is that it rural growth i s 1 percentage point lower will be difficult to attain the NSGRP and (4.0 percent), and another in which it i s 1 MDG targets for poverty reduction and percentage point higher (6.0 percent). The that enhanced rural income growth i s a overall GDP growth rate i s maintained. prerequisite to achieving those objectives. For these scenarios to occur, urban growth mustbe 9.5 percent or 3.5 percent, respec- Mainutrition tively. Under the scenario with lower rural growth, poverty declines by only ap- Progress in poverty reduction i s typically proximately 1 percentage point a year. In not only expressed by the reduction o f the the higher rural growth scenario, the de- fraction o f people whose consumption cline i s much larger-almost 2 percentage falls below the poverty line, but also points per year. measured by the reduction in the number o f malnourished people. The reason i s that This analysis illustrates the sensitivity o f consumption poverty and malnutrition are poverty reduction to growth inrural areas. quite distinct aspects o f human welfare It also illustrates the sensitivity o f these that may well respond differently to projections to modeling assumptions. For changes in income growth. Hence, the instance, by takingGDP growth as a proxy MDG for poverty and hunger reduction for consumption growth, one ignores that aims not only at reducing consumption the two may not move in parallel. In the poverty by 50 percent but also at reducing face o f high-income growth, households malnutrition, particularly the proportion o f might changetheir propensity to save. And underweight children under five years o f even ifthe propensity to save remains con- age. Underweight i s a composite indicator stant, GDP increasesmay not translate into for malnutrition, and it generalizes stunt- income increases. GDP i s measured in ing (low height for age) and wasting (low constant prices-it i s an index o f the quan- weight for age). The under-five malnutri- tity produced-but real income i s a com- tion rate, in terms o f weight for age, in posite o f quantity and prices. When rela- 1991/92 was 29.0 percent, so the MDG tive prices change, as is the case when objective for Tanzania i s 14.5 percent by terms o f trade change, one may observe 2015. GDP increases but no associated income increases (Wuyts 2005). Likewise, if the The NSGRP also explicitly aims at reduc- relative purchasing power between rural ing the prevalence o f malnutrition, taking and urban areas changes because relative as its measurable outcome the reduction in prices o f agricultural products fall com- stunting among children under five from pared with prices o f goods and services 43.8 percent in 1990 to 20.0 percent in from urban areas, then increased rural 2010. Inthis section, the focus i s on attain- GDP may not translate into increased rural ingthe MDGonhungerreduction. income. Income growth i s clearly important for Other changes may affect the pattern o f reducing malnutrition. Greater incomes at poverty reduction as well. Migration from the household level allow families to rural to urban areas i s likely to accelerate spend more on food, clean water, hygiene, poverty reduction, just as increases in and preventive and curative health care. population growth rates diminish it. De- Greater incomes allow them to have more 70 Outlook on Growth and Poverty Reduction diversified diets and to obtain more effec- ingHBSdata collected from two regions in tive child care arrangements. At the com- Tanzania in 1998 and 1999. munity level, greater income eventually leads to better access to and higher quality For income growth, the growth scenarios health care, improved water and sanitation from the previous section are followed- systems, and greater access to information. that is, per capita GDP growth o f 1.1per- Higher incomes also mean that the gov- cent per year in the low-growth scenario, ernment can collect more revenues to 2.5 percent in the medium-growth sce- spend on nutrition-improving programs. It nario, and 5.1 percent in the high-growth seems therefore reasonable to expect that scenario, Figure 4.8 clearly shows that an income growth contributes to the reduction optimistic 5.1 percent growth rate of GDP o f malnutrition. per capita and a malnutrition elasticity o f per capita income o f 0.51 (also optimistic) For the malnutrition simulations, the focus would reduce the prevalence o f malnutri- i s on the MDG objective. It i s assumedthat tion from 22 percent in 1999 to about 17 the malnutrition elasticities of real per cap- percent in2015. A simulationthat uses the ita income are 0.51 in the optimistic sce- medium-growth scenario and the elasticity nario and 0.30 inthe conservative scenario o f 0.30 reduces malnutrition from 22 per- (Mkenda 2004). These elasticities are de- cent in 1999 to 20 percent in2015, a rather termined on the basis o fresults reported by marginal decrease. This scenario seems to Haddad and others (2003) on the basis of be more realistic, ifa bit conservative, and their cross-country study. Haddad and oth- suggests that income growth alone cannot ers' (2003) findings are inthe range o f elas- resolve the malnutrition problem in Tan- ticities o f demand for nutrients calculated zania. Furthermore, the simulation did not by Abdulai and Aubert (2004a, 2004b) us- take into account the effect o f HIV/AIDS. Figure 4.8. Projected Reduction in Malnutrition (Underweight) in Tanzania, 2004-1 5 25.0 820.0 2 .-00 E . 15.0 ..._ -0 e, 5 10.0 - - . .- . Slow,0.51 -.----. Medium, 0.5 1.--0 - -.Fast, 0.51 -Slow, 0.30 -Medium,0.30 -Fast, 0.30 5.0 8 0 s0 g 0 z0 g0 g0 z0 Z0 20 20 z0 Y0 N N N N N N N N N N N N Year Source:Author's calculations. Note: The horizontalline presentshows the MDG objective of halving hunger by 2005. 71 Poverty Reduction and Growfh The disease i s likely to increase malnutri- trolling for common determinants o f nutri- tion because o f the increase inthe number tional status such as community factors, o f orphans and the decrease in the effec- individual and household characteristics tive household labor force. (including log per capita consumption), and possible endogenous factors o f program The simulations suggest that income placement (such as mean log per capita growth alone i s not sufficient to attain the consumptiodincome o f the village, nutri- MDG on hunger reduction. However, the tional status o f parents, major disaster in drop in the prevalence o f malnutrition be- the preceding 10 years, and the like). Re- tween 1999 and 2004 from 29 percent to sults o f the study show that both income 22 percent (National Bureau o f Statistics growth and the availability o f a nutrition and ORC Macro 2005) shows that large program significantly improve the nutri- declines in the prevalence o f malnutrition tional status o f children by reducing stunt- can be attained in a relatively short period. ingandunderweight (table 4.11). The observed decline couldbe the result o f much higher income-nutrition elasticities The projections carried out by Alderman, than the ones used here, but this explana- Hoogeveen, and Rossi (2006) show that it tion seems unrealistic. To attribute the ob- i s feasible to achieve the MDG on con- served decline in nutrition to income sumption poverty in Kagera (as it is na- growth alone would require an elasticity o f tionally) and that the MDG on nutrition 2.1. It i s more plausible that the decline i s cannot be achieved even under the highest the result of a combination o f income income growth scenario. Only when nutri- growth and nonincome factors such as tion interventions are available in the ma- more effective management o f malaria, jority o f villages will it be possible to improved breastfeeding, and vitamin A achieve the MDG on nutrition. Consider- supplementation. Between 1999 and 2004, ing the difficulty o fcovering all communi- the use o f mosquito nets increased from 21 ties inthe short term, the only solution for percent to 36 percent, vitamin A supple- meetingthe MDG target is to focus on per mentation went from 14 percent to 46 per- capita income growth while initiating or cent,6 and the proportion o f children who expanding nutrition interventions. are exclusively breastfed increased from 58 percent to 70 percent. At an average This analysis raises the question o f GDP growth rate o f 5.4 percent and an in- whether further declines in the prevalence come-nutrition elasticity o f 0.5 1, about o f malnutrition-beyond those attributable one-third o f the decline in the prevalence to income growth-can be expected and o f malnutrition between 1999 and 2004 whether the MDG on hunger reduction is can be attributed to income growth and the attainable. There seems scope for a posi- remainder to nonincome factors. tive response as long as nutrition interven- tions accompany income growth. Despite A study carried out by Alderman, Hoogev- the noted improvements, only 36 percent een, and Rossi (2006) for the Kagera region o f children under five slept under a mos- also concludes that income growth in com- quito net in 2004, vitamin A supplementa- bination with nutrition interventions is re- tion can be fkrther expanded, and 30 per- quired to attain the MDGs. In this study, cent o f babies younger than two months the effects of the availability o f commu- are not exclusively breastfed. Given these nity-based nutrition interventions (a child data, there certainly seems to be scope for feeding program) are measured after con- further improvements. Furthermore, the 72 Outlook on Growth and Poverty Reduction Table 4.11. Reduction in Malnutritionin Kagera Reduction in malnutrition (%) Per capita Reduction in Interventions income consumption No additional in 50% of Interventionsin all growth (%) poverty(%) interventions communities communities Underweight 37.0 42.0 46.7 Source:Alderman, Hoogeveen,and Rossi2006. Note:The simulations were done taking 1993as the baseyear. Becausethe per capita income growth rate between 1993and 2003 is known-O.7percentper year-the effectivegrowth rates requiredto attainthe 1993-2015 meangrowth ratesof 1 percent, 2 percent,and 3 percentfor the 2003-15 period are, respectively,1.3percent,3.1 percent,and 5.0 percent.Shaded cells indicate that the MDG and NSGRPobjectives are attained(that is,there is a 50 percent reduction). evidence from Kagera suggests that com- tors, such as the woman-man prevalence munity-based nutrition interventions have o f HIV/AIDS (of 1.22) show that gender considerable potential. Provided that an inequality persists . additional efhrt in nutrition-related inter- ventions i s made, the MDG on hunger re- Another encouraging fact i s the observed duction seems attainable. reduction in child mortality fkom 141 per 1,000 in 1991 to 112 in 2004. This drop Other MDGs occurs after stagnancy in child mortality rates during the 1990s (it was 137 in 1996 Progress toward attaining the MDG on and 147in 1999) and appearsto be areflec- universal primary education has been im- tion o f improved malaria management (im- pressive. As soon as universal primary proved treatment as well as increased use education was introduced in 2001, net en- o f bednets) and stronger health systems in rollment rates jumped and are now close general. The MDG target o f 47 deaths per (91 percent) to the target o f 100 percent 1,000 is ambitious, however, and meeting enrollment. this objective will remain a challenge. Close to complete enrollment also implies With respect to the reduction o f maternal gender equality in education. In primary mortality, no progress has beenmade. The education, such inequalities are small (ta- data in table 4.9 even suggest an increase ble 4.9), but in secondary education, gen- inmaternal mortality rates, but the differ- der inequalities persist. These inequalities ence i s not statistically significant. The may be on the way out: in 2004, the girl- lack o f progress i s reason for concern be- boy ratio in form Iwas 0.98, while it was cause the MDG i s ambitious (a 75 percent 0.55 in form IV. However, further moni- reduction) and maternal mortality rates toring is needed. Other non-MDG indica- high(578 per 100,000 births). 73 Poverty Reduction and Growth A recent blood sample survey suggests based growth projections provide a more that the prevalence rate o f HIV/AIDS i s 7 optimistic picture and suggest that a per percent, which i s less than was expected capita growth rate o f about 4 percent on the basis o f tests performed on blood should be sustainable inthe mediumterm, donors. There i s much variation in infec- given Tanzania's current economic poli- tion rates by age group, gender, and loca- cies and institutions. Further improve- tion. Women (7.7 percent) are more af- ments in policies and institutions are nec- fected than men (6.3 percent). Prevalence essary to scale up economic growth to the i s 5.2 percent in the 20 to 24 age cohort level o f about 8 percent targeted in the butmore than double that (10.9 percent) in NSGRP. On the input side, achieving the 30 to 44 age cohort. Prevalence rates these growth rates requires continued in- vary from a low o f 2.0 percent inManyara vestment in human resources with greater and Kigoma to a high o f 13.5 percent in concentration on secondary and tertiary Mbeya and Iringa. education, building on the significant achievements inrecent years in expanding Finally, some limited progress was made access to primary education. Sustaining in improving access to drinking water, higher growth also requires scaled-up pri- mainly because of increased access to pro- vate sector investmentand complementary tected wells and springs. Still, some 54 investment in infrastructure. Although in percent o f all rural households do not have the past gains in productivity have been access to improved drinking water, and primarily reform driven, inthe future gains much needs to be done to attain the MDG in productivity must come more often objective o f 33 percent. As far as sanita- from innovation andtechnological change. tion i s concerned, almost all urban (96 percent) and rural (92 percent) households Progressinreducing poverty depends criti- have access to a toilet. Inurban areas, one cally on the pattern o f economic growth. observes an increase in the fraction o f Because poverty i s concentrated in rural households without a toilet from 2 percent areas, a growth strategy needs to focus on in 1991to 4 percent in2000, possibly as a rural growth (although urbangrowth also i s result from rapidurbanization. important), both by reducing urban poverty and by providing opportunities for rural Conclusions populations through migration. Achieving the NSGRP target o f halving poverty by Our analysis o f Tanzania's historical 2010 requires a rural growth rate o f at least growth performance suggests caution as to 4.5 percent per year and an urban growth the likelihood that recent highgrowthrates rate o fabout 10percent. can be sustained or even further increased. To the extent that the demand-side stimu- Hunger is another important dimension o f lus emanating from increasing government poverty. Economic growth alone i s likely spendingdisappears, growth may revert to to be insufficient to reach the NSGRP and the range o f 4 to 5 percent per year, unless MDG targets for reducing malnutrition. there is a marked improvement in Tanza- Additional interventions, such as micronu- nia's international competitiveness and trient programs and community-based in- ability to diversify its economy. Policy- terventions, are necessary. 74 Outlook on Growth and Poverty Reduction 'Middle-income economies are those with a gross national income per capita o f more than US$765 butless thanUS$9,386. The CPIA is basedonBank economists' and sector specialists' ratings o f 20 items in four areas: man- agement o f inflation and current accounts, structural policies, policies for social inclusion and equity, and public sector management andinstitutions. Institutional Investor credit ratings are based on a survey o f leading international bankers, who are asked to rate each country on a scale o f 0 to 100 (in which 100 represents maximum creditworthiness). Institutional Investor averages these ratings, attaching greater weights to re- spondents with greater worldwide exposure and more sophisticated country analysis systems. The ICRG compiles monthly data on 13 political, 6 fmancial, and 5 economic risk factors to calculate risk indexes ineach o f these categories, as well as a composite risk index. Political risk assessment scores are based on subjective staff analysis o f available information. Economic risk assessment scores are based on objective analysis o f quantita- tive data. Financial risk assessment scores are based on analysis o f a mix o f quantitative and qualitative information. The political risk measure i s given twice the weight o f the fmancial and eco- nomic risk measures. Euromoney country risk scores are based on the weighted average o f quantitative indicators in nine categories: political risk (25 percent), economic performance (25 percent), debt indicators (10 per- cent), debt in default or rescheduled (10 percent), credit ratings (10 percent), access to bank fmance (5 percent), access to short-term finance (5 per- cent), access to capital markets (5 percent), and discount on forfeiting (5 percent). For items for which no data are available, the rating is 0. This method might introduce a downward bias for coun- tries such as Tanzania, for which data availability i s often poor. A nationally representative survey carried out by the Tanzania Food and Nutrition Center in July 2004 and implemented immediately after the vita- minA supplementation campaign shows that vita- min A coverage for children between 6 and 59 months i s 85 percent (TFNC 2004b). 75 PART 2 Sectoral Perspectives on Growth CHAPTER 5 Agricultural Productivity and Shared Growth At an average annual rate of 3.5 percent, Growth in Tanzania's agricultural gross Tanzania's agricultural growth exceeds domestic product (GDP) accelerated over the Sub-Saharan African average o f 3.3 the period, reaching an average rate o f 4 percent.* Sub-Saharan Africa's average percent for 1996 to 2003. Sectoral per- growth rate, although lower than the formance showed variation around the av- growth rates targeted by many African erage, but the variation was considerably countries, exceeded growth in other re- less in Tanzania than in many o f its gions o f the developing world between neighbors, indicating a somewhat greater 1990 and 2003 (figure 5.1). However, degree o f protection against severe shocks. with a population growth rate of 2.9 per- The growth was accompanied by structural cent during the 1990s, agricultural growth and policy changes that created new op- was insufficient to make a significant dif- portunities for producers and brought ference in per capita incomes and rural benefits to consumers. For example, Tan- poverty. zanian grain and horticultural producers 79 SectoralPerspectiveson Growth Figure 5.1. Average Annual Agricultural Growth, 1990-2003 4 1 3.3 3.1 3 - 2.7 2.8 2.2 2 - 1 - 0 -' -0.4 -1 I EastAsia and Europeand LatinAmerica Middle East SouthAsia Sub-Saharan the Pacific CentralAsia and the and North Africa Caribbean Africa Source: World Bank20059 Table 5.1. Sectoral and SubsectoralContributionsto GDP Growth, 1995-2003 National Contribution to accounts Annual Share of GDP growth components Subcomponents growth (%) GDP (%) (To, Manufacturing Nonagriculturalmanu- 5.8 32 39 and services facturing and services Agribusinessa 5.8 19 23 Primary Crops 3.8 36 28 agriculture Livestock 3.3 6 4 Forestryand hunting 3.1 3 2 Fishing 5.1 3 3 Totala 100 100 Source: Author'scomputationsfrom nationalaccountsdata with agribusinessshares taken from Jaffee and others 2003. a. Numbersdo not add becauseof rounding. participate in regional markets to a much it is not the fastest-growing sector. Why? greater extent than inthe past, when bam- Because its share in the economy i s large, ers to trade were higher. Because o f its andunlike tourism andthe mineral sectors, growth and its increased opennessto trade, it has a relatively low share o f intermedi- Tanzania was able to manage recent vari- ate inputs in gross value. Moreover, pri- ability inweather with minimal assistance, mary agriculture contributes indirectly to avoiding the severe food shocks that ad- growth in other sectors by stimulating de- versely affected its neighbors. mand for supplies o f goods and services to farmers (agricultural transport, storage, Agriculture has been a substantial con- and marketing, as well as consumer tributor to overall national growth. Table goods); by providing raw materials to 5.1 shows that between 1995 and 2003, processors; and by providing urban work- primary agriculture directly accounted for ers with food at affordable prices. Food 37 percent o f the country's very strong takes a high share o f urban household annual growth in GDP.* Agriculture's budgets, rangingfrom 58 percent inDar es contribution to growth is higheven though Salaamto 65 percent inother towns. Some 80 Agricultural Productivityand Shared Growth Figure 5.2. Labor ProductivityLevels in Tanzania and Comparators, 1990-2002 380 360 340 -Asia 320 - - - Developing 300 -Sub-Saharan Africa 280 -Tanzania 260 240 220 -I , I Source:World Bank World DevelopmentIndicatorsdata. Note: Countrieswith incompletedata for series are excluded.Asia developingincludesChina and India.Sub-SaharanAfrica excludes SouthAfrica. but not all o f those secondary effects o f rica, growth in Tanzania has depended on farm growth are captured in the contribu- expanding the area cultivated, but labor tion o f agribusiness to GDP growth, esti- productivity increases have been insuffi- mated at 23 p e r ~ e n t .Together, primary ~ cient to support faster growth and poverty agriculture and agribusiness contributed reduction. As shown in figure 5.2, labor roughly 60 percent o f to Tanzania's productivity in Africa has trended up only growth inGDP between 1995 and 2003. since 1994, with Tanzania's average level below that o f the Sub-Saharan Africa av- The positive performance o f agriculture, erage. Tanzania's 1.1 percent per year in- particularly in the later part o f the period, crease between 1990 and 2003 was lower i s all the more remarkable because it was than that recorded by some o f immediate achieved when global prices for major ex- neighbors, such as Malawi (5.5 percent), port commodities (coffee, cotton, cashews, Mozambique (2.3 percent), or Uganda (1.8 tea, and tobacco) were low and when the percent), as illustrated infigure 5.3. appreciation o f the shillingwas depressing producers' incentives. At 1.1 percent, Tanzania's labor productiv- ity gowth falls far short o f the levelneeded Despite the positive performance overall, to reduce rural poverty. Given the country's Tanzania's recent agricultural growth i s 2.3 percent annual increase inits agricultural not sufficient to meet the ambitious goals labor force, labor productivity must grow at embodied inthe national poverty reduction least 2.7 percent per year to reach a rate o f strategy. The strategy sets a target o f sus- 5.0 percent annual growth in agri~ulture.~ tained agricultural growth o f 5 percent per As figure 5.3 shows, such increases are rare year. Achieving this target requires a among the country's regionalneighbors. growth process that i s quantitatively faster and qualitatively different from that o f the Tanzania's experience and its strategy for late 1 9 9 0 ~even though that period was ~ agricultural growth should be seen more relatively successful. As in much o f Af- broadly in light o f global experience. 81 Secforal Perspectives on Growth Figure 5.3. Labor ProductivityTrends in Tanzania and Region, 1990-2003 percentage change in agricultural value added per worker T r e n d in agricultural v a l u e a d d e d p e r w o r k e r E a s t e r n a n d S o u t h e r n Africa: 1 9 9 0 - 2 0 0 3 - 1 . 5 0 % - 1 . 4 3 % C o n g Congo, R e p . Z i m b a b w e Mozarn bique S o u t h Africa 5 . 4 6 % 1 Source:World BankWorld DevelopmentIndicatordata. Countries that have achieved sustainedagri- countries, labor productivity rose sharply as cultural growth have done so by generating additional landwas brought into cultivation. andadopting technological change, resulting Growth inthose countries was accompanied inincreasedjoint productivity ofland, labor, by markedstructural change infarming and and capital (total factor productivity). by rapid technological change, largely in Whether the pattern o f technological change mechanical technology, that reduced labor has been labor saving or landsaving has de- requirementsinagriculture. pended on which factor is relatively scarcer.6 Countries with abundant land or Although several African countries have rapid expansion o f off-farm work have ex- an endowment o f land sufficient to follow panded the area cultivated per worker by this course, few have done so successfully, adopting labor-saving technologies, typi- at least on a widespread basis. In some cally involving improved machines, imple- cases, the property rights regime con- ments, and animal power. Giventhe relative strained access to land; in others, me- abundance o f land in Tanzania, a sectoral chanical or animal draft innovations were growth strategy reliant on expansiono f area limitedby poor access to output markets, could be considered consistent with the re- too limited a range o f appropriate and af- source endowment. It would follow the his- fordable technologies for fanners' condi- toric path o f other land-rich countries, such tions, and thin or nonexistent markets for as Argentina, Australia, Canada, the Russian long-term credit and for nonbank financial Federation, and the United States. In those services such as leasing. 82 Agricultural Productivity and Shared Growth A second path involves adopting tech- ever, a significant portion o f agricultural nologies that increase the productivity of growth involved keeping up with popula- land. Yield increasesincrops were the de- tion growth by expanding cultivated area fining characteristic o f the Green Revolu- using existing production methods. Had tion in land-scarce Asia, transforming the this growth path not been an option, food rural sectors between the 1960s and the security would have declined as the labor 1990s. Those increases were preceded by force expanded, rather than holding its hundredso f years of investment in irriga- own or improving modestly. But for the tion. The investment in irrigation comple- sector as a whole, a more widespread pat- mented the new biological technologies tern o f technological change is necessary, and reduced the risks associatedwith their one that i s adapted to regionally varied adoption, while creating a homogeneous factor endowments and to Tanzania's ex- production environment more amenable to traordinarily diverse production environ- accepting "broad spectrum" seed pack- ments. Moreover, land expansion using ages. New breeds of livestock and more existing techniques carries environmental efficient feeding regimes correspondingly costs as forests and wildlife areas are en- increased the yield o f meat and milk per croached on, as increasingly marginal land unit of land devoted to fodder. In Tanza- comes into cultivation, and as fish stocks nia, as elsewhere in Africa, increases in are depleted. the productivity o f land on the scale o f the Asian Green Revolution have been elu- Tanzania's agricultural growth path needs sive, although some progress has been to combine features o f the land-intensive achieved in the uptake o f improved varie- and labor-intensive models that conserve ties o fmaize, beans, and cassava. the resource base and thus, of necessity, will differ from the experience o f the In either of the growth paths described 1990s. Because o f the diversity o f Tanza- above, the productivity o f scarce factors in- nia's endowments and agroclimatic condi- creases, raising the returns to and the value tions, growth paths deriving from better o f those factors. A dynamic process o f ad- cultivation o f larger tracts will be optimal justment starts whereby rural households inmore land-abundant partsofthe country investineducation, implements, technology, such as the southern highlands, whereas and land. Farmsize and structure change in those associated with high yields and in- responseto economic signals. tensive cultivation will suit such areas as Morogoro and Kilimanjaro. Where an in- The agricultural growth process in the crease in area per worker i s possible (for 1990s, with its modest increases in labor example, inrelatively land-abundant areas productivity and stagnant yields of most such as Rukwa or where growing urban crops, derived largely from area expan- centers draw workers away from farms), sione7Data from the 2002/03 agricultural yield increaseswill be less necessary. The sample survey suggest that between converse applies to areas where land and 1998/99 and 2002/03 there has been no off-farmjobs are scarce. This situation ex- Wher expansion ofarable landunder cul- ists because o f an adding-up requirement: tivation. Examples o f labor-saving innova- the change in labor productivity must tion exist-for example, land expansion equal the change in land productivity plus through increased use o f animal traction in the change in area per worker. Table 5.2 the land-abundant Rukwa region. How- illustrates the trade-off, showing different 83 SectoralPerspectiveson Growth Table 5.2. Labor ProductivityGrowth: ContributingFactors Target yearly increasein Requiredannual agricultural labor Assumption regardingannualchange in land increasein land productivity(%) cultivated per agriculturalworker (%) productivity(%) 2.7 0 2.7 (no future growth in area per worker) 1.1 1.6 (futurearealworker growthsame as in past) 1.7 1.o (futurearea/workergrowth higherthan past be- cause of labor-savinginnovations) Source: Author's calculations. Table 5.3. Contributionsof Subsectors to 5.3 Percent Growth in Agricultural Gross Value of Production,1995/96-2002103 Share of agricultural gross value of Growth Subsector Growth rate production contribution Crops 5.5 87 4.8 Livestock 4.2 13 0.6 Total 100 5.3 Traditionalexports -6.9 7 -0.5 Nontraditionalexports 7.5 40 3.0 Importables 6.5 26 1.7 Nontradables 4.1 26 1.1 Total 100 5.3 Source: United Republicof Tanzania2004a. Note: Numbersmay not sum becauseof rounding.Some readerswill note that the 5.3 percent rateof growth in this table differs from the 4 percent ratecited in chapter 1. Herethe figure is basedon a differentsectoral aggregate,the gross value of production.The rate in chapter 1, which is the morewidely cited agriculturalgrowthstatistic, is basedon the growth rate of agricul- tural GDP, which netsout purchasedinputsand covers a broaderarray of productsthan does gross value of production. combinations o f change inlandper worker Such a process underpinnedKenya's suc- and change in land productivity that are cesshl smallholder-based growth in the required to meet a 2.7 percent rate o f in- 1970sand 1980s. Table 5.3 showsthat such crease inagricultural labor productivity. a process i s beginning in Tanzania, as pro- duction has shifted to nontraditional ex- Agriculture's value added can grow not ports andimport-competing commodities. only through expansion of cultivated area or increases in yield but also through A significant portiono fthe contribution of changes in the composition o f output that nontraditional exports to sectoral growth shift production out o f activities with low derives from maize (a regional export) and or negative value added into existing or pulses. These products are classified as new activities with higher profitability. nontraditional exports simply because in 84 Agricultural Productivity and Shared Growth the past they were produced largely for o f revival since 2003 (particularly cotton home consumption and trade was o f low and tobacco). This recovery may be due in volume and informal. The performance o f part to the recent depreciation o f the shil- maize i s particularly notable, because the ling in real terms. Finally, nontradables decision to open the borders to trade in (bananas, potatoes) appear to have a robust maize was very controversial, potentially domestic market. exposing the country to vulnerability in a commodity important to domestic food Assessment o f the sources o f Tanzania's security. The growth in exports o f maize agricultural growth since 1990 thus leads has not brought about domestic shortages, to the conclusion that, although aggregate because supply has grown commensu- growth i s commendable by regional and rately. Productiono f fish from Lake Victo- even global standards, that growth has ria is not included inthe aggregate, but its been o f a magnitude and nature insuffi- inclusion would increase further the con- cient to meet the country's objectives for tribution o f nontraditional exports to sec- the future. Agriculture has not lednational toral growth. One nontraditional export growth, even though the sectoral endow- crop, pyrethrum, had an impressive annual ment is quite favorable. Past growth de- growth trend o f 23 percent over the period rivedinlarge part from expanding the area from 1995/96 to 2002/03, but because o f under cultivation to meet growing domes- its low share in the total value o f agricul- tic and regional food demand. What tech- tural production, its contribution (meas- nological change occurred has not been ured in the table as rate o f growth times sufficiently widespread to induce higher share in gross value o f production) was sectonvide incomes. Improved opportuni- small. ties for trade, particularly in cereals, have led to some increase in food security and Import-competing commodities with high some shifi o f area away from products for income elasticities constituted a surpris- household consumption toward products ingly high share o f sectoral growth,' ac- destined for markets. This change in the counting for almost one-third. Examples composition o f output represents an im- o f commodities with strong performance portant target o f opportunity for the future. include rice and livestock products (milk But income per worker has not yet in- and dairy-especially yogurt and poultry). creased appreciably, and the gap in in- This finding illustrates the need for an ap- comes betweenfarming families and those proach to supporting agricultural growth inmore dynamic sectorshaswidened. that recognizes the diversity o f the sector. Maize is clearly a dominant crop, account- The lessons o f the agricultural growth ex- ing for more than 30 percent of the gross perience of the recent past have important value o f agriculture production, and it i s implications for the future. Expansion o f singularly important for food security. Yet cultivated area i s likely to remain impor- no single commodity or group o f com- tant, but it cannot be the sole source o f modities can be preselected as the one that growth, because land and labor endow- will deliver the desired sectoral perform- ments vary considerably between and ance. Traditional export crops must not be within regions. Research and marketing forgotten. Although not performing well in strategies need to be more location spe- recent years because o f low global prices cific, and the decentralization o f research (particularly for coffee, cotton, and cash- and public administration i s beginning to ews), traditional crops have shown signs support this shift. Futureexpansion o f area a5 Secforal Perspectives on Growth cultivatedmust be accompanied by greater tive o f broad-based growth. Continued investment in land improvements for both technological and financial assistance existing and newly cultivated area. In ar- from development partners will be needed, eas with higher population densities and as will reductions in industrial countries' adequate market access, more emphasis farm subsidies, which distort international needs to beput on generating and adopting agricultural markets. The private sector technologies that increase land productiv- must seize new opportunities opened in ity. Shifts inthe composition of output to- agricultural production, trade, processing, ward products with higher value added and input supply and, in most cases, can will be important to drive growth-in be counted on to do so ifthe business and some cases, the most important shifts. regulatory environment improves. These shifts will depend heavily on im- proving the outreach o f agricultural mar- These requirementsunderscore two endur- kets, as well as the capacity o f farmers to ing themes o f Tanzanian agriculture that coordinate among themselves to expand provide a strong basis for future growth. their marketingoptions andtake advantage First, farmers are a diverse group, not lim- o f economies o f scale in transport and itedto the stereotypical image o f the peas- storage. ant household with hoe technology, pro- viding its members with food and a small The imperatives of the growth process de- commercial surplus. Rather, a range o f termine the priorities for public support. households operate indiverse farming sys- Improvements in land require attention to tems, with varying degrees o f market in- land tenure issues and mobility o f land volvement. Second, Tanzanian farmers- through fbnctioning factor (especially land wherever they lie on the continuum be- andlabor) markets. Accelerated adoption of tween subsistence and commercial orienta- technology requires public investments in tion-have continually proven themselves its generation and dissemination through resourceful, market oriented, and eager to support for effective agricultural research respond to market opportunities. and advisory services, as well as public- private cost sharing for early phases o f A successfbl tripartite partnershipbetween adoption. Shifts in the composition o f out- the government, the development partners, put require well-functioning markets and andthe private sector can nurturea growth declining transactions costs. Costs o f trans- process that is more protective o f the natu- portation and communications are impor- ral resource base than inthe past and more tant, but so is the quality o f the business effective in increasing labor productivity, environment in rural areas, which either the key indicator o f improved farm in- attracts or repels competitive enterprises comes and poverty reduction. Government that can interact profitably with primary policy actions and decisions on institu- producers. tional reform and public expenditure set the context inwhich the agricultural sector To achieve ambitious sectoral targets for can grow to meet the high national expec- growth, the government needs to persevere tations-r fall substantially short. Given with policy reforms started in the 1990s the stated goal for sustained sectoral and complete the unfinishedagenda with- growth o f 5 percent per year, only an am- out regression or reversals. In addition, it bitious agenda o f reforms and well-chosen i s critical that the government rebalance public expenditure can be expected to suc- public expenditure to align with its objec- ,ceed. More limited actions require 86 Agricultural Productivityand Shared Growth amendment of the growth and poverty re- sociated rural services, including duction targets or entail recognition from finance the outset that they will not bemet. Management of Land for Agricultural Removing Constraints Growth on Agricultural Growth As shown intable 5.4, Tanzania is a large Although microlevel studies repeatedly em- country, primarily rural, with only a mod- phasize the financial and economic profit- est proportion o f land currently used for ability of farminginTanzania, as well as the agriculture. Although land appears to be benefits associated with technology adop- ample, rural areas suffer from a frequency tion, the agricultural sector appears to be far o f disputes over land and insecurity o f from a productivity take-off. Yet there are tenure that one would expect to see only in signs o f real progress in the forms o f mar- a country with a higher density o f settle- ket- and farm-level innovation andemerging ment and, hence, greater competition for institutions that canprovide a foundation for land. These circumstances derive in part accelerated productivity. The analysis o f from the colonial legacy of land admini- sources o f growth has documented some o f stration, inpart from the disruptions o f the the more visible factors. The tremendous villagization campaign o f the 1970s, and responsiveness o f Tanzanian farmers to in part from the subsequent delay in ad- changing market conditions has accounted, justing land policy and administration to inpart, for the good agricultural growth re- the needs o f a changing market economy. cord recently, inthe face o f highly unfavor- Factors limitingthe mobility o f farm fami- ableprices for traditional exports. lies from areas o f greater to lower popula- tion density may also be important. Owing to the many farm- and market-level constraints on smallholder producers, there To address the growing problems, the i s a vital, positive role for national govern- country passed the national land policy o f ment and localinstitutions inenabling agri- 1995. The policy provides more expedi- cultural growth andrural poverty reduction. tious access, enhanced security of tenure, Removal o f constraints on agricultural andbetter management of landas anatural marketing, processing, and farm productiv- resource. Implementation of the policy i s ityrequires focus onthe following: definedunder three main pieces o f legisla- tion: the Land Act, No. 4 o f 1999;the Vil- 0 Improved implementation o f land ten- ure andreforms lage Land Act, No. 5 o f 1999; and the LandDisputes Courts Act, No. 2 of 2002. 0 Expansion o f agricultural research ef- Commonly referred to as the new land fort, continued research and extension laws o f mainland Tanzania, these laws re- focus on client responsiveness and en- place the Land Ordinance o f 1953. Given gagement o f fanners in the research the large-scale movement of people under process, and strong emphasis on sus- the villagization program and the separa- tainable use o f landandwater resources tion o f households from their traditional lands, the current law recognizes that land 0 Imgation improvements rights are confirmed through long-term 0 Support for improved functioning o f occupancy, use, and development o f the output and input markets and for as- land. 87 Sectoral Perspectives on Growth Table 5.4. Land Use and Potentialfor Agricultural Land Expansion, Mid-I990s Share of Hectares total land Land use (millions) area (%) Urban 0.065 0.1 Rural protected land Protected forestlwoodland 13.838 15.0 Other protected(wildlife, national park) 13.291 14.0 Agricultural land currently used (10.8 million ha) Temporary crops 3.700 4.0 Pasture 6.150 6.5 Permanent crops 0.950 1.5 Rural unprotected land-available but not used Unprotectedforestlwoodland 26.321 28.0 Suitable for cropping but unused 7.000 7.0 Grasslandlbushland not suitablefor cropping; may be suitable for grazinga(may include some water area) 23.221 25.0 Total land area 94.536 100.0 Source: FA0 and World Bank 2001. a. As much as 25 percentof potentially suitable pasture land is affected by tsetse fly and cannot be used for cattle at present. A plan o f actions and investments to im- The planof action inthe reserve lands aims plement the laws was drawn up in 2005.9 at clarifylng and demarcating boundaries of Under the laws, all land is state property conservation areas, game reserves, and na- under the trusteeship of the president. It is tional parks to resolve concerns o f adjoin- allocated for use under varying forms o f ing villages. Land rights for pastoral and tenure, foremost o f which is the 99-year nomadic peoples remain controversial; the leasehold. The plan identifies separate ac- strategic plan for land states simply that tions for the three main categories o f land: nomadic livelihoods are not consistent with village land, urban land, andreserve land. sustainable and undisputedland rights. For all categories o f land, the program outlines For village land, attention i s focused on a process o f documentation, public educa- clarifying village boundaries to reduce tion, staffing, and capacity buildingto pro- conflicts over encroachment, surveying vide adequate attestation of landrights and and demarcating plots within the village, resolution ofdisputes. distributing certificates o f customary rights o f occupancy, establishing village The program emphasizes affirmation o f land councils, and introducing formal existing land rights and reduction o f con- working links and vertical-horizontal re- flict. This emphasis emerged from broad porting relationships between community- consultation with stakeholders leading up based organizations (CBOs), nongovern- to the formulation o f the policy and legisla- mental organizations (NGOs), and district tion, and it clearly responds to immediate landoffices. The program also seeks to set needs. The framework supports transac- upper and lower limits on plot sizes, to tions in rural land, in particular by clarify- prevent excessive fragmentation or con- ing boundaries to prevent disputes that centration o fvillage landholdings. would otherwise stymie transactions. Little 88 Agricultural Productivity and Shared Growth emphasis is accorded to detailed planning sources that the participant invests accord- for the increased volume o f transactions ing to a formula for matching grants. Par- that can be expected as the agricultural ticipants could bring their investment as economy becomesmore dynamic. their share of the matching grant and com- mit to working the land for a period of Although it calls for setting limits on the time, after which the right to conditional upper and lower sizes of holdings, the pro- use would be converted to a 99-year lease- gram does not address the practical details hold andwould be recorded as such. associated with limits. Will limits be uni- form, or will they vary by locality? Accord- The landprogram document refers broadly ing to what criteria will they be set? How to a national village resettlement scheme will they be enforced? Upper limits on the without further elaborating on the size o f holdings have been difficult to en- scheme's design. Such a program could force. Where lower limits have been en- successfully draw on experience in Africa forced, as in the restrictions on subdivision and in Latin America and would benefit in South Afi-ica under apartheid, they have from the other activities foreseen under the had negative economic and social effects land program. Those activities will assist andhavebeendifficult to reverse, evenafter inclarifying which landscouldbebrought a change inthe political order that imposed into the program on the supply side with- them. InTanzania, the proposedrestrictions out compromising commitments to re- on subdivision are intended to prevent trap- serves or the rights o f current users. Facili- ping people on small and diminishingplots tating the mobility o f land through in an environment where land is relatively transactions could forestall the need to set abundant. Yet because land i s abundant, le- legal restrictions on plot size and simulta- gal restrictions on plot size may not be nec- neously serve the growth agenda. essary if land markets can be made fluid enough and if resources can be made avail- Technological Change to Foster ableto facilitate voluntary migration. Growth: Generation, Dissemination, and Adoption The program recognizes that some aspi- rants will not be able to get land in their Given the sustained rapid growth o f the own localities if lower limits on plot sizes population, the limited creation o f em- are introduced. It thus recommends estab- ployment outside agriculture, and the rapid lishment o f a voluntary resettlement pro- increase in area farmed per worker gram to accommodatethose demands. Such (through out-migration or increased off- a program could be regarded as a home- farm rural employment), it will be a chal- steading program, open to any interested lenge to achieve more rapid growth in la- and qualified participant seeking to move bor productivity. Sustained increases in where landis available for agricultural pro- labor productivity and rural incomes there- duction. The program could be designed as fore depend on more rapid technological a public-private partnership in which the change that i s either labor saving or land public sector makes land available; pro- conserving, depending on local endow- vides aright to conditional use for aninitial mentsandmarketingoptions. period; assistswith costs o f relocation; pro- vides infrastructure to meet defined mini- A successful dynamic o f technological mal standards (roads, water, public services change in the smallholder sector starts if not already offered); and leverages re- with increased productivity in the prod- 89 Sectoral Perspectives on Growth ucts that farm households produce for Producers must have information their own consumption. Often a distinc- about the availability o f the varieties tion i s drawn between food crops and and guidance onhow to use them. cash crops, but this distinction i s no longer helpful because most food crops Money must be available to cover the now serve a cash crop function. Maize costs o f early adoption o f the new and cassava are often considered food varieties. crops, as opposed to cash crops such as cotton and coffee. Yet maize and cassava All three elements-the varieties, the have considerable importance as cash knowledge, and the financing-must be crops because market demand exists for linked in a way that provides access for them inraw and processed forms. smallholders. As producers realize productivity gains in Tanzania has invested in elements o f the products for their own consumption, they technology system with some success. In- can sequentially reduce the area devoted deed, past investments in research and ex- to that consumption and increase the area tension explain part of the good recent per- devoted to products for markets: This formance o f the sector. During the early shift may mean simplymarketingmore of 1990s, agricultural services consisted pri- the same product or taking advantage of marily of centralized, supply-drivenpublic opportunities to shift into products with services through the then Ministry o f Agri- higher profitability. Increased productiv- culture and Cooperatives (MAC)." Crop ity allows smallholders to move incre- and livestock services were integrated and mentally from subsistence to commercial organized around three main domains: re- farming. search, training, and extension. Technical services were handledseparately. Higher profitability translates into higher earnings, with increased savings, invest- As the 1990s progressed, shortcomings in ment in land improvement, and ability to the technology system became clear, and finance inputs for profitable production in adjustments were made in particular as- the next season. Demand for locally pro- pects. The technology transfer model inher- duced nonfarm goods and services also ent inthe Training and Visitation extension increases-for tin roofs, pumps for water system was found to be unsustainable be- management, implements for conservation cause o f the high costs o f service delivery. tillage, and other items. The initial in- Extension approaches barely took into ac- crease in productivity thus multiplies to count the concerns, needs, and involvement support greater dynamism in agriculture o f farmers. As a result, the majority o f the and localrural growth. farmers either did not access the services or, if they did, often found them irrelevant. Three elements must be present in the Decentralization o f research was initiated technology system to trigger the initial inthe early 1990swiththe establishmentof round o f increasedproductivity: seven agricultural research zones. The farming systems approach was adopted in 0 Researchers must have developed new research operations to strengthen the link varieties and production techniques between researchers and farmers. That ap- that are profitable under the conditions proach was followed by the introduction o f that smallholders face. the client-oriented research management 90 Agricultural Productivity and Shared Growth approach in selected zones. Research for and substations, which cover the main ar- coffee, tobacco, and tea was privatized and eas o f crops and livestock research in the hnded largely by a combination o f direct country. The main research effort comes donor support and a cess on producers. In fiom seven zonal research and develop- 2000, the former MAC was divided into ment centers (ZRDCs), each comprising a three ministries: the Ministryo fAgriculture lead station and associated centers and andFood Security (MAFS), the Ministryo f substations. The ZRDCs are located in the Water and Livestock Development seven agroecological zones and are re- (MWLD), andthe MinistryofCooperatives sponsible for both applied and location- and Marketing (MCM). This division specific adaptive research and training. fiagmented some o fthe research efforts and There are also a few institutes that are not introduced new requirements for coordina- under the zonal research setup and that tion (for example, livestock and crops for undertake specialized research work. They integrated farming systems were handled include the Animal Diseases Research In- by different ministries, as were crops and stitute in Temeke, Dar es Salaam, and the investments inirrigation schemes). In2005, Tsetse and Trypanosomiosis Research In- the three ministries were consolidated into stitute in Tanga. These institutes have a two, namely the Ministry for Agriculture, national mandate and are responsible for Food Security, and Cooperatives and the the design and evaluation o f nationwide Ministryfor Livestock. research programs in collaboration with the zonal centers. Research undertaken in the 1990s ad- dressed varietal improvements in the crop Throughout much o f the 199Os, with the sector, new breeds in livestock, pest and support o f several development partners, disease management, improved manage- including the World Bank, Tanzania's ag- ment o f soil fertility, and reduction o f ricultural extension program followed the postharvest losses. A summary o f selected traditional and centralized model o f train- achievements appears intable 5.5. ingandvisitation. By the end o fthe 1990s, evidence emerged that the cost o f the pro- These and other research results were pro- gram was high and unsustainable and that duced by the National Agricultural Re- its effectiveness was limited. At the same search System (NARS), which comprises time, commitments to decentralize service various public organizations-namely, the delivery under the local government re- Department o f Research and Development form program made the traditional admin- (DRD), the Tropical Pesticides Research istrative framework for extension unten- Institute, several universities, and the Tan- able. In December 2000, a new approach zania Forestry Research Institute-along to extension was developed under the Vi- with private sector organizations, which sion and Strategy Outline to Year 2010. include crop research institutes for to- (MAFS 2001). It involved an incremental bacco, coffee, and tea. The DRD o f the shift to an extension service that is respon- Ministry o f Agriculture, Food Security, sive to clients, pluralistic inthe delivery o f and Cooperatives is the lead institution o f services, cost-effective, and sensitive to NARS. It is tasked with the public role o f the gender mix in Tanzania's agricultural conducting, coordinating, and directing sector. agricultural research in the country. The DRD operates a network of more than 50 Several pilot activities helped to make the research institutes and associated centers vision a reality. The pilot initiatives were 91 Sectoral Perspectives on Growfh Table 5.5. Inventoryof TechnologiesComing Out of the Research System in the 1990s Innovation Existingor pipelinetechnologies Maizevarietal improvement Maize gray leaf spot-tolerant and high-yieldingvarie- ties (UH 615, UH 6303) have been developed and adopted widely in high-altitudeareas with severe crop loss. Cassava Cassava mosaic disease-tolerant clones have been developed. Cassava mealy bug reducesyield by 80- 100%. Cassava postharvestlosses Improved processingequipment has been tested and recommended and is now available. Bean disease resistance Improvedvarieties Uyole 94, Uyole 96, Uyole 98, and Kabanimaare high yielding and tolerate diseases. Bean multiplication Bean ImprovementProgramsupports seed multipli- cation. Bean maggotcontrol Combinationof integrated pest managementand cultivation techniques reduces populationof maggots and magnitudeof loss. Tomato seed development Tanya and Tengeru 97 tomato varieties increase yields and incomes and reduce dependence on im- portedseeds. Cashew disease Plant protectionagents (sulfur, organic fungicide An- vil, Bayfidan,and Topas) reduce powdery mildew disease; 20 clonal varieties have been identified as potentially high yielding and tolerant of diseases and pests. Banana pests Nematodesand weevils are managedthrough appli- cation of plant protectionagents, varietal improve- ments, and cultural practices. Deforestationin tobacco areas Varieties suitable for fuelwood and enhancementof soil fertility have been developed and introduced through agroforestry. Cattle breeds Dual-purposeMpwapwacattle have been introduced in pilot villages, and milk production increasedfrom 1-2 liters/day to 5-7 literdday under farmer man- agement. Goat breeds Dual-purposegoats with live weights of up to 45 kg and milk yields of 1.5 liters/day have been intro- duced. Newcastlediseasefor chickens Thermostablevaccine (1-2) is available for control in rural areas. Ricevariety Popular aromatic variety has been developed. World Bank 2004e. designed to test modalities for training, de- vice, as staff at national and district levels livery o f technical advice, and strengthening sorted out the implications o f decentraliza- o f support for marketing and economic deci- tion and pilot efforts in alternative modali- sion making, as well as to provide technical ties of service delivery were designed and expertise. The period fi-om 2000 to 2005 implemented. Despite some institutional was a challenging one for the extension ser- c o d h i o n and scarcity o f resources because 92 Agricultural Productivity and Shared Growth o f mismatches between the fiscal and the vate partners was limited. Disproportion- administrative decentralization, extension ate emphasis was placedon generating and continued to function and, together with the disseminating technology, and less on em- research establishment, contributed to the powering farmers (in terms o f both skills growthinoutput over this period. and finance) to adopt the technology. As a consequence o f weak links in the system, Farmer organizations, together with re- research did not always focus on tech- search and extension services, are key con- nologies that had the greatest potential ef- tributors to the technology system. MVI- fect on production systems. Technical WATA (Mtandao wa Vikundi vya breakthroughs did not yield good eco- Wakulima Tanzania), the country's main nomic returns. And promising technolo- farmer group network, has built an impres- gies remained on the shelf because o f lack sive organization up to the national level. o f knowledge or financing for adoption. Since its formation in 1993, MVIWATA Ruptured links in the technology chain has expanded to cover 120 local networks reduced returns to investments in each o f with some 1,000 affiliated groups in more the elements. Moreover, the system was than 80 districts (representing some 50,000 underfinanced-but given the somewhat to 70,000 households). A number o f other depressed returns resulting from the insti- groups operate at the community level, ei- tutional deficiencies, underfinancing was ther independently or with NGOs. They an appropriate response. have playedan active role inthe generation and transfer o f technology." Farmer Under the Agricultural Sector Develop- groups improve access to technology (for ment Programme (ASDP), a major reform example, through experiential learning, as o f the institutional structure o f the agricul- infarmer researchgroups and farmer field tural technology systemhas beendesigned school groups); to access funding (for ex- and is being implemented. The reform ample, through credit and saving groups); embodies the following guidingprinciples: for crop processing and marketing(through commodity marketing groups); for live- 0 Client and farmer empowerment. stock production (through dairy or poultry Through knowledge, control o f funds, groups); for gender-basedactivities; and for influence on organizations, and institu- support to members in case o f need tional change, farmers acquire the ca- (through indigenous and traditional, reli- pacity to analyze their constraints and gious, and cultural-based groups). identi@ opportunities, articulate their needs, exchange knowledge, access the The contribution o f agricultural research, services they need, become active extension, and empowerment o f farmers in partners, improve their bargaining the 1990s andearly 2000swas positive but power, and have final jurisdiction over also limited by several factors. Services the disposition o f funds provided to generally focused on increasing produc- ensure that they receive the services tion through short-term technical pack- they need and demand. Farmers' or- ages, without paying sufficient attention to ganizations and CBOs and networks farmers' circumstances, markets, and sus- are promoted and strengthened to be- tainability. Despite various attempts to come key development partners. strengthen them, the links between re- search, extension, and training were weak, 0 Demand-driven and market-led tech- and collaboration between public and pri- nology development and adaptation. 93 Sectoral Perspectives on Growth Farmers select, test, compare, and ing is made available at each level, as, for adapt appropriate technological, ser- example, through the competitive Zonal vice, and market options. Agricultural Research and Development Funds. Pluralism of providers of services and approaches. Diverse methodologies, At the village level, farmers are encour- processes, and funding, as well as ser- aged to organize into voluntary groups and vice providers, are supported. Public are empowered to articulate demand for- funding for the system remains impor- and control over-agricultural services. tant, but services can be provided by The interventions build on successful public extension workers, NGOs, and grassroots, bottom-up initiatives and op- private advisers. erations, such as MVIWATA groups, Subsidiarity. A constructive division o f farmer field school groups, and the like. labor between the national, district, Farmer groups will continue to access and local levels is maintained. At the public services, while also contracting national level, the extension service more private service providers. feeds knowledge into the system At the ward level, a Ward Farmer Forum through provision o f training, identifi- will be established to aggregate, priori- cation o f new approaches and tech- tize, and present demands for services. nologies, and preparation o f materials. Service standards are also defined and The ward extension team will typically enforced at the national level. At the comprise one to three government exten- sion staff members based at the ward or local level, organizations contract the village level, as appropriate, depending most suitable service providers, both on the degree o f coverage o f private and public and private. NGO services. The extension staffwill be Focus on economics and natural re- trained to facilitate agricultural develop- source management, HIV/AIDS and ment, both providing services directly malaria, and technical solutions. Agri- according to their expertise and assisting cultural service providers assist in is- groups in accessing other services as sues related to economic decision mak- needed. ing and management o f soil, forests, and water. At the district level, each district council and administration develops a District Ag- Transparency and accountability. Ac- ricultural Development Plan (DADP), to countability is built inthrough perform- be funded in part through a conditional ance contracts, performance monitoring, grant called the District Agricultural De- andthe ability o ffanners to choosetheir velopment Grant (DADG). The grant will providers. Farmers' feedback on ser- pay for locally financed services, for in- vices i s integrated into the periodic vestment in local infrastructure o f high evaluation o fserviceproviders. priority (such as roads and bridges to im- prove marketing), and for cost sharing for The reforms in the system operate at the adoption o f new technologies. The staffing village and ward levels, the district level, o f the District Agricultural Sector Office the zonal level (through the zonal agricul- will be revised to reflect the new func- tural research stations), and the national tions. The role o f the DADGs i s very level (through the line ministries). Financ- promising because it brings together the 94 Agricultural Productivity and Shared Growth administrative and fiscal architecture o f The National Irrigation Master Plan con- the local government reform and institu- firms that abundant groundwater i s avail- tional reforms inagricultural services. able in several regions-for example, in the volcanic areas o f northern and south- At the national level, the agricultural sector ern Tanzania as well as inthe sedimentary line ministries are responsible for policy, coastal basins (JICA 2002). regulatory, and planning functions. Jn the institutional restructuring to support agricul- Tanzania's ample water i s matched by tural services and investment at the district ample land suitable for irrigation. O f the level under the ASDP, the fragmentation of 44 million hectares suitable for agricul- the line ministries at the national levelis dif- tural production, only 10million are under ficult to justify (because core elements of cultivation and only 200,000 are irrigated. their previous responsibilities have shifted to This fraction represents a mere 2 percent the district levels) andi s costly. o f the cultivated area in the country. It i s estimated that up to 2 million hectares Irrigation to Raise Productivity could be irrigated. Approximately three- and Incomes fourths o f the area currently irrigated i s Tanzania i s well endowed with water, both farmed by smallholders in about 600 on the surface and below it, but the coun- small-scale irrigation schemes, usually us- ing diversions and hrrows in one o f the try suffers nevertheless from water short- nine major river basins. Very little irriga- ages because o f insufficient capacity to store and access water. Cumulatively, the tion i s drawn from groundwater; this lack lakes, wetlands, and aquifers provide huge o f activity i s a promising area for future natural storage capacity. The country also development, with direct and affordable has 2.7 million hectares o f wetlands (in benefits to the poor. Rice is by far the most important crop irrigated in Tanzania, Usangu and Malagarasi). Total renewable water resources are estimated at about 80 but sugarcaneis also irrigated. cubic kilometers per year, o f which 30 cu- Irrigation i s constrained by the afforda- bic kilometers per year i s groundwater bility of the investments required and by (FA0 2004). Several hydrological studies the profitability o f their use. Even the indicate locations where water tables are relatively modest implements needed for shallow andwater yield i s potentially quite localized access to groundwater are more significant (JICA 2002). Some o f the areas expensive in Tanzania than in, for exam- with highgroundwater potential include ple, India-by a factor o f about three (FA0 1997). Researchers from the United Makutupora inthe Dodoma region and Kingdom's Cranfield University found Ruvubasin inthe Coast region that "In Africa, the cost o f a borehole Sanya-Hale plain inthe Pangani basin drilled by a truck-mounted rig can be ex- tremely high, costing as much as 10 to 20 Arusha and the Karoo sandstone inthe times the cost o f the drilling and pump in Tanga region Asia. High unit costs mean that too few wells are drilled and communities and Fault zones around Kilimanjaro farmers remain dependent on interna- Parts o f Morogoro, Iringa, Mbeya, tional aid programs for this form o f infra- Mtwara, and Lindi structure development" (Carter 1999). And to compound the adverse impact o f 95 SectoralPerspectiveson Growth high initial costs, producers face difficul- Drillinga borehole, along with install- ties in accessing high-yielding varieties ing pumps and pipes, costs up to and moving products to market. Irrigation US$8,000 inTanzania. and agricultural productivity are clearly intimately linked, and neither can ad- Rainwater harvesting. Rainwater har- vance substantially independently from vesting, which has received greater at- the other. tention in recent years, refers to the small-scale concentration, collection, A suitably designedgroundwater irrigation storage, and use o frainwater runoff for system could reduce reliance on large bod- both domestic and agricultural use. ies o f water, including rivers and lakes, Rainwater runoff can be stored behind andcouldpromotemore sustainable use of bunds or in tanks. In areas with vast locally sourced and managed irrigation terrain and gentle slope, it is possible systems. Surface water available varies to construct small and mediumtanks to with rainfall, so open wells and borehole collect and store water during heavy tubewells can be constructed to spread the rains for supplemental irrigation dur- availability o f water throughout the grow- ing the rainy season and full irrigation ing season. Compared with large surface duringthe dry season. These tanks can irrigation schemes-the design o fwhich is also help recharge aquifers, which feed driven by topography and hydraulics- into open wells or tubewells. The aver- groundwater development i s often much age cost o f a tank to store about 30,000 more amenable to poverty targeting and i s cubic meters o f water i s about generally less capital intensive. US$4,000; that tank will irrigate about 2.5 hectares o fpaddy or 5.0 hectareso f Groundwater irrigation can complement h i t s and vegetables. irrigation using surface water. Integrating groundwater extraction with rainwater These costs are very high. In South Asia, harvesting and watershed management, rapid groundwater development has sup- along with efficient water distribution sys- ported a booming pump industry, which i s tems, leads to reliable, cost-effective irri- characterized by both economies o f scale gation systems. and intense competition. As a result, South Asia's rural poor have benefitedfrom low Groundwater can be extracted and distrib- costs for drilling and pumps. In Tanzania, utedinthree ways: however, pump irrigation development is limitedand the costs ofdrilling andpump- Open wells. Over a broad part o f the ing equipment are beyond the reach of country groundwater development has smallholders. In Nigeria, by contrast, concentrated mainly on shallow open technical innovations reduced the cost of wells for domestic purposes. The Min- constructing shallow tubewells by about istry o f Agriculture estimated in two-thirds between 1983 and 1990, with a 1996/97 that the cost o f digging a well commensurate increase inactivity. i s about US$3,000. Tubewells and open wells are easy to Borehole tubewells. Shallow tubewells manage and can irrigate up to 5 hectares. can be drilled by hand with simple The maintenance o f the pump or well is tools similar to soil augers, by power local, and farmers have a direct stake inits rotary drilling, or by a drilling method. upkeep and usage. Individuals or groups o f 96 AgriculturalProductivityand SharedGrowth Table 5.6. Institutional Frameworkfor Sustainable Development of Smallholder Irrigation Systems Requiredconditions Targets Technicalself-reliance Capacityof IrrigationDepartmentstaff, local governmentauthoritiesstaff, and extensionworkers Farmers knowledgeableabout water managementand operationsand maintenance Appropriatechoice of technology Attention to environmentalissues Financialself-reliance Rationalizationof tax regimefor small farmers Betteraccessto financialservices, especially savings Privatefirms active in supplying equipmentand implements Institutionaland Clarityon rolesand responsibilitiesof public servantsat district and organizationalsupport nationallevels Strengtheningand reformof IrrigationSection,zonal irrigationunits,and local governmentauthorities Legal attentionto landtenure,water rights,and ownershipof and re- sponsibilityfor irrigationinfrastructure Improvedaccessto advisory services Capacityto collectwater fees and pay operationsand maintenancecost Investmentclimate to supportgrowingconstellationof small firms manufacturingequipmentand providingservices Source: ASDP Working Group 2 2004. farmers can invest on their own initiative. tions to provide technical know-how, sup- Efficient distribution methods such as port agriculture research and extension, hand or treadle pumps and drip irrigation improve land tenure, and develop markets can lift and distribute water. Because andinfrastructure (table 5.6). pumpingwater, whether it is done manu- ally or with electricity, involves direct Ifsmallholder groundwater irrigation is to costs to farmers, the expense encourages be sustainable, watersheds must be well efficient use. managed to ensure adequate recharge. Groundwater recharge may also involve To promote the expansion o f smallholder large areas and several communities. irrigation, poor farmers must have access to cost-effective irrigation technologies Access to land and security o f tenure re- that provide a rapid return on investment. main important constraints on the expan- They must also have a reliable supply o f sion o f irrigation. The efforts noted above improved crop varieties and other inputs, to implement land laws, enhance tenure, as well as land tenure rights and markets and reduce conflicts over land all support to absorb increased production. Public and the expansion o f irrigated area. Expanding private investments in the assessment o f the area irrigated before clarifying the ten- the supply and quality of groundwater, in ure issues that now spawn conflict will the regulation o f groundwater extraction, increase the level o f conflict along with and in technical support to farmers are the value o fthe underlyingasset. needed to enable groundwater irrigation- and thereby help reduce poverty. These Inview ofthe importance attachedto irriga- efforts may require institutional interven- tion, the MAC formulated and adopted the 97 Secforal Perspectives on Growth National Irrigation Development Plan in nent, linking the various subsectors that 1994, with support from several develop- use water (such as domestic supply, irriga- mentpartners. The planrecognizes the defi- tion, hydropower, and environment) with ciencies o f past interventions and assigns water resourcesmanagement. A new water highest priority to rehabilitation o f and low- sector development strategy, along with cost improvements to existing schemes and legislation and an institutional framework, traditional water harvesting systems. The i s being finalized. planstresses (a) according the highest prior- ity to the rehabilitation or upgrading of ex- Policy reforms in several key areas are isting irrigation schemes; (b) upgrading tra- neededto underpinthe expansion of irriga- ditional water harvesting technology where tion. Administrative regulations andrestric- more intensive irrigationis not possible; and tions on marketing and trading irrigation (c) investing innew smallholder schemes in equipment should be simplified or re- those regions where conditions are appropri- moved. Procedures for importing irrigation ate andtraditional schemes do not exist. equipment4lling machines, pumps, and so forth-should be simplified(byreducing The institutional support for smallholder the import duty, providing an import sub- irrigationdevelopment inTanzania involves sidy, and so on). Instruments such as mi- key ministries, district authorities, agencies, crofinance lending, matching grants, and and CBOs. MAC support for smallholder joint ventures should include among allow- irrigation is provided through zonal irriga- able projects the smallholder groundwater tion units o f the Irrigation Department, and irrigationprojects. To support the pursuitof through a limited number o f irrigation tech- additional financing, the development part- nicians and agricultural extension staff ners and the government could join forces membersinthe districts. The units have lim- with NGOs such as International Develop- ited planning, designing, and supervision ment Enterprises to enhance the supply of capacity. The Irrigation Department is ex- low-cost irrigation equipment (for example, pected to provide coordination and policy treadle pumps and drip irrigation). Finally, guidance through its headquarters, as well as actions that affect the profitability o f irri- specialized technical services on site selec- gated agriculture, such as the ongoing ad- tion, survey and design o f irrigation infra- justment in the real exchange rate, contin- structure, and supervision o f scheme con- ued liberalization o f trade, and reforms o f struction. The MWLD is responsible for crop boards, will also acceleratethe expan- collecting and analyzing hydrological data sion o fthe areairrigated. for the development o f water resources and for the issuanceo fwater rights.However, its Irrigation presents a good array o f oppor- capacity to fulfill these functions at the dis- tunities for partnerships between the pub- trict leveli s extremely limited.Private sector lic and private sectors. Open well or involvement is expected to increase gradu- tubewell investments are a private good ally and to concentrate on infrastructure that should be the responsibility o f the construction by contractors and artisans and beneficiaries o f the investment. The public capacity building o f water user associations sector role should generally be limited to (WAS)by specialized training institutions establishing a conducive policy and an in- andNGOs. stitutional environment for investment. Direct subsidies for tubewell drilling and The new water policy, formulated in2002, operation are best avoided unless there i s a i s among the most advanced on the conti- compelling poverty reduction argument 98 Agricultural Productivity and Shared Growth for the subsidies. One-off matching grants bodies, drawing their membership from may be useful in situations o f great pov- the water users in a particular locality. erty and poorly functioning financial mar- They link with individual users and with kets. Sharing investmentcosts between the irrigation organizations that represent their private and public sectors through public- members. The role of catchment water private partnerships would reduce the of- user organizations i s to manage the alloca- ten heavy burden on the public budget. tion o f water resources at the local level Models for private investment include di- and the equitable allocation o f these re- rect investment in production (including sources during droughts. They also assist investments by smallholder farmers), pro- inresolving local disputes. Inareas where vision o f irrigation services (including op- W A Shave been formed, they have been eration and maintenance o f infrastructure), extremely useful in finding solutions to leasing o f irrigation technologies, and in- water management issues. frastructure development. Agricultural Marketing and Producer Water management can best be structured Prices when governments are responsible for the main infrastructure and farmers are re- Agricultural marketing and intermediary sponsible for management and oversight costs have decreased over time as policy through local bodies, such as W A Sand reforms during the 1990s reduced price distribution boards. In areas where interventions and eliminated monopoly groundwater irrigation predominates, purchases by government bodies, allowing W A S are valuable for organizing hard- greater scope for private sector trade and ware and infrastructure maintenance. investment. W A Scould hold community water rights and oversee water use among their mem- Traditional export crops. Despite the bers. Aquifer management organizations increase in market entry and the invest- under the umbrella of the river basin ment associated with reforms, real pro- committee or authority can complement ducer prices for traditional export crops effective local organizations for managing didnot increase significantly from the pre- groundwater. reform and early reform period (1987-94) to the period o f most rapid reform (1995- Many W A S in Tanzania are compara- 2002). The recent crop board study, which tively new and still weak. Such associa- identified the various factors influencing tions require significant investments in producer prices, nonetheless documented a capacity strengthening and in defining reform-induced increase in the producer's roles and responsibilities. Transfer o f irri- share o f the world price, ranging from 21 gation management responsibility from the percent to 26 percent for the four largest government to W A Scan be done effec- traditional export crops (coffee, cotton, tively only when accompanied by suffi- tobacco, and tea). The reason producer cient support for capacity building. price levels did not increase i s that this gain was offset by a 26 percent apprecia- W A Sprovide the mechanism for regis- tion in the real exchange rate and re- tering the demands and rights o f all users inforced in some cases (tea and cotton) in a particular catchment. Under the new by smaller reductions in the real world water policy, they are legally constituted price. Sectoral policies thus had a sijpifi- 99 Sectoral Perspectives on Growfh Figure 5.4. Producerand IntermediaryReturns as a Percentage of the Border Price Cashews Coffee Cotton Tobacco 1 Coast Mlwara Kagera KillmanJaro Shinyanga Tabora hlkua nraka a Mowo Nyeni itund lpung G 100 8 80 'E 42 P - 60 8 40 c 08 20 0 Source: Nyange2005. cant positive effect on producer incentives, costs that might be reduced even further offsetting the negative effect o f real ex- through improved public investment, poli- change rate movements. The reforms cies, and public-private sector collabora- changed the institutional environment in tion. Although the motivating hypothesis the marketing o f traditional export crops, o f the survey was that hightransport costs leading to an influx o f private traders, in- drive a significant wedge between produc- creased direct payment o f cash to produc- ers and the border, the survey found great ers, and in some cases (cotton and coffee) variability across crops and locations. Be- a significant increase inprivate investment cause cost structures vary, so do the spe- at the processor level. The chief negative cific remedies for highor unstable market- effect was a disruption o f input supply and ing margins, and so will the role o f public financing. and private bodies inrelieving constraints. A variety o f crops were analyzed, with Overall, the reforms helped producers farm-to-border marketing chains assessed cope with a challenging external environ- for cashews (inMtwara and Coast), coffee ment. A World Banl-supported survey in (in Kagera and Kilimanjaro), cotton (in December 2004 analyzed the cost ele- Shinyanga), andtobacco (inTabora). ments in marketing four traditional export crops: coffee, cotton, tea, and cashews. Reducing the large margins between the Figure 5.4 shows that despite the im- farm and the border will have a large posi- provement in the producer's share o f the tive effect at the fann level when markets border price, which implies a narrowing o f are reasonably competitive, passing the margins, marketing margins remain sig- savings through to farmers. Similarly, an nificant relative to final prices, accounting increase in the border price has a poten- for 30 percent to 50 percent o f the border tially large positive effect on producers, value. The results point toward a variety o f because a given change in the border price 100 Agricultural Productivity and Shared Growth Figure 5.5. Cost Componentsof Marketing Margins Arabica coffee Robustacoffee Tobacco Cashews Cotton Kilirnanjaro Kagera Tabora Mtwara Shinyanga 100% nn'nn'nnI 0 Traders' Margin 90% ITaxes 80% 70% IFeesand commissions 60% 0 Finance 50% 40% &I Processing 30% Packaging 20% material 10% ITransport 0% Source: Nyange 2005. represents a much larger percentage prices for farmers through cost-reducing change inthe farm price. measures, as well as the scope for such re- ductions, depends on local conditions and The components o fmarketingmarginsvary the specific crop under discussion. Inmany across crops and locations (figure 5.5). cases, the specific remedy requires knowl- Large items can include taxes and fees, the edge of the local cost constraint and of trader's margin (trading revenue minus out- trade-offs between the removal o f the con- of-pocket costs), the processor's margin, straint and other objectives (such as tax transport, finance, or packaging materials. revenue needs). Thus, district-level growth I s there scope for reduction in these items? strategies play an important role inidentify- In some cases, perhaps not: for example, ing market-levelconstraints on growth and packaging material costs may simplyreflect possible solutions. Insome cases, localbod- the full import cost o f the required materi- ies can address the problems on their own. als. But for other items, it may be possible Butinmanyothers, theyneedto collaborate to reduce costs through public sector in- with outside service providers and work vestments (intransport or inpower and wa- withnational or regionalbodies. ter, which are particularly important for processing) or policy (tax and regulatory). The importance o f largetrading marginsin Clearly, the public sector role inimproving some marketing chains requires special 101 Sectoral Perspectives on Growth comment. Large margins, especially im- an important instrument for identifying portant for cotton and tobacco, do not rep- constraints and strategies to remove them. resent traders' profits. They represent the trader's revenue net o f out-of-pocket costs, Transport costs were hypothesized in this but they do not subtract the costs of risk; study to be the most important cost ele- the costs o f time spent on coordinating, ment, but it is clear that, although they are searching for, and collecting market in- significant in some cases and require at- formation; and, more generally, transac- tention (as for tobacco, cashews, and cot- tions costs. Those other costs, which are ton), they are not the primary cost factor. likely to be significant, must be accounted For cashews, transport costs assume more for in order to assess the profits and rates importance in Mtwara than in Coast, but o freturnto traders. field visits in Lindi and Mtwara indicated that processing improvements that enable Public decisions or actions may be neces- the export o f value-adding cashew kernels sary to reduce transactions costs. Such a are likely to be important as well. The im- reduction could occur through some com- portance o f such innovations i s not re- bination o f better market information, im- flected in the current cashew marketing proved transport and communications, and chain, because most exports are o f raw better enforcement o f contracts. Rural law nuts.But improvedprocessing, packaging, and order is a factor as well.'* Here too, and handling and expanded end uses are knowledge o f the local business environ- important avenues for increasing producer ment andthe constraints on rural trade and returns inthe future. investment are required. District growth strategies rooted in local consultations Producer margins (figure 5.6) are also cru- with farmers, processors, and traders are cially affected by labor costs. Producer Figure 5.6. Cost Componentsof Producer Margins irabica coffee Robusta coffee 1 1 Tobacco 1 Cashews Cotton Kilimanjaro Kagera Tabora Shlnyanga 0Gross 100% margidacre 0 Other physical inputs 75% Pesticides 50% 0 Packaging 25% Other labor 0% w Weeding labor I I I I w Spraying labor -25% Harvesting labor -50% Source: Nyange2005. 102 Agricultural Productivify and Shared Growth margins are sometimes very high, but as New acts passed in2000 and 2001 for cof- with traders there may be nonpecuniary fee and cotton enhanced the regulatory and marketing costs that are hard to observe discretionary powers o f crop boards. The (for example, the search for a buyer). Fur- new acts responded to concerns that tradi- thermore, discussion with the survey or- tional exports had weakened during the fi- ganizer indicated that because the farmer nal years o f the 1990s and reflected a desire typically travels a considerable distance to to take strong measures to set things back sell at market, transport costs between on course. Crop boards were empowered to farm and market are an important unac- inspect, monitor, register, regulate, and li- counted-for component o f the producer's cense varieties, producers, traders, grades, margin. Inthat sense, the main hypothesis standards, processors, storage facilities, and o f this study regarding the importance o f exporters. The tea industrywas not subject transport costs i s valid, but for farm-to- to firther regulation. New regulatory meas- market transport rather than for transport ures for cashews were introduced under the from district markets to the border. Cashew Act o f 1994, its amendments in 1997, andrelatedregulations o f 1996. Sustained 5 percent growth i s feasible but will require renewed and increased atten- The additional regulatory powers were in- tion to the reform agenda and to the public troduced largely to address a perceived need sector's role in supporting traditional ex- to enhance competition, promote fair trade, port crops. Exports are likely to diversify, improve quality, and augment value. Inpur- and the relative importance o f the present suit o f those objectives, crop boards have "big five" may decline, but traditional ex- applied buying rules; introduced a "one- ports could be important future contribu- license rule" to maintain the integrity o f the tors to growth and smallholder income. coffee auction; announced indicative prices The challenge i s to raise productivity in for cashew, tea, and cotton; and required the face o f increasingly demanding inter- that growers be registered. The Cashew nationalmarkets. Board applied resources fiom the develop- ment fund to revive processing factories in What must be done? For coffee, potential an effort to improve access to global mar- exists to meet the increased demand for kets for kernels. The factories subsequently lower-quality robusta as well as specialty closed because they were not viable in the coffees. Getting ahead in international cot- specific circumstances o fthe investments. ton markets requires cotton o f consistently good quality, high yields, and high ginning Vertically integrated firms that were en- ratios. For cashews, improved farm produc- gaged in purchasing, processing, and ex- tivity, product quality, andvalue addition at porting emerged following liberalization. the processor level are central to improving Firms specializing only inpurchasing pri- market share. Furthermore, as local proc- maryproducts have limited scope to man- essing capacity grows, more cashews will age risks and access capital through prefi- be sold into the competitive and remunera- nancing. Yet regulations that require tive global market for kernels, rather than traders to give a two-week notice o f any the more restricted market for unprocessed downward changes in price weaken their nuts, as at present. Improving tea exports ability to manage risks. Vertical integra- requires a continued focus on quality in or- tion can be understood, therefore, as an der to capture the significant price premi- institutional response to conditions in the ums for higher-quality tea. markets. The one-license rule, which was 103 SectoralPerspectiveson Growth introduced to enhance competition and ganizations allows other entities to take on maintain the integrity o f coffee auction, this role. has served, perversely, to reduce competi- tion. Because coffee buyers cannot export, Price setting (in various forms) that seeks the number o f coffee buyers has declined. to inform producers o f prices that they More than 50 percent o f coffee transac- should expect has not helped them under- tions took place with only one buyer inthe stand price movements or make informed market duringthe 2004 season. decisions. Producers seem to have little access to information on the global price Producer organizations have also emerged movements that ultimately determine their and grown since the passage o f the acts. opportunities. A systematic effort to make They include farmer groups, cooperative market intelligence available to producers societies not affiliatedwith the cooperative and to assist producer groups inmarketing unions, cashew groups, and tea associa- would be a constructive step toward inte- tions. Most important among them are grating Tanzanian producers into global groups o f coffee producers, which were markets. initiated by private companies and NGOs to bring together smallholders. These Restrictive regulations, particularly in the groups empower producers, but they need coffee sector, have prevented organiza- technical and organizational support, ei- tions from evolving to meet the needs o f ther from public or private advisers or changing conditions inthe market. For ex- from contractual arrangements with verti- ample, coffee exporters cannot make in- cally integrated firms. Farmers in groups vestments upstream in central pulperies or are better able than individual farmers to buy unprocessed cherries fiom growers to access and understand information about improve quality and cater to niche mar- prices, technology, and risks and to par- kets. Such restrictions also have hampered ticipate directly inmarkets. investments that are necessary to improve quality. In addition, the cost o f entry into Information and institutional arrangements crop buyingi s high, resulting fiom regula- that help farmers process information are tions and the tendency to use buyers to particularly important, given the wide- collect taxes-a cashew buyer must de- spread suspicion o f traders and the history posit cesses, levies, and taxes on 100 tons o f stable panterritorial prices under the before getting a buying permit. This high previous marketing regime. Many produc- cost prevents potential buyers who have ers perceive the normal fluctuations o f limited resources from entering the mar- prices inherent in open market economies ket. Removing restrictive regulations and as attributable to the machinations o f trad- reducing the costs o f entry will encourage ers or others acting to manipulate markets. more agents to enter markets and new Information about the movement o f prices forms o f exchanges to emerge, thus en- over space and time i s particularly impor- hancing competition. The role that pro- tant because producers are still gaining ducer organizations can play incountering experience with regard to expectations of the monopsony power o f traders and the price behavior. The management o f price potential to enhance competition by facili- movements and expectations was among tating their development-rather than re- the original functions assigned to the crop stricting the scope o f activities o f trad- boards, and the growth o f producer or- ers-are still not adequately recognized. 104 Agricultural Productivity and Shared Growth Quality management remains a key issue cashew quality. Inthe case o f cashews, the that institutions seeking to improve mar- concern is largely about producers not be- keting opportunities must address. Al- ing compensated for quality rather than though quality i s specific to the products any decline in quality. And in tea, the and even to individual lots, it also has an quality has improved. element o f public good through its link with reputation. When a country's coffee There has been a decline in grading prac- or cashews are known to be o f higher- tices. For coffee, the extent o f adoption o f than-average quality, international agents improved grading practices i s mixed. Ex- seeking high-quality specialty and niche porters who established large buying op- products are more likely to invest in the erations indicated that it was difficult to country than iftheir quality i s perceived to enforce grading because they bought cof- be low. Investment inquality can thus lead fee through agents. For cashew nuts, to higher unit values and to improved proper grading to assess kernel outturn prospects for growth. Regulations on grad- requires cutting tests, which are rarely ing and quality premiums have been intro- done. Traders seem to rely more on re- duced under the assumption that the indi- gional differences in outturn rates, mixing vidual actions o f firms may damage nuts from different regions to achieve the reputations or that the firms may not act in quality required by importers. For tea, ways that build or maintain reputations. some factories insist that at least 80 per- This assumption is based on the perceived cent o f tea leaves have two leaves and a decline inquality following liberalization. bud. The requirement that tea leaves be d processed within four hours i s apparently The quality o f cotton lint appears to have rarely followed, but the situation has im- begun a long decline before liberalization; proved since privatization. the decline in premiums for cleanliness began in the 1980s. More recent evidence Whether quality problems exist in a sector i s not conclusive. The quality o f cotton depends on the unique aspects o f the crop, exports to Europe appears to have declined the nature o f the commodity market, and marginally, but that decline may be a re- the industry structure. For each o f the sponse to the demand for cotton for a crops, cultivation and processing practices broader range o f purposes. The quality o f have clear and measurable effects on qual- coffee does not appear to have declined. ity, but relative effects may differ, with An examination o f the class o f mild ara- implications for when quality differentia- bica coffees produced in the north zone tion should take place. For coffee and cot- and sold at the Moshi auction for four ton, producers' practices largely determine years before and four years after liberali- quality, with secondary processing having zation suggests that coffee quality interms marginal effects. Any loss o f incentives o f class alone remained roughly constant, for producers would have serious conse- with estates producing higher-quality cof- quences for quality, because damage by fee than other producers. For more recent producers i s irreversible. Secondary proc- years, available information suggests that essing has a greater effect on the quality o f it i s primarily estates that produce coffee tea and cashews (kernels), although the falling into the top five classes and that the source and quality o f green tea leaves i s a proportion o f their coffee that i s o f highest major determinant o f tea quality. Reputa- quality varies considerably from year to tion for cotton may be associated with the year. Fewer concerns exist about tea and country as a whole, whereas for coffee, it 105 Sectoral Perspectives on Growth may be associated with an agroclimatic only intea production, and even there lev- niche; for cashews, with processors; and els o f application may not be optimal. for tea, with individual factories. Finally, Surveys o f coffee and cotton producers excess processing capacity may be an im- indicate that only about 13 percent o f the portant factor in making firms bid up the growers use inorganic fertilizers. The use price o fpoorer grades o fproducts. o f pesticides i s somewhat higher but per- haps not high enough, considering the The current institutional structure in the level o f incidence o f disease in coffee. cotton sector is such that it i s not feasible Among producers o f the four crops, only for one or more firms to offer a premium smallholder tea producers have significant for quality andthereby enhancethe quality access to credit. Cashew producers often o f their exports. The growth inprocessing can obtain credit for fbmigation. Small- capacity may have led to a situation in holders producing cotton and coffee report which the processing o f large quantities o f little or no access to credit. low-quality cotton i s superior to other processing alternatives. There are clear Although access to credit i s recognized as indications that this situation does not ex- an important contributor to successful out- ist inother sectors, although one sees mul- comes, the crop boards have appropriately tiple strategies in operation in coffee. The refrained from getting involved directly in presence o f farmer groups that can directly the provision o f credit. A number of capture quality premiums in the market emerging innovations show promise in also ensures that quality differentiation improving the supply o f inputs. Groups o f does not disappear. The incentive struc- coffee farmers and independent societies tures and strategies o f some firms suggest are exercising strong voluntary control that the private sector markets Tanzanian over membersto produce high-quality cof- coffee, cashews, and tea in ways that en- fee; to market collectively; and, inturn, to hance reputation. gain access to credit markets. Contract farming in cotton allows the ginner to of- Regulating grading practices has not been fer credit and extension services to indi- effective, because it i s difficult to enforce viduals who can offer collateral. Agents and may be counterproductive when done familiar with borrowers are employed to without understanding the level o f quality recover credit (say, for cotton) and, hence, differentiation that markets are likely to serve to inculcate a culture o f discipline in support. A better strategy would be to pilot repayment. Outgrowers producing tea un- market activities that support a greater der contract to processors receive credit as level o f quality differentiation and to re- part o f the contractual relationship. The duce constraints on coordination among coffee vouchers and cotton passbooks are buyers that may prevent them from deal- each prepayment schemes under which a ing effectively with quality issues. Exam- portion o f the growers' sale i s retained for iningways to createmore distinct levels of inputs for the following season; however, quality would help further this process. these schemes are reported to be problem- atic for a numbero freasons. Access to and use o f inputs affects the costs o f production and the quality o f In an environment in which contract en- products. Plant protection is critical for forcement i s difficult, the innovative coffee, cotton, and cashews. Application credit arrangements that have developed o f inorganic nutrients i s widely practiced involve self-selected groups that are ca- 106 Agricultural Productivity and Shared Growfh pable o f exercising control over their boards the authority to control the import members or personal relationships and o f varieties in order to safeguard product the social capital that may exist in com- quality and reduce the chance o f disease or munities.Inmost cases, the private sector pest outbreaks, but the boards and partners has organized producers and developed inthe private sector have not succeededin relationships that are informal but effec- identifying successful strategies for ensur- tive to give them access to information ing the commercial supply o f plant mate- and inputs and to gain greater control rial. Scarce resources that go into exten- over supplies and quality. Vertically inte- sion for these commodities are used grated firms that can better manage risks mostly for regulatory rather than advisory and gain access to resources at reasonable purposes. costs appear to have incentives to enter into contract farming arrangements. They These findings have significant implica- need flexibility to fashion relationships tions for regulations and for the activities and make exchanges inways that may not o f boards. Since these recommendations be feasible under many o f the current were put forward, the government o f Tan- crop acts, as noted earlier. zania has been moving toward greater spe- cifics on the reform timetable. However, The crop acts authorize the establishment to date much o f the government's reform o f industry development and input funds, emphasis has been on financing issues supported by contributions from produc- rather than on clearer delineation o f public ers, to organize supply o f inputs and un- and private sector functions. More specif- dertake development activities deemed ics are needed on redefinition o f board necessary. Cotton and cashew boards have functions; measures to increase account- operationalized them. The funds are man- ability; and reduction in the number, cost, aged by trustees, who are less accountable and scope o f licensing requirements. to stakeholders than are the boards. The These issues need to be clarified and funds are also centralized. Some of the agreed on before amendment o f the programs (such as passbooks) were intro- boards' respective acts. duced recently and are therefore difficult to evaluate, but problems associated with Nontraditional agricultural exports. Past centralized decision making and organiza- agricultural growth was maintainedinpart tion o f supplies are apparent. The pro- because farmers substituted other produc- grams would benefit from greater respon- tion activities for traditional export crops. siveness to the needs o f producers and The export o f nontraditional products, in- from stronger participation by traders and cluding horticultural products, maize, rice, private input suppliers. and beans, has played an important role. The contribution o f these products to Tan- Research remains underfunded and not zanian exports i s undervalued because a fully able to meet the needs o f the sectors, significant portion occurs as informal particularly with regard to developing cross-border trade. Greater attention to the suitable varieties and organizing the sup- development o f these regional exports ply o f planting materials. Industry support would heighten appreciation o f their to privatized coffee and tea research insti- income-raising and stabilizing roles, while tutes i s not adequate to fulfill their man- increasing the benefits that flow to small- dates. Planting material i s supplied at sub- holder farmers, who are actively engaged sidized prices in all cases. The acts give inthis trade not only as producers but fre- 107 Sectoral Perspectives on Growth quently as traders. Fortunately, a base of The challenge for the Tanzanian industry donor support inthese areas has facilitated i s to strengthen its competitive position. expansion of the trade. This support, along The government and donors have taken with strategic infrastructure planning specific actions to facilitate orange pro- and research system support from local duction and marketing, whereas the onion government, will be essential to future trade has emerged more spontaneously. development. But both appear to have been helped by the general regulatory environment and Tanzania's horticultural and floricultural infrastructure planning in northern Tanza- exports consist o f a broad-based but small nia. As the Tanzania Diagnostic Trade trade with Europe, as well as a growing Integration Study (World Bank 2005f, trade with neighboring countries- Vol. 2, 36) noted, "Improved infrastruc- especially Kenya-in onions, tomatoes, ture and fewer regulations in Tanzania potatoes, and oranges. A background study may be one key reason that it i s able to (Sergeant 2004) on horticulture for the re- export more successfully to Kenya than cent Tanzania Diagnostic Trade Integra- Kenya i s to Tanzania. Unlike Kenya, Tan- tion Study (World Bank 20059 calculates zania has continued to tarmac roads to its that in 2004/05 officially recorded exports border posts. Tanzanian authorities have to Europe totaled US$24.4 million. The made issuance o f trade permits administra- exported items included cut roses, flower tively easy and cheap ... whereas in cuttings, bean seeds, and fresh fruits and Kenya, procedures for issuing permits and vegetables. Unofficial exports to Kenya o f import brokerage are complex and only oranges, onions, and tomatoes are esti- known traders can clear goods" Box 5.1 mated at US$37million over the same indicates that fbture efficiencies are possi- period. ble, but they depend on continued ef- forts-some o f which would benefit from Horticulture can reduce poverty through government attention, including infrastruc- existing links with small farmers or poten- ture and support for farmer marketingini- tial links through the expansion o f out- tiatives, as well as reducing constraints on grower schemes. Seasonal employment in output and input access. larger production facilities oriented to the European market i s also significant. It i s Concerning exports bound for Europe, estimated that more than 7,000 workers, demand prospects are good, but Tanza- mainly women, work seasonally in vege- nia's future competitive position depends table packing operations, floriculture on continued reform o f the regulatory en- greenhouses, and bean seed preparation vironment and upgrading o f infrastructure, units. The informal trade with Kenya has along with continued diversification into even more direct links to small, resource- production o f seeds and cutting materials poor farmers, many o f whom are actively for export. l3Interestingly, market partici- engagedintrading (box 5.1). pants do not list access to finance as a sig- nificant constraint. Specific steps to im- Future potential for export growth varies prove competitiveness include improving by commodity. The regional horticultural customs efficiency (that is, more speedy trade with Kenya i s based on a true com- recovery o f value added tax payments for parative advantage; however, it could be exporters and payment o f duty drawback); threatened if Kenyan orange and onion increasing access to air freight capacity production shifted to more suitable areas. from Kilimanjaro International Airport; 108 Agricultural Productivity and Shared Growth Orange Marketingand Competitiveness Most o f the oranges imported into Kenya originate from Tanga and Morogoro. Tanzania's com- petitiveness inKenyanmarkets i s due inpart to the absence o f greening disease, a problemthat affects Kenya's higher-elevation growers. Inaddition, Kenyanproducers face tighter landcon- straints, and orange production competes with more profitable export crops. External Support to Farmers to Facilitate Competitiveness Government provisiono fnew plantingmaterial inthe Muheza district, inthe Tanga region, helpedestablish significant orange production inthe 1970s.Morerecently, the Development Al- ternative Inc. Private Enterprise Support Activities project has helped farmers form marketing associations that streamline the marketing chain, with farmer associations now selling straight to Nairobi-basedtraders. Over time, farm-gate prices have almost doubled as farmers' marketing options have increased, and the bargainingpower o f local traders has declined. The project con- tinues to document and assess the effect oftaxes and fees leviedduringtransport to the border, I The which appear to be a continuing constraint on the trade. MarketingChain for Onions Tanzanian exports now account for a significant portion o f the Kenyan onion market. Tanza- nia exports throughout the year, mainly from Mang'ola inthe Arusha region. The marketing chain includes rural brokers who, for a fee, introduce wholesale traders to farmers who have onions for sale. The trader buys the crop from the farmer, packs it, and hires transport to the Arusha market, where the produce i s unloaded and sold by the wholesaler. Some efficiencies could perhaps be achieved by having the Nairobi trader work directly with the Tanzanian trader who buys from the farmer, rather than through the Arusha market. This approach would save handling costs, market fees, and the wholesaler's margin and would allow a greater portion o f the sales price to be returnedto the farmer. Reducing the number o f times that onions are handled would also improve quality further. Duringinterviews, traders stated that there was a law or rule that all produce destined for Nairobi had to be traded through the Arusha market. This information needs verification. The Mang'ola onion producers have not yet received support from a donor-fundedproject, but their competitiveness is based on both their proximity to Nairobi and their higher yields, which are probably due to lower incidence o f fungal diseases inTanzania's drier production areas. Tanzanian producers also have a quality edge in the Nairobi market: farmers grade the onions, and the better ones are exported. Farmers and traders note that their competitive posi- tion could be improved further by improvements inonion seed quality and an increase in the range o f seed varieties. Source: Sergeant2004. improving the supply o f skilled middle Kenya cannot be importedbecause they are managers and supervisors; and establishing not approved in Tanzania, thus putting an open regulatory environment for seeds, Tanzania at a disadvantage because it can- equipment, and other inputs. Some agro- not always use the safest, most effective, or chemical products used and approved in cheapest agrochemical inputs. 109 Sectoral Perspectiveson Growth Horticulture i s not the only nontraditional infrastructure and information systems, export that requires increased attention. along with efforts to improve the transmis- Over the past decade, cross-border trade in sion o f prices and information back to Tan- staple foods between the countries o f east- zania's smallholding maize producers. ern and southern Africa, with Tanzania as a crucial balance wheel, has become an A particular problem is the virtual absence increasingly important base o f food secu- o f information andanalysis on grain storage rity for regional consumers and source of systems and behavior. Aiiica's seasonal income for smallholding producers. The grain price instability is the highest in the Regional Agricultural Trade Intelligence world, and Tanzania's price instability is Network, a donor-funded project that also considerable, often moving from export monitors informal cross-border flows, re- to import parity levels within a season. This ports that in 2004 Kenya imported 88,000 instability is undoubtedly a factor behind metric tons o f maize and 18,000 metric recent efforts to expand the ambit o f the tons o f rice from Tanzania, while import- strategic grain reserve, but a coherent pro- ing 75,000 metric tons o f maize and gramfor reducing the source ofinstability is 53,000 metric tons o f beans from Uganda. lacking. Doing so depends on establishing a The Democratic Republic o f Congo is an- better information base about the entities other important destination for Tanzania's that store grain, their decision-making exports. The Central Bank o f Uganda now framework, the accuracy o f the information considers its own unofficial cross-border on which they make decisions, and so on. exports sufficiently important that it has Assessment o f grain marketing performance established a bimonthly monitoring system inaregional context andthe developmentof at dozens o fborder crossing points. Recent a coherent policy framework based on this discussions within the newly formed East assessment are essential. Africa Community on common grain grad- ing standards also attest to the trade's cur- Public Expendituresto rentimportance to the three members. Support Agricultural Growth Critical factors behind the East Afiica Competent management o f public expen- flows are the trade and marketingreforms diture to support agricultural growth is o f the late 1980s and 1990s and the move- critical for achieving Tanzania's strategic ment of Kenya into a position of long-term economic objectives. Most agricultural structural deficit inmaize and other grains. activity i s the business o f the private sec- The entry o f the major surplus producer, tor. Nonetheless, private agents-from the South Aiiica, into Kenyanmarkets i s likely most modest smallholders to medium- and if and when Kenya reduces its 35 percent large-scale commercial producers, to trad- import tariff on maize. Reform o f Kenya's ers and processors-rely on publicly sup- National Cereals and Produce Board would ported goods and services not adequately also have important repercussions for re- supplied by the private sector. Decisions gional production and storage. This policy on public spending for recurrent costs and instability, along with the market instability investment to support agricultural growth, associated with poor infrastructure, con- along with definition o f national and local tracting mechanisms, and information sys- institutions' roles and priority activities, tems, makes for a volatile market. The are the most important issues the govern- long-term viability o f this market will de- ment must tackle inpursuit o f its strategic pendonfocused efforts to improve regional objectives. 110 Agriculfural Productivityand Shared Growth Balance, Efficiency, and Timing of and rural growth as long as investments Agriculture-Related Expenditures that enhance competitiveness accompany investments inhealth and education. Although the management o f public ex- penditure i s important for all developing Efficiency o f expenditure requires that the countries, those like Tanzania face a spe- public sector confine itself to activities that cial macroeconomic challenge because o f are genuinely public in nature and that it the large flows o f assistance from devel- allow and encourage the private sector to opment partners. Large inflows can help in expand into the private sector's appropriate meeting key objectives for reducing pov- domain. Public goods and services are erty, but even with careful management, those that will be undersupplied by the pri- the flows will have real exchange rate ef- vate sector even under circumstances in fects that encourage the country to import which the private sector i s mature, well de- more and export less. These effects will be veloped, and operating with a full array o f more pronounced as more assistance is markets. Classically defined public goods channeled to nontradable services and ac- and services are those with environmental tivities with high domestic content and externalities or poorly defined property low import content. Increased spending on rights (air quality and biodiversity) and labor-intensive activities in the social sec- those for which nonpaying beneficiaries tors tends to have a relatively large effect cannot be excluded at reasonable costs on the real exchange rate in the short run, (much agricultural research, primaryeduca- butthey are also the activities that over the tion, basic health care and prevention o f longer runwill enhance growth and ensure disease, and roads). The array o f public sustainability. Investments in education goods and services that governments must (particularly for girls), in primary health provide to support growth in developing care, in control o f malaria and HIV/AIDS, countries i s so large and costly-and the and in clean water are clearly important resources for providing them are so con- for growth and for poverty reduction. But strained-that efficiency i s a fairly obvious difficult trade-offs must be faced when the requirement. Yet it i s often not achieved in financing for these activities brings practice. Finally, the timing o f expenditures changes in the real exchange rate that i s closely linked to balance. Investments in lower prices for the grain, coffee, cotton, the social sectors and competitiveness must meat, skins, and other products on which be timed to complement each other. When the incomes o f the ruralpoor depend. they do, a healthier and better-educated la- bor force has new economic opportunities, Management o f the trade-offs requires at- and more prosperous rural households are tention to three dimensions o f the decision better able to undertake measures on their process on public expenditure: balance, ef- own to sustain the health and creativity o f ficiency, and timing. Expenditures must be their members. sufficiently balanced so that as the real ex- change rate shifts, competitiveness in the Managing public expenditure to achieve tradable sectors increases through reduced balance, efficiency, and timeliness i s chal- costs o f transportation, more reliable and lenging, and no technical algorithm exists efficient provision o f power and water, to guide the process. Even if the technical generation and dissemination o f improved knowledge existed, it is not obvious that it technologies, and better public services. would be applied, because decisions on Economies can improve social indicators public expenditure are profoundly politi- 111 Secforal Perspectives on Growth cal. The budget i s allocated innegotiations production on state farms and marketing), between ministries, implemented by the but it also included investments in agricul- civil service, monitored through the public tural researchand extension, expenditure review process, and evaluated ultimately by voters through relations with During the 1990s, the agricultural sector their parliamentary representatives. De- adjusted well to the reforms and shift in velopment partners take an active interest expenditure. Until 1997, global prices for in providing information to underpin the the key export commodities were quite process, in helping clarify options and strong, and Tanzania's liberalization pro- trade-offs, and inmonitoring outcomes, vided producers with greater access to world markets at a propitious time. As re- Inthe impreciseprocess o fallocating pub- forms in marketing proceeded, producers lic expenditure, the lessons o f the past can received a high share o f border prices, and be very informative. In the period o f the the liberalizationimprovedincentives. 1980s leading up to the reforms o f the 1990s, public expenditures for agriculture At the same time that sectoral reforms as- were quite substantial, but they failed the sisted producers, movements in the real test o f efficiency, and much of the money exchange rate associated with the shift in was spent on direct production on state patterns o f public expenditure acted farms, marketing through parastatal against them. The increase in the real boards, and input subsidies. Research re- value o f the shilling o f about 40 percent ceived support, but scientists or public between 1996 and 2001 compounded a servants determined the priorities for re- weakening o f global prices for Tanzania's search without much input from produc- chief agricultural exports over the same ers. Agriculture's share of the economy at period. The increased efficiency in the the time was even greater than at present, marketing system associated with the lib- and other sectors could not balance the eralization and the rising share o f border slow growth in agriculture. The costly and prices passed back to producers insulated inefficient public expenditure for agricul- rural households somewhat from what ture could not be sustained at the observed would otherwise have been a severe shock. rates o f sectoral growth, and the poor per- Sectoral reforms largely offset the adverse formance o f agriculture was a key deter- impacts o f movements in the exchange minant in the decision to undertake hn- rate and global prices. damental reforms inthe 1990s. Between 2001 and 2003, the real value of The reforms of the 1990s attracted the sup- the shilling adjusted back toward the level port o f development partners at a time it held before the appreciation o f the mid- when priorities in the development com- 1990s. Realignment of the exchange rate munitywere shiftingtoward investments in improves competitiveness and assists agri- the social sectors. Because o f a shift inpri- cultural producers and exporters. The cur- orities and because expenditure on agricul- rent attention to the performance o f the ture under the previous economic order had crop boards and proposed changes in the yielded disappointing results, the budgetary regulatory environment for coffee, cotton, share of agriculture declined markedly. The and cashews will, if implemented, repli- drop inspending included areas that should cate some o f the benefits o f the reforms o f have been cut or eliminated because the the 1990s. But exchange rate pressures can private sector could do them better (such as be expected to remain, and regulatory re- 112 Agricultural Productivity and Shared Growth forms deliver mostly one-time benefits, technical services, and training); (c) physi- albeit over a period o f time. Continuous cal investment; (d) private sector devel- public investment to enhance competitive- opment, market development, and agricul- nesswill be needed to position the agricul- tural finance; and (e) cross-cutting issues, tural sector for dynamic adjustment in a such as gender rights. growing economy with a strong currency. The ASDP framework and process docu- Agricultural Policy Framework ment estimated five-year projections of and Public Expenditure public expenditures needed for implemen- tation. A 20 percent annual increase inde- The government, with the support o f de- velopment expenditures was projected velopment partners, developed the Agri- over the five-year period, with an empha- cultural Sector Development Strategy sis on research, extension, the policy and (ASDS) and adopted it in2003 to enhance regulatory environment, and marketing sectoral growth and competitiveness. The and institutional support. This emphasis i s objective o f the ASDS i s to achieve a sus- consistent with that o f the ASDP on com- tained agricultural growth rate o f 5 percent mercializing agriculture and on decentral- per year, primarily through transformation izing support through districts (through from subsistence to commercial agricul- DADPs) (table 5.7). The specific propos- ture. The private sector i s to lead the trans- als imply that research and extension, formation. The government has committed marketing and finance, policy and regula- to improve the enabling environment tory work, and institutional support be- through policy and institutional reforms come a larger share o f the agriculture and appropriate public expenditure. De- budget over time. In addition, 75 percent velopment partners have been asked to o f public resources for the sector are en- harmonize their assistance with agricul- visaged to be channeled through districts. tural growth and to provide it in support o f implementation o f the ASDS. Core fea- According to the projected financing for the tures o f the strategy are to strengthen the ASDP, the budget for 2005/06 should be policy and regulatory framework at the allocating T Sh72 billion to ASDP priority national and local levels, to provide effi- areas. That amount is expected to rise to cient and relevant public services, to sup- T Sh 104.5 billion by 2007/08. Becausethe port public-private partnerships across all categories inwhich the budget lines are re- levels o f the sector, and to support formu- ported do not correspond directly to those lation and implementation o f DADPs as o f the ASDP, it i s difficult to assess the de- the comprehensive tool for agricultural gree to which the 2005/06 budget shown in development at the district level. table 5.8 accords with the planningunder- lying the ASDP. The amounts allocated to The ASDP framework and process docu- the MAFS, MCM, and MWLD for house- ment is an operational response to the holds without ruralwater supply in2005/06 ASDS. The ASDP identifies five key op- appear to be approximately T Sh 115 billion. erational components as a focus for im- If the agricultural allocation under the plementation at both the national level, President's Office-Regional Administra- including zones, and the district level tion and Local Government (PORALG) i s through DADPs: (a) policy, regulatory, included, the total increases to T Sh 121bil- and institutional arrangements; (b) agricul- lion. The government appears able to spend tural services (research, advisory and fully enough to finance the ASDP, because 113 Sectoral Perspectives on Growth Table 5.7. Proposed Growth in Development Expendituresfor ASDP, 2002103 Amount of expenditure Currentshare of Annual increase Area of intervention (T Sh million) total cost (%) (%) Technicalarea Research 6,527 15 20 Extensionand advisory ser- 9,404 22 20 vices Livestockproduction 792 2 10 Crop production 5,164 12 10 Irrigation 6,919 16 10 Marketingand finance 7,996 19 30 Policy and regulatory work 1,805 4 40 Foodsecurity 488 1 10 Institutionalsupport 3,749 30 Training institutions 97 -9 10 Total 42,940 100 20 Amount of ofTarget(%) byyear 5 expenditure Currentshare of implementation Area of intervention (T Sh million) total cost (%) National vs. district National level 19,900 61 25 District level 31,600 39 75 Recurrent vs. development Total recurrent 41,300 45 33 Total development 51,500 55 67 Source: United Republicof Tanzania 2002b. Note: -=not available. current expenditures are already more than be elucidated in a detailed public expendi- the amounts required to support the core ture review focusing on agriculture. With activities. the government's commitment to the Present Expenditure Patterns ASDP, the correspondence between the budget categories for agriculture and the The share o f spending o n a g r i ~ u l t u r e 'in ~ program categories within ASDP should total government expenditure has fluctu- be clear and straightforward, so that the ated widely over the past decade, between degree o f consistency is readily apparent. 1.5 percent in 1995/96 and 5.5 percent in At present, this is not the case. Allocations 2000/01 and reaching 3.7 percent in to forestry and fisheries fall outside the 2005/06. Nine ministries hold votes on ASDP as defined, although inside the categories of expenditure directly relevant broad category of agriculture as it appears to agricultural performance. A s shown in ingovernmental aggregates. Discrepancies table 5.8, the MAFS holds the largest in categories for this reason are readily budget, at just over US$93 million. understandable. But even within agricul- ture, narrowly defined, the alignment o f The spending displayed in table 5.9 raises the budget with the ASDP priorities is dif- a number of questions, some o f which will ficult to discern. If the government is 114 Agriculfural Productivity and Shared Growth Table 5.8. 2005106 Budget Proposals Share T Sh US$ in total Ministry Votes (million) (million) (%) Ministry of Agriculture and Food Security 43 102,414 93.1 40 Ministry of Water and Livestock Development 49 45,961 41.8 18 Ministry of Works 47 44,010 40.0 17 Ministry of Natural Resources and Tourism 69 38,059 34.6 15 Prime Minister's Office 37 8,335 7.6 3 Ministry of Cooperatives and Marketing 24 5,974 5.4 2 President`s Office-Regional Administration and Local Government 56 6,321 5.7 2 Ministry of Energy and Mining 58 89 0.1 0 Ministry of Lands, Housing, and Settlement Development 48 3,284 3.0 1 Source: United Republic of Tanzania 2005a. allocating significant resources to agricul- T Sh2 billion in 2003/04 to T Sh7 billion tural expenditures that are not part o f the in2004/05 and T Sh 14billion in2005/06. ASDP, what does this spending imply The Medium-Term Expenditure Frame- about the degree o f commitment to the work guidelines suggest that, over the me- program and the extent to which develop- dium term, the government will focus on ment partners should support it? minimizing the costs o f production through subsidies on the price o f inputs. Thus, the The budgetary lines for the MAFS under fertilizer subsidies should be seen as part o f crop development include support to re- the expenditure in pursuit o f competitive- search and extension accounting for about ness; however, the record o f experience 16 percent o f expenditure o f the MAFS. globally shows that they do not pass the The allocation represents an increase rela- tests for efficiency o f expenditure. More- tive to historic levels and i s consistent with over, resorting to input subsidies to in- the commitment to support for generation crease competitiveness is not consistent and dissemination o f technology under the with the strategic directions underlying the ASDP. But even with this substantial in- ASDS. crease in spending on research and exten- sion, Tanzania's investment remains low The profitability o f fertilizer use varies by international standards. Space must be widely. Insome areas, the subsidies can be made for rising allocations to research and phased out only at great hardship to the extension inthe future. This consideration, former beneficiaries. Subsidies are diffi- in turn, suggests that less critical or pro- cult to target, and benefits are modest even ductive expenditures should include sunset for the recipients. Often, the least needy clauses that will allow them to be phased farmers capture the benefits. Administra- out over time. tive costs (including rent seeking) often outweigh the benefits. Subsidized fertilizer Among the items that should be phased out has often arrived too late to be effective i s the allocation to input subsidies. Most o f and has often been o f the wrong type. In this allocation goes to fertilizer subsidies, 2003/04, the fertilizer subsidy to farmers and the amount has increased from was relatively small, ranging from 115 Sectoral Perspectiveson Growth Table 5.9. Budget Ceilingsfor Rural Developmentand Agriculture in Line Ministries,2005/06 T Sh Ministrytotal (million) US$ (million) (%) Ministry of Agriculture andFoodSecurity Total budget 102,414 93.1 488 Increasedcrop productionthrough improved technology 14,135 12.9 67 Inputsubsidies on selectedfood crops 7 0.0 0 Increasedcrop productionthrough improvedtechnology 621 0.6 3 Strategic grain reserves 543 0.5 3 Irrigation 9,712 8.8 46 Harmonizedagriculturaltaxes to support ex- port crop efficiency 0.0 0.0 0 Ministry of Water and Livestock Development Total budgetfor ruralwater and livestock activities 45,961 41.8 100 Ruralwater supply 38,530 35.0 84 Animal production 2,639 2.4 6 Veterinary services 2,465 2.2 5 Livestock researchand training institutes 2,327 2.1 5 Ministry of Cooperatives andMarketing Total budget 5,974 5.4 197 Supportto cooperatives 2,001 1.8 66 Community credit (savings and credit coop- eratives, revolvingfunds) 545 0.5 18 Marketingdevelopment 1,333 1.2 44 Identificationof markets, promotionof value- adding products(Policy and Planning, 30%; Marketing Development, 70%) 1,260 1.1 42 Ministry of NaturalResourcesand Tourism Total budgetfor forestry and fisheries activities 38,059 34.6 100 Forestry and beekeepingactivities 20,985 19.1 55 Fisheriesactivities 15,240 13.9 40 Land management in wildlife reserves 1100 1.o 3 Agricultural researchand extension 634 0.6 2 Rural small and medium-sizeenterprises 100 0.1 0.3 Ministry of Works Total for rural roads activities 44,010 40.0 100 Rural roads construction, maintenance,and community roads 21,979 20.0 50 Ministry of Energy andMining Total for rural energy activities 89 0.08 100 Ruralenergy 89 0.08 100 Ministry of Lands, Housing, and Settlement Development Total for land managementactivities* 3,284 3.0 100 District land tribunals-34 functioning by July 2007 746 0.68 23 continued 116 Agricultural Productivity and Shared Growth Table5.9 continued T Sh Ministrvtotal (million) US$ (million) (%) Prime Minister`s Office Total for rural food security, finance, and dis- aster planning 8,335 7.6 100 Ruralfinancial services 3,030 2.75 36 Coordination of Agricultural Marketing Ser- vices Development Program 4,638 4.22 56 President's Office-Regional Administrafion and Local Government Total budget 126,412 114.9 100 Budget to strengthen, extend, and monitor resource allocation formula ensuring equity among local authorities 91,901 83.5 73 Source: United Republic of Tanzania 2005a. T Sh 950 per bag for fertilizer destined for December 2003 and March 2004, the date Makambako depots to T Sh 2000 per bag o f refund). for consignments going to Songea depots. The reduction in the retail price o f fertil- Subsidies crowd out the private sector. izer was only 8 to 10 percent. In addition, Even when programs are implemented fertilizer prices per bag ranged from through the private sector, they tend to T Sh 13,000 to T Sh 14,000, depending on send the signal that the government's be- the type o f fertilizer, an amount that was havior may be unpredictable and, hence, not significantly different from the increase the risk associated with a change 2002/03 prices. Delays in delivery o f the in policies. Under those circumstances, subsidized fertilizer caused panic among private input suppliers who could enter the farmers. As demand grew, rent-seeking market are reluctant to do so. Intervention behavior emerged, and in some parts of affects the expectations o f farmers. Pro- the southern highlands, fertilizer prices grams have continued to perpetuate de- reached T Sh 18,000 per bag. pendency and result in fbture demands for assistance. Assumingthat intervention will Subsidies are expensive to administer. be forthcoming, farmers underinvest in Reimbursement o f transport costs by the strategies to cope with risk. government i s reported to be time con- suming. Transporters were required in Subsidized fertilizer displaces more sus- 2003/04 to produce four sets o f copies o f tainable and profitable land-use practices. notary-certified cash sale receipts before Fallowing and the use o f organic matter submitting reimbursement claims to the become less attractive when free or subsi- MAFS. Importers were concerned about dized fertilizer is available. Other tech- the frequent changes inprocedures, which nologies, such as minimum tillage and caused more than three months o f delay conservation farming or low-input agro- in reimbursement. The slow reimburse- forestry, are in some cases superior ments led to exchange rate losses for im- alternatives. porters (the Tanzanian shilling depreci- ated from T Sh 1,050 to T Sh 1,120 Fertilizer subsidies played a significant against the U.S. dollar at import between role during the early years o f the Green 117 SectoralPerspectiveson Growfh Revolution inIndia, although they also left creased investment in water management a legacy o f costly and inefficient public would reduce the need for subsidized inputs intervention that has been politically diffi- byraisingreturns to their application. cult to divest. But India's subsidized fertil- izer was applied to fields well watered by The management o f the strategic grain re- subsidized irrigation. Africa has lower serve will be evaluated underthe enhanced levels o f irrigation and correspondingly public expenditure review that i s under higher production risk associated with in- way. The costs have risen over recent put use, lower levels of rural literacy, gen- years, for reasons that are not well under- erally weaker research and extension sys- stood, at the same time that Tanzania has tems, lower population densities, less experienced a reduction inthe volatility o f infrastructure, and fewer well-developed food production and regional grain mar- institutions for credit and contract en- kets have become more active. The costs forcement than India had at the time. Sub- to hold and rotate a modest stock should sidies in India were preceded by public be low, and significant savings may be investment in infrastructure (roads and ir- available through different approaches to rigation) and inresearchand extension. the strategic grain reserve. The steeply rising expenditures on fertil- Spending on the livestock sector is small izer subsidies in Tanzania could be more inabsolute numbers, relative to the legiti- productively spent on irrigation, research, mate public agenda for livestock, and rela- extension, roads, and rural infrastructure. tive to the potential contribution o f the Rather than spending to subsidize one in- subsector. Tanzania has the second-largest put, the government would be well advised herd o f cattle in Africa, and substantial to spend to increase productivity and eam- numbers o f sheep and goats. Ready avail- ings so that farmers can afford to purchase ability o f poultry feed and a suitable cli- an appropriate array o f inputs, including mate can support the growth o f poultry fertilizer, plant protection agents, seeds products for local markets and for export. and seedlings, tools, draught power, breed- Demand for meat i s rising globally, par- ing stock, improved feeds, and veterinary ticularly in Asian and Middle Eastern medicines. Alternative approaches to in- markets accessible to Tanzania. Opportu- creasing productivity and profitability are nities in global poultry markets are likely consistent with the ASDP, and fertilizer to increase as the intensive production re- subsidies are not. gimes in Asia are forced to adjust to con- cerns about avian flu. Spending on irrigation is very small. The MWLD manages a substantial budget and Legitimate public expenditures that would allocates significant amounts to rural water support growth in the livestock sector in- supply-but for human use, not for produc- clude research on breeds and improved tion. The comprehensive water strategy un- feeds; pasture and rangeland management; der preparationmay provide some guidance training o f veterinarians, paraveterinarians, on how much spending on imgation would and farmers to improve animal health and be appropriate, butcurrent levels are clearly quality; investment in reliable power and too low. The disproportion i s all the more water supply to attract private meat proc- evident in that underallocation to imgation essors and packers; disease control; and falls within the same budgetvote as overal- assistance in meeting sanitary and phyto- location to fertilizer subsidies-when in- sanitary standards. 118 Agricultural Productivity and Shared Growth The adequacy o f spending on rural roads with the commitments underlying the and energy should be judged against the ASDP i s not readily apparent in the cost o f meetingstandards o f service over a budget. five-year horizon. For example, the 5 per- cent growth target for agriculture implies The alignment o f expenditures around core that targets must be set for the density o f and complementary categories to support all-weather roads, so that farmers can the ASDP would contribute to efficiency. reach markets at a lower cost. The ade- The core elements would be those cur- quacy o f the present allocation should be rently included under the MAFS mandate, judged against the annualized costing to as well as livestock under the MWLD and meet the target. elements under the MCM. Activities sup- portedunder these mandates currently take Forestry i s well funded, as i s appropriate place at the national and local levels, and for a country with as rich an endowment spending at the local level will increas- as Tanzania has. But the allocation to and ingly be channeled through the PORALG. effectiveness o f the use o f public money As local governments increase in capacity for forestry should be assessed inconjunc- and as the DADG becomes operational, tion with the targets for establishing the the major distinction in managing public mandated Forest Service and its perfom- support for agricultural development will ance, as well as the increased constructive be not along ministerial lines, but rather role o f the private sector that i s envisaged. according to subsidiarity. That is, the na- In forestry, public and private contribu- tional-level institutions will undertake ac- tions can combine for effective manage- tivities with national implications, and dis- ment, and Tanzania's experience can be tricts and local institutions will take enhancedonboth counts. responsibility for the local activities. By 2007/08 three-fourths of expenditures to An assessment of the allocation o f public support the ASDP are expected to flow spending on agriculture for 2005/06 ac- through district governments. cording to criteria o f balance, efficiency, and timeliness thus suggests several con- As participants inthe political process be- clusions that can be checked further inthe come more familiar with the ASDP and pendingpublic expenditure review. Spend- the implications for the organization o f ing on agriculture is increasing, and this public support, a reconfiguration o f re- trend is a positive indication o f the will- sponsibility at the national level may be ingness to rebalance the budget in favor o f appropriate and feasible. Such a reconfigu- competitiveness. At the same time, the ef- ration would imply mergers o f the depart- ficiency o f expenditure inthe current allo- ments and entities that are at present re- cation, relative to the goals for growth and sponsible for elements o f the ASDP at the competitiveness, appears low. Too little i s national level into one ministry. That min- spent on research, extension, irrigation, istry would coordinate closely with the livestock services, roads, land administra- ministries responsible for complementary tion, and energy. Too much is spent on activities (roads, energy, and water), par- input subsidies and probably on the strate- ticularly at the planningstage. gic grain reserve. Crop development war- rants careful evaluation, because this cate- Such a reconfiguration o f responsibility at gory o f spending i s quite large and the national level and improved coordina- opaque. The consistency o f expenditures tion with both local and complementary 119 Sectoral Perspectiveson Growfh counterparts would facilitate adequate o f Standards (TBS) quality inspection funding o f exciting elements o f the ASDP at 0.25 percent o f the cost, insurance, agenda, such as the following: andfreight price. Any delay bythe TBS means port charges accumulate, thereby Share costs of adopting improved tech- increasing the cost o f fertilizers to nologies, rather than subsidizing a sin- farmers. In addition, importers are re- gle input. Tanzania is already using a quired to produce certificates o f quality public-private cost-sharing mechanism from manufacturers. Abolishing inspec- (through a matching grant) for tion requirements and retaining the cer- adopting new technology under the tificate o f quality from manufacturers, Participatory Agricultural Develop- together with spot-checks by the TBS, ment and EmpowermentProject within would significantly reduce the cost o f the ASDP. The selected technology i s fertilizers to farmers. one that i s profitable in the longer run, and the matching grant makes fertil- Improve the road network. As much as izer, seeds, seedlings, plant protection 40 percent o f the cost o f grains in the agents, implements, and any other major urban markets i s attributed to needed inputs affordable during the pe- transportation costs. Reducingthis cost riod o f adoption (in most cases, two would increase farm profitability. years). Inputs are purchased fiom the 0 Develop and disseminate more profit- private sector, and fanners save in- able technologies. This effort includes creased earnings while receiving the ensuring that agricultural research and matching grant so that they can con- extension focuses on developing more tinue to purchase inputs on their own. profitable technologies and varieties This approach can be scaled up more responsive to the application o f through the DADPs. inputs. It also means ensuring that Continued reform of taxes. The gov- farmers have access to information and ernment has already undertaken a num- recommendations specific to the tech- ber o f commendable reforms to im- nologies (rather than blanket prescrip- prove the tax regime for agriculture, tions) and that profitability i s part o f including setting limits on local taxes the calculus in deciding which tech- for traded agricultural commodities and nologies are recommended. reducing customs fees and value added 0 Ensure consistency in the policy envi- tax. Enforcement o f those measures ronment. The costs o f inputsdecline as should be pursued and their effects the volume o f transactions increases. monitored, because they can improve The number o f private input dealers profitability if well implemented. The and distributors i s more likely to ex- value added tax on port charges and pand when the policy environment is transport costs adds to the cost o f fertil- consistent and the government stays izer and couldbe reevaluated. away from direct intervention. Reduce cumbersome importation pro- Improve the opportunities for viable cedures for fertilizer. Such procedures systems to finance technology and in- include double inspection o f consign- puts. Vertically integrated private sys- ments, preshipment inspection by terns o f extension, credit, and inputs Cotecna at 1.2 percent o f the fkee-on- linked to output markets (as for to- board price, and the Tanzania Bureau bacco) are functioning in some areas. 120 Agricultural Productivity and Shared Growth They can be encouraged by reducing structive role here, eventhough the mar- inefficiencies in the banking systems ket transactions are betweenprivate par- (reducing interest on borrowing), ties. For example, through an agreement strengthening institutions for contract with TechnoServe, the U.S. Agency for enforcement, improving taxation (as InternationalDevelopment (USAID) has noted above), and improving the flow helped coffee growers respond to Octo- o f information. For example, small ber 2003 changes in the policy regime, farmers inMsowero and Sonjo villages whereby Tanzanian premium coffee were assisted under the PASS (Private producers and specialty roasters can ex- Agricultural Sector Support) pro- port high-quality coffee directly. This gram-supported by DANIDA (the change inthe marketingregulations (the Danish International Development contribution o f the public sector) creates Agency)-in forming groups and ac- new opportunities for Tanzanianfarmers cessing loans that were repayable over to supply roasters with specialty coffee a three-year period. Initial reports on for premiumprices. TechnoServe's con- the program are positive. servative estimate is that farmers' in- Reduce risks. Periodic weather shocks comes will be boostedbyUS$20million and external price shifts, together with over the next 10 years as a result o f the household-level production risk, signifi- combined effect o f direct export o f pre- cantly affect technology adoption, prof- miumcoffee and a new value added tax itability, and incomes. New methods to reclamationpolicy. manage external shocks, such as weather insurance, forward contracts, and op- Concerted pursuit o f those steps and others tions for price risk, are being pilotedand through adequate funding o f activities un- should be evaluated carehlly to deter- der the ASDP will provide a reasonable mine the scope for scalingthem up. likelihood that the growth targets can be met. Five percent annual growth is ambi- Improve market access and product tious but not impossible. Meeting this tar- quality. Investment in, awarenessof, and get, however, requires moving ahead deci- compliance with sanitary andphytosani- sively on the agendas o f policy r e f o m and tary standards for high-value products alignment o f public expenditure. Half- and stronger linkage o f groups o f small- measures and compromises to accommo- holders with supermarket chains will date special interests require sober reas- improve marketability and profitability. sessment o f the growth targets or admission The public sector can play a very con- fkom the outset that they will not be met. Notes here is the average o f year-to-year changes during the period inquestion. The sectoral aggregate measured in the World Primary agriculture, as defined in the national Bank Development Data Platform dataset used for accounts, includes crops, livestock, forestry, and this calculation is agricultural value added (gross fisheries. value of agricultural production minus purchased The full effects o f farm growth on the economy as intermediate inputs). The Tanzanian agricultural a whole are capture in intersectoral growth multi- value added measure includes fisheries and forestry plier models, such as the one implemented by Ma- or hunting, in addition to crop and livestock pro- hamba and Levin (2005) for thls Country Economic duction. The average rate o f increase measured Memorandum. Mahamba and Levin's analysis indi- 121 Sectoral Perspectives on Growth cates substantial fam growth multipliers that exceed grants from the Program for Cooperation with those fi-om nonfarmgrowth. Emerging Markets to at least four o f the compa- This calculation is based on an adding-up re- nies. Enza Zaden and Q-SEM Ltd. have been es- quirement: sector growth must equal the sum o f tablished to produce hybridseed. These companies labor force growth and labor productivity growth. have invested on the basis o f cheaper labor (com- The 2.7 percent labor productivity growth require- pared with Europe) and the climate, which allows ment i s simply the difference between the target all-year production with no heating costs. As with sector growth o f 5.0 percent per year and the labor the cuttings industry, a viable cut flower sector force growth o f 2.3 percent per year. gave these investors confidence that sufficient Uncertainties in data for Malawi came to light skills and services were available to establish a during the food crisis o f 2001 and 2002 and may be higher-value floricultural industry. Freight is not an affecting the historic averages. issue, because the output from these companies i s only a few kilograms o fhigh-value seed that canbe Total factor productivity, the productivity associ- exportedby courier. ated with improved use o f all factors (land, labor, capital) together but not attributable to any one l4That is, spending on agriculture, forestry, fish- 'factor,subsequent should increase ineither case. ing, andhuntingservices and affairs. A section deals with successes inthe agricultural research system over the past decade. Manyof the successes involved generating higher- yielding seed varieties. The key point here is that progress has been made in the recent past but has not been sufficient to induce the desired growth. This point argues for greater support for research, as discussed inthe section onpublic expenditures. Fromtable 5.3, this number is computed by divid- ing the growth contribution o f importables, 1.7 percent, by the agricultural gross value o f produc- tion rate o f 5.3 percent). The SPILL (Strategic Plan for Implementation o f Land Legislation) o f 2005. lo The discussion o f the historical experience with agricultural services is drawn from the program document for the Agricultural Sector Development Programme, developed by Task Force 3 o f the Ag- ricultural Sector Development Programme formu- lation team. I' For example, Mogabiri Extension Micro-Project in T h e district, Eastern Zone Client-Oriented Research and Extension Project inKilosa and Mu- heza districts, and the farmer field schools in Morogoro, Bukoba, and Muleba districts. l2 In Mtwara region, for instance, traders must carry cash fi-om Mtwara town to collection areas. Robbery and injury are ever-present threats that impose costs on marketing agents. l3 One o f the keys to the success o f diversification in the cuttings and seed propagation industry has been the involvement o f Dutch partners, who pro- vide plant material on the condition that all the output is sold back to them for retailing. Diversifi- cation efforts have been helped by the provision o f 122 CHAPTER6 Fostering Growth, Export Competitiveness, and Employment in Tanzania's Manufacturing Sector Background gests that, in general, output, employment, and the capital stock declined during that Tanzania's manufacturing sector is still period. The authors also document large small and contributes only about 8 percent rates o f turnover o f f m s in Tanzanian to gross domestic product (GDP). However, manufacturing, with new firms having in recent years, the sector has experienced higher productivity levels than older incum- fast growtho f about 8 percent per year, and bent firms. They conclude that reforms, in- exports o f manufactured goods have also cluding introduction o f a market-based for- recovered (figure 6.1). The analysis ofpanel eign exchangesystem, liberalization o ftrade enterprise surveys covering 1992 to 2000 by policy, privatization o f state-owned enter- Harding, Soderbom, and Teal (2002) sug- prises, and fiscal policy reform, have 123 SectoralPerspectiveson Growth Figure 6.1. Growth of ManufacturingSector Output and Exports (a) Real growth of the manufacturingsector (b) Exportsof manufactured goods Real growth ofthe manufactudng sector Expods of manubctund QoodS (US$) 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% 4 0 % 1991 1992 1993 1994 1985 1996 1997 1888 18992000 2001 20022003 2004 Source: United Republic of Tanzania various years. promoted a more effective production struc- In 2001, manufacturing's contribution to ture through a reallocation o f capital and total employment was 245,000, or 1.5 per- labor into more productive plants. Harding, cent o f all employees. Manufacturing ac- Soderbom, and Teal (2002) suggest that counted for about one-third o f nonagricul- such a structure should provide the basis for tural private employment (industry and improved growth potential, which seems to services). Although small at present, the sec- have materialized in recent years with tor has the potential to create better-paying growth rates o f around 8 percent in the jobs relative to those inagriculture. Incomes manufacturing sector. o f workers in agricultural typically fall be- low the poverty line. Labor demandin2001 Production i s concentrated in Dar es Sa- inmanufacturing reflectedthe sector's need laam in three types o f firms: (a) agro- for workers with technical skills. Manufac- processing and food and beverage pro- turing employed 34 percent o f all workers duction (fishprocessing, beer, spirits, and with craft andrelated skills, 32 percento fall cigarettes), which are nontraditional natu- machine andplant operators, and 11percent ral resource-based activities; (b) textiles o f all clerks. The manufacture o f labor- and other light industry such as furniture; intensive goods such as textiles, garments, and (c) heavy industry-producing metals furniture, andnonmetallic products occurs in (aluminum and iron sheets), cement, both the formal and the informal sectors. paints, and plastics. Recent growth has Between 2000 and 20002, large firms con- occurred mostly in consumer goods, such tributed the most to employment growth as food products and beer, edible oils, in manufacturing. However, employment textiles and garments, and metals, be- growth was not even across sectors. Agro- cause they have attracted new investment. industryandchemicalsexperiencedjob cuts; Growth in industries such as wood and incontrast, employment inmetalsandplas- paper, furniture, and construction has de- tics grew significantly. For policy makers clined. Other signs o f progress are de- interested in sustained poverty reduction, scribed by increased industrial capacity creating goodjobs through growth inmanu- use. facturing has clear advantages. 124 Fostering Growth,Export Competitiveness,and Employment in Tanzania'sManufacturingSector Tanzania needs a vibrant manufacturing (30 to 99 employees); and 25.3 percent are sector today for several reasons. Histori- large (100 plus employees). cally, worldwide manufacturing has been the foundation for a modern economy. A O f the surveyed firms, 72 percent o f the growing manufacturing sector triggers the firms are located either in the capital (39 development o f ancillary activities and percent) or in other urban areas (33 per- better-paying jobs. A large manufacturing cent). Most o fthe firms inthe larger urban sector can enable export diversification centers are engaged in heavy manufactur- that i s necessary to reduce Tanzania's vul- ing, such as the manufacture o f plastics, nerability to external shocks. It can smooth construction materials, chemicals and incomes at the household level through the paints, and metals. More labor-intensive creation o f nonfarmjobs that are more sta- industries are located mostly in smaller ble and provide higher incomes. On aver- towns with populations o f 1million or age, the monthly income in 2001 from a less. About 72 percent o f the food and manufacturing job was T Sh 103,407 agroindustries, 66 percent o f the furniture compared with T Sh76,277 in mining, and wood products industries, and 65 per- T Sh49,693 in construction, T Sh31,301 cent o f the textiles and garments industries intrade, and only T Sh 15,234 in agricul- are located outside the larger cities. ture, which is at present the largest source o f livelihood for Tanzanians. The average age o f a manufacturing firm i s about 13 years. Except for 2 percent that This chapter analyzes the results o f a re- are publicly held and 1 percent that are cent enterprise survey to identify the de- government owned, the remaining firms are terminants o f enterprise growth, invest- privately owned. Only 16 percent o f all ment, exports, and employment. The firms were previously state owned. O f objective i s to identify what more needs to those remainingfirms, 75 percent were pri- be done to unleashTanzania's manufactur- vatized after 1996, 40 percent after 1998, ing potential in order to achieve its vision and 8 f m s after 1999. Moreover, 20 per- o f a modem economy. cent o f Tanzania's manufacturing firms have some degree o f foreign ownership; the Determinants of Manufacturing average share o f foreign equity i s 72 per- Sector Growth cent. About 5 percent o f all firms are 100 percent foreign owned. The majority o f the The 2003 survey o f Tanzanian firms firms' shareholders are Tanzanian nationals (WED2004) covers the period from 2000 (60 percent). The share o f non-Afican na- to 2002 and consists o f a sample o f 276 tionals i s 14 percent; the share o f Kenyans, manufacturing firms in eight sectors, in- Ugandans, and other African nationals i s cluding food and agroindustry; chemicals between 2 and 3 percent each. The ethnicity and paints; construction materials; metals; o f the principal owners is predominantly furniture and wood products; paper, print- African (44 percent), followed by Asians ing, and publishing; plastics; and textiles, (26 percent), Europeans (6 percent), and garments, and leather products. The sam- Lebanese (4 percent). ple can also be disaggregated by firm size as measured by the number o f employees: Start-up capital for about 72 percent o fthe 6.1 percent o f the firms are micro (1 to 5 firms came from the owner's personal sav- employees); 41.3 percent are small (6 to ings or internal business funds from some 29 employees); 27.5 percent are medium other source. Bank loans, equity or sale o f 125 Sectoral Perspectives on Growth stock, and family and friends play a small related to higher growth in construction, role in financing start-up capital. Money mining, andquarryingin2002 to 2003. lenders and informal sources are also less important. These findings are quite similar Besides macroeconomic stability, what are to those o f Harding, Soderbom and Teal the main determinants o f manufacturing (2002). growth in Tanzania? Empirically, for the period from 2000 to 2002, the key factors When the links between private invest- drivingmanufacturing growth inTanzania, ment and growth were examined at the measured by growth in sales, included microlevel usingfirm survey data for 2000 these characteristics o f a firm: to 2002, a positive association between private investmentand growth inmanufac- Age turing surfaced. That association is en- couraging and underscores the necessity o f Size policies that can boost private investment for faster growth in the sector. Empirical Rate o f investment growth micro analysis i s used to identify the main Rate o f export growth constraints to investment, growth, and em- ployment in Tanzania's manufacturing Use o fnewer technologies. sector. In general, a firm's sector or location During 2000 to 2005, aggregate growth in (large cities or smaller towns) did not af- GDP was 6 percent per year; inmanufactur- fect growth. Similarly, a firm's other char- ing, it was about 7 percent. In2002, growth acteristics, such as the nationality or eth- in manufacturing was 8 percent. Firm sur- nicity o f its owner, did not affect its vey datafor value addedinmanufacturingis performance. unreliable, but trends in sales growth from 2001 to 2002 are consistent with aggregate Sales inolder firms grew more slowly than growth trends. Firm sales grew at 9 percent those in younger ones. But the greatest in2001 andat 11percentin2002. The vari- boost to sales growth came from larger ance insales growthwas large. firms and exporters. About one-third o f medium firms and over one-half of the Between 2000 and 2002, growth in sales large firms are exporters. Higher export increased in correspondencewith firm size performance was an important source o f as measured by the number o f employees. growth. For every 5 percent increase inthe Large firms and exporters outperformed proportion o f sales exported, overall sales smaller firms with sales growth rates o f grew by 1 percent. The strong link be- about 19 percent in 2002. Incontrast, sales tween the proportion of sales exported and in micro firms grew at only 2 percent. growth indicates that exports are a critical Growth across manufacturing subsectors source o f growth for globally competitive was not uniform. The fastest growing sub- firms as they relax the demand constraint sectors were metals (21 percent), agroin- that otherwise limits the size o f the market dustry (18 percent), plastics (17 percent), for nonexporters. The percentage o f the and textiles and garments (16 percent). workforce that could use computers served Furniture sales grew at only 3 percent. as a proxy for the technical skills o f a Faster growth inmetals and machinery was firm's employees. Sales grew by 0.67 per- 126 Fostering Growth,Expod Competitiveness,and Employment in Tanzania'sManufacturingSector cent for each 1 percent increase in the sistent with the theory o f firm growth, number o f employees usinga computer. firms age 10 years or less invested almost 10 times more than the average. Invest- Empirical tests confirm a positive relation- ment growth in smaller firms was nearly ship between investment and sales growth. zero compared with 2.1 percent in large For each percentage point increase in in- firms. Exporters invested substantially vestment, sales grew by 0.08 percent. more than nonexporters; exporters with Compared with the past when that link positive export growth outperformed all was missing, its emergence between 2000 other firms, with investment growth rates and 2002 provides cause for optimism. o f 2.15 percent. Perhaps the investor response was delayed because even though the transition to a Investment decisions are influenced by a market economy began much earlier, it variety o f intricately linked factors. Be- was only after a long lag that investors cause the firm dataset contains more than gained confidence in the government's 100 o f those factors, the latter were listed commitment to reform and began invest- under five broad categories, with each ing inmanufacturing. category containing about 15 to 40 different but strongly correlated variables. The lead- Although policy makers can do little about ingconstraints to investment are as follows: a firm's location, sector, or characteristics o f the owners, at least in the short term, 1. Access to andprice o ffinancial capital policy levers are available and can influ- ence the other determinants o f firm 2. Access to superior technology growth: size, investment, export, and tech- 3. Infrastructure nical skills accumulation in the workforce to support the adoption o f newer technolo- 4. Labor skills andproductivity gies such as those that enable the use o f computers to raise labor productivity. 5. Investment climate A better understanding o f investment in Econometric testing shows that firm char- acteristics defined by age, size, sector, lo- Tanzanian firms can be gained from the study o f Harding, Soderbom and Teal cation, and nationality o f ownership ex- (2002). The authors find that from 1992 to plain some o f the investment growth across all firms. Because policy makers 2001 investment rates were generally quite can do little directly about those character- low. They explain that the low rates could be due to the lumpy nature o f investment, istics, we control for them in our analysis and turn to the other policy-responsive de- which leads many firms to make large in- terminants o f investment. vestments once every four years. Micro- evidence from the 2003 Regional Program Constraint 1:Access to and Price on Enterprise Development firm survey of Financial Capital data validates shorter trends: on average, investment infirms grew at 0.2 percent per Empirical analysis indicates that access to year from 2001 to 2002, and the variance financial capital i s largely determined by was large. About one-half o f all manufac- the nationality or ethnicity o f the firm's turing firms invested in 2001 and 2002. owner and the share of foreign equity in Growth in investment increased corre- the firm. Easier access to financial capital sponding with a firm's age and size. Con- enabled exporters and larger firms to tran- 127 Secforal Perspectiveson Growth scend constraints such as availability o f cial sector. Chapter 10 discusses financial bank financing and higher interest rates sector issues inmore detail. that otherwise constrain the majority o f domestically owned firms. Typically, Constraint 2: Access to Superior firms with access to bank loans for start-up Technologies to lmprove Productivity capital could also use bank financing for investmentandworking capital. Empirically, the three key technology- related determinants of investment in the Most o f the start-up capital for nearly all manufacturing sector are (a) the proportion firms i s sourced from internal savings or o f a firm's machinery that i s no more than retained earnings rather than from bank 10 years old (the larger the share, the bet- financing. Among nonexporters, about 77 ter the performance and effect on invest- percent o f the equity was held by the do- ment growth); (b) the ratio of a firm's mestic private sector, 16 percent by for- capital assets to employment, which re- eign owners, and less than 5 percent by the flects the capital intensity o f production government. In contrast, exporters ob- (the higher the ratio, the more negative the tained almost twice as much equity from effect on investment growth); and (c) the foreign firms, which also facilitated access proportion o f firms that invest to acquire to investment and working capital. They newer technologies. The latter contributes relied less on bank loans despite consid- the most to investment growth and is erably lower interest rates. From 2001 to strongly related to the proportion o f em- 2002, the interest rates available to them ployees using a computer, as well as to the were 10 percent compared with more than level o f investment in research and devel- 16 percent for nonexporters. In general, opment (R&D) to develop or acquire new use o f bank loans was limited. Loan appli- technologies. Capacity utilization and the cations had a high transaction cost, and proportion o f firm machinery that i s no most domestic applicants were rejected. more than 10 years old are strongly corre- Only 20 percent o f the domestic firms had lated factors because older machines re- access to bank loans, which they used to quire more maintenance and lead to fre- finance about 14 percent o f their total in- quent operational disruptions compared vestment and 10 percent o f their working with younger machines. capital during those years. Exporters used bank loans more for working capital. Em- About 48 percent o f all firms, including 65 pirically, a positive nexus existed between percent o f the exporters, invested in the bank loans and investment growth at that purchase o f newer technologies in the pe- time, especially among younger firms. riod from 2000 to 2002, especially inagro- However, because most firms did not have industry, chemicals, andplastics. About 37 access to bank capital, private finance re- percent o f the large firms licensed newer mained the foremost constraint to start-up technologies from a foreign-owned com- capital. pany, and 22 percent o f all firms invested in R&D to develop new technologies. Policy priorities. Enhancing access to Compared with exporters, only 19 percent credit requires the pursuit o f macroeco- o f the nonexporters indicated that export nomic policies and financial sector re- orientation introduces incentives for firms forms that support a deepening o f the fi- to maintain their competitive edge glob- nancial sector and reduce the cost o f credit ally. Average spending on R&D varied by increasing the efficiency o f the finan- from TSh300,OOO in small firms to 128 Fostering Growth,Export Competitiveness,and Employmentin Tanzania'sManufacturingSector T Sh 5.0 million inlarge firms. Relative to days per year. Around 54 percent o f the average o f T Sh 3.9 million for all sec- the firms have installed private power tors, firms in agroindustry and metals in- generators at costs ranging from vested T Sh 10.0million each. On aver- TSh172,700 to TSh812,500 per em- age, only 5 percent o f the workforce i s ployee. Private provision o f water entails computer literate; in larger firms, com- additional costs o f construction in chan- puter literacy is more frequent-about 8 neling water to factories. About 40 per- to15 percent o f the employees use com- cent o f industrial water i s obtained from puters. Relative to exporting firms, more private sources. About 34 percent o f the employees in nonexporting firms use firms own a well, 31 percent have in- computers. On average, 12 percent o f all vested in their own water infrastructure, firms were certified by the International and about 12 percent share a water source Standards Organization (ISO), with the with the local community. For a country highest proportion among large firms (37 with a rich rural hinterland that supports percent) and exporters (24 percent). Firms agroindustry, the shortage o f reliable manufacturing plastics, textiles, construc- freight transportation by road, rail, and air tion materials, and agricultural products i s also a direct constraint to growth. had higher levels o f I S 0 certification than About 41 percent o f the firms lose be- others. tween land 50 percent o f their cargo in transit because o f spoilage, breakage, and Policy priority. Given the critical need for so forth, At least 10 percent o f the firms access to newer technologies to raise pro- invest in private roads; 18 percent have ductivity and export competitiveness, policy invested in freight transportation to cart makers should explore what, ifanythmg, the goods back and forth. About 26 percent government can do to promote the adapta- o f the firms have invested in transporta- tion andadoption o f superiortechnologiesat tion for employees to compensate for lack the firm level. Inthat context, the experience o f good public transportation to and from o f other developing countries such as Chile the workplace. and those in East and South Asia may be instructive. Lessonsfrom FundacihChile, a Another area with problematic infrastruc- nonprofit institution, and from India's insti- ture i s telecommunications. Only 49 per- tutes that facilitate technology for exporting cent o fthe firms have Internet connectivity, grapes may be useful (see Chandra 2006). but power outage-related spillovers dimin- Chapter 9 provides a more in-depthdiscus- ish the ability of those firms to access in- sion o fthe drivers o f innovation andtechno- formation in a timely manner. Firms pay logicalchange. about T Sh 11,428 per employee per year in telecommunications bills. Although that Constraint 3: infrastructure amount may seem meager, the opportunity cost o f regular and easy access to informa- Firms located in industrial estates do not tion technology i s yet another factor that appear to benefit from superior infrastruc- constrains Tanzanian firms from develop- ture although such benefits were the in- ingand maintaining better links with global tent o f development o f such infrastruc- markets inpursuito ffaster growth. ture. Measured interms of a single eight- hour period, power outages amount to at Empirical analysis indicates that the main least 18 working days per year. For some constraints in infrastructure are related to firms, the disruptions last 56 working electricity, water, road transportation, and 129 Sectoral Perspectives on Growfh factors such as telecommunications, Inter- a skilled labor force proxied by higher edu- net connectivity, and waste disposal facili- cation andtraining. Compared with the less- ties. Overall, infrastructure-related factors skilled individuals, skilled workers arebetter explain a significant proportion o f invest- at adapting to the new technologies o f pro- ment growth, especially for younger firms, duction that canraise firmproductivity even mostly because o f the exacerbated effect o f for simple goods such as T-shirts. The larger water- and transportation-related con- the share o f the workforce with secondary straints. Poor public infrastructure has and vocational education, graduate and driven many firms to invest in and substi- postgraduate education, and formal training, tute private infrastructure for public infra- the more likely that a fm will invest in a structure, but at a high opportunity cost as particular year. measuredby forgone investment for expan- sion. That situation is evident from the Unfortunately, data problems prevent a negative effect o f electricity, water, freight, meaningfbl analysis of the effect o f and cargo transport constraints on private HIV/AIDS on firms' investment rates. investment. Private provision o f those pub- However, other sources o f information lic goods imposes higher fixed costs per suggest that the pandemic is inflicting se- unit of installation necessary to operate vere costs on firms through decreased la- electric generators, water wells, and trans- bor productivity. portation, among others. In contrast, the provision o f private roads, Internet access, Between 2000and 2002, about 44 percent and telephone connections affects firms' o f the manufacturing workforce had com- decision to invest positively. It is plausible pleted primary education, 25 percent had that public goods such as private roads that secondary education, 12 percent had some require high upfront investment but low vocational education, and 7 percent had maintenance costs serve as assets and en- tertiary education or a diploma. In fast- courage fbture investment positively, growing firms such as exporters, the pro- whereas goods such as generators, water portion o f graduates or postgraduates (10 infrastructure, and freight transport, which percent) was twice as high as in firms that have high recurrent and replacement costs, sell domestically. Exporters also offer crowd out investments that expand firms. more training: 71 percent o f their work- force had received some type o f formal Policy recommendations. Improving training compared with only 47 percent o f the supply and regularity o f electricity that o f nonexporters. The high cost o f and water and the supply and quality o f firm-level training discourages large-scale freight transport would develop and en- training, especially o f workers who have hance the infrastructure needed for only primary education. The larger the growth. Chapter 10 provides an analysis share o f employees with graduate, techni- o f the challenges Tanzania faces in ex- cal, or tertiary education is, the more will- panding access to infrastructure. ing firms are to invest in their training, which probably explains why firms in Constraint 4: Labor Skills Tanzania and other Sub-Saharan African and Producfivify countries invest less in training than firms inEastAsia. Empirical analysis indicates that the key human capital-related determinants o f Ifthe creation ofhuman capital is not ac- firms' decision to invest are associated with celerated, Tanzania's manufacturing sec- 130 FosteringGrowth,Export Compefitiveness,and Employment in Tanzania'sManufacturingSector tor i s at risk o f either not attracting inves- Constraint 5: lnvesfment Climate tors or attracting investors only shopping for low-skilled workers. Bothpossibilities Although Tanzania's manufacturing sector are detrimental to investment growth in faces a plethora o f problems in areas such manufacturing, likely to generate low- as access to financial capital and to supe- wage employment, and a deterrent to rior technologies, infrastructure, and hu- economic diversification through an ex- man capital, the investment climate- pansion in manufacturing to move work- related constraints are less problematic. ers out o f poverty. The latter requires Whereas delays for most processes for firms to graduate into higher value-added production that needs skilled workers. starting or renewingbusinesspermits seem reasonable, bribes and excessive delays for More importantly, Tanzania's advantage many exporters and larger firms are prob- o f low labor costs i s not permanent. With lem areas. For example, instead of a nor- increasing integration into global mar- kets, its low- skilled, low-cost workers mal period o f 20 days for customs clear- ances, the delays lasted 97 to 150 days for will have to compete with high-skilled, many exporters and larger firms. Exporters low-cost workers from other competitors, wait much longer than nonexporters for especially those in Asia. Early signs o f inspection certificates, utility connections, that competition are apparent from the and construction permits. The bribes paid stagnant investment rates in manufactur- ranged from T Sh 35,000 for a water con- ing, especially in the garments subsector, and highlight further the sector's unsus- nection, to T Sh50,000 for a telephone line, to T Sh 150,000 for an import license. tainable growth. Delays for registration or re-registration Policy recommendations. Although ex- for business permits for larger firms last as pansion o f primary and secondary educa- long as 180 days. The bribes paid for such tion i s necessary, it i s not sufficient to at- permits are correspondingly high, ranging tract investment in manufacturing. Public from T Sh 50,000 to T Sh 180,000, with investmentintertiary and technical educa- the outliers inthe range o f T Sh 500,000 to tion i s critical for Tanzania to create the T Sh2.4 million. Eighty-four percent o f all large pool o f skilled labor needed to en- firms hire a company or an agent to assist courage firms to invest, to diversify pro- with registration. duction, and to increase export competi- Ongoing reforms are already redressing tiveness. Public investment is also the following areas: necessary to facilitate the use o f superior technologies, promote technological leam- Transactions costs o fnumerous licenses ing amongworkers and firms, and enhance and permits are being reduced as gov- productivity at the labor and firm levels. ernment officials are trained to operate Although the effect o f HIV/AIDS does not a one-stop shop for all investors. emerge in our empirical analysis, we would be unrealistic to ignore timely pub- Legislative reforms to protect private lic interventions that can increase preven- property rights and an efficient judicial tion and treatment o f HIV/AIDS among system are necessary for attracting lar- workers and their families. The opportu- ger investors. nity cost of not doing enough is high for Longer-term leases (100 years) are an both the government andthe firms. option for attracting fixed investment 131 Secforal Perspectives on Growfh inland and buildings, as well as com- constraints to production are relaxed fur- mercial courts that can speedily re- ther, manufactured exports can yield some solve disputes. rapid short-term gains, as evidenced by the double-digit growth experienced in recent Enhancing the Export years. In the medium- to longer-term, the Performance of the ability to export will be determined, as in Manufacturing Sector any other country, by the global competi- tiveness o f manufacturing firms. Their Tanzania's exports o f manufactured goods presence today is no guarantee o f their fu- have recovered since 1999. The share o f ture place inthe global market. nongold exports intotal exports i s 52 per- Existing exporters rely more on foreign cent. The share o f manufactured exports in sources o f financing, especially from parent nongold exports increased to 17 percent in companies or private capital. For most ex- 2004 (US$120 million) from 6 percent in isting exporters, 61 percent of the equity is 1999. Although a favorable trend in the private and domestically owned, and 31 real exchange rate between 2001 and 2004 percent i s foreign owned. For nonexporters, supported the growth in manufactured ex- 75 percent o f the equity is private and do- ports, an equally if not more critical factor mestically owned, and only 16 percent i s was the wide-ranging reforms to generate foreign owned. Clearly, local firms that are efficiency. These reforms led to acceler- limited to private sources o f start-up and ated production and boosted the competi- investment capital are also potentially con- tiveness o f Tanzanian exports. strained from entry into the exporting busi- ness. Existing exporters enjoy increased If exporters are defined as firms that ex- access to bank financing for working capi- port at least 10 percent o f their production, tal, commercial loans through overdraft only 19 percent o f Tanzanian firms are privileges, and lower interestrates. exporters. They are the key drivers o f the manufacturing sector, which grows at 0.2 Incontrast to nonexporters, exporting firms percent for every 1.0 percent o f additional employ a workforce with relatively higher production exported. Only 11 percent o f education levels-graduate, technical, and small firms are exporters as compared vocational-a proxy for skills. The ratio o f with 55 percent o f large firms. In2001 and skilled to unskilled workers in exporting 2002, the growth o f exporting firms was firms i s 1.6 times that o f nonexporters. Ex- about 24 percent per year as compared porters pay a premium for higher skills, with 8 to 11 percent per year for nonex- averaging 20 percent for managers, 37 per- porters. A large variation exists inthe pro- cent for professionals, and 19 percent for portion o f the output that firms export. those with technical skills. They pay about Approximately 25 percent o f the exporters 6 percent less for unskilled workers. On channel almost 100 percent o f their pro- average, they have five times more (10 per- duction to foreign markets, thus under- cent) workers with computer skills than do scoring the point that market demand i s nonexporters (2 percent). Given the higher not the bindingconstraint for firms that are share o f skilled workers, more exporters competitive. In Tanzania's small open (71 percent) invest in formal training than economy, with all else remaining the do nonexporters (47 percent). Exporters same, exporters can sell, in the short term, also have a larger proportion o f foreign all that they can produce. If the domestic managerswith more experience. 132 Fostering Growth,Export Competitiveness,and Employment in Tanzania'sManufacturingSector Inthe absence of publicly provided infra- made possible because o f (a) a larger structure and basic services, at least twice share o f own-financing from retained as many exporters as nonexporters, espe- earnings and private capital, (b) greater cially smaller firms, are able to privately foreign equity, and (c) easier access to finance their idrastructure. Exporters and bank loans for start-up and working large firms have equal ability to compen- capital. Their size creates spillovers that sate for the poor public infrastructure more thancompensate for the longer de- through the private provision o f roads and lays they face at ports and in obtaining transport and water wells, although this licenses and permits to do business. By ability must come at the cost o f investment. exploiting economies of scale, they can pay higher remuneration to attract more Seventy percent of Tanzania's manufac- skilled workers. Higher rates of com- tured exports are destined in equal propor- puter skills in the skilled workforce en- tions to two main regional markets: (a) able larger firms to license modern tech- Western Europe and (b) the regional Afri- nologies that boost competitiveness and can market, which is dominated by Kenya increase their capability to export. and Uganda. About 66 percent o f the ex- porters each sell nearly 35 percent of their Relative to firms in the garments and exports to those two regions, suggesting, textiles sector, firms in agroindustry at the very least, that a sizable share o f ex- contribute significantly to exports. isting Tanzanian exporters are as globally competitive in Western Europe as they are Larger firms, especially Kenyan-owned inthe African market, where competition firms, are the main drivers of exports from low-income Asian exporters such as relative to all other nationalities. Eth- China is rising. Considering that only 19 nicity also matters; relative to European percent of all firms exported more than 10 firms, the export performance o f Afri- percent of their output, one is encouraged can-owned firms is below average. to find that exports to markets inindustrial Export experience is important. Firms countries are not dominated by one or two that acquired exporting skills in the large firms, as is often the case in many past five years are able to export a lar- low-income countries. ger proportion of their output. Simi- What are the determinants of manufac- larly, given the complexity of export- tured exports? By measuring the share of ing, new entrants in the Tanzanian output exported, an empirical analysis re- market with prior foreign experience veals that the key determinants of export as exporters perform better. growth are as follows. High interest rates reduce use of bank loans for investment and start-up capital Firm size and sector matter. As export- andnegatively affect export expansion ers are mostly large firms, sales growth inlarge firms also translates into export The availability o f workers with growth. Compared with smaller firms, graduate and postgraduate education is large firms in Tanzania's manufacturing important and indirectly supports the sector have several advantages that en- development of computer skills. This able higher rates of productivity, which, is natural as more manufacturing proc- in turn, raise efficiency and growth. esses, especially in heavy industries They have a higher rate of investment, such as chemicals, paper, metals, and 133 Sectoral Perspectives on Growth machinery, are capital intensive and tained growth o f the manufacturing sec- need more sophisticated skills, includ- tor, there i s no alternative to export ingcomputer capabilities. growth. We recommend the following policies: 0 Exports destined for the Southern Af- rican Development Community or the 0 Two types o f discrete export promo- local regional markets inKenya do not tion incentives are needed: (a) those grow as fast as those destined for mar- that can retain existing firms and fa- kets outside Africa, such as Western cilitate their ability to export more and Europe, Eastern Europe, the United (b) those that can attract new entrants States, and Asian countries. The size into the export business, especially ex- and purchasing power of the latter i s porters equipped with globally com- much larger than low-income African petitive technologies needed to pene- markets, where competitionfrom other trate new markets and grow faster. regional African exporters selling Targeting incentives at large firms, similarproducts is high. both domestic and foreign, should be most effective in raising export 0 The government offers about 12 export promotion programs. However, less growth. With appropriate design, such than 25 percent of the exporters use a policy should also encourage learn- any ofthem. Among the export promo- ingbetweenexisting firms and interna- tion programs that are most popular tionally competitive ones to promote and used by at least 10 percent of the export competitiveness. firms are (a) retention o f export pro- 0 Scaling up exports requires more and ceeds in a foreign country, (b) the larger exporters. However, With the ex- bonded warehouse scheme, (c) the cus- ception o f foreign firms that have access toms duty drawback, (d) duty certifi- to cheaper capital fiom retained earnings cates, and (e) the profit tax exemption or parent companies, domestic manufac- scheme. turing firms' access to start-up and in- vestment capital is constrained by pri- Policy Recommendations vate savings. For firms that have access, highinterest rateslimituseofbank capi- Reinforcing the ability to produce and tal. Financialpolicies that increasecredit export i s key to overall growth in Tanza- for fixed investments for large firms are nia's manufacturing sector. For every 5 required. The government could learn percentage point increase in the propor- from policies adopted inEast and South tion o f output exported, manufacturing Asia: using the ability to export as the growth increases by 1percent. An impor- measuring rod, proactive governments tant aspect o f export growth i s the strong were able to preempt government failure link between large firms and exporters. by using credit expansion and lower in- Empirically, though most exporters are terest rates as effective policy levers for large firms, it would inpractice be incor- export expansion. rect to consider domestic sales growth as a substitute for export growth. Relative to 0 Given the weak state of infrastructure Tanzania's small domestic market, its ex- countrywide, an option for the gov- port markets offer unlimitedopportunities ernment i s to target infrastructure pro- for growth. In fact, for rapid and sus- vision to export processing zones and 134 Fostering Growth,Export Competitiveness,and Employmentin Tanzania'sManufacturingSector the main industrial areas to maximize marketing, compliance with global agglomeration economies and reduce standards, facilitation o f newer tech- the cost o f private infrastructure for nologies, and technological capacity exporters. As industrial areas house building. Third, for labor-intensive ex- large firms, such a policy should sup- ports such as garments and furniture, port the entry o f new exporters from policies that enable firms to link with larger firms as well as the scaling up o f global supply chains such as those that existingexporters. attract foreign multinational corpora- tions into local assembly and manufac- 0 The role o ftechnical andtertiary educa- turing are needed. Lesotho and Mada- tion in enhancing export competitive- gascar are good examples. A proactive ness i s critical. Policies with a longer- policy stance to signal the govern- term vision to substantially scale up ment's new commitment to promote technical and tertiary education are the exports-apecially exports to non- most important levers for improving la- African markets-is needed. bor productivity and technological up- grading to achieve global competitive- 0 Selectivity is recommended: strengthen ness and sustained export growth. This the few export promotion incentive pro- recommendation also implies more in- grams that f m s actually use (bonded vestments in primary and secondary warehouse scheme, retention o f foreign education to enhance the quality o f the exchange, duty drawbacks, profits tax throughput to higher education. exemption scheme, and duty certifi- cates) andrationalize the remainder, but 0 The prevailing policy attention on ex- most importantly, design special pro- port competition with neighboring grams to support export penetration in Kenya i s myopic when the true compe- non-Afiican markets. Additionally, a tition in the home and neighboring concerted monitoring o f the one-stop marketsis with low-income largeAsian shop to eliminate delays in issuingper- exporters such as China. A shift inpol- mitsto exporterswill help. icy focus is warranted to think outside o f the AfZca box-away from export markets within Africa and toward non- Conclusions and African markets. Three differentiated Recommendationsfor a export promotion policies tailor-made Manufacturing Sector Growth for the various export sectors are re- Strategy quired. First, for heavy industrythat re- lies on local natural resources, public A growth strategy basedon supporting the policies to support expansion and raise manufacturing sector-and within it larger productivity should be o f high priority. firms, especially exporters-has nontrivial Second, agro-processing o f local non- poverty implications for the large propor- traditional produce (such as fish and tion o f unemployed and relatively less- fruits) i s also natural resource-based skilled Tanzanians. Faster rates o f growth but requires meeting complex phyto- inthe sector contributed to higher growth sanitary standards to gain entry into in the overall economy and to the largest marketsinindustrial countries. Moving proportion o f better-payingjobs relative to up the value chain requires policies that jobs in the agricultural and informal sec- reduce the high costs o f information, tors, which fetch incomes that typically 135 Sectoral Perspectives on Growth fall below the poverty line. For example, Tanzania and i s recommended. The ra- in 2001, manufacturing generated about tionale for that selective approach i s moti- one-third o f nonagricultural private em- vated by today's global reality: if a firm ployment. How can policy makers maxi- cannot compete in the global market (that mize the sector's growth potential? Em- is, export), it i s unlikely to survive for too pirically, larger firms are the drivers o f long in Tanzania's domestic or Africa's growth and employment within the manu- regional markets, which are flooded with facturing sector. In absolute terms, in a cheaper imports from low-cost, high-skills sample o f 264 f m s , larger firms with producers such as those from East and more than 30 employees generated 29,000 South Asia. jobs relative to only 1,700 in smaller firms. Employment in larger firms grew at Firm size is a critical determinant of firm 8.5 percent year relative to only 3.0 per- growth in Tanzania. And, as many larger cent firms. Unsurprisingly, larger firms firms are also exporters, and most export- systematically paid significantly higher ers are larger firms, policies targeting lar- wages. The median wage for professionals ger firms should have large payoffs in in larger firms was T Sh240,000 relative helping to expand existing firms and pro- to T Sh 100,000 in smaller firms. For mote new entrants in the export sector. skilled workers, the median wage was Although policies that favor large firms T Sh 90,000 inlarger firms compared with also favor exporters, export promotion T Sh 60,000 in smaller firms, and for un- strategies are important in their own right, skilledworkers, it was T Sh60,000 in lar- given the limitedpurchasing power in the ger firms relative to T Sh 50,000 in Tanzanian market. This recommendation smaller firms. Larger firms also paid sig- requires a strategic, two-pronged ap- nificantly higher compensation rates. proach: one that targets large firms, and another that targets existing and potential Empirical analysis o f the determinants o f exporters . manufacturing firm growth shows that the two leading contributors to firm growth 0 Policies are needed that redress domes- are a technically skilled labor force and tic supply constraints associated with growth o f large firms, which also represent disproportionately higher transaction the majority o f the exporters. A 1 percent costs o f investment faced by large firms increase inthe technical skills o fthe work- and exporters. Policies that increasein- force increases firm growth by 0.7 percent, vestment and a program that reduces and a 5 percent increase in the output ex- obstacles associatedwith finance, infra- ported delivers almost a 1percent increase structure, technology, and skills are re- in firm growth, which, in turn, raises em- quired. All o f that may seem tanta- ployment growth. The third important de- mount to recommending everything- terminant of growth in manufacturing that is, redressing all barriers to produc- firms i s investment growth (growth elas- tion currently facing all manufacturing ticity i s 0.08). The jump in firm growth i s firms in Tanzania. It i s not. To circum- intricately tied to growth in exports be- vent the high financial and time costs cause o f limited purchasing power in the and the government's weak implemen- domestic market; therefore, an aggressive tation capacity, we recommend focus and proactive policy stance for promoting and pragmatism in catering to larger manufactured exports i s likely to have the f m s . In delivering physical inputs biggest effect on manufacturing growth in such as infrastructure and financial in- 136 Fostering Growfh,Export Competitiveness,and Employment in Tanzania'sManufacturing Sector puts such as more accessible bank fi- tiveness i s best tested. That challenge nancing, the government could identify requires a different set o f externally ori- spatial locations where manufacturing ented policies that exploit Tanzania's and export activity i s most prevalent. It latecomer advantageto leapfrog into the could then reinforce the delivery o f global marketplace. The starting point public inputs and services to those loca- should be direct public support to facili- tions. Spatial identification helps intar- tate the delivery o f the two key public geting exports. Similarly, to attract po- goods: (a) superior technologies o f pro- tential exporters, the government could duction through adaptation and (b) de- identify special areas, such as industrial velopment o f technical, tertiary, and districts and export processing zones, managerial skills neededto apply them. where larger firms interested in export- Public-private partnerships have served ing would be able to benefit from im- as the best mechanism to deliver those proved infrastructure and service deliv- two public goods. The ability o f Fun- ery, as well as financial support. That daci6n Chile to promote technology approach would render public support transfer and adaptation and that o f the in a financially feasible and timely Indian Institutes of Technology and manner for fast-growing exporters and Management to deliver critical skills of- potential new entrants into the export fer usefbl lessons. Additionally, finan- business. The manufacturing sector cial incentives are needed to reduce the cannot afford to wait until the con- highfixed cost ofentry into export mar- straints to investment are resolved kets and to attract firms and sustain economywide. them inthe export business. Special in- centives to promote exports out o f M- 0 Although the entry o f f m s into the ex- rica are likely to have the highest payoff port sector can be facilitated, much and sustainability. Those challenges re- more is needed to sustain them in the quire creativity (that is, thinking outside business. Creating and nurturing firms' o f the Afkica box) and political com- ability to export implies grappling with mitment. Lessons from East and South the challenge o f improving productivity Asia are good starting points, especially and spurring competitiveness to export in the area of agroprocessingand light to non-Afkican markets,where competi- manufacturing. 137 CHAPTER 7 Tourism in Tanzania Tanzania i s an up-market tourism destina- the principal destination for wildlife- tion. The country is endowed with a variety viewing safaris. However, the government o ftourism assets, including six World Heri- i s encouraging the development o f the tage sites and numerous wildlife parks, Southern Circuit, including the Selous Re- beach resorts, coral reefs, and spectacular serve, which is among the world's largest mountain scenic views. Twenty-eight per- natural reserves, to prevent overexploita- cent o f Tanzania's landmass i s protected tion o f the Northern Circuit. Other princi- area, consisting o f 15 national parks, the pal tourist destinations include the beach Ngorongoro Conservation Area, 31 game resorts, mainly on the island o f Zanzibar. reserves, and 38 game-controlled areas. Wildlife safaris and beach resorts are of- fered as single-destination attractions and Currently, wildlife i s the prime tourist at- combination packages. traction. The Northern Circuit, including the Ngorongoro Crater, the Serengeti Na- Starting in the early 199Os, leadership o f tional Park, and Mount Kilimanjaro, i s still commercial development in the tourism 139 Sectoral Perspectives on Growth Figure 7.1. Tourism Receipts in Tanzania, 1991-2005 900 800 Source: Data from the National Bureau of Statistics, Tanzania. industry shifted from the government to plays an important role in attracting for- the private sector. The government formu- eign direct investment (FDI). In 1999, it lates policies, regulates and promotes in- brought in 13 percent o fthe total FDI. vestment and services, and facilitates the supporting infrastructure. In 2005, Tanzania's average hotel occu- pancy rate was 48 percent, down from a Economic Contribution peak o f 64 percent in 1999 (table 7.1). of Tourism However, there are regional differences. In general, the Northern Circuit rates (70 to The Tanzanian government regards tour- 80 percent) are higher than the Southern ism as a priority sector. The contribution Circuit rates (40 to 45 percent). In 2001, o f the tourism industryto the gross domes- Europeans accounted for 31 percent o f the tic product (GDP) rose from 7.5 percent in visitors to the mainland and for 75 percent 1995 to 12 percent in 2001 and to around o f the visitors to Zanzibar, whereas Afii- 16 percent in 2004. In 2004, the tourism cans made up 40 percent o f the visitors to industry generated nearly 25 percent of the mainland. total export earnings. Throughout the An estimated 40 to 50 percent o f holiday 199Os, the tourism sector has performed visitors come overland from Kenya to very well and shown high growth rates Tanzania. That percentage constitutes a (see figure 7.1). From 1990 to 1999, tour- significant change from 1996, when it was ist arrivals and foreign exchange earnings estimated that some 60 percent o f holiday from tourism increased by an average an- visitors came through Nairobi. According nual rate o f 15.15 percent and 27.41 per- to the European tour operator survey, cent, respectively. However, in 2000, the three-fifths o f respondents indicate that tourist arrival rate fell by 2.2 percent, Tanzania i s now sold as a standalone des- partly as a result o f the terrorism attack in tination, rather than an add-on to a Kenya Dar es Salaam in August 1998. Since program (Ministry o f Natural Resources 2000, tourist arrivals and foreign exchange andTourism 2002). earnings have grown only modestly at av- erage annual growth rates o f 4.1 percent Tourism contributes not only directly to and 2.1 percent, respectively. Tourism also growth but also indirectly through its links 140 141 Sectoral Perspectives on Growth with other sectors o f the economy. The ef- those sectors. A study o f Northern Circuit fect o f an increase in tourism expenditure hotels and lodges conducted by the Multi- on economic activity in a country (such as lateral Investment Guarantee Agency output, income, or employment) can be (MIGA 2002) found that each room is es- measured by using multipliers. For Tanza- timated to create two jobs directly. In ad- nia, the output multiplier is estimated at dition, the study found that 27 percent o f 1.83, meaning that an increase o f revenues go to imports, 36 percent to ex- T Sh 1million in tourism output causes the penditureson goods and services produced output in the economy to increase by inTanzania, 15 percent to income, and 24 T Sh 1.83 million, implying that other sec- percent to government taxes. The average tors expand to service the tourism industry. daily expenditure per tourist was The backward multiplier for tourism in US$172.58 in2001, which has risen stead- Tanzania is 1.16, which measuresthe stim- ily from US$122.00 in 1995. uli given to supplying sectors because of increased tourism demand. A backward Given the backward links of tourism, that link of 1.16 requires the output of supply- sector has a large potential for increasing ing industriesto rise by T Sh 1.16 million, value added in the economy. Therefore, with a T Sh 1million increase in tourism tourism is important for reducing poverty in output (Kweka, Morrissey, and Blake Tanzania. Households that are involved in 2003). Both multipliers are higher than the tourism are 10percent less likely to bepoor respective multipliers for agriculture, andtherefore show lower poverty ratesthan manufacturing, and other services. Tourism food crop or fishproducers andmining sec- requires 44 percent o f its inputs fkom other tor households. Households close to pro- sectors, a rate that i s above the average o f tected wildlife areas also have a potential to all sectors. The most important input sec- earn tourism income. A study found that tors for tourism inTanzania are agriculture, the number ofhouseholds receiving income livestock, poultry, fisheries, dairy, manu- fkom tourism varies from 3 percent for facturing, nonperishable foods and dry households near the Loliondo Game- goods, ground transportation, and handi- Controlled Area to 12 percent for those crafts. Many o f the products are sourced near the Ngorongoro Conservation Area locally but are not necessarily produced in (Homewood and others 2001). However, Tanzania. Furthermore, even though some tourism i s rarely the principal source o f in- products are locally produced, tourism op- come, and the earnings are highly skewed erators often choose to import such prod- toward the elite. Another study found that ucts because o f limited domestic availabil- the combined (labor and nonlabor) income ity o f variety and quantity and because o f multiplier for tourism is 0.66, the lowest relatively lower quality. compared with other sectors in Tanzania (Kweka, Morrissey, and Blake 2003). The tourism sector employed 160,750 However, the tourism sector has the highest people in 2004 compared with 96,000 in value o f the labor income multiplier(0.45), 1995. Overall, tourism has a relatively which indicates the relative prevalence o f high employment multiplier, 5.39, which paid labor in that industry. The generally i s the number o f employees for each low values o f income indicators reflect low T Sh 1million increase in final demand levels o fwages, employment, or both. for tourism. About 75 percent o f the effect on employment i s to the benefit o f other Tanzania has a tax multiplier o f 0.08 for sectors, given the high links o f tourism to tourism-again higher than in agriculture, 142 Tourismin Tanzania Figure 7.2. Total Visitor Arrivals in Kenya and Tanzania, 1996-2004 Total Visitor Arrivals inKenya andTanzania (1000) 1600.0 1400.0 1200.0 1000.0 800.0 600.0 400.0 200.0 I 0.0 1996 1997 1998 1999 2000 2001 2002 2003 2004 --tKenya +Tanzania Sources: Data from the National Bureauof Statistics, Tanzania, and Central Bureau of Statistics, Kenya. manufacturing, and other services. That Tanzania has more than four times the multiplier i s probably rather high because landmass conserved as national parks. In o f the relative ease o f taxing hotels and addition, Tanzania's wildlife i s considered restaurants, which implies that the most to be superior interms o f quality, quantity, efficient way to tax tourism i s through tax- diversity, and visibility to that in compet- ingtourist expenditure inthe country. An- ing destinations. other study found that tourism has an un- ambiguously favorable effect on tax One reason for the differential in visitor revenue (Kweka 2004). A 10 percent tax arrivals i s that Tanzania targets higher- on all tourist expenditures increases gov- income tourists and tries to avoid the style ernment revenue by 2 percent, real GDP o f mass tourism fostered in Kenya. Mass by 0.3 percent, and total welfare by 0.2 tourism i s not seen as an option for Tanza- percent. nia because o f the fragility o f its natural The Tanzanian Tourism assets. In addition, targeting low-volume, high-yield tourism allows Tanzania to Industry Compared with That keep the image o f exclusivity, which is of Other Countries feasible because o f its assets. In terms of tourism assets, Tanzania can Tanzania's tourism industry shows better be compared with Kenya because both results if compared in terms o f earnings countries feature a number o f game parks, from tourism. In 2004, Tanzania received mountains, and lakes, as well as beaches US$746 million intourism earnings, while on the Indian Ocean. However, in 2004, Kenya earned US$495 million. Tanzania's Kenya received about 780,000 visitors performance interms o f growth o f tourism more than Tanzania (figure 7.2), although receipts by far surpassed that o f Kenya 143 Secforal Perspectiveson Growth Table 7.2. Hotel Occupancy Rate Time of occupancy Kenya(%) Tanzania (%) Uganda (%) Average 2002103 50.10 50.85 47.50 High season 75.78 72.92 72.30 Low season 27.61 34.52 24.04 Source: World Bank 2005d. Figure 7.3. World Travel and Tourism Council Competitiveness Index, 2004 Indexvalue: Least Competitive is 0; Most Competitive is 100 Source: World Travel and Tourism Council database. Note: Indexvalues range from 0 (least competitive) to 100(most competitive) during the 1990s:from 1990 to 1995,Tan- have an occupancy rate of about 50 per- zania showed an average annual growth o f cent. However, Tanzania shows better receipts of 31.8 percent, while Kenya's results during the low season: 34.5 per- receipts grew by 1.9 percent annually. cent as compared with 27.6 percent in From 1995 to 2000, Tanzania's average Kenya (table 7.2). annual growth o f receipts slowed to 23.3 percent, and Kenya marked an average According to the World Travel and Tour- annual decrease in receipts o f 10.7 per- ism Council Competitiveness Monitor cent.' Kenya picked up growth in the pe- 2004,* an online database (http://www riod from 2000 to 2004, experiencing an .wttc.org/Nu_compmon/compmonO4/Intro. average annual growth rate o f receipts of htm), Tanzania is less competitive overall 18.9 percent, far surpassing Tanzania's thanKenya, Mauritius, or Thailand (figure slow average annual growth rate o f 0.24 7.3). Tanzania's price level i s higher than percent. that o f Kenya and Thailand. Even though the tourism infrastructure in Tanzania i s In terms of the hotel occupancy rate, slightly betterthan inKenya, it is still very Tanzania shows similar figures to Kenya. weak. The levels o f technology, education, On average, both Kenya and Tanzania openness toward trade and visitors, and 144 Tourism in Tanzania Table 7.3. Perceptionof Infrastructure Services Percentage of respondents rankingservices poor or very poor Service Kenya Tanzania Uganda Roads 64.70 37.88 28.60 Waste disposal 61.80 33.33 17.90 Security 38.20 18.18 17.90 Water 35.30 28.79 10.70 Land telecommunications 32.40 6.06 14.30 Railways 29.40 1.52 7.10 Electricity 23.50 46.97 21.40 Mobiletelecommunications 17.60 4.55 3.60 Postalservice 14.70 1.52 11.10 Air freight 11.80 4.55 33.30 Trucking 2.90 1.52 18.50 Source: World Bank 2005d. Note: Boldtypeface indicates the three poorest services in each country. social factors are very low in Tanzania, nesspeople surveyed ranking that problem resulting in international rankings in the as major and very severe, the provision of lowest third. electricity i s a significantly greater prob- lem in Tanzania than in Kenya and Competitiveness in tourism i s determined Uganda (table 7.4). Tax administration i s by price, product, infrastructure, and ena- also rankedmore often as a major or very bling environment. A market survey of severe obstacle in Tanzania than in Kenya tour operators found that the quality o f or Uganda. Tax rates, however, constitute tourist services was perceived as a major a major obstacle in Tanzania as well as in weakness o f the product (Murphy and Kenya andUganda. Henegan2002). Furthermore, respondents indicated that a lack o f training could be Growth Potential detected and that service orientation should be more professional. The government o f Tanzania has set a tar- get o f 1million tourists by 2010 bringing Infrastructure services are a major obstacle US$1.5 billion in receipts. In the medium to the tourism sector in Tanzania. Almost term, the objective i s to increase the aver- 47 percent o f the surveyed tour operators age annual growth rate o f the tourism sec- perceive electricity services as poor or tor to 8 percent by 2005/06. very poor (table 7.3). In Kenya, roads, waste disposal, and security are regarded Tanzania's tourism potential i s largely un- as biggerproblems than electricity. derexploited; the sector can therefore make a greater contribution to growth and The recent Investment Climate Survey for poverty reduction. As noted earlier, Kenya Tourism in East Africa found that the in- has received more visitors than Tanzania, adequate provision o f electricity i s per- but Tanzania has bothmore landmass con- ceived as a major obstacle in the business served as national parks and superior wild- environment. With 62.1 percent o f busi- life compared with Kenya. Tanzania also 145 Sectoral Perspectives on Growth Table 7.4. Obstacles Encounteredin the Business Environment Percentageof respondentsrankingobstacles major andvery severe Obstacle Kenya Tanzania Uganda Telecommunications 23.5 10.6 0.0 Electricity 32.4 62.1 37.0 Transportation 26.5 27.3 14.8 Access to land 9.1 33.3 37.0 Tax rates 61.8 69.7 63.0 Tax administration 29.4 56.1 37.0 Customs and trade regulations 11.8 28.8 14.8 Labor regulations 11.8 18.2 11.1 Skills and educationof workers 11.8 28.8 37.0 Businesslicensingand operatingpermits 11.8 33.3 11.1 Access to financing 23.5 45.5 33.3 Cost of financing 61.8 47.0 55.6 Economicand regulatory policyuncertainty 32.4 40.9 22.2 Macroeconomicinstability 44.1 51.5 22.2 Corruption 50.0 48.5 44.4 Crime, theft, and disorder 47.1 28.8 14.8 Anticompetitive or informal practices 23.5 18.2 25.9 Source: World Bank 2005d. Note: Boldtypeface indicatesthe worst five obstacles in each country. targets higher-income tourists and avoids Ngorongoro Conservation Area, the Seren- mass tourism, which allows it to protect its geti National Park and the island o f Zanzi- fragile natural assets. In addition, Tanza- bar, Because o f saturation, growth in the nia's low-volume, high-yield tourism package holiday segment is limited. To in- strategy maintains its image o f exclusivity. crease growth in the tourism industry, the restructured TourismMaster Plan identifies Nevertheless, much unexploited potential investments in infiastructure, enhanced exists in new destinations and activities, in products, improved efficiency andcompeti- particular in niche markets. Moreover, the tiveness o f suppliers, and an improved ena- capacity o f tourist services in the southern bling environment (Ministryo fNaturalRe- part o f the country i s not fully used. Like- sources and Tourism 2002). The plan, wise, the potential that marine assets on the however, lacks a specific focus while offer- coast area offer has not yet been fully trans- ingseveralgrowth strategy options. formed into tourism products. Any further exploitation o f Tanzania's tourism potential The tourism industry has great potential to must ensure long-term sustainability, requir- increase its indirect contributions to eco- ingactivity bypublic andprivate actors. nomic growth and poverty reduction be- cause significant opportunities exist for Several structural constraints hamper the strengthening and increasing tourism realization o f the tourism industry's growth backward links inthe agriculture, manufac- potential. The tourism sector is highly con- turing, andservices sectors. Backwardlinks centrated. Few operators control demand can be increased both in volume and and volume to products such as the through inclusion o f additional industries. 146 Tourismin Tanzania Many of the tourism sector's suppliers are Management Area regulations to enable small, often informal operatorswith limited participation of local communities in con- capacity and limited access to capital and serving wildlife. Community-based tourism expertise. For example, in the existing and offers great potential to reduce rural pov- important h i t andvegetable link, suppliers erty. Wildlife-based tourism can be used as are mostly informal with little opportunity a revenue source for rural communities. to expandtheir business. Thus, the potential This form of tourism could be combined for increasing the value added o f the prod- with cultural tourism andsold as apackage. ucts by processing fruits and vegetables remains largely unrealized. Similarly, im- Investment in Supporting Infrastructure proved techniques, increased capacity, and better-trained employees would strengthen The tourism sector would benefit from in- sector-to-sector links. The Tanzania Diag- vestments made in transportation, telecom- nostic Trade Integration Study (World munications, electricity, andhealth services. Bank 20050 has identifiedanumber ofde- tailed and specific recommendations aimed Transportationinfrastructure.Tanzania's at increasing the economic contribution o f international accessby air i s inadequateand tourism by strengtheninglinks. expensive. Many visitors to the Northern Circuit fly to Nairobi, Kenya. KLMRoyal Furthermore, the potential exists to in- Dutch Airlines has had a direct flight from crease the number o f households receiving Amsterdam to Kilimanjaro for more than income fiom tourism. For example, in 30 years, andBritishAirways serves Dar es Talek, Kenya, near a gate to the Masaai Salaam. Inits current condition, the airport Mara Nature Reserve, 86.4 percent o f the on Zanzibar is inappropriate for landing households earn income from tourism, jumbo jets from Europe. The lengthening compared with 12 percent o f those near o f the runway i s under way; however, air- the Ngorongoro Conservation Area. port facilities and safety are inadequate. A public-private partnershipis recommended Recommendations to improve management and efficiency o f airport facilities and safety. Domestic Several recommendations can be made on transportation also needs great improve- the basis ofthese findings. ment. The road fiom Arusha to Lake Manyara andthe Serengeti National Park i s Development of Innovative Tourism dilapidated. Furthermore, to exploit the Packages and Niche Products tourism potential o f remote destinations, in particular inthe Southern Circuit, Tanzania Tanzania has not yet filly exploited its must improve roadand air transportation to potential in value-added niche products allow better access. such as adventure tourism, including climbing and trekking, deep-sea fishing, Telecommunicationsand electricity. The scuba diving, cultural tourism, bird watch- costs for telecommunication services and ing, andhunting. electricity are high in Tanzania. In general, only about 7 percent of Tanzania's popula- Currently, community awareness and par- tion has access to electricity, whereas the ticipation are almost nonexistent. The Tan- Dar es Salaam area consumes about half of zania Ministry o f Natural Resources and the country's electricity. The Tanzania Elec- Tourism has recently enacted Wildlife tric Supply Company's predominantly 147 Sectoral Perspectives on Growth Table 7.5. Electricity Provision Indicators Indicator Kenya Tanzania Uganda Outages (days) 82.90 91.85 - Duration (hours) 31.OO 4.55 - Firmswith generator (%) 88.89 74.24 60.71 Share of electricityfrom generator (%) 17.54 35.33 3.26 Firms with damaged equipment (%) 63.89 50.00 35.71 Value of damaged equipment (US$) 21,I14.04 8,565.51 7,274.12 hydroelectric system is prone to shortages ism industry, in particular for up-market caused by conditions such as poor rainfall. tourism. A better provision o f health clin- Moreover, the southern part o f Tanzania has ics and medical evacuation facilities i s no access to the national electricity grid. needed to meet the requirements o f the Most lodges in Tanzania's national parks tourism business (for example, decom- and game reserves rely on generators and pression chambers for diving). Tourists alternate solutions such as photovoltaic must protect themselves from malaria by cells. However, they are not likely to have using nets treated with insecticide and access to the grid inthe medium term. Even other prophylactic measures. where there i s access to electricity, the pro- vision is not reliable. Tanzania experiences Investment in Human Resources on average nearly 92 days o f outages, al- most 10 days per year more than Kenya. In Tanzania, no written policy and objec- However, the duration o f outages is shorter tives toward human resource development inTanzania (table 7.5). in tourism exist. Investment inhumanre- sources i s indispensable if the quality o f Health sewices. The availability o f heath tourism services and the professionalism services i s also very important to the tour- o f the industry i s to improve. Notes The Human Resources Index is proxied by using ' Average the EducationIndex obtained from the 2004 United annual growth rates o f receipts are cal- Nations Development Programme report, consist- culated by the World Tourism Organization. ing o f the adult literacy rate and the combined pri- The tourism Price Competitiveness Index is com- mary, secondary, and tertiary gross enrollment ra- puted using the Hotel Price Index and Purchasing tios. The Openness Index is an aggregate index Power Parity Index. The Infrastructure Index includingthe Visa Index, Tourism Openness Index, shows the level o f infrastructure development, Trade Openness Index, and Taxes on International combining the Road Index, the Sanitation Index, Trade Index. The Social Index i s a combination of and the Water Access Index. The Environment the HumanDevelopment Index, Newspaper Index, Index combines the Population Density Index, CO2 Personal Computer Index, and Television Index. Emission Index, and Environmental Treaties Index. 3See chapter 8 on the informal economy for a more The Technology Index unites the Internet Index, detailed discussion of the business environment for Telephone Index, Mobile Index, and HiTech Index. small andmediumenterprises inTanzania. 148 CHAPTER 8 The Informal Economy in Tanzania International estimates that include infor- economic activities in Tanzania were mal activities in agriculture suggest that within extralegal boundaries in the infor- Tanzania's informal economy accounts for mal economy and that 89 percent o f all about 60 percent o f the Tanzanian gross such activities are held extralegally. Ac- national income (Schneider 2004). The cording to ILD, the Tanzanian informal informal sector i s thus relatively large in economy has assetsworth US$29 billion. both regional and international comparison (figure 8.1). Data from the Tanzanian "In- Many small enterprises inTanzania operate tegrated Labour Force Survey, 2000/01" under a semiformal legal status without the (National Bureau o f Statistics 2001) that necessity o f registration with state authori- exclude informal sector activities in rural ties (table 8.1). Semiformal operators ap- agriculture suggest that the informal sector pear on a list of operators at the local au- employs about 16 percent o fthe total labor thorities, and they pay taxes that are force. The Instituto Libertady Democracia collected by local authorities. How long in- (ILD 2005) found that about 98 percent of formal operators may remain at any level o f 149 Sectoral Perspectives on Growth Figure 8.1. Size of the Informal Economyfor Selected Countries, as a Percentageof Gross National Income Source: Schneider2004. Table 8.1. Typology of Forms of Enterprisein Tanzania Formal status Legal form Description and subcategories lllicit None An enterprisefor which there is no legally permitted, licensed, or registeredcounterpart. Informal None An activity (for which there is a formal counterpart) that does not comply with requirementsof the regulatory sys- tem regarding licenses, permits, certificates, notification, or registrationof the activity. Semiformal Local authority- An activity carried on by an operator who appears on a licensed enterprise local authority list of licensed operators of enterprises but who is not registeredwith the state registrar. Existing forms of licenses include those for (a) hawkers, (b) busi- nesses, and (c) market stalls. Formal State-registered Persons registeredto conduct business activities under a enterprise registered business name. For example, an activity enu- merated on a state registeras a sole ownership enter- prise, a private limited company, or a joint stock company open to public subscriptionshareholdings. Source: Nelson and de Burijn 2005. informal or formal status differs widely in the National Bureau o f Statistics (2001), Tanzania. An operator mayprogressthrough people whose main economic activity i s in a semiformal stage or move directly fiom the informal economy are most often em- informal status to the official registration. ployed inthese sectors: retail trade o f agri- cultural products, meat, and chicken (20.7 Informal economy operations can be found percent); stationery, photography, and inmost sectors in Tanzania. According to general retail (18.8 percent); retail trade o f 150 The Informal Economy in Tanzania processed food (10.5 percent); and restau- The National Bureau o f Statistics (2001, rants and hotels (12.4 percent). The major- 58) suggeststhat the growth o f the informal ity of people whose informal activity is a economy during the past decade "is possi- secondary activity are employed in crop bly a result o f economic hardships house- growing (94.4 percent).' holds have been facing that have forced them to join the sector as a survival strat- Generally, for households engaged in in- egy." Accordingly, 44.5 percent o f people formal economic activities in urban areas, whose main activity is in the informal such activity tends to be their main activity, economy stated that the major reason for whereas such activity is more likely to be involvement inthat sector i s the inability to secondary in rural areas. According to the findwork. Forthose for whom the informal National Bureau o f Statistics (2001), one in economy i s a source o f secondary activity, three households was active inthe informal the main reason i s the need for additional economy in 2000/01, as opposed to one in income for families (43.9 percent). Such four households in 1990/91. The survey data indicate that survivalists constitute a shows that the number o f households with considerable part o fthe Tanzanian informal informal economy activities grew during economy. A study by the International La- the 1990s fiom 42 percent o f the total bour Organization (ILO), the United Na- households inurban areas to 61 percent. In tions Industrial Development Organization rural areas, 27 percent out o f the total (UNIDO), and the United Nations Devel- households had informal economy activi- opment Programme (UNDP) therefore dif- ties in 2000/01 as compared with 21 per- ferentiates between survival types o f opera- cent in 1990/91. Also, interms o f employ- tors and growth-potential types o f ment, the informal economy holds a bigger businesses inthe Tanzanian informal econ- share in urban areas than in rural areas. In omy (ILO, UNIDO, andUNDP 2002). Dar es Salaam, for instance, 36 percent o f the total labor force is employed inthe in- The National Bureau o f Statistics (2001) formal economy. further notes that informal economy ac- tivities as main employment are more The vast majority o f persons in the infor- concentrated in urban areas because o f mal economy are self-employed without problems o f unemployment, whereas in employees. Eighty-two percent2 are self- rural areas the informal economy pre- employed in the main activities, while 88 dominantly provides opportunities for percent are self-employed inthe secondary secondary activities. As stated before, activities. crop growing i s the sector employing the majority o f the people whose informal The level o f education o f informal busi- activity i s a secondary activity. The Na- ness operators i s generally low. Accord- tional Bureau o f Statistics' (2002) ing to the National Bureau o f Statistics "Household Budget Survey 2000/01" (2001), 64.2 percent o f operators in the finds that 75.8 percent of the people in main activities and 53 percent o f those in rural areas are mainly active in farming, the secondary activities have completed raising o f livestock, or fishing. In rural primary school. In the main and secon- areas, agricultural income accounts for dary activities, 16.7 percent and 43.5 per- the primary source o f income with a share cent, respectively, o f the operators have o f 60.4 percent, which means that almost no education or have dropped out o f pri- 40 percent o f the income i s derived from mary school. sources outside farm production. Sixty- 151 Sectoral Perspectives on Growth five percent o f rural households report es Salaam, the growth o f expenditures by more than three income sources. households headed by a self-employed person was even more dramatic at 65 per- Ingeneral, income tends to bemuch lower cent, which was evenhigher than the aver- in the informal sector than in the formal age growth in expenditures o f 60 percent sector. Tanzania i s no exception. Accord- experienced by households headed by a ing to the National Bureau o f Statistics person in paid employment. (see table (2001), the average income for paid em- 2.16). ployment o f households in Dar es Salaam that do not undertake informal activity i s T Sh 191,662, while it i s T Sh 85,960 for Constraints to Growth of households with informal activity. Income Enterprises in the Informal in rural areas is much lower, averaging Sector and Formalization T Sh 76,800 for households not involved in activities in the informal economy and Micro and small enterprises in Tanzania, T Sh 47,874 for households with informal most o f them in the informal sector, are activity. not only an important means to generate income; they are also an important entry Nonetheless, the informal economy thrives point for the development o f a strong pri- inTanzania becauseit provides opportuni- vate sector in Tanzania. As such, it i s im- ties for income generation to the poor and portant to consider how growth o f such unemployed and because it offers a low- enterprises and their transition into the cost ground for experimentation with formal sector canbe facilitated. business ideas that might lead to growth and formal enterprises. Forced formaliza- Informal operators state the existence o f a tion risks incurring the cost o f damaging number o f constraints to growth o f their fragile enterprises and livelihoods for very businesses. Most importantly, they are of- little benefits and suppressing business ten not in a position to afford permanent experimentation and development. The premises for their businesses. Second, they decision o f a small-scale informal operator lack access to credit. A lack o f business to formalize should be a voluntary one. management skills and very limited access to new technology also are detrimental to Our poverty analysis presented in chapter the growth o f informal businesses. Often, 2, as well as other research such as Owens informal entrepreneurs face harassment by and Teal (2005), suggests that the informal local authorities. That situation has im- sector-especially urban self employ- proved considerably since 2003, but it i s ment-has been an important path out o f still aproblem. Somebusinessesare demol- poverty for many Tanzanians. Not only ished, property is taken away, and in the did the share of self-employed persons in worst cases, entrepreneurs face charges. the adult population increase from 4.8 per- cent to 8 percentY3but expenditures by Furthermore, advocacy i s needed for the households headed by a self-employed in- informal economy. Currently, three asso- dividuals grew by 18 percent between ciations represent the interests o f informal 1991/92 and 2000/01, compared with a economy operators: the Small Industries much more modest growth of expenditure and Petty Traders Association (Vikundi o f agricultural households, which grew by vya Biashara Ndogondogo, or VIBINDO), only 7 percent during that period. In Dar an umbrella organization; the Tanzania 152 The Informal Economy in Tanzania Small Industrialists Society (TASISO); whereas the respective figures for Dar es and the Tanzania Food Processors Asso- Salaam are 95 steps, 283 days, and US$ ciation (TAFOPA). However, sectoral as- 3,816 (ILD 2005). ILO, UNIDO, and sociations are still weak. The I L O Syndi- UNDP (2002) calculated that the cost o f coop project facilitated the formation o f a coping with regulatory and nonregulatory national steering committee that includes constraints would amount to as much as 75 the Trade UnionCongress o f Tanzania, the percent o f monthly sales o f informal op- Tanzania Federation o f Cooperatives, the erators in the firewood and charcoal sec- Savings and Credit Cooperative Union, the tor, which explains the decision to remain government, and individual informal informal. economy groups. To register a business, entrepreneurs must VIBINDO states that the biggest constraint travel to Dar es Salaam, where the Busi- to formalizing a business i s access to af- ness Registration and Licensing Agency is fordable permanent premises. There are a located. ILO, UNIDO, and UNDP (2002) number of success stories o f informal op- show that business licensing i s equally erators formalizing their businesses after cumbersome, because it requires various the municipal authorities provided or as- procedures in different offices at regional sisted inthe access to permanentpremises. and district levels. Both processes, licens- ingandregistration, are requiredand often The costs o f startingand operating a formal involve bribingofficials, further increasing business are highinTanzania. According to the cost o f starting a formal business. the World Bank's Doing Business 2007 Costs o f formalization are often increased (World Bank 2006), entrepreneurs in Tan- even more because some officers are re- zania can expect to go through 13 steps to ported to be unhelpful, obstructive, and launch a business, which on average take uncaring (ILO, UNIDO, and UNDP 2002). 30 days and cost 92 percent o f the US$340 per capita income, compared with 11 per- D e Soto (2001) explains that a major prob- cent, 46 percent, and 114 percent o f per lem faced by the informal economy is its capita income in Botswana, Kenya, and inability to convert what he refers to as Uganda, respectively, or 6.5 percent in Or- dead capital (untitled assets) into capital ganisation for Economic Co-operation and that can be used, for example, as collateral Development (OECD) countries. However, for loans. Registering property i s an ex- the minimum capital required to obtain a pensive and time-consuming undertaking business registration number in Tanzania i s inTanzania: it takes 12 steps and 61 days 6 percent o f per capita income, which i s compared with 4 procedures and 33 days considerably lower than the regional aver- inOECD countries. The cost o fregistering age (297.2 percent) or the,OECD average property amounts to 12.2 percent o f the (28.9 percent). property value, which i s considerably higher than in OECD countries (4.7 per- The ILD(Instituto Libertady Democracia) cent). ILD (2005) found that the procedure found that the costs and burdens o f the to allocate land for urban purposes and to procedure to legally incorporate a private obtain a building permit takes 13 stages Bureau o f Change in Mbeya include 10 and includes 68 steps, which take eight stages, 103 steps, 379 days, and US$5,506, years to complete and cost US$2,252. 153 Sectoral Perspectives on Growth Figure 8.2. Estimated Unreported Revenuefor Tax Purposes 40 33 30 20 10 I 0 Kenya Tanzania Zambia Uganda Mozambique Source: World Bank 2005d. Also, enforcing formal contracts costs on The marginal effective tax rate i s higher average 5.3 percent o f the debt, compared for small businesses than for large firms. with the OECD average o f 10.9 percent. Local governments still levy a large num- According to ILD (2005), claims for a debt ber of taxes, fees, and charges. Local taxa- at Tanzania's Commercial Court Division tion is seen as a major constraint to for- take 9 stages, 96 steps, and 390 days and malization. In addition, bribes are cost US$11,964. Likewise, collection o f a estimated to add up to more than 5 percent debt by executing a court decree takes o f total sales. Comparative investment 1,286 days and costs US$1,022. In addi- climate survey datasets collected by the tion, hiring or firing employees is compara- World Bank (2004e) suggest, however, tively difficult in Tanzania. However, be- that the share o f revenue that Tanzanian cause employment regulations are not businesses deliberately fail to report for enforced, ILO, UNIDO, and UNDP (2002) tax purposes i s more than 30 percent, didnot identifythemas a severeconstraint. compared with around or less than 20 per- cent in Kenya, Uganda, and Zambia (fig- Tax rates, which are perceived as too high, ure 8.2). However, one should note that often act as a disincentive to formalize. A informal entrepreneurs do pay taxes to lo- recent study shows a number o f anomalies cal authorities. For example, inDar es Sa- with the tax regime for small businesses, laam, a local tax o f T Sh 100 per day i s in particular that the system is regressive collected from every informal operator. for nonrecordkeepers (FIAS 2006). More- over, the study states that the "tax envi- Insum, another reason for the large size of ronment encourages expansion o f the in- Tanzania's informal economy i s that for formal economy," because "small micro enterprises the benefits o f formality businesses face a proportionately higher are dwarfed by its costs. "Whether in time and financial costs to comply with monetary terms (direct cost or income administrative requirements and therefore foregone), or in terms o f time and energy, may not see any benefit o fjoining the tax the cost o f compliance turns out to be too net" (DFID and FIAS 2006, 18). Further- highfor most starting businesses, who are more, newly registered businesses must therefore obliged to start informally" pay taxes up front, which further increases (ILO, UNIDO, and UNDP 2002, 3). In the requiredstarting capital. other words, the state fails to provide an 154 The Informal Economy in Tanzania Figure 8.3. MedianValue Added per Worker (a) Medianvalue added in selected countries (b) Medianvalue added by size of firm MedianValue Addedpr worker inUs dollars Medianvalue addedperworker bysize (inUS 4,397 dollars) $4,500 $4,500 $4,011 $3,000 $3,000 $1,500 $1,500 $0 Tanzania Uganda Zamia Kenya China $0 Mcro (1-9) SmIl(10-49) Medium(50- Large (loot) Source: World Bank 2005d. 1 institutional environment that is conducive diaries for access to credit, the public sec- to investment in the formalization o f in- tor for the provision o f public services, formal enterprises. Therefore, the "current and clients and suppliers for contract- regulatory set up (a) fails to meet [the] ob- based transactions. Given the rigidities, jective o f ensuring quality control for the costs, and attitudes o f formal sector regu- majority; and (b) traps the entrepreneurs in lators and service providers, the access o f low quality settings, puts their upgrading informal entrepreneurs to important public and growth too far out of reach, and limits goods, such as electric and water utilities, the contribution o f the subsector to pov- andto other inputsand services is limited. erty reduction and national growth" (ILO, UNIDO, andUNDP 2002,35). Furthermore, operators in the informal economy cannot convert dead capital in Benefits of Increasing the form o f untitledassets into productive Formalization economic currency, such as collateral for loans to start or expand a business. Ac- The most obvious benefit to the govern- cording to ILD (2005), there i s ment of increasing formalization is higher US$29 billion in dead capital inTanzania. tax revenue. Furthermore, the govern- Moreover, operators' access to the capital ment's ability to implement policies and market i s very limited. Formal registration the effectiveness o f government programs i s generally a prerequisite for access to aimed at the private sector will rise be- credit and small business loans cause informal enterprises operate outside (US$lS,OOO to US$30,000). Therefore, o fthe government system o fregulation. informal sector operators are destined to remain small without being able to exploit From the perspective o f business opera- potential economies o f scale. As a result, tors, formalization o f their business in- productivity tends to be low. In general, creases the trustworthiness for customers, the labor productivity is lower inTanzania a benefit that has been mentioned by sev- than inmany other African countries-the eral entrepreneurs who formalized their median value added per worker is business. Furthermore, formalization cre- US$2,061, as shown in panel (a) o f figure ates the basis for formal transactions with 8.3. Panel (b) o f figure 8.3 indicates that, other entities, including financial interme- in the case o f micro enterprises (most o f 155 Sectoral Perspectiveson Growth Box 8.1. Examplesof Voluntary Formalization Alan Mungo experimented with operating an informal insurance agency and a clothes shop. He eventually progressedto producingwine and operating a safari tour company. Having developed a marketable product and identifiedmarket opportunities for expansion, he decided it would be in his interest to legitimize his enterprise. However, after preliminary inquiries, he found that the premises he used, an automobile garage inthe grounds o fhis dwelling house, would not be ap- provedby health inspectors, which meant he would not be able to apply for an operator's license. Around this time, he attended a meeting o frepresentatives o fthe Tanzanian FoodProcessorsAsso- ciation, the localtrade licensing officer, and arepresentative o fthe Ministryfor Agriculture, Food Security, and Cooperatives. That meetingproducedan agreementthat small-scale food processors who couldnot comply withregulations would not be forcedto cease their operations. Insteadthey could continue to operatewithout licenses,but they would still be subject to supervision byhealth inspectors. That condition incidentally impliedthe advantageo fbeingoffthe databaseo f operators liable for formal taxes. The acknowledged reason for the ministry'sdecision was that Tanzania's food-processing industrywas important; that it relied mainly on small-scale producers, most o f whom were not licensed; and that it would be forcedto close ifregulations were enforced. Mungo was able to continue operating, andlicensing costs were deferreduntilafter hispremises reached the standardabout one year later. At that point, his formalization costs were very low. Dan Himba formalized hisbusiness after an activity period o f 29 years. H e conductedinformal enterprise activities as a supplement to his salaried employment. Duringthat period, he enjoyed a long runo f starting informal enterprises, maximizing profits while he could, and abandoning those enterprises when they showed signs o f havingruntheir most profitable life cycle. When the conditions o fhis employment deteriorated, a full-timeentrepreneurial career became more attrac- tive to himand increased the stimulusto formalize. H e subsequently abandoned hisjob for a ca- reer inprivate enterprise. Source: Nelsonand de Burijn2005. which are informal), labor productivity i s the Tanzanian government should tolerate evenlower. and support informal economic activities. Governments can reduce informality by The decision to formalize an informal reducing the costs and increasing the bene- business depends on expected benefits fits associated with becoming formal. By from formalization (box 8.1). Often, in- reducing corruption, for example, they can formal businesses are too small to be able increase enterprises' willingness to deal to pay the costs o f formalization. with public institutions. In addition, gov- ernments can encourage firms to become Implicationsfor Policy formal by reducing the burdenimposed on formal enterprises by such factors as bani- To allow for a low-cost arena for business ers to entry, business regulations and in- experimentation, as well as a means o f in- spections, and labor regulations. come generation for poor and unemployed people in the absence o f a state-provided Inaddition, providing for anadequateinsti- safety net such as unemployment benefits, tutional framework that i s conducive to and 156 The Informal Economy in Tanzania provides incentives for voluntary formal- banks. In the long run, SACCOs could ization o f informal businesses i s important. function as a way to connect informal sav- Possible incentives are the abolishment o f ings with the formal financial sector. upfront payment o ftaxes or free training on procedures o f formalization. Alternatively, The government has already startedto im- an agency to assist entrepreneurs in regis- plement measures to reduce the cost o f tration procedures could be set up. More- doing business and to facilitate the formal- over, local governments should provide ization o f businesses and property. In Oc- more permanent premises at low rents tober 2004, it launched the Property and (such as market stalls in Ilala, Dar es Sa- Business Formalisation Programme. The laam) for informal entrepreneurs. The objective o f the program i s to identify as- Small and Medium Enterprise Develop- sets and guarantee property rights. In fall ment Policy of 2002 addresses the problem 2005, at the conclusion o f the diagnostic o f infrastructure requirements and calls on phase o f the program, ILD (2005) pre- local authorities to allocate and develop sented a comprehensive report that com- land for small and medium enterprises, to mented on legislation governing property develop industrial clusters and trade cen- ownership and compiled data on the pat- ters, and to identify and allocate underused terns o f movable- and fixed-asset owner- public buildings to such enterprises. A ship. Formalization reforms will be de- higher degree o f realization o f this policy signed on the basis o f those findings. would be very beneficial. The government has also taken important The focus of government has beenprimar- measures to reduce the burdeno f business ily on encouraging formalization of busi- licensing and registration, and a business nesses, while paying relatively less atten- activities registration bill was submittedto tion to supporting small-scale, informal the Tanzanian parliament in 2005. The sector activities. In addition to the review goal i s to create a business licensing sys- of laws and regulations that may unduly tem that is transparent and efficient, aim- impinge on small-scale informal activities, ing at registration rather than revenue col- that effort also requires a change in mind- lection and control in place. Furthermore, set o f public officials to recognize the the license fee for small businesses was value and importance o f informal sector abolished in 2004/05. By simplifying and activities for Tanzania's economic devel- harmonizing legislation and streamlining opment. In particular, local authorities, regulations o f business and property regis- who deal with the informal sector on a tration, the government should cut all costs daily basis, would benefit from capacity o f formalization in order to create an ena- buildinginthat area. bling environment. For example, the Busi- ness Registration and Licensing Agency Access to credit poses severe constraints should open branches throughout the to informal entrepreneurs. Often, informal country so that travel expenses o f entre- businesses find even microfinance preneurs who wish to license a business schemes hard to comply with. Alternative are kept at a minimum. credit schemes include rotating savings and credit societies, such as savings and Officers o f local and national government credit cooperatives (SACCOs). Develop- agencies should be well informed and ing and improving occupational SACCOs trained to assist potential licensed business can be an effective alternative to formal owners. They should pass on correct and 157 Secforal Perspectives on Growth u s e h l information in order to facilitate formalization process, as well as the business licensing. In particular, officers benefits and opportunities o f running a should be aware o f legislative changes formal business. Informal workers should such as the new taxation schedule and be encouraged to form cooperatives. comply with them. Training for informal business operators inmanagerial skills through seminars and Inadditionto government officers, poten- workshops i s o f the utmost importance tial and current informal entrepreneurs and i s very beneficial to fostering private should have a fair understanding o f the sector development. Notes ' Those figures are calculated on the basis o f the national definition o f employment. * The figure o f 82 percent was reached according to the standard International Labour Organisation definition o f labor force, or 73 percent using the national definition. Data from the Integrated Labour Force Survey (National Bureau o f Statistics 2001) and the Household Budget Survey (National Bureau o f Statistics 2002) provide different measures o f in- formal sector activities, but the magnitude and trends provided by these two surveys present broadly similar pictures. Regulatory constraints include complicated, lengthy, and unpredictable procedures; inadequate institutional arrangements; rent-seeking civil ser- vants; unreasonable specifications and standards; and a multiplicity o f taxes and levies. Nonregula- tory constraints include, for example, poor clients, lack o f access to financing, poor infrastructure, and unfair competition. The 75 percent figure is the highest average cost for coping with constraints that occurred in the firewood and charcoal sector. However, inthe cloth-making sector, for example, the highest average cost calculated was 5.8 percent o fmonthly sales. 158 PART 3 Elements of a Strategy for Shared Growth CHAPTER 9 Fostering Innovation, Productivity, and Technological Change Drivers of Innovation, Productivity, and Technological Change The application o f knowledge, as mani- change and improving productivity across fested inareas such as entrepreneurshipand all sectors o f the economy. Innovation in innovation, research and development Tanzania concerns not just the domestic (R&D), and people's education and skill development o f frontier-based knowledge; levels, i s now recognized as one o f the key more importantly, it relates to the applica- sources o f growth and competitiveness in tion and use o f existing knowledge to the the global economy. Developing countries localcontext. Itrequires a climate favorable such as Tanzania have ample scope to use to entrepreneurs that i s free from bureau- new and existing knowledge and innova- cratic, regulatory, and other obstacles and tion to develop new products and proc- that fosters interactions between the local esses, thereby enhancing technological and outside business worlds, with different 161 Elements of a Strategy for Shared Growth sources o f knowledge, including universi- Effective use o f knowledge in any country ties, public laboratories, users, think tanks, requires appropriate policies, institutions, industries, and indigenous communities. investments, and coordination across these four pillars. A country's economic and in- Four pillars are generally considered to be centive regime i s critical because it de- important for countries to make effective scribes the framework within which a so- use o f knowledge for their overall eco- ciety and an economy work-in other nomic and social development: words, the rules o f the game, both formal and informal. And the basic enabler for An economic and institutional regime any country to use knowledge effectively that provides incentives `'for the eff- i s education-encouraging learning and cient use o f existing knowledge, the the exploration o f new knowledge, with creation o f new knowledge, and the innovation being a driver o f technological flourishing o f entrepreneurship change and information and communica- An educated and skilled population tion technologies (ICTs) providing the that can create, share, and use knowl- mechanisms to reduce transaction costs. edge well These three functional pillars-education, A dynamic information infrastructure innovation, and ICTs-are the focus o f that can facilitate the effective com- this section. Box 9.1 provides a snapshot munication, dissemination, and proc- o f how Tanzania fares interms o f its com- essing o f information petitiveness on a global scale. An efficient innovation system o f Education firms, science and research centers, universities, think tanks, consultants, Well-educated and skilled people are key and other organizations that can tap for creating, sharing, disseminating, and into the growing stock of global usingknowledge effectively to spur growth knowledge, assimilate and adapt it to and innovation. Ideally, countries need to local needs, andcreate new knowledge develop flexible education systems, starting Box 9.1. BenchmarkingTanzania in the Global Context Some recent indexes that have beendeveloped to benchmark countries' performance interms o f competitiveness or knowledge readiness on a global basis include the following: WorldEconomic Forum's Global Competitiveness Report 2006-2007. This report (World Eco- nomic Forum 2007) highlights World Economic Forum's new Global Competitiveness Index (GCI), which provides an overview o f factors critical for drivingproductivityandcompetitive- ness. The factors are grouped into ninepillars: institutions, infrastructure, macroeconomy, health andprimary education, higher education andtraining, market efficiency, technological readiness, business sophistication, and innovation. Tanzania is ranked 104th(of 125 countries) on this new 2006 GCI, behindKenya (94) but aheado fUganda (113). World Economic Fo- rum's Business Competitiveness Index (BCI) is another index that focuses onmicroeconomic factors that determine economies' current productivity and competitiveness. Tanzania is ranked 73rd (of 121countries) on the B C Iin2006, aheado fUganda (88) but again behindKenya (68). continued 162 Fostering Innovation, Productivity, and TechnologicalChange Box 9.I (continued) WorldEconomic Forum s Africa Competitiveness Report 2004. This report (World Economic Forum 2004) highlightsthe prospects for growth and the obstacles to improving competitive- ness in25 African economies. Tanzania ranks 9th o f 25 countries on the overall global com- petitiveindex, surpassing Uganda, which i s ranked 14th, and Kenya, which is placed 15th. WorldBank Institute's Knowledge AssessmentMethodology. The figure below compares Tanzania's performance on the Knowledge Economy Index (KEI)with the performance o f the African region, Tanzania's neighbors, andBotswana and SouthAfrica, as well as with that o f well-performing East Asian economies such as Malaysiaand Thailand. It shows that between 1995 and the most current period (2004-05) Tanzania has made a substantial im- provement inits overall knowledge readiness, as evidencedby positive changes inthe KEI, particularlyfor the economic and incentive regime and the innovationpillars. Inaddition, it has made some improvement inICTs. Also, Ugandahas made strides inimproving its eco- nomic incentive regime and inthe ICT pillars, and Kenyahas strengthened its information in- frastructure over the past decade or so. Kenya's performance surpassesthat o f the African re- gion, whereas the performance o f Tanzania and Uganda does not. Botswana and Malaysia have slightly improvedtheir recent performance over that in 1995, but South Africa and Thai- landhave not. Thus, this relative comparison shows that eventhough a country can make progress, it can still fall relatively behindbecausethe world as a whole may have made much more significant improvement inthe variables that are used to track knowledge- and innova- tion-related performance. Tanzania and Comparators, 1995 and Most Recent Period Malaysia 1995 South Africa 1995 Thailand 1995 Botswana 1995 Kenya 1995 Africa 1995 Uganda 1995 Tanzania 1995 0.0 2.0 4.0 6:0 8.0 11 10Economic Regime 0 Innwation 0 Education ICT Source: World Bank's Knowledge Assessment Methodology, http://www.worldbank.org/kam. Note: The two bars representthe aggregate KEI score for a selectedcountryfor the most recent year for which data are available and for 1995,split into four pillars: economic incentive regime (in light blue), innovation(in yellow), education (in dark blue), and ICT (in green). Each color band representsthe contribution of a particular pillar to a country's overall knowledgereadiness. 163 Elements of a Strategy for Shared Growfh Figure 9.1, Adult Literacy Rates, 1970-2002 Adult Literacy Rates Source: World Bank internal database. with basic education, which provides the higher than that o f Uganda (69 percent) foundation for learning; moving next to but lower than that of Botswana (79 per- secondary and tertiary education, which cent), Kenya (84 percent), and South Af- can develop core skills, including techni- rica (86 percent). Inaddition, according to cal ones that encourage creative and criti- figure 9.2, Tanzania's average number of cal thinking for problem solving and inno- years o f schooling in 2000 (3.4) was vation; and moving finally to a system of higher than in Uganda (3.22), lower than lifelong learning. Developing countries inKenya(5.08), and far below the average such as Tanzania face many challenges in inSouth Africa (7.22). developing such systems. They include expanding coverage to achieve universal The recent focus on investment inprimary access to basic and secondary education; and secondary education, if sustained, providing tertiary education, which i s gen- promises accelerated increases in literacy erally weak; improving the links between and average years of schooling in the me- formal and informal education systems dium to long term. In recent years, the and the labor market; and raising the over- Tanzanian government has recognized the all quality o f learning. need to raise educational levels in the population as a necessary condition for Tanzania's economy today i s largely mar- enhancing economic growth. The general ket oriented and has in place many ele- education system in Tanzania includes ments required for private sector-led seven years o f primary education, four growth. However, it does not have the years o f lower secondary education, and sound base o f an adequately qualified and two years o f upper secondary education. trained workforce, which i s essential for Appropriate programs for primary and rapid economic growth and effective di- secondary education have been put in versification o f its production and export place to enhance access to and increase the bases. Figure 9.1 shows that in 2001 Tan- quality o f education. Key measures so far zania's adult literacy rate (77 percent) was have included the abolition o f primary 164 Fostering Innovation, Productivity, and Technological Change Figure 9.2. Average Years of Schooling, 1960-2000 Average Years of Schooling (Cohen-Soto) I Source: Cohen and Soto 2001, school fees in 2001, significant increases Bank 200%). The PEDP has strengthened in budgetary fbnding for primary educa- institutional capacity and management o f tion, and the implementation o f the Pri- education, as measured by the enhanced mary Education Development Program capacity in the Ministry o f Education and (PEDP). Under this program, Tanzania's Culture to provide policy and guidelines gross enrollment ratios for primary educa- and monitor education delivery. It has led tion increased from 78 percent in 2000 to to decentralization and delivery o f primary 106 percent in 2004. The net enrollment education through regional administration ratio increased from 59 percent in 2000 to offices, district wards, and schools and has 91 percent in2004. Girls represent 49 per- strengthened community participation and cent o fthe total. school-level management and accountabil- ity. Resource availability and use of re- A key challenge for government is to fo- sources have also improved, as measured cus more on improving the quality o f pri- by increased nonsalary expenditures inthe mary education. In terms o f inputs, the primary school budget. availability o f textbooks has also im- proved. On average, before the PEDP was With respect to secondary education, Tan- launched, eight students shared one book zania has one o f the lowest net enrollment on each subject. In2003, the book-to-pupil ratios in Sub-SaharanAfrica. Only about 9 ratio improved to 1:4. The government's percent o f the relevant age group attends target was to reach a ratio o f 1:l by 2006. secondary education, compared with an Teachers' knowledge and mastery o f the average o f 27 percent for Sub-SaharanAf- curriculum have also improved through rica in2000, including about 11percent in preservice and in-service teacher training lower secondary school and less than 2 interventions. The proportion o f grade A percent in upper secondary school. Only teachers (those with secondary-level edu- 22 percent o f primary school leavers in cation) increased from 46 percent in 1999 Tanzania have a chance to continue their to 58 percent in 2004. However, more education at the secondary level, com- qualified teachers are still needed (World pared with 50 percent in Uganda in 2001. 165 Elements of a Strategy for Shared Growth Secondary enrollment ratios are low for all The results o f the Southern and Eastern population groups, but especially for low- Africa Consortium for Monitoring Educa- income youth and students in rural areas. tional Quality (SACMEQ) I1(2000-03)' Few government schools have been estab- for primary schools show high reading lished, and inadequate incentives exist to and math scores for Tanzania's mainland provide nongovernment schools in rural compared with other countries. In read- communities, because households are un- ing, Tanzania's mainland placed third, able to pay the fees required(World Bank behind Seychelles and Kenya, while on 2004d). math scores, it i s fifth, behind Mauritius, Kenya, Seychelles, and Mozambique Three main challenges face Tanzania in (figure 9.3). secondary education: increasing access, raising quality, and reducing costs. To sup- In tertiary education, Tanzania's per- port reforms in secondary education, the formance i s very weak: tertiary gross en- government has launched the Secondary rollment ratios stood at 0.94 percent in Education Development Program (SEDP), 2002, compared with 3.24 percent for which has among its aims increasing the Uganda and 3.52 percent for Kenya in proportion o f the relevant age group com- 2001 and 4.69 percent for Botswana and pleting lower and upper secondary educa- 15-05percent for South Africa in 2002.* tion, expanding enrollment with equity, Inthe 2000/01 academic year, there were improving the learning outcomes o f stu- 6,117 students at the University o f Dar es dents (especially among girls), and ena- Salaam and 13,442 in total in the coun- bling the public administration to manage try's three universities (University o f Dar secondary education more effectively. To es Salaam, Sokoine University o f Agri- expand enrollment with equity, the SEDP culture, and the Tanzania Open Univer- includes measures to make more efficient sity). In April 2001, an education fund use o f resources, provide development was established to sponsor children from grants to schools and communities (mainly very poor families to complete higher those inunderservedareas), expand teacher education. But in the past 10 years, a supply, lower household costs for secon- number of private universities have also dary education, expand the scholarship emerged, and today the country has nine program for students fiom poor families, private universities, mostly small and de- and enhance the partnership with the non- nominational, which award diplomas in governmental sector. The program for qual- financial and business management, wild- ity improvement includes curricula and ex- life management, community develop- amination reforms, provision o f textbooks ment, social welfare and cooperatives, and teaching materials through capitation and transport and media operations grants to schools, and quality improve- (ESRF2002). ments in preservice teacher training to- gether with establishment of a system for Soderbom and others (2004) examined the professional in-service teacher develop- returns to education in the manufacturing ment. The program also includes institu- sector in Tanzania. Their findings suggest tional reforms and capacity building at the that the marginal returns to education for central, region, district, and school levels primary and secondary education are for more efficient operation o f the secon- rather limited.However, the data suggest a dary education system. sharp increase in the returns for people 166 Fostering Innovation, Productivity, and Technological Change Figure 9.3. Reading Scores and Mathematics Scores (a) Reading SEYCHELLES KENYA TANZANLA MALRlTlW SWAZILAND BOTSWANA MOZAMBDUE SOUTHAFRICA WANDA ZANZIBAR LESOTHO NAMlBlA ZAM0lA M A M W 0 100 200 300 400 500 500 700 I (b) Mathematics SEYCHELLES MOZAMBIQUE TANZANLA SWAZILAND BOTSWANA WANDA SOUTHAFRICA ZANZIBAR LESOTHO ZAMBlA MALAM NAMIBY 0 100 200 300 400 500 600 700 Source: Southern and EasternAfrica Consortiumfor MonitoringEducationalQuality, http://www.sacmeq.org/. who have a tertiary education (figure 9.4). were female and 23 percent resided inru- One o fthe implications o fthese findings is ral areas. The average hourly wage for that the effect on growth in the manufac- these wage earners amounted to approxi- turing sector o f more primary and secon- mately T Sh400, but it ranged from dary education is likely to be small, but roughly T Sh200 for those with primary greater investment in tertiary education education to more than T Sh700 for those has higher payoffs (Soderbom and others with secondary education. Disparities be- 2004). tween rural andurbanearners and between genders were sizable (World Bank2004d). The analysis o f the 2000/01 Integrated La- bor Force Survey in Tanzania shows the The difference in the profile o f marginal value o f secondary education to the indi- social returns to education for workers in vidual Tanzanian (figure 9.5). In a sample the manufacturing sector and for workers o f more than 3,000 wage earners between inthe overall labor force may suggest that the ages o f 18 and 65, more than half had at the tertiary level certain degree pro- completed primary education, a quarter grams are well rewarded by the manufac- had completed lower secondary (form IV), turing sector but that many other degree and 5 percent had completed upper secon- programs result in lower-paying jobs. The dary (form VI). O f the sample, 33 percent implication would be to shift the supply o f 167 Elements of a Sfrategy for Shared Growth Figure 9.4. Predicted Earnings in ManufacturingSector Based on Manufacturing Firm Surveys (a) Old cohort i) Cohort Old 1 -A- %Year3 --+-Year Year 2 Year 4 3 d 1 4 Q T 3 ed " lh 44 (b) Young cohort -+- Year d)YoungCohcat Year 2 -E- Year 31 -+- Year 4 Source: Soderbom and others 2004. 168 Fostering Innovation, Productivity, and TechnologicalChange Figure 9.5. Marginal Social Returns per Year of Education Based on Integrated Labor Force Survey 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% Primary Low Secondary Up Secondary univ Source: Authors' calculation based on World Bank 2004d. higher education to those programs that TVET system had failed to produce seem to be in demand by the manufactur- graduates who were suited for the labor ingsector. The difference inthe profiles of market, the government introduced policy social returns to education also suggests changes in 1996 that emphasize its contin- that, for the manufacturing sector, a lim- ued responsibility for the provision and ited supply o f workers with relevant terti- financing o f more and better basic educa- ary education i s a constraint, while for tion, coupled with a reduction in untar- other sectors the limited supply o f workers geted subsidies through increased cost with secondary education may be more of sharing, liberalization o f private education a constraint. and training at all levels, and decentraliza- tion o f authority. The Vocational Training The government has been the major finan- and Education Authority set up in 1994, is cier o f technical and vocational education working to ensure that the training that i s and training (TVET), with assistance from provided is responsive to the labor market donors. But the TVET system faces sev- (Gill andDar 1998). eral problems interms of inefficient use o f resources, inequitable distribution o f edu- Africa i s a capital-scarce region, and the cational opportunities, poor links to the loss o f this limitedresource i s widely con- labor market, and lack o f coordination be- sidereddetrimental to the prospects o f sus- tween donors and the government. The tained growth and development. There is a unsustainable costs o f training appear to significant parallel to this problem on the be causednot only by low use o f capacity, side o f human capital. Weakness inhuman but also by low student-to-faculty ratios, capital-and particularly skill defi- whereas the inequitable distribution o f ciency-is a drag on investment and education opportunities i s biased toward growth in Africa. Progress in overcoming primary schools for students from wealth- shortages o f skilled and trained workers ier backgrounds. Recognizing that the seems to be disappointingly slow, despite 169 Elements of a Strategy for Shared Growth the substantial resources devoted by both majority who left were either lecturers or governments and donors to this effort dur- senior lecturers (ESRF2002). ing the past three decades. This deficiency is sustained at a time when Africa i s losing In medicine, Tanzania is also facing a a very significant proportion o f its skilled massive skills loss, especially o f doctors and professional workers to other markets and scientists. Low salaries for doctors are and increasingly depending on expatriates the principal reason driving the brain for many vital functions. Although com- drain-and those who remain seek higher paratively Africa i s the smallest source o f wages in private hospitals in large urban immigration to the industrial world, a high centers, leading to a lack o f doctors in proportion o f its migrants i s made up o f some o f the country's district hospitals. In highly skilled professionals. For example, a bid to increase the number o f healthpro- it has been estimated that more than 30 fessionals in the country, the Tanzanian percent o f highly skilledprofessionals in a government has recently promised to number ofAfrican countries are lost to the cover all training costs for medical stu- countries o f the Organisation for Eco- dents in both public and private universi- nomic Co-operation and Development ties (Balile 2003). (OECD). Nearly 88 percent o f adults who emigrate from Africa to the United States The brain drain also has an effect on the have at least a high school education. development o f human resources in Tan- More African scientists and engineers zania, especially given the AIDS crisis. A work in the United States than in Africa. recent article (Hazlewood and Prakash The emigration o f technically skilled peo- 2005) reiterates the fact that Tanzania ple has left 20,000 scientists and engineers faces an acute shortage o f health care inAfrica serving600 millionpeople. workers. Low pay, poor working condi- tions, and limited training programs con- Tanzania i s no stranger to brain drain. The tribute to the problem, which i s exacer- most vulnerable professions at the national bated by the rising burden o f treating level include medicine, accounting, law, HIV/AIDS patients. Hazlewood and engineering, and science. As a proxy for Prakash estimate that Tanzania will have the national picture, data from two premier to find nearly 10,000 more workers to ad- institutions o f higher learning provide dress the rising needs o f HIV/AIDS pa- some interesting evidence. Out o f a teach- tients and three times that number to meet ing staff o f about 861, about 149 staff the Millennium Development Goals members (17.3 percent) left the University (MDGs) (figure 9.6). However, improving o fDar es Salaambetween 1990 and March the productivity o f individual health care 2002. The majority o f those who left were workers and o f the health care system as a from the Faculty o f Arts and Social Sci- whole could increase Tanzania's capacity ences (38), followed by the Faculties o f by two-thirds, even without additional Medicine (17), Engineering (13)' Law workers. Some improvements, such as (ll), (lo), and Commerce (9). Science providing telephones and motorbikes for Most o f them exited at the senior lecturer better communication, would be relatively and lecturer levels. The same was true at easy to make; others such as managing the the Sokoine University o f Agriculture. Out flow o f patients more satisfactorily and o f a staff o f 239 people, about 50 (21 per- implementing planning and accounting cent) left in the same period. Again, the tools would require more investment and 170 Fostering Innovation,Productivity, and TechnologicalChange Figure 9.6. Projected Shortfall of Health Care Workers in Tanzania " Number of public-sector health care workers inTanzania To meet requirements of MDG2 To meet needs t30.000 78 300 Shortfall of 130% 35,200 34,000 'If care and trcatincnr plan implcincntcd 100%throughpublic sector. 2W?r"5 .l.illenniuni Ilevelr~pinent( d r . Source: Hazlewoodand Prakash2005. Note: MDG = Millennium Development Goal. training. Making these changes requires Thus, innovationcan contribute to poverty that health organizations and the govern- reductionby ment increase the capacity of their training programs by at least half to ensure a sus- Fueling economic growth, as it forms tainable workforce and aggressively re- the basis of new activities, industries, cruit trained workers to alleviate the im- or services that can generate wealth mediate shortages. andjobs Inducing productivity gains, which are Innovation also a source o f wealth as well as a Innovation can be understood as the diffu- means o f maintaining jobs from for- sion o f a new or improved technology or eign or other types of competition practice in a given environment. Two lev- Maintaining the self-sustainability of els o f innovation are particularly relevant local communities, for example, for developing countries: at the micro through the adoption o f appropriate level, there i s the diffusion o f available technologies technologies for use by firms, individuals, or households to help improve their pro- Role of the Government ductivity, welfare, living conditions, and A recent study by the World Bank so on; at the sectoral level, there i s the de- (Chandra 2006) focuses on examples o f velopment o f new industries, generally innovation from low- and medium-income based on foreign technologies, that can be countries and identifies a host of govern- a source o f new jobs, income, and exports. ment actions that are essential to the de- 171 Elementsof a Strategyfor SharedGrowth velopment o f competitive ind~stries.~ In ported technologies with local capacity for order o f importance, first are actions that R&D.However, in1985, it enactedthe first make available appropriate skills and pro- National Science and Technology policy, vide support for technology acquisition which was revised in 1995. The major and development. Next are actions o f a thrust o f this policy was to establishrelative regulatory nature, such as those relating to priorities and programs to generate new standards and quality control. Then come knowledge and to determine strategies for the various types o f support that can be science and technology (S&T) develop- provided to enterprises and industry or- ment. The government also established the ganizations for export promotion, invest- Tanzania Commission for Science and ment, and so on (table 9.1). Technology in 1986 and the Centre for the Development and Transfer o f Technology Major issues in Innovation in Tanzania (CDTT) in 1994, in an effort to institute workable mechanisms for coordinating ca- The innovation climate in Tanzania pre- pacity-building efforts, adopting new tech- sents serious weaknesses. First and fore- nologies, strengthening R&D, and facilitat- most is a generally poor business environ- ing information exchange and extension ment coupled with mediocre infrastructure, service^.^ Although these are laudable ini- bureaucratic hurdles, and corruption. The tiatives, the reality i s that these institutions, technical culture o f the population i s also especially the CDTT, lack adequate re- not very developed, as evidenced by high sources, infrastructure, equipment, and illiteracy rates. Inthe past, Tanzania had no trained personnel to respond to the in- deliberate strategies or plansfor appropriate creased needs o f the local entrepreneurial selection, acquisition, and transfer o f tech- society and to develop a coherent science nology or for effective integration o f im- andtechnology policy. Table 9.1. Role of the Public Sector in Fostering Innovation Technical Export and acquisition Regulation Support to Technical investment and and industry skills Industry Spinoffs promotion development compliance organizations development Electronics, Malaysia X X X X Electronics, Taiwan, China X X X IT, India X X X X X Maize, India X X X Grapes, India X X X X Oil palm, Malaysia X X X X X X Salmon, Chile X X X X Wine, Chile X X X X X Nile perch, Uganda X X X Cut flowers, Kenya X X X Source: Basedon Chandra2006. 172 Fostering Innovation, Productivity, and Technological Change Box 9.2. The United Nations Development Programme'sTechnology Achievement Index Inits HumanDevelopment Report 2001,the UnitedNations DevelopmentProgramme (UNDP) introducedthe Technology Achievement Index (TAI), which tries to captures how well a country i s creating and diffusing technology andbuildinga human skill base-reflecting its capacity to participate inthe technological innovations o f the network age. This composite index is not a measure o f which country is leading inglobal technology development but focuses on how well the country as a whole is participating increating and usingtechnology. The index recognizes that a nation's technological achievements are larger and more complex than any index can cap- ture.Itis impossible to reflect the fullrange oftechnologies-from agriculture to medicine to manufacturing. Many aspects o f technology creation, technology diffusion, and human skills are hardto quantify. And even ifthey couldbe quantified, a lack o f reliable data makes it impossible to fully reflect them. For example, important technological innovations occur inthe informal sec- tor and inindigenous knowledge systems, but they are not recorded and cannot be quantified. Thus,the TAIfocuses onthree dimensions ofinnovationat the country level: creation ofnewprod- ucts andprocessesthrough R&D, use o fnewtechnologies and old inproduction and consumption, and availability o f skills for technological learningand innovation. Countries are rankedinfour categories: leaders, potential leaders, dynamic adopters, and the marginalized. The results show three trends: a map o f great disparities among countries, with TAI values rangingfrom 0.744 for Finlandto 0.066 for Mozambique; diversity and dynamismintechnologicalprogress among devel- oping countries; and a map o f technology hubs superimposed on countries at different levels o f de- velopment. Tanzania and Kenya are bothlisted as marginalized (below 0.20>--which meansthat technology diffusion andskillbuildinghave a long way to go inthese countries andthat largeparts o fthe population have not benefited from the diffusion o ftechnology. Source: UNDP 2001. c The low level o f R&D as a percentage o f health care (11); wildlife and fisheries (4); gross domestic product (GDP-nly 0.2 and universities and other higher learning percent (comparable with other African institutions (9) (ESRF 2002).5 Most o f CountriesFreflects the modest nature o f these entities are government institutions; Tanzania's research and innovative effort. objectives include conducting scientific Other more sophisticated indexes such as research and designing and manufacturing the United Nations Development Pro- machinery and equipment for agriculture, gramme's Technology Achievement Index, as well as appropriate technologies for ru- which includes measures o fthe diffusion o f ral businesses and small and medium-size new technologies (such as computers) as industrialenterprises. well as o f the technical qualifications o f the population, c o n f m the low technological However, these institutions lack real incen- capabilities o f the country (box 9.2). tive schemes for researchers to conduct this type o f research; as a result, only a few re- Research and Development searchers tend to be very instrumental in in Tanzania R&D activities. Inthe main, there is a cru- cia1lack o f resources for R&Dinstitutes, as There are about 62 R&D institutions in a result o f a deliberate liberalization policy Tanzania, covering agriculture, including o f the government. Such institutes are in- livestock and forestry (28); industry (11); creasingly facing reduced support fkom the 173 Elements of a Strategy for Shared Growth government and are therefore undertaking cation system, Tanzania must develop reforms to become independent executive specific actions for the promotion o f inno- agencies. They are looking for new lines o f vation and technology diffusion inorder to business that could make a profit and thus put its R&Dinfrastructure at the service of enable them to meet their personnel and the country's development. There is aneed operational expenses. This trend is worry- for a systemic approach that provides ing because it distracts these institutions complementary support on three basic as- from conducting research o f a more public pects: financial, technical, andregulatory. goods nature that could benefit society as a whole. Mechanisms for technology diffu- Financial support. There is a crucial lack sion are also modestly developed, with a o f resources on the demand side for accel- near absence o f decentralized structures erating the design, testing, use, and dis- such as agriculture extension services. semination o f technologies. We suggest that Tanzania establish two complementary Foreign direct investment (FDI) is an im- schemes, which are basedon simplematch- portant source o f technological upgrading ing fund principles. The schemes would in developing countries. There is no doubt provide-in grant form-50 percent o f the that in Tanzania, FDI-which i s relatively funding required for the development high compared with other Afiican coun- (R&D phase) of small and medium-size tries-has played a key role in the mod- projects (for example, upto US$20,000): ernization o f important sectors o f the econ- omy such as trade (retail), banking, 0 User scheme. This scheme would al- tourism, and the telecommunication net- low particular groups and communities works. It has also been crucial inthe take- to buy needed technologies and pro- off and growth o f new industries such the vide, ifappropriate, complementary in- fishing and gold mining.But, as shown by kind funding (labor for community the international experience, it takes time to purposes). build an indigenous innovative capability Developer scheme. This scheme would through foreign investment. It requires ex- fund 50 percent o f technical services plicit mechanisms such as the employment or R&D projects undertakenby small o f large contingents o f local cadres in and medium-size enterprises with managerialpositions as well as programs to R&D institutes (public, academic, and closely link local suppliers o f components so forth). It would thereby help- and materials to upgrade the suppliers' indirectly, but more effectively-the equipment andthe quality oftheir products. R&D institutes use their competences Such mechanisms do not exist inTanzania; to serve communities. (This type o f consequently, the transfer o f knowledge scheme is in place in a number o f in- and technology fiom foreign sources re- dustrial countries.) mains modest. Box 9.3 provides an exam- ple o f constraints to technology access in It is important that the management o f the horticulture sector inTanzania. such support schemes be carried out pri- marily at subnational levels. Regional and Policy Proposal for Developinga local commissions would screen and select Nationwide InnovationSupport Scheme projects with the support o f appropriate experts (including foreign ones). This In addition to improving the overall busi- screening and selection process i s a pri- ness environment and upgrading the edu- mary condition that i s necessary to have 174 Fostering Innovation, Productivity, and TechnologicalChange Box 9.3. Constraints to Technology Access in the Horticulture Sector in Tanzania The horticulturalsector has become an important contributor to exports inTanzania. Nonetheless, Tanzania's performance significantly lags behindthat o f Kenya, which has made important strides inincreasing its market share for horticulturalexports. It i s useful to analyze the con- straints to technology access inthis sector and, inparticular, the barriers to the introductiono f improvedvarieties o f seeds and pest controltechnologies, with the aimnot only o f identifying constraints, but also o f derivingsome lessons for other sectors. The most important findingi s that it is not formal import restrictions that pose constraints to the introductiono f technologies usedinother countries to Tanzania, but rather the underlyingregula- tions and institutions that impede such access, as well as a lack o f information. Inthe case o f ac- cess to pesticides, for example, the most impedingfactors are the unclear institutional framework governing registrationprocesses and the cumbersome, long, and expensive registrationprocess. The registrationo f agrochemicalpesticides can cost upto US$7,550 and can take as long as two years. Furthermore, the horticulturalsector inTanzania, as compared with that inKenya, suffers from a backlogo f registeredpesticides that are urgentlyneeded. The introductiono f biological control agents i s even more cumbersome and time consuming. Differentissues are at stake con- cerning the introduction o fnew seedvarieties. The transfer o f knowledge regarding new seedva- rieties often does not take place because smallholder farmers buy from stockists, who have little knowledge o fhow to handle these seeds. Another constraining factor is the lack o f awareness regardingthe recently adopted Plant Breeder's Rights (PBRs). Some recommendations that can help to improve the introduction o f improvedvarieties o f seeds and pest control technologies include the following: Streamline legislation. M e r the reviewbythe NaturalResourcesInstitute, Universityo f Green- wich, hasbeencompleted, legislation shouldbe streamlined inline withrecommendations inor- der to closeloopholes andmakethe process ofregistrationmoretransparent. Clear institutional arrangementsregardingthe Tropical PesticidesResearchInstitute andthe registrar are needed. Create harmonization within the East African Community. Working groups onpesticides ex- ist in East African Community (EAC). Reciprocal recognition o fregistration within the EAC should be encouraged as an important outcome o f this process. Recognize Kenyan registration. Given that the process o f reciprocalrecognitionamong EAC countries may take time, Tanzania should automatically register pesticides that are already registeredinKenya. Implement quality control inspections of andprovide training to stockists. Stockists should be frequently inspected to ensure quality of seeds andpesticides. These inspectionswould help to eliminate the circulation o f counterfeit products. Stockists and smallholder farmers should receive training to increase awareness o f the benefits o fhybridseeds and use o f improved va- rieties. Public-private partnerships are an option, giventhe interest o f seedimporters inin- creasing sales o f their products. Dissemination o f information about the PBRs. Given the domestic as well as international lack o f awareness that Tanzania now has the PBRs inplace, efforts to informstakeholders are essential. Efforts should also be increased at an international level to attract foreign direct in- vestment andto stimulate innovationinthis area. Source: Skof 2006. 175 Elements of a Strategy for Shared Growth efficient management and to reduce the by business enterprises, and intellec- bureaucracy that would ineluctably affect tual property rights. schemes administered at the central level. The central level, however, would have to 0 Inmany sectors, there is a need to de- have oversight and control the overall velop quality awareness and quality process. control, as well as related accreditation and certification procedures. A pro- Technical support. On the basis o f ex- gram should be implemented to raise perience accumulated in industrial coun- awareness o f these issues, because do- tries, we suggested that a network o f locally ing so could yield important results in based and owned structures be established a short time span. to serve the needs o f rural and urban com- 0 Firms or individuals that are first pro- munities for technical advice, information, ducers face major financial problems and assistance (indesign, marketing, and so and frequently do not get access to forth). These structures should be adapted credit from the banking system. The to different sectors (for example, extension government's recently established services for agriculture and design and credit guarantee mechanism, which i s manufacturing workshops for industry). supposed to mobilize the banking and They should also be conceived and oper- financial sector, does not seem to be ated as antennas o f central bodies to which working well. An audit needs to be they would be strongly connected through undertaken to examine in detail the informationtechnology (IT), databases, and mechanisms that can be put in place the like. They should be established on a to complement this incentive, such as clearly expressed demand from local com- microcredit schemes; equity invest- munities and funded on a 50/50 cost- ment procedures (such as the Dutch sharing basis, with local organizations Program for Cooperationwith Emerg- (municipalities, business or farmer associa- ing Markets, which supports 30 per- tions, and so forth) matching the resources cent o f the investment o f individual putinbythe centralgovernment. firms in the flower industry); and the like. Regulatory support. Regulatory-related actions must be implemented to deal with An lnnovation Multipurpose Facility several issues: To efficiently implement and fund the 0 Proposed actions that aim to stimulate multicomponent program proposed here, service-based contracts and formalize Tanzania could do well to establish a new new links between the business sector ad hoc facility-an innovation multipur- and the R&D infrastructure require the pose facility-that i s endowed with a criti- establishment o f clear legal and admin- cal mass o f funds and that operates with istrative procedures. We therefore rec- maximum flexibility inthe use o fthe types ommend reviewing, adjusting, and o fpolicy actions outlined above. The facil- standardizing appropriate models for ity could be administered at the level o f such relationships. Many issues may the president's office. Funding needs can be involved, including the use o f pub- be estimated at US$10 million to lic or university laboratory equipment US$15 million per year. If matched by and personnel by firms, the temporary equivalent spending from the private sec- employment o f university researchers tor (which currently spends almost noth- 176 Fostering Innovation, Producfivity,and Technological Change ing), this amount would double the current port). Adequate actions then must be national S&T expenditure, reaching 0.5 undertaken, including food quality percent o f GDP. That expenditure could control and insurance, improvement o f reach full scale three to four years after an testing facilities, support o f packaging initial pilot phase, which would possibly enterprises and distribution channels, be focused on a few specific industries and systematic scouting o f foreign (discussed next). The effect on the econ- knowledge and technologies o f poten- omy would become evident within five tial use in Tanzania. These measures years or so. It i s suggested that this gov- should be closely coordinated with ernment support mechanism be modeled programs that aim to improve agricul- on the Tanzania Social Action Fund, tural productivity and diversity, nota- which has been efficiently administered at bly as funded by the World Bank(such the president's level. as the Second Tanzania Agricultural ResearchProject). Promotion of Specific Industries The involvement o f foreign enterprises in The scheme for an innovation multipur- both sectors i s crucial. These enterprises pose facility could be focused on specific provide access to foreign markets, provide priority industries, including the tourism management competencies, and introduce sector and the agro-foods industry. The up-to-date technologies. It i s important to first i s already an important source o f establish a liaison with Tanzanian enter- income; the second would take advan- prise associations, which are efficient or- tage o f the large agriculture base. Sys- ganizations for negotiating and partnering temic, well-focused action i s needed in with foreign businesses. Clear contracts both industries. regarding technology licensing, personnel training, and access to export markets Intourism, such actions include a gen- eral campaign o f quality control and should be developed. On the whole, a two- pronged action strategy combining both accreditation o f hotels and related fa- upgrading and developing domestic capa- cilities throughout the country, effi- bilities and involving foreign actors would cient enforcement mechanisms for en- be essential for a successful innovation suring the respect for and compliance intervention to help improve Tanzania's with the defined standards, an interna- growth prospects. tional campaign o f marketing and promotion o f the tourism advantages o f Tanzania inselected countries o fthe Information and industrial world, and adequate finan- Communication Technologies cial support to Tanzanian firms and Rapid advances in ICTs are dramatically organizations intending to improve or affecting economic and social activities, as create new services. well as the acquisition, creation, dissemi- For the agro-foods sector, it i s impor- nation, and use o f knowledge. These ad- tant to consider the different types of vances are affecting the way in which production (vegetables, cereals, h i t s , manufacturers, service providers, and gov- meat, and so forth) and the missing ernments are organized and how they per- links that need to be constituted or form their functions. As knowledge and consolidated, from the basic producer innovation become increasingly important level to the market (both local and ex- elements o f competitiveness, the use o f 177 Elemenfs of a Sfrafegy for Shared Growfh ICTs i s reducing transaction costs and 9.7. Mauritius, by contrast, has beendoing time and space barriers, thereby allowing better than Botswana and South Afi-ica on the mass production o f customized goods telephony and personal computer (PC) and services and substituting for limited penetration andi s close to the level o f Ma- factors o f production. The pervasive and laysia. Inthe case o f Internet hosts, South global ICT revolution i s disrupting all Africa and Malaysia are the undisputed kinds of relationships, helping build new leaders. types o f organizations, widening the knowledge and productivity gap, and pos- A recent report by the United Nations ing serious risks for the unprepared. Itthus Conference on Trade and Development calls for countries such as Tanzania to de- (UNCTAD) also provides some insights velop capabilities to master the new tech- into the international digital divide. I t nologies and harness the full potential o f evaluates ICT development using a range ICTs for all sectors o f the economy for of indicators to benchmark connectivity, education, innovation, and learning; public access, ICT policy, and overall ICT diffu- sector management; private sector com- sion in 165 countries. Inthe benchmarking petitiveness; and capacity building. analysis, OECD countries continue to dominate the upper rankings, while South Research also shows strong links between Asian and African countries occupy the ICTs and growth. Compelling evidence lower half o f the rankings. The more de- exists that strengthening telecommunica- veloped African countries enter the rank- tions infrastructure and service is pivotal ings relatively early, with Mauritius 52nd in promoting trade and economic growth. and South Africa 66th. Botswana comes in It is estimated, for example, that a 10 per- 80th, while Kenya i s ranked 115th, Tanza- cent decrease in the bilateral price o f nia 135th, and Uganda 154th, indicating phone calls i s associatedwith an 8 percent that many Sub-Saharan Africa countries increase in bilateral trade (Fink, Mattoo, have a considerable way to go in terms o f and Neagu 2002). In Africa, significant ICT connectivity and difhsion (UNCTAD evidence suggests that if the telephone 2005). growth rate were 10 percent instead o f 5 percent (and if growth in electricity gen- Tanzania has been making some progress. eration were 6 percent instead o f 2 per- In 2003, it publisheda cross-sectoral Na- cent), the increase in Africa's growth rate tional ICT Policy (http://www.moct.go would be at least 0.9 percent higher (Es- .tz/ict) that relates ICTs to relevant sectors tache 2005). o f the economy, such as education, manu- facturing, health, and tourism. The policy Information infrastructure consists o f tele- was developed in response to the poor communications networks, strategic in- harmonization o f initiatives that has led to formation systems, and policy and legal random adoption o f different systems and frameworks affecting their deployment, as standards, unnecessary duplication o f ef- well as skilled humanresources neededto fort, and waste o f scarce resources, espe- develop and use it. In the ICT domain, cially through the loss o f potential syner- Kenya, Tanzania, and Uganda are all at a gies. The policy notes that the weak ICT very nascent stage o f application and use. infrastructure and the lack o f adequately It is not surprising that all three countries trained and skilled personnel are the main lag Botswana, Malaysia, and South Africa barriers to increased adoption o f ICTs in by a huge margin, as can be seen infigure Tanzania. This broad-based national strat- 178 Fostering Innovation, Productivity, and Technological Change Figure 9.7. ICT Infrastructure:Telephones, PersonalComputers, and Internet (a) Telephones(fixed mainlinesand mobile phones) (b) Personalcomputers (c) Internethosts -- miom- InternetHosts -yuwu Source: World Bank internal database. egy is designed to correct these weaknesses and coordination, it fails to loop in the by addressing Tanzania's developmental productive economic sectors. Expenditure agenda and calling for the creation o f ap- on building infrastructure and training propriate institutional arrangements to en- should be strategic investments after the sure that all stakeholders can rise to the needs and demands o f the private sector challenge of implementingthe ICT policy. are addressed-specifically which firms It is worth mentioning that stakeholder dis- can benefit the most from ICT, leading to cussions on ICT policy and related issues appropriate allocation o f scarce resources. were held mainly through an e-mail list o f e-think tanks-an ICT fraternity, compris- Inthe telecommunications sector,the policy, ing representatives from government, the legal, and regulatory framework has been private sector, and civil society. encouraging private sector participation.6 This sector i s regulated by the Tanzania Eventhough the government's ICT policy Communications Regulatory Authority addresses issues related to rationalization (TCRA). The performance o f the Tanzanian 179 Elements of a Strategy for Shared Growth Telecommunications Company Limited ture sharing. Thus, in general, Tanzania's (TTCL) has improved considerably since postal and telecommunications services February 2001, when a Dutch-German con- are weak, and the provision o f fixed tele- sortium, Celtel, took a 35 percent stake init. phone lines has been meager. An inade- The remaining shares were allocated to local quate regulatory framework persists, and financial institutions (14 percent), interna- competitionhas been hampered by various tional financial institutions (10 percent), and issues, such as inadequate interconnection TTCL employees (5 percent); the govern- agreements and directives, high fees and ment retained a 36 percent stake. At present, royalties levied by the TCRA, and the ab- TTCL has about 250,000 operational lines sence or nontransparency o f regulatory (InfoDev 2005). oversight. The sector liberalization and the privatiza- Progress made under the 1997 National tion o f TTCL have had a significant effect Telecommunications Policy is expected to on the market dynamics, particularly inthe continue. The mobile telephone market i s supply o f telecommunications services. fully competitive. Significant liberalization Market revenues grew from US$143 mil- has also taken place in various segments: lion in 1998 to US$389 million in 2003. private operators now provide basic, mo- The compound annual growth rate i s 19 bile, data, paging, Internet, payphone, and percent. Overall teledensity grew from 0.3 other value-added services. The mobile in 1998 to 2.57 in 2003, and the mobile telephone market involves a number o f op- market has grown 21.19 percent since lib- erators and i s growing rapidly. New mobile eralization and the introduction o f compe- operators have committed significant finan- tition, new products, and new services. cial resources to the development o f a state- of-the-art telecommunications infkastruc- Despite the competition, tariffs remain ture. Four providers operate under 15-year highandteledensity is one o fthe lowest in licenses: MIC Tanzania, Zanzibar Tele- the South African Development Commu- coms, Vodacom, and Celtel. As a result, nity (SADC), inpart due to the poor inter- overall teledensity-mainlines plus mobile connection framework and the lack o f phones-increased to 24 per 1,000 between regulatory independence and also because 1996 and 2003 (table 9.2). Tanzania's o f other issues, such as lack o f infkastruc- mainline and mobile phone penetration Table 9.2. ICT Indicatorsfor Tanzania and Comparators 1996-2003 2002 2000 2003 Internet Internet Mainlines Mobile Personal hosts users (per telephones Radios Televisions computers (per (Per Country 1,000) (per 1,000) (per 1,000) (per 1,000) (per 1,000) 10,000) 1,000) Botswana 87 241 150 44 40.7 13.99 50a Kenya 10 42 221 26 6.4 0.32 200a SouthAfrica 107 304 336 177 72.6 41.94 2,890a Tanzania 5 19 406 4 4.18 0.16 240 Uganda 2 16 122 18 3.32 0.07 125 Sub-Saharan 15 37 198 69 11.90 3.10 6,233 Africa, average Source:World Bank2004a, 2005a. a. Denotes data for 2001. 180 Fostering Innovation, Producfivity, and Technological Change i s higher than Uganda's. Anecdotal evi- crease intransparency and professionalism dence also exists that mobile phones are inthejudiciary system (Accenture, Markle increasingly being used inTanzania to get Foundation, and UNDP 2001). business-related information and to reduce transaction costs. For example, traders in Summary of Issues Dar es Salaam now can place orders with and Recommendations producers o f bananas throughout the coun- try-thus linking demand and supply in Some measures that can help strengthen real time and enhancing the efficiency o f Tanzania's performance in the three h c - markets. tional pillars are highlightedhere. Tanzania also has a comprehensive Inter- Education net service, including three licensed data- service providers and 21 Internet service Key challenges facing Tanzania in this providers. Most users access the Internet domain includethe following: from urban Internet cafks-the Internet i s not very accessible inthe rural areas. Tan- Sustaining and improving the quality zania in 2003 had twice as many Internet o f education as enrollments increase users (250,000) as Uganda (125,000). The by recruiting teachers, constructing government has developed a fairly com- classrooms, increasing preservice prehensive national Web site (http://www teacher training, and providing subsi- .tanzania.go.tz/), which provides consider- dies for purchasing teaching and learn- able background information on the econ- ingmaterials. omy and political structure o f the country; Ramping up secondary education, in- it hopes the site will help raise the coun- cluding improving its quality and rele- try's international profile and attract for- vance to the needs o f the economy. eign investment. In addition, a number o f ministries, state institutions, and embassies Inhigher education, strengthening the have their own sites. governance and administration o f the country's three public universities in Interms o fdeveloping humanresources in terms o f financial sustainability, up-to- IT, training centers that focus on the de- date content, and teacher training. velopment o f ICT knowledge workers are Usingthe potential o fdistance education only now beginning to emerge. For exam- to expand access to education services ple, the Soft Tech Training Center, estab- while improving equity. The Open Uni- lishedin 1993, i s committedto developing versity o f Tanzania offers degree pro- local expertise through ICT skills en- grams by correspondence and also in hancement. The government has also initi- regional centers. The costs are low, be- ated plans to encourage Tanzanians to de- cause the state covers tuition. But en- velop content that i s relevant to local rollments are low, partly because o f lack interests, and Tanzania has implemented o f content and partly because o f a several ICT applications relevant to its na- dearth o f partnerships with international tional objectives. Examples o f such initia- academic institutions that could provide tives include an information system to degree programs online. Combining strengthen the capacity o f wildlife institu- distance-education modalities with ex- tions and a computerized case-flow man- tended face-to-face interactions with agement system that has facilitated an in- Tanzania's other public universities may 181 Elements of a Strategy for Shared Growth be one way to boost enrollments andin- nical skills at all levels (from primary crease access to higher education. schools to university colleges), with Reforming teaching methods and the particular attention to vocational and curriculum at all levels to include professional schools. skills and competencies (communica- Facilitating access to and use o f foreign tion skills, problem-solving skills, knowledge and technology by attracting creativity, and teamwork) to meet the foreign investors and stimulating ap- new needs o fthe economy. propriate transfer through personnel Increasing the interface between indus- employment and training, links with lo- try and education and offering differ- cal suppliers o f components, and the entiated curricula that better meet like. the new skill demands o f industry, Increasing the resources o f the R&D generated by changing markets and institutes (public and universitybased), technologies. linking them to users' demands and Harmonizing the technical education needs. offered in secondary schools with that Strengthening structures, incentives, offered in technical colleges and then and regulations to facilitate the difh- linkingthese schools with the proposed sion o f up-to-date and appropriate zonal and regional institutes and col- technologies throughout the economy. leges. These institutes and colleges should offer differentiated products to Creating an ad hoc mechanism to pro- meet the differing needs o f industries, mote technological innovationthrough- such as mining, fisheries, major cash out the country and facilitate the dis- and food crops, external trade, and semination o f improved and new metal. technologies. This mechanism would make concrete and focus the overall ef- Devising strategies to proactively deal fort o f innovation promotion. Such a with problems o f skills lost through scheme, based on decentralized struc- braindrain. tures and initiatives, should address al- together the financial, technical, and lnnovation regulatory needs o f enterprises as well as those o f local communities. At the To encourage innovation, Tanzania needs same time, it should help support the to focus on the following efforts: R&D infrastructure by orienting it to- wardthe service o fthe country's needs. Improving the overall business envi- The scheme should give priority to a ronment (regulatory, tax, bureaucracy, few important industries such as agro- and other aspects) and the basic infra- foods andtourism. structure (especially for transport and power), both o f which currently create Conducting a comprehensive nation- considerable hindrances to any form o f wide innovation andR&D survey to es- innovative activity, even the most tablish concrete factors that either fa- modest ones. cilitate or hinder innovative activities in the country. The outcome o f the survey Improvingthe overall technical culture can serve to put inplace concrete inno- o f the population and notably the tech- vationpolicies andstrategies. 182 Fostering Innovation, Productivity, and Technological Change lCTs experiences in operating in a more To strengthen its information infrastruc- competitivemarket. ture, Tanzania should continue to pursue 0 Supporting the development rural tele- the following goals: communications infrastructure, such as bydeveloping universal access schemes. Finalizing and adopting the new elec- Rural areas lack telecommunications tronic communications bill, which i s services or have only limited access in keyto definingthe groundrules for sec- areas adjacent to maintowns andonma- tor development (including ruralareas). jor trunk roads. This effort requires de- veloping content in local languages Implementing the new converged li- (such as Swahili). censing framework, which will ensure hrther liberalizationofthe market. 0 Enhancing technical and business- related skills development among the Reviewing and modernizing telecom- population using ICTs through techni- munications policies andregulations to cal institutes and vocational centers. generate fair competition and reduce For example, the University o f Dar es high communication and operational Salaam i s offering I T training in its costs. computer center to the public. Building capacity to undertake such 0 Continuing to use global experiences reforms, including through the estab- to enhance the efficiency o f the tele- lishment o f systems and processes to communications sector. In many areas review the performance o f the regula- o f telecommunications reform, Tanza- tory institutions. For example, given nia has benefited by adopting best the great demands and expectations practices from both industrial and de- placed on the regulator (TCRA) by veloping countries. The hnctions and telecommunications sector reforms, roles o f the national regulator (TCRA) the Swedish government, through the are the best example. Further benefits Swedish International Development from global experience and best prac- Cooperation Agency, i s helpingTCRA tices depend on the capacity o f TCRA create capacity to meet its existing and and other institutions to learn from the future challenges and learn from its experiences o f other countries. Notes postsecondary nonuniversity courses; perhaps these ' schools are not counted inthe official statistics. In terms o f monitoring education quality, the SACMEQ monitors and evaluates the quality o f The studies cover the extremes o f the industrial education in selected southern and East African spectrum: wheat and maize in India, salmon and countries. The SACMEQ I1Project (2000-03) has wine in Chile, Nile perch in Uganda, oil palm in been completed in 13 countries: Botswana, Kenya, Malaysia, cut flowers in Kenya, medium-tech elec- Lesotho, Malawi, Mauritius, Mozambique, Na- tronics in Malaysia, high-tech electronics inTaiwan mibia, the Seychelles, South Africa, Swaziland, (China), and software exports from India. The indus- Tanzania (separate assessments on both the tries are chosen based on exceptional comparative mainland and Zanzibar), Uganda, andZambia. performance in the past decade, large contributions to overall growth, and the role that technological These statistics should be viewed in perspective: change played inits success (Chandra 2006). the countries differ intheir systems o fpostsecondary education. Tanzania has many students enrolled in For more information, see the science and tech- nology section o f the Tanzania Country Profile on 183 Elements of a Sfrategy for Shared Growth Tanzania's national Web site: http://www.tanzania .go.tz/science-technology.html. Tanzania's S&T infrastructure includes education infrastructure and R&D institutions such as the University o f Dar es Salaam; Sokoine University of Agriculture; the University College o f Lands and Architectural Studies; Muhimbili University Col- lege o f Health Science; Rwegalulira Water Re- sources Institute; and the National College of Mbeya, Arusha, and Dar es Salaam Institute of Technology. For example, Draka Comteq, which consists o f 11 companies inDenmark, Finland, France, Germany, the Netherlands, Norway, Singapore, the United Kingdom, and the United States has recently won a US$30.2 million turnkey project in Tanzania, in- volving the supply and installation o f two long- distance links, covering some 2,300 kilometers of optical fiber cable and 500 kilometers of optical power ground wire (Economic Intelligence Unit 2004a). 184 CHAPTER 10 Enhancing the Business Environment The quality of the business environment Figure 10.2 compares the cost structure of affects the cost o f doing business and thus firms in a sample o f selected countries, a country's attractiveness to investors and dividing the cost o f businesses into labor, its international competitiveness. The costs capital, input, and indirect costs. The ef- of an inefficient business environment are fect o f a poor business environment o n estimated to be very high in Tanzania in firms is often reflected in high indirect international comparisons. They amount to costs. Many African firms incur heavy 25 percent of sales and include the cost of costs for transport, logistics, telecommuni- contract enforcement difficulties, regula- cations, water, electricity, land and build- tion, bribes, crime, and unreliable infra- ings, marketing, accounting, security structure (figure 10.1). bribes, and so forth. Indirect costs as a 185 Elements of a Strategy for Shared Growth Figure 10.1. Cost of Inefficiencies in the Business Environment as a Percentage of Sales, Various Countries 30 w Poland China Brazil Algeria Tanzaala Source: World Bank 20059. Figure 10.2. Cost Structure of Firms byAverage Percentage of Total Costs china I I I Source: Eifert,Gelb, and Ramachandran2005. 186 Enhancing the Business Environment Figure 10.3. Percentageof Enterprises Rating Problems as Major or Very Severe Constraintson EnterpriseOperations and Growth in Tanzania, 2003 80 i73 Source: World Bank 2004c. percentage o f total costs on average are highest gains interms o f economic growth more than 20 percent inTanzania, equal to ifappropriate actionistaken: theprovision the average cost o f labor. Ina global econ- o f complementary infrastructure, the cost omy, where Tanzanian products compete o f and access to finance, and the cost aris- with those of countries such as China and ingfrom bureaucracy and corruptioninthe India, high indirect costs are a severe im- interaction o f the private with the public pediment to economic activity. In China, sector. indirect costs are only about 8 percent o f total costs; in Tanzania, they are about 24 Scaling Up Access percent o f total costs. In addition, factors to Infrastructure that affect indirect costs also affect the costs o f other inputs and thus lead to a loss This section assesses the extent to which o f competitiveness in the economy that infrastructure i s a binding constraint on exceeds the loss at the firm level. growth in Tanzania, its role in promoting growth and poverty reduction, and the key An investment climate assessment carried issues in scaling up investment in infra- out in 2003 (WED 2004), suggests that structure. the tax system, high-cost credit and limited access to credit, limited availability and Investingin infrastructure is important for poor reliability o f infrastructure services, growth. The experience o f fast-growing and red tape in the public sector are the developing countries such as China shows principal constraints to even higher growth that infrastructure can contribute signifi- rates (figure 10.3). cantly to growth. Infrastructure affects growth through its effect on enterprise This chapter focuses on three aspects o f productivity, the cost o f doing business, the business environment that promise the market access, and profitability. Analysis 187 Elements of a Strategy for Shared Growth Figure 10.4. Effect of Low Levels of Infrastructure on Economic Growth (a) Roads (b) Electricpowerconsumption Roads, paved (%of total roads) Electricpower consumption (kwh per capita) 100 90 9000 80 70 8000 7000 60 6000 50 5000 40 4000 30 3000 20 2000 10 1000 0 0 Source: World BankWorld DevelopmentIndicatorsdata. o f firm-level data identifies access to in- constraints on business enterprise growth. frastructure services as a key determinant Power supply, in particular, i s perceived o f enterprise growth and investments. For as the most serious infrastructural con- a low-income country such as Tanzania, straint. These constraints have driven where the majority o f the rural poor are business enterprises in Tanzania to invest smallholder farmers, reliable and afford- in their own infrastructure. The Invest- able infrastructure (particularly rural ment Climate Assessment has reported roads) i s a critical factor in improving that about 55 percent o f enterprises are market access and enhancing the capacity more likely to own generators to provide o f farmers to commercialize and diversify backup power supply. These generators into higher-value economic activities to represent a significant additional cost o f improve incomes. doing business, given that they are expen- sive to buy and run. The indirect costs o f Tanzania's infrastructure i s weak and in- poor infrastructure and private provision adequate. Although Tanzania has started o f infrastructure have been estimated at reforming its infrastructure sectors and 25 percent o f sales, compared with 7 to public spending (especially on roads) has 10 percent in Asia and 18 percent in increased, the country's infrastructure in- Uganda.' As pointed out in an independ- dicators are still among the lowest in the ent operations evaluation report by the world (figure 10.4). International Finance Corporation (World Bank 2000a), unreliable and expensive Thus, infrastructure continues to pose a electricity, high transportation costs, and major policy challenge. According to the poor communications are major factors 2004 Investment Climate Assessment, underlying Tanzania's lack of external infrastructure ranks among the three top competitiveness. 188 Enhancing the Business Environment The extent to which infrastructure is abind- Several studies have been carried out to ing constraint on Tanzania's growth can be estimate the contribution o f public spend- partly assessed from returns to infrastruc- ing on infrastructure to growth and pov- ture. Because infrastructure i s widely be- erty reduction in developing countries on lieved to be such a binding constraint, one the basis o f marginal returns and cost- would expect high economic rates o freturn benefit ratios. A recent study on public to this type o f public investment. To get an investment and poverty reduction in Tan- idea o f the magnitude o f the returns to in- zania attempted to compute marginal re- frastructure inTanzania, we looked at esti- turns andcost-benefit ratios for the various mates o frates o freturn on 26 roadrehabili- types o f public investments, using house- tation and upgrading projects that the hold data.3 The marginal returns and the government has identified as part o f its benefit-cost ratios were disaggregated by transport sector investment program. The geographic zone and compared across the rates o freturnranged from 9 percent to 103 various categories o f public spending. The percent, and the average was 32 percent, preliminary conclusion was that invest- which i s high. The ex post economic rates ment inrural roads has a highpayoff, with o f return on completed infrastructure pro- an average benefit-cost ratio o f 9:l. The jects financed by the World Bank, as pre- study also found that roads have diverse sented in its Implementation Completion effects across regions. The policy implica- Reports and Operations Evaluation De- tion i s that targetedpublic infrastructure is partment reports, however, show a some- required if the effect on growth is to be what mixed picture. A number o f the se- maximized. Further analytical work on lected projects had exceptionally high ex rates o freturnto infrastructure inTanzania post rates o f return: for example, the Sixth usingmore robust data and methodologies Highway Project yielded a rate o f return o f would provide greater insight into the 37 percent; the Fourth Power Project, 23 growth and poverty effects. To shed more percent; the Telecommunication Project, 50 light on the infrastructure challenges, the percent; and the road and storm water com- following section elaborates on the con- ponent o f the Urban Sector Rehabilitation straints inthe infrastructure sectors and the project, 33 percent. A few had low or even relatedpoverty andpolicy issues. negative rates o f return. The Highway IV Project, for example, yielded an economic Transport i s o f strategic importance to rate o freturno f4 percent. growth and poverty reduction: Given Tan- zania's geography and the dispersal o f ar- The infrastructure constraint inTanzania i s eas o f economic activities, roads are par- region and geography specific: The rates ticularly critical for the country's growth. o f returnto infrastructure in Tanzania are, The total road network i s estimated at generally, high.2However, the large varia- 85,000 kilometer^.^ Although road main- tion in the size o f the returns in each sec- tenance has improved since the creation o f tor-especially roads-suggests that the the Road Fund Board and the Tanzania infrastructure constraint inTanzania i s ge- Roads Agency (TANROADS), the overall ography specific. The magnitude o f re- condition o f the road network remains turns and the growth effect also dependon poor because o f underhnding and capacity the type and quality o f infrastructure. Gen- constraints. Table 10.1 depicts the condi- erally, returns on rural roads linking pro- tion o f the road network. O f the 85,000 duction areas with markets are higher than kilometers o f road, only 27 percent i s on other categories o froads. judged to be in good and fair condition. 189 Enhancing the Business Environment This finding i s relevantto Tanzania, where poverty alleviation into maintainable condi- the quality o f infrastructure i s closely re- tion, about US$3 billion is requiredover the lated to agricultural growth and the inci- next 10 years (World Bank 2004b). Cur- dence o f poverty. Those zones where pov- rently, Tanzania spends about US$lOOmil- erty incidence is the highest-the south lion to US$l50million per year in the road and the central zones-are the least con- sector, which is far from adequate. The nected, while poverty i s considerably funding problems have been compounded lower in well-connected Dar es Salaam by weak road administration capacity at the and the southern highlands. local government level and weak local pri- vate contracting capacity. For Tanzania to Road quality has a significant effect on pro- achieve its goal o f having about 45,000 ducer prices, the length of the supply chain, kilometers inmaintainable condition, a large competition, and access to basic social ser- amount o f additional financing i s required. vices. A recent study ofthe logistics costs o f However, it i s important that Tanzania ac- rural marketing (Nyange 2005) has shown celerate road sector reforms to ensure effec- that villages inareaswithpoor roads are less tive use of the funds earmarkedfor road re- well served by freight and public transport, habilitation and maintenance. In particular, as evidenced by fewer traffic arrivals, the the government needs to (a) enact a new limited tonnage of trucks, and more pro- Roads Act to transform TANROADS into nounced seasonality o f traffic. Villages with an independent authority and (b) redefine poor roads do not have regular transport and the responsibilities o f institutions inthe sec- are accessedby transporters only during the tor to preventduplication. peak season for crop marketing, thus deny- ing farmers the opportunity to benefit from Improved rail services would lower freight higher prices in the early and late seasons. transport costs. Although the rail system Moreover, there i s far less competition in share o f freight traffic i s far less than the villages withpoor roads, as evidencedbythe road system share, the rail system provides smaller number o f crop agents and export an important link between inland regions companies operating inthose villages. Con- and ports and i s a cheaper alternative for sequently, the producer prices received by transporting agricultural produce over long farmers in villages with poor roads are less distances. For more than a decade now, the thanthe prices receivedbytheir counterparts railway has faced major infrastructural and invillages with fair roads.This situationhas operational problems. Large sections o f a disincentive effect on production in the track are obsolete, and the available wag- less accessible villages. The survey also un- ons and locomotives are not sufficient to derlines the importance o f roads inimprov- sustain a reliable and efficient rail service. ing access to basic social services. As ex- Exporters complain o f delays and the un- pected, well-connected villages are better certainty in allocation o f locomotives, es- served with basic social and economic ser- pecially for crop haulage. The low rail vices, such as schools, health facilities, and transport capacity i s a source o f major lo- grainmillingplants. gistical problems in the economy and has contributed to congestion at the port. It is, Thus, more investment in the rehabilitation therefore, vital that the privatization or and maintenance o f the road network is concessioning o f the railways and rehabili- neededfor Tanzaniato achieveits long-term tation o f the infrastructure be undertaken growth objectives and poverty reduction as planned in order to improve efficiency goals. To get the core network required for and reduce freight transport costs. Fur- 191 Elements of a Strategy for Shared Growth thermore, improving rail services will help There i s anurgentneedto expandthe power reduce pressure on the road infrastructure, supply as well as its reliability and the effi- leading to lower costs for road rehabilita- ciency o f the energy sector through sector tion and maintenance. reform and investment, to meet the growing demand by industry and service sector. The The power sector in Tanzania has prob- governmenthas been implementinga power lems. For many years, Tanzania has ex- sector reform program aimed at improving perienced the effects o f droughts, high efficiency in power supply through com- technical and commercial losses in the mercialization and private sector participa- grid system, and (until recently) poor col- tion. Since 2002, the management o f the lection o f receivables and low tariffs. Per Tanzania Electric Supply Company capita power consumption in Tanzania i s (TANESCO) has been contracted to a pri- estimated at 62 kilowatts, much lower than vate company, Net Group. This arrangement in comparator countries such as Kenya has resulted in strengthened financial man- (120 kilowatts per capita), India (380 agement of TANESCO, which, however, kilowatts per capita), and China (987 remains vulnerable to external shocks such kilowatts per capita). Hence, there i s scope as the drought experienced in 2005/06. Un- for a significant increase in power con- fortunately, the technical turnaround has not sumption. currently, hydropower i s the yet been achieved-hence the continued major source, accounting for 70 percent o f problems inpower supply. the country's grid generation capacity of 773 megawatts. The remaining 30 percent To improve in the performance of the comes from thermal plants. The thermal power sector, further reforms and substan- plants that use imported fuel (diesel and tial investment in transmission and distri- heavy fuel oil) for power generation, bution are needed. Inthis regard, it i s im- mainly the Tegeta plant, are very expen- portant, that the government develop a sive to runcompared with those now using clear strategy and detailed implementation indigenous natural gas. plan for executing the remaining aspects o f the power sector reform. The key ac- Unreliable power supply i s perceived by tions to be undertaken include (a) promul- investors as a severe constraint on enter- gation o f the revised electricity legislation prise operations and growth. The poor per- that would allow private participation in formance in the power sector i s reflected the extension o f electricity services; (b) in frequent power outages. According to staffing and operationalization o f the En- the recent InvestmentClimate Assessment, ergy and Water Utilities Regulatory Au- the median number o f power outages ex- thority (EWURA); and (c) eventual con- perienced by enterprises in 2002 was esti- cessioning o f TANESCO, depending on mated at 48 compared with only 21 in the market situation. Tanzania i s already Kenya and 20 in Uganda. The high fre- encouraging private sector participation in quency o f power outages and poor quality generation expansion and needs to build o f supply caused by low voltage impose on this effort. In addition, the country can highfinancial costs on businesses through benefit from regional power integration loss o f production, wastage in the produc- through transmission interconnectors, no- tion process, and damage to equipment. tably with Kenya and Uganda. Perhaps not surprising, electricity i s the infrastructure constraint enterprises are Currently, less than 5 percent o f the rural most concerned about. population has access to electricity, and 192 Enhancing the Business Environment the overwhelming majority continues to o f waterborne disease, especially among rely on helwood for energy.The low elec- women and children. Apart from its direct tricity access rate for the rural population effect on social outcomes, poor water constrains the development o f nonfarm supply also affects rural productivity, be- activities and the improvement o fthe qual- cause the time spent by women fetching ity of life inrural areas. However, extend- water could be spent on more productive ing the main grid to rural areas is expen- activities. As in other infrastructure sec- sive, and the rural poor are unlikely to tors, meeting the long-term targets for afford electricity without subsidies, at least service provision will be challenging. The to cover the capital portion. To improve Millennium Development Goals needs access, Tanzania should consider investing assessment for Tanzania shows that, to in independent grids using small hydro reach the water and sanitation target, in- systems (pico, micro, or mini), as well as vestment per capita must double, to natural gas, as in Somanga. But for more US$12 in 2015 from the projected US$6 intensive uses than lighting for house- in 2006. Further institutional reforms in holds, a better substitute for biogas could the water sector are also critical for be liquefied petroleum gas. The other op- achieving the long-term sector goals. tion i s to promote the use o f solar photo- voltaic systems in small communities. Policy Recommendations Such systems would meet rural areas' for Scaling Up Investment lighting needs with respect to basic social in Infrastructure services such as health and education and allow the runningo f small water pumps. There i s no doubt that scaling up invest- ment inthe rehabilitation and expansion of Inadequate water supply affects both existing inf?astructure i s necessary to pro- growth and human development. Al- mote sustainedgrowth. However, inscaling though Tanzania i s endowed with abun- up infrastructure investment, the govern- dant freshwater resources, the provision ment must consider some policy issues to o f water for domestic andindustrial use i s ensure that the investment yields positive inadequate. More than 15 million o f the effects on growth and onpovertyreduction: 35 million Tanzanians lack a safe water supply. Despite recent improvement, only 0 While public financing, including do- 50 percent o f the rural population has ac- nor financing, will be critical for im- cess to clean sources o f water, and only proving basic infrastructure such as 70 percent o f the urban population does. roads and rural water, sectors such as Inadequate water supply in Tanzania railways, ports, and energy offer stems from underinvestment, past neglect greater potential for private sector par- o f maintenance o f facilities, and weak ticipation through the public-private water resource management,and institu- partnershipmodel. However, to attract tional capacity. Currently, 30 percent o f private investors, the regulatory rural water supply facilities are unreliable framework must be strengthened and or not functional. Some 20 to 40 percent the regulatory bodies must be provided o f urban water i s unaccounted for because greater autonomy. Strengthening Tan- o f technical and commercial losses. The zania's capacity to monitor the effect low domestic water supply coverage- o f infrastructure privatization on ex- especially in rural areas-affects quality panding access to infrastructure ser- o f life, as evidenced by the highincidence vices will also be important. 193 Elements of a Strategy for Shared Growth 0 Reforms inthe road and power sectors sity in economic activities, population will be critical in ensuring that invest- densities, remoteness, and other fac- ment in infrastructure contributes ef- tors. The policy implication i s that in- fectively to efficiency-based growth frastructure interventions must be re- andpoverty reduction. It is particularly gional or area specific to better reflect important that a new Road Act be local growth needs and priorities. Part passed, along with the Electricity Act. o f the regional infrastructure develop- The benefit incidence o f electricity ment strategies could entail clustering consumption clearly shows that elec- infrastructure investments in areas o f tricity is almost exclusively consumed growth potential, which would enable by people in the higher income quin- producers to exploit economies o f tiles. Thus, the rationale for subsidies scale and lead to faster growth. to the power sector i s weak, which un- derlines the potential for private sector 0 The infrastructure planning process investment inthe power sector to meet must take into account synergies be- the growing urban demand. tween the different types o f infrastruc- ture as well as links with other growth Tanzania's infrastructure requirements and poverty reduction initiatives. The are large and must be properly priori- effect o f infrastructure on growth i s tized. Givenbudgetary constraints, it is likely to be enhanced if an integrated not possible for Tanzania to improve approach in planning is adopted. For the quantity and quality of infrastruc- example, in the case o f transport, a lo- ture across the board. Therefore, it is gistical corridor approach integrating important that the large infrastructure ports, maritime and coastal shipping, investment requirements be properly railways, road freight, terminals, and prioritized through the existing sector warehouses and distribution centers i s and the medium-term expenditure likely to promote greater efficiency by planning and budgeting process. We lowering logistical costs. suggest that, in prioritizing infrastruc- 0 The selection o f infrastructure projects ture projects (particularly roads), the must be based on economic efficiency main considerations should include the criteria and demonstrated poverty ef- need to (a) strengthenconnectivity be- fects: The pressure to scale up invest- tween potential high-growth areas and ments in public infrastructure may domestic and regional markets and (b) compromise the quality o f public in- connect the poor to emerging growth vestments. It is therefore important opportunities and improve access to that the selection o f public infrastruc- basic social services. Priorities in road ture projects be guided, first and fore- sector investments should thus focus most, by efficiency criteria. Priority on upgrading rural roads (to open up should then be given to infrastructure areas of high economic potential) and projects that meet the economic effi- rehabilitating and maintaining major ciency criteria andbenefitthe poor. transport corridors and regional roads to enhance integration and connec- 0 Appropriate balance needs to be main- tivity to domestic markets. tained between new investment, reha- bilitation, and recurrent maintenance. 0 Infrastructure needs and constraints in Although road sector spending has in- rural areas vary, reflecting the diver- creased, it i s still insufficient to main- 194 Enhancing the Business Environment tain the existing road network. There- to financial sector development (Beck, Le- fore, as Tanzania scales up investment vine, and Loayza 2000). The analysis o f in infrastructure, an appropriate bal- both the agriculture sector and the data fiom ance needs to be established between the enterprise survey confirm that this find- new investment and maintenance. ing holds true for Tanzania, with a signifi- Building new roads is expensive; ap- cant positive relationship between access to propriate maintenance o f existing finance and enterprise investment and roads i s necessary to ensure the cost- growth. The International Monetary Fund's effectiveness and long-term sustain- "Financial System Stability Assessment" ability o f the infrastructure benefits. (IMF 2003) concluded that the depth and efficiency o f Tanzania's financial system 0 Scaling up investment in infrastructure may have adverse short-term macro- fell well short o fwhat i s needed to help sup- economic consequences. The high do- port economic growth. mestic content o f large-scale infrastruc- ture construction such as roads may The Tanzania investment climate assess- affect relative prices between tradables ment (WED 2004) identified the cost o f and nontradables. In the longer run, financing and access to finance as two o f however, the effect i s likely to be more the main obstacles to enterprise operation thancompensated for by the economy- and growth. Only 20 percent o fthe firms in wide productivity gains fiom the in- the investment climate survey reported hav- vestment. Nonetheless, this potential ing loans from a financial institution. In- downside needs to be managed care- vestment i s financed primarily through re- fully. To avert these short-term "Dutch tained earnings (68 percent o f new disease" effects, those choosing infia- investments). Close to two-thirds o f enter- structure spending need also to consider prises that invested didnot use the financial the import content and the gestation pe- sector. Financing enterprise growth through riod o f the investment. A higher import retained investment results in lower levels content reduces the likelihood and o f investment and, therefore, reduced busi- magnitude o f adverse Dutch disease ef- ness growth andcompetitiveness. fects. Investments with a short gestation Tanzania has undertaken substantial finan- period that result in a quick improve- cial sector reforms since 1991. These re- mentofTanzania's competitiveness can forms included liberalization o f the bank- also contribute to staving o f any nega- ing system to allow the entry o f private tive short-term effects. and foreign banks, enhanced financial sec- tor legislation, and strengthened supervi- Scaling Up Access to Capital sion o f banking by the Bank o f Tanzania. and Finance State ownership o f financial institutions has been significantly reduced with the Access to capital and finance is a critical sale o f the National Bank o f Commerce determinant o f investment and economic and the conclusion o f the first phase of the growth. Countries with well-developed fi- privatization o f the National Micro- nancial systems (banks, stock markets, and Finance Bank, involving the sale o f 49 bond markets) tend to grow faster than percent o f its shares to a consortiumledby countries with less well-developed systems. Rabobank o f the Netherlands. The legal Causation appears to runfrom financial sec- and regulatory environment for microfi- tor development to growth, not fiom growth nance has been reformed, including the 195 Elemenfs of a Strafegy for Shared Growth approval o f new microfinance regulations and 2005. This growth i s primarily attrib- in 2004. In addition, the payment and utable to the rapid increase ininternational clearance system has been strengthenedin reserves heldby the Bank o f Tanzania. By recent years with the establishment o f an contrast, net domestic assets o f the finan- electronic clearinghouse, an electronic cial sector declined from 12 percent o f hnds transfer system, and real-time gross GDP in 1997 to 3 percent in 2003, mainly settlement facilities offered by the Tanza- because o f the sharp reduction in lending nia interbank settlement system. A credit to the government by the Bank o f Tanza- reporting system i s under development, nia. Subsequently, the rapid increase in and a credit bureau has been launched. credit to the private sector led to a recov- Recent reforms o f the land legislation, ery o f domestic assets o f the banking sec- which now provides for landto be used as tor to 10percent o f GDP. collateral, represent important steps in supporting access to credit. Financial intermediation by commercial banks has increased from 14 percent o f These reforms have resulted in a diverse GDP in 1997 to about 17 percent in 2005. financial system and significant changes in The most dramatic developments occurred financial and monetary indicators. The fi- with respect to the structure o f credit. nancial system comprises 21 banks; 9 Overall credit by commercial banks in- nonbank financial institutions; several creased from 9 percent o f GDP in 1997 to pension funds, 2 o f which invest in finan- 15 percent in 2005. This expansion o f cial assets; 14 insurance companies; 63 credit was accompanied by a dramatic foreign exchange bureaus; about 650 sav- change in the composition o f credit. In ings and credit cooperatives (SACCOs); 1997, the bulk o f credit (5 percent o f several other microfinance institutions; GDP) still went to the public sector, while and a stock exchange. Foreign equity par- credit to the private sector amounted to ticipation accounts for about two-thirds o f only 3 percent o f GDP. By 2005, credit to banking system capitalization, and 57 per- government was still about 5 percent o f cent o f total banking assets are in banks GDP, but credit to the private sector had that are majority owned by foreign banks. increased to 10percent. Access to financial services for households Table 10.3 shows trends in the sectoral has declined over the past decade. In the composition o f lending by commercial early 1990s, 19 percent o f all households banks between 1997 and 2005. In 2005, had a member with a savings or current mining and manufacturing and trade ac- account. The restructuring o f the banking counted for almost 50 percent o f bank system, during which a number o f bank lending. Agricultural production and the branches were closed, contributed to a de- transportation sector accounted together cline in access in2001 to only 6.4 percent for another 20 percent. Credit between o f all households and 3.8 percent o f 1997 and 2005 grew at an average annual households inrural areas. rate o f 29 percent. Credit to agricultural production, building and construction, Table 10.2 presents monetary statistics for tourism, and specified financial institu- 1997 to 2005, using gross domestic prod- tions saw the fastest increases. Credit for uct (GDP) as a scaling device. Net foreign marketing and exporting agricultural pro- assets more than doubled between 1997 duce completely disappeared during that 196 Enhancing fhe Business Environment Table 10.2. FinancialVariables as a Percentage of GDP, 1997-2005 Percentageof GDPa Monetary survey 1997 2000 2001 2002 2003 2004 2005 Net foreign assets 8 9 12 13 17 18 19 Net domestic assets 12 8 6 6 3 4 10 Domesticcredit 9 9 8 7 8 9 12 Net claims on government 5 5 3 2 2 1 2 Credit to private sector 3 4 5 5 6 8 10 Other items (net) 3 -1 -2 -2 -4 -5 -2 Liquidity paper (issued by the Bank of Tanzania) 0 -4 -2 -1 -1 Broad money (M3) 20 17 18 19 21 22 29 Currency in circulation 6 5 5 5 4 5 6 Deposits 10 8 9 10 10 11 14 Foreign currency deposits 4 4 5 5 6 6 8 Commercial banks Net foreign assets 5 6 6 6 6 6 6 Foreign assets 5 6 7 6 6 6 Foreign liabilities 0 0 0 0 0 0 Net domestic assets 9 8 8 11 11 Reserves 1 2 2 2 2 2 Domesticcredit 9 10 8 10 10 11 12 Claims on government (net) 5 5 3 4 2 5 Claims on private sector 3 5 5 6 8 10 Other items net -1 -4 -2 -1 -1 5 Demand deposits 4 4 4 5 5 5 8 Time and savings deposits 6 5 6 6 6 6 8 Foreign currency deposits 4 4 5 - 6 - 6- 0- Source: Bank of Tanzaniavarious years. a. End-of-yearvalue as percentage of GDP for the precedingyear. period. It is also worthwhile to note that ings were higher, although still negative in the increase in credit to the private sector most years. The development in the real is not reflected in an increase in private savings rate seem to be closely related to sector investment, because most o f the domestic saving, which also has risen as credit finances working capital. The sec- real interest rates have increased. Interest toral pattern o f increases in credit is also rate spreads have declined in recent years, consistent with the sectoral GDP growth from 18.4 percent in 1997 to 11.5 percent patterns, where the fastest-growing sectors in2005. also experienced the largest increase in credit. Despite these signs o f progress, including growth in lending and increased competi- Figure 10.6 shows real interest rate devel- tion in microfinance, credit to the private opments between 1993 and 2005. Real in- sector remains very small andmostly short terest rates o n savings were around -5 per- term, interest spreads are high, and banks cent until 1998. Subsequently, between accumulate extensive holdings of govern- 1998 and 2005, real interest rates on sav- ment paper and sizable offshore dollar 197 Elements of a Strategy for Shared Growth Table 10.3. Commercial Bank Lending to Some Sectors, 1997-2005 Average growth rate Share in total domestic lending(%) (%I Sector 1997 2000 2003 2005 1997-2005 Public sector 3 2 2 1 14 Agricultural production 8 6 12 12 37 Mining and manufacturing 24 31 26 23 28 Buildingand construction 2 3 5 6 45 Transportation 8 13 9 8 28 Tourism 1 1 2 2 38 Marketing of agricultural produce 1 0 0 0 Exportof agricultural produce 2 0 0 0 Trade in capital goods 0 0 0 0 All other trade 24 26 23 24 29 Specifiedfinancial institutions 0 2 4 6 84 Other 27 13 17 19 23 Total 100 100 100 100 29 Source: Bank of Tanzania various years. Figure 10.6. Real Interest Rates for T-Bills, Lending,and Saving, 1993-2005 20 15 + 10 c P) $ 5 -5 -10 -+-Realinterestrate,T-bills +Realinterestrate,lending Realinterestrate, saving Source: Data from IMF various years. 198 Enhancing the Business Environment placements. Unfinished privatization and, an interinstitutional committee. Imple- most o f all, a number o f structural im- mentation o f the reformprogram involves pediments to lending (including a poor a three-pronged approach: (a) strengthen- credit culture, difficult and slow enforce- ing the lending environment and the fi- ment o f creditor rights, and lack o f suit- nancial infrastructure, primarily through able collateral) are the main factors limit- policy changes and institution strengthen- ing financial intermediation. ing; (b) facilitating the increase inlending to SMEs and long-term lending by com- The system i s ineffective in supplying mercial banks through selective interven- long-term hnds to the private sector. A tions by the government; and (c) giving large proportion o f the long-term liabilities direct support to providers o f financial o f the pension hnds are invested in bank services for micro and small enterprises. deposits, treasury bills, or short-term bonds. As such, one important supply o f Strengthening the lending environment long-term hnds to finance private sector and financial infrastructure. This aspect development fails to be passed through. includes completing the task o f divesting state-controlled entities in banking and A key issue is the distribution o f liquidity. insurance; strengthening the legal and ju- Although there i s a perception o f excess dicial framework that supports lending; liquidity, this is not the case for all the clarifying and deepening the regulatory banks. Ninety percent o f the excess liquid- andinformation and technology infrastruc- ity is concentrated in three banks, and a ture for households and micro enterprises; large part o f this liquidity comes from and encouraging long-term pension and government deposits. Moreover, a large insurance funds to finance longer-term portion o f this liquidity i s invested in gov- private investments. ernment se~urities.~ This situation implies that the commercial banks are not being Facilitating the increase in SME and provided the opportunity to compete for long-term lending. The government is government deposits and that intermedia- undertaking efforts to further improve the tion usingthese funds i s not taking place. availability o f medium- and long-term credit to key sectors o f the economy. With Overall, the picture reflects the low level the assistance o f the World Bank, it i s tak- o f development o f the financial sector in ing initiatives infive areas: Tanzania. In particular, access to credit i s limited to a small number o f enterprises It has launched an SME Credit Guar- with solid collateral in key urban areas, antee Scheme to encourage commer- while small and medium enterprises cial bank lending for SMEs. (SMEs) and firms located outside the main urban areas are virtually excluded. It is expected to launch a development finance guarantee facility, which will To address the challenges, the govern- be managed initially by the Bank o f ment has developed a roadmap for the Tanzania and will provide partial gov- Second-Generation Financial Sector Re- ernment guarantees to commercial form Program. This program includes an banks for their loans to development- implementation plan that i s based on the andexport-oriented projects. recommendations o f the Financial Sector It hopes to facilitate the creation o f a Assessment Program and was prepared by privately owned and managed long- 199 Elements of a Sfrafegy for Shared Growth term financing facility that would chan- ment of links between banks and microfi- nel funds from nonbanks or banks and nance institutions. potentially development partners (with- out government guarantee) to be lent to Achieving the plan. To achieve the ob- commercial banks on a long-term basis. jectives o f the Second-Generation Reform Itwill introduce a development finance Program, the government must fully institution, most likely incorporating commit to timely implementation o f the the Tanzania Investment Bank. The in- action plan. Considerable efforts have stitution would channel multilateral gone into developing the strategies and and bilateral donor finds and perhaps building consensus. The challenge is now use government seed money from the to beginimplementingthe plan. budget, but it would not take any new With continued reform, it is reasonable to deposits from the public. expect the expansion o f credit to continue. It will advance reforms in the pension The solid expansion of credit, both in ag- fund sector in order to unify the legal gregate terms and to a wider range o f and regulatory framework for all pen- small-scale borrowers, can be expected to sion funds, along with investment continue given the growing confidence by guidelines. It is expected that this ef- lenders that the credit environment has im- fort, particularly the development o f proved. Increasingly vigorous competition investment guidelines, will facilitate and the availability of deposit resources the channeling of pension funds' re- point to the likelihoodthat banks will reach sources into longer-term lending down Mher into serving small-scale cli- through commercial banks. ents. These efforts will be supported by continuing the thrust o fthe overall financial Giving direct support to providers of fi- sector policy environment. nancial services for micro and small en- terprises.The SecondGenerationFinancial Enhancingthe Public-Private Sector Reform Plan also includes initiatives Interface related to micro and rural finance. These initiatives are designed to respond to the The public-private interface covers two government's vision for the development o f important aspects o f the business envi- pro-poor finance in Tanzania, as articulated ronment. The first relates to the quality o f inthenationalmicrofrnancepolicy. Some of regulation o f private sector activities. The these initiatives, such as strengthening the second, whose importance for economic regulatory framework, will be addressed as growth has only recently been stressed in part o f strengthening the lending environ- the literature (for example, Rodrik 2004) ment and financial infrastructure. Others relates to the quality o f collaboration and will require direct supportto theproviderso f coordination between the public and the financial services. It is expected that a large private sector in an effort to foster eco- part o f this support will be provided by the nomic growth. Financial Sector Deepening Trust, funded by four bilateral developmentpartners.6 The Regulatory agencies, tax revenue authori- trust will provide assistance for the trans- ties (including customs), business and land formation o fnongovernmental organizations registries, and the judicial system all form focused on microfinance, the strengthening part o fthe public interface with the private o f networks o f SACCOs, and the develop- sector, which has an important bearing on 200 Enhancing the Business Environment - Box 10.1. Aspects of Governance Many`researchers and practitioners have triedto produce aggregate statistics that make it possible to compare the quality o f governance across countries and over time. Few o f these studies cover the entire world or all topics. Furthermore, the questions usedto elicit responses are usually not comparable across surveys. To increase country coverage, Kaufmann, Kraay, and Mastruzzi (2005) combined informationfrom as many as 60 mostly subjective indexes from other sourcesto produce six measures that capture differentaspects o f regulation, corruption, and governance: Voice and accountability. The extent to which citizens o fthe country are able to participate in the selection o f the government. Political stability. The likelihoodthat the government will be destabilized or overthrown by possibly unconstitutional or violent means, includingterrorism. Government effectiveness. The quality o f public service provision and the government bu- reaucracy, the competence and independence o fthe civil service, and the credibility o f the government's commitment to announced policies. Regulatory quality. The quality o f government policies. This measure focuses on the preva- lence o f market-unfriendly policies, such as price controls or inadequate bank supervision, and on perceptions o fthe regulatory burdenfacing businesses. Rule of law. The extent to which individuals have confidence inand abide by the rules o f so- ciety. This measure includes perceptions about the incidence o f crime (both violent and non- violent), the effectiveness andpredictability o f the judiciary, and the enforceability o f con- tracts. Control of corruption. The extent o f corruption (that is, the illegal use o fpublic power for private gain). the costs, risks, and barriers to business in nities, and ensuring that the provision o f Tanzania. The costs influence the range o f public goods and services (especially infka- opportunities that are profitable. Because structure) is well aligned with private sector investments are forward looking, risks and needs. An efficient public-private interface uncertainty determine the types and nature i s part o f the second-generation reforms o f investments. Entry restrictions limit in- that will determine the private sector's re- novation and the efficient provision o f sponse to productivity-based opportunities. goods and services. Factors that affect the public-private inter- The quality o f the public-private interface face overlap with the broader concept o f also i s critical for ensuring that the gov- governance. Box 10.1 describes the differ- emment can play a supportive role in ent aspects o f governance: voice and ac- growth efforts led by the private sector. countability, political stability, govern- This role relates to the flow o f information ment effectiveness, regulatory quality, rule between the private and the public sector, o f law, and control o f corruption. Figure allowing the government to play a suppor- 10.7 shows that for each o f these aspects tive role by removing obstacles, collaborat- Tanzania performs better than other low- ing inthe identification o f growth opportu- income countries. For all indicators except 201 Elements of a Strategy for Shared Growth Figure 10.7. Governance Indicatorsfor Tanzania, 1996-2005 (a) Comparisonbetween1996and 2005 TFlNZRNIFl (2005) Volce and Rccountabllity Political Stabilltylwo Violence tovernnent Effectiveness Regulatory Puallty Rule of Lau Control of Corruption 0 21 7s 1 9 Cantry's PercentileConpsison betueen 2005. 1996 (bop-botta order) Rank (0-100) (b) Comparisonwith low-incomecountryaverage,2004 TFlNZFlNIFl (2005) Voice and Rccountabllity Political Stabllity/No Violence Governncnt Effectlvoness Regulatory Quality Rule o f Lau Control of corruption e 25 50 1 e Conparison with inwne category averawe (Lou Inconc) (louer bar) Country's Pcrcsntlle Rank (0-188) Source: Kaufmann, Kraay,and Mastruzzi 2006. Note: The panel (b) depictsthe percentilerank on each governance indicator.Percentilerank indi- cates the percentageof countriesworldwidethat rate below the selectedcountry (subjectto margin of error). The selected comparator(if any) is depicted by the lower bar for each governanceindica- tor. Inthe bar chart, the statisticallylikelyrange of the governance indicator is shown as a thin black line. For instance,a bar of length75 percentwith the thin black lines extendingfrom 60 percent to 85 percent has the following interpretation:an estimated 75 percent of the countries rateworse and an estimated 25 percentof the countries rate better than the country of choice. However,at the 90 percent confidence level, only 60 percentof the countries rate worse, whereas only 15 percentof the countries rate better. Highervalues imply bettergovernance ratings. 202 Enhancing the Business Environment political stabilityho violence, Tanzania interchange systems, and border coopera- registered improvements between 1996 tion agreements. and 2005. With respect to control of cor- ruption, Tanzania moved from being per- The indicators reported inDoing Business ceived as among the most corrupt coun- 2007 provide an objective assessment o f tries to a position in which about 29 the regulatory environment. Information percent o f the 200 countries assessed are from the investment climate assessment perceived as doing worse than Tanzania in carried out in2003 (PED 2004) provides controlling corruption. The slight deterio- complementary information on the extent ration in Tanzania's rating o f political sta- to which weaknesses in the regulatory en- bilityho violence i s likely related to con- vironment constrain enterprise growth. tinuedpoliticaltensions betweenthe ruling andthe main opposition party. Barriers to Entry The World Bank's annual Doing Business Although some entry restrictions are nec- report provides an objective assessment o f essary-requiring enterprises to register the regulatory environment in 175 coun- with tax authorities, for example-others tries by measuring the number o f proce- dures, time required, and cost imposed on do more harm than good. Several recent businesses through government regulation. studies have shown that entry restrictions encourage firms to remain informal by In 2006, Tanzania ranked 142 among the making it difficult and expensive to enter 175 countries, which suggests that doing the formal sector (see figure 10.8).7 Entry business in Tanzania i s handicapped by restrictions can also lead to corruption severe weaknesses in the regulatory envi- and reduce competition, thereby hurtin ronment. The business licensing regime, economic performance in other ways. !i employment regulations, and registration Within the Organisation for Economic o f property are particularly problematic Co-operation and Development (OECD), areas. The situation i s somewhat better for example, multifactor productivity and with respect to the administrative cost and investment are significantly lower in procedures related to trading across bor- sectors that have greater restrictions on ders and the enforcement of contracts, al- entry. though even inthese areas there i s signifi- cant scope for improvement in order to Most enterprises in Tanzania did not rate enhance the attractiveness o f Tanzania for business licensing as a serious obstacle to investment. However, Doing Business enterprise operations and growth. Only 2007 (World Bank 2006) also recognizes about 27 percent o f enterprises said that it the efforts Tanzania i s undertaking to im- was a major or very severe problem. But prove its business environment. Indeed, these results probably understate the effect Tanzania i s assessed as being among the o f such regulations. Businesses that are top 10 reformers in 2005/06. Recent re- already operating are much less likely to forms include simplification o f the busi- view entry restrictions as an obstacle. In- ness licensing regime; reduction in the deed, they may even welcome regulations cost o f registeringproperty; revision o f the that limit competition. Companies Act to give investors greater protection; and reform o f customs admini- According to World Bank data on the cost stration through the introduction o f risk o f doing business," in2006, it took about management techniques, electronic data 30 days to fulfill all legal requirementsfor 203 Elements of a Strategy for Shared Growth Figure 10.8. Informality in Tanzania and in Comparator Countries Tanzania China India Kenya Uganda Source: Data from Schneider2006. starting a business in Tanzania (table lays (30 days) than large enterprises (21 10.4). Although the same process takes days) and very large enterprises (3 days). five days or fewer in some OECD coun- Obtaining a water connection also took tries, such as Australia, Canada, or Den- longer inTanzania than inany o f the com- mark, or the United States, this period i s parator countries, but the mediandelay-7 not exceptionally burdensome when days-was far shorter than the wait for compared with the time required in other electricity. Getting a telephone connection developing countries. For example, busi- was somewhat more time consuming, but ness licensing takes about 35 days in the medianwait o f 14 days was considera- China and India, 54 days in Kenya, and bly shorter than inKenya (60 days). 30 days inUganda. But the monetary cost o f this process as a percentage o f per cap- Business Regulations and Inspections ita gross national income (GNI) i s higher in Tanzania (92 percent) than in some of Business regulations and inspections may the comparator countries. In contrast, impose significant cost on businesses and business licensing costs 9 percent o f per encourage enterprises to operate infor- capita GNI in China, 74 percent in India, mally. Tanzania's licensing regime i s 46 percent in Kenya, and 114 percent in among the most burdensome, according to Uganda. the Doing Business 2007 data (World Bank2006). For example, obtaining all the Legal restrictions are not the only banier permits to build a warehouse involved 26 to entry that new firms face. Access to in- procedures that requiredtogether 313 days fiastructure-especially electricity-is with a cost o f 3,797 percent o f GNI. In another important obstacle. The median most other countries, fewer procedures wait for an electricity connection was 30 that consume less time are required, and days inTanzania, compared with 3 days in the cost for obtaining the requiredpermits China, 14 days in Uganda, and 15 days in i s significantly lower. InKenya, the cost i s Kenya. Micro, small, and medium-size only 38 percent o f GNI, and inThailand it enterprises faced much longer median de- i s only 11percent. 204 'D a m m b c o a! gC 8 S h h 0 V 9 0 Q ) W N c9 o o o h o m m o o o h m t m 7 0 7 r m m N rr r N 5 7 r c o N 8 .-m 0 r d L 5a 0 v) U -.-em S m I m h N.. d N . I N m o d m r 0 InFa.5 ( u c o h r n d c \ j c o t o V\ j ( c :ah N o oNororm c o r c o z m m to Y 7 >r I m m d c b0 r h m F O G b r coN00 VNZI E co r r n d r n 0 O c o Vd cohn+ O N m N 7 i2 m C I h Q) I rn 8 z .-m cw S C 205 206 Enhancing the Business Environment Figure 10.9. Percentage of Firms Inspectedand Number of Inspections per Year, by Government Agency 100% 92% 80% 66% 60% 50% 52% 4.5 48% 40% 20% 0% Tax Labor ai Fireand Sanitation Municipal Environment social building police security safety Government agency I W Firms reportinginspections/requiredmeetings 0Mediannumberof meetings I Source:World Bank 2004c. Note: Numberof visits is the mediannumber for firms with inspections. According to the information from an en- (47 and 20, respectively) than for nonex- terprise survey (World Bank 2004c), on porters (28 and 14). Figure 10.9 shows average enterprises in Tanzania reported the types o f inspections that may take that senior managers spent about 15 per- place. cent o f their time dealing with govern- ment inspections, regulations, and paper- Overall, inspections do not appear to vary work. This figure was slightly higher than greatly by sector. The vast majority o f es- inKenya (13 percent) and China (12 per- tablishments are inspected by government cent) and significantly higher than in agencies. For example, 92 percent of Uganda (4 percent). Similarly, the median manufacturing firms, 85 percent o f tour- enterprise inTanzania reported 15 inspec- ismestablishments, and 71 percent o f con- tions or meetings with government agen- struction firms reported tax inspections cies (including tax officials) per year, (table 10.5). But firms in the construction significantly more than in Uganda and sector were less likely to be inspected than India (5 and 6 inspections, respectively). enterprises in other sectors, while firms in However, enterprises in China and Kenya the manufacturing sector were most likely each reported a slightly greater number o f to be inspected. For firms that were in- inspections (16). Finally, the bureaucratic spected, the average number o f inspections burden associated with regulation i s much was similar across sectors. For example, higher for exporters. The numbers o f in- the median firm in all sectors was in- spections and days spent dealing with spected seven times by the tax inspectorate regulatory issues are higher for exporters in2002. 207 Elements of a Strategy for Shared Growth Table 10.5. Percentageof Firms Inspectedand Median Number of Inspections, by Sector Type of inspection Laborand Fire and social building Municipal Tax security safety Sanitation police Environment Percentaae of firms with insDecfions ConstrucFion 71 30 19 19 21 19 Tourism 85 58 23 48 18 32 Manufacturing 92 66 50 52 31 48 Median number of inspections for2firms with inspections Construction 7 2 3 5 2 Tourism 7 3 1 3 5 2 Manufacturing 7 4 2 2 5 2 Source: World Bank 2004c. Customsand Trade Regulations days, respectively, inKenya and seven and three days, respectively, inIndia. Exporting and importing in Tanzania are both hampered by poor customs admini- Poor customs administration and overly stration. Among enterprises that engaged restrictive trade and customs regulations in foreign trade, the median firm re- discourage enterprises from exporting. ported that it took 14 days on average for Indeed, as noted previously, trade and imports to clear customs and 7 days for customs regulation partially explain why exports, once they had reached the point enterprises in Tanzania export less than o f entry or exit. l1The median wait was similar enterprises in Kenya. Because ex- considerably shorter than the average porting has been linked to improved pro- typical wait because several enterprises ductivity-as well as an improved bal- reported exceptionally long delays.l2 ance o f trade-these delays and Managers also reported that clearance restrictions can have a real effect on en- times were unpredictable, forcing them terprise performance. to hold additional inventory in anticipa- tion o f worst-case scenarios. For im- Land Registration ports, the median enterprise reported that the longest delay was 21 days- Land constitutes one o f the most important considerably longer than the typical 14- asset bases for investments. The provision day wait. For exports, the median enter- o f information on land in the form o f a ti- prise reported that the longest delay was tle and a certificate o f occupancy or own- 12days. ership can transform potential assets into tradable assets o f capital. However, the Port and customs delays are considerably value o f these assets and their economic longer inTanzania than in any o f the com- potential is ofien jeopardized by the inse- parator countries. The median delays for curity o f property rights. The transaction imports and exports in China were less costs o f using land as collateral are high. than half as long-five and three days, re- The title registration system i s inefficiently spectively. Similarly, reported delays for administered and maintained, and poor imports and exports were seven and four security o f the physical files has provided 208 Enhancing the Business Environment opportunities for fraudulent and corrupt the Legal Sector Task Force in 1996. activity, compromising the integrityo f the Since then, the government has designed title registry. A poorly functioning land strategies for implementation to ensure registry makes it costly to verify the status that priority needs in the legal sector are o f a land title, which in turn affects the addressed accordingly. Problems identi- ability to sell land or associatedreal estate. fied by the Legal Sector Task Force, Existing title holders and banks have re- which are widely acknowledged, include ported that it takes an average o f three inordinate delays in resolving disputes months to register a title for mortgage in and dispensing justice in the justice sys- the landregistry. tem, inaccessibility of the justice system for the majority o f poor and disadvan- Legal System taged Tanzanians, low levels o f public trust in the justice system, and excessive Since embarking on economic reforms in prevalence o f unethical behavior (United the mid-1980s, Tanzania has witnessed Republic o f Tanzania 2004b). The key significant social and economic changes, challenges identified for the sector in- including the emergence and growth o f a clude the rapid social, political, eco- vibrant private sector that has been able to nomic, and technological changes taking stimulate economic growth and increase place in the country and internationally the country's competitiveness, as well as and, in particular, the need to rapidly de- to provide a variety o f goods and services velop a legal system to facilitate the effi- to citizens-previously the domain o f the cient development o f a private sector, state or state-run institutions. However, market-led economy and provide due pro- reform o f the country's legal sector to suf- tection o f consumer rights. ficiently support and enhance economic growth and efficiency led by the private Taxation sector has not taken place at the same pace as economic reform. The present legal sys- Tax administration was the area o f regula- temis mainly basedon Englishlaws intro- tion that enterprises were most likely to duced when Tanganyika was a British identify as a serious problem, more than mandate; in many aspects, it is outdated. half o f them rating it as a major or very Inaddition, a substantial part of the legis- severe obstacle. Some 92 percent reported lation and the institutions, even where sub- that they had required meetings with tax sequently amended, were fashioned for the inspectors; the median firm within this centrally planned economy and therefore subset reported seven meetings per year. cannot be efficacious in a competitive Fewer enterprises reportedhaving required market system. With globalization, re- meetings or inspections from other agen- gionalization, and technological advances, cies, and the number o f requiredmeetings the problem is more compounded as- was lower for those that did. inevitably-new demands are placed on the economy and the conduct o f its agents, Almost 73 percent o f enterprises in Tan- creating the need for a more conducive zania rated tax rates as a major or very environment for competitiveness. severe constraint on enterprise perform- ance and growth-considerably more Reform o f the country's legal sector be- than rated any other obstacle as a major gan in 1997 when the government ac- constraint and more than in any o f the cepted the recommendations submitted by comparator countries (figure 10.10).l3 209 Elements of a Strategy for Shared Growth Figure 10.10. Rating of Problems by Enterprisesin Tanzania and in ComparatorCountries 80% 1 73% 70% 68% 60% 50% 40% 30% 20% 10% 0% Tanzania China Uganda Kenya Source:World Bank 2004c. Among the countries where investment both exporters and nonexporters. In addi- climate assessments were completed by tion, each group ranked tax administration the end of2003, in only two (Brazil and among the top obstacles it faced. Ethiopia) were enterprises more likely to rate tax rates as a serious obstacle. Fewer The data from Doing Business 2007 enterprises in Tanzania (58 percent) rated (World Bank 2006) put the information tax administration as a serious problem. from the enterprise survey into perspec- However, tax administration ranked high tive. InTanzania, entrepreneurs there must among enterprises' serious concerns, fal- make 48 payments, spend 248 hours, and ling just below tax rates, electricity, and pay 45 percent of gross profit intaxes. Al- costs o f financing. though the number o f payments i s higher inTanzania thanthe average for the region Problems with tax rates and tax admini- and for OECD countries (41 and 15, re- stration were not unique to any one group spectively), the amount o f time spent and o f enterprises. In manufacturing, small, the effective tax rate are not too different subset reported seven meetings per year. from the regional and OECD averages Fewer enterprises reported having required (336 and 203 days and 71 and 48 percent, medium, large, and very large enterprises respectively). This information suggests all rated tax rates as the greatest obstacle that while entrepreneurs (not unexpect- to enterprise operations and growth (that edly) feel that taxation i s a major burden is, tax rates were the obstacle rated as ma- on their businesses, the system i s not sig- jor or very severe by the largest number o f nificantly more burdensome than that o f enterprises in each group). Enterprises in many other countries. tourism and construction also rated tax rates as a serious problem, as did foreign- lnformalify and evasion. It is difficult to owned and domestically owned firms and draw strong conclusions about tax eva- 210 Enhancing the Business Environment sion from investment climate surveys- reduce costly tax exemptions and incen- enterprise managers being unlikely to be tives (IMF 2003). Another would be to forthcoming in this respect. However, the encourage informal firms to enter the for- evidence from the survey i s consistent mal sector-and in so doing broaden the with the macroeconomic evidence. Eva- tax base-by reducing the regulatory bur- sion appears to be a greater problem in den on formal enterprises. Tanzania than in the comparator coun- tries. Managers in Tanzania estimated As a first step inimproving revenue mobi- that the typical firm in their area o f activ- lization, the government must improve tax ity reported 69 percent o f its sales for tax administration, which remains problematic purposes.l4By contrast, the average esti- (IMF 2003). For example, the value added mates inUganda and Kenya were 77 per- tax (VAT) efficiency ratio (the ratio o f cent and 86 percent. The high level o f VAT revenues to GDP divided by the evasion i s especially striking given that VAT rate) in Tanzania (0.20) lags the av- enterprises in Tanzania face a greater erage for Sub-Saharan Africa (0.27). The number o f tax inspections and required government has already taken some steps meetings than enterprises in most o f the to strengthen administration, including the comparator countries. establishment o f the Large Taxpayer Unit in2001, the introductiono ftaxpayer iden- Estimates o f amounts reported for small, tification numbers in 2000, and the five- medium, large, and very large enterprises year corporate plan for the Tanzania ranged between 68 and 83 percent o f Revenue Authority. sales. Reporting rates were highest among very large enterprises and lowest among Tax administration. Improving collec- micro enterprises (about 25 percent). tion should not be the only goal o f tax Consistent with the observation that in- administration reform. The government formal micro enterprises were far less should also reduce the burden that tax concerned about tax rates and administra- administration imposes on enterprises. tion than other enterprises, they reported Enterprise managers in Tanzania reported the smallest share o f sales to tax authori- that in 2002/03, they spent about seven ties-19 percent, compared with 28 per- days dealing with inspections or required cent for formal micro enterprises. The meetings with tax officials. In compari- relatively small gap betweeninformal and son, enterprise managers in China, formal micro enterprises may reflect the Kenya, and Uganda reported spending fact that informality i s a continuum, with two to three days on such tasks. Despite many enterprises being informal to some recent reforms, managers generally did degree. not report any reduction in this adminis- trative burden. Although 36 managers re- One approach that the government could ported fewer meetings with tax officials take to combat evasion would be to reduce in 2002 than in 2001, 59 managers re- the tax burden on formal enterprises. ported more meetings, and 149 reported However, since the government's fiscal the same number. discipline has underpinned Tanzania's im- pressive macroeconomic performance, rate The burden o f tax administration i s par- cuts would need to be combined with steps ticularly high on more productive firms, to improve revenue mobilization. One way which tend to be larger and are more for the government to do this would be to likely to be both foreign-owned and ex- 211 Elements of a Strategy for Shared Growth porters. This acts as a tax on efficiency. since 1996, when a presidential commis- In addition to discouraging firms from sion led by Judge Joseph Warioba pro- becoming more efficient, it also discour- duced the Warioba Report.17 It i s imple- ages them from taking actions that might mentinga national anticorruption strategy signal their efficiency, such as entering and has strengthened the institutional export markets. framework-notably through the Finance Act o f 2001 and the Public Procurement Inaddition, the government should com- Act o f 2002-and adopted a clear zero- bat corruption in tax administration. De- tolerance position on corruption (ESRF spite recent reform efforts, 21 percent o f and FACEIT 2002). Nonetheless, corrup- enterprises that had required meetings tion continues to have an important effect with tax inspectors reported that gifts or on businesses. informal payments to inspectors were expected or requested. l6The correspond- Grand corruption. Enterprise managers ing figures for the comparator countries in Tanzania see grand corruption- were 7 percent in Uganda, 21 percent in payments made to policy makers or senior China, and 38 percent in Kenya. The bureaucrats to win government contracts median value o f the gift or informal and influence lawmaking-as a serious payment was T Sh 400,000 (about problem (figure 10.11). Approximately 45 US$400 inmid-2003). percent o f managers said that payments to government officials that affected the con- Corruption tent o f government decrees had at least a moderate effect on their businesses, while The government appears to have made 47 percent said the same o f payments to some progress in reducing corruption members o f parliament that affected their Figure 10.11. Enterprises Reporting That Bribes to the Government Affect Their Businesses in Tanzania and Comparator Countries rn 60%- Tanzania Uganda Pakistan Kenya Source:World Bank 2004c. 212 Enhancing the Business Environment Table 10.6. Likelihoodof ReportingThat BribesWere Neededto Get Things Done, by Size of Enterprise Size of enterprise Micro Small Medium Large Very large Percentage reporting bribes neededto "get things done" 23 30 44 41 36 Percentage reporting gifts or pay- ments required Mainline telephone connection - 26 27 14 0 Electricalconnection 27 32 35 27 14 Water connection 33 22 24 13 0 Import license - 0 19 12 9 Operating license 34 15 30 14 6 Source: World Bank 2004c. Note: -=not available. votes. These figures were higher than in Petty corruption. Enterprises are also Kenya (which reported 40 percent and 26 affected by petty corruption-payments percent, respectively) and Uganda (which made to lower-level government officials reported 30 percent and 20 percent, to "get things done" in connection with respectively). customs, taxes, licenses, and other ser- vices. About 35 percent o f enterprise Informal payments to secure government managers said that informal payments contracts also appear common in Tanza- were typically needed for firms like nia. Of enterprises that did business with theirs. Of the enterprises that reported the government, about 33 percent re- that informal payments were needed, the ported that unofficial payments were median payment was about 0.3 percent o f needed to secure government contracts. sales. Bribes were fairly common for Within the subset o f firms reporting that many transactions, such as getting utility payments were needed, the median firm connections and applying for import li- reported that about 10 percent o f the con- censes (table 10.6). tract value was needed to secure govern- ment work. This result was similar to the Informal payments and gifts were also median percentages reported in Uganda common during inspections and manda- and Kenya (both 10 percent), but signifi- tory meetings with government officials. cantly higher than the median percentage About 21 percent o f enterprises that had reported in China (2 percent). The inspections or meetings with officials amount paid in bribes i s strongly corre- from the Tanzania Revenue Authority lated with firm performance: firms with reported that bribes were requested, with faster sales growth and higher productiv- the median amount being T Sh 375,000 ity pay larger bribes. Because corrupt of- (figure 10.12). Similarly, 25 percent o f ficials often target the most productive enterprises dealing with the municipal and profitable enterprises, corruption can police reported requests for bribes. The act as a tax on efficient firms. median payment in that case was 213 Elements of a Strategy for Shared Growth Figure 10.12. Requestsfor Bribes during Tax Inspections 30% 400 U)- a = 20% 300 200 ee2 226 10% 100 o m 5; 0% 0 Government agency ~ Source: World Bank 2004c. Note: "Percentagereporting bribe requests"includesonly enterprisesthat were inspected."Me- dian amount"only includesenterprisesthat report positiveamounts. T Sh 80,000. About 18 percent o f enter- bribes were needed to get things done, prises reported requests from labor or so- compared with 38 percent o f formal mi- cial security officials, 13 percent from cro enterprises. health and sanitation inspectors, 8 percent from environmental inspectors, and 7 Although micro enterprises were less percent from fire and building safety in- likely to report that bribes were needed spectors. Payments to these other inspec- to get things done, they were no less tors were generally lower than payments likely to report that bribes were needed to tax officials. for typical transactions. For example, 34 percent o f micro enterprises reported that Micro enterprises are less likely than lar- bribes were needed to get an operating ger enterprises to find corruption a major license, compared with 15 percent of or very severe constraint. Only 18 percent small, 30 percent o f medium, 14 percent o f informal micro enterprises and 30 per- o f large, and 6 percent o f very large en- cent o f formal micro enterprises found terprises. The most plausible explanation corruption to be a serious problem, com- is that micro enterprises are far less pared with 48 to 58 percent o f small, me- likely to interact with bribe-taking insti- dium, and large enterprises. Micro enter- tutions but at least as likely to encounter prises were particularly unlikely to face corruption when they do. For example, requests for bribes. Only 15 percent o f only 5 percent o f micro enterprises re- informal micro enterprises reported that ported getting an electricity connection 214 Enhancing the Business Environment Figure 10.13. Interactionwith InstitutionsThat Demand Bribes: Micro Enterprisesversus Small, Medium,and Large Enterprises 67% CI Mainline Electrical Water Import license Operating C Q) telephone connection connection license rc s0 connection Source: World Bank 2004c. inthe two years before the survey, corn- were also less likely to get operating li- pared with 30 percent o f larger enter- censes, water connections, and fixed-line prises (figure 10.13). Micro enterprises telephone connections. Notes Canada, the Netherlands, Sweden, and the United Kingdom. 'These estimates are drawn fiom Chandra, Kacker, 'Using data from the World Bank's Doing Busi- andLi(2005). ness database for 85 developing and developed The ex post rates o f return o f the selected infra- economies, Djankov and others (2002) find that structure projects should be interpreted in the con- informality i s greater in countries with greater bar- text o f the policy and institutional environment riers to entry. during the particular time period. It is also impor- Broadman andRecanatini (2001) findthat cormp- tant to note that economic rates o f return may not tion i s higher in transition economies, where entry fully capture all the benefits o f infrastructure in- barriers, which they measure using subjective data vestment, especially if those benefits occur inform from the World Business Environment Survey, are o f externalities (see Canning andBennathan2003). higher. Djankov and others (2002) find similar The preliminary results o f this study are con- results using different data. tained inFan,Nyange, andRao (2005). Using data from 17 high-income OECD econo- The core road network required for poverty alle- mies for several manufacturing and business ser- viation (that is, the network that provides reliable vice industries between 1984 and 1998, Scarpetta access for roughly 90 percent o f the population) is and Nicoletti (2003) show that multifactor produc- 45,000 kilometers long. This figure represents tivity is lower in sectors where regulation is more than half o f the existing network- stricter. Alesina and others (2003) present results emphasizing the importance o f roads for poverty for investment. Because the measures o f regulation reduction. are sector specific and are measured over time, 'It i s estimated that in April 2005, almost 50 per- both this study and Scarpetta and Nicoletti (2003) cent o fthe T-billswere heldby two banks. control for country- and sector-specific effects that mightaffect investment andproductivity. 215 Elements of a Strategy for Shared Growth lo Data are from the World Bank's Doing Business possible had the agency remained part o f the civil database. service (Fjeldstad 2002). l1 World Bank (20054 %ked enterprise managers See ESRF and FACEIT (2002) for a general how long it takes from the time that goods arrive at l7 summary of the Warioba Report and progress the Point of entry O rexit to the time that they clear against comption Over the preceding decade. customs. It did not ask how much time was attrib- utable to customs processing and how much was attributable to port operations because managers typically do not have access to this information. However, for the most part, the problems in Tan- zania appear to be related to customs, rather than port performance. For example, the Economist Intelligence Unit (2004c, 28) notes, "One o f the more successful privatizations carried out by the government is that o f Dar es Salaamcontainer port. This is now being run by Tanzania International Container Terminal Services (TICTS), a company that is financially controlled by a Hong Kong- based company.. .. While TICTS has benefited from the long-term project embarked on by the government in dredging the port's entry channel, its own investment in facilities has been a key fac- tor inimproving services and it now has the capac- ity to handle 250,000 containers annually." l2 For example, nine enterprises reported typical delays o fmore than 50 days. l3 For example, enterprises inTanzania were more likely to rate high tax rates as a major problem than firms in almost two-thirds o f the countries in the 1999 World Business Environment Survey. The survey asked firms in all surveyed regions about infrastructure, access to finance, policy in- stability and uncertainty, inflation, exchange rates, the functioning o f the judiciary, street crime, and organized crime. The survey also asked about regulatory constraints, including high tax rates and tax administration. l4 Enterprise managers were asked about the "typi- cal fm" in their industry rather than their own firm so that they could avoid implicating them- selves (WED 2004). Is It may also reflect the way that the question was asked. If registered enterprises felt that unregis- tered enterprises in the same sector and o f similar size were an "enterprise like yours," they may have includedthose enterprises intheir responses. l6 Reforms included the establishment o f the inde- pendent Tanzania Revenue Authority in 1996, and measures designed to limit political interference in tax administration andto allow the authority to pay salaries that were higher than would have been 216 CHAPTER 11 Harnessing Natural Resources for Sustainable Growth The Contribution of Natural Resourcesto Growth Natural resources in Tanzania constitute a o f decline in the catch per boat in Lake wealth asset. Since 1996, mining, fisher- Victoria and the catch o f fish and prawn in ies, and tourism have been the most dy- the coastal zones, which point toward a namic sectors in the economy. Although deceleration o f growth in the medium and tourism development i s a success story in long terms. macroeconomic terms, local development spinoff effects could be explored more Forestry, wildlife, and marine fisheries fully. Most known mineral deposits are resources, though declining, are still rela- already being tapped, but new mineral tively abundant and have largely untapped stocks are being discovered. The fisheries growth potential. Although these natural sector i s still growing, but there are signs resources, like labor and capital, contribute 217 Elements of a Strategy for Shared Growth to the economy and the subsistence baseo f (GDP) more fully, giving equal weight to the ruralpopulation, their value andpoten- the need for pro-poor growth and the tial i s underestimated. This underestima- maintenance o f a sustainable natural re- tion i s partlybased on missingmarkets for source base. Because o f policy failures, public goods, imperfect competition, dis- Tanzania's natural resource endowments tortions caused by government interven- are not harnessed in an optimal way to tions, and pricing o f natural resources be- achieve both economic growth and pov- low market value. The result of all these erty reduction. market failures leads to suboptimal eco- nomic decision making and loss o f income On the contrary, owing to weak govern- to the country. ance regimes in revenue-generating sec- tors, resources are offered below market The National Strategy for Growth and Pov- price to the benefit o f a few powerful erty Reduction o f 2005 subscribes to the winners and the loss o f the majority o fthe principles o f sustainable and equitable de- rural population. Yet these natural re- velopment. The operational starting points sources provide substantive potential for o fthese principles include the following: income to communities in rural areas. The weaknesses ingovernance regimes in 0 Renewable resources should be ex- forestry, wildlife, and fisheries include ploited on the bases o f maximizing primarily (a) the lack o f transparency and profits and sustaining yields. Resources accountability in issuing rights to extract should not be driven to extinction, re- resources and accrue revenues from them, gardless o f the dictates o f present value (b) inequitable sharing o f benefits with maximization. Hence, harvesting rates communities, and (c) weak monitoring should not exceed regeneration rates and surveillance o f stocks. In all four and waste emissions should not exceed principal sectors providing natural capital absorptive capacities. in the growth equation-forestry, wild- life, fisheries, and mining-royalties are 0 Nonrenewable resources should be ex- ploited at a rate equal to the creation o f set arbitrarily and do not reflect scarcity. renewable substitutes. Revenue from Royalties are, hence, not used as a policy the exploitation o f nonrenewable re- instrument of intertemporal resource pric- sources should contain an income ing and sustained yield management. component and a capital component. As long as these weaknesses are not ad- The capital component should be used dressed, a substantial base o f economic to invest in building up a new renew- growth will slowly erode and poverty re- able asset to replace the nonrenewable duction objectives are unlikely to be one at the point o f its exhaustion. achieved. 0 Revenue generated fiom natural re- sources should be shared equitably, in Contributions,Accounted and Unaccounted for particular with the rural comqunities on whose landthese resourcesare located. Commonly for forestry, wildlife, and fisheries, a great share o f the economic The macroeconomics o f sustainability re- contribution does not enter GDP and ex- quire integrating qualitative development port statistics and i s hence not taken into and growth in gross domestic product account in analyses o f growth. A general 218 Harnessing Natural Resources for Sustainable Growth Table 11.1. Ministry of Natural Revenues and Tourism Annual Revenue, 2002103 and 2003104 Amount of revenue (T Sh billion) Revenue source 2002103 2003104 Forestry 5.29 5.82 Wildlife 9.17 9.55 Fisheries 6.99 9.70 Tourism 0.83 0.96 Source: Ministry of Natural Resources and Tourism 2004. Note: Amounts include revenue collected and retained at source. problem i s the unavailability and poor ing and fish smoking. Forests provide wa- quality o f data. tershed functions for major rivers feeding into the national hydropower dams. The An overview o f annual revenue earned by lack o f reliable power and water supplies the Ministry o f Natural Resources and can hamper growth inthe long term and i s Tourism (MNRT) from its key depart- already being cited as a serious constraint ments in 2003 and 2004 is presented in inattracting private investment. table 11.1. Although revenue i s an impor- tant measure for growth, it does not cap- Further to their "source" functions, forests ture all contributions to economic and ru- also have "sink" functions, absorbing and ral development by the respective sectors. neutralizing the negative externalities o f economic growth-most importantly pol- Forestry provided more thanT Sh 5 billion lution. The value o f carbon sequestration ingovernment revenue in2003 and 2004, services provided by Tanzanian forests is as table 11.1 shows. It officially contrib- estimated to be between US$700 and utes 2 to 3 percent to GDP and 10 to 15 US$1,500 per hectare. Additional envi- percent to export earnings. Estimates that ronmental service functions include inputs include unaccounted-for services and non- from land and forests into agricultural industrial forestry reach 10 to 15 percent production. o f GDP. Revenuegenerated from wildlife resources Forests provide about 75 percent o f build- accrues to the MNRT mainly from hunting ing materials and 100 percent o f indige- licenses; it was more than T Sh 9 million nous medicinal plants and supplementary during 2003 and 2004 (table 11.1).An in- food products. In addition, forests provide dependent study (Baldus and Cauldwell an important component o f value added to 2004) o f the sector cites annual earnings in national income through their ecosystem 2001 o f about US$30million from tour- service functions-providing for industrial ists' hunting and an additional US$9 mil- and domestic water and energy supply. lion generated by the private companies Some 95 percent o f Tanzania's energy that lease hunting concessions from the consumption i s in the form o f fuelwood; government. In 2002, earnings from live that consumption includes major input fac- animal exports amounted to roughly tors inrural industries such as tobacco cur- US$170,000. 219 Elements of a Strategy for Shared Growth Table 11.2. Budget of Ministry of Natural Revenues and Tourism, as Distributed by Subsector, 2002103-2003104 2002103 2003104 ~~ Subsector T Sh thousand Percent T Sh thousand Percent Forestry and beekeeping 4,897,656 24 7,633,912 29 Wildlife 6,593,025 33 7,586,736 29 Fisheries 3,688,280 18 4,648,202 18 Tourism 2,208,073 11 2,880,761 11 Other 2,856,131 14 3,507,741 13 Total 20,243,165 100 26,257,352 100 Source: Ministry of Natural Resources and Tourism 2004. Note: Total includes other subsectors not listedhere.The amounts are only recurrentexpenditures. The largest income earner is the noncon- A great share of the marine catch does not sumptive use o f wildlife resources: game enter GDP and export statistics but plays viewing by international tourists. In 2001, an important role in livelihood support. Tanzanian national parks drew more than The official number o f artisanal fishermen 100,000 international visitors. This tourism has doubled since 1995, reaching close to generated receipts o f almost 5 percent o f 115,000 in2005. GDP, equivalent to aboutUS$400 million. Although the contribution o f mining to Inaddition, wildlife provides unaccounted- GDP was still not more than 2.8 percent in for subsistence values. Well over two-thirds 2005, it i s the single most important earner o f the people eat wild game, with up to 95 o f foreign exchange. About 50 percent o f percent o f the ruralpopulation claiming it as export earnings accrue from minerals, pre- their most important meatprotein source. dominantly from gold mining by large- scale, foreign-owned operators. In addi- Tanzania's production inthe fisheries sec- tion, mineral resourcesare important to the tor has grown at 4 percent annually be- artisanal miningsector. tween 2000 and 2005. In 2005, the value o f caught fish amounted to T Sh 339 bil- Public Investment in Natural lion compared with T Sh 78 billion in Resource-Based Growth 2000. About 75 percent o f revenue comes from freshwater fisheries, and only 25 It is obvious from these data that these percent from marine fisheries. However, natural resource-based sectors make an the number o f foreign vessels licensed to important contribution to both the formal operate in the Exclusive Economic Zone and the subsistence economies. However, (EEZ) on the mainland and Zanzibar has o f the three natural resource sectors, only increased from fewer than 10 in 1998 to fisheries represent a net contributor to the more than 170 in 2004, corresponding to Treasury. Forestry and wildlife are subsi- revenue o f US$3.3 million. In terms o f dized through government allocations to export earnings, fisheries contributed 8 cover their recurrent expenditures and percent o f total exports in 2005 through foreign grant allocations to fmance (US$142 million), the export value o f Nile their operational budgets. Table 11.2 shows perch alone beingUS$129 million. the government recurrent budget alloca- 220 Harnessing Natural Resources for Sustainable Growth tions to the sectors: forestry and wildlife still a long way from seeing the effects o f each received 29 percent o f the MNRT its implementation. Increased public in- budget in 2003/04, followed by fisheries vestments are needed to support these (18 percent) andtourism (11percent). broad-based environmental management activities. There is a mismatch between foreign re- source allocation for sectoral development UntappedGrowth Potential activities and national funding allocation for recurrent expenditures. The large de- With regard to nonconsumptive use o f gree o f underspendingo f the development wildlife for game-viewing tourism, the budgets in both sectors is a possible indi- potential in the southern parks remains cation o f constraints on capacity to absorb untapped. Although the Northern Circuit foreign funding and institutional ineffi- has supposedly reached maximum carry- ciencies, aggravated by the uncoordinated ing capacity in terms o f numbers o f visi- policies o f the development partners. In tors, places such as Ruaha and Katavi Na- the forestry sector, a planned sectonvide tional Parks are still fairly unknown. approach is supposed to address the last Shifting marketing and infrastructure de- problem. velopment to those areas would provide new growth potential for the tourism sec- There is a tendency to draw government tor. In addition, there i s scope to increase allocation away from these sectors, because the concession fees o f international tour they should finance themselves and move operators and lodges, which currently ac- toward privatization. However, there are count for only 2 percent o f the revenue o f trade-offs to this trend. There is a need for Tanzania National Parks (TANAPA). government to control and regulate, setting andenforcing fiscal andmarketinstruments Marine fisheries haverecorded a sharp rise to ensure sustained growth and incorpora- in revenue from licensing o f foreign ves- tion o f externalities. Although the recurrent sels in the EEZ (figure 11.1). Some government budget allocations to these sec- sources estimate that the present revenue tors are getting smaller and there i s a trend does not reflect the amount that the gov- toward privatizing some government func- ernment could earn and that the real catch tions, the need for sectonvide environ- i s much higher than what has been as- mental managementfunctions is increasing. sumed as the basis for setting license fees Environmental impact assessments; mar- (between 200 and more than 400 tons per ket-based instruments for environmental day per boat). Notably, there i s no catch- protection (such as taxes, subsidies, stan- based license or fee, and vessels are al- dards, and permits); monitoring o f stocks; lowed unlimitedcatch if they have a valid and legal enforcement are becoming more license. Although the annual revenue, as important as the economy grows. Because shown in figure 11.1, i s considerable, it i s o f institutional failures, these overarching low compared with the estimated value o f environmental management functions have fish caught by foreign vessels in Tanza- basically been lacking in Tanzania for the nian waters and sold in foreign markets. past decade . One must thus assume that there i s scope for revenue increases. - Although the new Environmental Man- agement Act provides the necessary envi- With regard to freshwater fisheries, past ronmental framework law, the country i s growth rates are based mainly on Nile 221 Elements of a Strategy for Shared Growth Figure 11.1. Annual License Revenue and Number of Foreign Vessels in EEZ, 1998-2004 4,O00,O00 180 3,500,000 160 3,000,000 140 120 p 5 u 2,500,000 m 2,000,000 100 g 1,500,000 80 % 60 2 1,000,000 40 500,000 20 0 0 1998 1999 2000 2001 2002 2003 2004 I+US$ +NO of vessels1 Source: Data from Fisheries Department. perch exports from Lake Victoria. Other managed, commercial fisheries can have lakes, such as Tanganyika and Nyasa, are positive effects on economic growth and commercially underdeveloped, as i s the poverty reduction at the same time. Princi- harvestingo f other species. ples o f management need to include reten- tion and reinvestment o f revenue into the Diversification could also be sought in sector and safeguarding o f the artisanal terms o f exploring additional export mar- fisheries to protect their rights and access kets. Risks and vulnerability increase to the resource. when export earnings in a sector depend entirely on a single market. For example, Potential for Local Spinoff fisheries exports from Lake Victoria are Effects destined mainly for the European Union. Following an unfavorable assessment o f Local spinoff effects are missing for ma- sanitary standards by the European Union, rine fisheries in the EEZ. While new fish- Tanzanian fish-processing plants had to eries agreements are being negotiated with halt all production for several months in foreign countries, no fish are expected to 1999 because o f a temporary ban. be landed ashore, and few supplieswill be sourced from within Tanzania. If no such Despite highgrowth in the fisheries sector spinoff effects are created, the net effect o f anddespite localproductiono f fishnets, 95 commercial fisheries on poverty reduction percent o f fishnets are imported. Hence, an may be negative, increasing competition important backward link to the industry with artisanal fisheries over the same re- and an employment opportunity remain source. unexploited. The fact that Tanzania i s a net importer o f Commercial fisheries present an impor- forest products i s a sign o f lost opportuni- tant, emerging revenue source for the ties for income generation for the local country and the sector. Ifthe sector i s well economy. Similarly, the mining sector 222 Harnessing Natural Resources for Sustainable Growth Figure 11.2. Income to OlolosokwanVillage, NgorongoroDistrict Council, 1999-2 003 12,000,000 10,000,000 8,000,000 I 6,000,000 Village (Tnb.) Diatrict (Tah.1 4,000,000 z,000,000 0 1999/2ooo 2OOOH)l 2001/02 2002/03 Source: Reconstructedfrom Kallongaand others 2003. Note: Figuresare not exact. seems to have had limited influence on ing operations generally are detached from reducing poverty in the local economy. local supply chains and therefore primarily Employment inthe large-scale miningsec- create employment inthe services sector. tor i s limited, although younger employees especially may receive significant salaries. Potential for Poverty The majority o f those employed in the Reduction mining sector are self-employed in the small-scale sector, typically as artisanal Inadditionto their potentialfor generating miners. Returns are very low, especially government revenue, wildlife, fisheries, when one considers the hardship associ- and forestry resources provide the nonag- ated with this kind o f employment. It ricultural subsistence base for rural com- seems, furthermore, that an increasing in- munities in remote locations. Increased come disparity i s emerging between those emphasis on natural resources-related en- employed in the small-scale mining sector terprises has the potential to create addi- and those employed inthe large-scale one. tional income opportunities for the rural To the extent that those recruited by the population. large-scale miningsector are recruited out- side the local community, the local com- For example, in Loliondo Division in munity is thus restricted to opting for Ngorongoro District, seven villages earn poorly paid employment opportunities in more than U S $ l10,000 per year fromjoint the small-scale sector. ventures with wildlife tour operators. In Ololosokwan village, tourism revenue to- Large-scale mining may have positive ef- tals about US$55,000 per year. The in- fects for local communities through the come from payments by one o f four tour improvement o f basic infrastructure. There operators in Ololosokwan i s shown in fig- is, however, no indication that expansion ure 11.2. If the effects o f elite capture are in the mining sector triggers significant avoided and income i s equitably distrib- growth inthe local economy, because min- uted within the communities, this income 223 Elements of a Strategy for Shared Growth Figure 11.3. Village Incomes from Hunting in Lunda-Mkwambi (Game-ControlledArea), Idodi, and Pawaga Divisions, 1996-99 18,000,000 16,000,000 14,000,000 12,000,000 211110,000,000 tourist hunting 8,000,000 village income from 6,000,000 resident hunting 4,000,000 2,000,000 0 1996 1997 1998 1999 Source: Data from Walsh 2000. Note: Village receives 25 percent of licensefees for tourists' hunting. has large poverty reduction potential in a The income from hunting quotas was suf- dryland area that does not offer many ficient to triple village-level communal other opportunities for diversification. income, enabling villages to pay district- level taxes that would otherwise be levied Although Ololosokwan i s an exceptional on households, as well as to carry out so- example, the potential for local develop- cial infrastructure investments. One o f the ment from wildlife-related tourism has not success factors identifiedwas that the pro- been fully tapped in other areas. In the ject has emphasized institutional capacity Mara-Serengeti ecosystem, the number o f buildingat village and intervillage levels. households earning any income from tour- ism varies from 86 percent in Talek, Similarly, community-based forest man- Kenya, to 12 percent and 3 percent at the agement has provided revenue to villages Ngorongoro Conservation Area and Lo- across Tanzania. The 2002 Forest Act au- liondo Game Reserve on the Tanzanian thorizes villages to sell timber from their side. own forest reserves, potentially providing a new source o f forest revenue that would In the Southern Circuit, tourism is grow- accrue directly to the communities. ing, offering potential scope for positive effects on local economic development. Despite the conducive policy framework Participatory wildlife management in in both the wildlife and forestry sectors, communities close to Ruaha National Park weak governance systems at both the cen- in Iringa District generated T Sh 15 mil- tral and local levels have so far limitedthe lion in local income in 1999, accrued realization o f the potential for poverty re- through earnings from the residents' hunt- duction through community-based natural ing quota. An additional T Sh4.1 million resources management. The main focus in was earned from the 25 percent share o f community wildlife management has been license fees from tourists' hunting (figure on institutions and the distribution o f 11.3). benefits rather than on enterprise opportu- 224 Harnessing Natural Resources for Sustainable Growth nities at the household level. Fear o f ineq- tion o f growth comes at the expense of uity has led to the relative neglect o f en- pricing resources below market value, trepreneurship in Tanzania, reflecting a which leads to loss of income, erosion o f persistent and much broader philosophical critical stocks, and an associated decelera- bias against private enterprise. tion o f growth inthe long term. SustainabiIity of Growth Estimates o f resource rents from marine fisheries, computed from license fees as a In the context o f sustainability o f natural percentage o f the value o f revenue through resources-based growth, several con- licenses for foreign vessels fishing in the straints are emerging that lead to revenue EEZ, show that gross resource rent is ap- loss and possible deceleration o f growth in proximately 2.2 percent. That percentage the long term: i s less than half o f what might be expected in a Western industrial fishery. Although the current license fee arrangements o f 0 Underpricing of resources, thereby not allowing the capture o f resource rents private fisheries agreements in the EEZ generate not insignificant amounts o f 0 Weak environmental governance sys- revenue, the level is too low to result in a tems reasonable return of revenue (more than 5 to 7 percent o f gross revenue) to capture a 0 Limited knowledge o f stocks, their resource rent. values, and changes over time Similarly, in forestry, royalties have been Underpricing of Resources fixed arbitrarily. The 2002 Forest Act de- mands the determination o froyalties based Sustainable growth based on renewable on market value, profitability, and princi- resources requires that the cost o f extract- ples o f sustainable harvesting. Improve- ing a resource and the notional cost o f re- ment o f the Forest Produce Pricing System placing a unit o f the resource, commonly should include market-based pricing of known as resource rent, be evaluated so forest produce and public auctions or ten- that the wealth base i s not eroded. Al- dering for timber lots. Royalties could also though royalties are the most important be used as an instrument to divert harvests source o f government revenue in forestry from pressured species toward lesser- (83 percent), wildlife (96 percent from known species. huntinglicenses), and fisheries (84 percent from royalties and 15 percent from export Inthe wildlife sector, the concession com- licenses), they are set arbitrarily and cap- ponent o f TANAPA's earnings i s only 2 ture neither market values nor resource percent, which most likely underrepresents rents. the value o f these concessions compared with the income they generate for the for- Similarly, in the mining sector, licenses to eign investor. Loss o f revenue and unsus- foreign investors do not take the capital tainable use i s also fostered through hunt- component into account. Tax incentives ing quotas that do not reflect true market for foreign investors have been generous, values and are not based on ecological to attract capital investment and to open monitoring to maintain critical stocks. the market, at the expense o f sustainability Presently, concessions are leased at rates principles. In this scenario, the accelera- far below true market value irrespective o f 225 Elements of a Strategy for Shared Growth size, quality, or income potential. This usually benefits a few powerfbl players, it situation represents a massive loss o f in- usually disadvantages the majority o f the come to the Wildlife Division (estimated population. It leads to loss o f income and at more than US$7million). The system livelihoods for rural communities. promotes subleasing to foreigners, with the result that much o f the income gener- Limited Knowledge of Resources ated by the industry never enters the coun- Stock Valuesand Stock Changes try and substantial tax revenue is lost. The optimal scale o f natural resource- Weak Environmental Governance based economic growth must be at a sus- tainable level. Hence, a general macrolevel In forestry, an undercollection o f 5 to 10 constraint o f growth i s that the optimal percent o f revenue i s reported to be due to scale i s the one at which the long-run mar- inefficiencies in revenue collection and to ginal cost o f expansion equals the long- corruption in the sector. The 2004 logging term marginal benefits o f expansion. This scandal in Rufiji revealed that illegally constraint cannot be operationalized if the harvested logs were valued at true costs o f resource extraction are un- T Sh 382.65 million. known. Inmarine fisheries, it is alleged that Zan- Commonly in fisheries, forestry, wildlife, zibar licenses for foreign vessels are regis- and mining, there are neither inventories tered in Muscat, Oman, the fees thus es- o f the fbll availability o f stocks nor com- caping the Zanzibar authorities. In EEZ plete information about their value. In ad- fisheries, the lack o f transparency i s attrib- dition, stock changes are not monitored uted to a large degree to the lack o f catch comprehensively. In the absence o f stock reporting by foreign vessels. The govern- and flow data, limits o f extraction and ance regime in EEZ fisheries i s unique in quotas associated with licenses can be set that it imposes responsibilities for trans- only arbitrarily. Hence, they are not based parency and accountability on other na- on sound ecological calculations and real- tions whose fisheries cover distant water. istic projections. For example, in marine fisheries, there are no catch limits attached In the wildlife sector, a nontransparent to licenses, allowing vessels to take as system o f quota setting for the hunting in- much fish as are available, and foreign dustry by the government leads to imper- fishing vessels return only scant informa- fect competition inthe market. There i s no tion on actual catch. Similarly in forestry, competitive bidding for hunting conces- land coverage, deforestation, and values sions, but distributionthrough autonomous represented in the country's forest estate govenunent decision making. Effective are a matter o f speculation. market forces are hence not applied to op- timize revenues. This policy intervention There i s already government effort in the leads to a monopoly o f knowledge by the fisheries and forestry sectors to address Wildlife Division and an oligopsony in some o f these problems. For example, the access to the resource, a situation inwhich Fisheries Department has lately increased a small number o f large buyers controls its monitoring, control, and surveillance the market. Consequently, quotas are sold with support from a South African Devel- below market value, leading to a loss in opment Community regional project. The revenue. Although imperfect competition Forestry Department is in the process o f 226 Harnessing Natural Resources for Sustainable Growth developing a national forest monitoring penetration o f foreign vessels into ter- facility and database. ritorial seas affects the catch o f ar- tisanal fisheries. Externalities Mining poses a number of threats to Consideration and efficient control o f ex- andpossibilities for localcommunities, ternalities i s important to reflect the true as well as for the miners themselves. cost o f the use o f resources and to prevent The nature and extent o f the threats their overexploitation. In addition, exter- andopportunities cannot be assessedin nalities can cause trade-offs between eco- detail because o f the lack o f reliable nomic growth and poverty reduction be- data. There are concerns that large cause they can negatively affect local commercial mining crowds out the ar- people's access to natural resources. tisanal sector. Also, there are indica- Lastly, control o f externalities can realize tions o f a number o fnegative social ef- cost savings in monies otherwise spent on fects, notably child labor, HIV/AIDS, pollution control. Examples are abundant: andgender imbalances. Economic growth i s associated with The evidence about the environmental an increased need for energy and wa- effects o f large-scale mining suggests ter supply for domestic and industrial that mining communities may suffer a purposes. Currently, 95 percent o f en- number o f severe effects, from direct ergy supply comes from biomass en- and observable noise and erosion to ergy. Because o f incorrect pricing, the longer-term pollution o f air, water, and price o f charcoal does not represent soil, which in turn may have serious the full value o f the wood being har- health consequences. Still, the evi- vested. In terms o f providing value dence does not allow for extrapolation; added to growth through energy and more comprehensive analysis i s re- water supply, Tanzania's forests pro- quired to get a better idea o f the envi- vide critical capital. Catchment forests ronmental implications o f large-scale are an example, and their conserva- mininginTanzania. tion i s clearly a binding constraint to be addressed. The current policy framework does not provide for sound management o f natural Increased agricultural production and resources and mitigation o f externalities. intensification can create externalities. Instruments applied at present for revenue Large commercial rice farming in the generation do not address externalities, nor UsanguPlains has reduced the dry sea- are they used as instruments to capture son flow o f the Great Ruaha River rents from natural resources. Rather than through intensified year-round irriga- employing fiscal instruments to steer the tion, which i s negatively affecting wa- exploitation o f resources, there is, alleg- ter use by small-scale farmers down- edly, tax evasion within the revenue- stream. generating public sectors themselves, Commercial fish production for export Hence, in the present regime o f environ- markets at Lake Victoria erodes a base mental governance, increased growth will o f livelihood and food supply for local come at the cost o f running down the re- fishing communities. Similarly, the source stocks, impedinglong-term growth 227 Elements of a Strategy for Shared Growth opportunities. To ensure a positive net ef- tional Strategy for Growth and Poverty fect o f accelerated growth on positive Reductionthat policies should be designed poverty reduction, a careful balance needs so that benefits from high-growth sectors to be preserved between increasing export are transmitted to the poor in the form o f earnings and maintaining the resource base better livelihood opportunities-for exam- for the artisanal sector. In particular, in ple, supporting supply links with local fisheries, certain safeguards need to be put producers. in place for the artisanal fisheries to pro- tect their rights, access to the resource, and The recommendations below are divided livelihoods. into general recommendations that equally apply to all natural resources sectors and Recommendations sector-specific recommendations: Making sustainable development opera- General recommendations tional i s an international political chal- lenge. In particular, in the context o f - Strengthen capacity for data col- globally shared resources such as fisher- lection, recordkeeping, monitor- ies, responsibilities apply to both harvest- ing, control, and surveillance and ing and host countries. True factor pricing enforce punitive measures to and resource rent capture are policy in- control illegal practices. struments that even some Western coun- - Controlexternalitiesthroughfis- tries grapple with. Yet some basic princi- tal instruments, royalties, and re- ples o f governance are missing in source pricing and increase reve- Tanzania, which, if applied, could regain nue from rent capture rather than some o f the lost opportunities described uncontrolledexploitation. inthis chapter. - Increase efficiency inrevenue The single most important recommenda- collection and administration, as tion for capturing and maintaining natural well as full transparency and ac- resource-based growth in Tanzania i s to countability over revenue genera- reform environmental governance so as to tion and distribution. achieve good governance, rule o f law, and - Promotemarket-basedprinciples equity. Such reform includes ensuring when appropriate, ensuring local greater coherence between different na- spinoffs and allowing competition tional policies and instruments, particu- andentrepreneurial development. larly community-based wildlife manage- ment, tourism development, rural growth Fisheries strategies, investment regulations and in- - centives, andpoverty reductionstrategies. Put in place a regulatory frame- work and sound governance re- In addition, Tanzania needs to make in- gime for marine fisheries, com- vestments in the improvement o f its hu- prising the EEZ and near-shore man capacity and capital stock so that fisheries. value-added processing o f natural re- - Safeguardrightsandlivelihoods sources can take place more often within for coastal communities, for ex- the country. This investment i s required to ample, through demarcation o f a comply with the principle stated inthe Na- community territorial sea. 228 Harnessing Natural Resources for Sustainable Growth - Conduct a fisheries sector review balance to net exports o f forest to assess the economic and so- products. cial, ecological, and fiscal per- spectives and policy options. The - Introducenewrevenuesources, review could inform policy mak- such as watershed management ers and influence the strengthen- fees from hydropower stations, ingofthe regulatory framework. sale o f genetic resources, and carbon credits. - Establish some form o f EEZ in- spectorate patrol to build a more 0 Wildlife accurate picture o f available re- sources. - Encourageattitudinalchangeto- ward wildlife at the policy level - Investigate the potential for ex- so that policy makers see it as an ports o f marine products and for asset for rural development and value adding o f these products to poverty reduction rather than as promote growth in the coastal something looked after by con- zone. servationists. Such an attitude change will include a shift in the 0 Forestries emphasis o f community wildlife - Introduce taxes for wood lot and approaches to focus on creating plantation owners-in particular enterprise opportunities. an income tax based on timber - Ensurethat localcommunities sales and a property tax based on are the principal decision makers the average productive capacity for allocation o f concessions and o f different landcategories. quota setting for hunting on their - Enforce the collection o f royal- land and that they receive and ties and fees and eliminate the manage the hnds generated on exemption for industries such as their land. tobacco and fishing. - Reform the tourist hunting indus- - Improve the forest produce pricing try to realize its true revenue po- systemthrough market-basedpric- tential. Such reform includes the ing, public auctions or tendering introduction o f market-based for timber lots, and cheaper royal- competition in the commercial ties for lesser-known species. hunting industry through com- petitive bidding for concessions. - Increase domestic and foreign It may have the positive side ef- private sector investment through fect o f naturally controlling sub- reduction o f bureaucracy in the leasing and related revenue licensing system, clear invest- losses. ment guidelines, clearly defined ownership o f all forestland, tax - Introduceperformance-basedin- incentives, credit facilities, and dependentmonitoring o fthe hunt- technology transfer. ing industry, possibly through certification, to ensure that certain - Increasethecapacityutilization standards are adhered to. Criteria o f the sector to reverse the trade should be set to consider the 229 Elements of a Strategy for Shared Growth maximum income from the least gies; set objectives in terms o f number of animals hunted and local development effects, not contributions toward protection just numbers of tourists or for- andcommunity involvement. eign exchange earnings. - Revisethequota-settingsystem - Establish a pro-poor tourism on the basis o f more objective growth program to place attention criteria, computerization o f hunt- on company practices, destination ing data, and monitoring o f tro- management, infrastructure de- phy quality and age. velopment, procurement patterns, - Conduct a review (by the Minis- national training, andregulation. try o f Finance) o f the financial Mining management and taxation proce- 0 dures o f the Wildlife Division to - Improve data collection on ex- assess strengths and weaknesses. ternalities in the mining sector This review would include an in- (for example, through more rig- ventory o f the true value o f hunt- orous and systematic enforce- ing licenses. ment of environmental impact assessments. 0 Tourism - Revise the pricing system to cap- - Integrateopportunities forpro- ture the capital component o f poor tourism into tourism strate- nonrenewable miningresources. 230 CHAPTER 12 Enhancingthe Capacity of the Poor to Participate in Growth This chapter examines what strategies and fertility. Next, we discuss aspects that would enable poor people to participate in allow poor people to build physical capital growth. The most important assets o f poor and, in particular, the exposure to risk and Tanzanians are labor and, in rural areas, the role o f financial markets. Finally, and land. A strategy that aims to enhance the startingfrom the realizationthat certain peo- capacity o f the poor to participate in ple may not be able to benefit from growth, growth should therefore focus on an inten- we discuss socialprotection andinequity. sification o f the use o f labor and land while permitting poor Tanzanians to build Improving Human Capital uphumanandphysicalcapital. of the Poor We first focus on those aspects that build Levels o f human capital are low in Tan- human capital: education, nutrition, health, zania, and building human capital i s an 231 Elementsof a Strategyfor Shared Growth Table 12.1. Increase in Per Capita Consumption Relative to Households Headed by Individualswith No Education Level of educationof head of household Dar es Salaam (%)Other urban areas (%) Rural areas (%) Some primary education 25 19 17 Completed primary education 57 30 42 Some secondary education 36 42 48 Completed secondary education 49 52 70 Postsecondary education 90 73 90 Adult education only 43 -1 -2 Source:Authors' calculations based on National Bureau of Statistics 2002. important element o f the poverty- dary education have a per capita income reducing strategy for at least two reasons. that i s 49 percent higher than those in First, by building human capital, the households headed by someone with no foundation i s laid for higher growth inthe education. In rural areas, the difference i s future (Barro 1991; Mankiw, Roemer, even larger: 70 percent (table 12.1). and Weil 1992). Second, by building the human capital o f the poor, the pattern o f Education is associated not only with in- growth will be pro-poor. come but also with nonincome dimensions o fpoverty. Low levels o f education lead to Inthis section, we discuss three elements higher total fertility, lower levels of child o f human capital: education, nutrition, and nutrition (Lindeboom and Kilama 2005), population (health care and fertility). higher child mortality (Rafalimanana and Westoff 2001), and an intergenerational Education transfer o f poverty because children from poor households are less likely to attend Twenty-nine percent of Tanzanians age 15 school themselves. The total fertility rate and above are illiterate (National Bureau o f o f women age 40to 49 i s 6.5 ifthey do not Statistics 2002), and according to the 2002 have any education but drops to 4.9 ifthey census, the average number o f years o f completed at least primary education (Na- education o f the working population (age tional Bureau o f Statistics and Macro In- 20 to 64) is 5.1. Not only i s the lack o f edu- ternational 2000). Evidence from Kagera cation one o f the factors contributing to shows that children with educated parents Tanzania being one o f the poorest econo- have better nutritional outcomes (Alder- mies inthe world, but within Tanzania, dif- man, Hoogeveen, and Rossi 2006). And ferences in education are strongly associ- according to the HouseholdBudget Survey ated with income levels. In rural areas, (HBS), 52 percent o f Tanzanian children where poverty is highest, the average level age 7 to 10 attended school, but only 44 o f education o f heads o f household i s 4.3 percent from the first two quintiles did so years compared with 7.8 years in Dar es (National Bureau o f Statistics 2002). Salaam, where poverty incidence i s lowest (National Bureau o f Statistics 2002). Con- With the prevailing low levels of educa- sumption regressions show that individuals tion, one can see that educational attain- living in Dar es Salaam in households ment needs to be raised inorder to reduce headed by someone who completed secon- poverty and to raise income in general. In 232 Enhancing fhe Capacity of the Poor to Participate in Growth Figure 12.1. Primary EducationPerformance,1995-2004 1,800,000 100 1,600,000 90 1,400,000 80 8 1,200,000 70 3 1,000,000 60 2 2 8 800,000 50 40 600,000 30 5 400,000 20 g 200,000 10 0 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Year Source: Ministryof Educationand Culture 2004. recent years, important initiatives to that Not only did net enrollment increase dra- end have been undertaken, of which the matically with PEDP fkom 59 percent in Primary Education Development Program 2000 to 91 percent in 2004, evidence fkom (PEDP), introduced in July 2001, and the Kilimanjaro and Ruvuma suggests that ine- Secondary Education Development Pro- qualities in access to primary education gram (SEDP), started in September 2004, have disappeared. Figure 12.2 presents, for are the most important. PEDP's aim i s to rural Kilimanjaro, concentration curves for have all children between the ages o f 7 2001 and 2003. Inpanel (a), the concentra- and 12 enrolled into standard one by 2005. tion curve for 2001 shows how pre-PEDP In recognition that some parents fail to access to primary education was unequally send their children to school because of distributed: children fkom wealthier house- cost or distance, the programhas abolished holds attended school relatively more often all school fees and other mandatory paren- than children from poorer households. The tal contributions and started an investment distribution o f access was as unequal as the program in school buildings and class- distribution o f consumption, representedby rooms. SEDP's main objective i s to ensure the Lorenz curve. With PEDP, inequalities that more o f the increased numbers o f inaccess to education disappeared, and the primary school graduates can be absorbed concentration curve coincides with the 45- into secondary schools. degree line. Though not shown in the fig- ure, results for Ruvuma are comparable. Under PEDP's standard 1, enrollment in primary schools increased tremendously Despite major progress in enrolling chil- figure 12.1). With a net enrollment rate of dren and in addressing aspects o f the 90 percent (and a gross enrollment rate gender gap (box 12.1), implementation o f that exceeds 100 percent), Tanzania has PEDP lags behind in certain regions, par- now put in place one o f the essential pre- ticularly the poorer and more isolated conditions to attain the MillenniumDevel- ones, as i s evident from net enrollment opment Goal o f ensuring that, by 2015, figures. In 2004, enrollment was 90 per- children will be able to complete a full cent for the country as a whole, but in course o fprimaryschooling. Tabora it was only 68 percent and in 233 Elements of a Strategy for Shared Growth Figure 12.2. Changes in the Distributionof Access to Educationin Rural Kilimanjaro,2001 and 2003 (a) 2001 0I 20I 1 I I I Cum. distribution of population 40 60 80 100 (b) 2003 Sources:Authors' calculations basedon NationalBureau of Statistics 2002 and UnitedRepublic of Tanzania 2004c. Kigoma 77.2 percent, whereas in Dar es ronment for learning. The availability of Salaam it was 93.1 percent. Interestingly, textbooks leaves much to be desired as some poor regions do particularly well. well, although it improvedfrom 8 students For example, Ruvuma has a net enroll- per textbook in 2001 to 4 students per ment o f 99.3 percent; thus, poverty is not textbook in 2004. That lack o f quality is the only explanatory factor for the diver- reflected in the pass rate at the Primary gence inperformance. School Leavers Exam. At 49 percent in 2004 (up from 22 percent in 2000), the Concerns exist about the quality o f educa- pass rate remains low, which implies that tion. Teacher-pupil ratios o f 1:59 (up from more than one-half o f the students do not 1:46 in2001) are hardlyan enabling envi- learn the material expected for primary 234 Enhancing the Capacity of the Poor to Participate in Growth Box 12.1. Gender Differencesin Education With PEDP, some gender biases ineducation disappeared. Gender parityinenrollment has al- most been attained: 49 percent o f those attendingprimary schools are girls (Ministry o f Education and Culture 2004). Information from the 2002 census also suggests that the gap ineducation at- tainment has been closed. As the figure below indicates, although women age 60 and above re- ceived only 30 percent o f the education of men of the same age, girls below age 15-the youngest age cohorts-are actually better educated than boys o f the same age. Gender Differencesin Educationby Age Age category Source: Data from 2002 census. However, not all gender differences disappeared. Pass rates for girls at the Primary School Leav- ers Exam are lower than those for boys. In2004,48 percent o f all boys passed compared with 33 percent o f all girls. That 15 percentage point gap inpass rates has persisted over the past 10years and leads to a gender imbalance at the end o f the primary curriculum, which is corrected at en- rollment in secondary school in form I. Form Ienrollment data for 2004 show, again, a near gen- der balance at entry level. Yet gender imbalances reappear later. Between form Iand form IV, the girl-to-boy ratio drops gradually untilit reaches a ratio o f one grl to two boys in form VI (United Republic o f Tanzania, Vice President's Office 2005). school. However, pass ratios are only a the start o f the millennium. Inthe past two proxy o f quality and do not really reflect years, that number has increased further, what children learn. and approximately 30 percent o f those who now finish primary school get a place In secondary education, there have been in a secondary school (figure 12.3). As a improvements inrecent years. The number result o f the expansion o fpublic secondary o f primary school leavers entering secon- schools, the share o f private secondary dary school increased from a low o f 3.4 schools in total secondary education has percent in the mid-1980s to 22 percent at gradually dropped from 60 percent in the 235 Elements of a Strategy for Shared Growth Figure 12.3. Enrollment in Secondary Schools, 1990-2004 -Fraction of primary school Ieaverstransitioningto secondary school v1 -Market share ofprivate secondaryschools 0 % c 60.0 6 2 50.0 8 3 40.0 18 30.0 lo t I I!? 20.0 2 8 10.0 25 50'1/ r" 4 1 I I I 1 I I I 1 I I I I 1 I I I I I ~I c0.0 2 ~ ~ ~ Source: Ministry of Education and Culture 2004. early 1990s to about 33 percent at the The long lead time before current invest- moment. Nonetheless, secondary educa- ments in education will pay off in higher tion remains a relatively exclusive affair income and reduced poverty i s reason to that i s skewed toward the non-poor, which consider whether levels o f education o f the HBS data illustrate. The probability those in their productive years can be that a child age 14 to 18 attends secondary raised through adult or youth education. school i s 2 percent if the child i s from a According to the HBS, in 2000/01 26 household inthe first consumption quintile percent o f households were headed by and 13 percent if the child comes from a someone with no education, which dem- household inthe top quintile. onstrates a need for education (National Bureau o f Statistics 2002). And youth and Investinginhumancapital has a long lead adult education potentially have a large time, and it will take many years before effect on a household's earning capacity. the current investments ineducation trans- That effect can be illustrated by the con- late into higher income and reduced pov- sumption level o f those who benefited erty. Even if, over the coming 10 years, from some primary education: it i s 17 to those age 10 to 19 manage to raise educa- 25 percent higher than the consumption tion levels to 7 years and those age 5 to 9 level o f those who did not receive any raise education levels to 10 years, the av- education (table 12.1). In Dar es Salaam, erage level o f education in the working consumption levels in households in population (age 15 to 65) will have risen which the head received adult education from 5.1 years to only 6.8 years. Levels o f only are 43 percent higher than inhouse- education among the youth (age 15 to 29) holds inwhich the head has no education. will have risen considerably, however, A word o f caution seems justified how- from 5.6 years now to 8.1 years, which ever, as adult education has no significant should enhance their ability to escapepov- beneficial effect in rural and other urban erty considerably. areas. 236 Enhancingthe Capacityof the Poor to Participatein Growth Figure 12.4. Percentageof UndernourishedChildren under Age Five, 1991-2004 50i 45 40 -4535 30 25 c r ;20 15 10 5 0 Stunting Underweight Wasting Source: Data from DHS 1991, 1996, 2004; National Bureauof Statistics and Macro International 2000. Nutrition trition, butmany suffer from deficiencies o f micronutrients such as iodine, iron, and vi- Improving the nutritional situation o f Tan- tamin A (National Bureau of Statistics and zania's population i s another element in a Macro International 2000). According to strategy aimed at building human capacity the 2004 Tanzania Demographic and for growth and poverty reduction. The Health Survey (DHS), only 43 percent o f prevalence o f undernutrition inTanzania i s households used adequately iodized salt, high. As of 2004, nearly 40 percent of the and 46 percent of children age 6 to 59 children age 0 to 59 months are chroni- months benefited fiom vitamin A supple- cally undernourished or stunted (low mentation in the six months preceding the height for age) (figure 12.4). About 3 per- survey (National Bureau o f Statistics and cent are wasted (low weight for height), ORC Macro 2005). Approximately two- and 22 percent o fchildren are underweight thirds ofchildren and43 percent ofwomen (low weight for age), which i s a composite are anemic. measure o f long- and short-term undernu- trition and one o f the Millennium Devel- Both types o f nutrient deficiencies have opment Goal indicators. negative consequences for the ability to be economically active. Undernutrition re- Nutrition rates are worst among the poor. tards physical growth directly as well as According to the 1999 Tanzania Reproduc- indirectly by increasing the susceptibility tive and Child Health Survey (TRCHS), 50 to disease. It affects cognitive and mental percent o f children inthe bottom two quin- development and educational attainment tiles are stunted and 34 percent are under- and leads to reduced productivity and re- weight. In comparison, 23 percent of chil- duced income. Iron-deficiency anemia, for dren fiom the top quintile are stunted, and instance, has been shown to reduce pro- 22 percent are underweight. Tanzanians are ductivity by 5 to17 percent, with the not only affected by protein-energy malnu- higher percentage holding for the heavier 237 Elements of a Strategy for Shared Growth Figure 12.5. Fraction Stunted by Region (1992-99) and Under-Five Mortality Rate per 1,000 Live Births (2002) 70'0 M n I 8 W." 200 z B -0 1 50.0 40.0 150 * B 30.0 7! 2 100 ; 20.0 50 10.0 8$3 0.0 0 Source:Authors' calculations using data from DHS 1991, 1996; National Bureau of Statistics and Macro International2000 (stunting); and 2002 census (under-fivemortality rate). manual work such as farming, typically small stature as well (Alderman, Hoogev- carried out by the poorer segments o f the een, and Rossi 2006). And children who population(Horton 1999;Horton and Ross are stunted at a young age are less tall later 2003). inlife(figure 12.6). Undernutrition i s closely associated with Inthe face o f highrates o f undernutrition, the inability o f HIV/AIDS infected people large-scale interventions to address nutri- to undergo antiretroviral treatment. It i s tion deficiencies are limited even though also associated with under-five mortality. most nutrition interventions have attractive A high correlation exists at the regional benefit-cost ratios (table 12.2). level (a correlation coefficient o f 0.27) be- tween the fraction o f children age zero to Interventions that focus on micronutrients, five that are underweight and the under- improve infant and child nutrition, or re- five mortality rate. The correlation with duce low birth weight have the highest stunted children is even higher at 0.57 benefit-cost ratios. However, the more (figure 12.5). challenging interventions, such as mother- child care programs or integrated child Undernutrition i s a factor that stretches care programs, also have attractive bene- across generations. Evidence from Kagera fit-cost ratios. In fact, if one were to con- suggests, for instance, that parents o f small sider large-scale interventions, such chal- stature are more likely to have children o f lenging interventions would have to be 238 Enhancing the Capacity of the Poor to Participate in Growth Figure 12.6. Current Height of Children by Stunting Status 10 Years in Kagera Source: Datafrom Kagera Healthand Development Survey. Table 12.2. Benefit-Cost Ratios of Nutrition Interventions Benefit-costratio Type of Intervention Low High Improving infant and child nutrition Breastfeedingpromotion in hospitals in placeswhere infant formula is normally used 5.6 67.1 Integratedchild care programs 9.4 16.2 Intensivepreschool programwith considerablenutritionfor poor families 1.4 2.9 Reducing micronutrient deficiencies Iodine (perwoman of childbearingage) 15.0 520.0 Vitamin A (child under age six) 4.3 43.0 Iron (per capita) 176.0 200.0 Iron (pregnant women) 6.1 14.0 Source: Behrman,Alderman, and Hoddinott2004. part o f the package because o f the nature portion o f stunted children increases more o f Tanzania's nutritionproblems. than fivefold between 0 to 6 months and 13 to 24 months (from 9 percent to 53 Undernutrition in Tanzania takes shape percent). The proportion o f severely during pregnancy (many babies are born stunted children increases more than sev- underweight) and during the first months enfold (from3 percent to 21 percent) (fig- following birth. According to the 1999 ure 12.7). A strong negative association TRCHS (National Bureau o f Statistics exists between nutritional status and be- and Macro International 2000), the pro- ing born during the rainy season, when 239 Elements of a Strategy for Shared Growth Figure 12.7. Nutritional Status of Children by Age, 1999 0 20 40 60 Age in months Source: Authors' calculations based on National Bureau of Statistics and Macro International 2000. demands on labor are highest and illness Not only are community interventions at- i s most prevalent (Alderman, Hoogeveen, tractive for their rate o f return and from a and Rossi 2006; Lindeboom and Kilama human development perspective, but they 2005). There i s a strong positive associa- are also likely to be pro-poor. Because nu- tion o f nutritional status with breastfeed- trition problems affect poor households ing, but a negative one exists for low more severely, the poor stand to benefit birth weight, the birth interval, and long more from interventions. Evidence to that duration o f breastfeeding because chil- effect comes from Kagera, where it has dren need additional nutrients after six been shown not only that community- months o f age (Lindeboom and Kilama based interventions had a considerable 2005). Evidence also suggests that chil- beneficial effect on nutritional status (Al- dren living in households consuming derman, Hoogeveen, and Rossi 2006), but more milk are better nourished (Beegle, also that poor households benefit dispro- D e Weerdt, and Dercon 2006; Lindeboom portionately(figure 12.8). andKilama 2005). No correlation (even a negative one) exists between food insecu- The challenge for community interventions, rity as measured inthe HBS and the inci- however, i s less whether they can be suc- dence o f malnutrition at a regional level. cessful thanwhether they canbe introduced Because undernutrition rates are high in such a way that they are sustained over even in the wealthiest households, time. Evidence to date suggests that verti- undernutrition seems to be less the result cal, project-type interventions are difficult o f a lack o f food availability than one of to maintain and hard to scale up (TFNC dietary knowledge, hygiene, and care for 2004a). Interventions that are integrated pregnant women and young children within the sectors (health, agriculture, and (UNICEF 1990).' That pattern i s consis- education) and that facilitate increased hu- tent with global experience. man and financial resource allocation to 240 Enhancing the Capacityof the Poor to Participate in Growth Figure 12.8. Effect of Community Interventions on Average Nutrition Scores in Kagera Nutrition programpresent -----Nutritionprogramabsent F : II "-I 7 0 2Mxx) 40000 60000 m Percapita consumption Source:Authors' calculations basedon Kagera Health and DevelopmentSurvey data for 1992 to 1994. nutrition at subregional and community clined rapidly between 1999 and 2004 (fig- levels shouldbe considered. ure 12.4). Other health indicators saw less progress or even deterioration. Maternal Health mortality, which was 529 per 100,000 births in 1996, did not decline and may Health i s another important determinant o f even have increased, though the increaseto human capacity. Many factors affect the 578 in 2004 i s not statistically significant. health o f individuals, including where one The fraction o f blood donors infected by lives, the state o f the environment, genet- HIV/AIDS increased from 7 percent in ics, nutritional status, income, and educa- 1994 to 12 percent in 2003 for women and tion. Access to and use o f health care ser- from 5 percent to 8 percent for men.3 vices are another determinant o f health outcomes. The prevalence o f illness in Tanzania re- mains high. According to the 1999 Inrecent years, progress was made in im- TRCHS (National Bureau o f Statistics and proving the health o f Tanzanians. Accord- Macro International 2000), 35 percent o f ing to the 2002 census, life expectancy in- children under age five were affected by creased from 44 years in 1978 to 49 years fever in the preceding two weeks, 12 per- in 1988 and to 54 years for males and 56 cent were affected by diarrhea, and 14 years for females in 2002. Infant mortality percent experienced acute respiratory in- dropped in all regions (figure 12.9), and fections. The HBS 2000/01,which reports nationally, it fell from 115 er 1,000 in on adults and children, stated that 27 per- 1988 to 95 per 1,000 in2002. Inaddition, I: cent o f its respondents indicated having child malnutrition, which remained un- experienced illness in the preceding four changed over the course o f the 1 9 9 0 ~de- ~ weeks (National Bureau o f Statistics 241 Figure 12.9. Infant Mortality,1988 and 2002 :[ Elements of a Strategy for Shared Growth 140 I 120 -2g ,- 80 9 0 P 60 5rn 6 40 , 20 " 0 Source: Data from 1988and 2002 censuses. Table 12.3. Number of Days of School or Work Missed Because of Illness, by ConsumptionQuintile Consumption quintile Days missed Lowest Second Middle Fourth Highest Average None 0.30 0.31 0.29 0.30 0.32 0.30 One week or less 0.45 0.46 0.46 0.46 0.43 0.45 One to two weeks 0.13 0.11 0.11 0.12 0.13 0.12 More than two weeks 0.13 0.11 0.11 0.12 0.13 0.12 Source: Authors' calculations based on National Bureau of Statistics 2002. 2002). Malarial fever and diarrhea are the having consulted a healthprovider. And of most common types o f illness. Such high those who had been unable to work for at levels o f illness have economic conse- least two weeks, 17 percent did not consult quences. Almost one infour people missed a health provider. Among those who did at least one week o f school or work as a not consult a health care provider, 7 per- consequence o f illness, and that finding i s cent indicated that care i s too expensive evenly distributed across consumption and 2 percent stated that it i s too far quintiles (table 12.3). awayS4Households from the poorest quin- tile were more likely to mention cost as an One-third o f those who reported being ill obstacle to seeking health care than were in the preceding four weeks indicated not households from the top quintile, but the 242 Enhancing the Capacity of the Poor to Participate in Growth Table 12.4. Distance to Health Facilities, 1991 and 2000 Dispensaryor health care center (% of population Hospital (% of populationliving livina at distance] at distance) Distance(km) 1991 (%) 2000 (%) 1991(%) 2000 (%) Less than 2 34.3 37.9 13.6 13.3 2-5.9 41.O 37.5 18.4 19.1 6-9.9 15.3 15.9 10.2 13.5 1o+ 9.4 8.6 57.7 54.1 Total 100.0 100.0 100.0 100.0 Mean distance 4.4 3.9 19.7 21.3 Source: National Bureau of Statistics 2002. difference i s not very pronounced (8 per- cal across quintiles, suggesting that, in ab- cent versus 6 percent). And the percentage solute Tanzanian shillings, poor house- o f households from the poorest quintile holds spend considerably less on health that mentioned distance as an obstacle to thando wealthier households. Such spend- seeking health care was as high as that ing is unusual and deserves consideration, among the wealthiest quintile (2 percent). because poor households typically spend a greater share on health. The percentage o f households indicating that distance or cost i s an obstacle to More than one-half the individuals who seeking health care seems low but may consulted a health care provider visited a be underestimated. One could conceive government provider. Use o f private ser- that among the 91 percent o f respondents vices i s highest in Dar es Salaam, but in who indicated no need for health care rural areas, private providers also play an despite being sick, some may have done important role with traditional healers and so because o f the distance to health care missionary facilities. Poor and nonpoor providers, the opportunity costs o f time, households have a remarkably even use the confidence in the providers, and the o f private and public health care provid- quality o f care provided. Because 9 per- ers: poor households tend to visit tradi- cent o f the population lives more than 10 tional healers relatively more frequently, kilometers from a dispensary or health whereas the differences between visits to care center, the importance o f distance in public facilities and private or mission seeking health care i s likely underesti- facilities are negligible, as illustrated by mated (table 12.4). the fact that the concentration curves for visits to those institutions almost coincide Likewise, costs probably play a more im- with the 45-degree line infigure 12.10. portant role than the responses from the HBS suggest. The legislation o f private Though few differences exist between practice and the introduction o f user fees poor and non-poor households in their likely contributed to a doubling in the exposure to major diseases like malaria, share o f health expenditures in nonfood diarrhea, or respiratory infections or in consumption during the, 1990s to about 8 the fraction o f budgetary expenses for percent. That percentage i s almost identi- health services, health outcomes differ 243 Elements of a Sfrafegy for Shared Growth Figure 12.10. Concentration Curves for Different Health Care Consultations Public facility _ _ _ _ _ Private facility ............Traditional healer _ _ Mission facility I 0 .2 I Cumulative.4I .5 I .EI 1I distribution of the population Source:Authors' calculations based on National Bureauof Statistics2002. Table 12.5. Differences in Health Outcomes, by Quintile Quintile Indicator Lowest Second Middle Fourth HighestAverage Infant mortality rate (deaths per 1,000) 114.8 107.5 115.4 106.8 91.9 107.8 Under-fivemortalityrate (deaths per 1,000) 160.0 159.3 192.7 155.0 135.2 161.1 Stunted children under age five (%) 49.5 52.5 45.0 36.6 23.4 42.7 Underweightchildren under age five (%) 32.2 35.1 28.8 23.9 21.7 28.8 HIV prevalence(%) 3.4 4.5 5.6 9.4 10.5 7.0 Sources:Gwatkin and others 2003 from 1999 DHS data; HIV prevalence, from Tanzania HIV/AIDS IndicatorSurvey 2004 data. considerably by wealth status. Infant and age o f preventive health services. Poor child mortality is 15 to 20 percent higher children, for instance, are less likely to be among the poor than among those in the reached by vaccination services, and top quintile (table 12.5). The difference women from poor families are almost for nutritional indicators is even larger: three times less likely to have their birth 30 to 50 percent. The one indicator in attended by trained medical personnel than which poor households do substantially are non-poor women (table 12.6). better than non-poor households is HIV prevalence, which i s 3.4 percent among Differences in behavior or economic cir- those in the poorest quintile and 10.5 per- cumstances also explain differences in cent among individuals in the wealthiest health outcomes between poor and non- quintile. poor households. For instance, poor indi- viduals have less access to clean water and The health outcome differential may be are less well educated. Children from the explained in part by differences in cover- poorest quintile are five times less likely to 244 Enhancingthe Capacity of the Poor to Participatein Growth Table 12.6. Access to Preventive Health Services, by Quintile Quintile Preventive health Lowest Second Middle Fourth Highest Average service (%) (%) (%) (%) (%) (%) Bacille Calmette-Guerin coverage 88.8 96.9 87.3 93.7 99.9 92.7 Measles coverage 63.4 84.2 72.2 88.4 89.0 78.1 Diphtheria-pertussis- tetanus coverage 66.2 86.1 78.5 91.I 88.7 81.O Fully immunized (diphtheria- pertussis-tetanus) 53.1 74.3 61.7 80.8 78.4 68.3 Birth attended by a medicallytrained person 28.9 35.0 33.3 48.4 82.8 43.8 Source: Gwatkin and others 2003 from 1999DHS data. Table 12.7. SocioeconomicAspects of Health, by Quintile Quintile Lowest Second Middle Fourth Highest Average Socioeconomicaspect (%) (%) (%) (%) (%) (%) Access to pipedwater 30.9 31.3 34.2 43.6 56.1 39.2 Children under five who use a bednet 9.4 12.1 12.2 27.8 52.2 22.3 Availability of iodized salt in the household 52.1 60.0 61.7 71.3 86.0 66.9 Femalegenital cutting 29.2 16.4 16.7 18.2 11.0 17.7 Source:Access to pipedwater, National Bureau of Statistics 2002; female genital cutting, Tanza- nia HIWAIDS IndicatorSurvey 2004 data; other indicators,Gwatkin and others 2003 from 1999 DHS data. sleep under a bednet than children from shared effort across different sectors, in- the top quintile. Individuals living in poor cluding health, education,' andwater. households are less likely to consume io- dized salt, and women inthe poorest quin- The health sector faces several chal- tile are almost three times more likely to lenges. Funding i s one. Though fbnding have experienced female genital cutting has improved-the total per capita alloca- than are those from the wealthiest quintile tion o f public expenditure to health in- (table 12.7). creased from T Sh 5,100 in 2001 to T Sh7,374 in 2004 (Makundi and others In acknowledging that nonhealth factors 2004)6-it remains extremely low. Fur- play an important role in determining thermore, with the rising costs o f drugs, health outcomes, one can put into perspec- the observed increase in the budget (45 tive the role o fthe health sector inimprov- percent) overstates the possibility o f pro- ing the health o f the poor (and the popula- viding additional care. Drug resistance to tion at large). Improving the health antimalarials and increasingly to tubercu- outcomes o f Tanzanians requires abroadly losis treatment demands new, expensive 245 Elements of a Strategy for Shared Growth drugs, thereby inflating the cost o f health dren, is related to household size. A care without offering new services. strong association exists between nutri- tional outcomes and household size (Lin- Another challenge i s the high cost of deboom and Kilama 2005), between edu- treatment for HIV/AIDS, which creates cational attainment and household size, pressure this will create on the overall and between consumption poverty and budget. Moreover, because o f the pattern household size (Mkenda 2005). With re- o f HIV/AIDS, which affects nonpoor spect to the latter, children living in a households much more than poor house- household with three people have a 16 holds, increasing the share o f financing percent probability o f living in poverty, that supports treatment will make the whereas children living in a household healthbudget less p r o - p ~ o r . ~ with six people have a 35 percent prob- ability o f living inpoverty.* Some developments also work in the op- posite direction. The current allocation Living in a large household not only in- formula for the distribution o f the health creases the risk o f poverty, but it affects budget across districts, for instance, takes other dimensions o f welfare as well. Even the degree o f poverty into account. Other if one controls for the consequences of efforts to provide greater financial protec- household size on income, living ina large tion o f poor households that seek medical household has negative conse uences for a care are ongoing. child's educational outcomes. Those con- 9 sequences are illustrated in table 12.8, Inaddition, a serious human resource cri- which presents a regression o f the educa- sis affects the health sector. Only one-third tion gap" for children age 8 to 11 as a o f the positions for medical officers (going function o f household size. It shows that by the staffing norms) is filled and only 23 the education gap increases for children percent o f the positions for assistant medi- with more siblings age 0 to 6. The regres- cal officer and public health nurse are sion also illustrates the direct effect o f in- filled (Makundi and others 2004). That come on educational attainment: relative human resource crisis goes back to the to those in the poorest quintile, children mid-1990s, when the total health work- from households in the wealthier quintiles force was about 67,000. By 2002, how- have a smaller education gap. ever, it had decreased to 49,000, with the population increasing during the same pe- The strong correlation between human re- riod from 25 million to 33 million inhabi- source outcomes and family size i s reason tants. That decrease affects people living to consider one o f the key determinants o f in poor areas especially, because in the family size: fertility. One is not surprised absence o f additional incentives, motivat- to see that the strong association between ing medical personnel to take up positions income poverty and household size i s in areas with a lack o f houses, with poor found between household wealth and fer- schools, and with low-quality medical fa- tility as well (figure 12.11). The total cilities i s difficult. number of births is more than twice as high in the poorest quintile as in the Household Size and Fertility wealthiest quintile. The differences in the number of births is also distinct between Another important element in strengthen- the poorer rural and the wealthier urban ing the population, especially the chil- areas. 246 Enhancing the Capacity of the Poor fo Participate in Growth Table 12.8. Ordinary Least Squares Regression of EducationGap of Children Age 8-1 1 Coefficient T-statistic Number of household members age 0-6 0.0502 6.0 Number of household members age 7-15 0.0036 0.5 Number of household membersage 16 and above -0.0168 -2.9 Age of child 0.4961 50.3 D-secondwealth quintile -0.1998 -7.0 D-third wealth quintile -0.4484 -1 4.6 D-fourth wealth quintile -0.6218 -1 8.5 D-fifthwealth quintile -0.9130 -22.5 Constant -3.0194 -30.9 Source:Authors' calculations based on National Bureauof Statistics 2002. Note: The education gap measuresthe number of years missed by the child and is defined as the years of education minus the age of the child minus 7. Figure 12.1I. per Woman Age 15-49, by Wealth Quintile Birth 6.0 5.0 4.0 2.0 0.0 1.0 2.0 3.0 4.0 5.0 urban rural Wealthquintile Source: Gwatkin and others 2003 from 1999 DHS data. There are additional reasons to consider rather than actual birth spacing prevailed, the consequences of high fertility. Having then the fraction o f children with a birth a large family (or high dependency ratio) interval o f less than two years would drop lowers-all other factors beingthe same- by 27 percent (from 13.2 percent to 9.6 the savings rate, which means that the percent), which would lead to an estimated capital stock cannot expand as fast as it decline in neonatal mortality o f 7 percent could have. High fertility thus leads to a and in under-five mortality by 3 percent decline in per worker output, which, as (Rafalimanana and Westoff 2001). empirical analysis has shown, results in the decline inthe growth of per capita in- High fertility also contributes to inequali- come (Mkenda 2005). ties. It causes gender inequalities because, in a society where female life expectancy Highfertility is also associatedwith short at birthi s 56 years andwhere women from birthintervals, which, inturn, causehigher the poorest quintile have a total fertility child mortality. However, if preferred rate o f close to eight (not counting unsuc- 247 Elements of a Strategy for Shared Growth cessful pregnancies), about one-half o f a with a drop in infant mortality, the situa- woman's adult life i s spent either carrying tion causes concern as it suggests an ac- a child in her womb or breastfeeding it celeration in population growth, which (Dasgupta 1995). And because poor fami- underlines the case for an active popula- lies spend less on their children's human tion policy. There i s certainly scope for resource endowments than do wealthy improvement. Total fertility could fall families, and because well-endowed chil- considerably, even if only the difference dren stand a better chance of earning between total wanted fertility rate and higher income, high fertility contributes to actual fertility was closed. The poor the perpetuation and possibly even wors- would benefit most from the gap closing ening o f existinginequalities. because the difference for them i s largest: 5.6 versus 6.5." The fact that high fertility can engender negative consequences for human resource Building Physical Capital building, inequality, and economic growth of the Poor i s already recognized in policy. The Na- tional Population Policy states, "Rapid For the poor to participate in growth, population growth tends to increase out- building human capital alone is insuffi- lays on private and public consumption, cient. Complementary assets need to be drawing resources away from saving for created. productive investment and, therefore, tends to retard growth in national output In the longer run, one expects the econ- through slow capital formation" (United omy to transform from an agricultural one Republic o f Tanzania 1992). The policy to a service- andmanufacturing-based one. also recognizes the pressure on national A strategy that enhances human capital resources that population growth engen- prepares for that transition, and building ders: "the strains caused by rapid popula- assets outside agriculture seems a logical tion growth are felt most acutely and visi- first step. However, the majority o f people bly in the public budgets for health, work as own-account workers in agricul- education, and related fields o f human re- ture (figure 12.12), and only a small frac- source development. The need to feed a tion are employed in the secondary and rapidly growing population also means tertiary informal and formal sectors. So to that part o f the gains from increased agri- address poverty inthe meantime (and how cultural production [is] eroded" (United long does the meantime last?), a focus is Republic o f Tanzania 1992). needed on strategies that allow poor households to accumulate physical capital The contraceptive prevalence rate has in- so as to expand and improve their on- and creased. The length o f actual birth inter- off-farm enterprises. vals increased from 30 months in the 1970s to 33 months in the 1980s and 35 Such a rural development strategy has months in the 1990s. Yet after an initial proved to be very successful elsewhere drop in the fertility rate in the early (Klasen 2003). Evidence from rapidly 1990s, the decline stopped. Total fertility growing East Asian countries shows that per women declined from 6.3 in 1991 to poverty reduction i s largest when growth 5.8 in 1996 and remained statistically un- makes use o f the assets that the poor pos- changed thereafter. Total fertility was 5.6 sess (DrBze and Sen 1989; Ravallion and in 1999 and 5.7 in 2004. In combination Datt 1996). And when poor households 248 Enhancing the Capacifyof the Poor to Participate in Growth Figure 12.12. Fraction of PeopleWorking as Own-Account Laborers in Agriculture 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Source: Data from 2002 census. build their physical capital and increase Risk, Growth, and Asset Accumulation their incomes, a virtuous cycle is started o f enhanced resources that are used for Risks are pervasive in Tanzania. Disease, additional investments in physical and fluctuating prices, erratic availability o f human capital, which, in turn, lead to marketing opportunities, climate variabil- higher incomes. ity, uncertainty around governance, and the loss o f major assets through theft, fire, Inprevious chapters, we discussedvarious death o f livestock or otherwise have major ways to strengthenthe incentives to invest, effects on the lives o f poor and non-poor such as higher agricultural prices, im- Tanzanians. In Kilimanjaro, for instance, proved infrastructure, and greater access to serious adult illness was found to lead to a markets. In short, those improvements in reduction in per capita consumption o f up the business climate increase returns to to 17percent (Christiaensen, Hofmann, and income-generating activities and make in- Sanis 2004). For Kagera, evidence shows vesting inassets or new income-generating that chronic illness leads to a 6 percent de- activities more attractive. As box 12.2 il- cline inconsumption growth (Rossi 2004). lustrates, investments in assets are less likely to be made without such improve- Exposure to risk contributes to large vari- ments.This chapter focuses on one aspect ability in well-being over time. Consider, o f building physical capital that has re- for instance, table 12.9, which shows the ceived little attention so far: exposure to transition between consumption quintiles risk. We also discuss another aspect of for a sample o f individuals from Kagera buildingphysical capital: access to finan- that was interviewedin 1994 and tracked cial markets. and interviewed again in 2004.12 It shows 249 Elementsof a Sfrafegyfor SharedGrowth Without proper incentives, investments inassets and improvedproduction technologies will not be made. The adoption o f improvedbanana varieties inKagera was less successful inareas with limitedmarketing opportunities. Farmers livinginareas where they hadto travel far to sell their bananas preferredto hack their unsoldbananasto pieces rather than carry the bananas on the ex- haustingjourney home. That practice hinderedthe successful adoption o fnew banana varieties (personal communication, Joachim DeWeerdt, March 2004, inBukoba). A qualitative study on income mobility (Kessy 2004) illustratesthe consequenceso flack o fmar- keting opportunities well: Duringavillage transect walk, the research teamobservedbananas left to rot inthe farms. This was especiallythe casewith matooke, the staple type ofbananathat is not usedinlocalbrewing.Not only was the bananamarket aproblem:respondents mentionedthat sometimes maizei s usedto feed chickenbecausethere is no market. The market for cottoni s also aproblembut the situationi s improving throughpri- vate traders visiting the village. A market for tobacco is readily available. Table 12.9. Consumption Transition Matrix in Kagera, 1994 and 2004 Quintile,2004 Quintile, 1994 Lowest (%) Second (%) Middle (%) Fourth (%) Highest (%) Total (%) Lowest 7 3 4 4 2 20 Second 6 4 4 3 3 20 Middle 5 3 4 3 5 20 Fourth 5 3 3 3 6 20 Highest 1 2 3 5 9 20 Total 24 15 18 18 25 100 Source: Authors' calculationsbasedon Kagera Healthand DemographicSurvey data for 1994 and 2004. some persistence in that individuals re- There are at least three ways in which ex- main in their initial quintile. Yet the per- posure to risk affects income generation centages on the diagonal, which reflect no and the accumulation o f assets. First and change inwealth status, are relatively low. best known are the consequences (or ex Of those in the bottom quintile in 2004, post effects) that follow from the materi- one out o f four (6 percent) originated from alization o f risk as a shock. A shock can the top two quintiles in 1994. With such lead to direct losseso f assets (for example, movements across wealth classes, it fol- livestock die) or indirect losses because lows that considerable uncertainty exists the household is forced to lay off assets to about one's future well-being. Arguably, deal with the crisis (for example, cattle are exposure to risk i s such that considerations sold to buy grain after a harvest has o f risk inform many economic decisions, failed). l3 including the choice o f the income portfo- lio, the amount to save, or the assets to in- Inaddition to those ex post effects, expo- vest in. sure to risk has an effect on accumulation 250 Enhancing the Capacityof the Poor to Participate in Growth decisions even before it materializes. Risk affects poor households dispropor- Those ex ante effects are potentially even tionately. One reason i s that once shocks more costly, though less visibly so, than occur, poor households have less ability to the ex post effects. Inthe presence o f risk, cope. In Kilimanjaro, following the low households may try to minimize their ex- coffee prices, for instance, poor house- posure and, in doing so, allow the compo- holds with few assets for coping were sition o f their income portfolio to be in- forced to uproot their coffee trees to make formed less by profitability and more by room for other crops. However, wealthier the security provided by the resulting in- households were able to maintain their come stream. Farmers, for instance, prefer trees and wait for betterprices (Christiaen- to grow safe but low-return crops over sen, Hofmann, and Sarris 2004). high-return but risky crops. Households may opt to invest their savings in assets In addition, because the poor have less that can be easily sold during a crisis (cash ability to cope with shocks, ex post, they and livestock) rather than in the most pro- bear the high costs o f an ex ante risk- ductive ones (bicycle, plow, or sewing avoidance strategy. In Shinyanga, for in- machine). stance, households with limitedoptions for consumption smoothing have been found Exposure to risk not only has static conse- to grow lower-return, safer crops such as quences in that the loss o f assets pushes sweet potatoes, sorghum, and millet. people back on their accumulation path or Wealthier households with a greater num- inthat people opt for less profitable means ber o f options for ex post coping are like- o f income generation, but it also has dy- lier to cultivate more profitable but risky namic consequences. It may lead to lower crops such as cotton and paddy. The cost growth because, in anticipation o f risk, to the poor o f such a diversification strat- households may save and invest less than egy is high. Depending on the area they would do planted, some farmers forgo up to 20 per- cent o f their income (Dercon 1996). It Little empirical work i s available that es- should be noted, however, that shocks can timates the effect o f risk on asset accumu- have positive consequences, as box 12.3 lation, and no such evidence exists for illustrates. Tanzania. Elbers, Gunning, and Kinsey (2003) estimate the effect o f risk on small- Reducing Risk holder farm households in Zimbabwe that rely on rain-fed farming. They find that The negative consequences o f exposure to the mean o f the asset distribution i s 46 risk for asset accumulation and growth and percent lower than it would be in the ab- the disproportionate effect risk has on the sence o f risk and that the annual growth poor are reasons to consider what can be rate would be 20 to 50 percent higher in done to reduce risk and exposure to it. To the absence o f risk. The ex ante effect is that end, one should consider the risks that the most important and explains 33 per- affect households most. Information from cent o f the growth shortfall: the remaining Kagera i s informative in that respect be- 13 percent i s attributable to the ex post cause households were asked to identify effect. Other empirical work (Rosenzweig those shocks that had a major effect on and Wolpin 1993) similarly suggests large their well-being during the past 10 years costs o f exposure to risk. (table 12.10). 251 Elementsof a Strategy for Shared Growth Among the top two events that ledto increased economic prosperity inMkalanga village inRu- vuma was the outbreak o f a crop disease known as gray leaf spot (GLS). The GLS outbreak oc- curred in 1998 and lasted four years. Together with the failure by the Mbinga Cooperative Union (MBICU) to pay farmers in 1997/98 and the break up o f MBICU in 1998, GLS was a major con- tributor to the hunger inMkalanga village in 1998 and 1999. The GLS outbreak, however, also ledto innovations inthe farming system. Before the outbreak, cassava was grown only inthe bordering western lowlandbelt (Lake Shore) o f Lake Nyasa. Be- cause cassava appeared to be less susceptible to the disease, it became more widely adopted. Be- fore then, many believedthat cassavacouldnot grow inthe cold Livingstone Mountains, where Mkalangai s located. Since then, cassavahas been growing very well, and it i s now cultivated in large volumes. Source: Kessy and Mashindano 2005. Table 12.10. Shockswith Major Consequencesfor Well-Being in Kagera, by Quintile 1994-2004 percentage of households experiencing incident Quintile Consequence Lowest Second Middle Fourth Highest Average Death of family member , 31.4 29.6 31.1 30.7 34.3 31.3 Poor harvest becauseof weather 20.6 25.5 13.3 12.0 16.1 17.7 Serious illness 16.9 19.1 16.0 19.9 13.2 17.2 Loss of assets 6.1 4.9 5.8 6.4 7.1 6.0 Few opportunitiesfor wage employment 5.0 4.6 3.8 7.7 8.6 5.7 Poor harvest (for other reasons) 2.8 5.4 5.0 3.7 1.I 3.8 Low crop prices 1.4 2.8 7.3 2.8 2.9 3.5 Evictionor resettlement 3.1 2.1 2.3 2.5 3.2 2.6 Off-farm employment 1.7 1.o 2.0 0.9 1.8 1.5 Remittances 1.4 0.0 3.8 0.6 1.4 1.5 Other 9.7 4.9 9.8 12.9 10.4 9.4 Source: Authors' calculations based on Kagera Health and DemographicSurvey data for 2004. Death and illness make up about 50 per- vey (Beegle, D e Weerdt, and Dercon cent o f the major shocks affecting house- 2006). Those include governance risks holds; poor harvest and low crop prices (Kessy 2004; United Republic o f Tanzania explain another 25 percent o f all shocks. 2 0 0 4 ~which vary from sins o f omission, ) ~ Less important, but nonnegligible are such as substandard service delivery (clin- shocks related to the labor market (7 per- ics that run out o f medication, absence of cent). Some are risks that feature promi- veterinary services, extension workers nently inqualitative analysesbut less so in who do not show up, and roads that are the Kagera Health and Demographic Sur- not maintained), to sins o f commission, 252 Enhancingthe Capacityof the Poor fo Participatein Growth Table 12.11. Five Main Causes of Mortality by Age Group Children under 5 Age 5-14 Age 15-59 Age 60+ Malaria Malaria HIV/AIDS or Malaria Still birth Diarrhea tuberculosis Diarrhea Perinatalcauses HIV/AIDS or tuberculo- Malaria Heart problems Diarrhea sis Diarrhea Acute respiratory Acute respiratory Acute respiratory Heart problems infections infections infections Unintentional injuries Neoplasms Unintentionalinjuries Source: Lorenz and Mpemba2005. such as harassment by government o f f - stance, how the price o f cows expressed in cials or inhibiting rules and regulations. maize declines to one-third the normal value Theft-particularly o f movable assets, during drought years and to one-twelfth the livestock, bicycles, and cash-also fea- value duringextreme droughts. tures relatively prominently. Agricultural research can also contribute An important point is that many risks are to a reduced effect o f weather shocks if preventable or their effect can be reduced. high-yielding, drought-resistant crop va- Governance-related risks can be avoided rieties are developed. Experience from altogether because they are human made. Dodoma suggests that even if high- The most important diseases and causes o f yielding crop varieties are less marketable, death (table 12.11) can also be treated effec- farmers very much value the security pro- tively (malaria, diarrhea, acute respiratory vided, which, in turn, frees resources to infections, and even HIV/AIDS) or are pre- invest more inhigher-value activities. ventable (malaria, diarrhea, and HIV/AIDS) by stressing the importance o f promoting That experience shows that much can be insecticide-treated bednets and safe sex, by done to reduce risk through existing sector providing access to clean water, and by en- policies. It also shows that a key element of suring that the health sector can effectively a strategy that reduces downward economic deliver treatments for the most prominent mobility and enhances growth opportuni- diseasesandcauses o fdeath. ties for poor and non-poor Tanzanians alike i s an effort to reduce exposure to risk or its Weather and price risks cannot be pre- consequences in various sectors, including vented, but their effects can be mitigated. health, water, nutrition, agriculture, inf7a- One way is through irrigation. Another i s structure, extension, and education. through improved infkastructure, access to markets, and storage facilities. Especially in Asset Accumulation and Financial isolated markets, climatic shocks lead to Markets large changes in prices. Especially after a weather shock, the livestock and food terms Another element in a strategy to promote o f trade tend to deteriorate because failed investments i s the improvement o f financial harvests result inan excess demand for food markets. Financial markets play a crucial and an excess supply of livestock. The 2004 role in facilitating growth in at least three Participatory Poverty Assessment (United ways: (a) access to savings accounts per- Republic o f Tanzania 2004c) notes, for in- mits people to store their cash in a safe 253 Elements of a Strategy for Shared Growth Table 12.12. Access to Savings Services in Rural Areas, 1991 and 2000 Type of service 1991 (%) 2000 (%) Savingsor current account 12.9 3.9 Informalsavings group 3.6 2.8 Source: NationalBureauof Statistics2002. place and to slowly build sufficient capital forms o f insurance that cover actual losses. to buy expensive assets, (b) credit allows Large fixed costs and volumes make fu- people to buy assets even before they have tures markets unavailable for small farm- accumulated sufficient savings, and (c) ac- ers. The absence o f collateral also prevents cess to insurance reduces the effect o f risk credit markets from developing. The credit on economic decisionmaking. and insurance that is provided is mostly informal and i s based on high observabil- Unfortunately, access to financial markets ity andrepeated interaction. i s very limitedand virtually nonexistent in rural areas. Insurance and credit markets That situation does not mean that formal are ill developed, and even savings ac- forms o f financial services cannot be de- counts are used very little. At a mean dis- veloped. Greater access to formal savings tance to a bank o f more than 30 kilome- mechanisms are possible, for instance, by ters, households' limited use o f savings relying on local institutions, such funeral accounts and other financial services is groups, that already manage financial re- unsurprising.And with the restructuring o f sources (see Dercon and others 2004). the banking sector during the 199Os, lim- Mobile banks and cell phone technology ited access to formal savings declined also present possibilitiesto increase access even further and was not compensated by to savings. Numerous other initiatives ex- an associated increase in informal savings ist that try to overcome the constraints to groups (table 12.12). financial markets by focusing on group responsibility to overcome collateral con- Whereasthe loss inaccess to formal savings straints (microcredit initiatives), by offer- increased the obstacles to asset accumula- ing insurance contracts against indexes (as tion, low inflation meant that cash could be- i s the case o f weather-based insurance come a more attractive store o f wealth. contracts), by using lease constructions or Kessy (2004) reports how cash savings are by organizing rotating saving and credit an important means for consumption associations (ROSCAs) and savings and smoothing. Christiaensen, Hohann, and credit cooperatives (SACCOs). Sarris (2004) note that inKilimanjaromone- tary savingsarethe most important means of The literature on those various arrange- coping with the coffee price shock.l5 ments i s adequately surveyed elsewhere (for example, Larson, Anderson, and Va- Despite the great advantages that func- rangis 2004). Suffice it to say that finan- tional rural financial markets would bring, cial markets do matter; that great gains can there are structural reasons for the highly be reaped if they can be improved, even imperfect financial markets in rural areas. marginally; and that various very promis- Geographic isolation, moral hazard, and ing initiatives exist in that area that war- the high cost o f information preclude rant attention. 254 Enhancing the Capacityof the Poor to Participatein Growth Dealing with Vulnerability 1994, about one-half managed to improve their well-being and enter the middle quin- One vision o f economic growth considers tile or higher. Surely a social protection poverty a transitory phenomenon. In the strategy has to focus on those households long run, everyone will converge toward that are not able to move upward. But how an equilibrium steady state, and if invest- do we identify those households or indi- ment opportunities with sufficiently high viduals? A useful concept inthis context i s returns exist, then this steady state will lie that o f a poverty trap, which may be de- above the poverty line. However, certain fined as a condition inwhich an individual groups lack the ability to benefit from i s pushed below the poverty line and i s growth opportunities. In this section, we unable to climb out o f poverty without ex- consider those vulnerable groups, which ternal assistance. Yet once assistance i s could comprise orphans, people with dis- provided, the individual should be able to abilities, and households headed by the sustain a living above the poverty line elderly or a child, but which also could unless pushedback again. simply consist o f a high concentration o f poor people (Klasen 2003). The concept o f a poverty trap i s intuitive, untilone asks why individuals are not able Social Protection and Vulnerable to climb out o f poverty themselves. After Groups all, in an environment where investments provide sufficiently high returns, people A common reflex inthinking about vulner- have a strong incentive to benefit from able groups i s to call for a safety net. How- such opportunities. If that benefit requires ever, vulnerability i s probably best ad- investments, they will have a strong incen- dressed through a mix o f economic growth, tive to save to self-finance the investment. attention to risk reduction, and a selective Because marginal returns to capital are use o f safety nets. In an economy in which also high, especially when few assets are income i s low and inequality limited, in- available, one expects poor households to come growth i s a prerequisite to moving have a large incentive to save and invest. people away from the poverty line. By trying to understand how poverty traps However, as the poverty transitions o f ta- come about, one can gain insight in the ble 12.9 have shown, a large group o f type o f interventions that would allow peo- households experience downward mobility ple to move out o f poverty. Typically, the inwealth status. Ifsuch downward move- presence o f a poverty trap requires the exis- ment could be reduced, much poverty tence of one or more critical wealth thresh- could be prevented. Hence, reducing risk olds that people have difficulty crossing i s another important element in a strategy from below. The presence o f a threshold by to reduce vulnerability. Such interventions itself i s not sufficient to explain the pres- not only are part o f an economic growth ence o f a poverty trap, because even if in- strategy but also are essential to social pro- vestments are lumpy, households could tection. slowly accumulate wealth and purchase the investment good later. Hence, the presence Safety Nets o f a wealth threshold has to be com- pounded by something else-for instance, Returning to table 12.9, one sees that o f the lack o f safe savings instruments, the those households in the bottom quintile in inability to save from low income (because 255 Elements of a Strategy for Shared Growth Box 12.4. A Poverty Trap in Shinyanga InShinyanga, cattleare ahigh-returninvestment (25 to 30percent annually). Cattleare also aliquid asset that canbe usedfor consumptionsmoothing, which makes cattleownershipattractive.But they are also alumpyinvestment.Wealthier ruralhouseholdshavebeenfoundto specializeincattle rearing, while poorerhouseholdsderive alarger share oftheir income from off-farm activities. Dif- ferences incomparative advantagedo not offer a convincing explanationfor this phenomenon. Householdsspecializinginoff-farm activities havemuchlower incomesbut are unlikelynotto have the skills requiredbecausecattle rearing is atraditional activity inthe area. The lack of credit markets andthe indivisibility of cattle implythat householdsmust be able to putuprelativelylarge amountsofmoney to investincattlerearing. However, poor households with low initial endowments from which only low incomes are earnedfindit hardto save enough to investincattle. That problemis exacerbatedby the fact that, becauseof low endowments, the poor have limitedability to cope with shocks. Consequently, suchhouseholdsenter into safe, lower-return activities, making saving evenharder. That combination of factors explains why poorer households specializeinoff-farm activities (suchas weeding or casuallabor) that require few skills or investmentsbut are safe. That pattern effectively traps poor householdsinpoverty, despite the attractive investment opportunities that exist inthe area. Source: Dercon 1998. o f minimumconsumption requirements), or fect credit markets, and the absence o f safe the malfunctioning o f credit markets. assets. Ifthe causes for a poverty trap have been identified, one needs to decide whether A typical threshold is minimum require- to assist those trapped (for example, through ments in nutrition, education, and nonfood asset transfers) or to attack the causes o f the consumption items such as clothing that are trap itself. A few illustrations follow. needed before a person can participate in the labor market. Because borrowing i s dif- Lumpiness in combination with inadequate ficult in Tanzania in general-specially credit markets may prevent poor households for poor people-once they are destitute, fiom entering high-return activities (box poor people may be permanently excluded 12.4). One response could then be transfers, fiom entering a home-based growth path or inthis caseinthe form oflivestock. Another participating inthe labor market. would be to addressthe lumpiness constraint by improving the means o f capital accumu- Addressing poverty traps i s attractive be- lationby stimulation o f ROSCAs and SAC- cause the interventions are temporary and COS;through the provision o f credit or op- the benefits permanent. Social protection tions for leasing; or, as i s suggested in that manages to address poverty traps United Republic o f Tanzania, Vice Presi- strengthens the economic self-reliance o f dent's Office (2005), through vertically in- the poor and their ability to invest. tegrated agricultural production such as out- grower schemes, whereby a processor It requires, however, the identification o f a provides associated farmers with inputs and poverty trap, which is not always evident, access to technology. especially because poverty traps are almost always the result o f a combination o f fac- Lack o f a clear regulatory framework, un- tors, typically including a threshold, imper- derpricing, and weak enforcement lead to 256 Enhancing the Capacity of the Poor to Participate in Growth an irreversible decline o f common prop- tential interventions. For instance, ifdifficul- erty resources, such as the fishstock inthe ties inundertakingcollective action explain Exclusive Economic Zone, a decline that the limitedpresence of ROSCAs and SAC- threatens to permanently disenfranchise COS,public interventions to convenepeople the local fishing communities from their may help. But if the issue is one of lack of main source o f livelihood. Prevention skills or the absence o f safe places to store through the promotion o f good govern- cash, an entirely different course o f action ance, a clear regulatory framework that i s shouldbetaken. enforced, and proper pricing are policy interventions that may be preferred over Likewise, the high opportunity costs and the provisiono ftransfers (COW 2005). discount rates o f poor parents and the in- ability o f young children to borrow for Adults require a threshold of physical and their own capacity building affect human human capital to be productive. That capi- capital accumulation o f children. One way tal i s typically obtained during childhood, to intervene i s to eliminate any financial whenone is not yet able to decide for one- barriers to education, as was done with self or to borrow against future income. PEDP for primary education. Hence, there i s an economic argument to provide transfers inthe form o f education, Ifone is clear as to why and how to inter- nutrition interventions, or assistance to vene, then priorities may have to be set for orphans and homeless children.16 whom to target. The HIV/AIDS crisis has placed much attention on the plight o f or- Once one has identifiedthe causes o f a pov- phans. According to the 2002 census, erty trap and discovered why households are 493,000 children (or 1.5 percent o f the total not able to deal with the poverty trap them- population) between ages 7 and 14 are selves (after all, doing so would be very at- double orphans, and their educational at- tractive because it would allow people to tendance lags behind that o f nonorphans attain their productive potential and escape (figure 12.13). However, an almost equally poverty), one should obtain guidance on po- large group o f 438,000 children with dis- Figure 12.13. Disabilityand Orphanhood Relativeto Primary School Attendance i 1 I HDisabled 0Orphans HNondisadvantaged 7 9 I 1s 16 17 Source: Data from 2002 census. 257 Elements of a Strategy for Shared Growth Box 12.5. Analysis Helps Clarify Whom to Target Fewpeople would argue against the case for providingsafety nets to people with disabilities. Yet a studyby Lindeboom (2005) suggeststhat the incidence o f poverty among people with disabili- ties i s only somewhat higher than that among the population at large. Forty percent o f those living inahouseholdinwhichthe householdheadis disabledare poor compared with 34 percent of those livingina householdheadedby a personwho does not have a disability. As the figure below illustrates, disabilities often occur late inlife. For most people, disabilities occur after the age when they attend school; oftentimes, disabilities (for example, blindness) oc- cur toward the end o f one's productive life. The implicationi s that many people with disabilities are disadvantaged intheir ability to earn an income but are not disadvantaged intheir educational attainment or job experience. That situation helps explain, inpart, why poverty i s only slightly higher among households inwhich the householdhead i s disabled. Number of People in DifferentAge Groups Reporting Disability at the Time of the Census 50 - 45 - -Fhysically/leprosy 40 - -visually inpaired .P 35 - C C m ...__., durrb 8" 20 25 ."__"__^_Hearinglspeech r inpaired $ 15 n albino 10 5 mentally handicapped 0 0 10 20 30 40 50 60 70 80 90 100 rmltiple handicapped age Source: Data from 2002 census. Further analysis, presented inthe main text, shows that children with disabilities are severely dis- advantaged intheir schooling attainment. And inchapter 3, we have shownhow educational at- tainment i s one o f the key determining factors for one's ability to earn an income. That evidence is cause for concernbecause the combination o fhaving a disability and having low educational attainment probably presents apoverty trap that i s hard to escape. abilities receives a lot less attention despite dren with disabilities (see also box 12.5), their worse plight.17 If the conclusion is then specific measures in the context of that more attention should be paid to chil- PEDP may be called for. But again, under- 258 Enhancing the Capacity of the Poor to Participate in Growth standing the cause for nonaccess is most vides access to markets, and i s embedded important. Depending on the type o f dis- in an environment with predictable gov- ability, other measures, including medica- ernance. Risk and financial markets are tion (for diseases such as leprosy) or provi- argued to be additional, important ele- sion o fprostheses, may be more effective. ments to a rural investment strategy. Conclusions Risk matters because it is a structural de- terminant o f poverty and because its pres- Enhancing the capacity o f the poor to par- ence reduces the long-run value o f the ticipate in growth requires building human capital stock and, consequently, growth. capital as well as accumulating physical We have shown that much can be done to assets. With regardto buildinghumancapi- reduce exposure to risk through sector in- tal, considerable progress has been made in terventions. Dealing with risk should education-primary education especially. It therefore be considered an important will take time, however, for the current in- crosscutting issue o f relevance to many vestments in primary education to contrib- sectors, including health, infrastructure, ute to poverty reduction, which i s a reason agriculture, andwater. to pay attentionto adult education also. The financial sector plays an important role Some progress has been made in health, inpromoting investments, especially when but major challenges remain. Tanzanians insurance instruments to deal with risk and are very often sick. Illness and untimely access to credit can be provided. Because death have large social and considerable those products are difficult to develop, es- economic consequences. It has been ar- pecially in a rural context, other financial gued that improving the health o f Tanza- products such as savings promotions or mi- nians is a responsibility that goes beyond crocredit and leasing may be considered. the health sector and also involves sectors Andbecausemanypeople resort to keeping such as education, extension, or water. savings in the form o f cash, maintaining low inflation will allow poor households to Nutrition and population growth are areas accumulate through cash savings. receiving little policy attention, despite their importance for human capacity build- Finally, the discussion o f social safety nets ing. Whereas the trends with respect to suggests that reducing risk i s not only key fertility are encouraging, no improvements for a growth strategy but also for social in nutritional status have been observed protection: reduced exposure to risk will since the early 1990s. Because only lim- prevent households from being pushed ited benefits can be expected from the cur- back on their accumulation path. Social rent episode o f high economic growth, protection therefore needs to look beyond large-scale interventions in nutrition seem safety nets and to engage sectors so as to justified. International evidence suggests promote risk-reducing policies (World that such interventions can be economi- Bank 2001). It has been argued that reduc- cally attractive because they are associated ingthe plight o f vulnerable groups requires with very highbenefit-cost ratios. an understanding o f poverty traps. Such an understanding will allow one to identify A prerequisite to building the physical as- whether transfers are needed or whether set base of poor households i s a business bottlenecks should be addressed to allow climate that promotes competition, pro- vulnerable groups to participate in growth. 259 Elements of a Strategy for Shared Growth Notes household size disappears only at a scale parameter of 0.6. And as Deaton and Zaidi (2002) argue, in a The finding that caring practices, diet, and hy- developing county where as much as three-fourths giene are important determinants o f undernutrition o f the total household budget is spent on food, little explains why income growth alone will not be suf- scope exists for economies o f scale, and an econo- ficient to address Tanzania's undernutrition prob- mies o f scale parameter o f close to 1 is to be ex- lems (chapter 4). pected. (In contrast, for industrial countries, a pa- The decline in infant mortality is contested. Sur- rameter o f 0.75 appears reasonable.) Consequently, veys carried out during the 1990s (the DHS and one can safely assume that the relationbetweenpov- TRCHS) show a stagnation o f mortality rates up to erty andhousehold size is robust. 1996 and a slight increase thereafter. Data from Such an association was not found for under- surveillance sites, on the other hand, support the nutrition. decline in infant and child mortality. The 2004 I OThe education gap i s the difference between the DHS, which will be released soon, will shed more number o f years o f education a child should have lightonthe actual trend. receivedand those actually received. There is some evidence o f a decline in the infec- Being poor is approximated here by those with tion rate among blood donors. Whether this finding no education. i s a statistical artifact or a reflection o f an actual l2 Some o f the people interviewed in 1994 could decline inthe infection rate i s not clear. not be reinterviewed in 2004. Some had passed Those percentages are likely to be underestimates. away, could not be traced, or refused to be inter- Ninety-one percent indicated not needing health viewed again. Ranking i s done by the quintile in care, yet that percentage probably also comprises 1994 using consumption expressed in 1994 prices those who did not fmd it worthwhile to seek care for those who were present inboth rounds and thus becauseo fconcerns about distance, cost, or quality. does not refer to the overall population. Apart from an intervention to teach students l3 In an environment where negative and positive proper care practices, one low-cost nutrition inter- shocks occur and where the production function is vention to reduce anemia would be deworming o f concave, exposure to shocks has negative conse- all children on their first day o f school. quences for growth because o f the asymmetric im- Those amounts are in constant 2001 Tanzanian pact it has on production: a negative shock inassets shillings. The nominal amount for 2004 i s leads to a greater loss inproduction than a positive TSh8,815. The price index for 2001 to 2004 i s shock o f equal size would add to production. calculated from table 11 o f the United Republic o f l4 Whether households save more or less following Tanzania (various years). an increase inrisk depends on the curvature o f their 7 We are not saying that the use o f antiretrovirals utility function and is essentially an empirical mat- should not be promoted or that no pro-poor ele- ter. The evidence presented for Zimbabwe suggests ments to antiretroviral treatment exist. Apart from that households save less. the cost o f the life-saving benefits, HIVIAIDS and 15 Christiaensen, Hofmann, and Sarris (2004) and the prolonged illness that precedes death from Kessy (2004) report that cash and savings are used HIV/AIDS creates a large economic burden. And in the more affluent villages; however, in poorer because many infected urban people decide to go communities, households lack the monetary reserves home, poor, rural households carry a dispropor- to do so. Presumably, the opportunity costs interms tionate share o f this burden. of forgone consumption are too high for poorer Anand and Morduch (1996) and Lanjouw and Ra- householdsto keep unproductive monetary savings. vallion (1995) have urged caution ininterpreting the l6 Remarkably, no one disputes that investing inthe apparent positive relationship between household education o f children is wise, but investing in the size and poverty at the household level. They note nutrition or care for orphans or homeless children that a scope for economies o f scale exists at the i s less accepted. household level afforded by the presence o f public l7 Whereas orphans appear less disadvantaged in goods such that an increase in the household size school attendance, Alderman, Hoogeveen, andRossi leads to a less-than-proportionate need for consump- (2006) show that orphaned children are significantly tion. However, the association between poverty and moreprone to being stuntedor underweight. 260 PART 4 Managing Policies and Expenditures for Shared Growth CHAPTER 13 Scaling Up Public Expenditure National Strategy for Growth and Reduction in Poverty and Millennium Development Goal Financing Requirements Sustaining economic growth and reducing 2005e) produce significantly lower esti- poverty will require increased levels o f in- mates o f the resource requirements, there i s vestment in public infrastructure and hu- nonetheless a growing consensus that mancapital. The UnitedNations (UN)Mil- achieving the targets for the MDGs andthe lennium Project (2005) estimates the National Strategy for Growth and Reduc- annual cost o f achieving the Millennium tion inPoverty (NSGRP) will require a sig- Development Goals (MDG) targets in Tan- nificant increase in public expenditure. Per zania as increasing from US$82 per capita capita government expenditure in 2004/05 in 2006 to US$161 by 2015 (in constant i s about US$90, o f which about US$39 i s 2003 dollars), as shown intable 13.1. Even directly related to the achievement o f the though other methodologies (World Bank MDGtargets (both domestically and donor 263 Managing Policies and Expenditures for Shared Growth Table 13.1. Per Capita MDG Investment Needs and Financing Sources, 2006-1 5 Amount (2003 US$) 2006 2010 2015 MDG investment needs Hunger 4 7 14 Education 11 13 17 Gender equality 2 3 3 Health 24 33 48 Water supply and sanitation 4 5 12 Improvementof lives of slum dwellers 3 3 4 Energy 14 15 18 Roads 13 21 31 Other 8 9 13 Total 82 111 161 Sources of financing Householdcontributions 9 11 17 Government expenditures 24 32 46 MDG financinggap 50 67 98 Current ODA for direct MDG support 15 15 15 Shortfall of ODA for direct MDGsupport over 2002 level 35 52 83 Source: UN MillenniumProject 2005. financed), according to the Millennium eign aid. Each source has a different mac- Project estimates. roeconomic cost. Conceptually, as long as the marginal benefits o fgovernment spend- Although there is uncertainty about the ac- ing are higherthan the marginalcost o f the tual financing need andthe capacity o f Tan- resources, it i s advisable to increase spend- zania to scale up MDG-related activities, ing. The marginal benefit o f additional given human resource constraints, there is government spending i s likely to be declin- little doubt that significantly increased re- ing while the marginal cost of finance is source availability would be a necessary, increasing. Policy making in Tanzania i s although by no means sufficient, condition typically based on the assumption that the for accelerating progress toward the MDG marginal costs o f resources are relatively targets. The challenge is to progress toward low for foreign aid inflows (but constrained the targets while managing the macroeco- by the amount donors are willing to pro- nomic implications for an overall favorable vide) and for domestic revenue. Since development outcome, with options strongly 1995, when the government's primary ob- grounded inthe microeconomic and sectoral jective was to achieve fiscal stabilization underpinnings of the economy. In the fol- and sustainability, the costs o f domestic lowing sections, we look at the prospects borrowing (in the form o f crowding out and implications of scaling up spending and private credit and generating inflationary resourcemobilization. pressures) and o f foreign, nonconcessional borrowing were considered exceedingly Potential sources o f financing for govern- high. Thus, they were excluded fkom the ment expenditure are taxation, seignorage, financing package. However, as Tanzania domestic and foreign borrowing, and for- has achieved macroeconomic stabilization 264 Scaling Up Public Expenditure Table 13.2. PotentialContributionof Domestic Revenueto Finance MDGs, 2006-25 Revenue per capita MDG spending per capita Per capita GDP (2003 US$) (2003 US$) growth (%) 2006 2015 2025 2006 2015 2025 Scenario I: Constantrevenue to GDP ratio 2 45 55 67 24 29 36 4 45 66 98 24 35 53 6 45 80 144 24 43 77 Scenario 11:Increase in revenue to GDP ratio to 20% by 2015 2 45 78 95 24 42 51 4 45 95 140 24 51 75 6 45 115 205 24 62 110 Source: Author's calculations. and is puttingmore emphasis on achieving nomic growth by itself is an important high growth, there is also scope to recon- source o f increased revenue. GDP per capita sider the appropriateness o f fiscal deficit growth rates o f 2, 4, and 6 percent would targets, especially in light o f highneeds for lead to increases in per capita government spending on economic infrastructure. revenue to US$67, US$98, and US$144, respectively, by 2025. This projection as- In the following sections, we look at the sumes that the revenue-to-GDP ratio re- potential for expanding domestic resources mains constant at 14 percent throughout the and aid inflows. Cost sharing and private period. Ifrevenuewere to increase gradually participation inservice delivery, especially to 20 percent o f GDP by 2015 andremain at infrastructure, are important complements that level, available resources would be al- to public financing. However, these issues most 50 percent higher by 2025 than if the are not discussed inthis section. revenue-to-GDPratio holds constant. Domestic Resources These simulations underline the importance o f economic growth as a means for generat- The N S G W emphasizes the importance o f ing resources for investment and poverty domestic resource mobilization as the main reduction. Although these simulations have source o f financing its implementation. Do- focused on the effect o f growth on avail- mestic resource mobilization i s also seen as able government revenue, the effect o f important inthe context o f efforts to reduce growth on incomes for households and Tanzania's aiddependency inthe mediumto firms i s equally important. More resources long run.The amount o f domestic resources for households provide the means to spend available for reaching the MDG targets de- on health and education and to save for in- pends on three parameters: the rate o f vestment. Inan environment inwhich most growth o f gross domestic product (GDP), investment i s financed from own savings the share o f GDP collected as revenue, and and retained earnings, such resources are the share o f revenue spent on MDG-related particularly important. Efforts to raise the activities. Table 13.2 shows various scenar- ratio o f revenue to GDP should therefore ios o f the evolution o f domestic revenue and carefully consider the effects o f doing so on spending on MDG-related activities. Eco- the competitiveness o f Tanzanian busi- 265 Managing Policies and Expenditures for Shared Growth nesses and on the ability o f the private sec- velopment assistance (ODA) plays an im- tor to save and invest. It i s worth emphasiz- portant role inthe economy and i s likely to ing that economic growth generates addi- remain important as long as Tanzania re- tional resources for growth and poverty mains a low-income economy. The discus- reduction, whereas an increase in the reve- sion of the financing needs o f the MDG nue-to-GDP ratio only transfers resources implementation clearly presumes a posi- from the private sector to the public sector. tive effect o f ODA on economic growth The marginaleffect o fprivate sector spend- and poverty reduction. Inthe short to me- ing on growth and poverty reduction com- dium term, aid-financed expenditures can pared with that o f public sector spending i s provide an important demand-side stimu- thus critical informing views onthe appro- lusto the economy. Inthe mediumto long priate level o f revenue generation in the term, aid-financed investments in human Tanzanian economy. and physical capital are intended to strengthen the supply side o f the economy Domestic borrowing could be another as the basis for sustained growth. source o f finance. Untilrecently, during the period o f macroeconomic stabilization, con- However, higher aid flows do not neces- cerns about crowding out credit to the pri- sarily lead to higher economic growth and vate sector and generating inflationary pres- the achievement o f targeted development sures have led the government to refrain objectives. Examples abound o f countries fi-om domestic borrowing. Those concerns that saw economic stagnation or decline remain relevant. In the context o f the despite highinflows o f foreign aid.' Inter- NSGRP, however, with increased emphasis national experience suggests that it i s criti- on poverty reduction through higheconomic cal to manage two separate but related sets growth, there i s now scope to revisit the ap- o f problems that could undermine the propriate level o f domestic borrowing- positive effect o f aid: especially since domestic debt levels are relatively low in Tanzania. Because higher 0 Weakening o f institutions growth will require significant investments in i&-astructure, an assessment of the ap- 0 Weakening o f competitiveness through propriateness o fthe fiscal position should be Dutch disease effects undertaken. It should take into account the medium- to long-term effects o f debt- A weakening o f institutions may occur if fmanced infi-astructure investment on hture aid undermines accountability to domestic growth and government revenue, as well as stakeholders, distorts incentives for public the recurrent cost implications and the effi- sector performance, or removes the pres- ciency o f government expenditure. Domes- sure for an efficient revenue collection sys- tic resource mobilization has a critical role tem. It i s thus important that scaled-up aid to play, but even under relatively optimistic goes hand in hand with reforms that im- scenarios it will be insufficient to meet all prove governance and domestic account- investment needs. ability. Inaddition, the design o f aid deliv- ery mechanisms should support domestic Scaling Up ForeignAid accountability rather than replace it with accountability to donor agencies. In this Domestic revenue i s clearly insufficient to respect, processes such as the Tanzania As- finance the implementation o f the NSGRP sistance Strategy and the Joint Assistance and to reach the MDGtargets. Official de- Strategy have an important role to play in 266 Scaling Up Public Expenditure the development o f appropriate account- 1990 and 1995, the REER was relatively ability mechanisms. Efforts to fully inte- stable, while aid inflows increased from grate foreign aid into Tanzania's budget US$l.166 billion to US$1.259 billion and system and thus make aid subject to the then declined to US$745 million. Since same accountability process as domestic then, aid inflows have been gradually in- resources are also important. creasing, to US$1.450 billion in 2003. The REERappreciatedby about 50 percent be- The effect on competitiveness occurs tween 1995 and 1998 and remained at that through a spending effect and a resource level until 2001. Since then, the REER has movement effect, which are commonly depreciatedsignificantly and is now back to labeled Dutch disease effects. A spending its 1995 level, which i s estimated to repre- effect arises from the fact that, unless all sent equilibrium. Exports o f manufactured inflows are spent on imports or exportable goods increased from US$30million in goods, the real exchange rate will increase. 1999 to US$80millionin2004. This outcome implies a deterioration o f a country's competitiveness. In addition, The absence o f a clear relationship be- there i s a resource movement effect, as tween foreign aid inflows and the REER i s resources move into the booming sector partly attributable to efforts by the gov- (that is, the government and development ernment to sterilize these inflows through bureaucracy) and the nontraded sector. the sale o f liquidity paper andthe accumu- lation o f reserves. Recently, however, the The spending effect has typically received sale o f bonds has exerted pressure on the most o f the attention in the literature. Aid interest rate and has the potential to crowd inflows to Tanzania have fluctuated signifi- out private sector investment. It i s note- cantly since the 199Os, with no apparent worthy that sterilization through the sale statistical relationship between aid inflows o f liquiditypaper has beenonly partial and and the real effective exchange rate that aid flows have resulted in relatively (REER).As figure 13.1 illustrates, between high rates o f monetary expansion. This Figure 13.1. EffectiveExchangeRates, 1990-2004 1 140 I I 35 - 120 30 A 0 100 25 0 II 0 80 20 % c B 60 155 40 1084 -+-CPI-based REER(2000=100) +ODA (%ofGDP) Source: World Bank staff calculations. 267 Managing Policies and Expenditures for Shared Growth Figure 13.2. Exports of Manufacturesas a Percentageof GDP and Exports of Goods and Services, 1990-2002 I 2.5% 1 T 16.0% 14.0% 2.0% ,;$ 12.0% v) 7 a 1.5% 10.0% ; 0 U (3 8.0% 8 'c s0 1.0% (3 6.0% 'c 4.0% I3 20n 0.5% 2.0% 5 s 0.0% 1990199119921993199419951996199719981999200020012002 -ManufactureslGDP-ManufactureslExports Source: United Republicof Tanzania various years. rapid expansion o f the money supply has sectors versus the nontraded and so-called been consistent with a low and declining booming sectors. However, since 2000, ex- rate o f inflation, since it was accompanied ports o f manufactures have beenrecovering, by financial deepening o f the economy. andmore recently, agricultural exports have Continued financial deepening o f the also started to grow (figure 13.2). economy i s thus a crucial element in in- creasing the capacity o f the economy to The resource movement effect works absorb large inflows o f aid without direct through two principal channels. The first effects on the real exchange rate. concerns the direct involvement o f Tanzani- ans in managing donor-related activities. Although most o f the discussion o f Dutch Estimates suggest that there are more than disease focuses on real exchange move- 1,000 donor-funded projects on the ground, ments, the ODA may primarily affect the ranging fiom relatively simple technical as- relative growth o f the tradable and nontrad- sistance projects to complex, multimillion able sectors, without actually causing dollar projects. Management o f donor assis- movements in the real exchange rate.2 In tance requires a staff in the offices o f the Tanzania, sectoral and balance o f payments donor agencies in Tanzania and a project developments since 1995 display symptoms staff, as well as a staff in the government that are fiequently associated with Dutch that i s devoted to managing and monitoring disease. On the sectoral side, aside fiom the donor-funded activities and ensuring that a booming mining sector, growth appears to variety o f donor requirements for financial be heavily concentrated in the nontraded management andreporting are met. sector. The balance o f payments shows a relative sharp decline in merchandise ex- Salaries for local donor agency staff mem- ports over the past decade, which could in- bers and project staff members are typi- dicate a loss o fcompetitiveness o fthe traded cally highly competitive and thus draw the 268 Scaling Up Public Expenditure best-qualified staff to those activities. crease the real resources available to Government staff members are frequently the economy, but it increases the likeli- enticed with salary top-ups or specific hood that government will crowd out benefits that derive from involvement in the private sector. donor-funded activities. Not only do such enticements draw human resources away Further liberalize imports to help in- from regular private and public sector po- crease absorption and reduce the need sitions and tasks, but they likely also ex- for real exchange appreciation or re- erts upward pressure on wages in the pub- serve accumulation. lic and private sectors, with implications for Tanzania's competitiveness as well as Focus on expenditures that will quickly for the cost o fpublic service delivery. release supply constraints and have a higher import content, including trans- Simulations o f the effect o f aid-financed port investments. increases in spending on HIV/AIDS o f Continue to improve the efficiency o f US$300 million (or 2.9 percent o f GDP) the banking sector to ameliorate the and increases in spending on education o f need for highreal interest rates. US$500 million (or 4.9 percent o f GDP) show only a small appreciation o f the real Consider adopting a more relaxed exchange rate as a result o f these foreign monetary policy. exchange inflows. Coordinate exchange rate, monetary, Although it may be difficult to establish and fiscal policy with the implications whether the low level o f manufacturing ex- o f aid inflows inmind. ports and the poor growth o f the tradables sectors in Tanzania are indeed directly re- lated to the inflows o f foreign aid and gold Debt Sustainability exports, both clearly highlight an intema- To the extent that scaled-up aid is in the tional competitiveness problem. The recent form o f concessional credits, the issue o f recovery o f manufactures exports gives debt sustainability needs to be considered. some cause for optimism, but enhancing After Tanzania received HIPC (Heavily In- the competitiveness o f the economy must debted Poor Countries) Initiative debt relief be a central element o f Tanzania's strategy in 2001 and m h e r debt relief through the to enhance and accelerate economic Multilateral Debt Relief Initiative, the ratio growth. In particular, the following set o f o f the net present value (NPV) o f debt to measures should accompany scaling up o f exports declined to about 64 percent. Panel aid (Foster and others 2005): (a) o f figure 13.3 shows the baseline projec- tions o f the net present value o f exports-to- Ensure that increased aid-financed GDP ratio. The ratio is projectedto increase spending is accompanied by increased initiallyto 71percent by2010, butto decline absorption o f the foreign exchange, subsequently to 63 percent by 2026, assum- which (assuming that government ing average export growth o f 8.5 percent spending continues to have a high local and annual growth o f concessional credit content) will probably require accep- disbursements (primarily fiom World Bank tance o f exchange rate appreciation. and Afican Development Bank credits) by Spending the aid without absorbing the 2 percent. Although debt ratios arerelatively foreign exchange does nothing to in- low, the debt sustainability analysis also 269 Managing Policies and Expenditures for Shared Growth Figure 13.3. Multilateral Credit Disbursementsand Debt Sustainability: Net PresentValue of Debt-to-Export Ratio,2006-26 (a) Growthof multilateralcredit disbursementsby 2 % 800 72 p u 700 70 E -5600 68 1 500 1 66 2E400 9 69300 64k P) 200 62 100 60 20 0 58 (b) Growth of multilateralcredit disbursementsby5 percent 1400 74 .Q 1200 72 3 1000 70 1 800 68 66 B fff 600 v) 4 400 62 6 4U 200 60 2 0 58 NW of debt-to-export ratio (%) (c) Doubling of multilateralcredit disbursementsin 2011 1600 120 + 1400 100 E -s 1200 1000 80 800 60 8 tft 2 Ei 400 ; 600 40 8 + 200 20 0 O Z Source: InternationalMonetary Fund and World Bank staff estimates 270 Scaling Up Public Expenditure suggests that Tanzaniaremains vulnerable to 0 Exposure to shortfalls in disburse- negative shocks on exports and GDP ments as compared with commitments growth, which could lead to a significant deteriorationinthe debt indicators. The conventional approach to reducing aid dependency is to argue for an increased Panels (b) and (c) of figure13.3 show the revenue effort. However, given the magni- effect o f a significant scaling up o f multi- tude o f Tanzania's aid dependency, increas- lateral credit disbursements, which would ing the revenue-to-GDP ratio is not going allow a matching increase in inputs. As to address the issue. Indeed, the only way shown in panel (b) o f figure 13.3, if dis- to reduce aid dependency is to increase bursements were to grow annually by 5 economic growth and to reach a level o f percent instead o f 2 percent, the NPV o f income at which an adequate level o f infra- debt-to-exports ratio would increase from structure and service delivery can be fi- 64 percent in 2006 to 73 percent by 2021 nanced from domestic resources. Raising and decline thereafter. the revenue-to-GDP ratio may have a nega- tive effect on economic growth and thus A doubling o f multilateral credit dis- prolong rather than reduce aid dependency. bursements in2011and subsequent annual The key challenges for Tanzania and its growth by 2 percent would result in a development partners are thus gradually increasing NPV o f the debt-to- export ratio, which would reach 101 per- 0 To adopt aid modalities that reduce cent by 2021and decline thereafter. the potential negative effect o f aid dependency In summary, from a debt sustainability perspective, there seems to be significant 0 To develop mechanisms that allow Tan- scope to expand Tanzania's fiscal space. zania to dealwith fluctuations inaid However, doing so requires that expendi- tures financed through increased borrow- Tanzania has been a pioneer in developing ing are indeed used to ensure sustained modalities that reduce the potential negative growth o f exports and GDP. The primary effects o f aid dependency. Inparticular, the constraint i s the availability o f conces- N G S W provides the overall framework that sional funds rather than their effect on sets the priorities for both domestic and for- Tanzania's debt sustainability. eign resource use. The Tanzania Assistance Strategy sets out principles that would en- Reducing Aid Dependency hance ownership and aid effectiveness, and an increasing share o f aid resources is hlly Dependency on foreign aid for financing integrated into the budget. public expenditures poses various risks for Tanzania: With respect to managing fluctuations in aid, most o f the attention has been focused on reducing them up front. All develop- 0 Distortion o f Tanzania's development priorities ment partners have made efforts to en- hance the predictability o f aid flows by 0 Distortions in incentives in the public providing more accurate informationto the sector authorities and, in the case o f general budget support, to provide the resources at Exposure to fluctuations inaid flows the beginningo f the fiscal year. 271 Managing Policies and Expenditures for Shared Growth Relatively less attention has been paid to ploit the revenue potential but to focus on dealing with fluctuations after they hap- strengthening tax administration so as to pen. Indeed, with Tanzania's cash budget be able to scale up revenue collection to system, shortfalls in aid disbursements (or compensate for a permanent decline inaid. domestic revenue) result immediately in expenditure cuts. As argued in various Conclusions Public Expenditure Reviews, though this mechanism has been instrumental in Estimated resource requirements for achieving macroeconomic stability, it has achieving the MDG and NSGRP targets a severe negative impact on the implemen- exceed available resources from both do- tation o f expenditure programs. Domestic mestic and foreign sources. Increases in borrowing may be an appropriate way to domestic resource mobilization should rely deal with unexpected fluctuations in for- primarily on expanding the revenue base eign aid. through sustained economic growth and further improvements intax administration, Permanent changes in aid flows will re- with the objective o f further progress to- quire permanent adjustments in expendi- ward an equitable and efficient tax system. ture and revenue. Hence, on the expendi- Increasing revenue through tax policy ture side, the government needs to be measures needs to be approached carefully, cautious about the composition o f expen- considering the relative efficiency o f the diture and avoid a situation in which public and private sectors and o f house- spending that i s difficult to scale back, holds to invest and spend. Concerning ex- such as wages and salaries, dominates. ternal resources in the form o f grants and Similarly, there is also a need to evaluate concessional lending, under the assumption carefully the recurrent cost implications o f o f sustained growth o f exports and GDP, donor-financed projects. Outsourcing and the binding constraint i s likely to be the public-private partnerships may also be availability o f these resources rather than useful instruments to facilitate expendi- debt sustainability. Although the effect o f ture-side adjustments in case there i s a aid inflows on the REERseems to be mod- shortfall in aid. They would also be a erate, raising the competitiveness o f the means for developing the private sector, Tanzanian economy remains nonetheless by using aid inflows to provide business critical in order to offset any aid-induced opportunities. On the revenue side, an ap- loss in competitiveness and shifts to the propriate strategy may be to not fully ex- nontradables sector. Notes 'Rajan and Subramanian (2005) provide a concise 'summaryof the debate. See discussion inRajan and Subramanian (2005). Inthe standard Swan-Salter trade model with two traded goods and one nontraded good, aid reduces the size o f the traded good sector without any changes in relative prices between tradables and nontradables. This result underscores the need to focus on quantities rather thanprices. 272 CHAPTER 14 Coordination of Economic Policy Formulation and Implementation in Tanzania Duringthe past decade, Tanzania has made The first i s to continue these orthodox significant progress in implementing eco- reforms, especially in areas in which pro- nomic reforms aimed primarily at macro- gress has been limited and the marginal economic stabilization, the liberalization o f returns to reform promise to be high. the economy, and the withdrawal o f the They include a further reduction in rent- public sector fiom commercial activities. seeking opportunities for public officials The reforms have resulted in an accelera- who interfere with private sector activi- tion o f economic growth. To sustain eco- ties and the improvement o f the policy nomic growth, policy makers face two and institutional framework for infra- challenges. structure providers. 273 Managing Policies and Expenditures for Shared Growth The secondchallenge arises from the chang- joyed significant power, particularly because . ingrole of economic managementinTanza- itplayedapivotalrole inresourceallocation nia. To date, areas for economic reforms andprice setting. have been fairly easy to identify and imple- mentation has required primarily political The initial response to the economic crisis will to overcome vested interests and a that emerged in the early 1980s was the minimum of technical competence in gov- formulation o f a National Economic Sur- ernment, often supplemented by technical vival Program (1981) followed by a assistance. However, fbture economic man- homegrown Structural Adjustment Pro- agement i s likely to require (a) close col- gram (1982), and a campaign against eco- laborationbetweenthe private sector andthe nomic saboteurs (1983). These programs government, to make economic manage- triedto address Tanzania's economic crisis ment more responsiveto the proactive iden- by intensifyingthe control regime and the tification and pursuit of new opportunities degree o f government intervention. Along- for the private sector, and (b) stronger har- side these government-led efforts was an monization and coordination among gov- intensifyingpublic debate on measures to ernment agencies to ensure consistent poli- overcome the economic crisis, which re- cies and the efficient use o f scarce human sulted in recognition o f the failure o f the and financial resources inpursuit of higher centrally planned economy. The govern- economic growth. This chapter assesses the ment started a process of reforms that state o f affairs in these two areas and pro- would change Tanzania from a command vides recommendations on how to to a market-driven economy. The failure strengthen the institutional fiamework for o f the efforts to address the crisis through policy coordination and the dialogue be- more state interventionalso ledto a loss o f tween the private andpublic sectors. confidence in planning and the role o f the PlanningCommission. Review of Institutions for Economic Policy The introduction o f far-reaching economic recovery programs in 1986 replaced me- To understand Tanzania's situation with dium-term planning with short-term eco- respect to policy coordination and plan- nomic management focused on fiscal and ning, one may find it instructive to review monetary stabilization as well as liberali- approaches to policy coordination and zation o f the economy. This reform im- planningsince independence. plied a shift in institutional responsibility and power from the Planning Commission The first comprehensive statement o f the to the Ministry o f Finance and the Bank o f country's economic policy after independ- Tanzania, which held the key responsibil- ence i s found inthe First Five-YearPlan for ity for the implementationo fthe economic Economic and Social Development, pub- recovery programs. The initial efforts and lishedin 1965.1 Central planningwas intro- relative successes in economic stabiliza- duced in Tanzania in the 1970s following tion during the second half o f the 1980s the promulgation of the Arusha Declaration were not sustained during the second term in 1967. As inother socialist countries, the o f President Ali Hassan Mwinyi. The gov- Planning Commission, with the president as ernment's inability to control credit expan- its chair, was at the center o f planning, pol- sion to public enterprises, massive tax ex- icy formulation, and coordination. Untilthe emptions, poor revenue collection, and tax mid-l980s, the Planning Commission en- evasion resulted in severe macroeconomic 274 Coordination of Economic Policy Formulation and Implementation in Tanzania disequilibria, such as large fiscal and bal- Foundation was established to represent ance o f payments deficits, high inflation, the private sector. and a decline ingrowth. The dialogue between government and the In 1996, the new, third-phase government private sector takes place in various forms. under President Benjamin Mkapa made it a First, in most policy areas the government top priority to restore macroeconomic stabil- has adopted participatory processes that al- ity and to control corruption. Inaddressing low the private sector to feed its views into these issues, the government made bold re- government policy making. Recent exam- assignments for economic management in ples are consultations on the formulation o f an attempt to ensure that the capacity and a new income tax act and the preparation of credibility o f institutions matchedthe task at the rural development and agriculture sector hand. Specifically, an extraordinary amount strategies. In addition, the government has o f power and influence in the economic established the Tanzania National Business management o f the country was transferred Council (TNBC), chaired by Tanzania's to the Bank o f Tanzania. Not only was the president, as the main forum for consultation bankresponsible for monetary management, betweenthe government andthe private sec- but with the introduction ofa cashbudget in tor. Since its establishment, the TNBC has 1996, it was also given the responsibility for very successhlly organized annual interna- determining monthly aggregate expenditure tional investor round tables, during which ceilings for the government in line with re- the government receives private sector input sources available fiom domestic revenue but also commits to and reports on progress and foreign aid. The bank also took the lead inimplementing specificpolicyreforms. in managing Tanzania's public debt. The reforms also further weakened the Planning Another significant development with re- Commission, as the responsibility for pre- gard to economic management was the paringthe development budget shifted fiom adoption in2000 o f the Poverty Reduction the PlanningCommission to the Ministry o f Strategy Paper (PRSP) process, which Finance in an effort to unify the budget and aims to align government policy and pub- strengthen budgetary control. The Rolling lic resource allocation with Tanzania's ob- Planand ForwardBudget, which until 1996 jective o f reducing poverty. The key re- hadbeenpreparedby the PlanningCommis- sponsibility within government for the sion but had since become virtually irrele- preparation and monitoring o f the PRSP i s vant for economic management, was re- with the Vice President's Office, while the placed by the Medium-Term Expenditure Ministryo f Finance takes the lead role in Framework (MTEF), which is being pre- coordinating the PRSP implementation. paredbythe Ministryo fFinance. This assignment o fresponsibilities implied a hrther weakening o f the role o f the With the shift to a market-based economy, Planning Commission (which was then the the government rightly embarked on ef- President's Office-Planning and Privati- forts to replace government planning o f zation) and a hrther fiagmentation o f eco- the economy with a process based on nomic management inTanzania. However, greater dialogue between the government with the creation o f the Ministry o f Plan- and the private sector. These efforts re- ning, Economy, and Empowerment in quired strengthening private sector institu- early 2006, the responsibility for the PRSP tions that could effectively engage in this was shifted to the newly created ministry, dialogue. The Tanzania Private Sector which also took over the responsibilities 275 Managing Policies and Expenditures for Shared Growth o f the President's Office-Planning and dressed to ensure that economic policy Privatization. making can support and sustain high eco- nomic growth and react appropriately to Finally, ongoing decentralization efforts the evolution o f Tanzania's domestic and also have a significant effect on economic international economy. management in Tanzania as responsibili- ties for the formulation, coordination, and Challenge I;Strengthen Coordination implementation o f government programs of Economic Policy Formulation and are shifted from the national level to the lmplementation district level. An example i s the redefini- tion o f the role o f government in the agri- While the Ministryo fFinance andBank o f culture sector. Under the Agriculture Sec- Tanzania are doing a commendable job tor Development Program, responsibility managing the economy for stability, coor- for the design o f agriculture sector pro- dinating and formulating the broader grams has shifted from the national level growth agenda exceeds their institutional to the district level, which i s now respon- mandates and capacities and also threatens sible for preparing district agriculture de- to dilute their focus on their core responsi- velopment plans. Similarly, responsibility bilities. Discussions with a variety of and related resources for the development stakeholders reveal a clearly perceived and maintenance o f district roads has been lack o f an institutional setup that could shifted from the Ministry o f Works to lo- performthis coordination function. cal government authorities. Challenge 2: Create a Platform for a Challenges National Dialogue on Growth-Related Issues As we have mentioned, the focus of eco- nomic policy during the late 1990s and Inline with the transformation from a cen- early 2000s has been primarily on macro- trally planned to a market-oriented econ- economic stabilization and more recently omy, the process and content o f policy on poverty reduction, with relatively little formulation and coordination clearly need emphasis on the quality o f economic to encourage dialogue and participationby growth and structural transformation. In stakeholders. Hence, the process must be addition, economic management has been an open one, rather than the preparation o f fragmented among a variety o f institu- strategies and plans behind closed doors. tions, causing a certain lack of coordina- In various areas, Tanzania has been very tion o f policy formulation and implemen- successful in establishing processes that tation. The new PRSP (the National allow for continuous dialogue between the Strategy for Growth and Poverty Reduc- government and stakeholders and that tion, or NSGRP) rightly puts greater em- draw on resources outside government for phasis on economic growth as a key the formulation o f plans and strategies. mechanism for reducing poverty and sug- Such processes include the Public Expen- gests a moreproactive role for government diture Review and MTEF processes, in in the pursuit of economic growth. This which the Ministry o f Finance has opened increased focus on economic growth raises up the budgetprocess while establishing a a number o f important institutional issues framework in which analytic work by that need to be addressed. We see six key various parties i s coordinated and brought challenges that urgently need to be ad- to bear on government processes. Simi- 276 Coordinationof Economic Policy Formulation and lmplemenfation in Tanzania larly, the Vice President'sOffice has com- the nature, source, and magnitude o f the plemented the broad consultative proc- relevant market failures; (b) the possible esses that take place duringthe preparation capture o f policy interventions by f m s o f the PRSP or progress report with the whose behavior the interventions are aimed establishment o f the Research and Analy- at regulating; and (c) the ability o f the pri- sis Working Group. The group serves as a vate sector to game policy makers when platform for an ongoing dialogue between policies suffer fiom dynamic inconsistency government and stakeholders on poverty- (that is, when the promise to withdraw sup- related issues and also as an instrumentto port fiom poorly performing activities lacks coordinate analytic work inthis area. credibility. Inthe area of economic growth and struc- Thus, to enhance the likelihood o f success tural growth, such a platform i s currently o f government interventions, policy mak- not available. This lack produces a situation ers must put appropriate governance rules inwhich a variety of initiatives proceed in and monitoring mechanisms in place that parallel with minimal interaction or coordi- would allow the weeding out o f ineffective nation. A similar setup could be envisaged interventions. Box 14.1 provides a sum- inthe areas ofeconomic growth andpolicy mary o f such design rules, as proposed by formulation. An existing or a newly estab- Hausmann and Rodrik (2005). lishedinstitutionwouldprimarilyserve as a convener o f stakeholders for the coordina- Challenge4: Redefine the tion and formulation o f economic policy, Government-Private Sector rather than as a mechanism for establishing Relationship a capacity to carry out all aspects o f policy coordination and formulation. Recent research on economic growth (Rodrik and Hausmann 2003) highlights Challenge 3: Ensure Adequate the importance o f public-private sector in- GovernanceArrangements for teractions infinding appropriate approaches Growth-EnhancingGovernment that result in higher growth at the micro- lnterventions andat the macroeconomic level. Inparticu- lar, such an approachwould require revisit- Where market failures such as technologi- ing the private-public sector interface that cal externalities, coordination externalities, takes place through central government in- or informational externalities exist, gov- stitutions such as the Tanzania Investment ernment interventions can play an impor- Center, the Tanzania Revenue Authority, tant role in fostering private sector activi- the Ministryo f Industry and Trade, and the ties and economic growth. The government agriculture sector ministries, with a view o f Tanzania has recently launched several toward redefining the interaction fiom a such initiatives, including targeted credit purelyregulatory or administrative one to a guarantee schemes, export processing problem-solving one. zones, and targeted agricultural subsidies. However, experience with such interven- Challenge 5: Strengthen Private tions also has brought to the fore a range o f Sector lnstitutions potential government failures, which can reduce the effectiveness o f such interven- As mentioned above, an important aspect of tions. Government failures can be caused economic policy formulation is input by the by (a) a lack of complete information about private sector. Having such input requires 277 Managing Policiesand Expendituresfor Shared Growth Box 14.1. GovernanceArrangements to Strengthen the Effectiveness of Growth-Enhancing Interventions Hausmann and Rodrik (2005) suggest principles that provide an initial basis for formulating an effective program: Clear criteriafor success andfailure. Not all entrepreneurial investments innew activities will pay off. Infact, only a small fraction o fbusiness ideas are likelyto be successful. From the perspectivg o fthe program objectives, however, one successcan pay for scores o f fail- ures. The programmust therefore clearly define what constitutes success and identifyobserv- able criteria for monitoring it. Otherwise, recipients o f incentives can game public agencies and can continue to receive support despite poor outcomes. The criteria should ideally depend onproductivity-both its progress and its absolute level-and not on employment or output. Although productivity can benotoriously difficult to measure, project audits by business and technical consultants at set intervals can provide useful indications. Sunset clause. The program must contemplate a built-insunset clause. Financial andhuman resources should not remain tiedup for a long time inactivities that are not paying off. Every publicly supported project must have not only a clear statement ex ante o f what constitutes success and failure, but also an automatic sunset clause for withdrawing support after an ap- propriate periodhas elapsed. Targetingof activities rather than sectors. Public interventions should support activities that suffer from externalities, not the sectorsthat confront them. This approach facilitates structur- ingthe support as a corrective to specific market failures insteado f as generic industrial poli- cies. Rather than providing incentives, say, for electronics, tourism, or call centers, govem- ment programs should subsidize bilingual training, feasibility reports for nontraditional agriculture, infrastructure investment, adaptation o f foreign technology to Tanzanian condi- tions, risk and venture capital, and so on. The government should not promote specific sectors but should support growth-enhancingactivities that often span several sectors. Similarly, the deciding factor shouldnot be the size o f the recipient enterprises. A sectoral approach may be required, however, to get the right people aroundthe table when coordination i s an issue or when the relevant public goods and regulations have a sectoral nature. Inprinciple, interven- tions should be as horizontal as possible and as sectoral as necessary. Spillover and demonstration effects. Subsidized activities needto have a clear potential for providing spillover and demonstration effects. Public support must be contingent on an analy- sis o f the activity's ability to attract complementary investment or to generate information or technological spillovers. Moreover, supported activities should be structured insuch a way as to maximize the spillovers. Autonomous agencies. The agencies carrying out promotionmust be autonomous and, there- fore, musthave demonstrated their competence. Subject to certain constraints discussed be- low, the authorities responsible for carrying out promotionneedto have enough autonomy and independence that they can insulate themselves from lobbying, designtheir work agenda appropriately, and have the flexibility to respondto changing circumstances. This require- ment, inturn,meansthat the agenciesselected for the purpose musthave aprior track record o fprofessionalism, technical competence, and administrative effectiveness. When administra- tive andhumanresources are scarce, it maybe betterto lodge promotionactivities inagen- cies with demonstrated competence than to create new institutions from scratch, even ifdoing so restricts the range o f available policy tools continued 278 Coordinationof Economic Policy Formulation and lmplemenfafion in Tanzania Box 14.1continued Monitoring. The relevant agencies mustbe monitored closely by a principal who has a clear stake inthe outcomes and has political authority at the highest level. Autonomy does not mean lack of accountability. Close monitoring and coordination o fthe promotion activities by a cabinet-level politician-a principal who has internalizedthe agenda o f economic restruc- turingand shoulders the mainresponsibility for it-is essential. Suchmonitoringnot only guards against self-interested behavior on the part o f the agencies but also helpsprotect the agencies from capture by private interests. This principal might be the minister of the econ- omy, for example, or the president. Ifhe or she is not the president, the principal musthave the ear o fthe presidentandmustbe viewedas the latter's associaterather than rival. Communication with theprivate sector. The agencies carrying out promotion must maintain channels of communication with the private sector. Autonomy and insulation do not mean that bureaucrats should isolate themselves from entrepreneurs and investors. Ongoing contact and communication allow public officials to establish a good basis o f information on business realities, without which sound decision makingwould be impossible. This combination o fbu- reaucratic autonomy and connectednessi s what Evans (1995) terms embedded autonomy in hisdiscussion of successful economic strategies inEast Asia andLatinAmerica. Mistakes in the discovevyprocess.Public strategies of the sort advocated here are often de- ridedbecausethey may leadto picking the losers rather thanthe winners. However, an opti- mal strategy o f discovering the productive potential of a country will necessarily entail some mistakes o fthat type. Some promoted activities will fail. The objective should not be to minimize the chances that mistakes will occur, which would result inno self-discovery at all, butto minimize the costs ofthe mistakes whenthey occur. Ifgovernments make nomistakes, it only means that they are not trying hardenough. Flexibility in agency design. Promotionactivities needto have the capacity to renew them- selves so that the cycle o f discovery becomes an ongoing process. Just as there i s no single blueprintfor undertaking promotion, the needs and circumstances ofproductive discovery are likely to change over time. The agencies carrying out these policies musttherefore have the capacity to reinvent and refashion themselves to fit the changing circumstances. Source: Hausmann and Rodrik2005. the private sector to develop appropriate in- tion with the dialogue, and their capacity to stitutions to make its views heard. A variety effectively represent their constituents. o f such institutions exists, including the Tan- zania Private Sector Foundation; the Con- Challenge 6: Strengthen the Capacity federation of Tanzania Industries; the Tanza- of lnstitutions at the Regional and nia Bankers Association; the Tanzania District Levels Chamber of Mines; the Tourism Council of Tanzania; the Tanzania Oil Marketing Com- The Country Economic Memorandum high- panies; the Tanzania Chambers of Com- lights the regional diversity of Tanzania's merce, Industry, and Agriculture; the Tanza- economy with respect to potential sources o f nia Association of Consultants; and the growth, access to infiastructure, and natural Tanzania Chamber of Agriculture and Live- resource endowment. The lack of diversifi- stock. An important issue is to survey these cation o f regional economies also makes institutions about their roles in the govern- them vulnerable to external shocks such as ment-private sector dialogue, their satisfac- changes in commodity prices. The memo- 279 Managing Policies and Expendituresfor Shared Growth randum also highlights the importance of provides scope for redistribution o f re- agricultural and rural development for pov- sources across districts. However, a sys- erty reduction. A shared-growth strategy tem in which resource availability at the thus requires strong institutions not only at local level i s more closely linked to reve- the national levelbut also at the locallevel. nue generation at the local level might provide more incentives for a greater focus Tanzania's ongoing decentralization proc- on economic growth at the local level. ess touches on important elements o f the growth agenda. Inparticular, responsibility for service delivery inagriculture and infra- Implementation of the Growth structure (district roads, water) i s being Agenda of the NSGRP shifted to local authorities, while the role o f the central government in these areas is The NSGRP, Tanzania's PRSP, has limited to policy formulation and monitor- evolved as a broader instrument o f national ing.Fundingfor these activities isprimarily policy, with an explicit emphasis on the provided through earmarked transfers from broader agenda o f policies for shared eco- the central government. A formula-based nomic growth. Links between the NSGRP system for budgetary transfers to the dis- and other instruments have also evolved: tricts has been adopted that takes into ac- count demographic and social indicators. It 0 The government has continued to en- i s important to ensure that local authorities courage an open process o f consulta- play a supportive role for regional growth; tion on these instruments o f policy therefore, the use o f these resources must coordination. guided be by a regional growth strategy 0 The decentralization process that is un- developed by a partnershipo f local authori- der way gives greater responsibility to ties, the private sector, and other stake- local authorities, including for the man- holders. Strengthening o f accountability agement o f infrastructure andwater. arrangements at the local level needs to ac- company increased resource flows to local The task facing the government now is to authorities. Regional- or district-level translate the renewed emphasis on growth growth strategies combined with strong into specific strategies, detailed implemen- accountability arrangements also form the tation plans, and proposed resource alloca- basis for a successful switch from condi- tions. Several challenges arise inthis task: tional to unconditional transfers, which i s envisaged in the medium term. Uncondi- 0 The link between the NSGRP and the tional transfers will provide local authori- MTEF appears to have immediate po- ties with greater scope to implement a tential for translating individual sector growth strategy that is tailored to the spe- strategies into programs and expendi- cifics o fthe district or region. ture plans: a higher degree o f coordina- tion may be needed to make this work In this context, the division of revenue for the relatively difhse, cross-sectoral sources between the central government agendathat i s the key for growth. and local governments is also important. At present, the revenue sources for local 0 Growth requires increasing public in- authorities are limited to a closed list and vestment: it also requires enhanced ca- typically provide only 10 to 20 percent o f pacity for scrutiny o f public invest- a district's revenue. The current system mentproposals, both at the sector level 280 Coordination of Economic Policy Formulation and lmplemenfafion in Tanzania and inthe overall coordination of pub- structure o f the poverty monitoring sys- lic investment, ensuring maximum ef- tem, require reorientation to a broader ficiency, complementarity with the role so they can capture the renewed private sector, and so on. emphasis on growth. This effort may The open process o f consultation on also require evolution inthe roles o f the policy coordination has attracted sup- Vice President's Office, the Ministryo f port and interest, largely along sectoral Planning,Economy, andEmpowerment, lines, in support o f sector programs; it and other central agencies involved in is less clear that there i s a strong and organizing NSGRP coordination and vocal constituency pressingthe claims implementation. o f an effective cross-sectoral strategy Implementingthe NSGRP will require to promote growth. action plans based on its broad objec- Numerous channels exist for interfacing tives. Inmost cases, at the sector level, with the private sector on issues that are action plans can be derived from indi- highly relevant for growth but do not vidual sector strategies and programs, yet provide the forum for a coalitionfor but there appears to be a case for growth across central agencies, sectors, stronger coordination o f sector plans civil society, andthe private sector. within a cross-sectoral growth strategy that adequately reflects the objectives The Country Economic Memorandum o fthe NSGRP. work has underlined the specificity o f growth opportunities at regional anddis- The current budget guidelines do not trict levels. A key challenge is to support appear to reflect the strong emphasis regionally specific growth strategies that on growth-enhancing investment im- recognize these different opportunities plied by the NSGRP: more work i s while ensuring that institutions at re- needed to translate the NSGRP objec- gional and lower levels of government tives in this area into detailed propos- are consistent withthis objective. als, particularly for infrastructure in- vestment, preferably in the course o f The broadened scope o f the NSGRP offers developing the current year's MTEF. an opportunity to review the institutions for coordinating policy on economic growth. Reorientation o f public expenditure to- The following may be some immediate en- ward a greater focus on growth will try points for strengthening the institutional also have implications for the use o f ex- environment for steering the growthagenda: ternal financing. Current work on the proposed Joint Assistance Strategy i s an Institutions related to the earlier nar- important opportunity to ensure that rower scope o f the PRSP, such as the this strategic challenge i s addressed. Note 'This section draws on M j e m (2000) and Muganda (2004). 281 PART 5 Appendixes Statistical Tables APPENDIX A Population and Demographics Appendixes Table A.1. Summary of Population in Tanzania, 1988 and 2002 1988census 2002 census Mainland Zanzibar Tanzania Mainland Zanzibar Tanzania Number of population(in thousands) 22,486 641 23,127 33,585 985 34,569 Populationunder 15years old (%) 46 47 46 44 44 44 Sex ratio (males per 100female) 94 95 94 96 96 96 Male (in thousands) 10,907 311 11,218 16,428 483 16,910 Female (in thousands) 11,580 329 11,909 17,157 502 17,659 Infant mortality rate (per 1,000 Iive births) 115 120 115 115 120 115 Crude death rate (per 1,000 population) 15 15 15 15 15 15 Life expectancy at birth 49 48 49 49 48 49 Crude birth rate (per 1,000 population) 46 49 46 46 49 46 Total fertilitv rate (%) 6.2 7.1 6.5 6.2 7.1 6.5 Sources: National Bureauof Statistics, Populationand HousingCensuses for 1998and 2002; World Bank World Development Indicators2005. Nofe: The 1988and 2002 census results show a smaller population size than estimates of the United Nations. 286 Population and Demographics Table A.2. PopulationSize by Region, 1967-2002 (thousands) Region 1967 1978 1988 2002 Mainland 11,958.7 17,036.5 22,533.8 33,584.6 Arusha 610.5 926.2 1,351.7 1,293.0 Coast 428.0 516.6 638.0 889.2 Dares Salaam 356.3 843.1 1,360.9 2,497.9 Dodoma 709.4 972 1,237.8 1,699.0 Iringa 689.9 925.0 1,208.9 1,495.3 Kagera 658.7 1,009.8 1,326.2 2,033.9 Kigoma 473.4 648.9 854.8 1,679.1 Kilimanjaro 652.7 902.4 1,108.7 1,381.I Lindi 419.9 527.6 646.6 791.3 Mara 544.1 723.8 970.9 1,368.6 Mbeya 753.8 1,079.9 1,476.2 2,070.0 Morogoro 682.7 939.3 1,222.7 1,759.8 Mtwara 621.3 771.8 889.5 1,128.5 Mwanza 1,055.9 1,443.4 I,878.3 2,942.1 Rukwa 276.1 451.9 695 1,141.7 Ruvuma 395.4 561.6 783.3 1,117.2 Shinyanga 899.5 1,323.5 1,772.5 2,805.6 Singida 457.9 613.9 791.8 1,090.8 Tabora 502.1 817.9 1,036.3 1,717.9 Tanga 771.I 1,037.8 1,283.6 1,642.0 Manyaraa n.a. n.a. n.a. 1,040.5 Zanzibar 354.8 476.1 640.6 984.6 Total 12,313.5 17,512.6 23,174.3 34,569.2 Sources: National Bureauof Statistics, Populationand HousingCensusesfor 1967, 1978, 1998 and 2002; World BankWorld Development Indicators2005. Note: n.a. = not applicable.The 1988and 2002 census results show a smaller populationsize than estimates of the United Nations. a. Manyara is a newly created regionthat resulted from the division of the Arusha region into two parts: Arusha and Manyara. 287 Appendixes Table A.3. Dependency Ratios 1967 1978 1988 1991- 1994 1996 2002 2004 Agegroup census census census 92DHS KAPS DHS census DHS Under 15 years 43.9 46.1 45.8 46.8 49.3 47.2 44.2 46