52102 A FR IC A D EVELOPMENT FORU M Africa's Infrastructure A Time for Transformation Africa's Infrastructure Africa's Infrastructure A Time for Transformation Vivien Foster and Cecilia Briceño-Garmendia Editors A copublication of the Agence Française de Développement and the World Bank © 2010 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 12 11 10 09 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8041-3 eISBN: 978-0-8213-8083-3 DOI: 10.1596/978-0-8213-8041-3 Cover and interior design: Naylor Design Cover photo: Arne Hoel/World Bank; technician in a chlorination facility at a water treatment plant in Senegal. Library of Congress Cataloging-in-Publication Data Africa's infrastructure : a time for transformation. p. cm. ISBN 978-0-8213-8041-3 -- ISBN 978-0-8213-8083-3 (electronic) 1. Infrastructure (Economics)--Africa. I. World Bank. II. Africa Infrastructure Country Diagnostic. HC800.Z9C324 2009 363.6096--dc22 2009025406 Africa Development Forum Series The Africa Development Forum series was created in 2009 to focus on issues of significant relevance to Sub-Saharan Africa's social and economic development. Its aim is both to record the state of the art on a specific topic and to contribute to ongoing local, regional, and global policy debates. It is designed specifically to provide practitioners, scholars, and students with the most up-to-date research results while highlighting the promise, challenges, and opportunities that exist on the continent. The series is sponsored by the Agence Française de Développement and the World Bank. The manuscripts chosen for publication represent the highest quality in each institution's research and activity output and have been selected for their relevance to the development agenda. Working together with a shared sense of mission and interdisciplinary purpose, the two institutions are committed to a common search for new insights and new ways of analyzing the development realities of the Sub-Saharan Africa Region. Advisory Committee Members Agence Française de Développement Pierre Jacquet, Directeur de la Stratégie et Chef Économiste Robert Peccoud, Directeur de la Recherche World Bank Shantayanan Devarajan, Chief Economist, Africa Region Jorge Arbache, Senior Economist Contents Preface xix Acknowledgments xxi Abbreviations xxiii Overview Africa's Infrastructure: A Time for Transformation 1 Finding 1: Infrastructure Contributed over Half of Africa's Improved Growth Performance 2 Finding 2: Africa's Infrastructure Lags Well behind That of Other Developing Countries 2 Finding 3: Africa's Difficult Economic Geography Presents a Challenge for Infrastructure Development 3 Finding 4: Africa's Infrastructure Services Are Twice as Expensive as Elsewhere 4 Finding 5: Power Is Africa's Largest Infrastructure Challenge by Far 5 Finding 6: Africa's Infrastructure Spending Needs at $93 Billion a Year Are More than Double Previous Estimates by the Commission for Africa 6 Finding 7: The Infrastructure Challenge Varies Greatly by Country Type 7 Finding 8: A Large Share of Africa's Infrastructure Is Domestically Financed 8 Finding 9: After Potential Efficiency Gains, Africa's Infrastructure Funding Gap Is $31 Billion a Year, Mostly in the Power Sector 9 Finding 10: Africa's Institutional, Regulatory, and Administrative Reform Process Is Only Halfway Along 12 Key Recommendations 14 Note 26 References 26 PART 1 The Overall Story 29 Introduction The Africa Infrastructure Country Diagnostic 31 Genesis of the Project 32 Scope of the Project 33 vii viii Contents Note 41 References 41 1 Meeting Africa's Infrastructure Needs 43 Infrastructure: The Key to Africa's Faster Growth 44 Africa's Infrastructure Deficit 47 Africa's Infrastructure Price Premium 49 How Much Does Africa Need to Spend on Infrastructure? 52 Overall Price Tag 58 Notes 60 References 60 2 Closing Africa's Funding Gap 65 Spending Allocated to Address Infrastructure Needs 66 How Much More Can Be Done within the Existing Resource Envelope? 67 Annual Funding Gap 75 How Much Additional Finance Can Be Raised? 75 Costs of Capital from Different Sources 82 Most Promising Ways to Increase Funds 82 What Else Can Be Done? 83 Notes 84 References 85 3 Dealing with Poverty and Inequality 87 Access to Modern Infrastructure Services--Stagnant and Inequitable 88 Affordability of Modern Infrastructure Services--Subsidizing the Better Off 90 Alternatives to Modern Infrastructure Services--the Missing Middle 94 Policy Challenges for Accelerating Service Expansion 97 Notes 102 References 102 4 Building Sound Institutions 105 Institutional Reforms: A Glass Half Full 106 Does Private Sector Participation Work? 110 How Can State-Owned Enterprise Performance Be Improved? 117 Do Independent Regulators Make Sense? 120 Notes 122 References 122 Contents ix 5 Facilitating Urbanization 125 Viewing Cities as Engines of Growth 126 Strengthening Urban-Rural Links 126 The Costs of Providing Infrastructure--Sensitive to Density 130 Investment Needs 132 Infrastructure Financing 133 Policy Issues and Implications 137 Six Principles for Efficient Urbanization 139 Notes 140 References 141 6 Deepening Regional Integration 143 Why Regional Integration Matters 144 Opportunities for Regional Cooperation across Infrastructure Sectors 146 Meeting the Challenges of Regional Integration of Infrastructure in Africa 154 Notes 160 References 160 PART 2 Sectoral Snapshots 163 7 Information and Communication Technologies: A Boost for Growth 165 The African ICT Revolution 166 ICT Sector Developments 167 Institutional Reforms in the ICT Sector 168 Completing the Remaining Investment Agenda 172 Policy Challenges 174 Notes 179 References 179 8 Power: Catching Up 181 Africa's Chronic Power Problems 182 A Huge Investment Backlog 185 The Promise of Regional Power Trade 187 Improving Utility Performance through Institutional Reform 187 The Challenge of Cost Recovery 191 Policy Challenges 194 Notes 201 References 201 x Contents 9 Transport: More Than the Sum of Its Parts 203 Integrating Multimodal Transport 204 Developing Logistics Systems 205 Developing Transit Corridors for Landlocked Countries 206 Increasing Competition 207 Revisiting Attitudes toward Private Supply and Profit 209 Meeting Social Obligations 209 Notes 210 References 210 10 Roads: Broadening the Agenda 211 Road Infrastructure--Lagging Other Regions Somewhat 212 Road Infrastructure Institutions and Finance--Promising Developments 213 Road Expenditures--More Maintenance, Less Rehabilitation 215 Road Conditions--Reflecting Quality of Sector Governance 217 Infrastructure Spending Needs--an Average of 1.5 Percent of GDP a Year 217 Transport Services--the Forgotten Problem 219 Moving Forward--Broadening the Reform Agenda 221 Notes 226 References 226 11 Railways: Looking for Traffic 229 Africa's Rail Networks 230 The African Rail Market 233 How Much Investment Can Be Justified? 236 Institutional Arrangements and Performance 238 Key Issues for Governments 243 The Way Ahead 246 Notes 246 References 246 12 Ports and Shipping: Landlords Needed 249 The African Shipping Market 250 African Ports 252 Policy Issues and Implementation Challenges 255 Notes 258 References 258 13 Airports and Air Transport: The Sky's the Limit 259 The African Air Transport Market 260 Air Transport Policy in Africa 263 Contents xi African Air Transport Infrastructure 265 Policy Challenges 267 Notes 269 References 270 14 Water Resources: A Common Interest 271 Water Resources and Economic Development: Challenges for Africa 272 Addressing the Challenges 276 Investing in Africa's Water Security 279 Note 284 References 284 15 Irrigation: Tapping Potential 287 Agriculture and Poverty Reduction 288 Current State of Irrigation 289 Economic Investment Potential and Needs 290 Effect of Expanding Agricultural Water Development 293 Implementation Challenges 294 Notes 296 References 297 16 Water Supply: Hitting the Target? 299 The Millennium Development Goal for Water--Elusive for Many 300 Differing Patterns of Urban and Rural Access 300 Financing the MDG 302 Using Appropriate Technologies 305 The Challenge of Cost Recovery 307 Improving Utility Performance through Institutional Reform 309 Reforms in the Rural Space 313 Policy Recommendations 316 Notes 321 References 321 17 Sanitation: Moving Up the Ladder 323 The State of Sanitation in Africa 324 Challenges and Policy Options 329 Several Common Challenges Remain for All Countries 333 Notes 335 References 335 Index 337 xii Contents Boxes I.1 The AICD Geographic Information Systems Platform for Africa 35 1.1 Introducing a Country Typology 51 2.1 Does Deficit-Financed Public Investment in Infrastructure Pay for Itself? 77 3.1 Access to Basic Infrastructure and Time Use 95 3.2 Access, Affordability, and Alternatives--Urban Public Transportation 96 3.3 Are Connection Subsidies Well Targeted to the Poor? 98 4.1 Infrastructure's Institutional Scorecard 106 4.2 Privatization in African Ports 114 4.3 Lessons from the DAWASA Lease Contract (Tanzania) 116 4.4 Lessons from Successful SOE Reforms in Botswana Power Corporation 118 4.5 Performance Agreement for the National Water and Sewerage Corporation (Uganda) 119 4.6 Regulation by Contract in Senegal 121 5.1 Land Issues in Tanzania 136 6.1 Not So EASSy 148 6.2 One-Stop Border Posts to Facilitate Trade 159 8.1 South Africa's Power Supply Crisis 183 8.2 Kenya's Success with Private Participation in Power 189 8.3 Botswana's Success with a State-Owned Power Utility 190 8.4 CREST Spreading Good Practices 196 8.5 Rural Electrification in Mali 199 9.1 Impediments to Transit Traffic Growth in the Maputo Corridor 207 9.2 A New Attempt to Reform the Transit System in the Cameroon­Central African Republic­Chad Corridor 208 10.1 The Role of AGETIPs 215 12.1 Private Participation and Port Efficiency: The Case of Apapa Container Terminal, Lagos, Nigeria 257 14.1 The Metolong Dam and Water Supply Program in Lesotho 277 15.1 Assumptions for Irrigation Investment Potential Study 291 15.2 An Enabling Environment for Reform: Office du Niger 295 15.3 Nigeria's Fadama Water User Association: Expanding Irrigation 296 16.1 Legalizing Household Water Resellers in Côte d'Ivoire 302 16.2 Standposts in Kigali, Rwanda 306 16.3 Cost Recovery, Equity, and Efficiency in Water Tariffs in Africa 308 16.4 Senegal's Successful Experience with Private Sector Participation 312 16.5 Uganda's Successful Case of State-Owned Enterprise Reform 314 16.6 Issues Constraining Rural Water Supply in Cross River State, Nigeria 316 17.1 What Is Improved Sanitation? 324 17.2 Ethiopia's Success with a Community-Led Program 331 17.3 Burkina Faso's Sanitation Tax 334 Contents xiii Figures O.1 Access to Household Services 3 O.2 Underlying Causes of Africa's Power Supply Crisis 6 O.3 Burden of Infrastructure Spending Needs 8 O.4 Infrastructure Public Spending as a Percentage of GDP 9 O.5 Rehabilitation Backlog 10 O.6 Hidden Costs of Utility Inefficiencies 11 O.7 Infrastructure Funding Gap by Sector and Country Type 12 O.8 Institutional Progress across Sectors 13 O.9 Access to and Affordability of Household Services 23 I.1 Country Coverage of the Africa Infrastructure Country Diagnostic 34 I.2 Representativeness of Phase I Sampled Countries 34 1.1 Changes in Growth per Capita Caused by Changes in Growth Fundamentals 45 1.2 Changes in Growth per Capita Caused by Changes in Different Kinds of Infrastructure 45 1.3 Contribution of Infrastructure to Total Factor Productivity of Firms 46 1.4 Growth of Africa's Infrastructure Stocks Compared with Asia 49 1.5 Africa's Aggregate Infrastructure Spending Needs, by Country, 2006­15 59 2.1 Sources of Financing for Infrastructure Capital Investment, by Sector and Country Type 68 2.2 Split Investment Responsibilities between Governments and Public Enterprises, by Type of Country and Sector 69 2.3 Rehabilitation Backlog 73 2.4 Costs of Capital by Funding Source 83 2.5 Spreading Spending over Time 84 3.1 Coverage of Network Infrastructure Services, 1990­2005 88 3.2 Expansion of Access to Infrastructure Services Each Year, Mid-1990s to Mid-2000s 89 3.3 Projected Universal Access for Piped Water for Sub-Saharan African Countries, 2050 and Beyond 89 3.4 Coverage of Modern Infrastructure Services, by Budget Quintile 89 3.5 Share of Household Budgets Spent on Infrastructure Services, by Budget Quintile 91 3.6 Population with Service Connections Who Do Not Pay for Service 91 3.7 Affordability of Subsistence Consumption Priced at Cost-Recovery Levels 91 3.8 Extent to Which Electricity and Water Subsidies Reach the Poor, by Country 93 3.9 Amount of Subsidy Needed to Maintain Affordability of Water and Electricity Service, 2005 94 3.10 Access to Alternative Water and Sanitation Services across All Income Levels 96 3.11 Increased Industrial and Commercial Tariffs, Niger and Malawi 99 3.12 Residential Customers Using Prepayment Meters, by Utility 100 4.1 Institutional Progress across Countries, by Income Group, Aid Dependence, and Resource Richness 108 4.2 Institutional Progress across Sectors 109 xiv Contents 4.3 Institutional Progress on Reforms, Regulation, and Governance 110 4.4 Implementation of Private Participation across Sectors 111 4.5 Private Participation in Management and Investment across Sectors 111 4.6 Links between Private Sector Participation and Performance Indicators in Telecommunications 113 4.7 Links between Market Concentration and Performance Indicators in Telecommunications 113 4.8 Links between Port Concessions and Performance Indicators 114 4.9 Links between Rail Concessions and Performance Indicators 115 4.10 Links between Electricity Management Contracts and Performance Measures 115 4.11 Prevalence of Good Governance Practices among State-Owned Enterprises for Infrastructure 119 4.12 Prevalence of Performance Contracts in Electricity and Water 120 5.1 Access to Infrastructure by Location 129 5.2 Change in Urban and Rural Service Coverage, 1990­2005 130 5.3 Quality Differentials between Main and Rural Road Networks 130 5.4 Affordability of a Basic Package of Household Infrastructure 133 5.5 Spatial Split of Historic Infrastructure Investments 133 5.6 Spatial Split of Future Infrastructure Investment Needs 134 5.8 Municipal Budgets of Selected African Cities 134 5.7 Institutional Patterns of Water and Electricity Supply in Urban Areas 134 6.1 Africa's Regional Infrastructure Challenge 147 6.2 Savings from Power Trade for Major Potential Power-Importing Countries 150 6.3 Uneven Distribution of Airport Hubs across Africa: Traffic Flows between Top-60 Intraregional City Pairs 152 7.1 Growth of Mobile Subscribers in Africa, 1998­2006 166 7.2 Global System for Mobile Communications Coverage in Africa, 1998 to Third Quarter of 2006 166 7.3 Price of One-Minute, Peak-Rate Call to the United States, 2006 167 7.4 Price Basket for Internet Access, 2005 167 7.5 Net Change in Fixed-Line Market, 2001­05 168 7.6 Costs of Overstaffing for Fixed-Line Incumbents in Selected Countries 169 7.7 Value Added and Excise Taxes on Mobile Telephone Services, 2006 169 7.8 Status of Mobile Competition, 1993­2006 170 7.9 Tariff Rebalancing in African Countries with a Liberalized Telecommunication Sector, 1993­2006 170 7.10 Voice Coverage Gaps in 24 Sub-Saharan Countries 173 7.11 Broadband Coverage Gaps in 24 African Countries 173 8.1 Underlying Causes of Africa's Power Supply Crisis 183 8.2 Economic Cost of Outages in Selected Countries 184 8.3 Economic Burden Associated with Power Utility Inefficiencies in Selected Countries 188 Contents xv 8.4 Inefficiency in Utility Performance 188 8.5 Effect of Reform Measures on Hidden Costs 189 8.6 Underpricing of Power in Selected Countries 191 8.7 Electricity Costs and Revenues by Type of Power System, 2001­05 192 8.8 Past and Future Cost-Recovery Situation 192 8.9 Affordability of Subsistence Consumption of Power at Cost-Recovery Pricing 193 8.10 External Financing Commitments for the African Power Sector, 1994­2007 201 9.1 Africa Registers Low Scores on the Logistics Performance Index, 2007 205 10.1 Progress with Road Fund Reforms 214 10.2 Average Annual Spending on Road Transport, by Country, 2001­05 216 10.3 Rehabilitation and Maintenance Spending Relative to Norms 216 10.4 Distribution of Road Network Length across Condition Classes, by Country 218 11.1 Map of African Rail Networks 230 11.2 Rail Network Size and Traffic by Region 231 11.3 Average Railway Network Traffic Density, 2001­05 232 11.4 Average Railway Traffic Volumes, 2001­05 234 11.5 Private Participation in African Railways since 1990 239 11.6 Labor Productivity on African Rail Systems 241 11.7 Rail Concession Labor Productivity 242 12.1 Balance of Sub-Saharan African Container Trade, 2005 251 12.2 African Ports, by Size 252 12.3 Average Moves per Hour by Category of Port 254 13.1 Growth of Air Traffic, 1997­2006 260 13.2 Top-60 International Routes within Sub-Saharan Africa, 2007 261 13.3 Regional Growth Zones in Seats Offered, All Travel Categories 262 13.4 Airfares by Distance on African Routes, Including North Africa, 2008 263 13.5 International Civil Aviation Organization Analysis of Safety Implementation in Africa, 2004 268 14.1 Interannual Hydroclimatic Variability in Africa, by Selected Regions and Countries 273 14.2 Africa's International River Basins 274 14.3 Africa's Hydropower Potential 275 14.4 Sub-Saharan Africa's Irrigation Potential 276 14.5 Water Security and Growth 276 14.6 Water Interventions and Poverty 278 14.7 Water Reservoir Storage per Capita in Selected Countries, 2003 280 14.8 Degree of Regional Water Cooperation 283 15.1 Percentage of Cultivated Area Equipped for Irrigation, by Country 290 15.2 Investment Potential for Dam-Based and Small-Scale Irrigation 292 16.1 Increases in Access to Water by Source, 1990s to Early 2000s 303 16.2 Coverage of Water Services, by Budget Quintile 303 xvi Contents 16.3 Economic Burden of Water Underpricing, by Country 307 16.4 Affordability of Cost-Recovery Tariffs in Low-Income Countries 309 16.5 Economic Burden of Water Utility Operational Inefficiencies, by Country 310 16.6 Effect of Utility Inefficiency on Access Expansion and Water Quality 311 16.7 Hidden Costs and Institutional Frameworks 311 16.8 Overview of Reforms Affecting Urban Utilities 311 16.9 Overview of Rural Water Reforms 315 17.1 The Sanitation Ladder 325 17.2 Percentage of the Population Sharing Toilet Facilities 325 17.3 Coverage of Sanitation by Budget Quintile 326 17.4 Annual Growth in Coverage of Sanitation Types, 1990­2005 326 17.5 Moving Up to the Bottom Rung of the Sanitation Ladder: Côte d'Ivoire and Ethiopia, 1990­2005 327 17.6 Upgrading Latrines: Madagascar and Rwanda, 1990­2005 327 17.7 Mainstreaming Septic Tanks: Senegal, 1990­2005 327 17.8 Characterizing Patterns of Access to Sanitation across Urban and Rural Areas 328 Tables O.1 Africa's Infrastructure Deficit 3 O.2 Africa's High-Cost Infrastructure 5 O.3 Overall Infrastructure Spending Needs for Sub-Saharan Africa 7 O.4 Infrastructure Spending on Addressing Sub-Saharan Africa's Infrastructure Needs 9 O.5 Finding Resources: The Efficiency Gap and the Funding Gap 12 O.6 Overview of Private Participation in Infrastructure 14 I.1 AICD Background Papers 33 I.2 AICD Working Papers 41 1.1 Links between Infrastructure and Growth in Africa: What the Research Says 44 1.2 Evidence on Links between Infrastructure and MDGs in Africa 47 1.3 International Perspective on Africa's Infrastructure Deficit 48 1.4 Intraregional Perspective on Africa's Infrastructure Deficit 50 1.5 Africa's High-Cost Infrastructure 50 1.6 10-Year Economic and Social Targets for Investment Needs Estimates, 2006­15 52 1.7 Africa's Power Needs, 2006­15 53 1.8 Power Spending Needs, 2006­15 54 1.9 Irrigation Spending Needs, 2006­15 55 1.10 Water and Sanitation Spending Needs, 2006­15 55 1.11 Transport Spending Needs, 2006­15 57 1.12 ICT Spending Needs beyond the Purely Market Driven: Investment Only, 2006­15 58 1.13 Overall Infrastructure Spending Needs for Africa, 2006­15 58 2.1 Spending of Most Important Players Traced to Needs (Annualized Flows) 67 Contents xvii 2.2 Annual Budgetary Flows 69 2.3 Average Budget Variation Ratios for Capital Spending 70 2.4 Existing Disbursements above Those Directed to Infrastructure Needs, Annualized Flows 71 2.5 Economic Rates of Return for Key Infrastructure 71 2.6 Potential Gains from Increased Cost Recovery 72 2.7 Potential Gains from Greater Operational Efficiency 73 2.8 Finding Resources: The Efficiency Gap and the Funding Gap 74 2.9 Funding Gaps, by Sector and Country Group 75 2.10 Net Change in Central Government Budgets, by Economic Use, 1995­2004 76 2.11 Annualized ODA Investment Flows 78 2.12 Historic Annualized Investment Flows from China, India, and Arab Countries 79 2.13 Annual Private Participation Investment Flows 80 2.14 Outstanding Financing Stock for Infrastructure, as of 2006 81 3.1 Understanding Coverage of Infrastructure Services: The Role of Supply and Demand Factors 90 3.2 Monthly Household Budgets 91 3.3 Capital Cost of Best, Second-Best, and Traditional Alternatives 97 3.4 Potential Targeting Performance of Connection Subsidies under Various Scenarios 101 4.1 Correlation between Institutional Scores for Infrastructure and Measures of Broader Country Governance 109 4.2 Cancellation of African Private Participation Contracts 111 4.3 Overview of Experience with Private Participation in Infrastructure 117 4.4 Links between Governance and Performance Indicators for Electricity and Water 120 4.5 Links between Regulation and Performance Indicators for Telecommunications, Electricity, and Water 121 4.6 Links between Type of Regulator and Performance Indicators for Water 121 5.1 Link between Agricultural Productivity and Distance to Urban Centers 127 5.2 Distribution of Population by Type of Settlement and Country Type 127 5.3 Economic Differentials between Rural and Urban Populations, by Country Type 128 5.4 Sectoral Contributions to GDP and GDP Growth 128 5.5 Households with Access to One or More Modern Infrastructure Services 131 5.6 Capital Cost per Capita of Infrastructure Provision, by Density 131 5.7 Overview of Local Infrastructure Financing Mechanisms 135 5.8 Population Density across Country Types 135 6.1 Costs of Reaching Full Intercontinental and Intraregional Connectivity 148 6.2 Benefits of Access to an Underwater Cable 149 6.3 Profile of Top-Six Potential Power-Exporting Countries 150 6.4 Average Delivery Time for Containers from Vessel to Consignee 151 6.5 Key Transport Corridors for International Trade, Sub-Saharan Africa 151 7.1 Prices for Access to International Voice and Internet Connectivity 171 xviii Contents 7.2 Investments Needed to Close Gaps in Voice and Broadband Coverage in Sub-Saharan Africa 172 8.1 Economic Cost of Emergency Power Generation 184 8.2 Power Sector Spending Needs 185 8.3 Financing Flows to the Power Sector 186 8.4 Composition of Power Sector Funding Gap 186 8.5 Cost and Affordability of Monthly Power Bills at Cost-Recovery Prices: Past and Future 193 10.1 Average Daily Traffic on the Main Road Network 213 10.2 Road Sector Spending Needs 219 10.3 Financing Flows to the Road Sector 219 10.4 Overview of Key Road Freight Parameters on Main International Corridors 220 12.1 Traffic Trends for Container Trade, Sub-Saharan Ports, by Region, 1995­2005 250 12.2 Traffic Trends for General Cargo, 1995­2005 251 12.3 Private Transactions for All Port Sectors, 2000­07 254 12.4 Average Port Delays 255 12.5 Typical Gateway Container and General Cargo Handling Charges in World Markets 255 13.1 Air Traffic Growth in Sub-Saharan Africa, 2001­07 261 13.2 Air Service Liberalization in African Regional Groupings 264 13.3 Runway Quality in Sub-Saharan Africa, 2007 266 14.1 Water Availability and Withdrawal 275 14.2 Capital Investment Needs in Large Multipurpose Hydropower Storage by 2015 281 15.1 Selected Irrigation Investment Indicators for Sub-Saharan Africa 289 15.2 Potential Investment Needs for Large-Scale, Dam-Based, and Complementary Small-Scale Irrigation in Sub-Saharan Africa 291 15.3 Share of Crops under Irrigation, Irrigation Investment Needs Assessment 293 15.4 Sensitivity of Irrigation Potential to Assumed Investment Cost 293 15.5 Food Price Changes for Various Indicators, 2020 and 2050 294 16.1 Evolution of Water Supply Coverage in Africa, by Source 301 16.2 Services Provided by Utilities in Their Service Areas 301 16.3 Quality of Services Provided by Utilities in Their Service Areas 302 16.4 Average Price for Water Service in 15 Largest Cities, by Type of Provider 303 16.5 Estimated Annual Financing Needed to Meet the Water MDG 303 16.6 Existing Financial Flows to Water Supply and Sanitation 304 16.7 Composition of Water Sector Funding Gap 305 16.8 Overview of Private Sector Participation's Effect on Utility Performance 312 16.9 Relationship between Rural Reform Index and Success in Expanding Rural Service Coverage 315 17.1 Patterns of Access to Sanitation in Africa 326 17.2 Cost of Sanitation Facilities in Senegal 332 Preface This study is part of the Africa Infrastruc- the Democratic Republic of Congo, Ethiopia, ture Country Diagnostic (AICD), a project Ghana, Kenya, Lesotho, Madagascar, Malawi, designed to expand the world's knowledge of Mozambique, Namibia, Niger, Nigeria, Rwanda, physical infrastructure in Africa. The AICD Senegal, South Africa, Sudan, Tanzania, Uganda, will provide a baseline against which future and Zambia. Under a second phase of the proj- improvements in infrastructure services can ect, coverage is expanding to include as many be measured, making it possible to monitor of the additional African countries as possible. the results achieved from donor support. It The AICD was commissioned by the should also provide a more solid empirical Infrastructure Consortium for Africa (ICA) foundation for prioritizing investments and following the 2005 G8 (Group of Eight) sum- designing policy reforms in the infrastructure mit at Gleneagles, Scotland, which flagged the sectors in Africa. importance of scaling up donor finance for The AICD is based on an unprecedented infrastructure in support of Africa's develop- effort to collect detailed economic and techni- ment. The World Bank is implementing the cal data on the infrastructure sectors in Africa. AICD under the guidance of a steering com- The project has produced a series of origi- mittee that represents the African Union, the nal reports on public expenditure, spending New Partnership for Africa's Development needs, and sector performance in each of the (NEPAD), Africa's regional economic com- main infrastructure sectors, including energy, munities, the African Development Bank information and communication technologies, (AfDB), the Development Bank of South irrigation, transport, and water and sanitation. Africa (DBSA), and major infrastructure This volume synthesizes the most significant donors. Financing for the AICD is provided findings of those reports. by a multidonor trust fund to which the main The first phase of the AICD focused on 24 contributors are the United Kingdom's Depart- countries that together account for 85 percent ment for International Development (DFID), of the gross domestic product, population, the Public-Private Infrastructure Advisory and infrastructure aid flows of Sub-Saharan Facility (PPIAF), Agence Française de Dével- Africa. The countries are Benin, Burkina Faso, oppement (AFD), the European Commission, Cameroon, Cape Verde, Chad, Côte d'Ivoire, and Germany's Entwicklungsbank (KfW). xix xx Preface A group of distinguished peer reviewers from This and other volumes analyzing key infra- policy-making and academic circles in Africa and structure topics, as well as the underlying data beyond reviewed all major outputs of the study sources described above, will be available for to ensure the technical quality of the work. download from http://www.infrastructure The Sub-Saharan Africa Transport Pol- africa.org. Stand-alone summaries are avail- icy Program (SSATP) and the Water and able in English and French. Sanitation Program (WSP) provided technical Inquiries concerning the availability of data support on data collection and analysis per- sets should be directed to the volume editors taining to their respective sectors. at the World Bank in Washington, DC. Acknowledgments This report was undertaken by the director's (professor, John F. Kennedy School of Govern- office of the Department for Sustainable Devel- ment, Harvard University), Cheikh Kane (inde- opment in the Africa Region of the World Bank. pendent expert on infrastructure finance), and A number of directors oversaw the implementa- Xinzhu Zhang (professor, Chinese Academy of tion of the project throughout its life, including Social Sciences, Beijing). (in chronological order) Michel Wormser, John In order to ensure broad-based par- Henry Stein (acting), and Inger Andersen. ticipation and consultation of World Bank The task team leaders for the report were technical practices, a number of internal Vivien Foster and Cecilia Briceño-Garmendia, peer review groups were formed to provide and the core team for the project comprised guidance and feedback on earlier drafts of Aijaz Ahmad, Dominique Akele, Sudeshna the document. The individual groups and Ghosh Banerjee, Carolina Dominguez Torres, their members are as follows: ICT sector-- Sophie Hans-Moevi, Elvira Morella, Nataliya Mavis Ampah, Philippe Dongier, Clemencia Pushak, Rupa Ranganathan, Maria Shkaratan, Torres, and Mark Williams; irrigation sector-- and Karlis Smits. Barbara Miller, Stephen Mink, and Ashok The project team is grateful to a number of Subramanian; power sector--Philippe Benoit, World Bank colleagues who acted as advisers on David Donaldson, Vijay Iyer, Luiz Maurer, key cross-cutting aspects of the report. These Rob Mills, Lucio Monari, Kyran O'Sullivan, include Antonio Estache, Jose Luis Irigoyen, Prasad Tallapragada, Clemencia Torres, and and Jyoti Shukla, who provided advice on gen- Tjaarda Storm Van Leeuwen; transport sector-- eral infrastructure issues; Sarah Keener, who Pierre Pozzo di Borgo, Michel Luc Donner, provided advice on social issues; Paul Martin, Michel Iches, Marc Juhel, Cornelis Kruk, who provided advice on environmental issues; Alain Labeau, Charles Schlumberger, and and Stephen Mink, who provided advice on Kavita Sethi; water supply and sanitation rural and agricultural issues. sector--Ventura Bengoechea, Jaime Biderman, A technical advisory panel provided indepen- Matar Fall, Sarah Keener, Peter Kolsky, Alex dent, external peer review on the quality of the McPhail, Eustache Ouayoro, Christophe Pre- background papers on which this report is based. vost, Caroline van den Berg, and Meike van The panel was cochaired by Shanta Devarajan Ginneken; finance theme--Gerardo Corro- (chief economist, Africa Region, World Bank) chano, Michael Fuchs, James Leigland, Anand and Louis Kasekende (chief economist, African Rajaram, Sudhir Shetty, Jyoti Shukla, Clem- Development Bank), and comprised Adeola encia Torres, Marilou Uy, and Marinus Ver- Adenikinju (professor, University of Ibadan, hoeven; poverty and inequality theme--Judy Nigeria), Emmanuelle Auriol (professor, Univer- Baker, Douglas Barnes, Ellen Hamilton, Julian sity of Toulouse, France), Tony Gomez-Ibanez Lampietti, and Kenneth Simler; institutional xxi xxii Acknowledgments theme--James Leigland and Jyoti Shukla; An editorial team comprising Bruce Ross- urban theme--Jaime Biderman, Catherine Larson, Steven Kennedy, and Joseph Caponio Farvacque-Vitkovic, Matthew Glasser, Sumila contributed significantly to improving the Gulyani, and Uri Raich; and regional integra- quality of the final manuscript submitted to tion theme--Uwe Deichmann, Jakob Kolster, the World Bank Office of the Publisher for and Mark Tomlinson. publication. Abbreviations $ All dollar amounts are in IPP independent power producer U.S. dollars unless otherwise JMP Joint Monitoring Programme indicated. KenGen Kenya Electricity Generating ADF African Development Fund Company AFRICATIP Association Africaine des KPLC Kenya Power and Lighting Agences d'Exécution des Company Travaux d'Intérêt Public MDG Millennium Development AGETIP agence d'exécution des travaux Goal d'intérêt public NEPAD New Partnership for Africa's AICD Africa Infrastructure Country Development Diagnostic NWSC National Water and Sewerage AMADER Agence Malienne pour le Corporation Développement de l'Energie O&M operation and maintenance Domestique et l'Electrification ODA official development assistance Rurale (Malian Agency OECD Organisation for for the Development of Economic Co-operation and Domestic Energy and Rural Development Electrification) PPI private participation in BPC Botswana Power Corporation infrastructure CEAR Central East African Railways PSP private sector participation CREST Commercial Reorientation of SAT-3 South Atlantic 3/West Africa the Electricity Sector Toolkit Submarine Cable DHS demographic and health SEACOM South Africa­East Africa­ survey South Asia­Fiber Optic Cable EASSy Eastern African Submarine SODECI Société de Distribution d'Eau Cable System de la Côte d'Ivoire GIS geographic information SOE state-owned enterprise systems SSATP Sub-Saharan Africa Transport GNI gross national income Policy Program GSM global systems mobile TEAMS The East Africa Marine System IBNET International Benchmarking TEU 20-foot equivalent unit Network TIR Transports Internationaux IBT increasing block tariff Routiers ICT information and VoIP Voice over Internet Protocol communication technology WiMAX Worldwide Interoperability for IDA International Development Microwave Access Association WSS water supply and sanitation Overview Africa's Infrastructure: A Time for Transformation T he Africa Infrastructure Country Diag- diseconomies of scale in production and nostic is an unprecedented attempt to high profit margins caused by lack of collect comprehensive data on the infra- competition. structure sectors in Africa--covering power, · Power is by far Africa's largest infrastructure transport, irrigation, water and sanitation, and challenge, with 30 countries facing regular information and communication technology power shortages and many paying high pre- (ICT)--and to provide an integrated analysis miums for emergency power. of the challenges they face. Based on extensive fieldwork across Africa, the following main · The cost of addressing Africa's infrastruc- ture needs is around $93 billion a year, about findings have emerged: one-third of which is for maintenance-- · Infrastructure has been responsible for more than twice the Commission for Afri- more than half of Africa's recent improved ca's (2005) estimate. growth performance and has the potential · The infrastructure challenge varies greatly to contribute even more in the future. by country type--fragile states face an · Africa's infrastructure networks increas- impossible burden and resource-rich coun- ingly lag behind those of other developing tries lag despite their wealth. countries and are characterized by miss- · A large share of Africa's infrastructure is ing regional links and stagnant household domestically financed, with the central gov- access. ernment budget being the main driver of · Africa's difficult economic geography pre- infrastructure investment. sents a particular challenge for the region's · Even if major potential efficiency gains are infrastructure development. captured, Africa would still face an infra- · Africa's infrastructure services are twice structure funding gap of $31 billion a year, as expensive as elsewhere, reflecting both mainly in power. 1 2 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION · Africa's institutional, regulatory, and admin- For most countries, the negative effect of defi- istrative reforms are only halfway along, cient infrastructure is at least as large as that but they are already proving their effect on of crime, red tape, corruption, and financial operational efficiency. market constraints. For one set of countries, power emerges as the most limiting factor by far, cited by more than half the firms in more Finding 1: Infrastructure than half the countries as a major business Contributed over Half of Africa's obstacle. For a second set, inefficient function- Improved Growth Performance ing of ports and associated customs clearance is equally significant. Deficiencies in transport Africa's growth improved markedly in the last and in ICTs are less prevalent but substantial decade. African countries saw their econo- in some cases. mies grow at a solid 4 percent a year from Infrastructure not only contributes to eco- 2001 to 2005. Resource-rich countries, which nomic growth, but it is also an important input have benefited from rising commodity prices, to human development (Fay and others 2005). demonstrate the highest growth rates. Growth Infrastructure is a key ingredient for achieving overall still falls short of the 7 percent needed all the MDGs. Safe and convenient water sup- to achieve substantial poverty reduction and plies save time and arrest the spread of a range attain the Millennium Development Goals of serious diseases--including diarrhea, a lead- (MDGs), however. Infrastructure, significant ing cause of infant mortality and malnutrition. in Africa's economic turnaround, will need to Electricity powers health and education services play an even greater role for the continent to and boosts the productivity of small businesses. reach its development targets. Road networks provide links to global and local Across Africa, infrastructure contributed markets. ICTs democratize access to informa- 99 basis points to per capita economic growth tion and reduce transport costs by allowing from 1990 to 2005, compared with 68 basis people to conduct transactions remotely. points for other structural policies (Calderón 2008). That contribution is almost entirely attributable to advances in the penetration Finding 2: Africa's Infrastructure of telecommunication services. The deterio- Lags Well behind That of Other ration in the quantity and quality of power Developing Countries infrastructure over the same period retarded growth, shaving 11 basis points from per cap- On just about every measure of infrastructure ita growth for Africa as a whole and as much as coverage, African countries lag behind their 20 basis points for southern Africa. peers in the developing world (Yepes, Pierce, The growth effects of further improving and Foster 2008). This lag is perceptible for low- Africa's infrastructure would be even greater. and middle-income countries in Sub-Saharan Simulations suggest that if all African coun- Africa relative to other low- and middle-income tries were to catch up with Mauritius (the countries (table O.1). The differences are par- regional leader in infrastructure) per capita ticularly large for paved roads, telephone main growth in the region could increase by 2.2 per- lines, and power generation. For all three, Africa centage points. Catching up with the Republic has been expanding stocks much more slowly of Korea would increase per capita growth by than other developing regions; so unless some- 2.6 percentage points a year. In Côte d'Ivoire, thing changes, the gap will continue to widen. the Democratic Republic of Congo, and Sen- To what extent does Africa's current deficit egal, the effect would be even larger. date to a low starting point for infrastructure In most African countries, particularly stocks? Africa started out with stocks that the lower-income countries, infrastructure were generally not very different from those emerges as a major constraint on doing busi- in South or East Asia in the 1960s for roads, ness, depressing firm productivity by about in the 1970s for telephones, and in the 1980s 40 percent (Escribano, Guasch, and Pena 2008). for power. The comparison with South Asia, Africa's Infrastructure: A Time for Transformation 3 which has similar per capita incomes, is par- universal access to these and other household ticularly striking. In 1970, Sub-Saharan Africa services is more than 50 years away in most had almost three times the generating capac- African countries (Banerjee, Wodon, and oth- ity per million people as South Asia. In 2000, ers 2008). Even where infrastructure networks South Asia had left Sub-Saharan Africa far are in place, a significant percentage of house- behind--with almost twice the generation holds remains unconnected, suggesting that capacity per million people. Also in 1970, demand-side barriers exist and that univer- Sub-Saharan Africa had twice the main-line sal access entails more than physical rollouts telephone density of South Asia, but by 2000, of networks. As might be expected, access to the two regions were even. infrastructure in rural areas is only a frac- Since 1990, coverage of household services tion of that in urban areas, even where urban has barely improved (figure O.1, panel a). coverage is already low by international stan- Africa is unlikely to meet the MDGs for water dards (Banerjee, Wodon, and others 2008) and sanitation. Moreover, on current trends, (figure O.1, panel b). Table O.1 Africa's Infrastructure Deficit Sub-Saharan Finding 3: Africa's Difficult Africa Other low-income low-income Economic Geography Presents a Normalized units countries countries Challenge for Infrastructure Paved-road density 31 134 Development Total road density 137 211 Main-line density 10 78 Relative to other continents, Africa is char- Mobile density 55 76 acterized by low overall population density Internet density 2 3 (36 people per square kilometer), low rates of urbanization (35 percent), but relatively rapid Generation capacity 37 326 rates of urban growth (3.6 percent a year), a Electricity coverage 16 41 relatively large number of landlocked coun- Improved water 60 72 tries (15), and numerous small economies. Improved sanitation 34 51 A further complication is that the continent Source: Yepes, Pierce, and Foster 2008. experiences particularly high hydrological Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density in lines per thousand variability, with huge swings in precipitation population; generation capacity in megawatts per million popula- across areas, seasons, and time, which climate tion; electricity, water, and sanitation coverage in percentage of population. change is likely to exacerbate. Figure O.1 Access to Household Services a. Stagnant trends b. Rural-urban divide 40 80 30 60 % of population % of population 20 40 10 20 0 0 1990­95 1996­2000 2001­05 piped water electricity flush toilets landline telephones piped water electricity national rural urban flush toilets landline telephones Source: Banerjee, Wodon, and others 2008. 4 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Africa's atomized nation-states are reflected Africa's water resources are abundant, in the region's fragmentary infrastructure but because of an absence of water stor- networks. Sub-Saharan Africa comprises 48 age and distribution infrastructure, they are nation-states, many of which are very small. grossly underused. Therefore, water security-- The bulk of those countries have populations eliable water supplies and acceptable risks of fewer than 20 million and economies smaller from floods and other unpredictable events, than $10 billion. International frontiers bear including those from climate change--will little relation either to natural features (such require a significant expansion of water as river basins) or to artificial features (such as storage capacity from the current 200 cubic cities and their accessibility to trading chan- meters per capita (Grey and Sadoff 2006). In nels, such as ports). Intraregional connectiv- other parts of the world, such capacity is in ity is therefore very low, whether measured in the thousands of cubic meters. The cost of transcontinental highway links, power inter- expanding water storage is extremely high connectors, or fiber-optic backbones. Most in relation to the size of Africa's economies, continuous transport corridors are concerned suggesting the phasing of investments, with with providing access to seaports, whereas the initial focus on achieving water security for intraregional road network is characterized by key growth poles. major discontinuities. Few cross-border inter- Water also needs to be distributed for agri- connectors exist to support regional power cultural use. In a handful of countries, only exchange, even though many countries are too 7 million hectares are equipped for irrigation. small to produce power economically on their Although the irrigation-equipped area is less own. Until recently, the whole of East Africa than 5 percent of Africa's cultivated area, it lacked access to a global submarine cable to pro- produces 20 percent of the value of agricultural vide low-cost international communications production. An additional 12 million hectares and Internet access. The intraregional fiber- could be economically viable for irrigation as optic network is also incomplete, but growing long as costs are contained (You 2008). rapidly. Because of their geographic isolation, landlocked countries in particular suffer from the lack of regional connectivity. Finding 4: Africa's Infrastructure Both the spatial distribution and rapid Services Are Twice as Expensive migration of Africa's population create major as Elsewhere challenges for reaching universal access. In rural areas, over 20 percent of the population lives Not only are Africa's infrastructure networks in dispersed settlements where typical popula- deficient in coverage, but the price of the tion densities are less than 15 people per square services provided is also exceptionally high kilometer; hence, the costs of providing infra- by global standards (table O.2). Whether for structure are comparatively high. In urban areas, power, water, road freight, mobile telephones, population growth rates averaging 3.6 percent a or Internet services, the tariffs paid in Africa year are leaving infrastructure service provid- are several multiples of those paid in other ers severely stretched. As a result, urban service parts of the developing world. The explana- coverage has actually declined over the last tion for Africa's higher prices sometimes lies decade, and lower-cost alternatives are filling the in genuinely higher costs, and sometimes in resulting gap (Banerjee, Wodon, and others 2008; high profits. The policy prescriptions for the Morella, Foster, and Banerjee 2008). In addition, two cases are, of course, radically different. population densities in African cities are rela- Power provides the clearest example of tively low by global standards and do not benefit infrastructure with costs genuinely higher in from large economies of agglomeration in the Africa than elsewhere. Many smaller coun- provision of infrastructure services. As a result, tries have national power systems below the the costs of providing a basic infrastructure 500-megawatt threshold and therefore often package can easily be twice as much as in other rely on small diesel generation that can cost developing cities (Dorosh and others 2008). up to $0.35 per kilowatt-hour to run, about Africa's Infrastructure: A Time for Transformation 5 Table O.2 Africa's High-Cost Infrastructure higher than those without (Minges and oth- Other ers 2008). Sub-Saharan developing Infrastructure sector Africa regions Power tariffs ($ per kilowatt-hour) 0.02­0.46 0.05­0.10 Finding 5: Power Is Africa's Water tariffs ($ per cubic meter) 0.86­6.56 0.03­0.60 Largest Infrastructure Challenge Road freight tariffs by Far ($ per ton-kilometer) 0.04­0.14 0.01­0.04 Mobile telephony Whether measured in generation capacity, ($ per basket per month) 2.60­21.00 9.90 electricity consumption, or security of sup- International telephony ply, Africa's power infrastructure delivers only ($ per 3-minute call to a fraction of the service found elsewhere in the United States) 0.44­12.50 2.00 the developing world (Eberhard and others Internet dial-up service ($ per month) 6.70­148.00 11.00 2008). The 48 Sub-Saharan Africa countries Sources: Authors' estimates based on Africon 2008; Bannerjee, (with 800 million people) generate roughly the Skilling, and others 2008; Eberhard and others 2008; Minges and same power as Spain (with 45 million people). others 2008; Teravaninthorn and Raballand 2008; Wodon 2008a and 2008b. Power consumption, at 124 kilowatt-hours Note: Ranges reflect prices in different countries and various per capita annually and falling, is only 10 per- consumption levels. Prices for telephony and Internet service represent all developing regions, including Africa. cent of that found elsewhere in the developing world, barely enough to power one 100-watt lightbulb per person for 3 hours a day. twice the costs faced by larger countries typi- More than 30 African countries experience cally with coal- or hydropower-based systems power shortages and regular interruptions to (Eberhard and others 2008). service (figure O.2). The underlying causes High road freight tariffs in Africa have vary: failures to bring on new capacity to keep much more to do with high profit margins pace with the demands of economic growth, than high costs (Teravaninthorn and Rabal- droughts that reduced hydropower in East land 2008). The costs for Africa's trucking Africa, oil price hikes that inhibited afford- operators are not much higher than costs in ability of diesel imports for many West African other parts of the world, even when informal countries, and conflicts that destroyed power payments are counted. Profit margins, by con- infrastructure in fragile states. Africa's firms trast, are exceptionally high, particularly in report losing 5 percent of their sales because of Central and West Africa, where they reach 60 frequent power outages--a figure that rises to to 160 percent. The underlying cause is limited 20 percent for informal firms unable to afford competition combined with a highly regulated backup generation. Overall, the economic market based on tour de role principles, which costs of power outages can easily rise to 1­2 allocate freight to transporters through a cen- percent of GDP. tralized queuing method rather than allowing A common response to the crisis is to ten- truckers to enter into bilateral contracts with der short-term leases for emergency power. At customers directly. least 750 megawatts of emergency generation The high costs of international telephony are operating in Sub-Saharan Africa, which and Internet services reflect a mixture of cost for some countries constitute a large pro- and profit factors. Countries without access portion of their national installed capacity. to a submarine cable must rely on expensive However, emergency generation is expensive satellite technology for international connec- at costs of $0.20­$0.30 per kilowatt-hour, tivity and have charges typically twice those and for some countries, the price tag can be in countries that do enjoy such access. Even as high as 4 percent of GDP. Paying for emer- when access to a submarine cable is secured, gency leases absorbs significant budgetary countries with a monopoly on this interna- resources, reducing the funds for longer-term tional gateway still have tariffs substantially solutions. 6 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure O.2 Underlying Causes of Africa's Power Supply Crisis MAURITANIA CAPE VERDE MALI NIGER SE CHAD SUDAN ERITREA NE GA GAMBIA L BURKINA DJIBOUTI FASO BENIN GUINEA-BISSAU GUINEA NIGERIA GHANA D'IVOIRE CÔTE SIERRA LEONE ETHIOPIA CENTRAL ON AFRICAN REPUBLIC RO LIBERIA IA ME AL TOGO DEMOCRATIC CA M SO GO EQUATORIAL GUINEA REPUBLIC OF UGANDA CON CONGO GABON KENYA SAO TOME AND PRINCIPE RWANDA . OF ES LL REP BURUNDI HE C TANZANIA SEY MALAWI COMOROS Main cause or trigger MAYOTTE ANGOLA natural causes (droughts) ZAMBIA E IQU MB oil price shock AR ZA ZIMBABWE ASC MO system disrupted by conflict NAMIBIA MAURITIUS DAG high growth, low investment/structural issues BOTSWANA MA SWAZILAND SOUTH AFRICA LESOTHO Source: Eberhard and others 2008. Finding 6: Africa's Infrastructure · Interconnect capitals, ports, border cross- Spending Needs at $93 Billion ings, and secondary cities with a good- a Year Are More than Double quality road network. Previous Estimates by the · Provide all-season road access to Africa's Commission for Africa high-value agricultural land. · More than double Africa's irrigated area. Meeting Africa's infrastructure needs calls for · Meet the MDGs for water and sanitation. a very substantial program of infrastructure investment and maintenance: · Raise household electrification rates by 10 percentage points. · Develop an additional 7,000 megawatts · Provide global systems mobile voice signal a year of new power generation capacity and public access broadband to 100 percent (about half through multipurpose water of the population. storage schemes). · Enable regional power trade by laying Implementing such an ambitious program to 22,000 megawatts of cross-border transmis- address Africa's infrastructure needs would cost sion lines. around $93 billion a year (about 15 percent of · Complete the intraregional fiber-optic the region's GDP). Some two-thirds of this total backbone network and continental subma- relates to capital expenditure, and the remain- rine cable loop. ing one-third to operation and maintenance Africa's Infrastructure: A Time for Transformation 7 requirements (table O.3; Briceño-Garmendia, inadequate competition for tenders have all Smits, and Foster 2008). played their role, with the last factor by far That cost is well over twice the $39 billion of the strongest. infrastructure spending estimated by the Com- The global financial crisis of 2008 can be mission for Africa report in 2005. That figure expected to reduce demand for some types of was based on a cross-country econometric infrastructure, but it would not hugely alter study, rather than the more detailed country- the estimated spending needs. Planning and level microeconomic modeling (Estache 2005). social targets rather than economic growth A more recent update of the cross-country drive a large share of the spending needs, for model used for the Commission for Africa example, the transport spending needs (which report came up with revised estimates in the are largely based on connectivity objectives) range of $80 billion to $90 billion, much closer and the water and sanitation spending needs to those reported here (Yepes 2007). (which are based on the MDGs). The spending About 40 percent of the total spending needs with the strongest direct link to economic needs are associated with power, reflecting growth are those for the power sector. However, Africa's particularly large deficits. About one- because of the large investment backlog in the third of the power investment needs (some sector, the estimated spending needs contain a $9 billion a year) are associated with multipur- strong component of refurbishment and catch- pose water storage for hydropower and water up. Thus, even halving economic growth esti- resource management. After power, water sup- mates for the region would reduce estimated ply and sanitation and then transport are the power spending needs by only 20 percent. most significant items. The global recession could also be expected to Given recent escalations in unit costs, these affect demand for ICT services, as well as trade- estimates are a lower bound. Although the related infrastructure, such as railways and investment estimates here are based on the most ports. However, the weight of these infrastruc- accurate unit-cost data available, develop- tures in the total spending needs is not much ment agencies are reporting significant cost more than 10 percent. escalations on projects under implementa- tion. For road projects, these escalations have averaged 35 percent but in some cases have been as high as 50­100 percent. Closer inspec- Finding 7: The Infrastructure tion reveals that no single factor explains this Challenge Varies Greatly by escalation. Domestic inflation, tight construc- Country Type tion industry conditions, oil price hikes, and The infrastructure challenge differs mark- edly across African country groups (Briceño- Table O.3 Overall Infrastructure Spending Needs for Garmendia, Smits, and Foster 2008). Because Sub-Saharan Africa of the widely varying circumstances, distin- $ billions annually guishing among middle-income countries Operation Infrastructure Capital and Total (like Cape Verde and South Africa), resource- sector expenditure maintenance spending rich countries with economies heavily reliant ICT 7.0 2.0 9.0 on petroleum or mineral revenues (like Nige- Irrigation 2.9 0.6 3.4 ria and Zambia), fragile states emerging from conflict (like Côte d'Ivoire and the Demo- Power 26.7 14.1 40.8 cratic Republic of Congo), and the remaining Transport 8.8 9.4 18.2 low-income countries that are neither fragile WSS 14.9 7.0 21.9 nor resource rich (like Senegal and Uganda) Total 60.4 33.0 93.3 is helpful. Source: Authors' estimates based on Banerjee, Wodon, and By far the most daunting infrastructure others 2008; Carruthers, Krishnamani, and Murray 2008; Mayer and others 2008; Rosnes and Vennemo 2008. challenges are those facing the fragile states Note: Column totals may not add exactly because of rounding (figure O.3). The recent conflicts affecting these errors. ICT = information and communication technology; WSS = water supply and sanitation. countries usually resulted in the destruction 8 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure O.3 Burden of Infrastructure Spending Needs have not tended to do so. Resource-rich coun- 40 tries could meet their infrastructure spending 35 needs for a more manageable price tag of about 30 12 percent of GDP. Moreover, the large roy- alty payments they received during the recent % of GDP 25 20 commodity boom provide a ready source of 15 finance. Yet resource rich-countries actually 10 lag nonfragile low-income countries in their 5 infrastructure stocks and spend less on infra- 0 tri e structure. They have been devoting their added tri e tri h tri e a un m un m un m un ric ric wealth not to infrastructure development but to co inco co co co inco es es es es co rce- Af -in n - le- u ra ow ow paying off debts. The governance challenges in a so idd ha re el el Sa m gil gil b- resource-rich environment may thus prevent the fra ra Su nf no transformation of wealth into infrastructure. capital operation and maintenance Meeting the infrastructure needs of the middle-income countries looks to be much Source: Briceño-Garmendia, Smits, and Foster 2008. Note: Figures refer to investment (except public sector) and more manageable. These countries should include recurrent spending. Public sector covers general govern- be able to meet their infrastructure spending ment and nonfinancial enterprises. needs with 10 percent of GDP. They are also much stronger in asset maintenance and insti- or dilapidation of their (already modest) tutional efficiency. Their more urban popula- national infrastructure platforms. In the Dem- tions also facilitate network rollout. ocratic Republic of Congo, about 50 percent of infrastructure assets need rehabilitation. The fragile states' infrastructure spending needs Finding 8: A Large Share of Africa's are especially large, particularly when mea- Infrastructure Is Domestically sured against the size of their economies. Such Financed countries would, on average, need to devote 37 percent of their GDPs to infrastructure Existing spending on infrastructure in Africa is spending to build a solid infrastructure plat- higher than previously thought, amounting to form. With their difficult environments, they $45 billion a year when budget and off-budget attract relatively little external financing, cap- spending (including state-owned enterprises turing only 10 percent of overseas development and extrabudgetary funds) and external finan- assistance and 6 percent of private capital flows ciers are taken into account. The latter include allocated to infrastructure. In addition to their the private sector, official development assis- huge financing burden, the fragile states do not tance, and financiers that do not belong to use their current resource envelope well; they the Organisation for Economic Co-operation underspend on maintenance and have ineffi- and Development (OECD). As much as two- cient service providers. thirds of this overall spending is domestically Nonfragile low-income countries need to sourced: $30 billion of annual spending is allocate, on average, about 23 percent of their financed by the African taxpayer and infra- GDPs to build and sustain a basic infrastruc- structure user, and a further $15 billion is from ture platform, a level difficult to envisage in external sources (table O.4). practice. Therefore, these countries will have to The public sector remains the dominant make difficult choices about the prioritization source of finance for water, energy, and transport of their infrastructure investments, and most in all but the fragile states. Public investment is of them have a long way to go in improving the largely tax financed and executed through cen- efficiency of operating existing infrastructure. tral government budgets, whereas the operating The resource-rich countries are, in principle, and maintenance expenditure is largely financed much better placed to meet their infrastruc- from user charges and executed through state- ture spending needs, though in practice they owned enterprises. Current levels of public Africa's Infrastructure: A Time for Transformation 9 Table O.4 Infrastructure Spending on Addressing Sub-Saharan Africa's Infrastructure Needs $ billions annually Operation and maintenance Capital expenditure Infrastructure Public Public Non-OECD Private Total sector sector sector ODA financiers sector Total spending ICT 2.0 1.3 0.0 0.0 5.7 7.0 9.0 Power 7.0 2.4 0.7 1.1 0.5 4.6 11.6 Transport 7.8 4.5 1.8 1.1 1.1 8.4 16.2 WSS 3.1 1.1 1.2 0.2 2.1 4.6 7.6 Irrigation 0.6 0.3 -- -- -- 0.3 0.9 Total 20.4 9.4 3.6 2.5 9.4 24.9 45.3 Source: Briceño-Garmendia, Smits, and Foster 2008. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; ODA = official development assistance; OECD = Organisation for Economic Co-operation and Development; WSS = water supply and sanitation. -- Not available. finance are substantially higher relative to GDP Figure O.4 Infrastructure Public Spending as a in the low-income states, typically absorbing Percentage of GDP 5­6 percent of total GDP (figure O.4). In abso- 7.0 lute terms, however, spending remains very 6.0 low, no more than $20­$30 per capita a year (Briceño-Garmendia, Smits, and Foster 2008). 5.0 Looking only at investment, one finds that % of GDP 4.0 official development assistance, private partici- pation in infrastructure, and non-OECD finan- 3.0 ciers together exceed domestically financed 2.0 public investment (Briceño-Garmendia, Smits, and Foster 2008). The private sector is by far the 1.0 largest source, on a par with domestic public 0 investment. Much smaller, but still significant, tri e tri e es tri e a un m un m un m ric tri co inco co inco co inco es es es Af capital flows are provided by official develop- un n co le- - - ra ow ow idd ch ha ment assistance and, to a lesser extent, non- el el -ri Sa m gil gil ce b- ra fra OECD financiers, such as China, India, and the ur Su nf so no re Arab states. The focus differs markedly in each capital operation and maintenance case. Official development assistance makes an important contribution to water and transport, Source: Briceño-Garmendia, Smits, and Foster 2008. particularly in fragile states. Non-OECD finance is significant in energy and rail, especially in resource-rich countries. Private participation in further--to the tune of $17 billion a year. This infrastructure is heavily concentrated in ICT. is Africa's major infrastructure efficiency gap (Briceño-Garmendia, Smits, and Foster 2008). First, some countries are allocating more resources to some areas of infrastructure Finding 9: After Potential Efficiency than would appear to be warranted (Briceño- Gains, Africa's Infrastructure Garmendia, Smits, and Foster 2008). This Funding Gap Is $31 Billion a Year, "excess expenditure" amounts to $3.3 billion Mostly in the Power Sector a year overall. The largest share of this excess expenditure relates to public spending on ICT Addressing a wide range of inefficiencies could infrastructure that the private sector could pro- make the existing resource envelope go much vide, particularly in middle-income countries. 10 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Although some of this "overspending" may cost of sound preventive maintenance. For be justified by phasing or sequencing, at least example, spending $1 on road maintenance part of these resources could possibly be real- provides a savings of $4 to the economy. So located to underfunded sectors. A need exists some reallocation of resources from invest- to monitor infrastructure expenditure more ment to maintenance may be warranted, par- closely against identified needs and priorities ticularly in low-income countries with very low and considering expected economic returns. maintenance spending. For roads, an estimated Second, African countries are typically $1.9 billion of capital spending on rehabilita- executing only about two-thirds of the budget tion could have been avoided with sound pre- allocated to public investment in infrastruc- ventive maintenance. ture (Briceño-Garmendia, Smits, and Foster Fourth, Africa's power and water utilities 2008). Put differently, public investment could present very high inefficiency in distribution in theory increase by 30 percent without any losses, undercollection of revenues, and over- increase in spending, simply by addressing the staffing (figure O.6). Utilities typically collect institutional bottlenecks that inhibit capital only 70­90 percent of billed revenues, and dis- budget execution. Changes include better plan- tribution losses can easily be twice the technical ning of investment projects, earlier completion best practice. According to household surveys, of feasibility studies, more efficient procure- about 40 percent of those connected to utility ment processes, and a move to medium-term services do not appear to be paying for them, multiyear budgeting. Increasing capital budget a share that rises to 65 percent for a significant execution to 100 percent could capture an addi- minority of countries. Undercollection is also tional $1.9 billion a year in public investment. a problem for some of Africa's road funds Third, on average, about 30 percent of the (Gwilliam and others 2008). State-owned tele- infrastructure assets of a typical African coun- communication incumbents employ roughly try need rehabilitation (figure O.5). This share six times the number of employees per con- is even higher for rural infrastructure and for nection than do privately operated enterprises countries affected by violent conflict. The reha- in developing countries. For ICT, countries bilitation backlog reflects a legacy of under- retaining state-owned incumbents are often funding maintenance, a major waste given incurring significant losses from overstaffing that the cost of rehabilitating infrastructure that average 0.2 percent of GDP. Similarly, is several times higher than the cumulative though to a lesser extent, overemployment in power and water utilities ranges from 20 percent to 80 percent over benchmarks in other developing areas. Overall, the revenues Figure O.5 Rehabilitation Backlog lost through these inefficiencies can easily 50 exceed the current turnover of the utilities by several multiples. For power, these losses are % of assets in need of rehabilitation 40 also material at the national level, absorbing 0.5 percent of GDP on the Sub-Saharan Afri- 30 can average, or $3.4 billion annually (Brice- ño-Garmendia, Smits, and Foster 2008). For 20 water, the absolute value of the inefficiencies is smaller, with the average amount accounting 10 for 0.2 percent of GDP, or $1 billion a year. Fifth, underpricing of infrastructure services 0 is substantial. Although African infrastructure charges are high by international standards, so ion ge s ge n er er ge s ds ad ay tio t at oa ra ra ra wa at ilw ro iga lw e e ve lr er are the infrastructure costs. Even relatively high av av ain ra n la ra irr ra en ba al all ru ru ra m rg ur ur er ru tariffs can fail to cover more than the operat- we nr ov no po ing costs. The revenues uncollected because of Source: Briceño-Garmendia, Smits, and Foster 2008. underpricing of power and water amount to Africa's Infrastructure: A Time for Transformation 11 Figure O.6 Hidden Costs of Utility Inefficiencies a. Power b. Water 5 5 4 4 % of GDP % of GDP 3 3 2 2 1 1 0 0 tri e es tri e tri e a tri e es tri e tri e a un m un m un m ric un m un m un m ric tri tri co inco co inco co inco es es es co inco co inco co inco Af es es es Af un un n co n le- - - co le- - - ra ow ow ra ow ow idd ch ha idd ch ha el el el el -ri Sa -ri Sa m m gil gil gil gil ce b- ce b- ra fra ra fra ur Su ur Su nf nf so so no no re re unaccounted losses collection inefficiencies labor redundancies Source: Briceño-Garmendia, Smits, and Foster 2008. as much as $4 billion a year on aggregate, an irrigation. There is no significant funding gap implicit subsidy for infrastructure consum- for ICT or transport. ers, and that is without taking into account Looking across countries, the dollar sizable subsidies to large industrial customers amount of the funding gap split evenly across that cannot be so readily quantified (Briceño- income groups. Although the largest financing Garmendia, Smits, and Foster 2008). Because gaps relate to capital investment, shortfalls in of the very regressive access to infrastructure funding for operation and maintenance are services in Africa, about 90 percent of those substantial, particularly in fragile states. If who have access to piped water or electricity the infrastructure financing gap is expressed services belong to the richest 60 percent of the as a percentage of GDP, the level of difficulty population (see figure O.9, panel a; Banerjee, involved in closing the gap becomes immedi- Wodon, and others 2008). Thus, better-off ately apparent. The burden associated with the households largely capture any subsidy to resi- infrastructure financing gap is insurmountable dential services. In fact, targeting is so deficient for fragile states. They would need to spend an that a completely random process for allocat- additional 25 percent of GDP on infrastruc- ing subsidies across the population would per- ture to eliminate their infrastructure deficits. form three times better at reaching the poor. Relative to the size of economies, by far the The overall funding shortfall for meeting largest financing gaps are in the energy, trans- Africa's infrastructure needs is given by the port, and water sectors of fragile states (figure difference between estimated infrastructure O.7, panel b). spending needs and a potential resources enve- As shown, the size of the funding gap for lope that includes existing spending and the low-income countries in particular is prob- potential efficiency gains. Even if all these effi- ably more than they could conceivably raise ciency gains could be fully realized, a funding through available funding channels. For this gap of about $31 billion a year would remain particularly challenging group of countries, (table O.5). This gap can be addressed only by additional measures may need to be taken. raising additional finance or alternatively by One option is to extend the time horizon adopting lower-cost technologies or less ambi- for the proposed investment program. Simu- tious targets for infrastructure development. lations suggest that low-income countries Looking across sectors, about 60 percent of could achieve the proposed investment targets the funding gap relates to power (figure O.7, within a period of 20 years without increas- panel a). The remainder relates to water and ing existing spending envelopes, as long as they 12 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table O.5 Finding Resources: The Efficiency Gap and the Funding Gap $ billions annually Cross-sector Item Electricity ICT Irrigation Transport WSS gain Total Infrastructure spending needs (40.8) (9.0) (3.4) (18.2) (21.9) n.a. (93.3) Existing spending 11.6 9.0 0.9 16.2 7.6 n.a. 45.3 Efficiency gap 6.0 1.3 0.1 3.8 2.9 3.3 17.4 Gain from raising capital execution 0.2 0.0 0.1 1.3 0.2 n.a. 1.9 Gain from eliminating operational inefficiencies 3.4 1.2 -- 1.9 1.0 n.a. 7.5 Gain from tariff cost recovery 2.3 -- -- 0.6 1.8 n.a. 4.7 Potential for reallocation n.a. n.a. n.a. n.a. n.a. 3.3 3.3 Funding gap (23.2) 1.3 (2.4) 1.9 (11.4) 3.3 (30.6) Source: Briceño-Garmendia, Smits, and Foster 2008. Note: ICT = information and communication technology; n.a. = not applicable; -- = not available; WSS = water supply and sanitation. Parentheses indicate negative values. Figure O.7 Infrastructure Funding Gap by Sector and Country Type a. Sector b. Country type 4 30 25 3 20 % of GDP 2 % of GDP 15 10 1 5 0 0 r at ly n rt ch a nd tri e nt w- tri h tri e we un m un m tio un ric po nit pp te mun n a ou lo co inco co inco ion log n es s es es co rce- po iga ns rie sa su no tio e c gile m tio y tra irr d er - le- u ow co rma om fra so an t idd wa re el inc non o m inf gil fra capital operation and maintenance Source: Briceño-Garmendia, Smits, and Foster 2008. fully exploit efficiency gains. One cannot say Finding 10: Africa's Institutional, the same of fragile states, however. They would Regulatory, and Administrative still require a substantial increase in spending Reform Process Is Only to meet the investment targets in any reason- able time frame, even when inefficiencies are Halfway Along fully captured. Another possibility is to adopt lower-cost During the last decade, African states have technologies to trim investment needs. Sav- made concerted efforts toward institutional ings of approximately one-third of spending reform in infrastructure. One could probably requirements in transport and in water and san- fairly say that the institutional reform process itation are achievable in this way, by adopting is halfway along (Vagliasindi and Nellis 2009). lower-cost road designs or lower-end solutions They have made progress, but few countries for water and sanitation (such as standposts have a modern institutional framework for and improved latrines). Countries face a stark these sectors. Overall, the greatest progress has trade-off between the level of service provided been in telecommunications, whereas trans- and the speed with which they can serve their port lags furthest behind (figure O.8). The entire population. focus also varies. In telecommunications, the Africa's Infrastructure: A Time for Transformation 13 Figure O.8 Institutional Progress across Sectors percentage score on institutional scorecard a. Telecommunications, electricity, and water b. Ports and railways 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 telecommunications electricity water ports railways reforms regulation governance reforms regulation Source: Vagliasindi and Nellis 2009. Note: See Vagliasindi 2008c for the definition of the institutional indicators. emphasis has been on implementing sector businesses delivers cash flows high enough to reform, and in water on improving the gover- finance investment. However, these arrange- nance of state-owned enterprises. ments have often (though not always) been Private participation has varied enormously good for operational performance, even if (Vagliasindi and Nellis 2009). Since the mid- characterized by renegotiation and premature 1990s, many African countries have experi- cancellation. A growing area of experimenta- mented with various forms of private participa- tion is the multiyear performance-based road tion in infrastructure, with very heterogeneous maintenance contract with the private sector, results (table O.6). which shows promise in safeguarding mainte- The private sector has proved willing to nance activities and keeping costs down. invest only in mobile telephones, power plants, Some progress has occurred with gov- and container terminals. The number of ernance reform of state-owned enterprises, mobile subscribers and the share of the popu- where incentive-based performance contracts lation receiving mobile signals increased by a and external auditing seem to be paying off. factor of 10 in five years, the result of compe- Corporate governance reforms, including the tition among private operators. Private inves- establishment of a somewhat independent tors have also provided significant finance for board of directors, are becoming more prev- thermal power generation (3,000 megawatts) alent across sectors, even if few enterprises and for container terminals at ports, even if have full corporatization that includes limited the volumes fall substantially short of require- liability, rate of return, and dividend poli- ments. Toll-road concessions are confined to cies. Performance contracts with incentives South Africa; traffic volumes elsewhere are and independent external audits have become not enough to make such endeavors financially dominant features of the reform process for self-sustaining. governance of state-owned enterprises, for In power, water, and railways, the pri- both electricity and water. When combined vate sector has delivered improvements in with managerial performance incentives, these operational performance but no new finance. measures seem to be having a material effect The numerous concessions (and related con- on performance. The introduction of inde- tractual forms) covering railways, power, and pendent audits has also increased efficiency, water distribution have not delivered signifi- for both electric and water utilities. cant investment. Because of a combination Evidence on the links between introduc- of low tariffs and low volumes, none of these ing an independent regulator and improving 14 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table O.6 Overview of Private Participation in Infrastructure resulting from excessive discretion and overly Infrastructure Extent of private Nature of broad objectives (Eberhard 2007). Regulatory sector participation experience Prospects autonomy remains elusive: in some countries, ICT turnover among commissioners has been high, Mobile telephony Over 90 percent of Extremely beneficial Several countries still and the gap between law (or rule) and practice countries have licensed with exponential have potential to grant has been wide. For water, where the vast major- multiple mobile operators increase in coverage additional licenses and penetration ity of service providers are state-owned enter- Fixed telephony About 60 percent of Controversial in some Several countries still prises, no evidence exists of any benefit from countries have cases, but have potential to regulation. For power and telecommunications, divested state-owned has helped improve undertake divestitures some effect is discernible, but it is far from telecommunication overall sector efficiency incumbent unambiguous. Weak regulatory autonomy and Power capacity constraints undermine the credibility Power generation 34 independent power Few cancellations but Likely to continue, of independent regulators. Most African regu- projects provide 3,000 frequent given huge unsatisfied latory agencies are embryonic, lacking funding MW of new capacity, renegotiations; power demands and limited and in many cases qualified personnel. investing $2.5 billion purchase agreements public sector capacity have proved costly for utilities Key Recommendations Power distribution 16 concessions and 17 Problematic and Movement toward hybrid management or lease controversial; models involving local contracts in 24 countries one-quarter of private sector in similar Based on these findings, one can make the fol- contracts cancelled frameworks before completion lowing 10 key recommendations: Transport · Addressing Africa's infrastructure efficiency Airports Four airport No cancellations but Limited number of gap is a pressing policy priority with poten- concessions, investing some lessons learned additional airports tial dividends of $17 billion a year. less than $0.1 billion, viable for concessions plus some divestitures · One of the most flagrant inefficiencies is the Ports 26 container terminal Processes can be Good potential to failure to maintain infrastructure assets-- concessions, investing controversial, but continue maintenance needs to be understood as an $1.3 billion cancellations have investment in asset preservation. been few and results positive · Institutional reform remains essential for Railroads 14 railroad concessions, Frequent renegotia- Likely to continue but tackling utilities' operational inefficien- investing $0.4 billion tions, low traffic, and model needs to be cies, both through private participation costly public service adapted obligations keep and through governance reforms for state- investment below owned enterprises. expectations Roads 10 toll-road projects, No cancellations Limited because only · Institutional reform should also go beyond almost all in South Africa, reported 8 percent of road utilities to strengthen the planning func- investing $1.6 billion network meets minimum tions of the line ministries and address seri- traffic threshold, almost ous deficiencies in the budgetary process. all in South Africa Water · Reforms are needed to get full value from existing infrastructure, where widespread Water 26 transactions, mainly Problematic and Movement toward hybrid management or lease controversial; models involving local administrative and regulatory bottlenecks contracts 40 percent of private sector in similar prevent facilities from being fully used. contracts cancelled frameworks before completion · Regional integration can contribute signifi- Sources: Authors' elaboration based on Bofinger 2009; Bullock 2009; Eberhard and others 2008; cantly to reducing infrastructure costs, by Gwilliam and others 2008; Minges and others 2008; Mundy and Penfold 2008; and Svendsen, allowing countries to capture scale econo- Ewing, and Msangi 2008. Note: ICT = information and communication technology; MW = megawatts. mies and manage regional public goods effectively. performance is currently mixed (Vagliasindi · Development of infrastructure networks and Nellis 2009). Some critics argue that regu- needs to be strategically informed by the latory agencies have simply created additional spatial distribution of economic activities risks because of unpredictable decisions, and by economies of agglomeration. Africa's Infrastructure: A Time for Transformation 15 · Infrastructure's social policy needs to be infrastructure inefficiencies. As African coun- rethought, placing more emphasis on recov- tries begin to feel the pinch of the global finan- ering costs from those who can afford it and cial crisis, and as other sources of funding begin on recasting subsidies to accelerate access. to dry up, measures to improve the efficiency · Achieving universal access will call for of using existing resources become particularly greater attention to removing barriers that attractive. Such measures provide an addi- prevent the uptake of services and offer- tional internal source of finance at a relatively ing practical and attractive second-best low monetary cost. Of course, in some cases, solutions. significant investments may be needed before efficiency gains can be captured (for example, · Closing Africa's infrastructure financing reducing distribution losses in power or water). gap is critical to the region's prosperity, and the global financial crisis has only made In other cases, the economic context of the crisis infrastructure more relevant. may simply increase the political cost of taking such measures, such as raising cost recovery or laying off excess employees. Recommendation 1: Address Africa's Potential efficiency gains take a wide variety Infrastructure Efficiency Gap as a of forms, which are developed in the recom- Pressing Policy Priority mendations that follow. Briefly, they include The findings presented underscore the magni- the following areas: tude of inefficiency with which Africa spends its current infrastructure resources. Of Africa's · Safeguarding maintenance expenditure to overall infrastructure spending needs of about avoid wasting resources on the repeated $93 billion a year, as much as $17 billion could rehabilitation of existing assets, which could be met simply by using existing resources save $2.6 billion a year in avoidable capital more effectively. expenditure for the roads sector alone Reaping this efficiency dividend has to be · Reforming institutions to improve the oper- a major policy priority for the region, and ational performance of utilities and other efforts to scale up infrastructure finance need service providers that are currently wasting to be made in the context of genuine com- $6 billion a year on inefficiencies such as mitments to address efficiency. Pouring addi- overstaffing, undercollection of revenues, tional funding into sectors characterized by and distribution losses high levels of inefficiency makes little sense. · Addressing deficiencies in the public expen- However, postponing increases in finance until diture framework, where $3.3 billion a year efficiency improves is not a valid option: the of infrastructure resources appear to be cost to economic growth and human develop- poorly allocated across sectors and low bud- ment is simply too high. Rather, development get execution prevents $1.8 billion a year of partner efforts to secure additional resources public investment funds from being spent for infrastructure finance must be matched by government efforts to improve their efficiency · Modernizing administrative and regulatory frameworks to reduce bottlenecks that pre- in using such resources. Parallel progress is vent services from being provided effectively needed on both fronts. across existing infrastructure networks and Moreover, investment finance is needed in impose substantial costs on infrastructure some cases to allow inefficiencies to be captured users (for example, where roads must be rehabilitated before they return to a "maintainable" condition · Reaping the economies of scale and coor- or when meters must be installed to improve dination benefits associated with regional revenue collection). These kinds of efficiency- integration, which in the case of power related investments deserve to be prioritized alone can be as high as $2 billion a year because of the high returns they typically bring. · Securing the highest returns from new The current global financial crisis only infrastructure investments by using them strengthens the motivation for addressing to secure economies of agglomeration and 16 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION to facilitate the development of productive of multiyear performance-based contracts for activities along key economic corridors roads has further contributed to the efficacy and efficiency of road maintenance. These · Rethinking infrastructure social policy to place more emphasis on cost recovery findings illustrate that a combination of fund- from those who can afford to pay, and ing mechanisms, institutional capacity, and redirecting the current $4 billion a year of contractual incentives is needed to overcome subsidies to accelerate access among lower- the maintenance challenge. income groups Donors have traditionally eschewed fund- ing maintenance, arguing it is more sustainable · Reducing the costs of meeting key infra- for funding directly from country budgets. The structure targets by adopting lower-cost argument is a good one. However, the willing- technologies that provide reasonable lev- ness of donors to fund asset rehabilitation els of service at a price that is affordable to can create perverse incentives for countries both consumers and the government. to neglect maintenance, because governments face a choice between raising taxes today to Recommendation 2: Make Greater finance maintenance or simply waiting a few Efforts to Safeguard Maintenance years to obtain subsidized donor capital for Spending reconstruction. In low-income, low-capacity The traditional neglect of maintenance expen- environments where maintenance is unlikely diture needs to be reversed by rethinking to be forthcoming, donors may be well advised maintenance as asset preservation. One-third to take this choice explicitly into account in of Africa's infrastructure assets need rehabilita- project design, rather than simply assume that tion, indicating that historic neglect of main- maintenance will happen. One way of doing tenance is endemic. For fragile states and for so is to choose more capital-intensive, low- rural infrastructure, the share of assets needing maintenance technologies. Even if they rep- rehabilitation is much higher. The shortfall in resent a higher investment cost in the short road maintenance spending is costing Africa run, overall life-cycle costs may be lower if $1.9 billion a year in avoidable capital expen- reconstruction can be avoided or postponed. ditures. In fact, spending $1 on maintenance As donors move toward sectorwide budget can provide a savings of approximately $4 to support, they will have a greater opportu- the economy. nity to ensure that maintenance spending is Thus, Africa's infrastructure financing gap adequately supported in the budget envelope. is not only about raising investment capital; a In any case, as a general principle, the estab- substantial part of it relates to maintenance. Yet lishment of a sound framework for financing maintenance offers one of the highest returns maintenance should be a prerequisite for the to infrastructure spending, so it may be more funding of major capital programs. helpful to think of maintenance as a kind of investment in asset preservation. Recommendation 3: Tackle The road sector shows that maintenance Inefficiency through Institutional can be improved through suitable institutional Reform reforms. Since the mid-1990s, the majority of Since the mid-1990s, the institutional agenda African countries have established road funds has broadened and deepened (Vagliasindi and as a means of channeling road user charges Nellis 2009). In the 1990s, the emphasis of to network maintenance. Countries with institutional reform was on sector restructur- road funds do significantly better at raising ing and private participation, transplanting adequate maintenance funds as long as the to Africa experiences from other parts of the fuel levies paid into these funds are set high developing world. This approach yielded dra- enough to provide material financing. More- matic results in telecommunications, but else- over, countries with both road funds and road where the benefits were more limited and the agencies do significantly better in safeguard- experiences more problematic. Even so, private ing the quality of their road networks. The use finance to African infrastructure came from Africa's Infrastructure: A Time for Transformation 17 nowhere to provide a flow of funds comparable enterprises by their supervisory agencies, in scale to overseas development assistance. whether line ministries or ministries of finance. A more nuanced, less dogmatic perspec- Transparency and accountability of state- tive on the private sector has emerged. This owned enterprises depend on solid systems perspective values private financing in mobile of financial management, procurement, and telephony, power generation, and ports, while management information. Today, basic oper- recognizing its limits in roads, rail, power, and ational and financial data on firm perfor- water (see table O.6). Even for infrastructure mance are not produced, reported, or acted where the proven appetite for private finance on. Without information or, perhaps worse, is very limited, the potential contribution of without action on what information is pro- the private sector to tackling costly manage- duced, better outcomes cannot be expected. ment inefficiencies (undercollected utility Key measures include auditing and publishing revenues, low labor productivity, or neglected financial accounts and using comprehensive road maintenance) remains valuable. Indeed, cost-based accounting systems that allow the the efficiency gains from such performance functional unbundling of costs and a clearer improvements are themselves a significant sense of cost centers. After this foundation is source of sector finance. Moreover, the concept in place, contracting mechanisms can improve of private participation has undergone signifi- performance--within the public sector or with cant expansion. More emphasis has fallen on the private sector. the local (not international) private sector and Public sector performance contracts need on hybrid models that experiment with differ- strong performance incentives. Initial attempts ent ways of allocating responsibilities between to improve African state-owned enterprises public and private partners. through performance contracts with their line Another important way in which the insti- ministry or other supervisory agency were tutional reform agenda has broadened is the minimally effective. Recent efforts in water greater focus on the quality of governance for (Uganda), however, have had a much more enterprises that remain state owned (Vagli- positive effect. The key feature of these con- asindi and Nellis 2009). The recognition that tracts is to incorporate incentives for good the private sector will never be a ubiquitous managerial (and staff) performance and, more service provider has come with the realiza- rarely, sanctions for failure to reach targets. tion that state-owned enterprises are here to Creating effective performance incentives stay. Therefore, it is necessary to recommit to in the public sector can be challenging, mak- the difficult process of reforming state-owned ing management contracts with the private enterprises. sector a relevant option. Either expatriate or Renewed efforts on state-owned enterprise local management teams can be contracted reform should favor governance over technical with, each of which offers advantages. Clarity fixes. Fortunately, better governance of state- about what a contract can and cannot achieve, owned enterprises can improve performance. particularly given its short time horizons, is Past efforts at improving utility management essential. At best, a management contract can focused too heavily on technical issues at the improve performance in a handful of rela- expense of corporate governance and account- tively manageable aspects of efficiency, such ability. Future state-owned enterprise reforms as revenue collection and labor productivity. It seem justified as long as they focus on deeper cannot solve deficiencies in the broader insti- institutional issues. Key measures include tutional framework; ideally, these should be greater decision-making autonomy for the addressed beforehand. Nor can a management board of directors, more objective selection contract raise investment finance or deliver criteria for senior managers, rigorous disclo- major effects on service quality that require sure of conflicts of interest, and more trans- substantial investments or lengthy gestations. parent, merit-based recruitment processes. In principle, regulation can do much; but in Parallel efforts can strengthen financial practice, regulation has proved difficult. Regu- and operational monitoring of state-owned lators have been set up across Africa, precisely 18 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION to insulate utilities from political interference A clear example is power generation. Tra- while closely monitoring enterprises. Improving ditionally, planning and procurement of new regulatory performance is a long-term process power infrastructure were the province of the to be pursued where private participation and state-owned utility. With power sector reforms competitive pressures are significant. The chal- and independent power producers, those func- lenge of establishing new public institutions in tions were often moved to the ministry of developing countries is often underestimated. energy or electricity. The transfer of skills was Independent regulation requires a strong polit- not always simultaneous, however, so plans ical commitment and competent institutions were not adequately informed by the complex- and people. Where some or all are lacking, con- ities on the ground. In many cases, planning sidering complementary or transitional options has collapsed. New plants are rarely timely, that reduce discretion in regulatory decision thereby opening power gaps that prompt making through more explicit rules and proce- recourse to temporary power and discour- dures or by outsourcing regulatory functions to age investors. When procurement is (finally) advisory regulators and expert panels may be undertaken, the authorities may not take the wise (Eberhard 2007). trouble to conduct international competitive bidding. This outcome is unfortunate because Recommendation 4: Include Line a rigorous bidding process lends credibility Ministries and Budgetary Processes on and transparency to procurement and results the Institutional Reform Agenda in more competitively priced power. Much of the emphasis of recent reforms has Because domestic public spending finances been on restructuring the service provider or the bulk of Africa's infrastructure investments, utility, bringing in private management, applying development partners need a broader view regulatory oversight, and so on. Little attention of the quality of public spending. Across the has been given to institutional strengthening of infrastructure sectors, most investments are by the sector line ministries. These line ministries line ministries through the budgetary process. have responsibilities, which, if not adequately Shortcomings in the way the rest of the sec- discharged, can jeopardize the functioning tor budget is allocated and spent may offset of the sector. They take the lead in sector development finance that focuses too narrowly planning, participate in the formulation of the on specific project interventions. So donor public budgets, and execute investments. How- resources are best channeled programmatically ever, deficiencies exist in all those areas. Unless as budgetary support or through sectorwide they are tackled head on, the effect of reforms projects, and development partners need to on service providers will remain limited. take a broader interest in the overall quality of Stronger sector planning is needed in infra- public spending. Thus, infrastructure interven- structure line ministries to ensure that the tions must be grounded in a broader under- construction of critical new assets begins early standing of the public expenditure framework enough to come on stream when needed. Too in each sector. often overlooked or debilitated during the Ad hoc political priorities with little or no course of sector restructuring efforts, plan- economic screening too often characterize the ning is a critical sector function. It is essential budgetary process. The annual budget cycle to restore this vital planning capability in the prevents adequate follow-through on the fund- line ministries and to develop sound techni- ing of multiyear infrastructure projects. When cal methodologies for identifying and selecting it comes to implementation, many countries infrastructure projects. More rigorous project have significant problems with budgetary screening can ensure that infrastructure invest- execution, with procurement bottlenecks pre- ments are selected according to their expected venting the full budget allocations from mate- returns and are appropriately sequenced and rializing in actual spending. synchronized with one another and with Key aspects of the public expenditure broader development plans to maximize syn- framework need to be addressed. The budget- ergies and avoid costly bottlenecks. ing process needs to move to a medium-term Africa's Infrastructure: A Time for Transformation 19 framework and link sector objectives and Liberalizing the trucking industry can resource allocations, underpinned by clear reduce the exorbitant road freight costs in sector plans that go down to specific activities Central and West Africa. The regulation and and their associated costs. The careful incor- market structures of the road freight industry, poration of maintenance in medium-term not the quality of road infrastructure, are the sector-planning tools can prevent the growing binding constraints on international corridors need for asset rehabilitation. Project appraisal (Teravaninthorn and Raballand 2008). Road should underpin the budgetary process for freight tariffs, which can reach $0.08­$0.13 public investment to ensure that all invest- per ton-kilometer in Central and West Africa, ments under political consideration pass at reflect the high profit margins of trucking least a minimum threshold of economic via- services (60­160 percent). The tour de role bility. Administrative processes that delay the regulatory framework, based on market shar- release of budgeted funds must be overhauled, ing and centralized allocations of freight, lim- and procedures for procurement, disbursement, its vehicle mileage and undermines incentives financial management, and accountability to improve fleet quality. The alternative is to must be modernized and streamlined. combine free entry to the market and market Water provides interesting examples of pricing with regulatory enforcement of rules how bottlenecks in the budgetary process for quality and operating behavior. Already can prevent the use of available resources. In practiced in southern Africa, these reforms West Africa, the binding constraint is not the can reduce road freight tariffs to $0.05 per availability of budgetary resources in many ton-kilometer. Without such reforms, further instances but the capacity to disburse them in investments in upgrading road network qual- a timely fashion (Prevost 2009). In Tanzania, ity will simply lead to higher profit margins for steep increases in budget allocations to the sec- the trucking industry without lowering trans- tor followed water's identification as a priority port costs for consumers. in the country's poverty reduction strategy, One-stop border posts are essential to avoid but disbursements increased at a much slower extensive delays in transit traffic along interna- pace, thus impeding any immediately discern- tional road corridors. Road conditions along ible effect on access (Van den Berg 2009). Africa's major international corridors are good, Parallel improvements are also needed in the with trucks reaching speeds of 50­60 kilome- way donor finance is channeled. Given the rele- ters an hour, but long delays at borders slow vance of external funds, a solid public expendi- effective velocities to little more than 10 kilo- ture management system for African countries meters an hour. A journey of 2,500 kilometers requires that donors improve the predictability from Lusaka, Zambia, to the port of Durban of their support and streamline and harmonize in South Africa takes on average eight days-- their procedures. In that sense, a focus on mul- four days of travel time and four days spent tidonor initiatives that pool funds to provide at border crossings. Compare that total with general budgetary support for a sectorwide land border-crossing times of no more than program of interventions is preferable. half an hour for industrialized countries. The cost of delays for an eight-axle interlink truck Recommendation 5: Use Administrative has been estimated at about $300 a day. The and Regulatory Reforms to Get Full investments to develop one-stop border facili- Value from Existing Infrastructure ties and to modernize customs procedures are Africa is failing to get the full development relatively modest and would pay back in barely potential even from its existing infrastructure a year. Without such reforms, further invest- networks. Administrative and regulatory fail- ments in the road network will have little effect ures create bottlenecks and prevent infrastruc- on overall transit times. ture assets from delivering the services they More reliable interconnection services can are supposed to. These problems are particu- avoid even longer delays on international rail larly evident in transport, where high-impact corridors. Locomotives from one country are reforms are urgently needed. generally not allowed to travel on another 20 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION country's network, mainly because of the boosted handling rates. Modernizing customs inability to provide breakdown assistance administration requires modern information to foreign operators. As a result, rail freight technology and associated database systems. crossing borders must wait to be picked up Such soft infrastructure has traditionally been by a different locomotive. These delays can be underfunded, contributing to poor port effi- extensive. A journey of 3,000 kilometers from ciency. Governance issues may also afflict cus- Kolwezi on the Democratic Republic of Congo toms administration. border to the port of Durban in South Africa Port and land distribution infrastructure takes 38 days--including 9 days of travel time need to be integrated. The lack of an inte- and 29 days associated primarily with loading grated land distribution system, particularly and interchange of freight. This delay partly for transit traffic, further impedes container reflects the lack of reliable, well-maintained traffic. Making the most progress are dry and locomotives, but it also reflects the absence of liquid bulk exports, where many port facilities clear contractual incentives to service traffic are privately owned and integrated within a from a neighboring country's network. Reduc- comprehensive logistics system. Container- ing such delays would require total rethinking ized trade, in contrast, is often only skin-deep. of contractual relationships and access rights Containers are packed and unpacked near the linking the railways along the corridor. It ports, and the benefits of fully integrated mul- would also likely require a regional clearing- timodal transport corridors associated with house to ensure transparency and fairness in container adoption are not secured. As a result, reciprocal track access rights. little containerized traffic moves into the land- Slow movement of containers and cargo locked hinterland, and most of those countries' through Africa's ports imposes very high eco- imports are transported as general cargo. nomic costs. Many firms cite bottlenecks at Overall, the transport regulatory and ports as their most pressing infrastructure administrative framework needs to promote constraint in countries as diverse as Burkina seamless multimodal transportation networks Faso, Cameroon, Malawi, Mauritius, and South more consciously. Transport chains can be no Africa. Container dwell times in East and West stronger than their weakest links, which are Africa are 12­15 days, twice the international usually the interchanges between different best practice of 7 days. Most delays are caused modalities--such as road to rail or rail to sea. by long processing and administration times The weaknesses are partly physical, where no and poor handling in congested port areas, physical connection exists between the modes rather than by any real limitations in basic and no infrastructure is available for transship- quay capacity. These delays can be very costly. ment. However, they are also partly institu- One extra day in port costs more than $35,000 tional, with responsibility for the interchanges for a 2,200-TEU (20-foot equivalent unit) ves- not falling clearly to one modal agency or the sel in 2006 and proportionately more for larger other. Finally, they are partly operational, with ships. Shipping lines have responded by intro- the government collecting taxes and duties, ducing "congestion charges": for a 20-foot or staff collecting bribes, slowing movements, container in 2006, ranging from $35 a day in and pushing up costs. Even at the sector policy Dakar, Senegal, to $420 a day in Tema, Ghana. and planning level, Africa's transport modes The solution lies in modernizing customs are too often parceled out across separate line administration and improving efficiency of ministries, thereby preventing a cohesive inter- cargo handling. The two main bottlenecks modal transport framework from emerging. within ports are loading and unloading of cargo and customs administration--both need Recommendation 6: Pursue to be addressed simultaneously. Inadequate Regional Integration to Reduce cranes are part of the problem, but new equip- Infrastructure Costs ment alone will not deliver better performance Regional integration lowers costs across all unless staff practices are also modernized. aspects of infrastructure. The high cost of infra- Ports with container terminal concessions have structure services in Africa is partly attributable Africa's Infrastructure: A Time for Transformation 21 to fragmentary national boundaries preventing infrastructure involves a high level of trust achievement of scale economies. between countries, not least because of the In ICT, power, ports, and airports, regional implied dependence on neighbors for key collaboration essentially provides scale econ- resources, such as energy and water. For example, omies that reduce the cost of service. Most if regional power trade were pursued fully, 16 African countries are simply too small to African countries would import more than half develop infrastructure cost-effectively on their power needs. A large share of that power their own. In ICTs, regional collaboration in would come from fragile states, such as the continental fiber-optic submarine cables can Democratic Republic of Congo and Guinea. reduce Internet and international call charges Regional institutions are needed to facili- by half, relative to national reliance on satel- tate agreements and implement compensa- lite communications. In power, 21 countries tion mechanisms. Some countries have more have national power systems below the mini- to gain from regional integration than others mum efficient scale of a single plant. By shar- do. As long as regional integration provides ing large-scale, cost-effective energy resources a substantial economic dividend, one should across countries, regional trade can reduce be able to design compensation mechanisms electricity costs by $2 billion a year. The traf- that make all participating countries better off. fic flows to most of Africa's national ports and Benefit sharing was pioneered through inter- airports are too low to provide the scale econ- national river basin treaties, such as that for omies needed to attract services from major Senegal, and could be applied to other regional international shipping companies and airlines. infrastructure more broadly. Africa has an Regional collaboration in multicountry hubs extensive architecture of regional political and can help overcome this problem. technical bodies, but they have overlapping In road and rail corridors and transbound- memberships, limited technical capacity, and ary river basins, collaborative management of limited enforcement powers. Nor do they cur- these regional public goods reduces the cost. rently have the capacity to implement cross- Many of Africa's infrastructure assets and natu- border compensation mechanisms. ral resources are regional public goods that cut Moving on regional projects that deliver across national frontiers and can be effectively quick wins is important. Because of the daunt- developed and maintained only through inter- ing investment agenda, better sequencing and national collaboration. Road and rail corridors priority setting for regional projects are needed. need to be managed collaboratively to smooth Political, economic, and spatial approaches transport and trade services to Africa's 15 land- have all been widely discussed. Regional proj- locked countries, avoiding the extensive border ects range from bilateral cooperation on a delays that slow international road freight to transmission line or border post to vast and 10 kilometers an hour. Africa's 63 international complex interventions, sometimes with a con- river basins call for cooperative water resource tinental reach. Given the size of the challenges, management and coordinated investments to starting small with projects that deliver tangible increase basin yields of food, power, and other high returns and building incrementally on ini- economic opportunities, while strengthening tial successes may be advisable. environmental sustainability and mitigating Regulatory harmonization needs to go hand the effects of droughts and floods. in hand with physical integration. Unless regu- Reaping these benefits poses numerous latory frameworks and administrative proce- institutional challenges. Among them are dures are harmonized to allow the free flow of mobilizing political will, developing effective services across national boundaries, physical regional institutions, setting priorities soundly, integration of infrastructure networks will harmonizing regulatory procedures, and facili- be ineffective. Making progress on regulatory tating project preparation and finance. reform has a relatively low monetary cost, but Notwithstanding the economic case for it can have a very high return. A good example regional integration, the mobilization of politi- is the Yamoussoukro Decision: opening the cal will faces considerable obstacles. Regional skies for air transportation across Africa, it has 22 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION led to greater freedom in the negotiation of spatially uncoordinated manner. In Africa-- bilateral agreements. too often--the limited infrastructure available Greater efforts are needed to facilitate prepa- is thinly spread out, preventing such synergies ration of complex regional projects, which are from being captured. particularly costly and time-consuming to pre- The urbanization process calls for a regional pare. That is especially true when projects are development perspective on infrastructure that large in relation to the size of the host economy looks at each city and its rural hinterland as and when they essentially depend on financing an integrated economic unit. Africa is urban- from downstream beneficiaries. They also stretch izing fast, creating change that is predictable the donor financing systems that are more typi- and beneficial for both urban and rural areas. cally geared toward national investments. Prosperity and density go together, as changes in productivity require agglomeration econo- Recommendation 7: Take a Spatial mies, larger markets, and better connectiv- View of Infrastructure Development ity. Concentration and urbanization trigger Priorities prosperity in both urban and rural areas, and Infrastructure networks are inherently spatial, well-functioning cities facilitate the transition both reflecting and underpinning the spatial from subsistence agriculture by providing a distribution of economic activity. Infrastruc- large market for rural products and support- ture plays a key role in enabling cities to benefit ing nonfarm activities. The debate of rural from economies of agglomeration. Transport versus urban development should therefore networks interconnect urban centers with each be replaced by the understanding that rural other and with international trading networks, and urban development are closely linked and providing the basis for exchange between the mutually dependent--and that economic inte- urban and rural economies. Energy, water, and gration of rural and urban areas is the only way ICT all enhance productivity within urban and to produce growth and inclusive development. rural spaces. Therefore, infrastructure plans In urban areas, deficiencies in land policies and priorities should be strategically informed and planning have become a huge impediment by a clear understanding of the spatial dis- to extending infrastructure services. African tribution of economic activity and potential. cities are growing fast, but with insufficient A clear example of this approach is the Spatial infrastructure and poor institutions, most Development Initiative of the New Partnership new settlements are informal and not cov- for Africa's Development (NEPAD). ered by basic services. Urban planning should The spatial lens is a useful basis for pri- be strengthened to reduce sprawl, enhance oritization of infrastructure investments and densification, prevent development in pre- provides insight into cross-sectoral links. carious environmental zones, and provide Looking at infrastructure through a spatial the appropriate balance between public and lens allows identification of the key bottle- private land to safeguard key trunk networks. necks along various trading corridors, which Property rights must be clearly defined so that are typically the highest-return interventions. land markets can function. Cities frequently Cross-sectoral links also become more appar- lack the financial basis to develop the infra- ent through a spatial view, shedding light on structure critical to their success. The local the need for coordinating interventions across tax base, though potentially large, is typically infrastructure sectors and between infrastruc- unexploited, leaving municipalities reliant on ture and client economic sectors. An emerg- central government transfers, which are too ing literature suggests that because of synergy often inadequate or unpredictable. effects, the returns from bundling multiple Large agricultural sectors and rural econo- infrastructure interventions in a particular mies remain central to economic growth and spatial area (Torero and Escobal 2005) or along poverty reduction in Africa. Yet the access of a given spatial corridor (Briceño-Garmendia rural populations to infrastructure is extremely and Foster 2009a, 2009b) are higher than low. Rural roads and irrigation systems are those from making the same investments in a together perhaps the most pressing of rural Africa's Infrastructure: A Time for Transformation 23 infrastructure needs. The two go hand in hand, Figure O.9 Access to and Affordability of and their development should follow the value Household Services of agricultural land and the spatial proximity a. Access by quintile to urban markets. ICT has made huge strides 80 in expanding rural access, with one in two % of households 60 rural Africans now in range of a global systems mobile signal. This platform can contribute to 40 agricultural productivity through simple text- 20 message extension services, through bulletins on agricultural market prices and meteoro- 0 Q1 Q2 Q3 Q4 Q5 logical conditions, and as a vehicle for finan- budget quintile cial transactions. The possibilities are only just piped water electricity beginning to be explored. b. Affordability curve Recommendation 8: Rethink 100 Infrastructure Social Policy % of households falling affordability threshold beyond the 5 percent 80 Although Africa's infrastructure services are rel- atively expensive, costs remain even higher than 60 prices, and this lack of cost recovery has major 40 detrimental effects. Underpricing infrastructure 20 services is costing Africa $4.7 billion a year in forgone revenues. In addition, because of ineq- 0 2 4 6 8 10 12 14 16 uitable access to infrastructure services, these monthly utility bill (US$) subsidies are highly regressive, largely bypass- low-income countries ing the poor (figure O.9). The underrecovery middle-income countries of costs impairs the financial health of utilities Sub-Saharan Africa overall and slows the pace of service expansion. Concerns about affordability are usually Source: Banerjee, Wodon, and others 2008. Note: Q1 = first (or poorest) budget quintile; Q2 = second the pretext for underpricing services but do budget quintile; Q3 = third (or middle) budget quintile; not bear much scrutiny (figure O.9). A subsis- Q4 = fourth budget quintile; Q5 = top (or richest) budget quintile. tence-level monthly utility bill priced in cost- recovery terms typically amounts to $6­$10 a month. In the middle-income countries, bills The affordability of services depends of this magnitude do not appear to present not only on prices, but also on the type of an affordability problem anywhere across the payment arrangements that are made avail- income spectrum. Nor do bills of this mag- able to consumers. Prepayment (pioneered nitude pose affordability issues for the more in the mobile telephone sector) can help affluent groups in low-income countries, the households budget their consumption and main ones to enjoy access to services. Afford- reduce revenue risks for operators. The same ability would become a binding constraint in approach is technologically feasible for elec- low-income countries only when service cov- tricity, and a growing number of power utili- erage starts to exceed 50 percent. Only in the ties are adopting it. poorest of countries, and those with excep- Subsidies are important, but subsidy design tionally high infrastructure costs, does full needs major rethinking, with a sharper focus cost recovery seem unachievable for today's on subsidizing connections, which can be more affluent consumers. Even in these cases, more equitable and effective in expanding operating cost recovery should be a feasible coverage. The affordability problems with con- objective, with subsidies limited to capital nection charges are often much more serious costs. Simulations suggest that raising tariffs to than those with use-of-service charges. More- cost recovery would have only minimal effects over, the absence of a connection may itself on poverty rates in most cases. be a good targeting variable for identifying 24 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION disadvantaged households, although less so that Africa can widely and rapidly adopt in a low-access environment where coverage modern infrastructure services. Low charges may be far from universal, even among afflu- for initial connection make market entry ent households. affordable. Prepayment schemes eliminate An important test of the coherence of a credit risk and give customers full control over subsidy policy is to see whether it would be their spending. Services are well tailored to affordable for the country under universal customer demands. Other network services, access. The existing underpricing of utility notably power and water, have tended to view services that benefit just a small minority costs access as a matter of simply rolling out new many African countries as much as 1 percent networks, overlooking the fact that even where of GDP. As countries move toward universal networks are available, the hookup rates are access, that subsidy burden would increase relatively low. They need to pay greater atten- proportionately, rapidly becoming unafford- tion to demand-side issues that prevent cus- able for the national budget. Countries should tomers from making connections: connection thus consider how the cost of any proposed charges that are much higher than household subsidy policy would escalate as coverage incomes, as well as tenure and urban devel- increases. This test of the fiscal affordability of opment issues. The most cost-effective way a subsidy is an important reality check that can to increase access for many utilities may be prevent countries from embarking on poli- through densification programs that increase cies that are simply not scalable and will keep hookups to existing networks by using greater coverage low. community outreach to understand better the demand side of the market. Recommendation 9: Find Practical Second-best alternatives can be fine-tuned Ways to Broaden Access to to provide feasible and attractive infrastructure Infrastructure Services services to those otherwise unserved. The vast Universal access to infrastructure services majority of those without access to modern remains distant for most African countries. infrastructure services rely on traditional alter- The vast majority of African households today natives, such as candles, wells, or unimproved lack access to modern power, piped water, sew- latrines. Although doing the job, these tradi- erage, and even all-season roads that service tional alternatives tend to be inconvenient, their communities. The very slow progress in inferior, or unsafe. Second-best solutions, such expanding this access since the mid-1990s sug- as street lighting, solar lanterns, standposts, gests that universal access to infrastructure is and improved latrines, would provide house- more than 50 years away for most countries holds with superior services at a cost that is in Africa. somewhat higher than the traditional alterna- This situation calls for a different approach tives but still falls far short of modern services. to expanding modern infrastructure services Puzzlingly, these second-best solutions are not and for greater attention to second-best alter- very prevalent in Africa, and even where they natives. Business as usual will not bring about exist, they tend to be available primarily to the the acceleration of infrastructure access that more affluent. Africa needs. Moreover, even if access can be A key problem seems to be the public-good accelerated, many people will have to continue nature of many of these solutions (such as to rely on alternatives to modern infrastruc- standposts and street lighting), which makes it ture services for many years to come. There- difficult for service providers to recover costs fore, infrastructure social policies in Africa and greatly complicates the administration of need to give greater thought to improving and the facilities. Effective institutional arrange- expanding second-best alternatives. ments must be found to support implemen- In expanding modern infrastructure tation of these alternatives. Another problem networks, closer attention should be paid to is that some of these alternatives, although the demand side of the equation. The mobile cheaper, may simply not be cheap enough to telephone revolution has clearly demonstrated be widely affordable. Africa's Infrastructure: A Time for Transformation 25 Recommendation 10: Close the 2 percent. Growth in Latin America and Asia Infrastructure Funding Gap was compromised in a "lost decade." Notwithstanding the importance of all these Many countries, ranging from China and efficiency measures, a substantial infrastruc- India to Argentina and Mexico, have used ture financing gap of $31 billion a year remains. infrastructure-based fiscal stimulus in times of Such a large shortfall looked daunting even economic crisis. If well targeted to addressing before the onset of the global financial crisis. key economic bottlenecks and complemented As of year-end 2007, many factors had by policy reforms, infrastructure investments converged to bring about rapid and sustained can pave the way for the later resurgence of increases in all major sources of external economic growth. Furthermore, some kinds finance for African infrastructure. Following of public works contracts are labor intensive, the Gleneagles Summit, OECD development creating short-term employment to alleviate assistance placed greater emphasis on sup- poverty. Although Africa could benefit from porting African infrastructure. Official devel- such a program, the continent does not have opment assistance flows almost doubled, from the means to finance it without external sup- $4.1 billion in 2004 to $8.1 billion in 2007. The port. Estimates suggest that a $50 billion stim- resurgence of economic growth on the conti- ulus package would be needed to offset the nent led to an upswing in private participation. impact of the economic crisis on Africa, and Since the late 1990s, private investment flows that focusing such a package on infrastructure to Sub-Saharan infrastructure almost tripled, investments would have the largest short-term going from about $3 billion in 1997 to $9.4 effect on GDP growth, boosting projections for billion in 2006/07 (about 1.5 percent of regional 2010 to 4 percent, compared with the postcrisis GDP). In addition, non-OECD countries-- 1.7 percent. In the long term, Africa would see notably China and India--began to take a a permanent increase of 2.5 percent of GDP growing interest in financing infrastructure (ODI 2009). within a framework of South-South coop- Any increase in donor finance for African eration. Their commitments rose from almost infrastructure should pay particular attention nothing in the early 2000s to finance about to the power sector and to the fragile states. $2.6 billion of African infrastructure annually Donors have neglected power since the 1990s. between 2001 and 2006. Although disburse- Although the private sector can contribute to ments tend to lag commitments by several funding power generation, donors will still years, if the record commitments of 2007 are need to scale up substantially to address the fully honored, the disbursements of external current crisis in the sector. This scale-up was finance for African infrastructure may con- already under way before the onset of the cri- tinue to increase over the next few years. sis, with donor commitments that first topped In the absence of any offsetting measures, $1 billion a year in 2005 reaching a peak of domestic infrastructure spending would likely $2.3 billion in 2007. Fragile states stand out fall, compromising economic recovery and as receiving less than their fair share of donor deepening poverty. The existing gap of $31 finance for infrastructure. Given the magni- billion a year could widen further as public tude of the financing gap that these countries budgets are squeezed, external capital flows face relative to the size of their economies, as decline, and consumer ability to pay user well as the importance of infrastructure in charges is eroded. The ability to construct regenerating their development, a case exists new infrastructure, address regional bottle- for channeling incremental donor resources in necks, and maintain existing assets would be their direction. severely reduced. In Latin America during the Some of Africa's larger low-income coun- 1990s, some 50 percent of the fiscal compres- tries have the potential to raise a significant sion to balance the public books came from amount of local finance for infrastructure if cuts in infrastructure spending. In Indonesia suitable instruments can be developed. In a following the Asian crisis, public investment in handful of African countries, domestic capi- infrastructure fell from 7 percent of GDP to tal markets are beginning to look wide and 26 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION deep enough to provide significant volumes of Calderón, César. 2008. "Infrastructure and infrastructure finance, Nigeria being the most Growth in Africa." Working Paper 3, Africa Infrastructure Country Diagnostic, World salient example (Irving and Manroth 2009). Bank, Washington, DC. However, most of this finance takes the form Carruthers, Robin, Ranga R. Krishnamani, and of relatively short-maturity commercial bank Siobhan Murray. 2008. "Improving Connectiv- lending, often not the best suited for infra- ity: Investing in Transport Infrastructure in Sub- structure projects. A need exists to further Saharan Africa." Background Paper 7, Africa develop corporate bond markets and to create Infrastructure Country Diagnostic, World Bank, regulatory conditions for greater participation Washington, DC. by institutional investors in funding infra- Commission for Africa. 2005. Our Common Inter- structure investments. est: Report of the Commission for Africa. London: Commission for Africa. Note Dorosh, Paul, Hyoung-Gun Wang, Liang You, and Emily Schmidt. 2008. "Crop Production The authors of this chapter are Vivien Foster and Road Connectivity in Sub-Saharan Africa: and Cecilia Briceño-Garmendia. A Spatial Analysis." Working Paper 19, Africa Infrastructure Country Diagnostic, World Bank, References Washington, DC. Africon. 2008. "Unit Costs of Infrastructure Proj- Eberhard, Anton. 2007. "Matching Regulatory ects in Sub-Saharan Africa." Background Paper Design to Country Circumstances: The Poten- 11, Africa Infrastructure Sector Diagnostic, tial of Hybrid and Transitional Models." World Bank, Washington, DC. Gridlines, Note 23 (May), Public-Private Banerjee, Sudeshna, Heather Skilling, Vivien Foster, Infrastructure Advisory Facility, World Bank, Cecilia Briceño-Garmendia, Elvira Morella, and Washington, DC. Tarik Chfadi. 2008. "Ebbing Water, Surging Defi- Eberhard, Anton, Vivien Foster, Cecilia Briceño- cits: Urban Water Supply in Sub-Saharan Africa." Garmendia, Fatimata Ouedraogo, Daniel Background Paper 12, Africa Infrastructure Camos, and Maria Shkaratan. 2008. "Under- Sector Diagnostic, World Bank, Washington, DC. powered: The State of the Power Sector in Sub- Banerjee, Sudeshna, Quentin Wodon, Amadou Saharan Africa." Background Paper 6, Africa Diallo, Taras Pushak, Hellal Uddin, Clarence Infrastructure Sector Diagnostic, World Bank, Tsimpo, and Vivien Foster. 2008. "Access, Afford- Washington, DC. ability, and Alternatives: Modern Infrastructure Escribano, Alvaro, J. Luis Guasch, and Jorge Pena. Services in Africa." Background Paper 2, Africa 2008. "Impact of Infrastructure Constraints on Infrastructure Sector Diagnostic, World Bank, Firm Productivity in Africa." Working Paper 9, Washington, DC. Africa Infrastructure Sector Diagnostic, World Bofinger, Heinrich C. 2009. "Air Transport: Chal- Bank, Washington, DC. lenges to Growth." Background Paper 16, Africa Estache, Antonio. 2005. "What Do We Know Infrastructure Sector Diagnostic, World Bank, about Sub-Saharan Africa's Infrastructure and Washington, DC. the Impact of Its 1990 Reforms?" World Bank, Briceño-Garmendia, Cecilia, and Vivien Foster. Washington, DC. 2009a. "Democratic Republic of Congo: Pri- Fay, Marianne, Danny Leipziger, Quentin Wodon, oritizing Infrastructure Investments--a Spatial and Tito Yepes. 2005. "Achieving Child-Health- Approach." World Bank, Washington, DC. Related Millennium Development Goals: The ------. 2009b. "Republic of Congo: Prioritiz- Role of Infrastructure." World Development 33 ing Infrastructure Investments--a Spatial (8): 1267­84. Approach." World Bank, Washington, DC. Grey, David, and Claudia Sadoff. 2006. "Water for Briceño-Garmendia, Cecilia, Karlis Smits, and Growth and Development: A Framework for Vivien Foster. 2008. "Financing Public Infra- Analysis." Theme Document of the 4th World structure in Sub-Saharan Africa: Patterns, Issues, Water Forum, Mexico City, March. and Options." AICD Background Paper 15, Gwilliam, Ken, Vivien Foster, Rodrigo Archondo- Africa Infrastructure Sector Diagnostic, World Callao, Cecilia Briceño-Garmendia, Alberto Bank, Washington, DC. Nogales, and Kavita Sethi. 2008. "The Burden Bullock, Richard. 2009. "Taking Stock of Railway of Maintenance: Roads in Sub-Saharan Companies in Sub-Saharan Africa." Background Africa." Background Paper 14, Africa Paper 17, Africa Infrastructure Sector Diagnos- Infrastructure Sector Diagnostic, tic, World Bank, Washington, DC. World Bank, Washington, DC. Africa's Infrastructure: A Time for Transformation 27 Irving, Jacqueline, and Astrid Manroth. 2009. Background Paper 4, Africa Infrastructure Sector "Local Sources of Financing for Infrastructure Diagnostic, World Bank, Washington, DC. in Africa: A Cross-Country Analysis." Policy Teravaninthorn, Supee, and Gael Raballand. 2008. Research Working Paper 4878, World Bank, "Transport Prices and Costs in Africa: A Review Washington, DC. of the Main International Corridors." Working Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim Paper 14, Africa Infrastructure Sector Diagnos- Kelly, Richard Green, and Alvaro Federico Barra. tic, World Bank, Washington, DC. 2008. "Costing the Needs for Investment in ICT Torero, Maximo, and Javier Escobal. 2005. "Mea- Infrastructure in Africa." Background Paper 3, suring the Impact of Asset Complementarities: Africa Infrastructure Sector Diagnostic, World The Case of Rural Peru." Cuadernos de Economia Bank, Washington, DC. 42 (May): 137­64. Minges, Michael, Cecilia Briceño-Garmendia, Vagliasindi, Maria, and John Nellis. 2009. "Evalu- Mark Williams, Mavis Ampah, Daniel Camos, ating Africa's Experience with Institutional and Maria Shkratan. 2008. "Information and Reform for the Infrastructure Sectors." Working Communications Technology in Sub-Saharan Paper 23, Africa Infrastructure Country Diag- Africa: A Sector Review." Background Paper 10, nostic, World Bank, Washington, DC. Africa Infrastructure Sector Diagnostic, World Bank, Washington, DC. Van den Berg, Caroline. 2009. "Public Expenditure Review in the Water Sector: The Case of Tan- Morella, Elvira, Vivien Foster, and Sudeshna Ghosh zania." Paper presented at Water Week, World Banerjee. 2008. "Climbing the Ladder: The State Bank, Washington, DC, February 17­20. of Sanitation in Sub-Saharan Africa." Back- ground Paper 13, Africa Infrastructure Country Wodon, Quentin, ed. 2008a. "Electricity Tariffs Diagnostic, World Bank, Washington, DC. and the Poor: Case Studies from Sub-Saharan Africa." Working Paper 11, Africa Infrastructure Mundy, Michael, and Andrew Penfold. 2008. Country Diagnostic, World Bank, Washington, "Beyond the Bottlenecks: Ports in Sub-Saharan DC. Africa." Background Paper 8, Africa Infrastructure Sector Diagnostic, World Bank, Washington, DC. ------. 2008b. "Water Tariffs and the Poor: Case Studies from Sub-Saharan Africa." Working ODI (Overseas Development Institute). 2009. A Paper 12, Africa Infrastructure Country Diag- Development Charter for the G-20. London: ODI. nostic, World Bank, Washington, DC. Prevost, Christophe. 2009. "Benin Rural Water Pub- Yepes, Tito. 2007. "New Estimates of Infrastructure lic Expenditure Review: Findings, Impacts and Expenditure Requirements." World Bank, Wash- Lesson Learned." Paper presented at Water Week, ington, DC. World Bank, Washington, DC, February 17­20. Yepes, Tito, Justin Pierce, and Vivien Foster. Rosnes, Orvika, and Haakon Vennemo. 2008. 2008. "Making Sense of Sub-Saharan Africa's "Powering Up: Costing Power Infrastructure Infrastructure Endowment: A Benchmarking Investment Needs in Southern and Eastern Approach." Working Paper 1, Africa Infrastruc- Africa." Background Paper 5, Africa Infrastruc- ture Country Diagnostic, World Bank, Washing- ture Sector Diagnostic, World Bank, Washing- ton, DC. ton, DC. You, Liang Zhi. 2008. "Irrigation Investment Needs Svendsen, Mark, Mandy Ewing, and Siwa Msangi. in Sub-Saharan Africa." Background Paper 9, 2008. "Watermarks: Indicators of Irrigation Africa Infrastructure Country Diagnostic, World Sector Performance in Sub-Saharan Africa." Bank, Washington, DC. Par t 1 The Overall Story Introduction The Africa Infrastructure Country Diagnostic I n 2005, the Commission for Africa drew the 1990s, many donors shifted their priorities public attention to the magnitude and to social interventions focused on poverty alle- urgency of Africa's development challenges viation, overlooking the central importance of and sounded a new appeal to the international economic growth as an engine of poverty reduc- community to meet them. In its landmark tion. Moreover, private capital flows in the early report, Our Common Interest, the commission 2000s were weak in the aftermath of the Asian underscored infrastructure as one of the con- crisis. The commission's report stated that tinent's central challenges: despite its clear benefits, African govern- Infrastructure is a key component of the ments and their development partners sharply investment climate, reducing the costs of reduced, over the 1990s, the share of resources doing business and enabling people to access allocated to infrastructure--reflecting its lower markets; is crucial to advances in agriculture; is priority in policy discussions. In retrospect, a key enabler of trade and integration, impor- this was a serious policy mistake, driven by the tant for offsetting the impact of geographical international community, which undermined dislocation and sovereign fragmentation, and growth prospects and generated a substantial critical to enabling Africa to break into world backlog of investment--a backlog that will markets; and is fundamental to human devel- take strong action, over an extended period opment, including the delivery of health and of time, to overcome. (Commission for Africa education services to poor people. Infrastruc- 2005: chap. 7, para. 63, citations omitted) ture investments also represent an important The report estimated Sub-Saharan Africa's untapped potential for the creation of pro- infrastructure financing needs to be $39 bil- ductive employment. (Commission for Africa 2005: chap. 7, para. 61, citations omitted) lion per year, divided almost equally between capital expenditure ($22 billion) and spending In the years preceding the commission's on operation and maintenance ($17 billion). report, external capital flows for African infra- On this basis, it recommended a doubling of structure had reached a historic low. During infrastructure spending in the region, to be 31 32 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION supported by increased donor allocations of A steering committee chaired by the Afri- $10 billion up to 2010. can Union Commission was formed to over- Soon after the publication of the commis- see the AICD project. The committee included sion's report, the Group of Eight summit at representatives from the African Development Gleneagles expressed a firm political commit- Bank, the New Partnership for Africa's Devel- ment to scale up donor financing for African opment, and the regional economic com- infrastructure, which led to the formation of munities (Common Market for Eastern and the Infrastructure Consortium for Africa. The Southern Africa, Economic Community of consortium became a forum where major Central African States, Economic Community bilateral and multilateral donors could work of West African States East African Commu- with continental and regional institutions to nity, Economic Community of West African spearhead economic integration and maintain States, and Southern African Development the momentum behind the political commit- Community). Agence Française de Développe- ment of Gleneagles. ment, the U.K.'s Department for International Development, the European Commission, Germany's Kreditanstalt für Wiederaufbau, Genesis of the Project the Public-Private Infrastructure Advisory Facility, and the World Bank pledged resources From its inception, the consortium recognized to the project. The implementation of AICD that the paucity of information and analysis was delegated to the Africa Vice Presidency on African infrastructure severely constrained of the World Bank. The steering committee scaling up. Even the most elementary data--on also convened a technical advisory panel of quantity and quality of infrastructure stocks, academics from around the world to provide access to services, prices and costs, efficiency independent review of the studies. parameters, historic spending, and future Technical work on the project began in investment needs--were either nonexistent mid-2006 and proceeded in three stages. The or limited in coverage. Most standard global first stage, running from mid-2006 to mid- databases on infrastructure covered barely a 2007, was devoted to primary data collection handful of African countries. at the country level, and it produced a suite of A stocktaking paper concluded that the data new databases on African infrastructure. The situation seriously impeded the region's abil- second stage, from mid-2007 to mid-2008, ity to interpret and understand the state of its focused on data analysis. It led to the pro- infrastructure. It asserted that "[w]e don't know duction of a number of background papers precisely how well Sub-Saharan Africa is meet- analyzing key aspects of infrastructure at the ing its infrastructure needs, because the quality continental level (see table I.1). The third stage, and quantity of the data has become so poor. from mid-2008 to mid-2009, involved consul- Improving Africa's ability to monitor and bench- tation and outreach on preliminary findings mark its performance should be a top priority and focused on producing this report. for the international community and is likely to For the purposes of the diagnostic, infra- be a major challenge requiring significant coor- structure is defined to include all the main dination across countries and donors" (Estache networks, those associated with information 2005: executive summary, 1). The consortium and communication technologies (ICTs), irri- concluded that, without such information, gation, power, sanitation, water, and transport evaluating the success of past interventions, (including air, maritime, rail, and road). As prioritizing current allocations, and providing far as possible, the diagnostic aims to cover a benchmark to measure future progress would not only physical infrastructure but also the be difficult. Therefore, the consortium decided services it provides. The emphasis is on pub- to unite in a joint knowledge program, the Africa lic access infrastructure, so the study does not Infrastructure Country Diagnostic (AICD). The cover oil and gas pipelines or private port and goal of the AICD is to improve the knowledge rail infrastructure dedicated to the exclusive use base of the African infrastructure sectors. of particular mineral or industrial activities. The Africa Infrastructure Country Diagnostic 33 Neither does the diagnostic consider needs for Table I.1 AICD Background Papers water storage infrastructure required to protect Number Category and title Authors countries from droughts and floods beyond Cross-cutting topics those necessitated by particular downstream BP2 Access, Affordability, and Alternatives: Sudeshna Banerjee, Quentin Wodon, Amadou uses such as hydropower electricity generation, Modern Infrastructure Services in Africa Diallo, Taras Pushak, Helal Uddin, Clarence irrigation, and water supply. Tsimpo, and Vivien Foster The primary unit of analysis for the diag- BP11 Unit Costs of Infrastructure Projects in Willem van Zyl, Lynette Coetzer, and Chris Sub-Saharan Africa Lombard nostic is the country. The focus is on Sub- Saharan Africa, given the genesis of the proj- BP15 Financing Public Infrastructure in Cecilia Briceño-Garmendia, Karlis Smits, and Sub-Saharan Africa: Patterns, Issues, Vivien Foster ect as a response to the major infrastructure and Options deficits in that part of the continent. Owing Spending needs studies to budgetary and feasibility constraints, the BP3 Costing the Needs for Spending in ICT Rebecca Mayer, Ken Figueredo, Mike Jensen, diagnostic was originally limited to 24 of Infrastructure in Africa Tim Kelly, Richard Green, and Alvaro Federico the 48 countries in the Sub-Saharan region Barra (figure I.1). This Phase I sample covers almost BP5 Powering Up: Costing Power Orvika Rosnes and Haakon Vennemo Infrastructure Spending Needs all of the major countries, which together account for about 85 percent of the population BP7 Improving Connectivity: Investing Robin Carruthers, Ranga Rajan Krishnamani, in Transport Infrastructure in and Siobhan Murray and GDP of the region. They were carefully Sub-Saharan Africa selected to represent the economic, geographic, BP9 Irrigation Investment Needs in Liang Zhi You cultural, and political diversity that charac- Sub-Saharan Africa: A Matter of Scale terizes the region (figure I.2). Therefore, the State-of-the-sector reviews sample of 24 countries is statistically represen- BP1 Stuck in Traffic: Urban Transport in Africa Ajay Kumar and Fanny Barrett tative, providing an adequate basis for draw- BP4 Watermarks: Indicators of Irrigation Mark Svendsen, Mandy Ewing, and Siwa ing inferences about the overall infrastructure Sector Performance in Sub-Saharan Africa Msangi situation of Sub-Saharan Africa. BP6 Underpowered: The State of the Power Anton Eberhard, Vivien Foster, Cecilia Later, the project steering committee rec- Sector in Sub-Saharan Africa Briceño-Garmendia, Fatimata Ouedraogo, ommended extending the coverage of the Daniel Camos, and Maria Shkaratan diagnostic to as many of the remaining Afri- BP8 Beyond the Bottlenecks: Ports in Michael Mundy and Andrew Penfold Sub-Saharan Africa can countries as possible. Following further fund-raising, Phase II of the project was ini- BP10 Information and Communications Michael Minges, Cecilia Briceño-Garmendia, Technology in Sub-Saharan Africa: Mark Williams, Mavis Ampah, Daniel Camos, tiated in mid-2008. It incorporates 16 more A Sector Review and Maria Shkratan countries, raising the total to 40. Although BP12 Ebbing Water, Surging Deficits: Urban Sudeshna Banerjee, Heather Skilling, Vivien the focus remains on Sub-Saharan Africa, Water Supply in Sub-Saharan Africa Foster, Cecilia Briceño-Garmendia, Elvira Phase II includes greater coverage of North Morella, and Tarik Chfadi African countries in a number of areas to com- BP13 Climbing the Ladder: The State of Elvira Morella, Vivien Foster, and Sudeshna Sanitation in Sub-Saharan Africa Ghosh Banerjee plete the African picture and provide a point of BP14 The Burden of Maintenance: Roads in Ken Gwilliam, Vivien Foster, Rodrigo comparison with the Sub-Saharan region. Sub-Saharan Africa Archondo-Callao, Cecilia Briceño-Garmendia, Alberto Nogales, and Kavita Sethi BP16 Air Transport: Challenges to Growth Heinrich C. Bofinger Scope of the Project BP17 Taking Stock of Railway Companies in Richard Bullock Sub-Saharan Africa The results of Phase II were not available at the time of writing, so the results presented in this volume are based on the analysis of the 24 Phase I countries. However, all financial aggregates · The spending needs pillar estimates the cost in this report have been scaled up to cover the of future infrastructure requirements. whole of Sub-Saharan Africa. Financial estimates · The fiscal costs pillar documents existing were scaled to reflect the weight of the 24 sample patterns of infrastructure spending. countries in the overall GDP of the region. · The sector performance pillar clarifies the The country-level analysis has three pillars, scope for improvement in efficiency as well each of which is described below: as structural and policy reforms. 34 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure I.1 Country Coverage of the Africa Infrastructure Country Diagnostic of Africa's spending needs, and most of them had limited country coverage. The best-known TUNISIA global cross-country studies estimate spending MOROCCO needs using econometric techniques and mac- ALGERIA LIBYA ARAB REP. roeconomic panel data (Estache 2005; Fay and FORMER SPANISH OF EGYPT Yepes 2003; Yepes 2007). These studies iden- SAHARA tify historical relationships between GDP and CAPE VERDE MAURITANIA MALI NIGER physical infrastructure stocks to predict infra- ERITREA THE GAMBIA SENEGAL BURKINA CHAD SUDAN structure requirements given current growth GUINEA-BISSAU DJIBOUTI GUINEA FASO BENIN NIGERIA forecasts. Unit costs of infrastructure are then CÔTE ETHIOPIA SIERRA LEONE D'IVOIRE GHANA CENTRAL AFRICAN REPUBLIC used to convert these predictions into financial LIBERIA CAMEROON SOMALIA TOGO EQUATORIAL GUINEA estimates. These types of studies provide inter- UGANDA SÃO TOMÉ AND PRÍNCIPE GABON CONGO DEM. REP. KENYA nationally consistent, first-order approxima- RWANDA OF CONGO BURUNDI tions of investment needs. They are likely to TANZANIA SEYCHELLES underestimate requirements, however, because COMOROS they tend to focus on infrastructure quantity ANGOLA MALAWI ZAMBIA rather than quality; do not take into account Phase I Phase II repressed demand and social targets; and ZIMBABWE MOZAMBIQUE MADAGASCAR MAURITIUS Phase II (partial) NAMIBIA use single, global (as opposed to country- BOTSWANA specific), unit-cost parameters based on effi- SWAZILAND cient implementation. This map was produced by the Map Design Unit of The World Bank. SOUTH The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment AFRICA LESOTHO Country-specific or sector-specific engi- on the legal status of any territory, or any endorsement or acceptance of such boundaries. neering cost studies existed for particular infrastructure packages: the West Africa Power Figure I.2 Representativeness of Phase I Sampled Countries Pool Master Plan, for example, and the African Development Bank's study of the Trans-Africa 100 Highway Network, in addition to various 90 country or regional master plans. These stud- 80 ies tend to be accurate and internalize policy- 70 defined targets, but they have a number of dis- % of countries 60 advantages. They are costly to produce, are not 50 available for all countries and sectors, and tend 40 to adopt a wide variety of methodologies that 30 limit their comparability across countries. 20 The AICD project studied spending needs 10 in five sectors: ICT, irrigation, power, trans- 0 port, and water and sanitation. The objective of the studies was to develop a simple but robust e e ne idd ome e e ed ed d on on m om cte ho rn ck r m inco ph ph inc inc country-based microeconomic methodology ve ffe op dlo glo o go nc lus ta low le le lan an idd fra ly lic that would be significantly more accurate than or nf m po co er pe the "top-down" macro studies, yet substantially low up more straightforward and standardized than the "bottom-up" engineering studies. The meth- Sub-Saharan Africa Africa Infrastructure Country Diagnostic odology aims to capture both market-driven investments to keep pace with the demands Note: AICD = Africa Infrastructure Country Diagnostic. generated by a growing economy and politically determined investment targets to meet social Estimating Future Spending Needs needs that may not be commercially lucrative At the outset of the AICD project, only a without government subsidy. As important as small number of standardized cross-country estimating the magnitude of investment needs, studies focused on estimating the magnitude the models calculate spending requirements for The Africa Infrastructure Country Diagnostic 35 rehabilitation of existing infrastructure assets information systems (GIS) tools. Creation as well as maintenance needs to sustain opera- of an African GIS database collated from tional (existing and new) assets. diverse sources and permitting the overlay The goal was not so much to produce an of geophysical, agroecological, demographic, estimate as to create a model that would allow and economic features with infrastructure exploration of spending needs under a vari- networks made this approach possible (see ety of different assumptions about economic box I.1). The input parameters needed to run growth, social objectives, unit costs, and other the spending needs models could be derived relevant parameters. Projections were based on largely from an extensive desk review of avail- World Bank GDP growth projections for the able information. next decade and United Nations demographic Although efforts were made to develop forecasts. methodologies that were consistent across In most cases, no clear methodological sectors, the specifics of each sector raised par- precedents existed for producing country- ticular challenges that called for some adapta- level estimates of spending needs based on tion. In all cases, spending needs include new this kind of microeconomic modeling. A investment, rehabilitation of existing assets, technique adopted across many of these stud- and operation and maintenance associated ies was spatial modeling using geographic with new and existing assets. BOX I.1 The AICD Geographic Information Systems Platform for Africa Early in the AICD process, it became appar- The GIS platform was assembled from a wide ent that geographic information systems (GIS) variety of sources. Public domain data, available would be a key input to many aspects of infra- from the World Bank and other organizations, structure analysis. A decision was therefore such as the Food and Agriculture Organization, made to assemble all available geographic the International Food Policy Research Institute, databases of relevance to the African infra- the International Union for Conservation of structure sectors into a single GIS platform. Nature, the U.S. Geological Survey, the Center The platform includes data sets from dif- for International Earth Science Information Net- ferent scales, levels of detail, reference years, work, the U.S. National Geospatial-Intelligence and coding schemes. In all, more than 20 sep- Agency, and Oak Ridge National Laboratory, arate thematic layers of geographic informa- were a primary resource. In some cases, gov- tion cover each of the following topics. ernment transport or other agencies provided data. Databases were also purchased from · Infrastructure networks: power stations, private sector sources or constructed from pri- transmission lines, dam sites, irrigated mary country data collected as part of the AICD areas, roads (including type, condition, and project. Where possible, an effort was made to traffic), railways, ports, airports, submarine update existing data sets with information on cables, fiber-optic backbones, and global condition, status, or other characteristics, based systems mobile (GSM) signal coverage on expert assessment and other sources. · Physiographic features: topography, meteo- The AICD GIS platform is publicly avail- rology, watercourses, river basin boundaries, able on the project Web site, http://www soil type, land coverage, and agricultural .infrastructureafrica.org, where users may con- usage and potential sult preassembled infrastructure atlases for each · Socioeconomic features: cities, population country, regional economic community, and densities, mines, oil fields, poverty indica- the continent; make use of the GIS tool to cre- tors, travel time to nearest urban centers, ate their own customized maps; or download and household access to services. shape files for more technical GIS analysis. 36 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION statals were devoting resources to infrastructure · For ICTs, the spatial analysis was used to estimate the costs, revenues, and hence development and maintenance. The Interna- financial viability of rolling out services to tional Monetary Fund's Government Financial remote rural communities. Statistics reports central government budget- ary spending for a number of broadly defined · For irrigation, the financial viability of irri- infrastructure sectors but does not include gating crops at different locations was pre- the expenditures of state-owned enterprises screened as suitable for large- or small-scale and special nonbudgetary funds dedicated to irrigation development based on proximity infrastructure, both of which are highly sig- to large dams in one case and the road net- nificant for the sector. Moreover, it does not work in the other. break down expenditures according to specific · For transport, the spatial analysis was used infrastructure sectors and functional outlays, to measure the extent of the road network such as capital or maintenance and operat- needed to meet a set of regional, national, ing expenditures. Most important, even the urban, and rural connectivity standards. very limited data recorded through the Fund's Linking these directly to economic objec- Government Financial Statistics were available tives did not prove feasible. for only a handful of African countries. · For power, the model is based on a least-cost Using the limited information available at optimization model that selects the most the time, researchers made some first-order cost-effective expansion path for national estimates of Africa's public expenditure on or regional power sector development to infrastructure (see Estache 2005; Estache, meet a given projection of demand. Gonzalez, and Trujillo 2007). Notwithstand- · For water and sanitation, the model builds ing numerous caveats regarding the quality upon existing work (Mehta, Fugelsnes, and and coverage of the public finance data, the Virjee 2005; Water and Sanitation Program overall picture that emerged showed alloca- 2006) and uses demographic growth trends tion of a declining share of public budgets to to analyze the number of new connections infrastructure from 1980 to 2000. needed to meet the Millennium Development Without a detailed understanding of expen- Goals (WHO and UNICEF 2006) under a diture patterns of key public institutions-- variety of different technological choices. central governments and state-owned enterprises--pinpointing the magnitude and The results of the AICD spending needs nature of the region's infrastructure financing studies are presented in chapter 1 and further gap or assessing the efficiency and effectiveness detailed in the corresponding sectoral chapters of public spending is difficult. To overcome in part 2. Detailed background papers also doc- these limitations, the AICD project built a new ument the methodology and results for each database of standardized cross-country data sector in much greater detail (see table I.1). that seeks to give a detailed yet comprehensive All of the spending needs models developed picture of public infrastructure spending, both for the project are available online at the proj- within and beyond the bounds of central gov- ect Web site. The Web-based versions allow ernment budgets. Data collection was based users to apply sensitivity of spending needs on a standardized methodology and covers, as to varying assumptions over a wide range of far as possible, the period 2001­06. To ensure input parameters for specific countries. The the cross-country comparability of the data, results are displayed both numerically and spa- a detailed methodology including templates tially in the form of maps. guided data collection in the field and back- office processing and documentation (Briceño- Documenting Existing Spending Garmendia 2007). Patterns The methodology is designed to be compre- At the outset of the AICD project, almost no hensive insofar as it covers all relevant budget- information was available about the extent to ary and nonbudgetary areas of infrastructure which African governments and their para- spending. The collection of data on spending The Africa Infrastructure Country Diagnostic 37 was grounded in an overview of the institu- distinguish to some extent between current tional framework for delivering infrastructure expenditures, capital expenditures, and vari- services in each of the countries while aiming ous subcategories. to identify all of the channels through which Much of the necessary data could be lifted public resources go into infrastructure. The directly from the budget documents and work began with a detailed review of the cen- financial statements of the relevant parastatals, tral government budget. Thereafter, financial although in many cases, careful recoding of the statements were collected from all the para- data was necessary to align them with the proj- statals and special funds that had been identi- ect template. Local consultants undertook field- fied in the institutional review. work that was coordinated centrally to ensure In countries where infrastructure service quality control and data consistency. The focus providers are highly decentralized (as in munici- of data collection was on executed expendi- pal water utilities), financial statements could be tures, but wherever possible, the budgeted and collected from only the three largest provid- released expenditures were also collected. ers. Privatized infrastructure service provid- The targeted period for data collection was ers were included if a majority of their shares 2001­06, although a complete time series was remained government owned or if they con- not always available. All financial data are pre- tinued to depend on the state for capital or sented as annual averages over the period, to operating subsidies. Thus, telecommunication smooth out annual variations and maximize incumbents were typically included, whereas available data points. All data were denomi- mobile operators were not. nated in local currency and centrally normal- In some countries, local governments have ized using exchange rate, GDP, and population begun to play an increasingly prominent role data taken from the World Development Indi- in infrastructure service provision, but com- cators database of the World Bank. prehensive expenditure data at the local gov- Public expenditure data were complemented ernment level could not always be collected. by financing data from secondary sources to In some cases, however, the central govern- provide a comprehensive view of financial ment produces consolidated local government flows to African infrastructure and the rela- accounts. Otherwise, an alternative source of tive importance of the different players. These information was the fiscal transfers from cen- secondary sources included the World Bank's tral to local governments reported in the budget Private Participation in Infrastructure data- and on which local governments relied, given base, which documents trends in private capital limited alternative sources of revenue. In some flows; the Development Assistance Committee cases, transfers are earmarked for infrastruc- database of the Organisation for Economic ture spending; in others, the share allocated to Co-operation and Development (OECD), cov- infrastructure could only be estimated. ering external financial support from bilateral Data were collected to permit both classi- and multilateral OECD donors; and a new fication and cross-classification by economic database on non-OECD finance for African and functional categories. That is, a matrix was infrastructure (Foster and others 2008). To established so that spending on each functional make these financial flows methodologically category could be decomposed according to the consistent with those for public expenditure, economic nature of the expense and vice versa. researchers converted commitments made by Functional classification followed as closely as external financiers into disbursements using possible the four-digit category or class level typical disbursement profiles for infrastruc- of the functional classification (COFOG) pro- ture projects. Every effort was made to avoid posed in the International Monetary Fund's double counting between public expenditure Government Financial Statistics Manual 2001 and external finance. (IMF 2001), making possible identification of The results of the public expenditure analy- all major infrastructure subsectors. The eco- sis provide the foundation for chapter 2 of this nomic classification of expenses also followed report, but they are reported in much greater the Fund's framework, making it possible to detail in Background Paper 15 (see table I.1). 38 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION At this time, one can say that the level of pub- provide a snapshot of the situation prevailing lic expenditure on infrastructure in Africa is in 2006 at the time of data collection. A second substantially higher than previously thought-- block of quantitative indicators documents the and certainly several times higher than earlier operational, technical, and financial aspects of estimates. The resulting public expenditure sector performance, with particular focus on database is now available to the public via the infrastructure service providers such as utilities. project Web site and can be downloaded by users Wherever possible, the quantitative data cover for a variety of purposes. The database con- the period 2001 to 2006, and the most recent tains detailed information about expenditure available year is the one reported. patterns by institution, sector, and functional For each sector, manuals were developed category. to guide the data collection for the indicators. The analysis of public spending patterns The manuals map the rationale and conceptual was complemented by work on unit costs of structure of the data collection, provide detailed infrastructure projects and included a review definitions of the indicators, lay out question- of the costs and cost structures associated with naire formats to assist in eliciting information, a sample of donor-funded projects covering and map a database structure for coding of the roads, power, and water supply. The typical data. Such detailed manuals were designed to outputs of these projects were standardized to guide consultants responsible for data collection permit the creation of standardized unit costs. and to ensure comparability of indicators across Data were collected from the bills of quantities countries and ultimately over time, should the for the public works contracts of these projects process be repeated. and entered into the standardized template. For some sectors (power, railways, roads, The overall sample included 115 road projects, water and sanitation), the indicators could 144 water projects, and 58 power projects. The be collected only through detailed in-country resulting database of unit costs illustrates the fieldwork. For a number of other sectors (air dispersion that can be experienced in donor- transport, ICT, irrigation, ports), the data funded infrastructure projects depending on a could be collected remotely through the arm's- range of factors. length administration of questionnaires with telephone follow-up and the compilation of Understanding Sector Performance data from existing publications and sources. At the beginning of the project, relatively little The data collection involved contacting several systematic, comprehensive, and empirically hundred infrastructure institutions around grounded literature was available on the per- Africa, including more than 16 rail operators, formance of the five infrastructure sectors. 20 road entities, 30 power utilities, 30 ports, To develop a comprehensive and detailed 60 airports, 80 water utilities, and 100 ICT portrait of the infrastructure sectors in Sub- operators, as well as the relevant line ministries Saharan Africa, the AICD developed a set of in all of the countries. standardized performance indicators covering The data collection focused on the compi- both the consumer and the service provider lation of existing information available from perspectives. These indicators are collected the target institutions through their annual for the full range of infrastructure subsec- reports, internal databases, and knowledge tors, including air transport, ICT, irrigation, of their managers. Thus, the coverage of the ports, power, railways, roads, and water and databases reflects the state of self-knowledge of sanitation. In each case, a common conceptual the institutions. The project did not have the structure was adopted. resources to undertake primary survey work to A first block of qualitative indicators was gen- obtain data on missing indicators. erated through a substantial questionnaire that The resulting data were centralized, and two documents the details of the legal, institutional, forms of quality control were conducted. The and regulatory framework, which are summa- first was a review by specialists knowledgeable rized in a series of specially developed indexes about the countries in question. The second (see chapter 4 in this volume). Qualitative data, consisted of logic and consistency checks on The Africa Infrastructure Country Diagnostic 39 the database as a whole by examining data pat- and up-to-date information on the status of terns and outliers. ICT in Africa not available from any other The survey of infrastructure service provid- single source. ers was complemented by work on patterns of In transportation, the Sub-Saharan Africa household access to and expenditure on infra- Transport Policy Program (SSATP) has played structure services aimed to integrate all exist- an important role in developing the knowl- ing household surveys conducted in Africa edge base through an abundant literature of from 1990 to 2005. These sources included case studies and policy reports. The SSATP 67 demographic and health surveys (DHSs) has made some efforts to move toward a set of and multi-indicator cluster surveys contain- quantitative indicators for the transport sector, ing detailed information on household access although these remain limited in scope. The patterns, and 30 budget surveys containing program has played a leading role in develop- detailed information on household expendi- ing road sector modeling tools, most notably ture patterns. Data from all of these surveys RONET, that enable road maintenance costs were standardized (based on a careful com- to be estimated based on a detailed physical parison of questionnaires) and integrated into specification of the road network. In addition, a single meta-database, making consistent there have been important contributions to the analysis possible of time trends within coun- understanding of institutional reform in the tries and diverging patterns across countries. road sector (Benmaamar 2006) and some work A standardized approach was used to group on the performance of African rail concessions. households socioeconomically according to At the outset of the AICD, however, no unified asset quintiles in the case of the DHSs and database existed on road type, condition, and expenditure quintiles in the case of the budget traffic. These data were collected on a georefer- surveys. The meta-database covers 39 countries enced, link-by-link basis that allows the infor- in Africa; time trends are available for 23 of the mation to be presented graphically in a map countries. and that underpins detailed financial analysis The main source of telecommunications of the road network using the RONET model. data for Africa is the International Telecom- At the outset of the project, relatively little munication Union, which compiles time- continent-wide analysis of the African water series data on a number of indicators and utilities existed (Estache and Gassner 2004b). publishes information on telecommunications The starting point for water utility data col- regulation. In addition, a number of one-off lection was the databases developed by the reports had been written on the African tele- Water Utilities Partnership and the Interna- communication sector. These reports quickly tional Benchmarking Network (IBNET). Both become outdated and are often limited to cer- sources were sparse in their country coverage tain groups of countries. The AICD project and focused primarily on utility operational has improved the timeliness, detail, and scope performance without covering the institutional of these data sets, including compiling more framework in any great depth. Both initia- recent data than are available from intergov- tives informed the development of indicators ernmental sources, verifying the accuracy of under the AICD, which aimed to be consistent existing information, widening and completing with them in areas of overlap. The data col- coverage for all African countries, and enhanc- lection process was coordinated with IBNET ing data to incorporate more detailed and spe- to increase African country coverage for both cific indicators for tariffs, regulations, market projects. Generally speaking, the AICD opera- structure, and the user's perspective, among tional and financial performance indicators others. The project has also structured the data are a subset of those collected by IBNET, but into analytical categories and compiled several the qualitative indicators and tariff schedules indexes to facilitate comprehension of the vast collected through AICD go much further than amount of data. In summary, the information anything done before. Five modules of qualita- in the AICD data set for ICT provides a struc- tive data were collected for each country, cover- tured framework of comprehensive, inclusive, ing the institutional and regulatory framework 40 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION for water provision, governance arrangements indicators are now available to the public on the for specific water utilities, the status of the sani- project Web site, http://www.infrastructureafrica tation sector, the status of the rural water sector .org, and through the Development Data Plat- in each country, and the prevalence and charac- form of the World Bank. Containing detailed teristics of small-scale service providers in the information about institutional, operational, largest city in each country. The quantitative technical, and financial indicators relating to indicators aim to capture the operational and each of the sectors covered, the database can be financial performance of utilities from 2001 to downloaded for a variety of purposes. 2006, together with their tariff schedules. In The work on these three pillars and cross- countries where service provision is decentral- cutting issues resulted in the creation of 17 ized, comprehensive data could not be captured original background papers on which this on all utilities, but efforts were made to cover "Flagship Report" is based (table I.1). The the three largest utilities in each country. main findings that follow in this report refer In the case of irrigation, limited data were to these background papers. Readers seeking available at the country level. Sub-Saharan further technical details on any of these issues Africa has little experience with irrigated can find these papers through the project Web agriculture. Most performance indicators are site (http://www.infrastructureafrica.org). In limited to specific irrigation systems. The best due course, the background papers will be single source of data on comparable cross- repackaged as four sectoral volumes on ICT, country indicators was the global databases of power, transport, and water and sanitation, the Food and Agriculture Organization. They which will be technical companions to this were complemented where needed by data Flagship Report. from the World Bank and the International In addition to these three central pillars Food Policy Research Institute. of the data collection effort, more than 20 Power was undoubtedly the least docu- working papers were commissioned, cover- mented of Africa's infrastructure sectors at ing a range of ad hoc topics of relevance to the outset of the project (Estache and Gassner African infrastructure (table I.2). The top- 2004a). Some basic indicators on overall energy ics include linkages between infrastructure, balance and national power generation port- growth and fiscal sustainability, welfare effects folios were available from the International of infrastructure reforms, utility tariffs and Energy Agency and others, but coverage of subsidies, urban infrastructure services, local African countries remained quite limited, and private finance of infrastructure, impact of the available indicators did not provide any real inadequate power supply on firms, and the picture of power utility performance. Although role of small, independent suppliers of water. the Union of Producers, Transporters and Dis- The working papers are also available on the tributors of Electric Power in Africa (the Afri- project Web site. can power utilities association) has developed Beyond the initial data baseline estab- its own database of performance indicators, it lished here, the AICD project aims to estab- is not available to the general public. The Africa lish a sustainable basis for ongoing data Energy Commission is also developing a data- collection on Africa's infrastructure sectors. base of energy indicators for the continent, but This Flagship Report presents and analyzes it was not available in time for this project. the baseline information on the African The results of the various sector reviews infrastructure sectors collected because of provide the foundation for the corresponding this project. The long-term value of the effort sector chapters contained in part 2 (chapters depends on the sustainability of data collec- 7­17) of this report. In addition, the results of tion efforts to ensure that key infrastructure the household survey analysis are reported in trends on the continent can be tracked over chapter 3 on poverty and inequality, while the time and progress against this benchmark overall findings of the institutional analysis can be accurately measured. Plans are under are summarized in chapter 4 on institutions. way for the Statistical Department of the The resulting databases of sector performance African Development Bank to take over the The Africa Infrastructure Country Diagnostic 41 Table I.2 AICD Working Papers Number Title Author WP1 Making Sense of Sub-Saharan Africa's Infrastructure Endowment: A Benchmarking Approach Tito Yepes, Justin Pierce, and Vivien Foster WP2 Paying the Price for Unreliable Power Supplies: Own Generation of Electricity by Private Firms Vivien Foster and Jevgenijs Steinbuks in Africa WP3 Infrastructure and Growth in Africa César Calderón WP4 Electricity Reforms in Mali: A Micro-Macro Analysis of the Effects on Poverty and Distribution Dorothée Boccanfuso, Antonio Estache, and Luc Savard WP5 Electricity Reforms in Senegal: A Micro-Macro Analysis of the Effects on Poverty and Distribution Dorothée Boccanfuso, Antonio Estache, and Luc Savard WP6 Building Sector Concerns into Macro-Economic Financial Programming: Lessons from Antonio Estache and Rafael Muñoz Senegal and Uganda WP7 Cost Recovery, Equity, and Efficiency in Water Tariffs: Evidence from African Utilities Sudeshna Banerjee, Vivien Foster, Yvonne Ying, Heather Skilling, and Quentin Wodon WP8 Potential for Local Private Finance of Infrastructure in Africa Jacqueline Irving and Astrid Manroth WP9 Impact of Infrastructure Constraints on Firm Productivity in Africa Alvaro Escribano, J. Luis Guasch, and Jorge Pena WP10 A Tale of Three Cities: Understanding Differences in Provision of Modern Services Sumila Gulyani, Debabrata Talukdar, and Darby Jack WP11 Electricity Tariffs and the Poor: Case Studies from Sub-Saharan Africa Quentin Wodon WP12 Water Tariffs and the Poor: Case Studies from Sub-Saharan Africa Quentin Wodon WP13 Provision of Water to the Poor in Africa: Informal Water Markets and Experience with Water Sarah Keener, Manuel Luengo, and Sudeshna Banerjee Standposts WP14 Transport Prices and Costs in Africa: A Review of the Main International Corridors Supee Teravaninthorn and Gaël Raballand WP15 The Impact of Infrastructure Spending in Sub-Saharan Africa: A CGE Modeling Approach Jean-François Perrault and Luc Savard WP16 Water Reforms in Senegal: A Micro-Macro Analysis of the Effects on Poverty and Distribution Dorothée Boccanfuso, Antonio Estache, and Luc Savard WP17 Fiscal Costs of Infrastructure Provision: A Practitioner's Guide Cecilia Briceño-Garmendia WP18 Crop Production and Road Connectivity in Sub-Saharan Africa: A Spatial Analysis Paul Dorosh, Hyoung-Gun Wang, Liang You, and Emily Schmidt WP19 The Impact of the Yamassoukro Decision Charles Schlumberger WP20 Cost Recovery, Equity, and Efficiency in Power Tariffs: Evidence from African Utilities Cecilia Briceño-Garrmendia and Maria Shkaratan WP21 What Can We Learn from Household Surveys about Cooking Fuel Use in Daniel Camos Sub-Saharan Africa? WP22 Evaluating Africa's Experience with Institutional Reform for the Infrastructure Sectors Maria Vagliasindi and John Nellis long-term data collection effort based on the Briceño-Garmendia, Cecilia. 2007. "Fiscal Costs of methodological framework developed under Infrastructure Provision: A Practitioner's Guide." Working Paper 17, Africa Infrastructure Coun- the AICD project. The sponsors of the AICD try Diagnostic, World Bank, Washington, DC. project remain firmly committed to ensuring Commission for Africa. 2005. Our Common Inter- the sustainability of the data collection effort. est: Report of the Commission for Africa. London: Commission for Africa. Note Estache, Antonio. 2005. "What Do We Know The authors of this chapter are Vivien Foster about Sub-Saharan Africa's Infrastructure and the Impact of Its 1990 Reforms?" World Bank, and Cecilia Briceño-Garmendia. Washington, DC. Estache, Antonio, and Katharina Gassner. 2004a. References "The Electricity Sector of Sub-Saharan Africa: Benmaamar, Mustapha. 2006. "Financing of Road Basic Facts and Emerging Issues." World Bank, Maintenance in Sub-Saharan Africa: Reforms Washington, DC. and Progress towards Second Generation Road ------. 2004b. "Recent Economic Developments Funds." Discussion Paper 6, Road Management in the Water and Sanitation Sectors of Selected and Financing Series, Sub-Saharan Africa Sub-Saharan African Countries: Overview of Transport Policy Program, World Bank, Basic Facts and Emerging Issues." World Bank, Washington, DC. Washington, DC. 42 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Estache, Antonio, Marianela Gonzalez, and Mehta, Meera, Thomas Fugelsnes, and Kameel Lourdes Trujillo. 2007. "Government Virjee. 2005. "Financing the Millennium Devel- Expenditures on Health, Education and opment Goals for Water and Sanitation: What Infrastructure: A Naïve Look at Levels, Out- Will It Take?" International Journal of Water comes and Efficiency." Policy Research Working Resources Development 21 (2): 239­52. Paper 4219, World Bank, Washington, DC. Water and Sanitation Program. 2006. Getting Fay, Marianne, and Tito Yepes. 2003. "Investing Africa on Track to Meet the MDGs for Water and in Infrastructure: What Is Needed from 2000 Sanitation: A Status Overview of Sixteen African to 2010?" Policy Research Working Paper 3102, Countries. Joint Report of African Ministers World Bank, Washington, DC. http://ssrn.com/ Council on Water, African Development Bank, abstract=636464. European Union Water Initiative, and Water and Foster, Vivien, William Butterfield, Chuan Chen, Sanitation Program. Nairobi: Water and Sanita- and Nataliya Pushak. 2008. Building Bridges: tion Program­Africa, World Bank. China's Growing Role as Infrastructure Finan- WHO and UNICEF (World Health Organization cier for Sub-Saharan Africa. Trends and Policy and United Nations Children's Fund). 2006. Options no. 5. Washington, DC: Public-Private Meeting the MDG Drinking Water and Sanitation Infrastructure Advisory Facility, World Bank. Target: The Urban and Rural Challenge of the IMF (International Monetary Fund). 2001. Gov- Decade. Geneva: WHO and UNICEF. ernment Financial Statistics Manual 2001 Yepes, Tito. 2007. "New Estimates of Infrastructure (GFSM 2001). Washington, DC: IMF Statistics Expenditure Requirements." World Bank, Department. Washington, DC. Chapter 1 Meeting Africa's Infrastructure Needs I nfrastructure is central to Africa's develop- The estimated spending needs are $93 billion ment.1 Major improvements in information a year (15 percent of the region's GDP)--more and communication technology (ICT), for than twice the 2005 estimate by the Commission example, added as much as 1 percentage point for Africa.3 Total spending estimates divide fairly to Africa's per capita growth rate during the evenly among the middle-income countries, last decade, since the mid-1990s. However, the resource-rich countries, and low-income deficiencies in infrastructure are holding back nonfragile states (in the neighborhood of $28 the continent by at least 1 percentage point in billion­$30 billion a year), with low-income frag- per capita growth. In many countries, infra- ile states accounting for a smaller share of total structure limitations, particularly in power, needs (about $14 billion a year). The burden on depress productivity at least as much as red their economies varies dramatically per income tape, corruption, and lack of finance--the group, ranging from 10­12 percent of GDP for usual suspects in many people's minds when middle-income and resource-rich countries to they think of constraints on growth. 25 percent of GDP for low-income nonfragile In density of paved roads, capacity to gener- states and 36 percent for fragile states. The total ate power, and coverage of telephone main lines, cost splits two to one between capital investment both low-income and middle-income African and operation and maintenance expenses. countries lag behind their peers elsewhere in Over 40 percent of the expenditure needed the developing world.2 A few decades ago, in the is in the power sector, which must install 1960s to 1980s, Africa's infrastructure endow- 7,000 megawatts of new generation capac- ments were similar to those in East and South ity each year just to keep pace with demand. Asia, but those regions have since expanded their Slightly more than 20 percent is associated with infrastructure stocks more rapidly, surpassing achievement of the Millennium Development Africa's position. Meeting Africa's infrastructure Goals (MDGs) for water supply and sanitation. needs and developing cost-effective modes of A further 20 percent of the spending require- infrastructure service delivery will entail a sub- ment is associated with the transport sector to stantial program of infrastructure investment. In achieve a reasonable level of regional, national, addition to building new infrastructure, existing rural, and urban connectivity and to maintain facilities must be rehabilitated and maintained. existing assets. 43 44 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Infrastructure: The Key to Africa's allow isolation of the first of these effects with Faster Growth some precision. The estimated effect of rais- ing Africa's infrastructure to some regional or African economies have grown at a solid 4 international benchmark shows considerable percent annual average in recent years. The consistency of 1 or 2 percentage points in per fastest growth has been in resource-rich capita growth. countries, which have benefited from rising A key question for policy makers is how commodity prices. In almost all cases, how- much infrastructure development contributes ever, that performance still falls short of the to growth relative to other policy parameters. 7 percent growth needed to achieve substan- One study finds that expanding and improving tial poverty reduction and attain the MDGs. infrastructure contributed almost 1 percent- Although infrastructure has contributed to age point to per capita economic growth from Africa's recent economic turnaround, it will 1990 to 2005, compared with only 0.8 percent- need to do even more to reach the continent's age point for macroeconomic stabilization and development targets. structural policies (Calderón 2008). Stabiliza- Inadequate infrastructure impedes faster tion policies include measures to control price growth in Africa. This view, highlighted by the inflation and rein in fiscal deficits, while struc- Commission for Africa (2005), is supported tural policies include measures to enhance by considerable economic research (table 1.1). human capital, increase financial depth, pro- Based on a cross-country econometric analysis mote trade openness, and improve governance. and a handful of country studies, the research Central Africa is the region where infrastruc- confirms a strong and significant connection ture improvements have made the largest between infrastructure stocks and economic contribution to recent growth, totaling 1.1 growth. Although the relationship undoubt- percentage points. Only in West Africa did the edly runs in both directions--infrastructure effect of macroeconomic policies on growth supporting growth and growth promoting exceed that of infrastructure. Over the same infrastructure--modern research techniques period, infrastructure in East Asia contributed Table 1.1 Links between Infrastructure and Growth in Africa: What the Research Says Study Method Scope Sector Conclusions Easterly and Levine 1997 Multicountry Africa Telecommunications, Infrastructure is strongly and significantly correlated with growth. power Esfahani and Ramirez 2003 Multicountry Africa Telecommunications, Africa's growth per capita would be 0.9 point higher with East power Asia's infrastructure. Calderón and Servén 2008 Multicountry Africa Telecommunications, Africa's growth per capita would be 1.0 point higher with the power, roads Republic of Korea's infrastructure. Estache, Speciale, and Veredas 2005 Multicountry Africa Various Confirms earlier work and underscores equal relevance for coastal and landlocked countries. Calderón 2008 Multicountry Africa Telecommunications, Africa's growth per capita would be 2.3 points higher with power, roads Mauritius's infrastructure. Calderón and Servén 2008 Multicountry Africa Telecommunications, Extends earlier results to show infrastructure also has a negative power, roads effect on inequality. Fedderke and Bogetic 2006 Country study South Africa Various Finds long-term relationship between infrastructure and growth based on robust econometric techniques. Ayogu 1999 Production function Nigeria Various Finds strong association between infrastructure and output in panel data. Kamara 2008 Production function Various Africa Various Finds strong association between infrastructure and output in panel data. Reinikka and Svensson 1999a Enterprise surveys Uganda Power Unreliable power is a significant deterrent to private sector investment. Escribano, Guasch, and Pena 2008 Enterprise surveys Africa Various Infrastructure has a substantial effect on total factor productivity. Source: Authors' elaboration. Meeting Africa's Infrastructure Needs 45 1.2 percentage points to per capita growth Figure 1.1 Changes in Growth per Capita Caused by Changes in Growth Fundamentals (figure 1.1). 2.5 percentage points of per capita The substantial contribution of infrastruc- 2.0 ture to Africa's recent growth is almost entirely economic growth attributable to greater penetration of telecom- 1.5 munications (figure 1.2). In contrast, the defi- 1.0 cient infrastructure of the power sector has 0.5 retarded growth, reducing per capita growth 0 for Africa as a whole by 0.11 percentage point and for southern Africa by as much as 0.2 per- ­0.5 e s a a a a ca a er op ric ric ric ric ric centage point. The effect of road infrastructure fri Tig Af Af Af Af Af r lA Eu rth t st rn ian ra es rn is generally positive, if rather small, perhaps Ea he nt No W As te ut Ce es st So W because of the absence of a widely available Ea measure of road quality, which is the critical stabilization policies structural policies infrastructure variable affecting transport costs. Source: Calderón 2008. More detailed microeconomic work on the relationship between infrastructure and the per- formance of firms (see table 1.1) supports these Figure 1.2 Changes in Growth per Capita Caused by Changes in Different Kinds of macroeconomic findings. The data consistently Infrastructure show a strong relationship between infrastruc- 2.5 percentage points of per capita ture stocks and the output, productivity, and 2.0 investment behavior of firms. An exhaustive economic growth study analyzed the entire set of investment cli- 1.5 mate surveys in Africa (Escribano, Guasch, and 1.0 Pena 2008). The central finding was that in most 0.5 African countries, particularly the low-income countries, infrastructure is a major constraint 0 on doing business and depresses firm produc- ­0.5 a ca a a ca a tivity by about 40 percent. The study first looked ric ric ric ric fri fri Af Af Af Af tA lA rth st rn ra at the relative contribution of infrastructure es Ea he nt No W ut Ce So and noninfrastructure investment variables to telecommunications power roads firm productivity (figure 1.3). For many coun- tries, such as Ethiopia, Malawi, and Senegal, the Source: Calderón 2008. negative effect of deficient infrastructure is at least as large as that of crime, red tape, corrup- tion, and lack of financing. Safe water's effect on health is well docu- For a subset of countries--among them mented. Serious illnesses transmitted through Botswana, Ethiopia, and Mali--power is the unsafe water, such as infectious diarrhea, are most limiting infrastructure factor, cited as a a leading cause of infant mortality (Esrey and major business obstacle by more than half the others 1991). Moreover, better water and sanita- firms in more than half the countries (figure tion service is associated with less malnutrition 1.3). Poorly functioning ports and slow cus- and stunting. Waterborne illnesses can be a sub- toms clearance are significant constraints for stantial economic burden, affecting both adult Burkina Faso, Cameroon, and Mauritius. Defi- productivity and children's overall health and ciencies in broader transport infrastructure education. The economic gain of meeting the and ICTs are less prevalent but nonetheless MDG target for water is estimated at $3.5 billion substantial in Benin and Madagascar. in year 2000 prices, and the cost-benefit ratio Infrastructure is also an important input to is about 11 to 1, suggesting that the benefits of human development (Fay and others 2005). safe water are far greater than the cost of pro- As such, it is a key ingredient in the MDGs vision (Hutton 2000; Hutton and Haller 2004). (table 1.2). Household members, primarily women and 46 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 1.3 Contribution of Infrastructure to Total Factor Productivity of Firms a. Overall contribution of infrastructure b. Infrastructure contribution by sector Namibia Eritrea Botswana Ethiopia Swaziland Botswana Mauritius Swaziland South Africa Namibia Kenya Mali Madagascar Uganda Tanzania Zambia Niger Kenya Burkina Faso Senegal Mauritania Madagascar Cameroon South Africa Mali Malawi Eritrea Mauritania Zambia Niger Ethiopia Burkina Faso Uganda Cameroon Senegal Benin Benin Tanzania Malawi Mauritius 0 20 40 60 80 100 0 20 40 60 80 100 % contribution to total factor productivity % contribution to total factor productivity infrastructure others electricity customs clearance transportation information and cummunication water technology Source: Escribano, Guasch, and Pena 2008. children, face a substantial opportunity cost in 1990; Venkataraman 1990). Similarly, better travel time when they have to fetch water. More access to electricity lowers costs for businesses than 20 percent of the population in Cameroon, and increases investment, driving economic Ghana, Mauritania, Niger, and Tanzania must growth (Reinikka and Svenson 1999b). travel more than 2 kilometers to their primary Improved transportation networks enable water supply. Rural dwellers tend to travel far- isolated rural communities to move into com- ther than urban dwellers (Blackden and Wodon mercial agriculture, thereby increasing their 2005; Wodon 2008). income, and to use health and education ser- Better provision of electricity has impor- vices some distance away (Barwell 1996; Calvo tant benefits for health because vaccines and and others 2001; Davis, Lucas, and Rikard medications can be safely stored in hospitals 1996; Ellis and Hine 1998; World Bank 1996). and food can be preserved at home (Jimenez By reducing the time and money it takes to and Olson 1998). Electricity also improves move goods, better transportation improves literacy and primary school completion rates competitiveness, helping create more jobs and because students can read and study after boost incomes (Limão and Venables 1999; sundown (Barnes 1988; Brodman 1982; Foley World Bank 2000, 2001). Meeting Africa's Infrastructure Needs 47 Table 1.2 Evidence on Links between Infrastructure and MDGs in Africa Study MDG Sector Conclusion Calvo 1994 Promote gender Water In four African countries surveyed, women saved over 1 hour per day after they began using a new, equality improved water source in their villages. Eberhard and Van Eradicate poverty Electricity In Cape Town, South Africa, households with electricity spent 3­5 percent of their incomes on energy, Horen 1995 compared with 14­16 percent for those without access. Lanjouw, Quizon, and Eradicate poverty Electricity In Tanzania, the presence of electricity in a village increased income from nonfarm business activities by Sparrow 2001 61%. Nonfarm income in villages with electricity was 109 times that in villages without electricity. Kenny 2002 Eradicate poverty ICT In Zambia, a survey of 21,000 farmers found that 50 percent of farmers credited radio-backed farm forums with increasing their crop yields. Saunders, Warford, and Eradicate poverty ICT A survey of transportation costs of an agricultural cooperative in Uganda in 1982 demonstrated that Wellenius 1994 200 agricultural cooperatives would save an average of $500,000 per year because of telecommunica- tions as a result of avoided transportation costs. Aker 2008 Eradicate poverty ICT In Niger, introduction of cell phones reduced price dispersion of grains, improving farmer and consumer welfare. World Bank 2000 Eradicate poverty Transport In Ghana, after a rural roads rehabilitation project, costs for transporting goods and passengers fell by about one-third on average. Croppenstedt and Eradicate poverty Transport In rural Ethiopia, farmers with access to an all-weather road increased their probability of using fertilizer Demeke 1996 by 10­20 percent because of cheaper transport costs. Doumani and Listorti 2001 Achieve universal Water In Nigeria, Guinea worm, a parasitic infection caused by poor-quality drinking water, was responsible for education 60 percent of all school absenteeism. Jimenez and Olson 1998 Reduce child mortality Electricity Clinics in Uganda and Ghana with photovoltaic cells for power kept refrigerators running for three to four years, whereas in Mali, clinics without these facilities had refrigerator failure about 20 percent of the time. Telecommunication Devel- Reduce child/maternal ICT In Mozambique, telemedicine could save hospitals up to $10,000 a year due to savings in transportation opment Bureau 1999 mortality costs for inappropriate referrals. Davis, Lucas, and Reduce child/maternal Transport In Tanzania, between one-third and one-half of villagers affected by a rural roads project reported Rikard 1996 mortality improved access to health care. McCarthy and Wolf 2001 Reduce child/maternal Water Across 20 African countries, access to safe water was found to be the fourth most important determi- mortality nant of health outcomes, after access to health care, income, and fertility rate. Sources: Authors' elaboration based largely on Kerf 2003a, 2003b, 2003c, and 2003d. Note: ICT = information and communication technology; MDG = Millennium Development Goal. The expansion of ICT networks democra- The differences are particularly large for paved- tizes access to information. It can be particu- road density, telephone main lines, and power larly critical for rural populations otherwise cut generation. The gap exists for both low-income off from important technological know-how or and middle-income groups. critical information about market prices (Kenny Was Africa's current infrastructure defi- 2002; Saunders, Warford, and Wellenius 1994). cit caused by a low historic starting point? In many cases, telecommunication improve- Has it always been worse-off than the rest of ments also reduce transportation spending by the world? In the 1960s (roads), 1970s (tele- allowing people to avoid fruitless journeys or to phones), and 1980s (power), Africa's stocks perform transactions remotely (Telecommuni- were quite similar to those of South or East cation Development Bureau 1999). Asia. (The one exception was paved-road den- sity, in which South Asia already enjoyed a huge advantage over both Africa and East Asia Africa's Infrastructure Deficit as far back as the 1960s. For household cov- erage of electricity, both South and East Asia By just about every measure of infrastructure were already far ahead of Africa in the early coverage, African countries lag behind their 1990s, and this gap has widened over time.) peers in other parts of the developing world Africa expanded its infrastructure stocks (see table 1.3; Yepes, Pierce, and Foster 2008). more slowly than other developing regions, 48 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 1.3 International Perspective on Africa's Infrastructure Deficit African low-income Other low-income African middle-income Other middle- Normalized units countries countries countries income countries Paved-road density 34 134 284 461 Total road density 150 29 381 106 Main-line density 9 38 142 252 Mobile density 48 55 277 557 Internet density 2 29 8.2 235 Generation capacity 39 326 293 648 Electricity coverage 14 41 37 88 Improved water 61 72 82 91 Improved sanitation 34 53 53 82 Source: Yepes, Pierce, and Foster 2008. Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density in lines per thousand population; generation capacity in megawatts per million population; electricity, water, and sanitation coverage in percentage of population. opening a gap between Africa and Asia (figure sanitation, the differences between subregions 1.4). The comparison with South Asia--with have been relatively small. Today, the South- a similar per capita income--is particularly ern African Development Community region striking. In 1970, Africa had almost three times has a strong lead over all other subregions on more electricity-generating capacity per million almost every aspect of infrastructure. The weak- people than did South Asia. By 2000, South est infrastructure endowments are in Central Asia had left Africa far behind--it now has Africa (for roads, water, and sanitation) and in almost twice the generating capacity per mil- East Africa (for ICT and power) (table 1.4). lion people. Similarly, in 1970 Africa had twice To better portray the diversity that exists the main-line telephone density of South Asia, across Africa, this report classifies countries into but by 2000, South Asia had drawn even. And in four types: (a) middle-income countries, (b) the case of mobile density, low-income African resource-rich countries, (c) fragile states, and countries are actually ahead of South Asia. (d) other low-income countries. (See box 1.1 for China and India have largely driven the full definitions.) These categories were chosen rapid infrastructure expansion in South and because they capture differences in financing East Asia. In particular, China has pursued a capacity and institutional strength that are rele- conscious strategy of infrastructure-led growth vant in understanding infrastructure outcomes. since the 1990s, committing more than 14 per- Outcomes across these different types of cent of GDP to infrastructure investment in countries are strikingly diverse. The difference in 2006 (Lall, Anand, and Rastogi 2008). infrastructure stocks between African middle- At independence, substantial variations in income countries and other African countries is infrastructure existed across different subre- to be expected, although African middle-income gions in Africa. Southern Africa started with countries have only a narrow edge over low- relatively high infrastructure endowments and income countries elsewhere in the developing achieved some of the highest annual growth world. The lags associated with fragile states are rates in infrastructure stocks over the last four readily understandable, given the disruption of decades. In 1980, the subregion had more than conflict. three times the generating capacity per million Especially striking is the extent to which people of other subregions; in 1970, it had five resource-rich countries lag behind others in times the telecommunication density of the their infrastructure endowment, despite their other subregions. With regard to roads, West greater wealth. In recent years, resource- Africa was in a much stronger position than the rich countries have devoted their additional other subregions in the 1960s but was overtaken wealth not to infrastructure development by southern Africa by the 1980s. In water and but to paying off their debt. The governance Meeting Africa's Infrastructure Needs 49 Figure 1.4 Growth of Africa's Infrastructure Stocks Compared with Asia a. Paved-road density d. Generation capacity 25 5 20 4 15 3 index index 10 2 5 1 0 0 1960 1970 1980 1990 2000 1980 1990 2000 b. Total road density e. Electricity coverage 5 100 90 4 80 70 3 60 percent index 50 2 40 30 1 20 10 0 0 1960 1970 1980 1990 2000 1990­95 1996­2000 2001­05 c. Main-line density f. Piped-water coverage 25 100 20 80 percent 15 60 index 10 40 5 20 0 0 1970 1980 1990 2000 1990­95 1996­2000 2001­05 Africa South Asia East Asia Sources: Banerjee and others 2008; Yepes, Pierce, and Foster 2008. Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density, in lines per 1,000 people; generation capacity, in megawatts per 1 million people. challenges in a resource-rich environment telephone, or Internet services are several multi- may also prevent the transformation of that ples of those paid in other parts of the developing wealth into infrastructure. world. Two explanations exist for Africa's high prices. First, the cost of providing infrastructure services in Africa is genuinely higher than else- Africa's Infrastructure Price where because of the small scale of production, Premium the reliance on suboptimal technologies, or the inefficient management of resources. Second, the The prices paid by African consumers for infra- high prices reflect high profit margins caused by structure services are exceptionally high by the lack of competition in service provision and global standards (table 1.5). The tariffs charged inadequate price regulation. Of course, the two in Africa for power, water, road freight, mobile factors can be simultaneously at play. 50 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 1.4 Intraregional Perspective on Africa's Infrastructure Deficit Low income, Low income, Normalized units ECOWAS EAC SADC Central Middle incomea Resource richa nonfragilea fragilea Paved-road density 38 29 92 4 284 14 14 55 Total road density 144 362 193 44 381 66 106 197 Main-line density 28 6 80 13 142 14 7 16 Mobile density 72 46 133 84 277 105 46 53 Internet density 2 2 4 1 8.2 1.6 1.2 3.1 Generation capacity 31 16 176 47 293 67 39 40 Electricity coverage 18 6 24 21 37 26 16 12 Improved water 63 71 68 53 82 57 57 66 Improved sanitation 35 42 46 28 53 32 37 31 Source: Yepes, Pierce, and Foster 2008. Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density in lines per thousand population; generation capacity in megawatts per million population; electricity, water, and sanitation coverage in percentage of population. EAC = East African Community; ECOWAS = Economic Community of West African States; SADC = Southern African Development Community. a. Country groupings are discussed in box 1.1. Table 1.5 Africa's High-Cost Infrastructure Other developing Sector Africa regions Power tariffs ($ per kilowatt-hour) 0.02­0.46 0.05­0.1 Water tariffs ($ per cubic meter) 0.86­6.56 0.03­0.6 Road freight tariffs ($ per ton-kilometer) 0.04­0.14 0.01­0.04 Mobile telephony ($ per basket per month) 2.6­21.0 9.9 International telephony ($ per 3-minute call to United States) 0.44­12.5 2.0 Internet dial-up service ($ per month) 6.7­148.0 11 Sources: Banerjee and others 2008; Eberhard and others 2008; Minges and others 2008; Teravaninthorn and Raballand 2008. Note: Ranges reflect prices in different countries and various consumption levels. Prices for telephony and Internet represent all developing regions, including Africa. Power provides the clearest example of a margins are exceptionally high, particularly in sector with genuinely higher costs in Africa Central and West Africa where they reach lev- than elsewhere. Many small countries rely on els of 60 to 160 percent. The underlying cause small-scale diesel generation that can cost up is the limited competition in the sector, com- to $0.40 per kilowatt-hour in operating costs bined with a highly regulated market based alone--about three times higher than coun- on tour de role principles, whereby freight is tries with larger power systems (over 500 allocated to transporters through a central- megawatts), which are typically hydropower ized queuing method rather than by allowing based (Eberhard and others 2008). truckers to enter into bilateral contracts with In contrast, high road freight tariffs in Africa customers directly. are caused more by excessive profit margins than The high prices for international tele- by high costs (Teravaninthorn and Raballand phone and Internet service in Africa reflect a 2008). The costs that Africa's trucking operators mixture of cost and profit. In countries that face are not significantly higher than in other have no access to a submarine cable and are parts of the world, even when informal pay- forced to rely on expensive satellite technol- ments are taken into account. However, profit ogy, charges are typically twice as high as in Meeting Africa's Infrastructure Needs 51 BOX 1.1 Introducing a Country Typology Africa's numerous countries face widely diverse economic not classified as resource intensive, using this criterion.) situations. Understanding that structural differences in coun- Examples include Cameroon, Nigeria, and Zambia. tries' economies and institutions affect their growth and · Fragile states are low-income countries that face particu- financing challenges as well as their economic decisions larly severe development challenges, such as weak gov- (Collier and O'Connell 2006), this report introduces a four- ernance, limited administrative capacity, violence, or the way country typology to organize the rest of the discussion. legacy of conflict. In defining policies and approaches This typology provides a succinct way of illustrating the diver- toward fragile states, different organizations have used sity of infrastructure financing challenges faced by different differing criteria and terms. Countries that score less African countries. than 3.2 on the World Bank's Country Policy and Insti- tutional Performance Assessment belong to this group. · Middle-income countries have GDP per capita in excess of Some 14 countries of Africa are in this category. Examples $745 but less than $9,206. Examples include Cape Verde, include Côte de Ivoire, the Democratic Republic of Congo, Lesotho, and South Africa (World Bank 2007). and Sudan (World Bank 2005). · Resource-rich countries are countries whose behaviors are · Other low-income countries compose a residual category strongly affected by their endowment of natural resources of countries with GDP per capita below $745 and that are (Collier and O'Connell 2006; IMF 2007). Resource-rich neither resource-rich nor fragile states. Examples include countries typically depend on minerals, petroleum, or both. Benin, Ethiopia, Senegal, and Uganda. A country is classified as resource rich if primary com- modity rents exceed 10 percent of GDP. (South Africa is Source: Briceño-Garmendia, Smits, and Foster 2008. MAURITANIA CAPE MALI NIGER VERDE CHAD ERITREA SENEGAL SUDAN GAMBIA GUINEA-BISSAU BURKINA FASO GUINEA BENIN NIGERIA SOMALIA SIERRA LEONE CÔTE GHANA ETHIOPIA D'IVOIRE TOGO CENTRAL AFRICAN LIBERIA CAMEROON REPUBLIC EQUATORIAL GUINEA UGANDA KENYA O GABON NG CONGO, RWANDA CO DEM REP BURUNDI TANZANIA MALAWI ANGOLA ZAMBIA MOZAMBIQUE ZIMBABWE MADAGASCAR MAURITIUS resource-rich countries NAMIBIA BOTSWANA nonfragile low-income countries SWAZILAND fragile low-income countries LESOTHO SOUTH AFRICA middle-income countries 52 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION countries that enjoy cable access. Even when Power Spending Needs Are by Far access to a submarine cable is obtained, coun- the Largest tries with a monopoly on this international Africa's largest infrastructure needs are in the gateway have tariffs that are substantially power sector. Whether measured in generat- higher than those without a monopoly (Min- ing capacity, electricity consumption, or secu- ges and others 2008). rity of supply, Africa's power infrastructure delivers only a fraction of the service found elsewhere in the developing world (Eberhard How Much Does Africa Need to and others 2008). The 48 countries of Africa Spend on Infrastructure? (with a combined population of 800 million) generate roughly the same amount of power as Meeting Africa's infrastructure needs and Spain (with a population of 45 million). Power developing cost-effective modes of infra- consumption, which is 124 kilowatt-hours per structure service delivery call for a substantial capita per year and falling, is only 10 percent of program of investment, rehabilitation, and dis- that found elsewhere in the developing world, ciplined maintenance combined. The physical barely enough to power one 100-watt light- infrastructure requirements are the grounds bulb per person for 3 hours a day. Africa's firms for a new set of estimates for spending require- report that frequent power outages cause them ments that are the foundation of this report. to lose 5 percent of their sales; this figure rises to In all cases, the estimated spending takes 20 percent for firms in the informal sector that into account both growth-related and social are unable to afford backup generators. Chap- demands for infrastructure, and it incorporates ter 8 in this volume contains a more detailed the costs of maintenance and rehabilitation as discussion of Africa's power challenges. well as new investment. Addressing this power shortage will require The time horizon for estimating spend- enormous investments in infrastructure over ing needs is a decade. The assumption is that the next decade. Based on four economic mod- over a period of 10 years running up to 2015, els, covering the Central, East, Southern, and the continent should be expected to address West African Power Pools, potential generation its infrastructure backlog, keep pace with the projects in each power pool are identified and demands of economic growth, and attain a ranked according to cost-effectiveness. These number of key social targets for broader infra- models make possible estimating the cost structure access (table 1.6). of meeting power demand under a range of Table 1.6 10-Year Economic and Social Targets for Investment Needs Estimates, 2006­15 Sector Economic target Social target Information and communication Complete submarine cable loop around Africa and 36,000-kilometer Extend GSM voice signal and public access broadband to technology fiber-optic backbone network interconnecting national capitals to 100 percent of the rural population. each other and to submarine cable loop. Irrigation Develop all financially viable opportunities for large- and small-scale n.a. irrigation, potentially some 12 million hectares. Power Attain demand-supply balance in power production, developing Raise household electrification rate by about 10 percentage points 7,000 megawatts of new generation capacity annually within a over current levels, entailing an additional 57 million new house- regional framework entailing 22,000 megawatts of new cross-border hold connections. interconnections. Transport Attain 250,000 kilometers of good-quality road networks supporting Raise the Rural Access Index from the current level of 34 percent regional and national connectivity goals. nationally to 100 percent in highest-value agricultural areas. Place entire urban population within 500 meters of road supporting motorized access. Water and sanitation n.a. Meet the Millennium Development Goals for water and sanitation. Sources: Banerjee and others 2008; Carruthers, Krishnamani and Murray 2008; Mayer and others 2008; Rosnes and Vennemo 2008; You 2008. Note: GSM = global systems mobile. n.a. = not applicable. Meeting Africa's Infrastructure Needs 53 alternative scenarios that consider access tar- next decade (to 2015), to allow power to flow gets, fuel prices, unit costs of investment, and freely from country to country. The financial feasibility of cross-border trade (Vennemo and returns on these interconnectors can be as high Rosnes 2008). as 120 percent in the Southern African Power Demand for power is almost directly pro- Pool; it is typically 20­30 percent in the other portional to economic growth. Installed capac- pools. Regional trade can also put Africa on a ity will need to grow by more than 10 percent path to cleaner development, because it would annually--or more than 7,000 megawatts a increase hydropower's share of the continent's year--just to meet Africa's suppressed demand, generation portfolio from 36 percent to 48 per- keep pace with projected economic growth, cent, displacing 20,000 megawatts of thermal and provide additional capacity to support the plant in the process and saving 70 million tons rollout of electrification. Since 1995, expan- of carbon emissions each year. Finally, raising sion of the sector has averaged barely 1 per- electrification rates will require extending dis- cent annually, or less than 1,000 megawatts tribution networks to reach almost 6 million a year. Most of that power would go to meet additional households a year over the next dec- nonresidential demands from the commercial ade (to 2015). and industrial sectors. The overall costs for the power sector in The most cost-effective way to expand Afri- Africa are a staggering $41 billion a year--$27 ca's power generation is through regional trade billion for investment and $14 billion for that allows countries to pool the most attrac- operation and maintenance (table 1.8). About tive primary energy resources across national half the investment costs are for development boundaries. Regional trade shaves around of new generating capacity. Approximately 15 $0.01 per kilowatt-hour off the marginal cost percent is earmarked for rehabilitation of exist- of power generation in each of the power pools ing generation and transmission assets. About (and as much as $0.02 to $0.04 per kilowatt- 40 percent of the costs are for the Southern hour for some countries), leading to savings of Africa Power Pool alone. about $2 billion a year in the costs of develop- ing and operating the power system. Mobilizing Achieving Water Security Remains an the benefits of regional trade depends on devel- Unquantified Challenge oping major untapped hydropower projects One important infrastructure requirement in the Democratic Republic of Congo, Ethio- not explicitly estimated in the investment costs pia, and Guinea, which would become major is water storage capacity, which is required to exporters in the Southern, East, and West Afri- reach water security. Africa experiences huge can Power Pools, respectively (table 1.7). It also swings in precipitation across areas, across hinges on establishing some 22,000 megawatts seasons, and over time (Grey and Sadoff 2006). of interconnectors that will be needed over the Climate change will only exacerbate this Table 1.7 Africa's Power Needs, 2006­15 New generation New cross-border New household Pool capacity (MW) interconnectors (MW) connections (millions) CAPP 4,395 831 2.5 EAPP 17,108 3,878 20.0 SAPP 33,319 11,786 12.2 WAPP 18,003 5,625 21.5 Island states 368 n.a. 1.2 Total 73,193 22,120 57.4 Source: Adapted from Rosnes and Vennemo 2008. Note: CAPP = Central African Power Pool; EAPP = Eastern African Power Pool (including Nile basin but excluding the Arab Republic of Egypt); Island states = Cape Verde, Madagascar, and Mauritius; SAPP = Southern African Power Pool; WAPP = Western African Power Pool. n.a. = not applicable. 54 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 1.8 Power Spending Needs, 2006­15 $ billions annually Investment New transmission and Total operation and Total spending Pool Rehabilitation New generation distribution Total investment maintenance needs CAPP 0.1 0.9 0.3 1.3 0.2 1.4 EAPP 0.3 3.5 3.0 6.8 1.1 7.9 SAPP 2.6 4.5 2.9 10.0 8.4 18.4 WAPP 1.0 3.5 3.7 8.2 4.0 12.3 Island states 0 0.1 0.2 0.3 0.3 0.6 Total 4.0 12.5 10.1 26.6 14.0 40.6 Source: Adapted from Rosnes and Vennemo 2008. Note: CAPP = Central African Power Pool; EAPP = Eastern African Power Pool (including Nile basin but excluding the Arab Republic of Egypt); Island states = Cape Verde, Madagascar, and Mauritius; SAPP = Southern African Power Pool; WAPP = Western African Power Pool. Row totals may not add exactly because of rounding errors. variability. As a result, water security--defined it constitutes less than 5 percent of Africa's as reliable water supplies and acceptable risks cultivated area, the irrigation-equipped area from floods and other unpredictable events, represents 20 percent of the value of agricul- including those from climate change--will tural production. Chapter 15 in this volume require a significant expansion of water stor- contains a more detailed discussion of Africa's age capacity from the current level of 200 irrigation challenges. cubic meters per capita. The amount of storage The model suggests that a further 6.8 million needed to withstand both flood and drought hectares are economically viable for irrigation, risks has not yet been precisely modeled for based on local agroecological characteristics, most African countries; hence, the needed market access, and infrastructure costs (You investment could not be estimated. Even a 2008). Most of this area, more than 5.4 mil- simplistic approach, however, such as estimat- lion hectares, is ideal for small-scale irrigation ing the cost of bringing all African countries schemes, assuming that they can be developed from their current storage levels of around 200 for an investment of no more than $2,000 a cubic meters per capita to South Africa's level hectare. A further 1.4 million hectares has the of 750 cubic meters per capita, is enough to potential for large-scale irrigation schemes illustrate the hundreds of billions of dollars that could be retrofitted to dams already serv- that could be required. ing hydropower purposes or incorporated into Nevertheless, about half the new genera- the development of new hydropower schemes tion capacity outlined for the power sector foreseeable within the next decade, assuming relates to water storage infrastructure with that the distribution infrastructure needed for multipurpose benefits. These hydropower irrigation can be added for an investment of no schemes would therefore also contribute, to an more than $3,000 a hectare. Finally, 1.7 million unknown extent, toward achieving the water hectares equipped for irrigation have fallen into security objective. The increased storage capac- disuse but could be recovered by rehabilitating ity they represent could--under appropriate the infrastructure. Spreading these investments multipurpose management principles--help over a 10-year span would require $2.7 billion attenuate the shocks associated with floods annually, plus a further $0.6 billion a year to and droughts. See chapter 14 in this volume support maintenance of new and existing sys- for a more detailed discussion of Africa's water tems (table 1.9). resource challenges. Reaching for the MDGs in Water Scope for Expanding Irrigated Areas and Sanitation Only 7 million hectares, in a handful of coun- The MDG target for access to safe water is tries, are equipped for irrigation. Although 75 percent of the population by 2015; for Meeting Africa's Infrastructure Needs 55 Table 1.9 Irrigation Spending Needs, 2006­15 $ billions annually Investment Total Large-scale Small-scale Total maintenance Rehabilitation schemes schemes investment Total 0.6 0.6 0.3 1.8 2.7 3.3 Source: You 2008. improved sanitation, it is 63 percent. As of 2006, Table 1.10 Water and Sanitation Spending Needs, the last year for which official data have been 2006­15 $ billions annually published, the figures for Africa were 58 per- cent and 31 percent, respectively. To meet the Sector Investment Maintenance Total MDG goal, the number of people with access Water 11.0 5.5 16.5 to safe water would need to increase from 411 Sanitation 3.9 1.4 5.4 million to 701 million by 2015--an increase of Total 14.9 7.0 21.9 29 million a year compared with recent prog- Source: Banerjee and others 2008. ress of only 11 million per year. To meet the MDG sanitation goal, the number of people Transport Needs Are Substantial with access to improved service would need to Africa's road density seems sparse compared increase from 272 million in 2006 to 617 mil- with the vastness of the continent, but it is not lion by 2015--an increase of 35 million a year unreasonable relative to the continent's popu- compared with recent progress of only 7 mil- lation and income. A more detailed discus- lion a year. Chapters 16 and 17 in this volume sion of Africa's transport challenges appears offer more detailed discussions of Africa's water in chapters 9­13 in this volume. The adequacy supply and sanitation challenges, respectively. of Africa's current transport network can best The overall price tag for reaching the water be assessed by examining whether it provides and sanitation MDG access is estimated at $22 an adequate level of connectivity to facilitate billion (roughly 3.3 percent of Africa's GDP), the movement of people and goods between with water accounting for more than two- regions, within nations, out of rural areas, thirds (table 1.10). Capital investment needs and across cities. Using a spatial model, one can be conservatively estimated at $15 billion can assess the cost of linking economic and a year (2.2 percent of the region's GDP). These demographic nodes through transport infra- needs include both new infrastructure and structures so as to achieve regional, national, rehabilitation of existing assets. Estimates are urban, and rural connectivity. based on minimum acceptable asset standards. Regional connectivity within the African It is assumed that access patterns (or relative continent requires a network that links all prevalence of water and sanitation modali- capital cities and cities with over 1 million ties) remain broadly the same between 2006 inhabitants to deep-sea ports and interna- and 2015 and that services are upgraded for tional borders. This objective can be achieved only a minimum number of customers. The with a two-lane network of a little over 100,000 maintenance requirements stand at $7 bil- kilometers maintained in good condition. lion a year (1.1 percent of the region's GDP). About 70 percent of this network is already in Operation and maintenance of network and place, but about one-quarter of it needs to be non-network services, respectively, amount widened from one lane to two lanes, and about to 3 percent and 1.5 percent of the replace- three-quarters of it needs to be improved to ment value of installed infrastructure. Reha- good quality. The overall cost of meeting this bilitation costs have been estimated based on target amounts to $2.7 billion a year, or barely a model that takes into account the mainte- 15 percent of total spending needs for the nance backlog of network infrastructure in transport sector. The bulk of this expenditure each country. is for investment. 56 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Regional connectivity also requires a rail Urban connectivity is defined as ensur- network, ports with adequate capacity, and ing that the entire urban population lives no airports. For railways, the main costs are for farther than 500 meters from a paved road rehabilitation of the existing track. For ports, capable of supporting motorized access. Afri- more container berths are needed to keep pace can cities today have paved-road densities well with the growth of international trade. For air below the average for well-provided cities in transport, the model does not suggest any need other developing countries, which typically for new terminals, but some expansion is pro- have densities of 300 meters per 1,000 inhab- vided based on passenger traffic projects. For itants. Meeting the objective of 500 meters runways, the investments primarily relate to would require adding 17,000 kilometers to the improving the condition of existing runways. current urban road network, and upgrading No need was found for building new runways, and improving 70,000 kilometers of the exist- although in a few cases lengthening existing ing network, costing $1.6 billion a year, which runways to support the use of larger aircraft serves to underscore the significance of urban was relevant. roads within Africa's overall transport require- Connectivity within a country requires ments. Most of this sum is needed to widen extending the regional network to link capital and pave existing urban roads. cities to their corresponding provincial cen- To create a transport network that provides ters and to other cities with more than 25,000 adequate regional, national, rural, and urban inhabitants by at least a one-lane paved road. road connectivity complemented by adequate The overall regional network and such national rail, port, and airport infrastructure will require networks would encompass 250,000 kilome- significant spending--$18 billion a year, half ters to meet this objective. About half of this of which is related to maintenance (table 1.11). network already exists in the form of paved Investment requirements are driven primarily roads, whereas the other half would need to by spending needed to upgrade the category of be upgraded to a paved network. The cost of existing assets (for example, from a gravel to a meeting this target is $2.9 billion a year. A sub- paved road), to improve the condition of exist- stantial share of that amount is for upgrading ing assets (from poor to good or fair condition), existing unpaved roads to paved surfaces. and to expand the capacity of existing assets Rural connectivity is defined as providing (for example, from one lane to two lanes). Just accessibility to all-season roads in high-value over half of this spending would be directed agricultural areas. Only one-third of rural Afri- at nonroad transport modes, particularly for cans live close to an all-season road, compared their maintenance. The remainder is roughly with two-thirds of the population in other evenly spread among national connectivity, developing regions. Because of low popula- urban connectivity, and rural connectivity. tion densities in rural Africa, raising this Rural Access Index to 100 percent for Africa would ICT Spending Needs Look More be essentially unaffordable. An alternative Manageable approach is to provide 100 percent rural con- Africa's progress in ICT is close to that seen else- nectivity to those areas with the highest agri- where in the developing world. The percentage cultural land value. Limiting access attention of Africa's population living within range of a to areas with 80 percent of the highest agri- global systems mobile signal rose from 5 per- cultural production value, the cost would be cent in 1999 to 57 percent in 2006 (Minges and a significant $2.5 billion a year, or close to 13 others 2008). Over the same period, more than percent of the overall spending requirement. 100 million Africans became mobile telephone About half of that sum is for maintenance, subscribers. Indeed, in some countries, house- whereas the remainder is devoted to improving hold access to mobile telephone services now the condition of existing rural roads, upgrad- exceeds that of piped water. Internet penetra- ing road surfaces to ensure all-season accessi- tion lags considerably, with little more than bility, and to a lesser extent, adding new roads 2 million subscribers and a further 12 million to reach isolated populations. estimated to be making use of public access Meeting Africa's Infrastructure Needs 57 facilities. The ICT revolution has been accom- Table 1.11 Transport Spending Needs, 2006­15 plished largely through market liberalization $ billions annually and private sector investment, which will Investment continue to be the main driving force behind Improve Upgrade Add Total Total Overall future investments. The state will need to con- Sector/area condition category capacity investment maintenance total tinue investing in a few critical areas, however. Regional connectivity 0.5 1.1 0.2 1.8 0.9 2.7 Chapter 7 in this volume contains a more detailed discussion of Africa's ICT challenges. National connectivity 0.5 1.2 0.2 1.9 1.0 2.9 The private sector will undertake the major Rural expenditures in this sector to service growth in connectivity 0.8 0.4 0.1 1.3 1.2 2.5 market demand. The urban market for ICT ser- Urban vices is well established and profitable. Demand connectivity 0.3 0.4 0.4 1.1 0.5 1.6 for voice services in this market is expected to Railways, ports, grow as penetration rates continue to rise from and airports 0.2 0.6 1.9 2.7 5.9 8.6 20 to 46 lines per 100 inhabitants. In addition, Total 2.2 3.7 2.7 8.6 9.6 18.2 incipient markets for broadband services are Source: Carruthers, Krishnamani, and Murray 2008. expected to expand from 0.04 to 2.54 lines per Note: Railways, ports, and airports include investments by South Africa's Transnet and other demand-driven transport investment needs covered by the private sector. 100 inhabitants. These demands can be met Column totals may not add exactly because of rounding errors. entirely by private sector investment. Spatial models are used to simulate the com- mercial viability of further expanding cover- would be required. Private finance would likely age of voice and broadband signals into rural be forthcoming for the highest-traffic seg- areas using global systems mobile and WiMAX ments. However, the more ambitious the aspi- (Worldwide Interoperability for Microwave rations for extending connectivity, the larger Access) technologies (Mayer and others 2008). the component of public finance that would The models consider the cost of network rollout be required. based on topographical factors and local avail- A modest level of broadband service could ability of power. They also estimate local revenue be provided using WiMAX technology to potential based on demographic densities, per provide low-volume connectivity to a lim- capita incomes, and estimated subscriber rates. ited number of institutions and public access With no market barriers, the private sector telecenters in rural areas. Using this approach, alone could profitably extend global systems and again in the absence of market barri- mobile signal coverage to about 95 percent of ers, the private sector alone could profitably Africa's population (Mayer and others 2008). extend WiMAX coverage to about 89 percent The remaining 5 percent, living in isolated of Africa's population (Mayer and others rural communities, is not commercially viable 2008). The remaining 11 percent, living in and would require a significant state subsidy to isolated rural communities, are not commer- connect. The percentage of the population that cially viable and would require a significant is not commercially viable varies substantially state subsidy to support network rollout. As across countries, from less than 1 percent in with voice, the percentage of the population Nigeria to more than 20 percent in the Demo- that is not commercially viable to cover varies cratic Republic of Congo. substantially across countries, from less than Broadband service, by contrast, is still in 1 percent in Nigeria to more than 70 percent its infancy and will expand only if significant in the Democratic Republic of Congo. investments are made in rolling out high- Finally, Africa is in the process of complet- capacity fiber-optic cable across the continent. ing a network of submarine cables that links it Just interconnecting all Africa's capitals would to the global intercontinental network. Several require a network of 36,000 kilometers of projects are already under way to close the loop fiber-optic cable. If the network were extended around the eastern side of the continent. Some to cover all cities with 500,000 or more inhab- strengthening of the West African submarine itants, more than 100,000 kilometers of cable system is also needed, plus cable links to service 58 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION outlying islands, such as the Comoros, Mada- Comparison with the Commission gascar, and the Seychelles. The private sector is for Africa showing considerable appetite to take on this The $93 billion estimate is more than twice the kind of investment. estimate of the Commission for Africa in 2005, The investment costs of this additional ICT which was based on cross-country econo- infrastructure, beyond what would be purely metric studies, rather than the more detailed driven by market demand, are relatively mod- country-level microeconomic modeling of est when compared with other infrastructure the Africa Infrastructure Country Diagnostic sectors. Achieving universal rural access for (Estache 2006). A recent update of the cross- both voice service and limited broadband ser- country model used for the Commission for vice based on WiMAX technology could be Africa report came up with a revised estimate accomplished for an investment of $1.7 billion of $80 billion to $90 billion (Yepes 2007). a year, the bulk of which could come from the Some 40 percent of the total is for the power private sector, with additional public fund- sector, which requires about $41 billion each year ing amounting to no more than $0.4 billion a (6 percent of African GDP; Rosnes and Vennemo year. Completing the submarine and intrare- 2008). A significant share of the spending for gional fiber-optic backbone would entail an power is for investment in multipurpose water annual (private sector) investment of less than storage schemes and thus makes an important $0.2 billion, although this sum would more contribution to water resources management. than double if a more ambitious network con- The second-largest component is the cost of necting all cities with over 500,000 inhabitants meeting the MDGs for water and sanitation-- were envisaged (table 1.12). Factoring in the about $22 billion (3 percent of regional GDP). market-driven investments needed to keep pace The third-largest price tag is associated with the with demand in established urban markets, transport sector, which comes in at just over $18 the estimated ICT sector annual investment need billion (3.6 percent of GDP). rises to $7 billion a year, plus another $2 billion annually for operation and maintenance. Distribution of Spending among Countries Three groups of countries--the middle- Overall Price Tag income countries, the resource-rich countries, and the low-income nonfragile states--share Africa's overall cost to build new infrastruc- roughly equally in the bulk of total spending. ture, refurbish dilapidated assets, and operate Each of these groups needs to spend around and maintain all existing and new installa- tions is estimated at almost $93 billion a year for 2006 through 2015 (15 percent of African Table 1.13 Overall Infrastructure Spending Needs for GDP; table 1.13 and figure 1.5). Africa, 2006­15 $ billions annually Capital Operation and Total Sector expenditure maintenance needs Table 1.12 ICT Spending Needs beyond the Purely Market Driven: Investment Only, 2006­15 ICT 7.0 2.0 9.0 $ billions annually Irrigation 2.7 0.6 3.3 Universal Universal access Fiber-optic Power 26.7 14.1 40.8 Type of access to voice to broadband backbone linking Submarine investment signal platform capital cities cables Transport 8.8 9.4 18.2 Private 0.58 0.68 -- -- WSS 14.9 7.0 21.9 Public 0.20 0.23 -- -- Total 60.4 33.0 93.3 Total investment 0.78 0.91 0.03 0.18 Sources: Authors' calculations based on Banerjee and others 2008; Carruthers, Krishnamani, and Murray 2008; Mayer and Source: Mayer and others 2008. others 2008; Rosnes and Vennemo 2008. Note: In contrast to the preceding tables, the expenditure for operation and maintenance is excluded Note: ICT = information and communication technology; because of the difficulty of apportioning it across the different subcategories presented. WSS = water supply and sanitation. -- Not available. Row totals may not add exactly because of rounding errors. Meeting Africa's Infrastructure Needs 59 Figure 1.5 Africa's Aggregate Infrastructure Spending Needs, by Country, 2006­15 a. As a percentage of GDP b. In US$ billions annually Africa Africa fragile low-income countries resource-rich countries nonfragile low-income countries middle-income countries resource-rich countries nonfragile low-income countries middle-income countries fragile low-income countries Congo, Dem. Rep. of South Africa Ethiopia Nigeria Niger Sudan Madagascar Ethiopia Senegal Congo, Dem. Rep. of Mozambique Kenya Chad Tanzania Tanzania Côte d´Ivoire Kenya Senegal Uganda Madagascar Burkina Faso Ghana Sudan Cameroon Zambia Mozambique Ghana Uganda Benin Chad Côte d´Ivoire Zambia Rwanda Niger Malawi Burkina Faso Lesotho Benin Nigeria Namibia Cameroon Malawi Namibia Rwanda South Africa Lesotho Cape Verde Cape Verde 0 10 20 30 40 50 60 70 0 20 40 60 80 100 % of GDP US$ billions/year investment (new and rehabilitation) operation and management Sources: Authors' calculations based on Banerjee and others 2008; Carruthers, Krishnamani, and Murray 2008; Mayer and others 2008; Rosnes and Vennemo 2008. $28 billion to $30 billion to meet its infrastruc- amounting to no more than 10 percent to ture needs. The price tag for the fragile states 13 percent of their respective GDPs. For is only about half as much at $13 billion. The low-income countries, however, as much as largest spending needs for an individual coun- 25 percent of GDP would be needed, rising to try by far are in South Africa, which requires an implausible 37 percent for the low-income $27 billion a year. fragile states. Ethiopia, Madagascar, Niger, and The burden of spending relative to the above all, the Democratic Republic of Congo countries' GDPs is very different across face an impossible challenge--their infrastruc- groups. For middle-income and resource-rich ture needs range from 26 to over 70 percent of countries, the burden appears manageable, GDP (see figure 1.5, panel a). 60 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Distribution of Spending--Investment of the spending needs are driven by targets versus Maintenance rather than economic growth; this applies, The overall spending requirements break down for example, to the transport spending needs two to one between investment and operation (which are largely based on connectivity objec- and maintenance, with the balance between tives) and to the water and sanitation spend- them shifting across country groupings. In the ing needs (which are based on the MDGs). The middle-income countries, the spending needs spending needs with the strongest direct link are skewed toward maintenance, which absorbs to economic growth are those for the power more than half the total. These countries have sector. However, because of the large backlog already put in place much of the infrastruc- in that sector, estimated spending needs con- ture they need, and their main challenge is to tain a strong component of refurbishment and preserve it in good condition. Across the three catch-up. Thus, even halving economic growth other country groupings, almost three-quarters estimates for the region would reduce esti- of spending needs are associated with invest- mated power spending needs by only 20 per- ment and only one-quarter with operation and cent. The global recession could be expected maintenance. These countries have a vast con- to affect demand for ICT services and trade- struction (and reconstruction) agenda to com- related infrastructure, such as railways and plete before they will have much to maintain. ports. However, the weight of those infrastruc- tures in the total spending needs is not much Will the Price Tag Grow--or Shrink? more than 10 percent. These estimates of investment are based on costs prevailing in 2006, the base year for all Notes of the African Infrastructure Country Diag- The authors of this chapter are Vivien Foster and nostic figures. It is well known that the unit Cecilia Briceño-Garmendia, who drew on back- costs of infrastructure provision have escalated ground material and contributions from César significantly during the last few years (Africon Calderón, Alvaro Escribano, J. Luis Guasch, Paul 2008). Lombard, Siobhan Murray, Jorge Pena, Justin The most reliable evidence available comes Pierce, Tito Yepes, and Willem van Zyl. from the road sector, where cost overruns 1. Although the Africa Infrastructure Country Diagnostic project is limited to the study of Sub- reported on multilateral agency projects in Saharan African countries, this book sometimes 2007 averaged 60 percent. The higher costs are substitutes Africa for Sub-Saharan Africa. The not just from inflation in petroleum and asso- reader should bear in mind, however, that the ciated input prices, but they also reflect a lack information refers only to Sub-Saharan Africa. of competition for civil works contracts and 2. Road density is measured in kilometers per 100 the tight position of the global construction square kilometers; telephone density in lines per industry, as well as lengthy delays in project thousand population; electricity generation in implementation. Similar escalations in unit megawatts per million population; and electric- costs have been reported anecdotally in other ity, water, and sanitation coverage in percentage areas of infrastructure, notably power. Pos- of population. sibly, the recent upward pressure on the costs 3. Monetary figures are in U.S. dollars unless oth- erwise noted. of infrastructure may be reversed as the cur- rent global downturn takes its toll, but that is hard to predict. Based on the situation in References 2006, the preceding estimates likely represent Africon. 2008. "Unit Costs of Infrastructure Proj- a conservative lower boundary for the cost ects in Sub-Saharan Africa." Background Paper 11, Africa Infrastructure Sector Diagnostic, of developing infrastructure assets at today's World Bank, Washington, DC. prices. Aker, Jenny. 2008. "Does Digital Divide or Provide? The global financial crisis of 2008 can be The Impact of Cell Phones on Grain Markets expected to reduce demand for some types of in Niger." Center for International and Devel- infrastructures, but it would not hugely alter opment Economics Research, University of the estimated spending needs. A large share California, Berkeley. Meeting Africa's Infrastructure Needs 61 Ayogu, Melvin. 1999. "Before Prebendalism: A Collier, Paul, and Stephen O'Connell. 2006. Positive Analysis of Core Infrastructure Invest- "Opportunities and Choices." Draft chapter for ment in a Developing Fiscal Federalism." African the synthesis volume of the Explaining African Development Review 11 (2): 169­98. Economic Growth PRoject of Oxford University Banerjee, Sudeshna, Quentin Wodon, Amadou and Swarthmore College, Centre for Study of Diallo, Taras Pushak, Hellal Uddin, Clarence African Economies. Tsimpo, and Vivien Foster. 2008. "Access, Afford- Commission for Africa. 2005. Our Common Inter- ability, and Alternatives: Modern Infrastructure est: Report of the Commission for Africa. London: Services in Africa." Background Paper 2, Africa Commission for Africa. Infrastructure Sector Diagnostic, World Bank, Washington, DC. Croppenstedt, Andre, and Mulat Demeke. 1996. "Determinants of Adoption and Levels of Barnes, Douglas F. 1988. Electric Power for Rural Demand for Fertiliser for Cereal Growing Growth: How Electricity Affects Rural Life in Devel- Farmers in Ethiopia." Working Paper 96-3, oping Countries. Boulder, CO: Westview Press. Centre for the Study of African Economies, Barwell, Ian. 1996. "Transport and the Village: Oxford, U.K. Findings from African Village-Level Travel Davis, Tony, Kimberly Lucas, and Ken Rikard. and Transport Surveys and Related Studies." 1996. Agricultural Transport Assistance Program: Discussion Paper 344, World Bank, Impact Study. Washington, DC: United States Washington, DC. Agency for International Development. Blackden, Mark, and Quentin Wodon, eds. 2005. Doumani, Fadi M., and James A. Listorti. 2001. "Gender, Time Use and Poverty in Sub-Saharan "Environmental Health: Bridging the Gaps." Africa." Working Series Paper 73, World Bank, World Bank Discussion Paper 422, World Bank, Washington, DC. Washington, DC. Briceño-Garmendia, Cecilia, Karlis Smits, and Easterly, William, and Ross Levine. 1997. "Africa's Vivien Foster. 2008. "Financing Public Infra- Growth Tragedy: Policies and Ethnic Divisions." structure in Sub-Saharan Africa: Patterns, Issues, Quarterly Journal of Economics 112 (4): 1203­50. and Options." AICD Background Paper 15, http://ssrn.com/abstract=88828. Africa Infrastructure Sector Diagnostic, World Bank, Washington, DC. Eberhard, Anton, Vivien Foster, Cecilia Briceño- Garmendia, Fatimata Ouedraogo, Daniel Brodman, Janice. 1982. "Rural Electrification and Camos, and Maria Shkaratan. 2008. "Under- the Commercial Sector in Indonesia." Discus- powered: The State of the Power Sector in sion Paper D-73L, Resources for the Future, Sub-Saharan Africa." Background Paper 6, Washington, DC. Africa Infrastructure Sector Diagnostic, World Calderón, César. 2008. "Infrastructure and Growth Bank, Washington, DC. in Africa." Working Paper 3, Africa Infrastruc- ture Country Diagnostic, World Bank, Eberhard, Anton, and Clive Van Horen. 1995. Washington, DC. Poverty and Power: Energy and the South African State. London: UCT and Pluto. Calderón, César, and Luis Servén. 2008. "Infra- structure and Economic Development in Sub- Ellis, Simon, and John Hine. 1998. "The Provi- Saharan Africa." Policy Research Working Paper sion of Rural Transport Services: Approach 4712, World Bank, Washington, DC. Paper." Sub-Saharan Africa Transport Policy Program Working Paper 37, World Bank, Calvo, Christina Malmberg. 1994. "Case Study on Washington, DC. the Role of Women in Rural Transport: Access of Women to Domestic Facilities." Sub-Saharan Escribano, Alvaro, J. Luis Guasch, and Jorge Pena. Africa Transport Policy Program Working Paper 2008. "Impact of Infrastructure Constraints on 11, World Bank, Washington, DC. Firm Productivity in Africa." Working Paper 9, Africa Infrastructure Sector Diagnostic, World Calvo, Christina Malmberg, Colin Gannon, Bank, Washington, DC. Kenneth M. Gwilliam, and Zhi Liu. 2001. "Transport: Infrastructure and Services-- Esfahani, Hadi Salehi, and Maria Teresa Technical Notes." PRSP Toolkit, World Bank, Ramirez. 2003. "Institutions, Infrastructure, Washington, DC. and Economic Growth." Journal of Develop- ment Economics 70: 443­77. http://ssrn.com/ Carruthers, Robin, Ranga R. Krishnamani, and abstract=310779. Siobhan Murray. 2008. "Improving Connectiv- ity: Investing in Transport Infrastructure in Sub- Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff. Saharan Africa." Background Paper 7, Africa 1991. "Effects of Improved Water Supply and Infrastructure Country Diagnostic, World Bank, Sanitation on Ascariasis, Diarrhea, Dracunculia- Washington, DC. sis, Hookworm Infection, Schistosomiasis and 62 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Trachoma." Bulletin of the World Health Organi- ------. 2003b. "Linkages between ICT and Pov- zation 89 (5): 609­21. erty Alleviation." World Bank, Washington, Estache, Antonio. 2006. "Africa's Infrastructure: DC. Challenges and Opportunities." Paper pre- ------. 2003c. "Linkages between Transport and sented at "Realizing the Potential for Profitable Poverty Alleviation." World Bank, Washington, Investment in Africa" seminar, IMF Institute DC. and Joint Africa Institute, Tunis, February ------. 2003d. "Linkages between Water and Sani- 28­March 1. tation and Poverty Alleviation." World Bank, Estache, Antonio, Biagio Speciale, and David Vere- Washington, DC. das. 2005. "How Much Does Infrastructure Mat- Lall, Rajiv, Ritu Anand, and Anupam Rastogi. 2008. ter for Growth in Sub-Saharan Africa?" World "Developing Physical Infrastructure: A Com- Bank, Washington, DC. parative Perspective on the Experience of the Fay, Marianne, Danny Leipziger, Quentin Wodon, People's Republic of China and India." Asian and Tito Yepes. 2005. "Achieving Child-Health- Development Bank, Manila. Related Millennium Development Goals: The Lanjouw, Peter, Jamie Quizon, and Robert Spar- Role of Infrastructure." World Development 33 row. 2001. "Non-Agricultural Earnings in (8): 1267­84. Peri-Urban Areas of Tanzania: Evidence from Fedderke, Johannes W., and Zeljko Bogetic. 2006. Household Survey Data." Food Policy 26: "Infrastructure and Growth in South Africa: 385­403. Direct and Indirect Productivity Impacts of 19 Limão, Nuno, and Anthony Venables. 1999. "Infra- Infrastructure Measures." Policy Research Work- structure, Geographical Disadvantage and ing Paper 3989, World Bank, Washington, DC. Transportation Costs." Policy Research Working Foley, Gerald. 1990. Electricity for Rural People. Paper 2257, World Bank, Washington, DC. London: Panos Institute. Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim Grey, David, and Claudia Sadoff. 2006 "Water for Kelly, Richard Green, and Alvaro Federico Barra. Growth and Development: A Framework for 2008. "Costing the Needs for Investment in ICT Analysis." Theme Document of the 4th World Infrastructure in Africa." Background Paper 3, Water Forum, Mexico City, March. Africa Infrastructure Sector Diagnostic, World Hutton, Guy. 2000. Considerations in Evaluating the Bank, Washington, DC. Cost-Effectiveness of Environmental Health Inter- McCarthy, F. Desmond, and Holger Wolf. 2001. ventions. Geneva: World Health Organization. "Comparative Life Expectancy in Africa." Policy Hutton, Guy, and Laurence Haller. 2004. Evalu- Research Working Paper 2668, World Bank, ation of the Costs and Benefits of Water and Washington, DC. Sanitation Improvements at the Global Level: Minges, Michael, Cecila Briceño-Garmendia, Mark Water, Sanitation and Health Protection of the Williams, Mavis Ampah, Daniel Camos, and Human Environment. Geneva: World Health Maria Shkratan. 2008. "Information and Com- Organization. munications Technology in Sub-Saharan Africa: A Sector Review." Background Paper 10, Africa IMF (International Monetary Fund). 2007. Infrastructure Sector Diagnostic, World Bank, Regional Economic Outlook: Sub-Saharan Africa. Washington, DC. Washington, DC: IMF. Reinikka, Ritva, and Jakob Svensson. 1999a. Jimenez, Antonio, and Ken Olson. 1998. Renew- "Confronting Competition: Firms' Investment able Energy for Rural Health Clinics. Golden, CO: Response and Constraints in Uganda." In National Renewable Energy Laboratory. http:// Assessing an African Success: Farms, Firms, and www.nrel.gov/docs/legosti/fy98/25233.pdf. Government in Uganda's Recovery, ed. Paul Col- Kamara, Samura. 2008. "Restoring Economic lier and Ritva Reinikka. Washington, DC: World Growth and Stability in Fragile and Post- Bank. Conflict Low Income States: Views from the ----. 1999b. "How Inadequate Provision of Pub- Trench." Paper presented at World Bank Poverty lic Infrastructure and Services Affects Private Reduction and Economic Management confer- Investment." Policy Research Working Paper ence, Washington, DC, April 15­16. 2262, World Bank, Washington, DC. Kenny, Charles. 2002. "The Costs and Benefits of Rosnes, Orvika, and Haakon Vennemo. 2008. ICTs for Direct Poverty Alleviation." Develop- "Powering Up: Costing Power Infrastructure ment Policy Review 20: 141­57. Investment Needs in Southern and Eastern Kerf, Michel. 2003a. "Linkages between Energy Africa." Background Paper 5, Africa Infra- and Poverty Alleviation." World Bank, structure Sector Diagnostic, World Bank, Washington, DC. Washington, DC. Meeting Africa's Infrastructure Needs 63 Saunders, Robert J., Jeremy Warford, and Bjorn World Bank. 1996. "Socio-Economic Impact of Wellenius. 1994. Telecommunications and Rural Roads: Impact Evaluation Report for Economic Development. 2nd ed. Baltimore: the Fourth Highway Project in Morocco." Johns Hopkins University Press for the OED Report 15808-MOR, World Bank, World Bank. Washington, DC. Telecommunication Development Bureau. 1999. ------. 2000. "Ghana: Building a Stronger Telemedicine and Developing Countries--Lessons Transportation System." Precis 189, Learned. Geneva: International Telecommunica- Washington, DC. tion Union. ------. 2001. Cities on the Move: A World Bank Teravaninthorn, Supee, and Gaël Raballand. Urban Transport Strategy Review. Washington, 2008. "Transport Prices and Costs in Africa: DC: World Bank. A Review of the Main International Corri- ------. 2005. Global Monitoring Report 2005. dors." Working Paper 14, Africa Infrastructure Washington, DC: World Bank. Sector Diagnostic, World Bank, Washington, DC. ------. 2007. DEPweb glossary. Development Education Program, World Bank. http://www Venkataraman, K. 1990. "Rural Electrification in .worldbank.org/depweb/english/modules/ the Asian and Pacific Region." In Power Systems glossary.html#middle-income. in Asia and the Pacific, with Emphasis on Rural Electrification, ed. Economic and Social Com- Yepes, Tito. 2007. "New Estimates of Infrastructure mission for Asia and the Pacific, 310­32. New Expenditure Requirements." World Bank, York: United Nations. Washington, DC. Vennemo, Haakon, and Ornica Rosnes. 2008. Yepes, Tito, Justin Pierce, and Vivien Foster. "Powering-Up: Costing Power Infrastructure 2008. "Making Sense of Sub-Saharan Africa's Investment Needs in Africa." Background Paper Infrastructure Endowment: A Benchmarking 5, Africa Infrastructure Country Diagnostic, Approach." Working Paper 1, Africa Infra- World Bank, Washington, DC. structure Country Diagnostic, World Bank, Wodon, Quentin, ed. 2008. "Electricity Tariffs Washington, DC. and the Poor: Case Studies from Sub-Saharan You, Liang Zhi. 2008. "Irrigation Investment Needs Africa." Working Paper 11, Africa Infrastruc- in Sub-Saharan Africa." Background Paper 9, ture Country Diagnostic, World Bank, Africa Infrastructure Country Diagnostic, World Washington, DC. Bank, Washington, DC. Chapter 2 Closing Africa's Funding Gap T he cost of addressing Africa's infrastruc- Substantial evidence indicates that a lot ture needs is estimated at $93 billion, more can be done within Africa's existing some 15 percent of Africa's GDP--about resource envelope. Inefficiencies of various two-thirds for investment and one-third for kinds total about $17 billion a year. If appro- maintenance. The burden varies greatly by priately tackled, fixing these inefficiencies country type. About half of the capital invest- could expand the existing resource envelope ment needs are for power, reflecting the par- by 40 percent. ticularly large physical deficits in that area. First, countries and development institu- Existing spending is higher than previously tions allocate $3.3 billion in infrastructure thought. African governments, infrastructure spending to areas that appear surplus to the users, the private sector, and external sources basic infrastructure requirements (as defined in together already contribute about $45 billion chapter 1 of this volume), which suggests that to directly address the infrastructure needs public and aid flows can be redirected toward previously identified. About one-third of this areas of greater impact on development. amount is spent by middle-income countries, Second, because only three-fourths of the whereas fragile states barely account for 5 per- capital budgets allocated to infrastructure are cent of it (about $2 billion in total), mirroring actually executed, about $2 billion in public the weakness of their economies and the investment is being lost. enormous disparity in terms of financing and Third, underspending on infrastructure institutional capabilities across Sub-Saharan asset maintenance is another major waste of African countries. About two-thirds of the resources because the cost of rehabilitating existing spending is domestically sourced, infrastructure assets is several times higher from taxes or user charges, and channeled than the cumulative cost of sound preven- through public institutions, making the public tive maintenance. In the road sector alone, sector--governments and nonfinancial pub- addressing undermaintenance can save $1.9 lic enterprises together--the most important billion a year in rehabilitation, or spending $1 financier of capital investment, funding more on maintenance can be a savings of about $4 than half of total investment. to the economy. 65 66 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Fourth, Africa's power and water utilities and the countries benefit fully from the additional state-owned telecommunication incumbents finance. Otherwise, what is the use of pouring waste about $6 billion a year on inefficiencies water into a leaking bucket? such as overstaffing, revenue undercollection, and distribution losses. Fifth, underpricing infrastructure services Spending Allocated to Address accounts for $4.7 billion a year in lost revenues. Infrastructure Needs In all, with existing allocation patterns and even if potential efficiency gains are fully Africa is spending $45 billion a year to address captured, a funding gap of $31 billion a year its infrastructure needs. Existing spending on remains: three-quarters for capital and one- infrastructure in Africa is higher than previ- quarter for maintenance. About $23 billion of ously thought when budget and off-budget this gap relates to power and a further $11 bil- spending (including state-owned enterprises lion to water supply and sanitation (WSS). For and extrabudgetary funds) and external financ- fragile states, the funding gap is an implausible ing (comprising official development assistance 25 percent of GDP on average, almost equally [ODA], financiers from outside the Organisation divided among energy, water, and transport. for Economic Co-operation and Development How can Africa close such a sizable funding [OECD], and private participation in infra- gap, equivalent to one-third of the estimated structure [PPI]) are taken into account. This infrastructure needs? Additional funds will be level of spending is associated with allocations required, and in a few countries--mainly the directly targeted to cover the needs identified fragile ones--the magnitude of the funding in chapter 1. In practice, however, some coun- gap calls for considering taking more time to tries spend more on some infrastructure subsec- attain targets or using lower-cost technologies. tors than the estimated benchmark requirements Historical trends do not suggest much pros- while incurring funding gaps in other subsec- pect of increasing allocations from the public tors. This existing spending with potential for budget: even when fiscal surpluses existed, they reallocation is not counted here but is consid- did not visibly favor infrastructure. External ered later in this chapter. finance has been buoyant in recent years, and The four-way country typology introduced disbursements will likely continue to grow as in chapter 1 of this volume--comprising mid- projects committed move to the implementa- dle-income countries, resource-rich coun- tion stage. In light of today's financial crisis, tries, fragile states, and other low-income however, prospects for new commitments do countries--serves as a basis for summarizing not look good. Private capital flows, in particu- the diversity of infrastructure financing chal- lar, can be expected to decline. Fiscal pressure lenges (see box 1.1). Expressed as a percentage is growing in donor countries, and to judge by of GDP, infrastructure spending is comparable previous crises, foreign aid is likely to slow. across the different country types, at around By delaying investment timetables, and 5­6 percent of GDP, with the exception of assuming that efficiency gains are fully cap- nonfragile low-income countries, which spend tured, many countries could even attain the at 10 percent of their GDP. In absolute dollar infrastructure targets without increasing their terms, the middle-income countries spend the spending envelopes. Targeting a high level of most (roughly $16 billion), reflecting their service might not always work in the best inter- much larger purchasing power. Fragile states, est of a country. Lower-cost technologies can by contrast, account for a tiny amount of over- permit broadening the portion of the popula- all spending (about $2 billion), reflecting the tion with access to some level of service. weakness of their economies (table 2.1). Closing Africa's funding gap inevitably The public sector, with the lion's share of requires undertaking needed reforms to reduce spending, is by far the most important finan- or eliminate the inefficiencies of the system. cier. In the middle-income countries, domestic Only then can the infrastructure sectors become public sector resources (comprising tax reve- attractive to a broader array of investors and nues and user charges raised by state-owned Closing Africa's Funding Gap 67 Table 2.1 Spending of Most Important Players Traced to Needs (Annualized Flows) Percentage of GDP $ billions O&M Capital expenditure O&M Capital expenditure Country Public Public Non-OECD Total capital Public Public Non-OECD Total capital type sector sector ODA financiers Private expenditures Total sector sector ODA financiers Private expenditures Total Middle income 3.7 1.2 0.1 0.0 0.8 2.1 5.8 10.0 3.1 0.2 0.0 2.3 5.7 15.7 Resource rich 1.1 1.5 0.2 0.7 1.7 4.1 5.2 2.5 3.4 0.5 1.4 3.8 9.1 11.7 Low-income nonfragile 4.0 1.5 2.2 0.5 1.9 6.1 10.1 4.4 1.6 2.5 0.6 2.1 6.7 11.1 Low-income fragile 2.0 0.6 1.0 0.8 1.2 3.6 5.6 0.8 0.2 0.4 0.3 0.5 1.4 2.1 Africa 3.2 1.5 0.6 0.4 1.5 3.9 7.1 20.4 9.4 3.6 2.5 9.4 24.9 45.3 Sources: Africa Infrastructure Country Diagnostic (AICD); Briceño-Garmendia, Smits, and Foster 2008 for public spending; PPIAF 2008 for private flows; Foster and others 2008 for non-OECD financiers. Note: Aggregate public sector covers general government and nonfinancial enterprises. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. O&M = operation and maintenance; ODA = official development assistance; OECD = Organisation for Economic Co-operation and Development; Private = private participation in infrastructure and household self-finance of sanitation facilities. enterprises) account for the bulk of spending Patterns of specialization are clear across across all infrastructure subsectors. Across the the different sources of external finance (figure other country typologies, domestic public sec- 2.1). Across sectors, PPI is strongly concen- tor resources contribute approximately half trated on information and communication of total spending. One-third of this aggregate technology (ICT), which shows the highest public sector spending (or an equivalent of commercial returns. ODA has tended to focus 1.5 percent of GDP) can be traced exclusively on public goods with high social returns, nota- to capital investments. bly roads and water. Much non-OECD finance This level of effort by African governments has gone to energy and, to a lesser extent, to to develop their infrastructure pales when com- railways, both sectors with strong links to pared with what East Asian countries have done industry and mining. Across countries, PPI has in recent decades. China, for example, adopted a tended to go to middle-income and resource- determined and clear strategy to increase infra- rich countries, which have the greatest ability structure investment (publicly and privately to pay for services. Non-OECD finance has financed) as a means of achieving accelerated shown a preference for resource-rich countries, economic growth. Fixed capital formation in with a strong pattern linking infrastructure Chinese infrastructure more than doubled investment and natural resource extraction, between 1998 and 2005. By 2006, only infra- and ODA has preferred nonfragile low-income structure investment was higher than 14 percent states with limited domestic resources but ade- of GDP, perhaps the highest in the world. quate institutional capacity. The fragile states Excluding middle-income countries, external do not seem to have captured their fair share financiers contribute roughly one-half of Afri- from any of the external sources. ca's total spending on infrastructure. External sources include ODA from the OECD countries, official finance from non-OECD countries (such How Much More Can Be Done as China, India, and the Arab funds), and PPI. within the Existing Resource External finance is primarily for investment-- Envelope? broadly defined to include asset rehabilitation and reconstruction--and in most cases does not Africa is losing about $17 billion per year to provide for O&M. Since the late 1990s, PPI has various inefficiencies in infrastructure opera- been the largest source of external finance, fol- tions or spending. In this context, four distinct lowed by ODA and non-OECD finance, which opportunities can be identified for efficiency are broadly comparable in magnitude. gains. First, improving budget execution rates 68 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 2.1 Sources of Financing for Infrastructure Capital Investment, by Sector and Country Type a. Sector b. Type of country 10 10 9 8 8 7 US$ billions US$ billions 6 6 5 4 4 3 2 2 1 0 0 information and power transport water supply middle income resource rich nonfragile fragile communication and sanitation low income low income technology public official development assistance private participation in infrastructure non-OECD financiers Sources: Africa Infrastructure Country Diagnostic (AICD); Briceño-Garmendia, Smits, and Foster 2008 for public spending; PPIAF 2008 for private flows; Foster and others 2008 for non-OECD financiers. would increase the potential of fully using terms, however, middle-income countries resources allocated to public investment. Sec- have a much larger infrastructure budget, with ond, reallocating existing spending toward spending per capita at $150­$200, compared subsectors in greatest need, therefore with with $20­$40 in low-income countries. In highest economic returns, would allow the other words, per capita budgetary spending existing budget envelope to better cover exist- on infrastructure by middle-income coun- ing needs. Third, raising user charges closer to tries is about five times that of low-income cost-recovery levels would provide more effi- countries. cient price signals and help capture lost rev- Overall, spending on transport (notably enues. Fourth, reducing operating efficiencies roads) is the single-largest infrastructure item of utilities and other service providers would in general government accounts. It ranges from prevent waste of significant resources, support about half of all general government spending healthier utilities, and improve service quality. on infrastructure in middle-income countries to 60 percent in low-income countries. Water Raising Capital Budget Execution and sanitation spending is the second-largest African central governments alone allocate, on category, particularly in the middle-income average, 1.5 percent of GDP, or 6­8 percent of countries. Energy spending features heavily in their national budgets, to support the provi- resource-rich countries. sion of infrastructure (table 2.2). For Africa, From a functional perspective, more than 80 this effort translates into about $300 million percent of budgetary spending goes to invest- a year for an average country, which would ment. With the exception of middle-income not take many African countries a long way. countries and the ICT sector, the central To put this figure in perspective, an invest- government makes the bulk of public invest- ment of $100 million can purchase about 100 ment, even in sectors in which state-owned megawatts of electricity generation, 100,000 enterprises provide most services. Strikingly, new household connections to water and sew- relative to central government, nonfinancial erage, or 300 kilometers of a two-lane paved public institutions, such as utilities and other road. It runs well short of covering the invest- service providers, make little infrastructure ment needs estimated in chapter 1 of this vol- investment (figure 2.2). The state-owned ume (see chapter 1 for details). enterprises are essentially asset administrators. As a percentage of GDP, budget spending This spending pattern reflects government on infrastructure is comparable across low- control of some of the main sources of invest- and middle-income countries. In absolute ment finance, be they royalty payments (in Closing Africa's Funding Gap 69 Table 2.2 Annual Budgetary Flows Percentage of GDP $ billions Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income 0.0 0.1 0.1 0.6 0.7 1.5 0.0 0.2 0.2 1.7 1.8 4.0 Resource rich 0.4 0.0 0.1 0.8 0.3 1.6 0.8 0.0 0.3 1.7 0.7 3.6 Low-income nonfragile 0.1 0.1 0.3 0.7 0.3 1.5 0.1 0.1 0.3 0.8 0.4 1.7 Low-income fragile -- -- -- 0.6 0.1 0.7 -- -- -- 0.2 0.0 0.3 Africa 0.1 0.1 0.1 0.7 0.5 1.5 0.8 0.4 0.8 4.4 3.1 9.5 Sources: Africa Infrastructure Country Diagnostic; Briceño-Garmendia, Smits, and Foster 2008. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. -- Not available. resource-rich countries) or external develop- Figure 2.2 Split Investment Responsibilities between Governments and Public ment funds (in low-income countries). It also Enterprises, by Type of Country and Sector reflects to some extent the limited capability of 1.0 state-owned enterprises to fund their capital 0.9 0.8 investment through user fees. 0.7 % of GDP Because central governments are such key 0.6 players in infrastructure investment, inefficien- 0.5 0.4 cies within the public expenditure management 0.3 systems are particularly detrimental. By way of 0.2 example, central governments face significant 0.1 0 problems in executing their infrastructure bud- r e gi w i ich in e e r e gi w i ich in e e r e gi w i ich in e e r e gi w i ich in e e ra reso om lo om m ra so om lo om m ra so om lo om m ra so om lo om m gets. African countries are, on average, unable fra e lo ce r fra e lo ce r fra e lo ce r fra e lo ce r co co co co nc le nc nf re nc le nc nf re nc le nc nf re nc le nc ei ei ei ei gi u gi u gi u gi u w w w w dl dl dl dl to spend as much as one-quarter of their capi- id id id id m m m m l l l l tal budgets and one-third of their recurrent nf no no no no budgets in the corresponding fiscal year (table information and power transport water 2.3). The poor timing of project appraisals and communication technology late releases of budgeted funds because of pro- curement problems often prevent the use of government public enterprises resources within the budget cycle. Delays affect- Sources: Africa Infrastructure Country Diagnostic; Briceño-Garmendia, Smits, and Foster 2008. ing in-year fund releases are also associated with Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. poor project preparation, leading to changes in the terms agreed upon with contractors in the original contract (deadlines, technical spec- finding assumes, arguably a stretch, that budget ifications, budgets, costs, and so on). In other estimates are realistic and aligned with resources cases, cash is reallocated to nondiscretionary available. Either way, the associated saving sug- spending driven by political or social pres- gests that the resolution of these planning, bud- sures. Historically, the road sector is the worst geting, and procurement challenges should be offender of unused budget allocations, some- central to the region's reform agenda. times as much as 60 percent of the budget. Even if budgets are fully spent, concerns Improving the efficiency of budget execu- exist about whether funds reach their final tion could make a further $2 billion available destinations. A few public expenditure track- for infrastructure spending each year. If the bot- ing surveys have attempted to trace the share tlenecks in capital execution could be resolved, of each budget dollar that results in pro- countries could on average increase their capital ductive front-line expenditures. Most of the spending by about 30 percent without any existing case studies concern social sectors, increase in current budget allocations. This rather than infrastructure, but they illustrate 70 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 2.3 Average Budget Variation Ratios for Capital Spending Overall Country type Electricity Communication Roads Transport WSS Irrigation infrastructure Middle income -- 100 75 100 66 60 78 Resource rich 60 37 71 73 43 -- 65 Low-income nonfragile 75 64 72 72 72 68 76 Low-income fragile -- -- -- -- -- -- -- Sub-Saharan Africa 66 72 73 79 66 66 75 Sources: Africa Infrastructure Country Diagnostic; adapted from Briceño-Garmendia, Smits, and Foster 2008. Note: Budget variation ratio is defined as executed budget divided by allocated budget. Based on annualized averages for 2001­06. WSS = water supply and sanitation. -- Not available. leakages of public capital spending as high as the highest, averaging for the continent more 92 percent (see Pritchett 1996; Rajkumar and than a 100 percent economic rate of return Swaroop 2002; Reinikka and Svensson 2002, and well above returns for rehabilitation and 2003; Warlters and Auriol 2005; and references new construction (table 2.5). By favoring cited therein). investment over maintenance, African govern- ments have been implicitly equating public Reallocating Existing Spending to investment with productive expenditure, Subsectors in Need even though not all investment is productive About $3.3 billion a year is spent above the and not all current spending is wasteful.1 The estimated requirements to meet the identified maintenance of public goods under the juris- infrastructure needs (see chapter 1 in this vol- diction of general governments is essential ume). This spending--funded by (or through) to harness the economic returns to capital and public budgets--includes domestically raised to avoid costly rehabilitation. Highest returns to funds and international aid (OECD and non- maintenance are seen in networks already well OECD sources). Most of this apparent over- developed, particularly in middle-income coun- spending is in telecommunications in countries tries and nonfragile low-income countries. that have maintained state ownership of the From a sectoral perspective, economic fixed-line incumbent. State ownership not only returns to railway investments are the lowest uses expensive public resources in activities that among infrastructure interventions. Railway the already competitive telecommunication rehabilitation interventions are justified only market can provide but also forgoes future tax for a few higher traffic systems. Investment in revenues from expanded business activity. To a water supply and irrigation would bring very much lesser extent and only in middle-income solid returns in health benefits and productiv- countries, the other sector showing potential ity, but returns to power generation need to for reallocation is transport. The overspending be compounded by coordinated investment in in this case is driven by apparent overinvest- transmission and distribution networks. ment in road networks that, as will be seen later, paradoxically coexists with undermainte- Improving Cost Recovery from nance (table 2.4). User Charges How much of that spending in ``excess'' of Two-thirds of African power and water utili- infrastructure needs is influenced by politi- ties apply tariffs that comfortably cover oper- cal factors? How far are these politics-tainted ating costs, but only one-fifth of those utilities decisions from economic optimization? set tariffs high enough to recover full capital How should these resources be reallocated? costs. Achieving recovery of only operating Estimates of the economic rates of return to costs across all African power and water utili- key infrastructure interventions can provide ties would raise $2.5 billion a year (0.4 percent some answers. of the region's GDP). Revising tariffs to make Across infrastructure interventions in Africa, them equal to long-term marginal costs, and the rates of return to road maintenance are thereby enabling all African power and water Closing Africa's Funding Gap 71 Table 2.4 Existing Disbursements above Those Directed to Infrastructure Needs, Annualized Flows Percentage of GDP $ billions Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income -- 1.4 -- 0.0 0.1 1.5 -- 3.7 -- 0.0 0.3 4.1 Resource rich -- -- 0.0 0.4 -- 0.4 -- -- 0.0 0.8 -- 0.8 Low-income nonfragile -- 0.1 -- 0.2 -- 0.3 -- 0.1 -- 0.3 -- 0.4 Low-income fragile -- -- -- -- -- -- -- -- -- -- -- -- Africa -- 0.5 -- -- -- 0.5 -- 3.3 -- -- -- 3.3 Source: Africa Infrastructure Country Diagnostic. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals for Africa differ from the sum of the individual groups because reallocation is allowed only within groups. ICT = information communication and technology; WSS = water supply and sanitation. -- Not available. Table 2.5 Economic Rates of Return for Key Infrastructure Railway Road Road Road Country type rehabilitation Irrigation rehabilitation upgrades maintenance Generation Water Middle income 18.5 19.3 45.4 19.8 143.0 13.6 26.8 Resource rich 10.8 24.2 16.2 17.4 114.5 20.2 37.0 Low-income nonfragile 6.2 17.2 17.6 12.8 125.7 14.3 7.7 Low-income fragile 2.5 -- 9.2 12.0 67.6 24.7 36.9 Sub-Saharan Africa 5.1 22.2 24.2 17.0 138.8 18.9 23.3 Source: Africa Infrastructure Country Diagnostic. Note: -- Not available. utilities to recover capital costs also, would Underpricing user charges for roads costs the increase the potential for efficiency gains to region some $0.6 billion a year (0.1 percent $4.2 billion a year (0.7 percent of the region's of GDP). GDP; table 2.6). Although underpricing is equally prevalent in power and water utilities, Reducing the Operating Inefficiencies the value of the lost GDP revenues is slightly of Utilities higher for power (at 0.4 percent of GDP) than African state-owned enterprises are charac- for water (at 0.3 percent). terized by low investment and high operating Raising tariffs to cost-recovery levels is evi- inefficiency. State-owned enterprises account dently easier said than done and entails a host for between 80 percent (energy) and 40 per- of social and political challenges. Chapter 3 in cent (water) of total public expenditures this volume examines these issues in greater (general government and nonfinancial enter- depth and provides a realistic appraisal of the prises). Despite their large resource base, they feasibility of improving cost recovery for util- invest comparatively little--on average, an ity services in Africa. equivalent of between 15 percent (energy) and In the road sector, a widespread movement 18 percent (water) of the government resource exists for using fuel levies and taxes as indirect envelope. As a result, governments are typi- user charges (see chapter 10 in this volume). cally required to step in to assume most of the For this system to work, fuel levies need to investment responsibilities of state-owned be set high enough to cover the maintenance enterprises, which are relegated to undertak- costs imposed by the use of the road network. ing daily operation and maintenance. In many Comparing existing fuel levies with the levels cases, investment is unaffordable because of needed to secure road maintenance makes it the significant underpricing of services, which possible to estimate the underpricing in roads. barely allows the recovery of operating costs. 72 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 2.6 Potential Gains from Increased Cost Recovery Percentage of GDP $ billions annually Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income 0.0 -- -- 0.0 0.4 0.4 0.0 -- -- 0.0 1.0 1.0 Resource rich 0.8 -- -- 0.1 0.1 0.9 1.7 -- -- 0.2 0.2 2.0 Low-income nonfragile 0.8 -- -- 0.2 0.3 1.1 0.8 -- -- 0.2 0.3 1.3 Low-income fragile 0.0 -- -- 0.1 0.6 0.7 0.0 -- -- 0.0 0.2 0.3 Africa 0.4 -- -- 0.1 0.3 0.7 2.3 -- -- 0.6 1.8 4.7 Source: Africa Infrastructure Country Diagnostic. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. -- Not available. In addition, most state-owned enterprises Fourth, the undermaintenance of infrastruc- operate at arm's length from the central gov- ture assets is widespread but represents a false ernment, failing in practice to meet criteria economy because the rehabilitation of assets for sound commercial management. When is usually much more costly in present-value these enterprises run into financial difficul- terms than the preventive maintenance to ties, the general government--as the main avoid such asset deterioration. stakeholder--acts as the lender of last resort, absorbing debts and assuming by default the Overemployment. Overemployment reaches financial, political, regulatory, and misman- $1.5 billion a year (0.24 percent of GDP; table agement risks. Lumpy capitalizations and debt 2.7). Most overemployment was found in swaps that cover the cumulative cost of opera- telecommunication utilities in countries that tional inefficiencies are frequent events in the have retained state ownership of their fixed- African utility sector, which can potentially line incumbent. In Sub-Saharan Africa, such create a moral hazard that would perpetuate utilities achieve, on average, 94 connections operational inefficiencies if proactive reforms per employee, compared to developing- are not undertaken. country benchmarks of 420 connections per This section considers four types of oper- employee, an overemployment ratio of 600 ational inefficiencies and estimates their percent. Similarly, African power and water potential monetary value. First, state-owned utilities have overemployment ratios of 88 enterprises may retain more employees than percent and 24 percent, respectively, over is strictly necessary to discharge their func- non-African developing-country benchmarks. tions, often because of political pressure to These striking results for labor inefficiencies provide jobs for members of certain interest underscore the importance of strengthening groups. This issue affects state-owned enter- external governance mechanisms that can prises across the board, including those in impose discipline on the behavior of state- ICT, power, and water. Second, utilities incur owned enterprises. Overemployment par- substantial losses on their power and water tially explains why in African countries with a distribution networks. Both poor network publicly owned operator, the share of spend- maintenance, which leads to physical leakage, ing allocated to capital spending frequently and poor network management, which leads to remains below 25 percent of total spending clandestine connections and various forms of despite pressing investment needs. theft, contribute to these losses. Third, power and water utilities face serious problems in col- Distribution Losses. Distribution losses amount lecting their bills, largely a result of the social to $1.8 billion a year (0.3 percent of GDP). Afri- and political impediments to disconnecting can power utilities typically lose 23 percent of services, which lead to a nonpayment culture. their energy in distribution losses, more than Closing Africa's Funding Gap 73 Table 2.7 Potential Gains from Greater Operational Efficiency Percentage of GDP $ billions annually Operational Transport Transport inefficiencies Electricity ICT Irrigation (roads) WSS Total Electricity ICT Irrigation (roads) WSS Total Losses 0.2 -- -- -- 0.1 0.3 1.3 -- -- 0.5 1.8 Undercollection 0.3 -- -- 0.1 0.1 0.5 1.9 -- 0.5 0.5 2.9 Labor inefficiencies 0.0 0.2 -- -- -- 0.2 0.3 1.3 -- -- 0.0 1.5 Undermaintenance -- -- -- 0.2 -- 0.2 -- 1.4 -- 1.4 Total 0.5 0.2 -- 0.3 0.2 1.2 3.4 1.3 -- 1.9 1.0 7.5 Source: Africa Infrastructure Country Diagnostic. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. -- Not available. twice the best practice of 10 percent. Similarly, Figure 2.3 Rehabilitation Backlog African water utilities typically lose 35 percent 50 of their water in distribution losses, nearly twice the 20 percent benchmark. The finan- % of assets in need of rehabilitation 40 cial value of those distribution losses is much higher for power at $1.3 billion per year than 30 for water at $0.5 billion per year. 20 Undercollection of Bills. The undercollection of bills amounts to $2.9 billion a year (0.5 per- 10 cent of GDP). African power and water utili- ties manage to collect about 90 percent of the 0 bills owed to them by their customers, short of ion e s e n r er ge s ds te ad ay ag ag tio at a best practice of 100 percent. Again, although oa ra wa at ilw ro er er iga lw ve lr er av av ain ra n la ra irr ra en ba water utilities perform worse than power utili- al all ru ru ra m rg ur ur er ru we nr ov no ties at the enterprise level, the financial value po of the losses is much greater for power. In Source: Briceño-Garmendia, Smits, and Foster 2008. many African countries, public institutions are among the worst offenders in failing to pay for utility services. The undercollection of On average, 30 percent of African infra- fuel levies for road sector maintenance is also structure assets need rehabilitation (figure 2.3). an issue, although the absolute values for this Although documenting the exact magnitude inefficiency are smaller than expected. of undermaintenance is difficult, the share of today's assets in need of rehabilitation provides Undermaintenance. Deferring maintenance a good indicator of past neglect. In general, expenditures is perhaps the most perverse ineffi- the state of rural infrastructure is substantially ciency and the hardest to quantify. Given the pre- worse than the rest, with 35 percent of assets in carious financing position of the infrastructure need of rehabilitation, compared with 25 per- sectors, cutting back on maintenance is often cent elsewhere and 40 percent of roads. Wide the only way to make ends meet, but spending differences exist across countries. In the best too little on maintenance is a false economy. cases (Burkina Faso and South Africa), little Rehabilitating or replacing poorly maintained more than 10 percent of assets need rehabilita- assets is much more costly than keeping them up tion, and in the worst cases (the Democratic with sound preventive maintenance. Moreover, Republic of Congo, Nigeria, Rwanda, and consumers end up suffering as service quality Uganda), more than 40 percent do. gradually declines. Indeed, not providing main- For roads alone, undermaintenance over tenance and replacement investment is the most time leads to additional capital spending costly way of financing today's operations. of $1.4 billion a year (0.2 percent of GDP). 74 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Although undermaintenance affects all infra- transferring resources from areas that seem to structure sectors, only for roads were suffi- be overfunded to areas that are clearly under- cient data available to quantify the cost. Every funded. The next-largest potential gain of $1 that goes unspent on road maintenance $4.7 billion a year would come from raising leads to a $4 liability to the economy (Nogales user charges for infrastructure services. Again, 2009). Therefore, capital spending on roads is better pricing of power produces the greatest much higher than it would otherwise need to dividends. Finally, raising budget execution be--with continual reconstructing of the same ratios through improvements in the public assets rather than creating new ones. The vast expenditure framework could capture an addi- majority of Sub-Saharan African countries do tional $2 billion a year. not cover road maintenance costs; more than Addressing some of the operational defi- half of the countries have shortfalls of over ciencies may require substantial investments 40 percent of maintenance needs. However, in network rehabilitation or system upgrades. institutions seem to have an important role to Reallocating resources, raising user charges, and play. Countries with well-designed second- reducing overemployment all carry significant generation road funds seem to do much bet- political costs, which complicate their imple- ter in meeting their maintenance needs (see mentation. Therefore, expecting that all these chapter 10 in this volume). efficiency gains could be fully captured is unre- alistic. Given the magnitude of the needs, cap- Closing the Efficiency Gap by turing only half of them would much improve Promoting Reforms the financing and the perspectives for new In sum, $17.4 billion could be captured financing in the African infrastructure sectors. through improvements in infrastructure man- Even if all these efficiency gains could be agement and institutions. The largest potential fully realized, a sizable funding gap would gains of $7.5 billion a year come from address- remain. Chapter 1 of this volume identified ing operating inefficiencies. Some of the most spending requirements of $93 billion a year to pressing and most rewarding would be resolv- address Africa's infrastructure needs. Setting ing undermaintenance of roads and increas- these requirements against the $45 billion of ing the efficiency of the power utilities. The existing spending directly traced to these needs second-largest potential gains of $3.3 billion and the $17 billion of potential efficiency gains a year come from improving the allocation of still leaves an annual infrastructure funding existing resources across sectors, essentially gap of $31 billion (table 2.8). Table 2.8 Finding Resources: The Efficiency Gap and the Funding Gap $ billions annually Cross-sector Item Electricity ICT Irrigation Transport WSS gain Total Infrastructure spending needs (40.8) (9.0) (3.4) (18.2) (21.9) n.a. (93.3) Existing spending 11.6 9.0 0.9 16.2 7.6 n.a. 45.3 Efficiency gap 6.0 1.3 0.1 3.8 2.9 3.3 17.4 Gain from raising capital execution 0.2 0.0 0.1 1.3 0.2 n.a. 1.9 Gain from eliminating operational inefficiencies 3.4 1.2 -- 1.9 1.0 n.a. 7.5 Gain from tariff cost recovery 2.3 -- -- 0.6 1.8 n.a. 4.7 Potential for reallocation n.a. n.a. n.a. n.a. n.a. 3.3 3.3 Funding gap (23.2) 1.3 (2.4) 1.9 (11.4) 3.3 (30.6) Source: Briceño-Garmendia, Smits, and Foster 2008. Note: ICT = information and communication technology; n.a. = not applicable; -- = not available; WSS = water supply and sanitation. Parentheses indicate negative values. Closing Africa's Funding Gap 75 Annual Funding Gap Although the infrastructure funding gap is primarily for capital investment, a shortfall Existing spending and potential efficiency also exists for O&M. About two-thirds of the gains can be netted from estimated spend- infrastructure funding gap relates to shortfalls ing needs to gauge the extent of the shortfall. in capital investment. All together, Africa needs The result is that Africa still faces an annual to increase infrastructure capital investment by funding gap of about $31 billion (5.1 percent 5 percent of its GDP (approximately $28 bil- of GDP). Over 70 percent of the infrastruc- lion annually); nonfragile low-income coun- ture funding gap is for energy, representing tries need to invest an additional 8 percent, a shortfall of $23 billion a year. The rest of and fragile states an additional 18 percent. The the gap is related to WSS, where about an remainder of the infrastructure funding gap additional $11 billion is needed annually to relates to O&M: low-income countries cover at meet the Millennium Development Goals most two-thirds of their O&M needs. (MDGs), and to a lesser extent irrigation, Closing the $31 billion infrastructure fund- which accounts for roughly $2 billion annu- ing gap is partly about raising additional funds ally of the funding gap. No funding gap was but also about possibly taking more time to found for ICT and transport, where, on the attain targets or using lower-cost technolo- contrary, close to $1 billion and $2 billion a gies. The remainder of this chapter evaluates year, respectively, could be available if effi- the potential for raising additional finance and ciencies were captured within each of these very generally explores policy adjustments to sectors (table 2.9). reduce the price tag and the burden of the About 60 percent of the funding gap relates financial gap. to low-income fragile and nonfragile coun- tries combined. The resource-rich countries generate one-fourth of the funding gap, and a How Much Additional Finance further 18 percent of the gap is attributable to Can Be Raised? middle-income countries. As a percentage of GDP, the burden of the shortfall for resource- Only a limited number of financing sources rich and middle-income countries is smallest are available, and the current global financial at 2­4 percent of GDP. Nonfragile low-income crisis is likely to affect them all adversely. First, countries face a shortfall of 9 percent of GDP, domestic public finance is the largest source of and fragile states face an insurmountable funding today, but it presents little scope for 25 percent. By far, the largest funding gaps an increase, except possibly in countries enjoy- relative to GDP are for energy and water in ing natural resource windfalls. Second, ODA to fragile states. African infrastructure has grown substantially Table 2.9 Funding Gaps, by Sector and Country Group Percentage of GDP $ billions annually Potential Potential for for Country type Electricity ICT Irrigation Transport WSS reallocation Total Electricity ICT Irrigation Transport WSS reallocation Total Middle income 3.9 (0.3) 0.0 (0.1) 0.0 (1.5) 2.0 10.7 (0.9) 0.1 (0.3) 0.0 (4.1) 5.5 Resource rich 2.0 0.2 0.8 (0.6) 1.7 (0.4) 3.7 4.5 0.5 1.8 (1.4) 3.7 (0.8) 8.2 Low-income nonfragile 4.2 (0.2) 0.6 (0.4) 4.7 (0.3) 8.6 4.7 (0.2) 0.7 (0.5) 5.2 (0.4) 9.5 Low-income fragile 7.1 1.9 0.1 5.3 10.2 0.0 24.6 2.7 0.7 0.0 2.0 3.9 0.0 9.4 Africa 3.6 (0.2) 0.4 (0.3) 1.8 (0.5) 4.8 23.2 (1.3) 2.4 (1.9) 11.4 (3.3) 30.6 Source: Africa Infrastructure Country Diagnostic. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals do not add because efficiency gains cannot be carried across country groups. ICT = information and communication technology; WSS = water supply and sanitation. 76 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION in recent years in line with political pledges, but with infrastructure investment bearing much this assistance could slow down if countercycli- of that, falling by almost 1.5 percent of GDP. In cal assistance is put in place. Third, non-OECD middle-income countries, budgetary spend- finance has been rising steeply, but its future ing increased by almost 4.1 percent of GDP, is now unclear. Fourth, private participation, but the effect on infrastructure spending was also very buoyant during Africa's recent growth almost negligible, and the additional resources upswing, will be particularly vulnerable to the went primarily to current social sector spend- downturn in global markets. Finally, local ing. Only in the low-income countries did the capital markets have so far contributed little to overall increases in budgetary expenditure infrastructure finance outside South Africa, but have some effect on infrastructure spend- they could eventually become more important ing. Even there, however, the effect was fairly in some of the region's larger economies. modest and confined to capital spending. The nonfragile low-income countries have allo- Little Scope for Raising More cated 30 percent of the budgetary increase Domestic Finance to infrastructure investments. The fragile A key question is the extent to which coun- states, despite seeing their overall budgetary tries may be willing to allocate additional fiscal expenditures increase by about 3.9 percent resources to infrastructure. In the run-up to the of GDP, have allocated only 6 percent of the current financial crisis, the fiscal situation in increase to infrastructure. Sub-Saharan Africa was favorable. Rapid eco- Compared with other developing regions, nomic growth averaging 4 percent a year from public financing capabilities in Sub-Saharan 2001 to 2005 translated into increased domestic Africa are characterized by weak tax revenue fiscal revenues of just over 3 percent of GDP on collection. Domestic revenue generation around average. In resource-rich countries, burgeoning 23 percent of GDP trails averages for other resource royalties added 7.7 percent of GDP to developing countries and is lowest for low- the public budget. In low-income countries, income countries (less than 15 percent of GDP substantial debt relief increased external grants a year). Despite the high growth rates in the last by almost 2 percent of GDP. decade, domestically raised revenues grew by less To what extent were the additional resources than 1.2 percent of GDP. This finding suggests available during the recent growth surge allo- that increasing domestic revenues above what cated to infrastructure? The answer is surpris- is currently raised would require undertaking ingly little (table 2.10). The most extreme case challenging institutional reforms to increase the is that of the resource-rich countries, particu- effectiveness of revenue collection and broaden larly Nigeria. Huge debt repayments more the tax base. Without such reforms, domestic than fully absorbed the fiscal windfalls in revenue generation will remain weak. these countries. As a result, budgetary spend- The borrowing capacity from domestic ing actually contracted by 3.7 percent of GDP, and external sources is also limited. Domestic Table 2.10 Net Change in Central Government Budgets, by Economic Use, 1995­2004 percentage of GDP Sub-Saharan Low-income Low-income Use Africa Middle income Resource rich nonfragile fragile Net expenditure budget 1.89 4.08 (3.73) 1.69 3.85 Current infrastructure spending as a share of expenditures 0.00 0.02 0.03 0.00 0.09 Capital infrastructure spending as a share of expenditures (0.14) 0.04 (1.46) 0.54 0.22 Source: Africa Infrastructure Country Diagnostic, adapted from Briceño-Garmendia, Smits, and Foster 2008. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Totals are extrapolations based on the 24-country sample as covered in AICD Phase 1. Closing Africa's Funding Gap 77 borrowing is often very expensive, with inter- example, Indonesia's total public investment est rates far exceeding those on concessional in infrastructure dropped from 6­7 percent of external loans. Particularly for the poorest GDP in 1995­97 to 2 percent in 2000. Given countries, the scarcity of private domestic sav- recent spending patterns, every reason exists ings means that public domestic borrowing to expect that, in Africa, changes in the over- tends to precipitate sharp increases in interest all budget envelope will affect infrastructure rates, building up a vicious circle. For many investment in a similar pro-cyclical manner. Sub-Saharan countries, the ratios of debt ser- vice to GDP are more than 6 percent. Official Development Assistance-- The global financial crisis can be expected to Sustaining the Scale-Up reduce fiscal receipts because of lower revenues For most of the 1990s and early 2000s, ODA to from taxes, royalties, and user charges. Africa is infrastructure in Sub-Saharan Africa remained not exempt from its impact. Growth projections steady at a meager $2 billion a year. The launch for the coming years have been revised down- of the Commission for Africa Report in 2004 ward from 5.1 to 3.5 percent, which will reduce was followed by the Group of Eight Gleneagles tax revenues and likely depress the demand and willingness to pay for infrastructure ser- vices. Commodity prices have fallen to levels of the early 2000s. The effect on royalty revenues, BOX 2.1 however, will depend on the saving regime in each country. A number of oil producers have been saving royalty revenues in excess of $60 a Does Deficit-Financed Public Investment in barrel, so the current downturn will affect sav- Infrastructure Pay for Itself? ings accounts more than budgets. Overall, this Underinvestment in infrastructure, health, and education during much adverse situation created by the global finan- of the 1990s has ignited a lively debate on whether some countries cial crisis will put substantial pressure on public could tolerate a larger public deficit if the additional resources were sector budgets. In addition, many African coun- invested in growth-enhancing sectors. The analysis undertaken by tries are devaluing their currency, reducing the the International Monetary Fund does not explicitly take into account purchasing power of domestic resources. the potential link between public investment and growth--only its short-term costs. Nevertheless, running a short-term deficit now may Based on recent global experience, fiscal help produce the growth that will balance the budget later. adjustment episodes tend to fall dispropor- By incorporating this long-run growth effect into the standard tionately on public investment--and infra- models used to assess fiscal sustainability, one can see whether taking structure in particular.2 Experience from a longer-term perspective would lead to a more favorable stance for earlier crises in East Asia and Latin America deficit-financed infrastructure. The results turn out to be very country indicates that infrastructure spending is par- specific, underscoring the difficulty of generalizing in this area. ticularly vulnerable to budget cutbacks dur- In Uganda, investment in infrastructure leads to higher output, ing crisis periods. Based on averages for eight but also--because of its relatively low productivity--worsens the Latin American countries, cuts in infrastruc- debt ratio. A better way to finance infrastructure may be to improve ture investment amounted to about 40 percent the existing capital stock by prioritizing O&M expenditure over new of the observed fiscal adjustment between the investments. Although increased public expenditure on health and education also leads to higher output, the effect is not as large as for early 1980s and late 1990s (Calderón and infrastructure. Servén 2004). This reduction was remark- In Senegal, by contrast, public investment in infrastructure does able because public infrastructure investment not seem to be as effective in boosting growth. Both O&M spending already represented less than 25 percent of on infrastructure and public investment in other sectors such as health overall public investment in Latin American and education seem to have a stronger effect on growth. However, countries. These infrastructure investment no matter how spending is allocated, it seems to worsen the debt- cuts were later identified as the underlying to-GDP ratio, reflecting the low productivity of public expenditure in problem holding back economic growth in the this case. whole region during the 2000s (box 2.1). Sim- Source: Estache and Muñoz 2008. ilar patterns were observed in East Asia dur- ing the financial crisis of the mid-1990s. For 78 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Summit in July 2005, where the Infrastructure ODA commitments were also set to increase Consortium for Africa was created to focus further before the crisis, but prospects no longer on scaling up donor finance to meet Africa's look so good. The three multilateral agencies-- infrastructure needs. The main bilateral and the African Development Bank, the European multilateral donors committed to double by Commission, and the World Bank--secured 2010 the (already higher) 2004 flows to reach record replenishments for their concessional $10 billion a year, about 1.6 percent of Africa's funding windows for the three to four years GDP at that time. Donors have so far lived up beginning in 2008. In principle, funding allo- to their promises, and ODA flows to African cations to African infrastructure totaling $5.2 infrastructure almost doubled from $4.1 bil- billion a year could come from the multilateral lion in 2004 to $8.1 billion in 2007. Close to agencies alone in the near future. In practice, three-quarters of ODA comes from multi- however, the crisis may divert multilateral lateral donors (African Development Bank, resources from infrastructure projects and European Community, and International toward emergency fiscal support. Bilateral sup- Development Association [IDA]), while Japan port, based on annual budget determinations, and the United States drive the doubling of may be more sensitive to the fiscal squeeze bilateral commitments. in OECD countries, and some decline can be A significant lag occurs between ODA com- anticipated. Historical trends suggest that ODA mitments and their disbursement, suggesting has tended to be pro-cyclical rather than coun- that disbursements should continue to increase tercyclical (IMF 2009; ODI 2009; UBS Invest- in the coming years. The commitments just ment Research 2008; World Economic Outlook reported are significantly higher than the 2008; and references cited therein). estimated ODA disbursements of $3.8 billion (table 2.11). This gap reflects the normal Non-OECD Financiers--Will lags associated with project implementation. Growth Continue? Because ODA is channeled through the gov- Non-OECD countries financed about $2.6 bil- ernment budget, the execution of funds faces lion of African infrastructure annually between some of the same problems affecting domes- 2001 and 2006 (table 2.12).3 This sum is not tically financed public investment, including far short of the volumes from ODA; however, procurement delays and low country capacity the focus of the finance is very different. Non- to execute funds. Divergences between donor OECD financiers have been active primarily in and country financial systems, as well as unpre- oil-exporting countries (Angola, Nigeria, and dictability in the release of funds, may further Sudan). The bulk of their resources have gone retard the disbursement of donor resources. to power and to transport. In the power sec- Bearing all this in mind, if all commitments tor, primarily hydroelectric schemes received up to 2007 are fully honored, ODA disburse- $1 billion per year, and in the transport sec- ments could be expected to rise significantly tor, railways received nearly $1 billion a year. (IMF 2009; World Economic Outlook 2008). For electricity, that amounts to 0.17 percent Table 2.11 Annualized ODA Investment Flows Percentage of GDP $ billions Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income 0.01 0.00 0.00 0.03 0.04 0.08 0.03 0.01 0.00 0.09 0.10 0.23 Resource rich 0.03 0.01 0.00 0.11 0.11 0.25 0.08 0.01 0.00 0.23 0.24 0.56 Low-income nonfragile 0.50 0.03 0.00 1.12 0.71 2.36 0.55 0.04 0.00 1.24 0.78 2.61 Low-income fragile 0.10 0.01 0.00 0.64 0.29 1.04 0.04 0.00 0.00 0.23 0.10 0.38 Africa 0.11 0.01 0.00 0.28 0.19 0.59 0.69 0.06 0.00 1.80 1.23 3.77 Source: Africa Infrastructure Country Diagnostic. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. Closing Africa's Funding Gap 79 Table 2.12 Historic Annualized Investment Flows from China, India, and Arab Countries Percentage of GDP $ billions Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income 0.00 0.01 0.00 0.01 0.00 0.02 0.00 0.02 0.00 0.02 0.01 0.05 Resource rich 0.33 0.06 0.00 0.34 0.04 0.76 0.74 0.13 0.00 0.75 0.08 1.69 Low-income nonfragile 0.12 0.15 0.00 0.22 0.05 0.54 0.13 0.17 0.00 0.24 0.05 0.59 Low-income fragile 0.58 0.07 0.00 0.11 0.06 0.82 0.21 0.03 0.00 0.04 0.02 0.30 Africa 0.17 0.05 0.00 0.16 0.03 0.41 1.08 0.34 0.00 1.06 0.16 2.64 Source: Africa Infrastructure Country Diagnostic, adapted from Foster and others 2008. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. of African GDP, significantly larger than the For the three major sources of external finance, 0.11 percent coming from ODA. significant complementarity exists, despite China's official economic assistance qua- some overlap. PPI seeks the most commercially drupled between 2001 and 2005, reaching lucrative opportunities in telecommunications. more than 35 Sub-Saharan countries. Most Non-OECD financiers focus on productive of the inflows have gone to resource-rich infrastructure (primarily power generation and countries, in some cases making use of barter railroads). Traditional ODA focuses on financ- arrangements under the "Angola mode." 4 This ing public goods (such as roads and water sup- type of South-South cooperation builds on ply) and plays a broader role in power system economic complementarities between China development and electrification. and Africa. China takes a strategic interest in A similar pattern of specialization emerges Africa's natural resource sector, while Africa geographically, with different countries rely- harnesses China's strengths in construction to ing to differing degrees on the various sources develop its economic infrastructure. of finance. The countries most heavily reliant India has become a significant financier on PPI are Kenya and Nigeria, supplemented of energy projects in Africa. India's financial by ODA in Kenya and by Chinese financing in assistance focused initially on export credits to Nigeria. The countries that rely predominantly facilitate the purchase of Indian goods. However, on non-OECD financiers are often oil produc- India has signaled a bold commitment to sup- ers (Angola, Gabon, Guinea, Mauritania, and port big infrastructure projects, predominantly Sudan). Most of the remaining countries rely in energy, with up to $1 billion in Nigeria primarily on traditional ODA (Burundi, Mali, (including a 9-million-ton per year refinery, Niger, Rwanda, and Tanzania). Other coun- a 200-megawatt power plant, and a 1,000- tries (the Democratic Republic of Congo and kilometer cross-country railway) and close to Guinea) draw on a mixture of OECD and non- $100 million a year in Sudan (for a 700-kilometer OECD sources. oil pipeline from Khartoum to Port Sudan and The implementation process for ODA and four 125-kilowatt power plants). non-OECD finance is completely different. A The Gulf States, through their various devel- key difference between Chinese finance and opment agencies, have been funding African ODA is that whereas the latter is channeled infrastructure for some time. Infrastructure through the government budget, the former projects of a smaller scale than those funded tends to be executed directly by China, often by the Chinese and Indian governments char- with associated imports of human resources. acterize their portfolio, with strong support to Although this approach raises significant chal- such countries as Mali, Mauritania, Senegal, lenges, it does at least offer the possibility of and Sudan. Resources from the Gulf States circumventing some of the capital budget have been distributed almost equally among execution problems typically associated with water, roads, and small energy projects. public investment. 80 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Non-OECD finance also raises concerns sectors, such as WSS, attracted almost no about sustainability. The non-OECD financiers private activity. The same is true of longer- from China, India, and the Gulf States follow term and higher-risk projects. Through 2004, sectors, countries, and circumstances aligned greenfield and small projects accounted for with their national business interests. They offer 70 percent of all PPI, while concessions and realistic financing options for power and trans- divestitures of incumbent utilities accounted port and for postconflict countries with natural for less than 10 percent. Greenfield transac- resources. However, nongovernmental organiza- tions, with no long-term risks and little or no tions are voicing concerns about the associated investment, are much more prominent than in social and environmental standards. Non-OECD other regions, and they tend to be small. financiers also provide investment finance Africa's resource-rich countries have been without associated support on the operational, capturing the largest volume of private partici- institutional, and policy sides, raising questions pation. Relative to their GDP, Africa's middle- about the sustainability of the new assets. income countries are not doing that well, while How the current economic downturn will low-income countries--even fragile states-- affect non-OECD finance is difficult to pre- are capturing flows worth well over 1 percent dict because of the relatively recent nature of of GDP. these capital inflows. Coming from fiscal and Since the mid-1990s, a shift has occurred royalty resources in their countries of origin, toward projects with longer horizons. Conces- they will likely suffer from budgetary cutbacks. sions and existing assets increased to 20 percent The downturn in global commodity prices of the private partnerships in infrastructure. may also affect the motivation for some of the Sectors other than ICT have increased; the most Chinese infrastructure finance linked to natu- important recorded transactions are in trans- ral resource development. port, such as the concessions in Sudan for the Juba Port ($30 million) and Uganda's Rift Val- Private Investors--over the Hill ley railways ($400 million concession). More- Since the late 1990s, private investment flows over, larger greenfield power projects, beyond to Sub-Saharan African infrastructure tripled, concessions and management contracts, are increasing from about $3 billion in 1997 to starting to emerge. $9.4 billion in 2006/07. That is about 1.5 per- Private capital flows, in particular, are likely cent of regional GDP for all sectors, more than to be affected by the global financial crisis. In recent ODA flows (0.6 percent of GDP, or $3.7 the aftermath of the Asian financial crisis, pri- billion a year) but still less than half of general vate participation in developing countries fell government spending (table 2.13). by about one-half over a period of five years, Close to two-thirds of cumulative private following its peak in 1997. Existing transac- commitments between 1990 and 2006 were in tions are also coming under stress as they ICT-related projects (Leigland and Butterfield encounter difficulties refinancing short- and 2006). Power was second. Socially challenging medium-term debt. Table 2.13 Annual Private Participation Investment Flows Percentage of GDP $ billions Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total Middle income 0.00 0.60 -- 0.16 0.00 0.76 0.01 1.63 -- 0.44 0.00 2.08 Resource rich 0.13 1.13 -- 0.21 0.00 1.47 0.28 2.52 -- 0.47 0.01 3.28 Low-income nonfragile 0.15 1.19 -- 0.12 0.00 1.46 0.16 1.32 -- 0.13 0.00 1.61 Low-income fragile 0.02 0.72 -- 0.04 0.00 0.78 0.01 0.26 -- 0.01 0.00 0.28 Africa 0.07 0.89 -- 0.16 0.00 1.12 0.46 5.72 -- 1.05 0.01 7.24 Source: Adapted from PPIAF 2008. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation. -- Not available. Closing Africa's Funding Gap 81 Local Sources of Finance--a Possibility local financing was in corporate bonds, all to in the Medium Term finance transport. Local capital markets are a major source of Only 10 percent of outstanding bank loans infrastructure finance in South Africa, but are for financing infrastructure investments. At not yet elsewhere. Local infrastructure finance about $5 billion, this sum is a little less than the consists primarily of commercial bank lend- total for Malaysia alone. ing, some corporate bond and stock exchange However, a recent trend indicates new issu- issues, and a nascent entry of institutional ers (particularly of corporate bonds) are com- investors. Because the scale of local financ- ing to market in several countries, in some cases ing in South Africa and its advanced state of with a debut issue. More than half (52 percent) evolution are so far ahead of those elsewhere, of the corporate bonds listed on the markets at attention here focuses on prospects elsewhere year-end 2006 were by infrastructure compa- in the region. nies. The share of corporate bonds outstand- Outside South Africa, the stock of outstand- ing at year-end 2006 that had been issued to ing local infrastructure finance amounts to finance infrastructure exceeded half in 7 of the $13.5 billion (table 2.14). This figure comprises 11 countries with bond markets reporting these transport, the first-ranking sector of all local data. West Africa's regional exchange, the BRVM infrastructure financing, attracting 47 percent (Bourse Régionale des Valeurs Mobilières), had of the total, followed by ICT at 32 percent.5 the highest share of issues financing infrastruc- The low-income nonfragile countries were ture (more than 90 percent). The amount of the destination for 55 percent of all local infra- financing is still small, however. structure financing identified in this study. The Local financial markets remain underdevel- two low-income fragile countries (Côte d'Ivoire oped, shallow, and small. Long-term financing and the Democratic Republic of Congo) with maturities commensurate with infra- attracted just 3.5 percent ($474 million), nearly structure projects is scarce.6 The capacity of three-quarters of it in bank financing and the local banking systems remains too small and remainder in equity issues by companies in constrained by structural impediments to Côte d'Ivoire. For the resource-rich countries, finance infrastructure. Most countries' banks the $4.9 billion in local infrastructure financ- have significant asset-liability maturity mis- ing was a nearly equal mix of bank and equity matches for infrastructure financing. Bank financing. For the three middle-income coun- deposits and other liabilities still have largely tries, more than half of the $544 million in short-term tenors. More potential may exist Table 2.14 Outstanding Financing Stock for Infrastructure, as of 2006 $ millions % of total Outstanding financing outstanding for infrastructure WSS Electricity ICT Transport Public works Total stock Middle income (excluding South Africa) -- 82.0 -- 440.7 21.3 544.0 4.0 Resource rich 1.7 1,097.6 2,303.9 1,459.1 46.8 4,909.1 36.5 Low-income nonfragile -- 1,496.7 1,984.5 4,065.5 4.4 7,551.0 56.1 Low-income fragile -- 63.0 53.4 346.3 -- 462.7 3.4 Total 1.7 2,739.3 4,341.8 6,311.7 72.4 13,466.9 Share on total outstanding stock (%) 0.01 20.34 32.24 46.87 0.54 100.0 Source: Adapted from Irving and Manroth 2009. Note: Based on annualized averages for 2001­06. Averages weighted by country GDP. Figures are extrapolations based on the 18-country sample covered in AICD Phase 1. Totals may not add exactly because of rounding errors. Stock includes bank loans, government bonds, corporate bonds, and equity issues. Stock level reported under "Transport" may be an overestimate because many countries report this category together with elements of communications and storage. Based on data from the following 18 countries--middle income: Cape Verde, Lesotho, and Namibia; resource rich: Cameroon, Chad, and Nigeria; low income: Burkina Faso, the Democratic Republic of Congo, Ethiopia, Ghana, Madagascar, Malawi, Mozambique, Niger, Rwanda, Tanzania, Uganda, and Zambia. ICT = information and communication technology; WSS = water supply and sanitation. -- Not available. 82 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION for syndicated lending with local bank par- (figure 2.4). For public funds, raising taxes is ticipation--though the increase in new loans not a costless exercise. Each dollar raised and over 2000­06 occurred in a favorable external spent by a Sub-Saharan African government financing environment. has a social value premium (or marginal cost Harnessing the significant potential for of public funds) of almost 20 percent. That local capital markets to finance infrastructure, premium captures the incidence of that tax particularly local bond markets, is contingent on the society's welfare (caused by changes on their further development. It is also contin- in consumption patterns and administrative gent on further reforms, including those that costs, among other things).9 To allow ready would deepen the local institutional inves- comparisons across financing sources, this tor base. Well-functioning and appropriately study standardized the financial terms as the regulated local institutional investors (pen- present value of a dollar raised through each sion funds and insurance companies) would of the different sources. In doing so, it recog- be natural sources of long-term financing for nized that all loans must ultimately be repaid infrastructure because their liabilities would with tax dollars, each of which attracts the better match the longer terms of infrastructure 20 percent cost premium. projects. Private pension providers have begun Wide variation exists in lending terms. to emerge with a shift from defined-benefit to The most concessional IDA loans charge zero defined-contribution schemes, viewed as less interest (0.75 percent service charge) with costly, more transparent, and easier to man- 10 years of grace. India, China, and the Gulf age. Moreover, local institutional investors are States, respectively, charge 4 percent, 3.6 per- taking a more diversified portfolio approach to cent, and 1.5 percent interest and grant four asset allocation. years of grace.10 Regional integration of financial mar- The cost of non-OECD finance is some- kets could achieve greater scale and liquidity. where between that of public funds and ODA. More cross-border intraregional listings--of The subsidy factor for Indian and Chinese both corporate bonds and equity issues--and funds is about 25 percent and for the Arab more cross-border intraregional investment funds, 50 percent. ODA typically provides a (particularly by local institutional investors) subsidy factor of 60 percent, rising to 75 per- could help overcome national capital markets' cent for IDA resources. In addition to the cost impediments of small size, illiquidity, and of capital, the different sources of finance dif- inadequate market infrastructure. They could fer in the transaction costs associated with also facilitate the ability of companies and gov- their use, which may offset or accentuate some ernments to raise financing for infrastructure.7 of the differences. So far, this intraregional approach to rais- ing infrastructure financing remains largely untapped.8 Most Promising Ways to The African banking system did not feel the Increase Funds effects of the global financial crisis at first, but the crisis is slowly but surely affecting finan- Given this setting, what are the best ways of cial systems around the region, adding to the increasing the availability of funds for infra- already enormous challenge of developing structure development? The place to start is local financial markets. clearly to get the most from existing budget envelopes, which can provide up to $17.4 bil- lion a year of additional resources internally. Costs of Capital from Different Beyond that, a substantial funding gap still Sources remains. Before the financial crisis, the pros- pects for reducing--if not closing--this gap The various sources of infrastructure finance were reasonably good. Resource royalties were reviewed in the previous sections differ at record highs, and all sources of external greatly in their associated cost of capital finance were buoyant and promising further Closing Africa's Funding Gap 83 growth. With the onset of the global financial Figure 2.4 Costs of Capital by Funding Source crisis, that situation has changed significantly and in ways that are not yet entirely foresee- public funds 1.17 able. The possibility exists across the board that private sector 1.1 all sources of infrastructure finance in Africa India 0.91 may fall rather than increase, further widening China 0.87 the funding gap. Only resource-rich countries Gulf States 0.65 have the possibility of using natural resource 0.51 official development assistance savings accounts to provide a source of financ- International Development Association 0.33 ing for infrastructure, but only if macroeco- nomic conditions allow. One of the few things 0 0.20 0.40 0.60 0.80 1.00 1.20 cost of raising $1 of financing (US$) that could reverse this overall situation would be the agreement upon a major stimulus pack- Sources: Average marginal cost of public funds: as estimated by Warlters and Auriol 2005; cost of equity for private sector: as in Estache and Pinglo 2004 and Sirtaine and others 2005; authors' age for Africa by the international community, calculations. with a focus on infrastructure as part of the effort to rekindle economic growth and safe- guard employment. conclusion assumes they have first fully cap- tured efficiency gains. Without such efficiency gains, the targets could not be met even over 30 What Else Can Be Done? years without increasing spending above cur- rent levels (figure 2.5, panel b). Most of the low-income countries, and in Low-income nonfragile and resource-rich particular the fragile states, face a substantial countries would need to delay an additional funding gap even if all the existing sources decade to meet targets with existing spending of funds--including efficiency gains--are levels. By spreading the investment needs over tapped. What other options do these countries 20 rather than 10 years, these countries could have? Realistically, they need either to defer the achieve the proposed targets within existing attainment of the infrastructure targets pro- spending envelopes (figure 2.5, panel a). Again, posed here or to try to achieve them by using this outcome would be possible only if efficiency lower-cost technologies. gains are fully exploited. Otherwise, they would need more than 30 years to reach the target with Taking More Time existing resources (figure 2.5, panel b). The investment needs presented in this book Low-income fragile states would need to are based on the objective of addressing Africa's delay by more than two decades to meet infra- infrastructure backlog within 10 years. To meet structure targets within existing spending lev- this target, middle-income, resource-rich, and els. By spreading the investment needs over 30 low-income nonfragile states would need to rather than 10 years, low-income fragile states increase their existing infrastructure spending could achieve the proposed targets within by 50 to 100 percent, while low-income fragile existing spending envelopes (figure 2.5, panel states would need to increase their infrastructure a). However, without efficiency gains, these spending by an impossible 350 percent. Extend- countries would take much longer than 30 ing the time horizon for the achievement of years to meet the associated targets or alter- these goals should make the targets more afford- natively would still need to double their exist- able. But how long a delay would be needed to ing spending to reach the target in 30 years make the infrastructure targets attainable with- (figure 2.5, panel b). out increasing existing spending envelopes? By delaying only three years, spreading Using Lower-Cost Technologies the investment needs over 13 rather than 10 Many possible alternative technological solu- years, middle-income countries could achieve tions exist for meeting a given infrastructure the proposed targets within existing spending target, and each offers a particular combina- envelopes (figure 2.5, panel a). However, this tion of financial cost and quality of service. 84 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 2.5 Spreading Spending over Time a. Resource envelope plus potential efficiency gains b. Existing resource envelope (% deviation from current envelope) (% deviation from current envelope) 500 500 variation in resources needed 450 450 variation in resources needed 400 400 350 350 300 300 250 250 200 200 150 150 100 100 50 50 0 0 ­50 ­50 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 number of years needed to attain infrastructure targets number of years needed to attain infrastructure targets fragile low-income countries nonfragile low-income countries middle-income countries resource-rich countries Source: Authors' calculations. Where budgets are constrained, policy mak- unimproved latrines, remains the same as it is ers face a choice between providing a high today (see chapters 16 and 17 in this volume). level of service to a few people or a lower level Thus, as population grows, the number of peo- of service to a broader cross-section of the ple accessing high-level services will be larger population. Critical trade-offs must be con- in absolute terms. If instead, all additional sidered; thus, one cannot jump to the conclu- people served use cheaper solutions, such as sion that a high level of service is always in a standposts and improved latrines in urban country's best interest. The extent to which areas or boreholes and unimproved latrines cost-saving technologies are available var- in rural areas, the overall cost of meeting the ies considerably across sectors. Two of the MDGs would fall by 30 percent. clearest cases are water and roads, which are discussed in detail next. Unfortunately, the Using Alternative Technologies in Roads. In the power sector (which has by far the largest case of roads, the cost of reaching regional and associated investment tag) does not present national connectivity targets can be reduced many technological alternatives for reducing by 30 percent by adopting lower standards for the cost of electricity generation. trunk roads. Road connectivity targets can be attained by using different engineering stan- Using Alternative Technologies in WSS. In the dards. The scenario considered here is one in case of water and sanitation, the cost of achiev- which regional and national connectivity are ing the MDGs drops by 30 percent with greater achieved by a good-condition asphalt road reliance on lower-cost technologies. The network with at least two lanes for regional MDGs can be achieved using either higher- and at least one lane for national connectivity. end solutions, such as piped water and septic The same connectivity could be achieved at a tanks, or cheaper solutions, such as standposts cost reduction of 30 percent if a single-surface- and improved latrines. The scenario consid- treatment road in fair condition is substituted ered here is one where the MDGs are met by for an asphalt road in good condition. preserving the prevalent mix of high-end and Notes lower-end technologies. That is, the relative The authors of this chapter are Cecilia Briceño- share of the population enjoying access to a Garmendia and Nataliya Pushak, who drew direct water connection, sewers, or a septic on background material and contributions tank--all regarded as high-level services-- from William Butterfield, Chuan Chen, Vivien compared to the share of people with access Foster, Jacqueline Irving, Astrid Manroth, Afua to lower-end solutions, such as standposts and Sarkodie, and Karlis Smits. Closing Africa's Funding Gap 85 1. In particular, maintenance is essential to har- Calderón, César, and Luis Servén. 2004. "Trends ness the economic returns of capital, but good- in Infrastructure in Latin America, 1980­2001." quality data on how much of current expendi- Policy Research Working Paper 3401, World tures go to maintenance is hard to track. Bank, Washington, DC. 2. Servén (2005) and Hicks (1991) summarize the De Haan, Jakob, Jan Sturm, and Bernd Sikken. facts on Latin American and other developing 1996. "Government Capital Formation: Explain- countries. For industrialized countries, see also ing the Decline." Weltwirtschaftliches Archiv Roubini and others (1989); De Haan, Sturm, 132 (1): 55­74. and Sikken (1996) document the experience of Estache, Antonio, and Rafael Muñoz. 2008. "Build- industrialized countries. ing Sector Concerns into Macro-Economic 3. This section draws heavily on Foster and others Financial Programming: Lessons from Senegal (2008). and Uganda." Working Paper 6, Africa Infra- 4. Essentially, the Angola mode was devised to structure Country Diagnostic, World Bank, Washington, DC. enable African nations to pay for infrastructure with natural resources. In a single transaction, Estache, Antonio, and Maria Elena Pinglo. 2004. China bundles development-type assistance "Are Returns to Private Infrastructure in Developing Countries Consistent with Risks with commercial-type trade finance. A Chi- since the Asian Crisis?" Policy Research Working nese resource company makes repayments in Paper 3373. World Bank, Washington, DC. exchange for the oil or mineral rights. The China Export-Import Bank acts as a broker, receiving Foster, Vivien, William Butterfield, Chuan Chen, and Nataliya Pushak. 2008. Building Bridges: money for the sale and paying the contractor China's Growing Role as Infrastructure Finan- for providing the infrastructure. This arrange- cier for Sub-Saharan Africa. Trends and Policy ment safeguards against currency inconvertibil- Options no. 5. Washington, DC: Public-Private ity, political instability, and expropriation. Infrastructure Advisory Facility, World Bank. 5. Data are as of year-end 2006, or most recent Hicks, Norman. 1991. "Expenditure Reductions available, for the sampled countries, excluding in Developing Countries Revisited." Journal of South Africa. International Development 3 (1): 29­37. 6. Because South Africa's financial markets are so Irving, Jacqueline, and Astrid Manroth. 2009. much more developed than any of those of the "Local Sources of Financing for Infrastructure other 23 focus countries, this section excludes in Africa: A Cross-Country Analysis." Policy South Africa. Research Working Paper 4878, World Bank, 7. One new initiative is the Pan-African Infra- Washington, DC. structure Development Fund, a 15-year regional IMF (International Monetary Fund). 2009. The State fund for raising finance for commercially viable of Public Finances: Outlook and Medium-Term Pol- infrastructure projects in Africa, which raised icies after the 2008 Crisis. Washington, DC: IMF. $625 million in its first close in 2007, including Leigland, James, and William Butterfield. 2006. funds from Ghanaian and South African insti- "Reform, Private Capital Needed to Develop tutional investors. Infrastructure in Africa: Problems and Pros- 8. In addition, the lack of a benchmark yield curve in pects for Private Participation." Gridlines, Note the vast majority of those African countries that 8 (May), Public-Private Infrastructure Advisory have an organized bond market has limited cor- Facility, World Bank, Washington, DC. porate bond issuance, as has the general absence Nogales, Alberto. 2009. "The Cost of Postponing of credit ratings agencies and a lack of awareness Roads Maintenance." World Bank, Washington, among prospective issuers as well as investors. DC. 9. The marginal cost of public funds measures ODI (Overseas Development Institute). 2009. A the change in welfare associated with raising Development Charter for the G-20. London: ODI. an additional unit of tax revenue (Warlters and Auriol 2005). PPIAF (Public-Private Infrastructure Advisory Facility). 2008. Private Participation in Infrastruc- 10. See Foster and others (2008) for further details. ture Project Database, http://ppi.worldbank .org/. Pritchett, Lant. 1996. "Mind Your P's and Q's. The References Cost of Public Investment Is Not the Value of Briceño-Garmendia, Cecilia, Karlis Smits, and Vivien Public Capital." Policy Research Working Paper Foster. 2008. "Fiscal Costs of Infrastructure in 1660, World Bank, Washington, DC. Sub-Saharan Africa." Africa Infrastructure Country Rajkumar, Andrew, and Vinaya Swaroop. 2002. "Pub- Diagnostic. Washington, DC: World Bank. lic Spending and Outcomes: Does Governance 86 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Matter?" Policy Research Working Paper 2840, Sirtaine, Sophie, Maria Elena Pinglo, J. Luis World Bank, Washington, DC. Guasch, and Viven Foster. 2005. How Profitable Reinikka, Ritva, and Jakob Svensson. 2002. Are Infrastructure Concessions in Latin America? "Explaining Leakage of Public Funds." Empirical Evidence and Regulatory Implications. Discussion Paper 3227, Centre for Economic Trends and Policy Options no. 2. Washington, Policy Research, London, U.K. DC: Public-Private Infrastructure Advisory Facility, World Bank. ------. 2003. "The Power of Information: Evi- dence from a Newspaper Campaign to Reduce UBS Investment Research. 2008. "Global Capture." Policy Research Working Paper 3239, Economic Perspectives: The Global Impact World Bank, Washington, DC. of Fiscal Policy." Roubini, Nouriel, Jeffrey Sachs, Seppo Honkapohja, Warlters, Michael, and Emmanuelle Auriol. 2005. and Daniel Cohen. 1989. "Government Spending "The Marginal Cost of Public Funds in Africa." and Budget Deficits in the Industrial Countries." Policy Research Working Paper 3679, World Economic Policy 8 (4): 99­132. Bank, Washington, DC. Servén, Luis. 2005. "Fiscal Discipline, Public World Economic Outlook. 2008. "Estimating the Investment and Growth." World Bank, Size of the European Stimulus Packages for Washington, DC. 2009." Chapter 3 Dealing with Poverty and Inequality C overage of modern infrastructure ser- In the absence of modern infrastructure vices has been stagnant since the mid- services, the next best option would be to 1990s and remains strongly skewed reach households with lower-cost, second-best toward more affluent households. In urban solutions, such as standposts, improved areas, those who fail to hook up to nearby net- latrines, or street lighting. However, the preva- works form a significant share of the unserved lence of these second-best solutions is surpris- population, suggesting that demand-side barri- ingly low in Africa, and those that exist tend ers are also at work. In these circumstances, the to cater more to the higher-income groups key questions are whether African households than to the middle of the income distribu- can afford to pay for modern infrastructure tion. The majority of Africans resort instead services, and if not, whether African govern- to traditional alternatives, such as wells, unim- ments can afford to subsidize them. proved latrines, or kerosene lamps. Significant A subsistence power or water bill ranges challenges exist in increasing the coverage of between $2 and $8 a month. This cost is well second-best alternatives, particularly because within the affordable range for most house- their public good nature makes some of these holds in Africa's middle-income countries technologies more difficult for service provid- and for the more affluent segments that cur- ers to operate on a commercial basis. rently enjoy access to utilities in low-income The business-as-usual approach to expand- countries. However, affordability would defi- ing service coverage in Africa does not appear nitely become an issue for most people in the to be working. Turning this situation around poorest low-income countries should access be will require rethinking the approach to service broadened. expansion in four ways. First, coverage expan- African governments already spend $4.1 sion is not only about network rollout, but also billion a year (0.7 percent of GDP) on power about a need to address demand-side barriers and water subsidies that benefit mainly a small such as high connection charges or legal tenure. group of affluent customers. Expanding these Second, cost recovery for household services levels of subsidy to the entire population would needs to be improved to ensure that utilities be fiscally unsustainable for most countries. have the financial basis to invest in service 87 88 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION expansion. Third, rethinking the design of util- Figure 3.1 Coverage of Network Infrastructure ity subsidies to better target them and to accel- Services, 1990­2005 erate service expansion is desirable. Fourth, 35 any approach must consider the actual level of 30 % of households service that households can afford to pay and 25 that governments can afford to subsidize and 20 put greater emphasis on second-best alterna- 15 tives to modern infrastructure services. 10 5 0 1990­95 1996­2000 2001­05 piped water electricity Access to Modern Infrastructure flush toilets landline telephone Services--Stagnant and Inequitable Source: Banerjee, Wodon, and others 2008. Coverage of modern infrastructure services in Africa is very low by global standards (Estache percentage of the population added annually to and Wodon 2007). Coverage of electricity is the coverage of modern infrastructure services about 20 percent in Africa; 33 percent in South is a good measure of the intensity of effort in Asia; and more than 85 percent in East Asia, service expansion--and it differs dramatically Latin America, and the Middle East. Cover- across services (figure 3.2). Less than 0.5 percent age of piped water is 12 percent in Africa, 21 of the population is added each year to the net- percent in South Asia, and more than 35 per- work of piped water and flush toilets, whereas cent in other developing regions. Coverage of about 1.5 percent is added to that of electricity flush toilets is 6 percent in Africa, 34 percent in and cellular telephone services. For water and South Asia, and more than 30 percent in other sanitation, the rate of expansion of alternative developing regions. Africa's telecommunica- services such as latrines, standposts, and bore- tions coverage, however, compares favorably holes is significantly faster than that of piped with South Asia's and is not so far behind that water and flush toilets. These regional averages of other developing regions. Africa's low cov- mask outstanding performances by individual erage of infrastructure services in part reflects countries. For piped water, Benin, Côte d'Ivoire, its relatively low urbanization rates, because and Senegal reach an additional 1.5­2.0 percent urban agglomeration greatly facilitates the of their population each year, compared with extension of infrastructure networks.1 less than 0.1 percent for Africa as a whole. Household surveys show only modest gains Universal access to modern infrastructure in access to modern infrastructure services over services lies at least 50 years in the future for the period 1990­2005 (figure 3.1). This stag- most countries. Projecting current rates of nant overall picture masks two divergent trends. service expansion forward and taking into Service coverage in rural areas has seen modest account anticipated demographic growth, one improvements, whereas that in urban areas has can estimate the year countries will reach uni- actually declined. For example, urban cover- versal access to each of the modern infrastruc- age of piped water fell from 50 percent in the ture services. The results are sobering. Under early 1990s to 39 percent in the early 2000s, and business as usual, less than 20 percent of Sub- urban coverage of flush toilets from 32 percent Saharan African countries will reach universal to 27 percent. Although many new connections access for piped water by 2050, and less than are being made in urban areas, declining urban 45 percent will reach universal access to elec- coverage largely reflects the inability of service tricity (figure 3.3). In about one-third of coun- providers to keep pace with urban population tries, universal access to piped water and flush growth of 3.6 percent a year. toilets will not be reached in this century. The pace of service expansion differs dra- Coverage varies dramatically across house- matically across sectors and countries. The holds with different budget levels (figure 3.4). Dealing with Poverty and Inequality 89 Among the poorest 60 percent of the population, Figure 3.2 Expansion of Access to Infrastructure Services Each Year, coverage of almost all modern infrastructure Mid-1990s to Mid-2000s services is well below 10 percent.Conversely, 2.0 the vast majority of households with coverage % of population 1.5 belong to the more affluent 40 percent of the population. In most countries, inequality of 1.0 access has increased over time, suggesting that 0.5 most new connections have gone to more afflu- ent households (Diallo and Wodon 2005). This 0 r t es ts s les ty s es situation is not entirely surprising, given that, te ile ne ne os ici rin rin wa ho to ho ho dp ctr lat lat re sh lep ll p ed n ele even among households with greater purchas- bo sta ed al flu pip te ce ion ov ne pr dit ing power, coverage is far from universal, and dli im tra lan well under 50 percent in most cases. Relative to the other modern infrastructure services, Source: Banerjee, Wodon, and others 2008. electricity coverage is somewhat higher across the spectrum. Figure 3.3 Projected Universal Access for Piped Water Low coverage rates can reflect both sup- for Sub-Saharan African Countries, 2050 and Beyond ply and demand factors. On the one hand, 100 the household may be physically distant from % of countries 80 an infrastructure network (and thus face an 60 absence of supply). On the other hand, the 40 household may choose not to connect to a 20 network even when it is nearby (and thus 0 piped water flush toilets electricity landline express a lack of demand). Understanding telephones this difference is important because the policy by 2050 2050­2100 2100­2200 after 2200 implications differ radically. By exploiting the spatial distribution of household survey Source: Banerjee, Wodon, and others 2008. samples in urban areas, one can quantify the relative importance of these supply and Figure 3.4 Coverage of Modern Infrastructure demand factors in accounting for low ser- Services, by Budget Quintile vice coverage. Using this approach, one can 100 distinguish between the percentage of popula- % of households tion that has access to the service (those living 80 physically close to the infrastructure) and the 60 percentage of the population that hooks up to 40 the service when it is available. 20 Lack of coverage for urban electricity 0 supply is equally about demand and supply Q1 Q2 Q3 Q4 Q5 factors. The power infrastructure for elec- budget quintile tricity is physically close to 93 percent of piped water electricity flush toilets the urban population, but only 75 percent landline cell phones garbage telephones collected of those with access actually hook up to the service. This means that half of the popu- Source: Banerjee, Wodon, and others 2008. lation lacking coverage live close to power Note: The data are the latest available as of 2006. infrastructure. One can often observe this phenomenon in African cities, where informal settlements flanking major road corridors lack (table 3.1). The physical extent of the piped power service even though distribution lines water network is more limited, reaching only run overhead. 73 percent of the urban population, and Overall, the coverage gap for piped water hookup rates for those in proximity are only is primarily attributable to supply factors 48 percent. In general, the role of demand 90 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 3.1 Understanding Coverage of Infrastructure Services: The Role of The tenure status of households may also Supply and Demand Factors significantly impede hookup to modern infra- percent (population-weighted average) structure services. A study of slum households Decomposition of coverage Unserved due to factors in Dakar and Nairobi finds that coverage of Infrastructure Access Hookup Coverage Supply Demand piped water and electricity is more than twice Piped water as high among owner-occupiers as among ten- Low-income countries 68 42 31 86 14 ants (Gulyani, Talukdar, and Jack 2008). Even Middle-income countries 91 74 69 64 36 among owner-occupiers, lack of formal legal titles can affect hookup to services. Overall 73 48 38 81 19 Electricity Low-income countries 93 73 69 50 50 Affordability of Modern Middle-income countries 95 86 81 39 61 Infrastructure Services-- Overall 93 75 71 48 52 Subsidizing the Better Off Source: Banerjee, Wodon, and others 2008. Note: Access is defined as the percentage of the population that lives physically close to infrastructure. African households exist on very limited Hookup is defined as the percentage of the population that connects to infrastructure when it is avail- able. Coverage is defined as the percentage of the population that has the infrastructure service; it household budgets. The average African house- is essentially the product of access and hookup.. In calculating the proportion of gap attributable to demand and supply factors, one considers the hookup rate of the top budget quintile in each geo- hold of five persons has a monthly budget of graphical area to be an upper bound on potential hookup, absent demand-side constraints. less than $180, ranging from about $60 in the poorest quintile to $340 in the richest quintile (table 3.2). Thus, purchasing power--even in factors is higher in middle-income countries Africa's most affluent households--is modest than in low-income countries, reflecting the in absolute terms. Across the spectrum, house- fact that infrastructure networks are more hold budgets in middle-income countries are highly developed in the former and have a roughly twice those in low-income countries. broader geographical reach. Most African households spend more than It may appear paradoxical that households half their modest budgets on food, with little left do not universally take up connections to over for other items. Spending on infrastructure modern infrastructure services as networks services (including utilities, energy, and trans- become physically available. Clear economic port) averages about 7 percent of a household's reasons exist, however, why this might be so. budget, though this can be 15­25 percent in In some cases, households may have access to some countries and represents overall a signifi- cheaper substitutes, such as boreholes. More cant share of the nonfood budget. As household substitutes are available for piped water than budgets increase, infrastructure services absorb for electricity, which may explain the much a growing share, rising from less than 4 percent lower hookup rates for the former. In other among the poorest quintile to more than 8 per- cases, utility connection charges are set pro- cent among the richest (figure 3.5). In terms hibitively high for low-income households. For of absolute expenditure, this difference is even example, 60 percent of the water utilities sur- more pronounced: whereas households in the veyed for this study apply connection charges poorest quintile spend, on average, no more in excess of $100. Charges range from about than $2 per month on all infrastructure services, $6 in the Upper Nile in Sudan to more than households in the richest quintile spend almost $240 in Côte d'Ivoire, Mozambique, Niger, and $40 per month. South Africa. The average connection charge Given such low household budgets, a key across the region is 28 percent of gross national question is whether households can afford to income (GNI) per capita. For Niger, the charge pay for modern infrastructure services. One is more than 100 percent of GNI per capita. measure of affordability is nonpayment for Similarly, the five water utilities in Mozambique infrastructure services. Nonpayment directly charge more than 75 percent of GNI per capita. limits the ability of utilities and service provid- These comparisons illustrate how high connec- ers to expand networks and improve services by tion charges present a barrier to affordability. undermining their financial strength. Based on Dealing with Poverty and Inequality 91 Table 3.2 Monthly Household Budgets 2002 $ National Poorest Second Third Fourth Richest Income group average quintile quintile quintile quintile quintile Overall 177 59 97 128 169 340 Low-income countries 139 53 80 103 135 258 Middle-income countries 300 79 155 181 282 609 Source: Banerjee, Wodon, and others 2008. Figure 3.5 Share of Household Budgets Spent on Figure 3.6 Population with Service Connections Who Infrastructure Services, by Budget Quintile Do Not Pay for Service 9 40 100 8 35 90 expenditure (US$/month) 7 30 80 budget share (%) % of population 6 70 25 5 60 20 4 50 15 3 40 2 10 30 1 5 20 0 0 10 poorest second third fourth richest 0 quintile poorest second third fourth richest budget share (left axis) expenditure (right axis) quintile electricity piped water Source: Banerjee, Wodon, and others 2008. Source: Banerjee, Wodon, and others 2008. household surveys, one can compare for each Figure 3.7 Affordability of Subsistence Consumption Priced at Cost-Recovery Levels quintile the percentage of households that report paying for the service with the percent- 100 % of households spending more than age of households that report using the service. 90 5% of their monthly budget Those that do not pay include both clandestine 80 70 users and formal customers who fail to pay 60 their bills. Overall, about 40 percent of people 50 connected to infrastructure services do not 40 pay for them (figure 3.6). Nonpayment rates 30 range from about 20 percent in the more afflu- 20 ent quintile to about 60 percent in the poorest 10 quintile. A significant nonpayment rate, even 0 2 4 6 8 10 12 14 16 among the more affluent quintiles, suggests monthly bill (US$) that a culture of payment problems exists in upper bound for subsistence consumption low-income countries addition to any affordability issues. lower bound for subsistence consumption middle-income countries The cost of providing subsistence con- sumption of water or electricity ranges from Source: Banerjee, Wodon, and others 2008. $2 to $8 a month, depending on the extent of consumption and cost recovery (figure 3.7). A range between $2 (based on an operating more formal method of gauging affordability cost-recovery tariff of $0.40 per cubic meter is to measure the cost of utility bills against and an absolute minimum consumption of household budgets. The cost of a monthly 4 cubic meters) and $8 (based on a full cost- subsistence consumption of piped water can recovery tariff of $0.80 per cubic meter and a 92 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION more typical modest household consumption few poor consumers are connected to the ser- of 10 cubic meters). The cost of a monthly vice (Boccanfuso, Estache, and Savard 2008a, subsistence consumption of power can range 2008b, 2008c). As the consequences of higher between $2 (based on a low-cost country tar- power or water prices work their way through iff of $0.08 per kilowatt-hour and an absolute the economy, however, broader second-order minimum consumption of 25 kilowatt-hours) effects on wages and prices of goods in the and $8 (based on a high-cost country tar- economy as a whole can lead to more substan- iff of $0.16 per kilowatt-hour and a more tial effects on poverty (Boccanfuso, Estache, typical modest household consumption of and Savard 2008a, 2008b, 2008c). 50 kilowatt-hours). Notwithstanding these findings, most Afri- An affordability threshold of 5 percent can countries heavily subsidize tariffs for power of household budgets provides a gauge for and water services. On average, power tariffs measuring which utility bills might be afford- recover only 75 percent of full costs, and water able to African households. By looking at the tariffs only 64 percent. The resulting implicit distribution of household budgets, one can cal- service subsidies amount to as much as $4.1 culate the percentage of households for which billion a year (0.7 percent of Africa's GDP), such bills would absorb more than 5 percent divided evenly between power and water (see of their budgets and thus prove unaffordable. chapter 2 of this volume). Monthly bills of $2 are affordable to almost the Because electricity and water subsidies are entire African population. Monthly bills of $8 typically justified by the need to make ser- would remain affordable for the entire popula- vices affordable to low-income households, a tion of the middle-income African countries, key question is whether subsidies reach such indicating that cost recovery should not be a households. Results across a wide range of major problem for these countries. African countries, for both power and water Cost-recovery tariffs would also be afford- sectors, show that the share of subsidies going able for those currently enjoying access in low- to the poor is less than half of the share of the income countries, but not for the remaining poor in the overall population, indicating a very population. In low-income countries, monthly regressive distribution (figure 3.8). This result bills of $8 would remain perfectly affordable is hardly surprising given that connections to for the most affluent 20­40 percent of the power and water services are already highly population, the only portion enjoying access. skewed toward more affluent households. To However, such bills would not be affordable to put these results in perspective, one must com- the poorest 60­80 percent that currently lack pare them with the targeting achieved by other access even if services were extended to them. forms of social policy. Estimates for Cameroon, The affordability problems associated with a Gabon, and Guinea indicate that expenditures universal access policy would be particularly on primary education and basic health care large for a handful of the poorest low-income reach the poor better than do power and water countries--Burundi, the Democratic Repub- subsidies (Wodon 2008a, 2008b). lic of Congo, Ethiopia, Guinea-Bissau, Malawi, Can African governments afford to further Niger, Tanzania, and Uganda--where as much expand today's subsidy model to achieve uni- as 80 percent of the population could not versal access? Little justification currently exists afford a monthly bill of $8. for utility subsidies, given that they typically The immediate poverty-related effect of do not reach unconnected low-income house- raising tariffs to cost-recovery levels is gener- holds but rather favor more affluent, connected ally quite small, although it may have second- households that do not really need subsidies order effects. Detailed analysis of the effect to afford the service. However, the preced- of significant tariff increases of the order of ing analysis indicates that affordability would 40 percent for power and water services in become a major issue to the extent that Africa's Senegal and power services in Mali confirms low-income countries move aggressively toward that the immediate poverty-related effect on universal access. Given the very high macro- consumers is small, essentially because very economic cost today of subsidizing even the Dealing with Poverty and Inequality 93 Figure 3.8 Extent to Which Electricity and Water Subsidies Reach the Poor, by Country a. Electricity b. Water Nigeria Senegal Gabon Central African Republic Gabon Congo, Rep. of Congo, Rep. of Côte d'Ivoire Nigeria, Kaduna Cape Verde Togo Togo Congo, Dem. Rep. of São Tomé and Principe Nigeria, Fed. Cap. Ter. Senegal Cameroon Cameroon Côte d'Ivoire progressive progressive regressive regressive Mozambique Niger Ghana Chad Central African Republic Cape Verde Burundi Guinea Malawi, Blantyre Burundi Ghana Burkina Faso Guinea Chad Uganda Malawi Malawi, Lilongwe Uganda Burkina Faso Rwanda Rwanda 0 0.2 0.4 0.6 0.8 1 1.2 0 0.2 0.4 0.6 0.8 1 1.2 measure of distributional incidence measure of distributional incidence Sources: Banerjee, Wodon, and others 2008; Wodon 2008a, 2008b. Note: This figure presents a measure of distributional incidence, which captures the percentage of total subsidy value reaching the poor relative to the percentage of the poor in the population. A value greater than 1 implies that the subsidy distribution is progressive (pro-poor), because the share of benefits allocated to the poor is larger than their share in the total population. A value less than 1 implies that the distribution is regressive. minority of the population with access to power unconnected households over 20 years would and water, a legitimate question is whether be substantially low, only 0.35 percent of GDP African governments can afford to scale up this for power and about 0.25 percent of GDP for subsidy-based model to the remainder of their piped water. A key difference is that the cost of populations. this one-time subsidy would disappear at the Providing universal use-of-service sub- end of the decade, whereas the use-of-service sidies for power and water would absorb an subsidy would continue indefinitely. unaffordable 1.6 percent of GDP above exist- The welfare case is quite strong for one- ing spending, about two-thirds for power and time capital subsidies to support universal one-third for water. These values are high in connection. Households without access to relation to existing operation and maintenance utility services eventually pay much higher expenditure, so it is difficult to believe that prices and, as a result, limit their consumption they would be fiscally affordable for govern- to very low levels. Small-scale piped-network ments (figure 3.9). operators charge 1.5 times the formal network One-time capital subsidies could be pro- price, point sources charge 4.5 times the for- vided at a lower cost and if spread over 20 mal network price, and mobile distributors years, might just be affordable. The cost of can charge up to 12 times the formal utility providing a one-time capital subsidy of $200 tariff (Kariuki and Schwartz 2005). A recent to cover network connection costs for all survey of Accra, Dar es Salaam, and Nairobi 94 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 3.9 Amount of Subsidy Needed to Maintain Affordability of Water and Electricity Service, 2005 a. Ongoing use-of-service subsidy b. One-time connection subsidy 2.5 2.5 2.0 2.0 % of GDP % of GDP 1.5 1.5 1.0 1.0 0.5 0.5 0 0 electricity water electricity water operating subsidy needed to maintain affordability capital subsidy needed to maintain affordability operation and maintenance spending (2005) capital spending (2005) Source: Banerjee, Wodon, and others 2008. found that the price of utility-piped water Box 3.2 details some of the access and afford- ranges from $0.5 to $1.5 a cubic meter, whereas ability challenges arising for urban transpor- small water enterprises charge between $4 and tation in Africa's burgeoning cities. $6 (McGranahan and others 2006). Similarly, for electricity, the cost of providing basic illu- mination through candles or kilowatt-hours Alternatives to Modern is an order of magnitude greater than that for Infrastructure Services--the electricity per effective unit of lighting (Foster Missing Middle and Tré 2003). Interestingly, even though nonutility water Even with renewed efforts, most African coun- vendors charge higher unit prices, those pur- tries will not realize universal access to modern chasing water from vendors do not necessar- infrastructure services for some decades. In ily spend more than those purchasing water the meantime, most households will continue from the public utility--they simply adjust the to rely on alternative ways of meeting their quantity consumed. water and energy requirements (figure 3.10). Nonmonetary benefits of connection can For the most part, these methods comprise also be very significant. Beyond the potential traditional alternatives such as wells, unim- monetary savings, piped water and electricity proved latrines, and kerosene lamps. However, are associated with a wide range of health, edu- second-best alternatives also exist that provide cation, and productivity benefits (see chapter a significantly higher level of service than the 1 of this volume). Better water and sanitation traditional solutions but at a substantially service is associated with less malnutrition and lower cost than full-blown piped water or stunting, and it liberates women from the time- power connections. Examples include stand- consuming chore of collecting water, leaving posts for water supply, improved latrines for more time for income-generating activities sanitation, and street lighting for basic neigh- (box 3.1). Better electricity provision improves borhood illumination. literacy and primary school completion rates, Although traditional alternatives are in because better quality of light allows students widespread use, second-best alternatives have to read and study without sunlight. not yet become popular in most countries. The Finally, affordability concerns also exist for water and sanitation sectors illustrate this point urban transportation services. Transportation very clearly. In both cases, the traditional alter- represents one of the largest household bud- natives (whether wells or unimproved latrines) get expenditure shares among infrastructure provide the largest share of service across the services, absorbing 4­6 percent of the budgets income spectrum. However, the second-best of those households reporting expenditures. alternatives (whether standposts or improved Dealing with Poverty and Inequality 95 BOX 3.1 Access to Basic Infrastructure and Time Use It is often said that access to basic infrastructure could help infrastructure could reduce poverty through a reallocation increase the earnings of households both by making their of household members' time. work more productive and by freeing time allocated to Some emerging evidence from household surveys indicates domestic chores and allocating such time savings to pro- that the time saved from access to basic infrastructure can ductive work. Some authors have argued that households indeed be substantial. Using data from Sierra Leone, Wodon face a "time overhead" constraint--the minimum number and Ying (2009) show that women work more than men on of hours that household members must spend on basic domestic tasks and that the domestic workload of children is chores vital to the well-being and survival of the family. also high. Access to water and electricity helps reduce domestic This burden includes time spent preparing meals, washing work time by up to 10 hours a week. Using data from Guinea, clothes, cleaning, transporting water, and gathering fuel Bardasi and Wodon (2009) find effects on time use of a simi- for cooking and heating. Access to basic infrastructure can lar order of infrastructure for access to water. The question significantly reduce this time overhead, and thereby free remains whether household members could find employment time for productive work. Because households that lack opportunities using the time saved through access to infra- access to basic infrastructure tend to be poor and because structure. Even if time savings were remunerated at a fraction most of these households have members who want to of the minimum wage, the economic return of these projects work longer hours to increase their earnings, access to is often very large owing to the time savings of households. latrines) have coverage rates comparable to or conspire to limit the extension of the second- even lower than the best alternatives (whether best alternatives. piped water or flush toilets) despite their sig- On the demand side, the costs of the nificant cost advantages. Moreover, coverage second-best alternatives may still be relatively of the second-best alternatives is just as regres- high, given limited household budgets. Water sive as that of the best alternatives. Neverthe- from standposts, though relatively cheap to less, some countries have made significant provide, is often retailed by intermediaries progress in expanding coverage of second-best charging substantial markups that outweigh alternatives--such as the Democratic Republic the underlying advantages in construction of Congo, Mozambique, Tanzania, and Uganda costs. Improved latrines, though cheaper than for standposts; and Burkina Faso, Cameroon, flush toilets, are nonetheless substantially Ghana, Madagascar, and Rwanda for improved more expensive than unimproved latrines, and latrines. Although data are not available to uneducated households may not be aware of make a similar comparison for lighting, it is well the health benefits. known that coverage of street lighting lags. On the supply side, their public good The capital costs of the second-best alter- nature greatly complicates the implementa- natives are still only a fraction of those asso- tion of second-best alternatives. The provision ciated with the best alternatives, even if they of standposts and street lighting is unattractive are also significantly more expensive than the to utilities because of their limited scope for traditional alternatives (table 3.3). Thus, the revenue collection as well as the greater poten- second-best alternatives provide an opportu- tial for revenue loss from increased clandestine nity to make limited investment budgets go connections once networks are provided. For further and accelerate the expansion of service improved latrines, the limited experience of improvements. the local construction sector restricts the avail- Therefore, understanding the factors that ability of such designs and may keep their costs lie behind this "missing middle" is important. higher than they would be in a mass market Once again, both demand and supply issues (see chapter 17 of this volume). 96 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 3.2 Access, Affordability, and Alternatives--Urban Public Transportation Access to urban transportation services in Sub-Saharan to business practices, such as overcrowding, erratic schedul- Africa is constrained both by the limited reach of the urban ing, and price discrimination, that favor business interests paved-road network and by the limited size of the bus over those of consumers. fleet. Bus fares in African cities tend to be about $0.30 per Only one-third of the roads in African cities are paved, one-way trip, irrespective of bus size. Evidence from bud- ranging from barely 10 percent in Kinshasa (Democratic get surveys indicates that households spend on average Republic of Congo) and Kigali (Rwanda) to more than $12­$16 a month on urban transportation services. This 70 percent in Kampala (Uganda). Paved-road density is sum is enough to purchase about 20 round-trip bus tickets typically about 300 meters per 1,000 inhabitants, compared per month, which covers the essential travel requirements with more typical values of 1,000 meters per 1,000 inhab- of one commuter per household, leaving nothing to cover itants for developing cities around the world. Overall, the travel needs of other household members. Expenditure road network constitutes less than 7 percent of the land levels of the poorest households would be inadequate to area in most African cities, compared with 25­30 percent cover the transportation costs of even one commuter per in developed-country cities. The low coverage of the paved household. network limits the reach of bus services. In many African cit- A significant minority of urban households do not ies, numerous outlying neighborhoods can be reached only report any expenditure on urban transportation, suggest- by two-wheeled vehicles. ing that their transportation needs are met entirely by foot. The typical availability of bus services in African cities is Data on the modal split of urban journeys show, on aver- 30­60 bus seats per 1,000 residents, although the number age, 37 percent of urban trips taken on buses and another can be as low as 10 seats per 1,000 in Addis Ababa (Ethio- 37 percent on foot. The remainder of trips is spread across pia), Kinshasa, and Ouagadougou (Burkina Faso). In con- a range of private modes. The percentage of trips on foot trast to middle-income countries that typically have 30­40 can be 50 percent or more--Nairobi, Kenya (47); Douala, large bus seats per 1,000 residents, low-income countries in Cameroon (60); and Conakry, Guinea (78). Africa have only 6 per 1,000. The combination of low access and limited affordability The proliferation of informal minibus services has been a for service conspires to seriously constrain the mobility of private response to the financial demise of large, publicly run urban residents, preventing cities from realizing their full bus services in most cities. Although minibuses have been potential to bring together people, services, and economic very responsive to demand, they pose a variety of social opportunities. issues, including congestion, pollution, and road safety. Moreover, the lack of effective regulation of minibuses leads Source: Kumar and Barrett 2008. Figure 3.10 Access to Alternative Water and Sanitation Services across All Income Levels a. Alternative water services b. Alternative sanitation services 60 60 50 50 % of households % of households 40 40 30 30 20 20 10 10 0 0 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 budget quintile budget quintile piped water surface water open defecation flush toilet public standpost well water VIP/chemical/SANPLAT traditional latrine Source: Banerjee, Wodon, and others 2008. Dealing with Poverty and Inequality 97 Table 3.3 Capital Cost of Best, Second-Best, and Traditional Alternatives $ per capita Alternative Water Sanitation Lighting Traditional Well 21 Unimproved latrine 39 Kerosene lamp -- Second best Standpost 80 Improved latrine 57 Street lighting -- Best Piped water 283 Flush toilet 125 Electricity 133 Source: Chapter 5 of this volume. Note: Capital cost estimates are based on a density of 3,272 persons per square kilometer. -- Not available. Policy Challenges for Accelerating infrastructure investments. The challenge of Service Expansion reaching universal access is typically character- ized as a supply problem of rolling out infra- The business-as-usual approach to expand- structure networks to increasingly far-flung ing service coverage in Africa does not appear populations. However, household survey evi- to be working. The low and stagnant cover- dence shows that a significant segment of the age of household services comes with a major unserved population in urban areas lives close social and economic toll. Most African coun- to a network. The relatively low rate of hookup tries have tackled universal access by provid- to existing infrastructure networks leads to ing heavily subsidized best options, such as lower financial, economic, and social returns to piped water and electricity. This approach the associated investment, because the physical has tended to bankrupt and debilitate sector asset is operating below its full carrying capac- institutions without bringing about any sig- ity. This finding has implications for network nificant acceleration of coverage. The associ- rollout strategy. ated public subsidies have also largely bypassed First, hookup--rather than access--needs most needy groups. Few services and countries to be considered the key measure of success. are expanding coverage at rates high enough Interventions that aim to expand service cov- to outstrip demographic growth, particularly erage too often measure their outcomes by urbanization. the number of people who can connect to the Turning this situation around will require network provided. As a result, little attention rethinking the approach to service expansion in is given to whether these connections actu- four ways. First, coverage expansion is not just ally materialize after the project's completion. about network rollout. Demand-side barriers Unless the focus of monitoring and evaluation such as high connection charges or legal ten- shifts from access to hookup, those involved ure must be addressed. Second, cost recovery in service expansion will have little incentive for household services needs to be improved to think about the demand side of service to ensure a financial basis for utilities to invest coverage. in service expansion. Third, rethinking the Second, the most cost-effective way of design of utility subsidies to better target them increasing coverage may be to pursue densi- and to accelerate service expansion is desirable. fication programs that increase hookup rates Fourth, the level of service that households can in targeted areas. Unserved populations liv- afford to pay for and governments can afford ing physically close to infrastructure networks to subsidize must be considered realistically could (in principle) be covered at a much and greater emphasis placed on second-best lower capital cost than those living farther modern solutions. away, thereby providing the highest potential return to a limited investment budget. In that sense, they may deserve priority attention in Remembering the Demand Side of efforts to raise coverage. the Equation Third, expanding coverage requires com- Overlooking the demand side of network munity engagement. Dealing with the demand- rollout can lead to much lower returns on side barriers that prevent hookup requires a 98 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION more detailed understanding of the potential expansion of utility networks is hampered by client base of the utility. What are their alterna- the absence of legal tenure and by high turnover tives? How much can they afford to pay? What rates among tenants, not to mention inadequate other constraints do they face? This approach spacing of dwellings. Providing services to these in turn suggests a broader skill base than communities will require close cooperation utilities may routinely retain, going beyond with urban authorities because many of these standard expertise in network engineering to issues can be resolved only if they are addressed encompass sociological, economic, and legal in a synchronized and coordinated manner. analysis of--and engagement with--the target populations. Taking a Hard-Headed Look at Fourth, careful thought should be given Affordability to how connection costs might be recovered. Underrecovery of costs has serious implica- As noted earlier, high connection charges-- tions for the financial health of utilities and widespread across Africa--are one obvious slows the pace of service expansion. Many of demand-side barrier to hookup, even when Africa's water and power utilities capture only use-of-service charges would be affordable. two-thirds of the revenues needed to function In these circumstances, one may legitimately sustainably. This revenue shortfall is rarely cov- ask whether substantial one-time up-front ered through timely and explicit fiscal transfers. connection charges are the most sensible way Instead, maintenance and investment activities to recover network connection costs. Alterna- are curtailed to make ends meet, starving the tives can be considered, including repaying utility of funds to expand service coverage and connection costs over several years through an eroding the quality of service to existing cus- installment plan; socializing connection costs tomers (see chapter 8 on power and chapter 16 by recovering them through the general tariff on water utilities in this volume). and, hence, sharing them across the entire cus- Affordability, the usual pretext for under- tomer base; or directly subsidizing them from pricing services, does not bear much scru- the government budget (see box 3.3). tiny. Political economy likely provides the real Fifth, expansion of utility networks needs explanation for low tariffs, with populations to be closely coordinated with urban devel- currently connected to utility services tending opment. In many periurban neighborhoods, to be those with the greatest voice. The implicit BOX 3.3 Are Connection Subsidies Well Targeted to the Poor? Within the framework of Niger's poverty (CFAF) (US$39) plus CFAF 2,500 (US$6) of reduction strategy, some 11,200 branchement administrative fees. Moreover, the household sociaux (social connections) to the water net- had to be able to pay the bill each month in work were provided in 2002­04, about half of a single installment. The 11,200 connections them in Niamey. To be eligible for a social con- had been originally planned over five years, nection, household members had to be living but they were realized in only a year and a in a house within 20 meters of the main pipe, half, and some 600 requests could not be and the house had to be built of solid materials satisfied. (a permanent construction). Some geographic Households that benefited from a social targeting to the poor was achieved by making connection belonged to the poorer income the social connections in poor and suburban quintiles of the population, suggesting that quarters of the city. the connections were fairly well targeted. To obtain a social connection, a house- hold had to pay a deposit of 17,500 francs Source: Tsimpo and Wodon 2009. Dealing with Poverty and Inequality 99 subsidies created by underpricing are extremely Various tactical measures can improve the regressive in their distributional incidence. acceptability of tariff increases, but ensuring In all but the poorest African countries, ser- their sustainability is most important. Tariff vice coverage could be substantially increased increases can be phased in either gradually or before any real affordability problems would be instantly through a one-time adjustment. Both encountered. In the poorest of the low-income approaches have advantages and disadvantages. countries, affordability is a legitimate concern The public acceptability of tariff increases can for the bulk of the population and would con- be enhanced if they are part of wider measures strain universal coverage. Even in the poorest that include service quality improvements. countries, however, recovering operating costs One method to strengthen social accountabil- should be feasible, with subsidies limited to ity is adopting communication strategies that capital costs. link tariffs with service delivery standards and How would removal of utility subsidies suggest conservation measures to contain the affect poverty reduction? For most countries, overall bills. In any event, ensuring that the electricity and water spending accounts for realignment of tariffs and costs is sustained only a tiny fraction of total consumption. At by providing for automatic indexation and the national level, a 50 percent increase in tar- periodic revisions of tariffs is perhaps most iffs or even a doubling of tariffs has a marginal important. effect, with the share of the population living Countries have pursued different paths in poverty increasing barely 0.1 of a percentage to increase tariffs critical for operational and point. Among households with a connection to financial sustainability. In Niger, the standpost the network, the effect is larger but still limited. and low-volume tariffs have barely increased Indeed, an increase in the share of households since 2000, but the industrial and commer- in poverty larger than 1 or 2 percentage points cial tariffs have grown 6­7 percent in nominal rarely occurs. Because the households that terms (figure 3.11). In Lilongwe, Malawi, the benefit from a connection also tend to be bet- same increases have been applied across all tar- ter off than other households, the increase in iff categories. In addition, Botswana, Namibia, poverty starts from a low base. Thus, the small South Africa (Eskom), and Tanzania recently effect of a tariff increase on poverty could be increased electricity tariffs as a result of oil offset by reallocating utility subsidies to other price shocks in 2007­08. areas of public expenditure with a stronger A danger always exists that higher tariffs will pro-poor incidence. simply lead to lower revenue collection, but Figure 3.11 Increased Industrial and Commercial Tariffs, Niger and Malawi a. Niger (SPEN), 2000­06 b. Malawi LWB, 1999­2006 500 120 450 400 100 local currency unit local currency unit 350 80 300 250 60 200 150 40 100 20 50 0 0 2000 2001 2002 2003 2004 2005 2006 1999 2000 2001 2002 2003 2004 2005 2006 government commerce September industry residential 41­75 m3 institutional residential 0­10 m3 residential 0­15 m3 residential>75 m3 residential 10­40 m3 residential > 40 m3 residential 11­40 m3 industry > 100 m3 Source: Banerjee, Wodon, and others 2008. Note: LWB = Lilongwe Water Boards; SPEN = Société de Patrimoine des Eaux de Niger; m3 = cubic meter. 100 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION prepayment meters can help. In the absence Targeting Subsidies to Promote of a strong payment culture, customers who Service Expansion object to tariff hikes may "retaliate" by refusing Subsidies also have a role to play, but their to pay their bills. Thus, even before address- design requires major rethinking. Subsidies ing tariff adjustments, utilities must work on have a valuable and legitimate role in the proper raising revenue collection rates toward best circumstances. They may be appropriate when practice and establishing a payment culture. households genuinely cannot purchase a sub- At least for power, one technological solu- sistence allowance of a service that brings major tion is to use prepayment meters, which place social and economic benefits to themselves and customers on a debit card system similar to those around them, as long as governments can that for cellular telephones. For utilities, this afford to pay those subsidies. However, the approach eliminates credit risk and avoids design and targeting of utility subsidies must nonpayment. For customers, it allows them to be radically improved to fulfill their intended control their expenditure and avoid consum- role. As noted earlier, the utility subsidies pro- ing beyond their means. South Africa was at vided in Africa today largely bypass the poorest. the forefront in development of the keypad- African utilities typically subsidize con- based prepayment electricity meter. ESKOM sumption, but subsidizing connection is launched the first product, Cashpower, in potentially more equitable and effective in 1990. Tshwane in South Africa reports univer- expanding coverage. The affordability prob- sal coverage of its consumers with prepayment lems associated with connection charges are meters. In Lesotho, Namibia, and Rwanda, a often much more serious than those associated majority of residential customers use prepay- with use-of-service charges. Because connec- ment meters. In Ghana and Malawi, a clear tions are also disproportionately concentrated policy exists of rapidly increasing the share of among the more affluent, the absence of con- residential customers on prepayment meters nections is disproportionately concentrated (figure 3.12). among the poorest, which could potentially facilitate the targeting process. The targeting performance of connection subsidies ultimately hinges on how new connec- Figure 3.12 Residential Customers Using Prepayment tions are allocated. In African countries, where Meters, by Utility coverage is far from universal even among the Tshwane higher-income groups, connection subsidies LEC may be just as regressive as consumption sub- NORED sidies, essentially because the unserved higher- ESKOM income groups will likely be the first to benefit Electrogaz utility from coverage expansion. Simulations suggest VRA Escom that--if new subsidized connections mirror Sonabel the distribution of existing connections--the SBEE share of connection subsidies going to the poor TANESCO would be only about 36 percent of the share of 0 10 20 30 40 50 60 70 80 90 100 the poor in the population--a highly regressive residential customers with result no better than that of existing consump- prepayment meters (%) tion subsidies (table 3.4). Source: Eberhard and others 2008. Limiting subsidies to connections in new Note: Electrogaz = Rwanda's national electric utility; Escom = network rollout rather than densification Electricity Supply Corporation (Malawi); ESKOM = South Africa's national electric utility; LEC = Lesotho Electricity Company; of the existing network would substantially NORED = Northern Electricity Distribution Service (Namibia); improve targeting. The share of connection SBEE = Société Béninoise d'Énergie Électrique (Benin); Sonabel = Société Nationale d'Électricité du Burkina (Burkina Faso); subsidies going to the poor would rise to TANESCO = Tanzania Electricity Supply Company Limited; 74­95 percent of their share in the population, Tshwane = Tshwane Local Electric Utility (South Africa); VRA = Volta River Authority (Ghana). depending on the utility involved, but the Dealing with Poverty and Inequality 101 outcome would remain regressive. Providing a Table 3.4 Potential Targeting Performance of Connection Subsidies under Various connection subsidy equally likely to reach all Scenarios percentage of total poor getting connection subsidies relative to percentage of unconnected households would ensure that poor in the population the percentage going to the poor exceeds their New connections Only households beyond All unconnected share of the population by 112­118 percent-- mirror pattern of reach of existing network households receive finally, a progressive result. Improving the Utility existing connections receive connection subsidies subsidy distributional incidence beyond this modest Electricity 37 95 118 level would require connection subsidies to be Water 35 74 112 accompanied by other socioeconomic screens. Sources: Banerjee, Wodon, and others 2008; Wodon 2008a, 2008b. These findings illustrate that in the low-access environment in most African countries, the absence of a connection remains a relatively weak targeting variable. support the classification of beneficiaries, as Can anything be done to improve targeting well as a significant amount of administrative of use-of-service subsidies? The poor perfor- capacity. Both factors are often absent in Africa, mance of existing utility subsidies is explained particularly in the low-income countries. by pro-rich coverage and by the widespread An important test of the coherence of a use of poorly designed IBTs. Common design subsidy policy is whether the country could failures in water IBTs include high fixed afford the policy if it were scaled up to univer- charges and minimum consumption levels sal access. The underpricing of utility services that penalize small consumers, as well as the that benefit just a small minority of the popu- large size and universal applicability of the first lation costs many African countries as much as subsidized block (Banerjee, Foster, and oth- 1 percent of GDP. As countries move toward ers 2008). Common design failures in power universal access, that subsidy burden would IBTs include large subsistence thresholds increase proportionately, rapidly becoming that allow only consumers with exceptionally unaffordable for the national budget. Thus, high consumption to contribute fully to cost countries should consider how the cost of any recovery (Briceño-Garmendia and Shkaratan proposed subsidy policy would escalate as cov- 2008). Achieving major improvements in the erage improves. This test of a subsidy's fiscal targeting of use-of-service subsidies by over- affordability is an important consideration to hauling the design of increasing block tariffs help countries avoid embarking on policies (IBTs) is difficult. Some improvements in tar- that are simply not scalable. geting could be achieved by eliminating fixed Another potentially effective method of charges, reducing the size of first blocks to targeting is to limit the allocation of subsidies cover only genuinely subsistence consump- to lower-cost and lower-quality alternatives tion, and changing from an IBT to a volume- that encourage self-selection. For services such differentiated tariff in which those consuming as water, for which different modes of service beyond a certain level forfeit the subsidized provision exist, subsidies could possibly be first-block tariff completely. Even with these concentrated on second-best alternatives such modifications, however, the targeting of such as standposts while requiring full cost recov- tariffs would improve only marginally and not ery from private piped-water connections. The become strongly pro-poor in absolute terms. theory is that more affluent customers will Global experience suggests that the target- eschew second-best services and automatically ing of utility subsidies can be improved and opt to pay the full cost of the best alternative, become reasonably progressive, if some form of thus identifying themselves and leaving the geographical or socioeconomic targeting vari- subsidized service to less affluent customers. In ables can be used beyond the level of consump- Africa, however, the use of self-selection may be tion (Komives and others 2005). However, such less effective, because coverage of second-best targeting schemes hinge on the existence of alternatives such as standposts and improved household registers or property cadastres that latrines is just as regressive as coverage of best 102 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION alternatives such as piped water and flush alternative to standposts, although it is often toilets. not legally recognized. In addition, yard taps that serve four or five households--not several Giving More Consideration to hundred--can reduce costs while avoiding some Second-Best Solutions of the most serious public good problems. Second-best solutions appear to provide a happy compromise but face many implemen- Notes tation challenges. As noted earlier, second-best The authors of this chapter are Sudeshna Ghosh approaches provide modern services that are far Banerjee, Quentin Wodon, and Vivien Foster, preferable in welfare terms to their traditional who drew on background material and contribu- alternatives and are still much less costly than tions from Tarik Chfadi, Amadou Diallo, Sarah Keener, Taras Pushak, Maria Shkaratan, Clarence best modern infrastructure services. So why are Tsimpo, Helal Uddin, and Yvonne Ying. these second-best services fairly rare in Africa-- 1. The cross-regional figures for infrastructure and skewed toward more affluent households? coverage are unweighted simple averages. A key problem of many second-best solu- tions is that their public good nature compli- References cates their adoption. Both public standposts and Banerjee, Sudeshna, Vivien Foster, Yvonne Ying, street lighting are essentially public goods. This Heather Skilling, and Quentin Wodon. 2008. characterization makes it difficult for utilities to "Achieving Cost Recovery, Equity and Efficiency recover the costs of these services and exposes in Water Tariffs: Evidence from African Utilities." Working Paper 7, Africa Infrastructure Country them to theft by extending the network's reach Diagnostic, World Bank, Washington, DC. into lower-income areas. Thus, utilities have Banerjee, Sudeshna, Quentin Wodon, Amadou no real incentive to provide such loss-prone Diallo, Taras Pushak, Hellal Uddin, Clarence services. In addition, facilities are vulnerable to Tsimpo, and Vivien Foster. 2008. "Access, Afford- maintenance issues, because nobody is respon- ability, and Alternatives: Modern Infrastructure sible for preventing, reporting, or addressing Services in Africa." Background Paper 2, Africa problems. One solution is to introduce an agent Infrastructure Sector Diagnostic, World Bank, Washington, DC. responsible for managing the facility, charging for service, and soliciting maintenance activi- Bardasi, Elena, and Quentin Wodon. 2009. "Work- ing Long Hours and Having No Choice: Time ties. However, covering the agent's salary adds Poverty in Guinea." Policy Research Working significantly to the cost of the second-best alter- Paper 4961, World Bank, Washington, DC. native, and agents often exploit their control- Boccanfuso, Dorothée, Antonio Estache, and Luc ling position to charge excessive rates. Savard. 2008a. "Electricity Reforms in Mali: A The African experience with standposts pro- Micro-Macro Analysis of the Effects on Pov- vides pointers for improving the performance erty and Distribution." Working Paper 4, Africa of such public facilities (Keener, Luengo, and Infrastructure Country Diagnostic, World Bank, Washington, DC. Banerjee 2008). Where standposts are admin- istered by local agents, the management model ------. 2008b. "Electricity Reforms in Senegal: A Micro-Macro Analysis of the Effects on Pov- should be grounded in the prevailing culture erty and Distribution." Working Paper 5, Africa of the beneficiary community. Checks and bal- Infrastructure Country Diagnostic, World Bank, ances are needed to ensure that the delegated Washington, DC. manager behaves responsibly. The utility also ------. 2008c. "Water Reforms in Senegal: A must be closely involved in monitoring the sta- Micro-Macro Analysis of the Effects on Poverty tus of the standposts, regularly collecting the and Distribution." Working Paper 16, Africa water revenues, and providing technical assis- Infrastructure Country Diagnostic, World Bank, Washington, DC. tance to the standpost operators. Therefore, defining a useful set of incentives that bolster Briceño-Garmendia, Cecilia, and Maria Shkaratan. 2008. "Achieving Cost Recovery, Equity, and the utility's growing interest in participating Efficiency in Power Tariffs: Evidence from in the standpost business is essential. In some African Utilities." Working Paper 21, Africa environments, resale of water by households Infrastructure Country Diagnostic, World Bank, with private connections can be a practical Washington, DC. Dealing with Poverty and Inequality 103 Diallo, Amadou, and Quentin Wodon. 2005. "A Africa Infrastructure Country Diagnostic, World Note on Access to Network-Based Infrastructure Bank, Washington, DC. Services in Africa: Benefit and Marginal Inci- Komives, Kristin, Vivien Foster, Jonathan Halpern, dence Analysis." World Bank, Washington, DC. and Quentin Wodon. 2005. Water, Electricity, Eberhard, Anton, Vivien Foster, Cecilia Briceño- and the Poor: Who Benefits from Utility Subsidies? Garmendia, Fatimata Ouedraogo, Daniel Washington, DC: World Bank. Camos, and Maria Shkaratan. 2008. "Under- Kumar, Ajay, and Fanny Barrett. 2008. "Stuck in powered: The State of the Power Sector in Sub- Traffic: Urban Transport in Africa." Background Saharan Africa." Background Paper 6, Africa Paper 1, Africa Infrastructure Country Diagnos- Infrastructure Sector Diagnostic, World Bank, tic, World Bank, Washington, DC. Washington, DC. McGranahan, Gordon, Cyrus Njiru, Mike Albu, Estache, Antonio, and Quentin Wodon. 2007. Mike Smith, and Diana Mitlin. 2006. How Small Infrastructure and Poverty in Sub-Saharan Africa. Water Enterprises Can Contribute to MDGs: Evi- Directions in Development Series. Washington, dence from Dar es Salaam, Nairobi, Khartoum, DC: World Bank. and Accra. Leicestershire, U.K.: Water, Engineer- Foster, Vivien, and Jean-Philippe Tré. 2003. "Mea- ing and Development Centre, Loughborough suring the Impact of Energy Interventions on University. the Poor--An Illustration from Guatemala." In Tsimpo, Clarence, and Quentin Wodon. 2009. Infrastructure for Poor People: Public Policy for "Who Benefits from Electricity Consumption Private Provision, ed. Penelope Brook and Tim versus Connection Subsidies? Evidence from Irwin, 125­78. Washington, DC: World Bank. Niger." Development Dialogue on Values and Gulyani, Sumila, Debabrata Talukdar, and Darby Ethics, World Bank, Washington, DC. Jack. 2008. "A Tale of Three Cities: Understand- Wodon, Quentin, ed. 2008a. "Electricity Tariffs ing Differences in Provision of Modern Services." and the Poor: Case Studies from Sub-Saharan Working Paper 10, Africa Infrastructure Country Africa." Working Paper 11, Africa Infrastructure Diagnostic, World Bank, Washington, DC. Country Diagnostic, World Bank, Washington, Kariuki, Mukami, and Jordan Schwartz. 2005. DC. "Small-Scale Private Service Providers of Water ------. 2008b. "Water Tariffs and the Poor: and Electricity: A Review of Incidence, Struc- Case Studies from Sub-Saharan Africa." ture, Pricing, and Operating Characteristics." Working Paper 12, Africa Infrastructure Policy Research Working Paper 3727, World Country Diagnostic, World Bank, Bank, Washington, DC. Washington, DC. Keener, Sarah, Manuel Luengo, and Sudeshna G. Wodon, Quentin, and Yvonne Ying. 2009. "The Banerjee. 2008. "Provision of Water to the Poor Determinants of Domestic Work Time in Sierra in Africa: Informal Water Markets and Experi- Leone." Development Dialogue on Values and ence with Water Standposts." Working Paper 13, Ethics, World Bank, Washington, DC. Chapter 4 Building Sound Institutions I nstitutional competence and capacity are providers; outcomes below expectations; and important determinants of the performance a high degree of official and public skepticism of infrastructure providers in every sector. about whether the application of the standard That seems obvious, but systematic analysis has package is producing (or even could produce) been lacking on the nature and extent of the the desired results. A large part of the explana- links between stronger institutions and better tion for this situation is thought to lie in the outcomes: specifically, broader access, higher relative weakness of African practices, policies, service quality, and more financially efficient and agencies (that is, institutions) that guide service. This chapter looks at the different and oversee African infrastructure sectors and institutional models applied, the approaches firms, public or private. to strengthen infrastructure-relevant institu- The statistical analysis for this chapter tions, and the effect of the various approaches suggests that institutions make a difference. on performance. It reveals strong links between institutional The standard infrastructure reform and pol- reforms and enhanced governance in the coun- icy prescription package of the 1990s--market try, sector, and enterprise--and improvements restructuring, private involvement up to and in the quantity and quality of infrastructure including privatization, establishing indepen- services (with sectoral variation). Given the dent regulators, and enhancing competition-- link between institutional development and yielded a fair number of positive results in performance improvements, and the high costs Africa. This conclusion deserves stress: ben- of inaction, strengthening sectoral institutions eficial outcomes following the application of and country and sectoral governance is a very these reforms have often been unacknowledged worthwhile investment. or at least underappreciated. Nevertheless, this Most African countries have undertaken set of reforms has proved more difficult to preliminary institutional reforms, mainly the apply in Africa than in other regions. One finds broader sectoral policy and legal measures, in Africa numerous failures to implement, or many of which can be accomplished by the fully implement, the policy package; renegotia- stroke of a pen. What has lagged are regulatory tions or cancellations of contracts with private and governance reforms; they have taken much 105 106 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION more time to bear fruit. For instance, effective almost all African countries have embarked regulation requires building organizations that on institutional reforms, on average they have challenge established vested interests. Gover- adopted no more than 50 percent of good insti- nance improvements, particularly in state- tutional practices. The variation in performance owned enterprises (SOEs), require aligning across countries is roughly two to one, with the internal and external incentives, which again most advanced countries (Kenya) scoring about require broader reforms of the external envi- 70 percent and those furthest behind scoring ronment for infrastructure service providers. 30 percent (Benin). At the country level, progress in one infra- structure sector is no guarantee of progress in Institutional Reforms: A Glass another. That is, institutional development in Half Full infrastructure sectors is uneven both among and within countries. Countries that perform fairly Africa's institutional framework for infrastruc- well in one aspect of infrastructure do not neces- ture is no more than halfway along the path to sarily do so in another. This finding suggests that best practice. The components of the institu- sector-specific constraints may be as important tional performance indicators developed for this as country-specific constraints. It also points to study capture a wide range of characteristics of the potential for greater cross-fertilization of the institutional environment (box 4.1). A coun- experiences across sectors within a country. try's aggregate score on this index suggests the In addition, the quality of the institutional extent of institutional reforms. Overall, although framework differs across country groupings. BOX 4.1 Infrastructure's Institutional Scorecard To analyze the links between institutional factors related to competitive labor and capital market). Note that reform and infrastructure and performance outcomes at the sector and regulation are country-level indicators, whereas governance enterprise levels, this study devised a standardized survey- is measured at the enterprise level. based methodology that describes the nature of each insti- The Infrastructure Institutional Scorecard applied in this tutional reform and measures the intensity of the reform chapter derives from a detailed survey of African infrastruc- efforts. The methodology builds on, and is compatible with, ture sectors and enterprises. The reform and regulatory other recent literature on the subject. scorecards cover 24 countries for all sectors (except railways This methodology yields a "scorecard," a succinct snap- and ports, which included only 21 countries and 15 coun- shot of what has happened, sector by sector, in three key tries, respectively). The governance scorecards have been institutional dimensions: (a) broad sectoral policy reforms, collected for the 24 telecommunication providers and 21 (b) amount and quality of regulation, and (c) enterprise gov- railway providers. A sample of 30 utilities in the electricity ernance. First, reform is defined as implementing sectoral sector and a sample of 52 utilities in the water sector were legislation, restructuring enterprises, and introducing policy examined. oversight and private sector participation. Second, the qual- The resulting list of institutional reforms represents a ity of regulation entails progress in establishing autono- refinement and extension of previous attempts to generate mous, transparent, and accountable regulatory agencies a global scorecard of institutional reforms for infrastructure and regulatory tools (such as quality standards and tariff sectors. The choice of the indicators was made in consulta- methodology). Third, governance entails the implementa- tion with infrastructure sector experts. Operationally relevant tion of measures inside the enterprise (such as strengthening indicators were selected, each of which had to meet two shareholder voice and supervision, board and management conditions. First, an action was chosen if a consensus existed autonomy, and mechanisms for accounting and disclosure) that it represented "best practice" and was being applied in and measures aimed at improving the external environ- different sectors. Second, the data needed to calculate the ment in which the enterprise operates (including outsourc- indicator had to be relatively easy to obtain at the sectoral ing to the private sector and introducing discipline from a and enterprise levels. (continued) Building Sound Institutions 107 BOX 4.1 (continued) Reform Internal Governance Legislation Ownership and Shareholder Quality Existence of de jure reform Concentration of ownership Implementation of reform Corporatization/limited liability Restructuring Rate of return and dividend policy Unbundling/separation of business lines Managerial and Board Autonomy State-owned enterprise corporatization Autonomy in hiring/firing/wages/production/sales Existence of regulatory body Size of board Selection of board members Policy Oversight Presence of independent directors Oversight of regulation monitoring outside the ministry Dispute arbitration outside the ministry Accounting, Disclosure, and Performance Monitoring Tariff approval outside the ministry Publication of annual reports Investment plan outside the ministry International financial reporting standards/external audits/ Technical standard outside the ministry independent audit Audit publication Private Sector Involvement Remuneration of noncommercial activity Private de jure/de facto Performance contracts/with incentives Private sector management/investment/ownership Penalties for poor performance Absence of distressed/renegotiation/renationalization Monitoring/third-party monitoring Regulation External Governance Autonomy Labor Market Discipline Formal autonomy on hiring/firing Restriction on dismissing employees Financial autonomy (partial/full) Wages, compared to private sector Managerial autonomy (partial/full) Benefits, compared to private sector Multisectoral agency/commissioners Capital Market Discipline Transparency No exemption from taxation Publication of decisions via report/Internet/public hearing Access to debt, compared to private sector Accountability No state guarantees Existence of appeal Public listing Independence of appeal (partial/full) Outsourcing Tools Billing and collection Existence of tariff methodology/tariff indexation Meter reading Existence of regulatory review; length of regulatory review Human resources information technology Jointly, the three sets of indicators (reform, regulatory, Separately, each indicator serves as a basis for measuring the and governance) added together summarize the overall (aggregate and disaggregate) effect of progress in reforms level and type of institutional reforms in any given country. and enterprise performance. Source: Vagliasindi 2008c. Reflecting countries' broader characteristics, cantly further ahead with water reform, perhaps the extent of institutional reforms differs reflecting the strong role of donors in this sector. across these groups (figure 4.1). For example, For telecommunications reform, the resource- middle-income countries are significantly fur- rich low-income countries have higher scores. ther ahead with power sector reform, whereas A correlation exists between the quality aid-dependent low-income countries are signifi- of infrastructure institutions and the overall 108 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 4.1 Institutional Progress across Countries, the overall level of governance and control of by Income Group, Aid Dependence, and corruption,1 as well as the quality of public Resource Richness percentage score on institutional scorecard administration. A polity executive constraint indicator also measures the extent of checks a. Income group and balances within a government (Center for 100 Systemic Peace 2006). 90 80 Nevertheless, some countries do well on 70 infrastructure despite broader governance 60 50 limitations, and vice versa. For Kenya and 40 Niger, lower scores on the country governance 30 indicators have not impeded the achieve- 20 10 ment of good scores across all utilities' insti- 0 tutional reforms. For Zambia, low scores in electricity water telecommunications budget execution and financial management low income middle income have not prevented the country from earn- ing reasonably good scores in infrastructure b. Aid dependence 100 institutional reforms. By contrast, Benin and 90 Lesotho display high country governance and 80 70 polity executive constraint scores and have 60 decent budgetary and financial management 50 40 standards, but neither has a high overall insti- 30 tutional score in the utility sectors. 20 10 Institutional development in the utilities 0 sector is well ahead of that in the transport electricity water telecommunications sector (figure 4.2). Unsurprisingly, institutional aid dependent not aid dependent development is furthest ahead in telecommu- nications, where technological change and c. Resource richness 100 competition have driven momentous change, 90 bringing the overall average reform score to 80 70 just under 50 percent. Electricity and water are 60 not that far behind, with institutional reform 50 scores just over 40 percent. Although institu- 40 30 tional actions in power and water lag those in 20 telecommunications for implementing reform 10 0 agendas, they score somewhat higher in the electricity water telecommunications quality of regulation. Also, the governance not resource rich resource rich framework for the main service providers is significantly better than that for fixed-line tele- Source: Vagliasindi and Nellis 2009. communication incumbents. By contrast, insti- Note: See Vagliasindi 2008c for the definition of the institutional indicators. tutional scores for ports and railways are only about half those for the utilities. These sectors have made significant progress with reform but quality of institutions in the country, though lag in developing the regulatory framework. this correlation is much stronger for electricity Across countries and sectors, the greatest than for water (table 4.1). A key question is the progress has been in sector reform. Average scores extent to which a country can make progress in exceed 60 percent for reform legislation and 50 reforming infrastructure institutions if its wider percent for sector restructuring, policy oversight, governance framework is deficient (Levy 2007). and private sector participation (figure 4.3, Numerous indicators have been developed in panel a). Telecommunications, the most recent years (for example, by Kaufman, Kraay, advanced, scores about 80 percent of the best- and Mastruzzi 2008), attempting to capture practice index across all areas of sector reform. Building Sound Institutions 109 Table 4.1 Correlation between Institutional Scores for Infrastructure and Measures of Broader Country Governance Infrastructure Polity executive Budget and financial Public Overall Control of sector constraint management administration governance corruption Electricity 0.34 0.29 0.53 0.49 0.46 Water 0.08 0.33 0.3 0.18 0.08 Sources: Vagliasindi and Nellis 2009; Center for Systemic Peace 2006 for polity executive constraint scores; IDA 2008 for Country Policy and Institutional Assessment score; Kaufmann, Kraay, and Mastruzzi 2008 for governance and control of corruption. Figure 4.2 Institutional Progress across Sectors percentage score on institutional scorecard a. Telecommunications, electricity, and water b. Ports and railways 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 telecommunications electricity water ports railways reforms regulation governance reforms regulation Source: Vagliasindi and Nellis 2009. Note: See Vagliasindi 2008c for the definition of the institutional indicators. The equivalent score for electricity is about from government interference while remaining 60 percent, and for water about 50 percent. accountable to society. These aspects of regula- Transport scores about 50 percent on private tion have proved more challenging, with scores sector participation, but this development has remaining relatively low. not been accompanied by the broader legal and Governance lags behind other areas of structural reforms seen in the utilities sectors. institutional development, and the limited Interference from government continues to progress shows up mainly in internal mana- undermine regulatory independence in many gerial practices. Whereas the relevance of countries. Infrastructure regulation in Africa sectoral and regulatory reforms has generally is still in its early days. Typically, new laws and been well recognized, the governance regime regulatory bodies have been introduced for has received less attention from policy makers telecommunications and electricity, whereas and analysts. Almost all Sub-Saharan countries few countries have created water or transport ranked significantly and consistently lower on regulators. The quality of regulation can be this dimension of institutional development measured along several dimensions (figure 4.3, than on the others (figure 4.3, panel c). Most panel b). On the technical side, regulation needs countries are doing better on internal gover- to be founded on solid methodological tools, nance than on external governance. Internal and the resulting decisions need to be commu- governance relates to structures within the nicated to the public in a transparent manner. service provision entity, such as the extent African regulators score the highest on these to which its structure approximates standard dimensions, even if (in absolute terms) they still corporate forms; the qualifications and auton- have some way to go. On the political side, reg- omy of its senior management and board of ulation requires a certain degree of autonomy directors; the nature, quality, and timeliness 110 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 4.3 Institutional Progress on Reforms, Regulation, and Governance of the information it submits to its overseers; percentage score on institutional scorecard and the adoption of accounting and disclosure a. Sectoral reforms standards. External governance, by contrast, 100 refers to external market disciplines: being 90 subject to private rather than public sector 80 accounting and auditing systems, contract- 70 ing out noncore activities to private provid- 60 ers, and being obliged to raise debt or equity 50 funds on private capital markets, domestic or 40 international. 30 Only Kenya and South Africa have raised 20 much from external capital markets. Kenya 10 corporatized its power distribution utility 0 and more recently its generation firm, and n g ht r e to then issued a minority of shares in each on ag tio rin g ec rsi er a ctu s isl av ve te tru leg the Nairobi Stock Exchange. The 2006 initial yo iva s re lic pr po public offering of 30 percent of the shares of KenGen raised $35 million, a small but sig- b. Regulatory quality 100 nificant start. In South Africa, share issuance 90 was not considered feasible, but the national 80 utility, ESKOM, was corporatized, obtained 70 a credit rating, and then issued corporate 60 bonds--$120 million in 2007 alone. 50 The data generated in the sectoral chapters 40 and the institutional scorecard analysis shed 30 considerable light on the efficacy of the three 20 central pillars of infrastructure institutional 10 reform, namely, private sector participation 0 (PSP), state-owned enterprise governance, and regulators. y y ity ols e m nc ag bil to no re er ta pa to av un au ns co tra ac Does Private Sector Participation c. Governance quality Work? 100 90 The lessons learned from overall experi- 80 ence with PSP demonstrate sectoral nuances. 70 Whereas some sectors display a significant 60 extent of PSP that has brought about valuable 50 outcomes (mobile telephony, power genera- 40 tion, and ports), results are mixed in other sec- 30 tors (roads, power and water distribution). 20 The extent of private participation varies 10 significantly across sectors. Despite widespread 0 legislation in the region allowing private oper- ip y ity g et et ge om cin sh k k il ra ar ar ab r ators entry into infrastructure, implementa- ur on e ne m m av nt tso t ow au or al ou ou pit lab tion lags are common, particularly for water d c ac ar ca bo and railways (figure 4.4). telecommunications electricity water ports railways In Africa, most private participation in Source: Vagliasindi and Nellis 2009. water, electricity, railways, and ports has been Note: See box 4.1 for the definition of the institutional indicators. by methods other than full divestiture, that is, Building Sound Institutions 111 through management contracts, leases, and five percent of contracts in water have been concessions. Only in telecommunications has canceled, as have 15 percent in electricity divestiture been widely applied. The amount (table 4.2). Note that these cancellations do of private participation, of all sorts, varies by not include contracts that have undergone sector, with the most in telecommunications renegotiation because of the complaints of one and the least in water (figure 4.5). or both of the parties involved; nor do they As noted previously, private participation account for cases where anticipated renewals in Africa has had some problems. Twenty- of leases or, especially, management contracts have not occurred, leading to a resumption of state management. In all infrastructure sectors, Figure 4.4 Implementation of Private Participation contract negotiation, monitoring, and enforce- across Sectors ment have proved more time-consuming and 100 difficult than expected. 90 Despite these difficulties, the survey 80 undertaken for this chapter reveals significant gains from private participation in some sec- percentage of countries 70 tors and for certain aspects of performance. 60 A higher degree of private sector involvement 50 is associated with higher labor productiv- 40 ity (connections per employee), though the 30 link is statistically significant only in the case 20 of electricity and ports, and higher but not 10 statistically significant cost-recovery ratios. 0 In telecommunications, the countries with tio m- ity er s rts ay above-average private involvement display at ric po ica co ns ilw w ct un tele ra ele higher access in both the fixed and the mobile m allowed by law actually implemented segments of the market. More extensive private involvement in ports is associated with above- Source: Vagliasindi and Nellis 2009. average technical efficiency. Table 4.2 Cancellation of African Private Participation Figure 4.5 Private Participation in Management Contracts and Investment across Sectors Number of Percentage Type of contract contracts canceled 100 Water 90 Management contract 15 20 80 Lease contract 7 45 percentage of countries 70 Concession contract 4 50 60 BTO/BOO 1 0 50 Divestiture 1 0 40 Total 28 25 30 Electricity 20 Management or lease contract 17 24 10 Concession contract 16 31 0 Independent power project 34 6 tio m- ty r s rts te ay ici po ica co wa ns ilw ctr un tele ra ele Divestiture 7 -- m Total 74 15 management investment Source: Vagliasindi and Nellis 2009. Note: BOO = build-operate-own; BTO = build-transfer-operate. Source: Vagliasindi and Nellis 2009. -- Not applicable. 112 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Earlier empirical studies also find some Privatization of fixed-line incumbents has evidence of positive outcomes, albeit based affected access and productivity, and qual- on case study evidence and by no means in all ity of services somewhat, though the change instances. For example, a recent study of seven is not statistically significant (figure 4.6). The infrastructure privatizations in Africa assessed growth in the number of subscribers has been three factors: (a) efficiency gains and losses, low and in almost all countries negative, except (b) nature and competence of the transaction, Nigeria, the only country where competition and (c) who won and who lost (and by how has also been introduced in the fixed-line seg- much) in society because of the transaction ment of the market. Productivity is also low, (BIDE 2006). Three of the seven (Côte d'Ivoire compared with international benchmarks electricity, Senegal Airlines, and Senegal water) (lines per employee). were assessed as "unqualified success stories" Still, several private participation transac- in efficiency terms, according to a variety of tions in the fixed-line segment of these mar- financial and service quality measures. Three kets (the remaining natural monopoly) have others (Mozambique water and Uganda water run into problems. In the last few years, stra- and telecommunications) were assessed as tegic investors from developed countries have producing "some positive changes but less largely withdrawn from African telecommuni- than what most expect from privatization." cation privatizations. Only three such sales have Only Senegal electricity was classified as having occurred since 2001, and in those sales, no tra- "no significant effect" (BIDE 2006: 2). None of ditional strategic partner obtained a controlling the seven was assessed as negatively affecting stake. Recent telecommunication divestitures efficiency. have either been public offerings (South Africa Such studies found a close correlation and Sudan), sales to developing-country inves- between the competence in negotiating the tors (ZTE of China in Niger and Maroc Telecom transaction and the efficiency gains by the in Burkina Faso), or sales to domestic investors new private operator. The distributional effect (Malawi and Nigeria). In several instances, gov- could be assessed fully in two cases (Côte ernments have repurchased shares in incum- d'Ivoire electricity and Senegal water) and bent operators. That happened in Ghana and partially in the other five. Even with limited is planned in Rwanda. In Tanzania in 2005, information, the study found a correlation the government repurchased shares previously between institutional capacity and wider dis- privatized in its fixed-line operator, TTCL. In tribution of benefits. That is, the cases having May 2007, the government placed TTCL under the better institutional arrangements in con- a three-year management contract with a Cana- ducting the transaction show better outcomes dian firm, SaskTel. As of early 2009, the govern- for a broader range of stakeholders than do ment was considering canceling the contract, the cases where the transaction process was claiming that SaskTel had failed in its commit- rated lower. ment to raise nonguaranteed debt financing to Using the analysis of the sectoral chapters rehabilitate and expand the network. in this study, one can further assess the extent Access has spread quickly since 1998 because and effect of private participation in African of the rapid rise of mobile telephony, largely infrastructure, from the most extensive and from the combination of private participation successful involvement to the least. and increasingly intense competition. Private investment, the bulk of it greenfield, in cellular Telecommunications phone technology has allowed new providers to Private participation in telecommunications enter previously monopolistic markets, result- has taken place in the majority of Sub-Saharan ing in greater access and declining, though still countries. In 15 countries, at least partial priva- comparatively high, consumer prices. Analysts tization of the state-owned fixed-line telecom- generally argue that the dramatic increases munication incumbent has occurred. Licensing in African access to and coverage of telecom- new private mobile operators for greenfield munication services owe more to the entry of networks has been even more widespread. new mobile operators, thereby strengthening Building Sound Institutions 113 competition, than to the improved information Figure 4.6 Links between Private Sector Participation and Performance Indicators and incentives of private managers and owners. in Telecommunications Most new mobile operators are controlled by one of five multinational firms operating in total telephone subscribers (per 100 people) the region: France Telecom, MTC (Kuwait), MTN (South Africa), Millicom (Luxembourg), mobile subscribers (thousand or Vodacom (South Africa). of subscribers per employee) The salutary effects of competition are appar- ent. A strong link exists between liberalization in main telephone line this sector (and others) and better outcomes in (100 lines per employee) access and productivity (figure 4.7). Countries with lower market concentration in the mobile number of faults (per main line per year) segment of the market, measured by the Her- findahl-Hirschman Index,2 have much higher 0 20 40 60 80 100 penetration rates and productivity in the same segment of the market, as well as in the fixed- privatized state-owned enterprises line business, though none of these links are Source: Vagliasindi and Nellis 2009. statistically significant. Note: None of these performance differentials was found to be statistically significant at 5 percent. Ports By 2006, 20 port concessions were operating in Figure 4.7 Links between Market Concentration and Performance Indicators in Africa, with 6 more in process. Evaluations of Telecommunications these concessions indicate that delays, costs, and thefts were reduced and that port infrastructure total telephone subscribers started to improve. Cargo-handling rates and (per 100 people) the use of better handling systems in African "concessioned" ports are significantly higher mobile subscribers (thousand than in state-managed ports (figure 4.8). of subscribers per employee) Private sector involvement has grown greatly in container terminals in the region main telephone lines (100 lines per employee) since 2000 (now in eight countries and in pro- cess in several others). Case study evidence for Nigeria and Tanzania confirms the results of number of faults (per main line per year) the statistical evidence reported here (see box 4.2). Nonetheless, the level of private sector 0 20 40 60 80 100 penetration in the African port sector is low, compared with other regions. low market concentration high market concentration Source: Vagliasindi and Nellis 2009. Railways Note: None of these performance differentials was found to be statistically significant. Concessions to the private sector have been applied in 13 of the 24 countries reviewed, and the mechanism is presently under nego- The concession process is not always tiation in 3 countries and under consideration smooth. For example, the Kenya-Uganda con- in another 3 countries. Evaluations of several cessionaire, brought in with much fanfare in other longer-standing concessions, nonethe- 2006, has had difficulties in raising the prom- less, conclude that railways under concessions ised investment financing. The contract was perform more efficiently than those remain- renegotiated in late 2008 to reduce the share- ing in state hands, although the difference is holding of the original investor and to allow significant only in locomotive availability and the owning governments to seek new private coach productivity (figure 4.9). partners. 114 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 4.8 Links between Port Concessions and Performance Indicators Concessions have not resolved the key issue of mobilizing finance. The drawback annual (weighted average) of concessions is that they rarely produce the growth rate of container trade 1995­2000 anticipated (and in many cases, contracted) investments for network rehabilitation and expansion. The reason is that service revenues average moves per hour (unit: 100 moves)* are too low to support investment finance, partly due to low traffic volumes and intense intermodal competition and partly due to the container gantries as failure by governments to compensate conces- handling system sionaires for running loss-making obligatory (percent)** passenger services. Thus, chronic underinvest- 0 10 20 30 40 50 60 70 80 90 ment and dilapidated infrastructure remain as major railway problems. The limited invest- concessioned others ment capital forthcoming has been financed by international financial institutions and Source: Vagliasindi and Nellis 2009. Note: *Performance differential is statistically significant at the 10 percent level; **significant at passed through government to the private rail- the 5 percent level. way operators. BOX 4.2 Electricity The most common form of private participa- tion in the electricity sector has been inde- Privatization in African Ports pendent power projects (IPPs), with 34 in A lease in the container terminal in the port of Dar es Salaam, 11 African countries (Besant-Jones 2006). Tanzania, produced, within five years, a doubling of throughput, a Assessments of African IPPs have drawn 70 percent reduction in container dwell time, greater customer sat- attention to the lack of alternatives, given the isfaction, record profits, and vastly increased government revenues severe shortfalls in generation throughout the (from taxes, lease fees, and payments of $14 per container cleared). region and the very important fact that most The number of expatriate managers fell from 17 to 4; the number IPPs do produce the amounts of electricity of Tanzanian senior managers doubled. More than half the original called for in the contracts. Without IPP pro- workforce was dismissed, but salaries for those remaining increased by an average 300 percent--and postlease expansion in operations duction, the number and duration of service created 500 new jobs, far more than the number previously laid off. disruptions would be far higher than it has Dar es Salaam became the fastest container terminal in Africa, with been in recent years. performance exceeding that of many European and Australian ports. Low transparency in IPP negotiations In 2004, Nigeria started a major effort to reform its clogged, ineffi- translated into high costs. Many, if not most, cient, and very expensive ports. The government enacted "upstream" IPPs were negotiated hastily, in periods of cri- policy and legal reforms while hiring, through concessions, expe- sis, and competitive bidding processes were rienced private operators to manage, operate, and rehabilitate often amended or skipped entirely. That haste 26 ports. The new autonomous regional port authorities, now the resulted in extremely high costs that now "landlords" of the ports, negotiated the concession contracts. The pose a heavy financial burden on the balance Federal Ministry of Transport took on the role of sector policy maker. sheets of power purchasers: national distribu- Just a few months after the concessioning of the Apapa-Lagos con- tainer terminal, delays for berthing space had dwindled, and leading tion utilities (and their government owners). shipping lines reduced their congestion surcharge from $525 to $75, This situation, in turn, has led to widespread saving the Nigerian economy an estimated $200 million a year. Observ- suspicion that IPP negotiations were incom- ers credit the improvements as much to the upstream reform of the petently or corruptly managed (Gratwick and institutional setting as to bringing on board private operators. Eberhard 2008). Worth noting is that Tanzania and Nigeria are the top two coun- Regulatory benchmarking can be used to tries in institutional reforms. enhance transparency. A proposed method- Sources: World Bank 2005 for Tanzania; Leigland and Palsson 2007 for Nigeria. ology would facilitate a regulatory review of power purchase agreements by explicitly bench- marking them for price and risk allocation, to Building Sound Institutions 115 identify and ensure that the terms of the agree- Figure 4.9 Links between Rail Concessions and Performance Indicators ments are "fair and balanced" to all parties who will be directly and indirectly affected by these locomotive availability transactions (Besant-Jones, Tenenbaum, and (% of time)*** Tallapragada 2008). The Nigerian energy regu- lator has tested the methodology. cars (million ton- Management contracts are the second kilometer per car) most common form of electricity PSP, with 17 in 15 countries.3 Management contracts have coaches (million passenger- produced large and significant labor produc- kilometer per coach)* tivity gains (figure 4.10), though they have not proved sufficient, in and of themselves, staff (million units of traffic per staff) to overcome the broader policy and institu- tional deficiencies of the sector. The effects 0 0.5 1.0 1.5 2.0 2.5 3.0 of these contracts on cost recovery, system concession state-owned enterprises losses, and collection rates, however, turn out to be minimal and statistically insignificant Source: Vagliasindi and Nellis 2009. (figure 4.10). As with concessions in railways, Note: *Performance differential is statistically significant at the 10 percent level; ***significant at the 1 percent level. management contracts in electricity have not been instrumental in generating invest- ment funds. Several independent evaluations of man- Figure 4.10 Links between Electricity Management Contracts and agement contracts conclude that they pro- Performance Measures duced efficiency gains and improved financial performance (Davies 2004). The problem is labor productivity that they have not been sustained. Only 3 of (connections per employee)*** 17 negotiated contracts remain in operation. A few have been canceled. More commonly, implicit collection rates however, although technicians and donors (% on total electricity billed) recommend renewal of contracts after their initial phase, African governments, for mainly transmission and distribution loss sociopolitical reasons, choose to reestablish (% on total production) public management. Lessons from this experi- ence include the importance of setting targets cost recovery not only for commercial performance but also (ratio base 100) for improvements in quality of supply and ser- 0 20 40 60 80 100 120 140 160 180 200 vice, including expanded access, so that con- management contract other sumers experience tangible benefits. Effective contractual oversight is needed to track per- Source: Vagliasindi and Nellis 2009. Note: ***Performance differential is statistically significant at the 1 percent level. formance, to fairly assess and award incentive payments or penalties, and to reduce informa- tion asymmetry. Finally, postcontract manage- ment succession issues need to be addressed example, dating from the 1990s). The two in early (Ghanadan and Eberhard 2007). Gabon and Senegal have been judged as suc- cesses in producing service improvements, Water network expansions, and financial stability In water, leases have been applied rather widely, (BIDE 2006). The long-term leases in Côte with management contracts as the second most d'Ivoire and Niger have been producing posi- common form of private participation. Con- tive financial and operational results, despite cessions have also been used in several African difficult external conditions (Marin 2008). countries (in Chad, Gabon, and Senegal, for Only one divestiture has occurred, entailing 116 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION the sale of 51 percent of equity, which was in which are medium term in nature, broader the water company in Cape Verde in 1999. improvements in operational efficiency had Private participation in water has generated been observed. Access had also improved, much hostility and opposition. Because water although funding continued to come from the is a commodity essential for life, many feel that public sector. its distribution should be free or at a very low price. Despite considerable evidence of persis- What Have We Learned? tently poor performance by publicly owned and Any evaluation of the African experience with operated African water companies, many people private sector participation must be nuanced believe that the service cannot and should not by the wide variation in sector experiences be turned over to private delivery. To support (table 4.3), and both successes and failures their position, they point to several notable pri- considered in tandem. Public and official vate participation problem cases in water, such attention have focused more on the failures as Tanzania (see box 4.3) and Uganda. and contentious cases, particularly in water, A recent detailed empirical review of the but also in power and transport. However, the main experiences with private sector partici- lesson from African private participation is pation in the African water sector concluded not that the approach should be discarded, but that there had been beneficial effects on enter- that it should be applied selectively and care- prise performance in a number of cases but fully to those areas of infrastructure where it that the extent of these impacts depended on has a proven potential to contribute. the contractual forms that were used (Marin Expectations should also be kept realistic. 2008). In the case of management contracts, Experience has shown that there are only a few which are short term in nature, there had been niches where, either through raising investment benefits in terms of revenue collection and ser- finance or improving operational efficiency, the vice continuity but little effect on other aspects private sector can contribute significantly to of performance. In the case of lease contracts, investment finance--namely, ICT (particularly mobile networks), power generation, port con- tainer terminals, and a handful of high-traffic BOX 4.3 road segments (table 4.3). While the overall volume of private finance for infrastructure investment is limited, it is nonetheless sub- Lessons from the DAWASA Lease stantial, having (at least during the mid-2000s) Contract (Tanzania) exceeded the volume of ODA for these sectors The failed lease contract of DAWASA, the water authority for Dar es (recall table 2.1). But even in areas of infra- Salaam, Tanzania, is instructive. The contract was signed in August 2002. structure that have not proved attractive for It was supposed to run for 10 years, but the government canceled it in private finance--such as most roads, railways, May 2005, after only 21 months of operation. The government claimed power and water distribution systems--private the private provider had failed to meet water production and collection management can still make a significant contri- targets, pay the lease and other fees, meet service quality and quantity bution to improving operational performance, commitments, and pay the penalties assessed for noncompliance. For and thereby help to recover the very substantial its part, the private provider claimed that its bid and business plan were funds that are currently being lost to various based on inaccurate, out-of-date, or partial information provided by the government. Arbitration in April­May 2005 failed; the government ter- kinds of inefficiency (recall table 2.8). minated the contract, and the service returned to public management. Of course, one cannot deny the problems Critics note that the private provider's lawsuit against the government in African private participation, or that many for breach of contract was rejected (in 2008) by a British court. They of them can be attributed, in all sectors, to argue that this case shows how difficult it is to ensure that private provi- institutional deficiencies. Poor sectoral plan- sion of water will fulfill either the anticipated financial objectives or the ning, vague or absent sector policies, and distributional objectives. long-standing weak financial and operational Sources: BIDE 2006; Marin 2008. performance in the utilities helped create the crises of demand and insecurity that led to rushed decisions. In state-owned infrastructure Building Sound Institutions 117 Table 4.3 Overview of Experience with Private Participation in Infrastructure Infrastructure sector Extent of private participation Nature of experience Prospects ICT Mobile telephony Over 90 percent of countries have licensed Extremely beneficial with exponential Several countries still have potential to grant multiple mobile operators increase in coverage and penetration additional licenses Fixed telephony Some 60 percent of countries have Controversial in some cases, but has helped Several countries still have potential to undergone divestiture of SOE improve overall sector efficiency undertake divestitures telecommunication incumbent Power Power generation 34 IPPs provide 3,000 MW of new capacity, Few cancellations but frequent renegotiations; Likely to continue, given huge unsatisfied investing $2.5 billion PPAs have proved costly for utilities demands and limited public sector capacity Power distribution 16 concessions and 17 management or lease Problematic and controversial; one-quarter Movement toward hybrid models involving contracts in 24 countries of contracts canceled before completion local private sector in similar frameworks Transport Airports Four airport concessions, investing less than No cancellations but some lessons learned Limited number of additional airports viable $0.1 billion, plus some divestitures for concessions Ports 26 container terminal concessions, investing Processes can be controversial, but Good potential to continue $1.3 billion cancellations have been few and results positive Railroads 14 railroad concessions, investing $0.4 billion Frequent renegotiations, low traffic, and Likely to continue but model needs to be costly public service obligations keep adapted investment below expectations Roads 10 toll-road projects, almost all in No cancellations reported Limited because only 8 percent of road South Africa, investing $1.6 billion network meets minimum traffic threshold, almost all in South Africa Water Water 26 transactions, mainly management or Problematic and controversial; 40 percent Movement toward hybrid models involving lease contracts of contracts canceled before completion local private sector in similar frameworks Source: Authors' elaboration based on Bofinger 2009; Bullock 2009; Eberhard and others 2008; Gwilliam and others 2008; Minges and others 2008; Mundy and Penfold 2008; and Svendsen, Ewing, and Msangi 2008. Note: ICT = information and communication technology; IPP = independent power project; PPA = power purchase agreement; SOE = state-owned enterprise. firms, poor management, inadequate or non- 2008­09 will further deplete investor appetites existent record keeping (at both the firm and for comparatively high-risk ventures in emerg- higher levels), and lax monitoring created ing markets, heighten the reluctance of African organizational and informational chaos that officials to embark on innovative schemes, and reduced the interest of potential investors add weight to the notion of the primacy of the and severely complicated the due diligence public sector. Therefore, the existing level of processes of those who bid. Nontransparent state ownership is likely to persist, and may contract negotiation proceedings, substan- indeed increase, in the near to medium term. dard procurement practices, and inadequate This likelihood requires renewed attention to contract monitoring and enforcement mecha- a long-standing but recently neglected issue: nisms have been among the factors contribut- improving the financial and operational per- ing to higher prices, poorer-than-anticipated formance of state-owned firms (Nellis 2005; outcomes, renegotiated or canceled contracts, Gómez-Ibáñez 2007). and governance issues. These institutional deficiencies must be dealt with if private par- ticipation is to fulfill its potential. How Can State-Owned Enterprise Despite the inroads of private participation Performance Be Improved? over the past two decades, Africa remains the region with the highest state ownership of its Africa has the highest percentage of state- infrastructure utilities. The financial crisis of owned infrastructure utilities of any developing 118 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION region. Given the mixed record; the absence of Only limited success in achieving full cor- investor appetite; and the antipathy of African poratization, including establishing limited officials, nongovernmental organizations, and liabilities and introducing rate of return and many observers of development toward private dividend policies, has been recorded in Africa. participation--and the resulting increasing The telecommunication sector alone can be reluctance of the donor community to push reported as a success story, with electricity for privatization in infrastructure--state own- and water lagging, proving that even coun- ership is likely to be the norm for some time. tries with high scores compare poorly with The track record so far in governance other regions. reforms is not encouraging and varies sub- More limited corporate governance reforms stantially across sectors and countries. One have been started more evenly across all sec- must start with the recognition that a few tors and are becoming a dominant feature in African state-owned infrastructure firms have electricity and water. These changes include sustained good performance in the absence of the introduction of boards of directors (even if private participation. Botswana and Uganda the size tends to be either too large or too small show that fully state-owned African utilities compared with international standards), selec- can deliver high-quality performance (see tion of board members according to a com- box 4.4 and box 4.5, respectively). petitive process rather than direct appoint- The cost of inaction implies high hidden ments by line ministries, and the introduction costs. The estimated hidden cost of inefficiency of independent directors (figure 4.11). coming from mispricing, unaccounted losses, Performance contracts with incentives and and collection inefficiency is on average equal independent external audits have become to 0.6 percent of GDP in the water sector and dominant features of the governance reform pro- 1.9 percent of GDP in the power sector. The cess for both electricity and water (figure 4.12). inefficiency of SOEs can also be measured Independent audits have also been good for by excessive employment. In the telecommu- efficiency in both cases. nication sector, the hidden cost of excessive Of governance reforms that appear to be the employment is on average equal to 0.1 percent most important drivers of higher performance, of GDP. two appear especially promising: performance contracts with incentives and independent external audits (table 4.4). Uganda has had good experience with a performance contract BOX 4.4 in its water company, providing the utility with incentives for good performance and produc- Lessons from Successful SOE Reforms in ing greater accountability (see box 4.5). The introduction of independent audits has also Botswana Power Corporation positively affected efficiency for both electric- The state-owned and -operated Botswana Power Corporation ity and water utilities. has long provided reliable, high-quality service. Over the years, What can such cases teach? First, recast and Botswana Power has expanded its network in both urban and rural reapply the performance contract approach areas, covered its costs, posed no burden on the government bud- to SOE reform. Initial attempts to improve get, minimized system losses (10 percent), and earned a decent African SOEs using this device were mini- return on assets. Although the availability of cheap imported power from South Africa (now severely threatened) is part of the expla- mally effective, but recent efforts have had a nation for good performance, analysts give five institutional factors stronger and much more positive effect. The equal weight in explaining this success: (a) a strong, stable economy, more recent performance contracts applied (b) cost-reflective tariffs, (c) lack of government interference in mana- with some success in Uganda (and, reportedly, gerial decisions, (d) good internal governance, and (e) competent, in Kenya) should be studied and modified for well-motivated staff and management. broader application to African utilities across sectors. Source: PPA 2005. Second, renew efforts to strengthen the financial and operational monitoring of SOEs. Building Sound Institutions 119 BOX 4.5 Performance Agreement for the National Water and Sewerage Corporation (Uganda) Between 1998 and 2004, the National Water and Sewer- also increased the head office's commitment to provide age Corporation (NWSC) system operated under two man- financial and material resources to enable different areas to agement contracts with private providers. By the end of the implement rehabilitation and investment programs. second contract, neither party had an interest in continuing. In 2002, automatic tariff indexation was introduced. After 2004, public managers, operating under performance In addition, the Stretch-Out Program increased staff com- contracts, were responsible for the service. A review of per- mitment by improving internal communication and setting formance during the entire period concluded that the targets tougher performance targets and corresponding incentives. set for the private management contracts were fulfilled but A one-minute management system was introduced to further that the public management team furnished similarly good enhance individual staff members' accountability for targets. performance. The main stages in the enterprise reform pro- The government introduced a three-year performance cess are described below. contract in 2000. The NWSC's debt service obligations were From February 1999 onward, the management of the suspended in return for a commitment to operational and NWSC in Uganda has sequentially implemented a number financial improvements and an increase in coverage. of reform programs. First, local officials, called area service In 2003, a second performance contract continued the providers (areas), negotiated with central authorities a set of suspension of debt service and specified that NWSC's debt tightly defined performance targets. Second, area manag- would be restructured to a sustainable level. A review com- ers were given control over running the process. Third, they mittee monitored implementation of the agreement. The were held strictly accountable for specific results. main incentives of the agreements are bonuses for manag- A number of measures including the 100-day program ers and staff, if performance targets are achieved. and the service and revenue enhancement programs resulted in better specification of targets for the areas. The programs Sources: Baietti, Kingdom, and van Ginneken 2006; Vagliasindi 2008a. Figure 4.11 Prevalence of Good Governance Practices among State-Owned Enterprises for Infrastructure a. Ownership and shareholder quality b. Managerial and board autonomy 100 100 90 80 80 70 % of utilities % of utilities 60 60 50 40 40 30 20 20 10 0 0 telecommunications electricity water telecommunications electricity water corporatization limited liability size of the board selection of board members rate-of-return policy dividend policy presence of independent directors Source: Vagliasindi and Nellis 2009. Note: See box 4.1 for definitions of institutional indicators. Some of the structures implied in the Organi- function through an independent agency ver- sation for Economic Co-operation and Devel- sus a decentralized structure) have not yet opment's Principles of Corporate Governance been sufficiently "tested" in practice and may for SOEs (favoring a centralized ownership not suit all developing countries. A centralized 120 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 4.12 Prevalence of Performance Contracts in their economic performance, including the Electricity and Water detailed structure of subsidies and intersec- 100 toral arrears (Vagliasindi 2008b). 80 Do Independent Regulators Make % of utilities Sense? 60 Countries rank rather low on regulatory 40 independence across all sectors, confirming that the standard model has not fit the chal- 20 lenges in Africa well. Independence in several infrastructure sectors has been challenged 0 across all formal, financial, and managerial electricity water dimensions. Evidence on the links between performance contracts introducing an independent regulator and performance contracts with incentives improving performance is weak, but a signifi- penalties for poor performance cant positive effect is discernible in telecom- munications (table 4.5). Source: Vagliasindi and Nellis 2009. In water, where state-owned enterprises still predominate and are likely to for some time, countries with an independent regu- lator perform no better than those without Table 4.4 Links between Governance and Performance Indicators for Electricity and Water (table 4.5). This outcome may reflect the fact that donors have tended to be the largest pro- Connections Technical losses per employee Access moters of water regulatory agencies and they Reform/sector Yes No Yes No Yes No tend to assist on the most problematic situa- tions. In addition, many countries (particu- Performance contracts larly the francophone nations) have opted for Electricity 24.2a 23.3a 176.8** 103.0 14.6a 28.1a regulation by contract rather than create an Water 36.2 33.6 13.6 6.2 28.1 14.6 independent agency. Independent audit Nonetheless, hybrid regulatory schemes Electricity 22.9 28.3 164.3** 92.7 22.0 9.6 have not proved to be superior to traditional Water 35.2 35.7 7.6 6.0 9.6 a 22.0a forms of regulation in the water sector. Africa Source: Vagliasindi and Nellis 2009. Infrastructure Country Diagnostic data support Note: **Performance differential is statistically significant at the 5 percent level. no evidence on the superiority of regulation by a. The sign of the links between the variables is not as expected. contract over the traditional form of regulation by agency (see table 4.6 and box 4.6). For railways and ports, the regulatory func- structure, where the owner is the ministry of tion is generally entrusted to ministries of finance rather than an independent agency, is transport. Only Tanzania is establishing two more suited to the limited physical and human multisector regulatory agencies, one for pub- resource bases of most African countries. lic services and the other for transport. Mali Moreover, it has been implemented relatively and Senegal had railway regulatory agen- successfully in several developing countries. cies that were later converted into a common Under a decentralized or dual model--where railway-monitoring agency for both countries. the owner is the sectoral ministry or both a In railways, an independent regulator is still central authority (the ministry of finance or considered necessary, not so much to prevent treasury) and the sectoral ministry--the cen- exploitation of the monopoly power of the tral authority can collect and monitor infor- private sector as to protect the concessionaire mation about the state-owned enterprises and from the erratic behavior of governments, Building Sound Institutions 121 Table 4.5 Links between Regulation and Performance Indicators for Telecommunications, Electricity, and Water Technical losses Connections per employee Access Sector Regulation No regulation Regulation No regulation Regulation No regulation Telecommunications 0.2 0.3 0.38*** 0.03 0.2*** 0.1 Electricity 23.3 25.3 155.3 117.3 22.3 11.9 Water 35.2 34.8 6.8a 8.3a 36.1 35.9 Source: Vagliasindi and Nellis 2009. Note: ***Performance differential is statistically significant at the 1 percent level. a. The sign of the links between the variables is not as expected. Table 4.6 Links between Type of Regulator and Performance Indicators for Water Technical losses (% production) Connections per employee Access (% households) Regulation by Regulation by Regulation by Regulation by Regulation by Regulation by Sector contract agency contract agency contract agency Water 39.4 31.5 0.19 0.05 16.9 32.1 Source: Vagliasindi and Nellis 2009. Note: None of these performance differentials was found to be statistically significant. BOX 4.6 Regulation by Contract in Senegal In the highly successful case of Senegalese a specific escalation factor for key input prices, water, the officials negotiating the contract did and only included a general input price adjust- a world-class job of structuring a solid transac- ment factor. When oil prices roughly doubled tion and capturing almost immediate gains for shortly after transfer, the operator wanted consumers by using a creative compensation relief." The government was reluctant to allow structure with many of the efficiency properties the price increases demanded by the private of two-part utility pricing. By contrast, in the (minority) owner and operator, and the deal less-successful case of electricity, Senegalese collapsed. The firm returned to public hands negotiators of the electricity divestiture did not where, following a failed second attempt at anticipate a predictable problem: "Price was divestiture, it has done rather well. set by a perfectly reasonable price-cap [Retail Price Index ­ x] formula, but it failed to include Source: BIDE 2006. including the nonpayment of passenger service employees. Although regulatory requirements obligations to the concessionaires. differ with country size and income, that Weak regulatory autonomy and capac- difference does not fully explain the observed ity constraints undermine the credibility of variation in capacity. Contrast Sub-Saharan reg- the independent regulator (Eberhard 2007). ulatory budgets to those of the most developed Most Sub-Saharan regulatory agencies are countries: in 2005, the U.S. electricity regulator, embryonic, lacking funding and in many cases the Federal Energy Regulatory Commission, qualified personnel. Budgets vary consider- had a budget of about $240 million for 1,200 ably, ranging from less than $300,000 to about employees, and the United Kingdom's Office of $3 million for electricity. Staffing also varies the Gas and Electricity Markets had a budget of widely, from one or two to a couple of dozen $74 million for about 300 employees. 122 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Notes Bullock, Richard. 2009. "Taking Stock of Railway The authors of this chapter are Maria Vagliasindi Companies in Sub-Saharan Africa." Background and John Nellis, who drew on background mate- Paper 17, Africa Infrastructure Sector Diagnos- rial and contributions from Sudeshna Ghosh tic, World Bank, Washington, DC. Banerjee, Cecelia Briceño-Garmendia, Vivien Fos- Center for Systemic Peace. 2006. Political Regime ter, Yan Li, Elvira Morella, and Maria Shkaratan. Characteristics and Transitions. Fairfax, VA: 1. The World Bank's Country Policy and Institu- George Mason University. tional Assessment indicator reported in IDA Davies, Ian C. 2004. Management Contracts in 2008. the Electricity Sector: Case Studies in Malawi, 2. The index is calculated by squaring the market Lesotho, Tanzania and Rwanda. Washington, share of each firm competing in the market and DC: World Bank. summing the results. For example, for a market Eberhard, Anton. 2007. "Matching Regulatory consisting of four firms with shares of 30, 30, 20, Design to Country Circumstances: The Potential and 20 percent, the index is 2,600 (302 + 302 + for Hybrid and Transitional Models." Gridlines, 202 + 202 = 2,600). Note 23. Public-Private Infrastructure Advisory 3. Management contracts should not be confused Facility, World Bank, Washington, DC. with performance contracts. In a management Eberhard, Anton, Vivien Foster, Cecilia Briceño- contract, an owning government hires private Garmendia, Fatimata Ouedraogo, Daniel personnel to operate an SOE. The contractor Camos, and Maria Shkaratan. 2008. "Under- is paid a fee for service; normally, bonuses are powered: The State of the Power Sector in Sub- Saharan Africa." Background Paper 6, Africa awarded to the contractor if stipulated perfor- Infrastructure Sector Diagnostic, World Bank, mance targets are met. A performance contract Washington, DC. is a set of negotiations between a government and SOE managers from the public sector, spell- Ghanadan, Rebecca, and Anton Eberhard. 2007. "Electricity Utility Management Contracts ing out the obligations and responsibilities of in Africa: Lessons and Experience from the the two parties over a set period. Performance TANESCO-NET Group Solutions Manage- contracts can include incentives to SOE manage- ment Contract in Tanzania." Management Pro- ment (and staff); typically, the main issues speci- gramme in Infrastructure Reform and fied deal with the managers' obligations to meet Regulation Working Paper, Graduate School some targets and the government's obligation to of Business, University of Cape Town, allow price increases, provide investment capital, South Africa. settle past debts, pay bills on time, and so on. Gómez-Ibáñez, José A. 2007. "Alternatives to Infrastructure Privatization Revisited: Public References Enterprise Reform from the 1960s to the 1980s." Baietti, Aldo, William Kingdom, and Meike Policy Research Working Paper 4391, World van Ginneken. 2006. "Characteristics of Well- Bank, Washington, DC. Performing Public Water Utilities." Water Supply Gratwick, Katharine Nawaal, and Anton Eberhard. and Sanitation Working Note 9, World Bank, 2008. "An Analysis of Independent Power Washington, DC. Projects in Africa: Understanding Development Besant-Jones, John E. 2006. "Reforming Power and Investment Outcomes." Development Policy Markets in Developing Countries: What Have We Review 26(3): 309­38. Learned?" Energy and Mining Sector Board Dis- Gwilliam, Ken, Vivien Foster, Rodrigo Archondo- cussion Paper 19, World Bank, Washington, DC. Callao, Cecilia Briceño-Garmendia, Alberto Besant-Jones, John E., Bernard Tenenbaum, and Nogales, and Kavita Sethi. 2008. "The Burden Prasad Tallapragada. 2008. "Regulatory Review of Maintenance: Roads in Sub-Saharan Africa." of Power Purchase Agreements: A Proposed Background Paper 14, Africa Infrastructure Sector Benchmarking Methodology." Energy Sector Diagnostic, World Bank, Washington, DC. Management Assistance Program Formal Report IDA (International Development Association). 337/08, World Bank, Washington, DC. 2008. IDA: The Platform for Achieving Results BIDE (Boston Institute of Development Economics). at the Country Level, Fifteenth Replenishment. 2006. Impact of Privatization in Africa: Synthesis of Washington, DC: IDA. Eight Cases. Washington, DC: World Bank. Kaufmann, Daniel, Aart Kraay, and Massimo Bofinger, Heinrich C. 2009. "Air Transport: Chal- Mastruzzi. 2008. "Governance Matters VII: lenges to Growth." Background Paper 16, Africa Aggregate and Individual Governance Indica- Infrastructure Sector Diagnostic, World Bank, tors 1996­2007." Policy Research Working Paper Washington, DC. 4654, World Bank, Washington, DC. Building Sound Institutions 123 Leigland, James, and Gylfi Palsson. 2007. "Port Lessons from Private Participation in Infrastruc- Reform in Nigeria." Gridlines, Note 17. ture in Sub-Saharan Africa--Case Study of Bot- Public-Private Infrastructure Advisory swana Power Corporation." Report submitted to Facility, World Bank, Washington, DC. World Bank, Washington, DC. Levy, Brian. 2007. Governance Reforms, Poverty Svendsen, Mark, Mandy Ewing, and Siwa Reduction and Economic Management. Msangi. 2008. "Watermarks: Indicators of Washington, DC: World Bank. Irrigation Sector Performance in Sub-Saharan Marin, Philippe. 2008. Public-Private Partnerships Africa." Background Paper 4, Africa Infrastruc- for Urban Water Utilities: A Review of Experiences ture Sector Diagnostic, World Bank, in Developing Countries. Washington, DC: World Washington, DC. Bank. Vagliasindi, Maria. 2008a. "The Effectiveness of Minges, Michael, Cecilia Briceño-Garmendia, Boards of Directors of State Owned Enterprises in Mark Williams, Mavis Ampah, Daniel Camos, Developing Countries." Policy Research Working and Maria Shkratan. 2008. "Information and Paper 4579, World Bank, Washington, DC. Communications Technology in Sub-Saharan ------. 2008b. "Governance Arrangement for Africa: A Sector Review." Background Paper 10, State-Owned Enterprises." Policy Research Work- Africa Infrastructure Sector Diagnostic, World ing Paper 4542, World Bank, Washington, DC. Bank, Washington, DC. ------. 2008c. "Institutional Infrastructure Indi- Mundy, Michael, and Andrew Penfold. 2008. cators: An Application to Reforms, Regulation "Beyond the Bottlenecks: Ports in Sub-Saharan and Governance in Sub-Saharan Africa." World Africa." Background Paper 8, Africa Infrastruc- Bank, Washington, DC. ture Sector Diagnostic, World Bank, Vagliasindi, Maria, and John Nellis. 2009. "Evalu- Washington, DC. ating Africa's Experience with Institutional Nellis, John. 2005. "The Evolution of Enterprise Reform for the Infrastructure Sectors." Working Reform in Africa: From State-Owned Enter- Paper 23, Africa Infrastructure Sector Diagnos- prises to Private Participation Infrastructure-- tic, World Bank, Washington, DC. and Back?" Research Paper 117, Fondazione Eni World Bank. 2005. "Second Port Modernization Enrico Mattei, Milan. and Railway Restructuring Projects, Project PPA (Power Planning Associates Ltd.). 2005. Performance Assessment Report." World Bank, "Towards Growth and Poverty Reduction: Washington, DC. Chapter 5 Facilitating Urbanization A frica is urbanizing fast, a change that is to link national economies with regional and predictable and beneficial. Economic global markets. geography indicates that prosperity and African cities are growing fast, but because density go together because higher productivity of insufficient infrastructure and poor institu- requires agglomeration economies, larger tions, most new settlements are informal and markets, and better connectivity. Concentra- not covered by basic services. This situation tion and urbanization trigger prosperity in has severe consequences for health, incomes, urban areas as much as in rural areas, and and market integration. A combination of well-functioning cities facilitate trade and the institutional reform, land policy and planning, transformation of rural production and non- housing policies, and basic services is required farm activities. The debate over rural or urban for urban expansion that is more equitable and development should thus be replaced by the inclusive in nature. understanding that rural and urban develop- Many necessary investments are beyond the ment are mutually dependent and that eco- limited fiscal and financial base of African cities. nomic integration of rural and urban areas is Decentralization has increased the responsibili- the only way to produce growth and inclusive ties of cities, but not their powers and incen- development. tives to raise (and retain) revenues. Cities need Populated places in Africa need infrastruc- access to predictable streams of revenue and ture to enhance the competitiveness of their the flexibility to raise additional resources, to businesses and the productivity of their work- safeguard service provision to their constituen- ers. Energy, roads, water, and information and cies. They also need to improve their technical communication technologies (ICTs) give Afri- and managerial capacity to deal with priorities can economies the capacity to develop. Long- in investment and operation and maintenance, run growth requires an efficient system of to guide the inevitable expansion, to attract urban centers that includes small, medium, and private partners, and to understand their sur- larger cities that produce industrial goods and rounding neighborhoods to develop synergies. high-value services, along with well-functioning Africa's large agricultural sector and rural transportation networks (roads, rails, and ports) economy remain central for overall economic 125 126 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION growth and poverty reduction. Better infra- services. Addressing bottlenecks in city per- structure is crucial for raising agricultural formance is an effective entry point into this productivity and facilitating access to markets "virtuous" circle (Kessides 2006). for agricultural products. The Asian experi- Proximity to cities (neighborhood effect) ence suggests that successful economic growth is critical for enabling the shift from subsis- demands higher agricultural productivity, tence to commercial agriculture, for increasing which raises incomes and the demand for rural incomes, and for making living standards nonagricultural products, lowers food prices, converge. Areas within two hours' travel time and frees up labor for (mainly urban) indus- of cities of at least 100,000 people seem to trial and service employment (World Bank have diversified into nonagricultural activities 2008).1 (Dorosh and others 2008). Rural areas located The policy challenge is to harness market between two and eight hours' travel time from forces that encourage concentration and pro- such cities account for more than 62 percent of mote convergence in living standards between the agricultural supply and generate a surplus villages, towns, and cities. Policy makers will sold to urban areas. In areas farther than eight be more effective if they look at development hours from these cities, agriculture is largely strategies for broad economic areas that inte- for subsistence, and less than 15 percent of grate towns and cities with their surrounding the land's agricultural potential is realized rural hinterland. This chapter discusses and (table 5.1).2 Similarly, farmers closer to cities estimates the infrastructure needs of rural and tend to use more and higher-quality fertilizers urban areas in the context of rapid urbaniza- and pesticides and better equipment, result- tion and its challenges for infrastructure, insti- ing in clear improvements in productivity. So tutions, and targeted interventions. the growth of urban markets is a key factor in raising the income of the rural population in the hinterland. Viewing Cities as Engines of Growth Strengthening Urban-Rural Links The debate on growth strategies has often looked at urban and rural areas as compet- An integrated approach to development rec- ing for primacy in the national agenda and ognizes and facilitates the links between urban in investment allocations, but now is the time and rural areas. Urban centers consume rural to frame the debate differently. Cities exist products and offer inputs for rural produc- because of the economic and social advan- tion; rural areas serve as markets for goods and tages of closeness. Urban centers contribute services produced in urban areas. Migration to national economic growth by increasing produces social and economic links between individual, business, and industry produc- urban and rural areas. Migrants often remain tivity through agglomeration economies; by connected to their families, which they support increasing household welfare through social through remittances. In addition, rural people mobility and human development; and by often receive health services and education from promoting positive institutional change. Cit- nearby towns and cities. Institutional and fiscal ies also drive rural development, serving as links are often present as well. In most cases, primary markets for rural production and fiscal redistribution takes place--typically generating income that flows back to rural from cities to rural areas, given the cities' larger areas. Links between urban and rural areas tax base. constitute a virtuous circle, where access Rural-urban links are constrained by inad- to urban markets and services for nonfarm equate transport networks, poor electricity and production stimulates agricultural produc- water provision, and limited coverage of ICT. tivity and rural incomes, which in turn gen- Weak institutions add to the constraints. For erate demand and labor for more goods and example, in Ethiopia, the city of Dessie enjoys Facilitating Urbanization 127 a strategic location and is a main distribution In several fragile states, civil war has con- center for manufacturing products to the sur- tributed to urban expansion as people from the rounding regions (World Bank 2007). However, affected regions seek refuge in cities. One-third the lack of a developed agroprocessing industry of Africa's urban population is concentrated limits the market opportunities for higher-value in the region's 36 megacities with more than agricultural production and the benefits for the 1 million inhabitants. Much of the remain- surrounding agricultural areas. der is spread across 232 intermediate cities of between 100,000 and 1 million inhabitants Urbanization in Africa and in periurban areas. The largest cities are Africa's population remains predominantly growing fastest, suggesting that Africa's urban rural. About 66 percent of the inhabitants live population will become more concentrated. in rural areas, with significant variation across As is typical in the early phases of urbaniza- countries (table 5.2).3 In African middle-income tion, urban household incomes in Africa are countries, half of the inhabitants live in rural much higher than rural incomes, almost twice as areas, whereas in the landlocked low-income high. The 2009 World Development Report notes countries, they account for about 70 percent. that an economy's transformation is seldom The vast majority of Africa's rural population geographically balanced (World Bank 2009). (or half of the overall population) lives in the Productivity tends to increase where people and rural hinterland within six hours' travel time economic activities concentrate to take advan- of cities having at least 50,000 inhabitants, tage of agglomeration economies. The initial whereas about 16 percent of Africa's population growth spurt is typically associated with a diver- lives in isolated areas more than six hours' travel gence in living standards between leading regions time from the same cities. The continent is urbanizing rapidly, how- Table 5.1 Link between Agricultural Productivity and Distance to Urban Centers ever, and will become predominantly urban by 2020. The share of urban population rose from Percentage of Percentage of Per capita Percentage of total total crop production 15 percent in 1960 to 35 percent in 2006 and Travel time total area population production ($ per capita) will reach nearly 60 percent by 2020. Urban Less than 1.7 hours 10.0 41.4 23.6 57.00 growth is presently estimated at 3.9 percent a 1.7­7.6 hours 50.0 46.0 62.5 135.80 year. Rural migration accounts for one-quarter More than 7.6 hours 40.0 12.5 13.9 110.70 of that growth, with the remainder attributable to urban demographic growth and adminis- Total 100.0 100.0 100.0 n.a. trative reclassification (Farvacque-Vitkovic, Source: Dorosh and others 2008. Note: Totals may not add exactly because of rounding errors. Glasser, and others 2008). n.a. = Not applicable. Table 5.2 Distribution of Population by Type of Settlement and Country Type GNI per Percentage of total population capita Intermediate Secondary Periurban Rural Remote Country type ($ per capita) Megacities cities cities areas hinterlands rural areas Sub-Saharan Africa 875 13.4 10.2 0.2 10.3 49.8 16.4 Low-income countries, landlocked 245 8.3 8.3 0.2 7.5 56.5 19.2 Low-income countries, coastal 472 11.1 6.7 0.2 12.0 46.3 23.6 Middle-income countries 5,081 24.6 15.8 0.4 12.5 50.1 1.6 Source: Authors' compilation based on geographic information systems analysis of Global Rural-Urban Mapping Project population density data. Note: GNI = gross national income. GNI per capita is in current dollars using the Atlas method. Estimates are based on a panel of 20 coun- tries. Megacities have more than 1 million people; intermediate cities have between 100,000 and 1 million people; secondary urban areas have between 100,000 and 50,000 people; periurban areas are less than one hour from the nearest city with more than 50,000 people; rural hinterlands are between one and six hours from the nearest city with more than 50,000 people; and remote areas are more than six hours from the nearest city with more than 50,000 people. 128 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION (mostly urban) and lagging regions (mostly than 20 percent of the continent's GDP, despite rural), but as incomes increase, the divergence is accounting for more than 60 percent of the followed by convergence.4 Essential household population (table 5.4; Kessides 2006). Recent consumption converges soonest, access to basic work in Tanzania confirms these values: the public services next, and wages and income later. country's urban areas are home to 23 percent Convergence occurs because of the mobility of of the population yet account for 51 percent of people and resources across regions and declin- the GDP (Maal 2008). One could infer that the ing economic distances among regions. average productivity in urban areas is at least African countries stand at the beginning three times that of rural areas (Farvacque- of this process. The difference between urban Vitkovic, Glasser, and others 2008). and rural incomes explains the lower urban poverty rates (35 percent) vis-à-vis rural pov- Infrastructure in Urban and Rural Areas erty rates (52 percent) (table 5.3). In absolute For population centers to realize their full eco- terms, the African rural poor are almost three nomic potential, the provision of infrastructure times as numerous as the urban poor. This pic- and public services must be efficient. Basic ser- ture holds for every country regardless of its vices for households in both urban and rural geography. Similar differences are observed in areas can guarantee sustainable urbanization access to services. and social equity, enhance living conditions, The incomes in urban areas reflect the and prevent disproportionate flows of under- higher productivity possible thanks to agglom- served rural people to the city. Investment in eration economies--the gain in efficiency infrastructure can improve productivity in the from having many businesses and workers in modern sector and connectivity with and across proximity. In countries where traditional non- locations. Deficiencies in infrastructure and mechanized agriculture still dominates the services, which limit the potential for agglom- rural sector, the difference between urban and eration economies, hinder African economies rural productivity can be large. Assuming that and may explain the underperformance of agricultural output originates in rural areas businesses in Africa relative to other continents. and industry and services originate in urban One-third of African businesses report a worri- areas, the African rural sector contributes less some lack of electricity, and 15 percent identify Table 5.3 Economic Differentials between Rural and Urban Populations, by Country Type Monthly household budget Poverty rate National Rural Urban National Rural Urban Country type ($/month) ($/month) ($/month) (percent) (percent) (percent) Sub-Saharan Africa 144 106 195 48 35 52 Low-income countries, landlocked 86 75 139 49 32 53 Low-income countries, coastal 145 115 209 47 38 51 Middle-income countries 535 256 691 n.a. n.a. n.a. Source: Authors' compilation based on household surveys reported in Banerjee and others 2008. Note: n.a. = not applicable. Table 5.4 Sectoral Contributions to GDP and GDP Growth percentage 1990­95 1996­2000 2001­05 Item Agriculture Industry Services Agriculture Industries Services Agriculture Industries Services Contribution to GDP 17 31 52 17 30 53 19 31 50 Contribution to GDP growth 59 ­28 69 14 30 56 26 37 37 Source: Authors' compilation based on National Accounts data. Facilitating Urbanization 129 transportation as a major constraint. Poor- Every year since 1990, an additional 0.9 per- quality roads and other transportation infra- cent of the urban population has gained access structure endanger connectivity between rural to improved water and 1.7 percent to improved and urban areas, between products and mar- sanitation, whereas the corresponding rural kets, and between workers and labor markets. figures stand, respectively, at only 0.3 and 0.4. The difference in the coverage of basic Electricity service has been expanded to an infrastructure services is huge across urban additional 3 percent of urban residents but to and rural areas. For the spectrum of household only an additional 0.8 percent of rural residents. services, urban coverage rates are between 5 Even so, rampant demographic pressure in and 10 times those in rural areas (figure 5.1). urban areas has caused coverage rates to decline The absolute differences are largest for power for all urban services (particularly improved and smallest for ICT services. Electricity and water), whereas coverage of all rural services improved water supply (such as piped connec- has increased (particularly power and ICT). As tion or standpost) extend to a majority of the a result, the gap between urban and rural cover- urban population, but to less than one-fifth of age rates has narrowed slightly but at the cost the rural. An even smaller share in rural areas of leaving urban dwellers and businesses with- uses septic tanks or improved latrines; and access out infrastructure for domestic and industrial to ICT services remains negligible. In almost purposes (figure 5.2). This finding shows that half of the countries, energy coverage barely urban service providers have struggled to keep reaches 50 percent of the urban population and pace with accelerating urbanization. 5 percent of the rural. In addition, fewer than Africa's sparse road density often leaves 40 percent of African urban households enjoy rural areas isolated from urban markets. Only a private water connection, a septic tank, or an one-third of Africans living in rural areas are improved latrine, a share that falls to 5 percent within 2 kilometers of an all-season road. in rural areas. The paved-road density is 134 kilometers per Growth in urban and rural coverage of 1,000 square kilometers of arable land, and the network infrastructure tends to be positively unpaved, 490 kilometers. Moreover, the qual- correlated. Countries with faster expansion of ity of the rural network is perceptibly lower urban coverage of water and electricity also than that of the main network, with almost tend to have faster expansion of rural cover- half in poor condition (figure 5.3). The lack age, suggesting that an urban network eases of adequate urban transport is an obstacle for expansion toward rural areas. It may also sug- businesses and for labor mobility. gest that urban customers cross-subsidize rural The spatial footprint of infrastructure net- water networks and electrification. works is larger than the coverage rates would suggest. In the rural hinterlands, where the bulk of Africa's rural population lives, 40­50 percent Figure 5.1 Access to Infrastructure by Location of people live within range of an infrastructure network. Even in isolated rural areas, the share is 100 as high as 15 percent. This information suggests 90 80 that hookup rates to infrastructure networks are 71 lower in rural areas. In some cases, that likely % of population 70 63 60 reflects much lower rural purchasing power. In 50 42 others, technical limitations might prevent rural 40 inhabitants from connecting to infrastructure 30 22 20 20 14 networks, even when close to one. 12 10 2 3 7 Infrastructure investment (particularly in 0 rural areas) continues to focus on sector-specific landline cell phone improved improved electricity sanitation water interventions rather than spatially synchroniz- rural urban ing and concentrating the provision of different infrastructure services in larger "bundles." Avail- Source: Banerjee and others 2008. able evidence suggests that bundling of services 130 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 5.2 Change in Urban and Rural Service Coverage, 1990­2005 a. Improved water b. Improved sanitation 100 100 80 80 % of population % of population 60 60 40 40 20 20 0 0 1990­95 1996­2000 2001­05 1990­95 1996­2000 2001­05 c. Electricity d. Landline 100 100 80 80 % of population % of population 60 60 40 40 20 20 0 0 1990­95 1996­2000 2001­05 1990­95 1996­2000 2001­05 rural urban Source: Banerjee and others 2008. Figure 5.3 Quality Differentials between Main and costs because the interaction across services Rural Road Networks compounds positive effects, such as time savings 100 or increased connectivity. With barriers to pro- 90 ductivity reduced faster than when services are 80 taken individually, poor households have more 70 60 chances to access economic opportunities. percent 50 In the African context, not only are levels of 40 household service coverage low, but they are 30 also uncoordinated. As a result, the share of 20 10 households with access to a bundle of multiple 0 infrastructure services is very low, even among main network rural network the better-off (table 5.5). good fair poor Source: Gwilliam and others 2008. The Costs of Providing leads to higher returns among beneficiary Infrastructure--Sensitive households than when the services are pro- to Density vided individually. For example, in Peru joint access to two or more services generates a larger The cost of infrastructure network expansion increase in rural household welfare than when is highly sensitive to population density. For services are accessed separately (Torero and the exact same infrastructure bundle, in both Escobal 2005). This finding is valid for urban urban and rural spaces, the capital cost (per dwellers as well, at least for water, sanitation, capita) declines with density. At the highest electricity, and telephone services, regardless density, the cost of a bundle of high-quality of how they are combined (Chong, Hentschel, services is $325 per capita; for medium-density and Saavedra 2007). Access to multiple services cities, it is $665; for the rural hinterland $2,837; also generates a larger reduction of opportunity and for isolated areas $4,879 (table 5.6). These Facilitating Urbanization 131 Table 5.5 Households with Access to One or More Modern Infrastructure Services percentage Quintile Number of services National Rural Urban Poorest Second Third Fourth Fifth Any one modern infrastructure service 33 15 76 4 17 23 44 78 Any two modern infrastructure services 17 4 47 0 2 7 19 56 Any three modern infrastructure services 9 1 28 0 0 3 11 32 Any four modern infrastructure services 4 0 12 0 0 1 4 16 Source: Banerjee and others 2008. Note: Household coverage rates are population weighted for the latest available year. Modern infrastructure services include piped water, flush toilet, power, and landline telephone. Table 5.6 Capital Cost per Capita of Infrastructure Provision, by Density $ per capita except as otherwise noted Secondary Rural Deep Infrastructure type Large cities cities hinterland rural Density (people/km2) 30,000 20,000 10,000 5,008 3,026 1,455 1,247 38 13 Water Private tap 104.2 124.0 168.7 231.8 293.6 416.4 448.5 1,825.2 3,156.2 Standpost 31.0 36.3 48.5 65.6 82.4 115.7 124.5 267.6 267.6 Borehole 21.1 21.1 21.1 21.1 21.1 21.1 21.1 53.0 159.7 Hand pump 8.3 8.3 8.3 8.3 8.3 8.3 8.3 16.7 50.4 Sanitation Septic tank 125.0 125.0 125.0 125.0 125.0 125.0 125.0 125.0 125.0 Improved latrine 57.0 57.0 57.0 57.0 57.0 57.0 57.0 57.0 57.0 Unimproved latrine 39.0 39.0 39.0 39.0 39.0 39.0 39.0 39.0 39.0 Power Grid 63.5 71.2 88.5 112.9 136.8 184.3 196.7 487.7 943.1 Minigrid 87.6 95.2 112.5 136.9 160.8 208.3 220.7 485.8 704.2 Solar photovoltaic 92.3 92.3 92.3 92.3 92.3 92.3 92.3 92.3 92.3 Roads High quality 31.6 47.4 94.7 189.2 313.1 651.3 759.8 269.1 232.4 Low quality 23.6 35.4 70.7 141.2 233.8 486.3 567.3 224.3 193.6 ICT Constant capacity 1.1 1.7 3.3 6.6 10.9 22.8 26.6 39.8 129.7 Actual capacity 1.1 1.7 3.3 6.6 10.9 22.8 26.6 129.7 422.1 Total Variable qualitya 325 369 480 665 879 1,031 1,061 940 836 b Constant (high) quality 325 369 480 665 879 1,400 1,557 2,837 4,879 Source: Authors' compilation based on numerous Africa Infrastructure Country Diagnostic sources. Note: ICT = information and communication technology. a. For variable quality, technology differs by density and location as follows: (a) water--private tap in large cities, standposts in small cities, boreholes in secondary urban cities, hand pump in rural areas; (b) sanitation--septic tanks in large cities, improved latrines in small and secondary urban cities, traditional latrines in rural areas; (c) power--grid in urban areas, minigrid in rural hinterland, solar in deep rural areas; (d) roads--high-quality scenario; (e) ICT--constant capacity in urban and rural areas. b. For constant (high) quality, the same technology--the most expensive one--applies at any density except for power (grid at any level of density). 132 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION values illustrate the cost disadvantage of Investment Needs African cities (generally less dense), compared with their higher-density Asian counterparts. As noted in chapter 2 of this volume, Africa Africa's urban expansion is occurring with needs to spend some $93 billion per year to declining densities (urban sprawl), which in meet its infrastructure needs over the next itself will make per capita infrastructure costs decade to 2015. About two-thirds of this total even higher. relates to capital investment and the remaining For Africa's largest cities, with a popula- third to operation and maintenance (Briceño- tion over 3 million and a median density of Garmendia, Smits, and Foster 2008). These 5,000 people per square kilometer, water and investment needs can usefully be divided into sanitation represent the heaviest weight in the three categories. The first category relates to the infrastructure bundle (54 percent), followed investment needed to expand productive infra- by roads (28 percent), power (17 percent), and structure that underpins the national economy ICTs (1 percent). as a whole: for example, generation capacity Economies of density are so important that to serve industrial production;5 transmission the rollout of network infrastructure becomes lines; fiber-optic backbones; and the main prohibitive at low levels of density. In these components of the national transport system cases, it would make sense to shift toward a including trunk roads, railways, airports, and package of lower-cost technological alterna- seaports. The second category relates to infra- tives, such as solar panels, hand pumps, and on- structure specifically needed to service the site sanitation as population density falls. The urban space, including urban roads and urban cost of variable-quality technology rises more ICT, power, and water distribution networks. gradually--from $325 per capita in high- The third category relates to infrastructure density cities to $665 in medium-density cities specifically needed to service the rural space, to $940 in the rural hinterland; it then drops including rural roads, rural household services, back to $836 in isolated areas (table 5.6). The and irrigation. implication is that the highest per capita cost Historically, Africa has been investing around is found in secondary urban areas. Densities $26 billion a year in infrastructure,6 as reported there are high enough to demand higher- in chapter 2. Of this total, about 30 percent goes quality solutions but still not high enough to to productive infrastructure that underpins the benefit from significant economies of scale in national economy, 50 percent goes to servicing the delivery of services. the urban space, and the remaining 20 percent Population density affects not only the to servicing the rural space (figure 5.5).7 The cost of network expansion but also the avail- lion's share of investment in the power sector ability of resources to pay for it. Aggregate has financed the development of energy capac- household spending capacity per square ity for industrial production and transmission. kilometer ranges from some $3,500 a year Infrastructure linked to the national economy in deep rural areas to $2.5 million a year in also accounts for an important share of invest- cities with populations over 3 million and a ments in transport. Across all sectors, infra- density of 5,000 people per square kilometer. structure dedicated to servicing the urban space Therefore, in rural areas the cost of a high- absorbs a greater share of investments than quality infrastructure bundle is 10 to 20 that dedicated to servicing the rural space. The times the annual household budgets, making only exception is water and sanitation, where a it manifestly unaffordable (figure 5.4). This more even split is observed, although even that ratio falls steeply in urban areas, where the represents something of an urban bias, given cost of the bundle is one to three times the that the population is predominantly rural. annual household budget. For high-density Looking ahead, Africa will need to invest cities (beyond the range observed in Africa), $60 billion per year, according to the esti- this ratio falls to less than two-thirds of the mates presented in chapter 1 of this volume. annual household budget. The spatial pattern of these future investments Facilitating Urbanization 133 will look somewhat different from those of Figure 5.4 Affordability of a Basic Package of Household Infrastructure the past: 34 percent of the total will need to 18 ratio of investment cost to annual go to productive infrastructure that under- household budget by sq. km pins the national economy; another 32 per- 15 affordable not affordable cent will go to servicing the urban space, and 12 the remaining 34 percent to service the rural 9 space (figure 5.6). Thus, the share going to the national economy increases slightly, whereas 6 the share dedicated to servicing urban spaces 3 declines and the share dedicated to rural space increases; this pattern is observable across the 0 00 00 00 00 00 0 6 2 4 7 4 8 38 13 individual sectors also. 34 87 27 99 61 45 18 ,0 ,0 ,0 ,0 ,0 9, 4, 3, 2, 2, 1, 1, 30 25 20 15 10 population density (no. people/sq. km) Infrastructure Financing Source: Authors' compilation. Note: Affordable represents those population densities in which the capital cost per hectare of a basic The jurisdiction responsible for infrastructure package of infrastructure services represents no more than one year of income for the inhabitants of that hectare. Not affordable represents those population densities in which the capital cost per hectare financing and provision differs hugely across of a basic package of infrastructure services exceeds, by several multiples, the annual income for the inhabitants of that hectare. sectors and countries. Although ICT and power are usually national responsibilities, respon- sibility for the water supply in urban areas is widely decentralized (figure 5.7). Nevertheless, Figure 5.5 Spatial Split of Historic Infrastructure Investments in many countries, markedly the francophone 100 countries, operation is entrusted to utilities that remain national. Where municipal utili- 80 ties do exist, municipal governments own only percent 60 a few in whole or in part. Responsibility for transport infrastructure is divided between 40 national and local jurisdictions, with bound- 20 aries varying from one country to another. The 0 central government is typically responsible for power ICT transport water and the trunk road network, as well as railways, sanitation ports, and airports. Local governments are national urban rural typically responsible for local roads. Most local jurisdictions, whether urban or Source: Authors' compilation. Note: ICT = information and communication technology. National economy refers to productive rural, lack the resource base to provide ade- infrastructure that underpins the national economy and cannot be specifically attributed to servicing quate infrastructure services to households and inhabitants of urban or rural space (for example, the interurban trunk network, the national power interconnected system, major ports, and airports). businesses. Municipal budgets are very small in Urban space or rural space refers to infrastructure that is primarily oriented toward servicing needs of urban or rural inhabitants, respectively (for example, urban or rural household services, urban or relation to the cost of meeting infrastructure rural roads). requirements implied by fast urban growth. Data collected from a sample of cities suggest that the average expenditure per person per year rarely exceeds $10 (figure 5.8). South Africa is 80 percent of their operating revenues. This an exception, with Cape Town at $1,163 and dependence diminishes the incentives for Durban at $1,152. local governments to raise their own revenues. Transfers from central governments or Transfers are often unpredictable, hindering direct financing have become the most impor- long-term projects and planning. They often tant source of funding for local infrastructure favor small localities over larger cities with (table 5.7). Cities in most African countries serious infrastructure bottlenecks. On the depend on central transfers for more than revenue side, local governments have limited 134 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 5.6 Spatial Split of Future Infrastructure Investment Needs taxing authority, even though the potential 100 revenue base is large. For large cities, most of the revenues collected are transferred to the 80 national treasury. The use of local government budgets var- 60 ies widely across countries. In Ghana, revenues percent are mainly dedicated to capital expenses (78 40 percent), whereas in Côte d'Ivoire and Senegal, current expenditures occupy the largest account 20 (80 percent and 90 percent, respectively). 0 In Ghana, local taxes and local resources are power ICT transport water and irrigation almost nonexistent, whereas in Côte d'Ivoire sanitation and Senegal, they make up a significant share productive infrastructure urban rural of the revenue envelope. Own revenues col- lected by local governments are less than 1 Source: Authors' compilation. Note: ICT = information and communication technology. Productive infrastructure underpins the percent of GDP in all three countries. national economy and cannot be specifically attributed to servicing inhabitants of urban or rural space (for example, the interurban trunk network, the national power interconnected system, major ports, and airports). Spatial Planning, Land Regulations, Urban or rural refers to infrastructure that is primarily oriented toward servicing needs of urban or rural inhabitants, respectively (for example, urban or rural household services and urban or rural roads). and Housing Urban boundaries have expanded, and zones that were rural in the mid-1980s are now Figure 5.7 Institutional Patterns of Water and Electricity Supply in Urban Areas part of the metropolitan areas surrounding major African cities. This trend is particularly a. Water supply b. Electricity evident in the countries with faster urban 4% growth, such as Burkina Faso, Cameroon, and Ghana (Farvacque-Vitkovic, Glasser, and 39% others 2008). The pattern reveals strong physi- cal growth, typified by moderate and patchy densification within the inner-city core, as 61% residential areas give way to commercial users and peripheral growth occurs unguided and 96% decentralized not decentralized at low density. In many cases, spatial expansion of cities Sources: Banerjee and others 2008; Eberhard and others 2008. has been faster than population growth, reduc- ing density. Density in African cities ranges Figure 5.8 Municipal Budgets of Selected African Cities between 1,000 and 4,000 people per square kilometer (table 5.8). Density in rural areas is Arusha, Tanzania fewer than 100 people per square kilometer. Beira, Mozambique Africa's megacities are denser but still far from Maputo, Mozambique their Asian peers. Only a handful of African Nairobi, Kenya cities (such as Ifon Osun in Nigeria, Mbuji- Mayi in the Democratic Republic of Congo, or Porto Novo, Benin Mombasa in Kenya) attain population densities Lagos, Nigeria of 10,000 to 20,000 people per square kilometer, Dar es Salaam, Tanzania compared with densities of 20,000 to 40,000 in Kinshasa, Congo, Dem Rep. of Bangalore, Hyderabad, and Mumbai in India; Monrovia, Liberia Guangzhou and Shanghai in China; and Seoul in the Republic of Korea (Bertaud 2003). The 0 5 10 15 20 25 30 35 US$ per capita more typical densities for African megaci- ties of no more than 5,000 people per square Sources: UN-Habitat 1998; corresponding city budget documents. kilometer are comparable to those of cities of Facilitating Urbanization 135 Organisation for Economic Co-operation and land market and the monopoly by traditional Development member states, such as London landowners lead to shortages of urban land and New York. Built-in areas are growing faster supply and rising prices. Lack of land titles than urban populations in 7 of 10 African hurts business development and the estab- cities, suggesting falling densities (Angel, lishment of new firms. With no access to land Sheppard, and Civco 2005).8 As a result, the and located in underserved and peripheral already high costs for infrastructure in urban areas, the poor suffer from poor connectivity Africa will increase with further sprawl, hin- and low access to labor markets. The resis- dering the affordability of basic services and tance of landowners and the lack of registries adding to the environmental and carbon also hinder cities from raising revenues on footprints. urban land (box 5.1). In many African countries, land institu- African cities face shortages of housing and tions are still incipient, reflecting the political shelter. In most countries, the real estate mar- economy and colonial legacy. In Africa, land- ket and government agencies supply at most ownership is made difficult by the extreme one-fourth of the annual demand for hous- centralization of procedures, the costs of ing; the remaining three-fourths turn to the titling, and the rapid depletion of central gov- ernments' land reserves. In this context, spon- taneous settlements have developed. Limited Table 5.7 Overview of Local Infrastructure Financing Mechanisms land supply and high prices affect location Mechanism Urban Rural decisions and exclude low-income house- Direct central provision ICT, power, sometimes water ICT, irrigation, occasionally power holds from the official land market. Many gov- and water ernments have subsidized plots, but supply is Central government transfers Not earmarked and usually modest Not earmarked and usually modest; much below demand. Governments have tried growing use of rural funds to help residents excluded from land markets Local taxes Significant potential but requires Much more limited potential and have expanded infrastructure to new set- clear property rights and greater tlements, but the results have been disappoint- formalization of urban economy ing. Recent work in Burkina Faso, Ghana, and Municipal debt Limited number of No creditworthiness creditworthy cases Mali reveals the coexistence of traditional and public systems of property rights, complicating Property sales Significant potential but requires Much more limited potential clear prior definition of land titles registries and duplicating land rights, which User fees Significant potential Minor potential because of lower are difficult to enforce (Farvacque-Vitkovic purchasing power of households and others 2007, Farvacque-Vitkovic, Raghu- Source: Authors' compilation. nath, and others 2008). The limited size of the Note: ICT = information and communication technology. Table 5.8 Population Density across Country Types GNI per Density in Density in Density in Density Density capita Density in intermediate secondary periurban in rural in remote Country type ($ per capita) megacities cities urban areas areas hinterlands areas Sub-Saharan Africa 875 3,621 1,482 1,281 89 35 13 Low-income countries, landlocked 245 2,529 1,702 1,306 132 71 14 Low-income countries, coastal 472 4,083 1,661 1,492 100 35 13 Middle-income countries 5,081 1,229 574 824 58 45 19 Source: Authors' compilation based on geographic information systems analysis of Global Rural-Urban Mapping Project population density data and Henderson 2002 for megacities. Note: GNI = gross national income. GNI per capita is in current dollars using the Atlas method. Estimates are based on a panel of 20 countries. Density = number of people per square kilometer. Megacities have more than 1 million people; intermediate cities have between 100,000 and 1 million people; secondary urban areas have between 100,000 and 50,000 people; periurban areas are less than one hour from the nearest city with more than 50,000 people; rural hinterlands are between one and six hours from the nearest city with more than 50,000 people; and remote areas are more than six hours from the nearest city with more than 50,000 people. 136 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 5.1 housing, but the effect of these incentives has been minimal (Farvacque-Vitkovic, Glasser, and others 2008). Construction costs are very high, Land Issues in Tanzania especially for landlocked countries.9 Cement, iron, and other materials are imported, making Tanzania is at the early stage of its urbanization, and the population is growing in most urban areas. Since the 1960s, demand for urban housing prices unaffordable. Most new houses land with services has significantly and systematically exceeded what are captured by medium-income households. the government has supplied. Official demand for plots averages According to estimates, of five applications, 75,000, but the supply is below 6,000 a year. Most applicants lose only one housing unit is allocated to a needy hope and turn to informal markets to obtain land for their develop- household.10 ment needs. Serious difficulties also exist in accessing land for investment pur- Urban Growth and Informal poses. The Tanzania Investment Centre has so far registered 4,210 Settlements investment projects, 80 percent of which require access to land par- A major problem for Africa's growing cities is cels. However, the Tanzania Investment Centre estimates that only the rapid spread of informal settlements. Lack one-quarter of those registered can acquire land through the exist- of affordable serviced plots and zoning poli- ing formal system. Between 2004 and 2007, 440 applications for land allocation were received, but only 13 applicants received their cies have often excluded the poor from being titles, evidence that the formal land delivery system in urban areas is integrated with urban development, leaving not working. In 1998, informal settlements in Dar es Salaam covered them in underserviced shelters (slums) both 48 percent of the built-up area, and recent estimates place the share in and on the outskirts of major cities. The as high as 70­80 percent. UN-Habitat standard definition of people The scarcity of formal plots stems from the underperformance of living in slums is overcrowding, low access the public system, the lack of institutional or financial resources, and to water and sanitation, lack of secure ten- the competition of the formal system with informal land develop- ure, and poor housing quality. Based on this ment, the latter being much cheaper and easier. Institutional con- definition, as much as 70 percent of Africa's straints also exist. The 1999 Land Act concentrated most powers of urban population resides in slums (UN-Habitat land management in the central government, depriving local govern- 2003), and from 1990 to 2001, the slum pop- ment authorities of the institutional structures to deal with their land- related responsibilities. Goals for the land are not explicit, and laws ulation grew at 4.4 percent a year, faster than are not applied consistently. The division of responsibilities among the urban population. If this trend persists, an actors is unclear, and there is no coordination. additional 218 million Africans will be living in slums by 2020, and almost one-third of the Sources: Muzzini and Lindeboom 2008; Raich and Sarzin forthcoming. world's slum population will be in Africa. Great heterogeneity exists across cities in the living and income conditions of slum residents. Not all people living in these settle- ments are income poor, although significant informal market and their own construction. overlap occurs, and living standards differ In Burkina Faso and Mali, more than 45 per- across countries. In many cases, informal cent of the population in the capital city lives settlements are scattered across cities, side by in informal settlements (Farvacque-Vitkovic side with wealthier residences. In Tanzania, and others 2007). In Ghana, the market pro- urban dwellers in the periurban areas do well, duces only 20 percent of the annual demand enjoying the freedom of informality, even for shelter (Farvacque-Vitkovic, Raghunath, without land titles or finished walls. Indeed, and others 2008). In Accra, people's increased informal periurban areas are sometimes the purchasing power because of remittances has most dynamic, precisely because overzealous led to higher real estate prices, and zoning regulation does not affect them. A strong "city and legal restrictions artificially limit the land effect" also exists. Thanks to leadership, land available for housing. security, ownership, and civic participation, Many governments have tried to help the the inhabitants of Dakar's slums (Senegal) poor by subsidizing construction, providing tax have living standards far superior to Nairobi's incentives to developers, or producing public (Kenya) even though the latter have higher Facilitating Urbanization 137 incomes and education levels (Gulyani, Taluk- the creditworthiness to raise their own debt dar, and Jack 2008). finance, and the few examples (South Africa The main problem of slums and informal- and Zimbabwe) have illustrated the weakness ity is exclusion from basic amenities. Nai- of the system and the need for impartial credit robi's slum residents pay up to 11 times more ratings. for water sold by private vendors than those For larger cities, the economic base is larger, who have access to piped water (Farvacque- and they have greater autonomy to raise their Vitkovic, Glasser, and others 2008). In Africa own taxes. However, tax receipts are often sent as a whole, the price of piped water is $0.50 to the central governments, and political fac- per cubic meter, whereas water purchased tors hinder the use of property taxes. Although from private vendors in mobile carts is $4.75 African cities generate 80 percent of the coun- per cubic meter. Inadequate access to basic try's tax revenues, they end up with less than services also has implications for health and 20 percent of the resources.11 On the other human development. Moreover, spatial mis- hand, larger urban areas are likely to have matches and distance constrain accessibility to fewer spending responsibilities because they education and livelihood opportunities. are often covered by national service providers (such as power and water utilities), thus reliev- ing these municipalities of some budgetary Policy Issues and Implications expenses they would otherwise incur. In cities such as Dar es Salaam, Tanzania, Infrastructure Financing and Nairobi, Kenya, improvement in land Cities should be spending more and spending management institutions could open the door more wisely. Although trunk infrastructure to increased property revenues for munici- and services with substantial spillover effects palities, more land use and sales revenues for are clearly the responsibility of the central gov- the government, and complementary private ernment, cities are responsible for solid waste, finance. For example, Cairo has held sev- sewerage, drainage, and lighting. In many eral auctions of land for conversion, adding cases, they are called on to help with shelter as 10 percent to the city budget. It has also well. The purported benefits of decentraliza- swapped land permits for private infrastruc- tion are not realized because decentralization ture in public land. For large Chinese cities, policies have given cities more responsibili- land leases are the usual method to mobilize ties (notably in social sectors) but not more resources, as well as in Mumbai and Bangalore resources. Without independent or predict- in India. In 11 African countries, street address- able sources of revenues, African cities can ing (or addressage) has been established in the rarely plan or decide on the best way to allo- major municipalities, increasing municipal tax cate their resources. billings by about 50 percent, with 90 percent In principle, cities have greater potential collection rates. In Burkina Faso, Mauritania, to raise local revenues. First, the larger urban and Togo, street addressing has helped in economy provides a significant local tax base, inventorying the local tax base and implement- although its predominantly informal nature ing residential taxation (Farvacque-Vitkovic prevents the authorities from capturing taxes. and others 2007; Farvacque-Vitkovic, Glasser, Second, high-value urban properties constitute and others 2008; Kessides 2006). In Benin, a major potential tax base, although the lack decentralized management fostered increases of clear property titles prevents it from being in city revenue of 82 percent in Cotonou, realized. Third, the higher purchasing power 131 percent in Parakou, and 148 percent in and tighter agglomeration of urban house- Porto Novo, with better collection rates as holds make recovery of a significant propor- well (Kessides 2006). tion of infrastructure financing requirements In rural areas, transfers complemented by with user fees easier. centralized funds dominate. Many countries Long-term debt is always an alternative (at have tried to fund rural investment by intro- least in theory), but few African cities have ducing centralized funding mechanisms to 138 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION channel earmarked central government funds businesses and households to locate in the city and donor resources to rural infrastructure. and pushing prices up, thereby leading to a These mechanisms include rural water funds relocation of activities and residents to non- (90 percent of countries), rural electrification regulated places. Land acquisition delays are funds (76 percent), and rural telecommunica- very long in Ethiopia and Zambia. In Mozam- tion funds (29 percent). For power, rural funds bique, businesses pay on average $18,000 in bring faster expansions in the electrified rural processing fees for land, and in Nigeria, they population: an annual increase of 0.72 percent must register land to use it as collateral, a pro- is seen in countries with such funds, compared cess that can take up to two years and cost to an annual contraction of -0.05 percent else- 15 percent of the land value (Kessides 2006). where. For rural water and rural ICT funds, no Land institutions can improve information, significant differential was seen in the rate of strengthen property rights, record market trans- service expansion. actions, and steadily move toward more open Central funds can also support the mainte- land markets. With an endowment of 63 square nance of rural infrastructure. To deal with poor miles of land, the Tema Development Corpo- maintenance, many countries are allocating ration in Ghana is planning and laying out the part of their national road fund revenues to the Tema area, constructing roads and sewerage sys- maintenance of the rural network (60 percent tems, preparing and executing housing projects, of countries). This decision may be a good and managing rental units. Permits for housing strategy: countries that allocate at least $0.015 construction are submitted to the authority, per liter of their fuel levies to the rural road which charges a permit fee based on the value networks have a significantly higher share of the property to be developed (Farvacque- of their rural roads in good condition than Vitkovic and McAuslan 1993). those that allocate less, 36 percent versus Urban planning should guide urban expan- 21 percent. sion and the associated infrastructure needs. Because of their top-down approach and weak Land Policies and Urban and implementation, urban planning and master Territorial Planning plans have lost their meaning in many Afri- Adequate land policies and markets are the can cities. Urban dynamics are seldom cor- backbone of an efficient urban transition. Land rectly foreseen, and in most cases, the political management institutions include a compre- economy has the last word in determining the hensive land registry, credible mechanisms for location of infrastructure or major develop- contract enforcement and conflict resolution, ments. To be efficient and useful, planning flexible zoning laws, and versatile regulation should be flexible, participatory, and indica- of subdivisions that help rather than hinder tive (10­15 years). Urban reference maps the conversion of land for different uses. Prop- should lay out major roads and city services, erty rights embodied in land titles are essential the areas for urban expansion, and the reserves for converting assets into usable wealth. The for amenities. transformation of the agricultural sector from Planning should check sprawl, enhance communal land rights to individual property densification, prevent development in pre- rights is important for the rural-urban trans- carious environmental zones, and enhance the formation, but it may take a long time. delivery of affordable serviced land and infra- Land use and building regulations become structure. Ideally, planning should be rooted in more important as urbanization advances. participatory strategies and linked to local and Governments regulate land markets to ensure central budgets. Without realistic projections separation of land between different uses for resource availability, urban plans often fall and to ensure the integration of public and into discredit. Dakar (Senegal), Lagos (Nigeria), private uses of land, such as providing space and Maputo (Mozambique) recently prepared for transport infrastructure in densely popu- city development strategies as frameworks to lated areas. Land regulations can be overzeal- encourage participation from the community ous, however, distorting the incentives for in discussing challenges and opportunities. Facilitating Urbanization 139 Territorial planning is critical in rural areas basic infrastructure services, and shelters are to promote a more integrated vision of devel- upgraded into better and more durable con- opment and to enhance growth opportunities. structions. Investment in transportation and Rural development requires the coordinated social programs is also used to strengthen provision of infrastructure services to support links between slum areas and the rest of the agricultural production and off-farm activities, city and to facilitate social integration. The such as irrigation infrastructure, rural roads, and Accra District Rehabilitation Project in Ghana associated transport services, as well as storage is an example of successful upgrading, as are and distribution infrastructure for agricultural several national upgrading programs in Ethio- products. In rural areas, the limited administra- pia, Kenya, and Uganda. tive capacity hinders an integrated vision. Coor- The second, more controversial, approach dination may be further complicated because focuses on resettling slum dwellers, by moving some services, such as irrigation, may be a cen- them either to existing neighborhoods or to tral government competency, whereas others new, less crowded, and safer locations. In either (such as roads) are local. Chongqing, China's case, slum dwellers are compensated for reset- experience with implementing a territorial tlement and disruptions to their livelihoods. development plan on a regional scale provides Much can be done, starting with the provi- an example that may be of interest to Africa. sion of basic services and infrastructure com- Bundling infrastructure services can sub- bined with effective land policy. Legislation that stantially increase the return of infrastructure boosts land prices and excludes the poor will investments. Bundling infrastructure not only need to be revised. Basic packages of services secures larger welfare gains to households, (street lighting, paving, drainage, roads) should urban and rural alike, it also maximizes the be extended to the broadest number of people at economic and social effect of infrastructure the lowest possible cost. In Kenya, Mozambique, service provision in rural areas by granting and Nigeria, major improvements are possible better access to economic opportunities and for a low of $150 per capita, compared with reducing the gap between poor and nonpoor. $1,800 or more for finished solutions. Urban Therefore, investment policies, especially when transportation can bring the urban poor into rural infrastructure is concerned, should more the large labor markets. A consistent and exten- attentively seek complementarities across sec- sive policy of land titling would provide clarity tors. This requires that institutional coordina- and predictability to the land market, develop tion and planning and financial capacity are people's ownership, and promote private invest- deployed as needed. Although bundling is an ment. In many countries, however, land titling opportunity to realize larger returns on invest- will continue to be difficult (more difficult than ments, alone it is not enough to drive economic extending basic services) because of the political and social development in rural areas. A more economy and weak administrative capacity. far-reaching vision of rural development that maximizes coordination and complementa- rities across sectors beyond the infrastructure Six Principles for Efficient field--for which, however, bundling consti- Urbanization tutes an important tool--is needed. This is the scope of territorial development. Based on the foregoing discussion, there are six key principles for achieving efficient urbaniza- Informal Settlements tion. First, adopt a solid analytical framework to Preventing the formation of slums and upgrad- help define priorities and sequencing. In places ing existing ones are major concerns to policy that are mostly rural, governments should be makers; one of two possible approaches nor- neutral and establish the foundations for effi- mally prevails. The first approach focuses on cient urbanization (World Bank 2009). Good improving the living standards of slum dwell- land policies and universal provision of basic ers in their existing locations. They are given services are central. Where urbanization has land tenure; slums areas are equipped with accelerated, the priority should be investments 140 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION in connectivity to ensure that the benefits of broadly planning for the extension of urban rising economic density are widely shared. In settlements, taking into account transporta- highly urban places, targeted interventions may tion, connectivity, and environmental factors. be needed to deal with slums and exclusion. Sixth, improve the fiscal soundness of cities: Second, recognize that the political econ- (a) improve transparency and predictability of omy influences the urban transition. African transfers; (b) strengthen and simplify local cities are not very powerful. Unlike cities in taxation, changing the focus of property tax East Asia or Latin America, African cities have from ownership to occupancy; (c) take advan- little autonomy and depend on the central gov- tage of cost recovery from revenue-producing ernment for resources, infrastructure projects, services, such as markets and bus stations-- and even land development. Chinese mayors they can make up 70 percent of medium-size are appointed by the party, but their political city revenues; and (d) use municipal contracts careers depend on how well they develop their (between central and local governments) and cities. Especially if they are elected, U.S. and adressage to help local governments manage European mayors may see their cities as steps their resources. to higher political positions, including the presidency. Mayors in Africa have limited free- dom of action. Many of the difficulties Afri- Notes can cities have in collecting property taxes are The authors of this chapter are Elvira Morella, related to the political influence of the major Maria Emilia Freire, and Paul Dorosh, who landowners, who oppose such taxes. drew on background material and contribu- Third, be pragmatic. While the long-term tions from Alvaro Federico Barra, Catherine Farvacque-Vitkovic, Matthew Glasser, Sumila goal is to have well-defined property rights Gulyani, Darby Jack, Austin Kilroy, Barjor and land titling, in the short run, cities may Mheta, Stephen Mink, Siobhan Murray, Madhu need to "finesse" land titling and use occu- Raghunath, Uri Raich, Raj Salooja, Zmarak pancy as a basis for land registration and Shalizi, and Debabrata Talukdar. taxation. Resource-constrained governments 1. Note that the early stages of this transition need should invest in minimum packages of water- not involve movements of activities or people; sanitation-energy in informal underserviced rural households increasingly earn incomes quarters at the citywide level and resist the from rural nonagricultural activities (in agri- idea of transforming slums into perfect neigh- cultural processing, construction, commerce, borhoods. For about $1,200 per capita, many and private services). African slum dwellers can be provided with 2. This correlation between agricultural produc- tion and proximity (as measured by travel time) basic services, compared with $18,000 spent in to urban markets holds even after taking agro- more comprehensive and sophisticated proj- ecology into account. ects in Latin America (Farvacque-Vitkovic, 3. No internationally accepted standard exists for Glasser, and others 2008). identifying urban areas, and each country tends Fourth, focus on cities and areas important to use its own definition. This situation hinders for the economy. The priorities should be to any effort to make sensible comparisons across improve the institutional framework (espe- countries. In this chapter, urban areas, from sec- cially on land markets), to provide techni- ondary cities to megacities, are identified using cal and financial resources for planning and a subset of the GRUMP (Global Rural-Urban developing infrastructure and basic services, Mapping Project) urban extents layer (CIESIN to harness agglomeration economies, and to 2004). The GRUMP urban extents were joined to a data set of city populations compiled by Hend- deal with congestion. erson (2002) and of urban extents classified by Fifth, improve land policies so that mar- population size. To complete the urban-rural kets are more flexible and can respond to the gradient, nonurban areas were classified by dis- increase in demand. That requires compiling tance or travel time to the nearest city. The com- inventories of government land and of for- bination of urban extents and city populations mal and informal developers, gathering prices allowed creation of a density-based typology of and costs for land plots and construction, and cities. However, given the limitations associated Facilitating Urbanization 141 with these input data, the calculation of density Banerjee, Sudeshna, Quentin Wodon, Amadou is approximate at best. Even so, the density-based Diallo, Taras Pushak, Hellal Uddin, Clarence characterization of "urban" areas allows com- Tsimpo, and Vivien Foster. 2008. "Access, Afford- parisons across regions and reflects the relation ability, and Alternatives: Modern Infrastructure between density and agglomeration economies. Services in Africa." Background Paper 2, Africa Infrastructure Country Diagnostic, World Bank, 4. World Development Report 2009 expressed geo- Washington, DC. graphic transformation as the development of the leading and lagging regions. Although both Bertaud, Alain. 2003. "Order without Design." regions may consist of both urban and rural http://www.alain-bertaud.com. areas and hence agriculture and nonagriculture, Briceño-Garmendia, Cecilia, Karlis Smits, and in South Asia, lagging regions are predomi- Vivien Foster. 2008. "Financing Public Infra- nantly rural and agriculture remains the main structure in Sub-Saharan Africa: Patterns, Issues, source of livelihood. and Options." Background Paper 15, Africa Infrastructure Country Diagnostic, World Bank, 5. As mentioned, the most important hindrance Washington, DC. to economic production is the supply of energy. In many countries, it can account for half of the CIESIN (Center for International Earth Science value of the final output. Information Network). 2004. Global Rural-Urban Mapping Project (GRUMP): Urban Extents. 6. These funds include annual public investments, Palisades, NY: CIESIN, Columbia University. annualized official development assistance, annu- alized emerging financiers that do not belong to Chong, Alberto, Jesko Hentschel, and Jaime the Organisation for Economic Co-operation Saavedra. 2007. "Bundling of Basic Public Ser- vices and Household Welfare in Developing and Development, and annualized private par- Countries: An Empirical Exploration for the ticipation in infrastructure. Case of Peru." Oxford Development Studies 35 7. These figures include energy and roads; no data (3): 329­46. are available on the share of investment on ICT Dorosh, Paul, Hyoung-Gun Wang, Liang You, and water that serves industrial production. and Emily Schmidt. 2008. "Crop Production 8. Population figures are derived from national and Road Connectivity in Sub-Saharan Africa: censuses. A Spatial Analysis." Working Paper 19, Africa 9. The price of construction is estimated at $222 per Infrastructure Country Diagnostic, World Bank, square meter, so a 75-square-meter home would Washington, DC. cost about $17,000 (excluding land) (AGETIPE Eberhard, Anton, Vivien Foster, Cecelia Briceño- 2005). Average incomes are $850 a year, produc- Garmendia, Fatimata Ouedraogo, Daniel ing a ratio of housing price to income of over Camos, and Maria Shkaratan. 2008. "Under- 2,000 to 1, one of the world's highest. powered: The State of the Power Sector in Sub- 10. Efforts to develop mortgage systems for low- Saharan Africa." Background Paper 6, Africa income households in Africa have encountered Infrastructure Country Diagnostic, World Bank, several problems: lack of credit history, lack of Washington, DC. regular income, shallowness of the financial Farvacque-Vitkovic, Catherine, Alicia Casalis, market, lack of long-term funding, lack of land Mahine Diop, and Christian Eghoff. 2007. and house registries, high lending rates, and "Development of the Cities of Mali: Challenges high credit risks. and Priorities." Africa Region Working Paper 11. Too often, the central government appropriates 104a, World Bank, Washington, DC. the city-generated taxes and distributes them to Farvacque-Vitkovic, Catherine, Matthew Glasser, sectors that are not necessarily the most pro- Barjor Mehta, Madhu Raghunath, Austin Kilroy, ductive and certainly not the ones that will feed Alvaro Federico Barra, and Raj Salooja. 2008. into the urban economy. "Africa's Urbanization for Development: Under- standing Africa's Urban Challenges and Oppor- tunities." World Bank, Washington, DC. References Farvacque-Vitkovic, Catherine, and Patrick AGETIPE (Agence d'exécution des travaux McAuslan. 1993. "Politiques Foncière des Villes en d'intérêt public pour l'emploi). 2005. Contract Développement." World Bank, Washington, DC. documents. www.agetipe.org. Farvacque-Vitkovic, Catherine, Madhu Raghunath, Angel, Schlomo, Stephen C. Sheppard, and Daniel Christian Eghoff, and Charles Boakye. 2008. L. Civco. 2005. The Dynamics of Global Urban "Development of the Cities of Ghana: Chal- Expansion. Washington, DC: Transport and lenges, Priorities and Pools." Africa Region Work- Urban Development Department, World Bank. ing Paper 110, World Bank, Washington, DC. 142 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Gulyani, Sumila, Debabrata Talukdar, and Darby Empirical Base for Policy Dialogue. Washington, Jack. 2008. "A Tale of Three Cities: Understand- DC: World Bank. ing Differences in Provision of Modern Raich, Uri, and Zara Sarzin. Forthcoming. Services." Working Paper 10, Africa Infrastruc- Financing the Urban Expansion in Tanzania. ture Country Diagnostic, World Bank, Washington, DC: World Bank. Washington, DC. Torero, Maximo, and Javier Escobal. 2005. "Meas- Gwilliam, Ken, Vivien Foster, Rodrigo Archondo- uring the Impact of Asset Complementarities: Callao, Cecilia Briceño-Garmendia, Alberto The Case of Rural Peru." Cuadernos de Economia Nogales, and Kavita Sethi. 2008. "The Bur- 42 (May): 137­64. den of Maintenance: Roads in Sub-Saharan Africa." Background Paper 14, Africa Infra- UN-Habitat (United Nations Human Settle- structure Country Diagnostic, World Bank, ment Programme). 1998. "Global Urban Washington, DC. Observatory." http://ww2.unhabitat.org/ programmes/guo. Henderson, J. Vernon. 2002. "World Cities Data." http://www.econ.brown.edu/faculty/henderson/ ------. 2003. "Global Urban Observatory." http:// worldcities.html. ww2.unhabitat.org/programmes/guo. Kessides, Christine. 2006. The Urban Transition in World Bank. 2007. "The Challenge of Urbaniza- Sub-Saharan Africa: Implications for Economic tion in Ethiopia: Implications for Growth and Growth and Poverty Reduction. Washington, DC: Poverty Alleviation." Africa Region, Water and The Cities Alliance. Urban Development Unit 2, Washington, DC. Maal, Simen Jansen. 2008. "Measuring the Con- ------. 2008. World Development Report 2008: tribution of Urban Centers to the National Agriculture for Development. Washington, DC: Economy of Tanzania." Report to World Bank, World Bank. Dar es Salaam, Tanzania. ------. 2009. World Development Report 2009: Muzzini, Elisa, and Wietze Lindeboom. 2008. Reshaping Economic Geography. Washington, The Urban Transition in Tanzania: Building the DC: World Bank. Chapter 6 Deepening Regional Integration W ith many small, isolated economies, management and development of cross-border Africa's economic geography is public goods. Road and rail corridors linking particularly challenging. Regional landlocked countries to the sea are an example integration is likely the only way to overcome of such a regional public good, as are regional these handicaps and participate in the global airport and seaport hubs. The same can be said economy. Integrating physical infrastruc- of Africa's 63 international river basins. ture is both a precursor to and an enabler for Reaping these benefits, however, poses insti- deeper economic integration, thereby allowing tutional challenges: countries to gain scale economies and harness regional public goods. For successful regional · Building a political consensus. The politi- integration, countries must start small; build on cal obstacles can trump the economic case. successes; think globally, linking Africa to more Regional infrastructure involves a high external markets; and compensate the least for- level of trust between countries, not least tunate, recognizing that benefits are not always because of the implied dependence on evenly distributed. neighbors for key resources such as water The benefits of regional integration are visible and energy. across all aspects of infrastructure networks. · Establishing effective regional institutions. For information and communication technol- Regional institutions have to facilitate ogy (ICT) and power, regional infrastructure agreements and compensation. Africa has provides scale economies that substantially an extensive architecture of regional politi- reduce the costs of production. Thus, con- cal and technical bodies, but these face tinental fiber-optic submarine cables could problems because of overlapping member- reduce Internet and international call charges ships, limited technical capacity, and lim- by one-half. Similarly, regional power pools ited enforcement powers. that allow countries to share the most cost- · Setting priorities for regional investments. effective energy resources can reduce electric- Given the daunting investment agenda, ity costs by $2 billion a year. For transport and better sequencing and priority setting of water, regional collaboration allows optimal regional projects has been elusive. Political, 143 144 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION economic, and spatial approaches to prior- the world requires large up-front investments ity setting have all been widely discussed. in undersea cables or satellite communication. For private engagement in ICT or energy, the · Developing regional regulatory frameworks. Physical integration of infrastructure net- opportunity to serve a larger regional mar- works will be effective only with harmonized ket makes the extensive up-front investments regulatory frameworks and administrative more attractive. Airports and seaports must be procedures to allow the free flow of services organized as regional hubs to reach the scale across national borders. necessary to attract airline and shipping ser- vices from beyond the continent. · Facilitating project preparation and cross- Coordinated management and invest- border finance. The complexity of regional ment allow countries to reap the best from infrastructure projects makes them costly multicountry infrastructure systems. Some and time-consuming to prepare. This is regional infrastructure investments, such as particularly true when projects are large many types of transport investments, pro- in relation to the size of the host economy vide public goods, or they facilitate access and essentially depend on financing from to a common pool resource, as with water downstream beneficiaries. resource management for irrigation and other uses. Both public goods and common pool resources require strong coordination. Because Why Regional Integration Matters the quality of a transport network depends on its weakest link, broad participation is crucial, Regional approaches can address the infrastruc- even where benefits are unequally distributed. ture backlog in Africa and propel economic Water can be exhausted, and upstream users are growth, overcoming the region's difficult in a stronger position than downstream users. geography (Limão and Venables 1999). Sub- Collective agreements, effective monitoring, Saharan Africa has 48 countries, most with and conflict resolution mechanisms can ensure small populations--more than 20 countries a fair distribution of costs and benefits. have a population of less than 5 million. Econ- The goal of all regional infrastructure omies are also very small--20 countries have efforts is to facilitate the spatial organization a GDP of less than $5 billion. The small scale of economic activity as a catalyst for faster means governments have difficulty funding the growth. Lessons from the new economic geog- large fixed costs associated with infrastructure raphy, for which Paul Krugman received the development. In addition, 15 African countries Nobel Prize in Economics in 2008, explain this are landlocked, depending on their neighbors concept. Natural resource exports will remain for access to global markets. important, but they provide few job oppor- Most infrastructure investments share char- tunities and their benefits are seldom widely acteristics of public goods, and all benefit to shared. Dutch disease, greater macroeconomic varying degrees from scale economies. Infra- volatility, and weak governance have slowed structure sharing addresses the problems of growth in some resource-rich African coun- small scale and adverse location. Joint provision tries (Collier 2007). Fast employment growth increases the scale of infrastructure construc- and sustained welfare improvements in devel- tion, operation, and maintenance. It reduces oping countries require a move toward mod- costs, pools scarce technical and managerial ern, mostly export-oriented manufacturing capacity, and creates a larger market. Econo- activity. The shift in trade that allowed East mies of scale are particularly important in the Asia's rapid growth can also benefit Africa. In ICT and power sectors. Big hydropower projects the world's fastest-growing regions, the largest that would not be economically viable for a sin- increase in trade has been within industries, for gle country make sense when neighbors share parts and components produced in one loca- their benefits. Although new ICT systems-- tion and assembled in another. Manufacturing especially mobile telephones--allow provision is more about specialized "tasks" than finished at the local retail level, connecting Africa to products (Collier and Venables 2007). Deepening Regional Integration 145 Splitting up manufacturing processes several deep-water ports in neighboring coun- allows much more specialization, which gives tries when the lack of scale deters international rise to scale economies and, thus, to cost shipping firms from serving many African advantages. The result is the concentration of ports. Complementary connective infrastruc- specialized production in new manufactur- ture (roads, transport services, and smooth ing centers around the world, linked across border facilitation) encourages regional factor national boundaries in regional production mobility and trade in intermediate inputs. networks. Launching such processes in Africa With its many small countries, Africa has a will not be easy, but some general points can large coordination problem in managing net- be made. The World Development Report 2009: work infrastructure. For instance, to link the Reshaping Economic Geography identifies three major agglomerations in Ghana and Nigeria, principles for regional infrastructure: start regional transport links also pass through Togo small, think global, and compensate the least and Benin. fortunate (World Bank 2009). Compensate the Least Fortunate Start Small The benefits of concentration mean that Regional infrastructure is an ideal entry point growth will most likely be in a small number for integration processes, because the costs of existing cities that have location advantages and benefits and the rights and responsibilities and an existing economic base, such as coastal can be more easily defined. In the past, many cities with a good investment climate. Favoring regional agreements have failed because they these areas in planning regional infrastructure were overly ambitious, trying to achieve too investments, at least initially, makes economic much too fast. Regional infrastructure shar- sense. With the proper complementary poli- ing builds institutions that promote closer cies, other areas in the region will also benefit. economic integration, and mutual depen- Labor mobility will lead to remittances from dence encourages political stability. Countries migrants who find jobs in dynamic growth will be more willing to cede some sovereignty centers. Specialization means that even small in exchange for tangible benefits, such as players can find a niche. For instance, car water sharing or lower prices for power or assembly may be possible only in some large ICT services. African countries such as Nigeria and South Africa. However, smaller countries such as Think Globally Cameroon or Zambia can specialize in com- Regional integration should not simply adapt ponents. For this approach, regional transport failed import substitution policies to a regional and communication costs must come down. level. Instead, it is the means to greater global For some areas, however, no amount of infra- integration. African markets, even with regional structure investment will trigger growth. They pooling, are too small to sustain high growth. will need coordinated incentives, preferential Regional integration scales up supply by creat- allocations of aid in education and health, to ing larger production networks and beneficial create portable assets in the form of human agglomeration. However, the key objective is capital that can then migrate to where the to connect to world markets for intermediate employment opportunities exist. inputs and intermediate or final outputs. This Africa faces severe challenges in diversify- approach has implications for regional infra- ing from raw material exports and breaking structure development. into world markets for manufactured goods. For a connection to global markets, primary China and India have unified markets with production centers--more likely in coastal populations that are 70 percent and 50 percent areas--must become regional infrastructure larger, respectively, than all of Sub-Saharan hubs with efficient ports and airports. These Africa. Whereas Shanghai or Shenzen, China, large and lumpy investments must be con- draws on a catchment area for labor and prod- centrated where they promise the highest eco- ucts of several hundred million people, the nomic return. Little point exists in developing home market for most African growth centers 146 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION is limited to a few million. Enabling Africa to Africa Submarine Cable (SAT-3), which passes develop regional manufacturing clusters that along the Gulf of Guinea and down to South can compete globally requires lowering barri- Africa. The entire east coast of Africa lacks ers to both productive interaction and (at least access to an underwater cable. Intraregional temporarily) preferential access to world mar- connectivity by fiber-optic cables is also limited. kets with liberalized rules of origin. Regional Most countries rely on satellites for interna- integration is essential, and regional infrastruc- tional telecommunications, including Internet ture sharing must be a key priority. National access, leading to prices for dial-up and broad- infrastructure programs such as those in India band Internet at least twice as high as in other or China (for example, the Golden Quadrilat- regions. Transmission capacities are low, and eral highway program) will involve agreement costs are high. Countries without access to an by numerous countries in Africa. However, the underwater cable have only 3 bits of bandwidth payoffs to greater coordination and integration per capita, whereas those with access have 24 of infrastructure will be large. bits. The average cost of an international fixed- line telephone call within Sub-Saharan Africa is $1.23 a minute, almost twice the cost of a call Opportunities for Regional to the United States ($0.73 a minute). Intrare- Cooperation across Infrastructure gional call traffic is barely 113 million minutes, Sectors compared with intercontinental call traffic of 250 million minutes.1 Africa's regional infrastructure networks have Several projects are already under way major gaps that increase the costs of doing busi- to complete the loop of underwater cables ness and prevent the realization of scale econo- around Africa, with an estimated value of mies. Supply of infrastructure as a public good $1.8 billion (table 6.1). Most are commercially and management of common pool resources sponsored and privately financed, such as the have been deficient. More efficient regional inte- Eastern African Submarine Cable System proj- gration of infrastructure is needed in all sectors: ect to link South Africa and the Horn of Africa ICT, transport, power, and water (figure 6.1). (box 6.1). The cost of completing the nascent fiber-optic network interconnecting the capital ICT--Slashing Costs of International cities of Sub-Saharan Africa and the main sub- Voice and Internet Connectivity marine cables is modest at $316 million, based As in other parts of the world, mobile tele- on a cost of about $27,000 per kilometer. phones have greatly improved telecommuni- The most immediate direct benefits of cations in Africa. However, the benefits have enhanced connectivity are reduced prices and been limited to local and domestic communi- better service for international voice and Internet cations. The region's national telecommunica- connectivity. Prices for most services in countries tion networks are poorly integrated with each with underwater cable access are half those other and the rest of the world. Optical-fiber in countries without access (table 6.2). Such technology provides the least expensive and large price reductions could boost demand highest-capacity transmission of telephone, for these services and, ultimately, economic Internet, and other data traffic. Submarine and productivity. However, too often, access to land-based communication cables combine underwater cables remains with the incum- high speed and large capacity. Although initial bent operator, which (without adequate regu- investments are high, marginal transmission latory controls) charges monopoly prices that costs are very low. prevent consumers from reaping the full cost Access to the global network of submarine advantage of this technology. Countries with cables is low in Africa, especially for landlocked multiple international gateways see some com- countries that depend on neighbors for access. petitive pressure, which keeps service prices Both coordination and massive investments significantly lower than in countries where are required. The region's main international an underwater cable is the only international underwater cable is the South Atlantic 3/West gateway (table 6.2). Deepening Regional Integration 147 Figure 6.1 Africa's Regional Infrastructure Challenge a. ICT--closing the circle b. River basins--managing common resources TUNISIA TUNISIA MOROCCO MOROCCO ALGERIA ALGERIA LIBYA EGYPT LIBYA EGYPT MAURITANIA MAURITANIA MALI · MALI NIGER ERITREA ERITREA CAPE VERDE CAPE VERDE NIGER CHAD SUDAN SENEGAL · · DJIBOUTI SENEGAL DJIBOUTI · GAMBIA CHAD GUINEA- · · SUDAN · BISSAU · · · · GUINEA- NIGERIA BISSAU ETHIOPIA GUINEA · · · SOMALIA GUINEA NIGERIA SOMALIA SIERRA LEONE · ETHIOPIA SIERRA LEONE · · CENTRAL AFRICAN CENTRAL AFRICAN LIBERIA · · · REPUBLIC LIBERIA BENIN REPUBLIC CÔTE D'IVOIRE BENIN · · TOGO GHANA TOGO CAMEROON CÔTE D'IVOIRE GHANA BURKINA FASO CAMEROON · BURKINA FASO KENYA · EQUATORIAL GUINEA · EQUATORIAL GUINEA GABON GABON SAO TOME · · KENYA UGANDA SAO TOME UGANDA AND PRINCIPE AND PRINCIPE CONGO, DEM REP · CONGO, DEM REP SEYCHELLES RWANDA · TANZANIA RWANDA BURUNDI SEYCHELLES REP. OF. CONGO BURUNDI · REP. OF. CONGO TANZANIA · ANGOLA COMOROS COMOROS ANGOLA · MALAWI MAYOTTE ZAMBIA ZAMBIA · MALAWI SAINT HELENA · · MOZAMBIQUE MOZAMBIQUE · NAMIBIA REUNION NAMIBIA · ZIMBABWE Broadband network BOTSWANA · · · ZIMBABWE BOTSWANA MADAGASCAR SOUTH MADAGASCAR SWAZILAND existing or under construction · SWAZILAND AFRICA SOUTH international river basins LESOTHO missing AFRICA LESOTHO c. Roads--connecting the dots d. Power--trading electricity regionally · · · TUNISIA TUNISIA MOROCCO · MOROCCO · ALGERIA ALGERIA LIBYA LIBYA EGYPT EGYPT MAURITANIA ERITREA MAURITANIA MALI · ERITREA CAPE VERDE · NIGER CAPE VERDE MALI NIGER SENEGAL · DJIBOUTI SENEGAL · · · CHAD · · CHAD DJIBOUTI GUINEA- SUDAN · SUDAN · · GUINEA- · BISSAU · · · · · · BISSAU · · · · GUINEA NIGERIA GUINEA SOMALIA · SOMALIA SIERRA LEONE · · NIGERIA · · · · SIERRA LEONE · ETHIOPIA CENTRAL AFRICAN ETHIOPIA BURKINA FASO · BURKINA FASO · CENTRAL AFRICAN · · · REPUBLIC · · · · · REPUBLIC LIBERIA BENIN LIBERIA BENIN · · · GHANA TOGO CAMEROON · GHANA TOGO CÔTE D'IVOIRE CÔTE D'IVOIRE CAMEROON · · EQUATORIAL GUINEA EQUATORIAL GUINEA · · UGANDA · · KENYA UGANDA KENYA SAO TOME GABON SAO TOME GABON · · CONGO, DEM REP · AND PRINCIPE · AND PRINCIPE CONGO, DEM REP · · RWANDA SEYCHELLES · SEYCHELLES · RWANDA BURUNDI REP. OF. CONGO BURUNDI REP. OF. CONGO TANZANIA · TANZANIA · · · COMOROS COMOROS ANGOLA ANGOLA · ZAMBIA · MALAWI ZAMBIA MALAWI · · · · · · MOZAMBIQUE MOZAMBIQUE · NAMIBIA · NAMIBIA ZIMBABWE ZIMBABWE · · Main road corridors BOTSWANA · · · Power network BOTSWANA · · · SWAZILAND SWAZILAND SOUTH MADAGASCAR MADAGASCAR good or fair SOUTH · existing AFRICA · LESOTHO LESOTHO AFRICA poor or missing missing Source: Maps provided by the African Development Bank, 2008. Beyond direct benefits, enhanced con- specialized production that exploits econo- nectivity supports closer regional economic mies of scale--to move from low-technology, networks and integration with international standardized manufacturing to internationally markets (Leamer and Storper 2001). Good competitive production and to ensure access communication is a precondition for the to global markets for manufactured outputs emergence of buyer-supplier networks for and business services. 148 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 6.1 Costs of Reaching Full Intercontinental and Intraregional Connectivity $ millions Intercontinental connectivity Intraregional connectivity Required Required Region Project investment Project investment East Africa EASSy, TEAMS 260 Connect main hubs 132 within and between Southern Africa InfraCo, SRII 510 subregions and to 7 West and Central Africa Infinity, GLO-1, WAFS 1,010 underwater cables 177 Sub-Saharan Africa total 1,780 316 Source: Mayer and others 2008. Note: Intraregional data are for 24 countries. EASSy = Eastern African Submarine Cable System; GLO-1 = Globacom-1; SRII = Southern African Development Community Regional Information Infrastructure Initiative; TEAMS = The East Africa Marine System; WAFS = West African Festoon System. BOX 6.1 Not So EASSy The Eastern African Submarine Cable System (EASSy) is an has meant that prices remain high and the benefits for cus- underground fiber-optic cable that runs from South Africa tomers limited. to Sudan, allowing all countries along the route to connect Designed to ensure effective competition and regula- to the global submarine cable system. tion, EASSy is owned by a consortium, which also includes Developed and owned by a consortium of about a special-purpose vehicle owned by a group of smaller 25 telecommunication operators, mostly from eastern operators from the region. Development finance support and southern Africa, the cable is expected to cost $230 for EASSy is provided as loans to this special-purpose vehi- million. About one-third of the funding will come from cle, which can sell capacity in any market in the region to nonconcessional debt finance provided by five develop- licensed operators on an open-access, nondiscriminatory ment finance institutions (the International Finance Cor- basis, thus competing with other consortium members. poration, the European Investment Bank, the African As traffic volumes increase, the special-purpose vehicle is Development Bank, the French Development Agency, and required to pass on cost reductions to customers. the German Development Bank), with the remainder from Reaching consensus about these access arrangements has commercial equity. been difficult, leading to project delays. Meanwhile, Kenya The project took about three years to develop. The first moved forward with its own underwater cable, The East Africa stage involved discussions and negotiations with stake- Marine System (TEAMS), with links to the United Arab Emirates. holders to determine the project's structure. The second Technically much simpler, that project enjoys significant private stage focused on the technical and financial details of backing. Unless the system can be integrated in a regional net- implementation. The third stage, laying the cable, began work, however, costs will be higher and benefits less broadly in 2008. shared than with a regional effort. A third, privately funded Policy makers and development finance institutions have effort, South Africa­East Africa­South Asia­Fiber Optic Cable focused on not repeating the experience of the SAT-3 cable, (SEACOM), is planned to connect South Africa and several East which runs along the west coast. That project was also African countries to global networks by mid-2009. financed, built, and managed by a consortium of operators, but each member of the consortium has exclusive control over Source: Based on interviews with staff from the World Bank's ICT Policy access to the cable in its own country. Lack of competition Department, 2008. Power--Capturing Scale Economies have developed only a small fraction of that to Reduce Energy Costs potential. Some of the region's most cost- Although well endowed with hydropower and effective energy resources are far from major thermal energy resources, African countries centers of demand in countries too poor to Deepening Regional Integration 149 Table 6.2 Benefits of Access to an Underwater Cable Price for a Price for 20 Price for call within Price for a hours of dial- ADSL broad- Sub-Saharan call to the up Internet band Internet Share of Africa ($ per United States access ($ per access ($ per Access level countries (%) minute) ($ per minute) month) month) No access to submarine cable 67 1.34 0.86 67.95 282.97 Access to submarine cable 32 0.57 0.48 47.28 110.71 Monopoly on international gateway 16 0.70 0.72 37.36 119.88 Competitive international gateways 16 0.48 0.23 36.62 98.49 Source: Minges and others 2008. Note: ADSL = asymmetric digital subscriber line. raise the billions of dollars needed to develop Regional trade also puts Africa on a cleaner them. For example, 60 percent of the region's development path in terms of carbon emissions. hydropower potential is in the Democratic Regional power trade would increase the share Republic of Congo and Ethiopia. Because of hydropower in the continent's generation 21 of 48 Sub-Saharan African countries have portfolio from 36 percent to 48 percent, dis- total generation capacity of less than 200 placing 20,000 megawatts of thermal power in megawatts--efficient scale of production the process and saving 70 million tons a year below minimum--they pay a heavy penalty: of carbon emissions (8 percent of Sub-Saharan costs reach $0.25 per kilowatt-hour, compared Africa's anticipated emissions through 2015). with $0.13 per kilowatt-hour in the region's Applying the Clean Development Mechanism at larger power systems. $15 per ton of carbon would reduce the region's The desire to pool energy resources and lever- carbon emissions another 4 percent. Closely age scale economies in power sector develop- integrating power grids will also help balance ment led to formation of regional power pools loads when other renewable energy resources, in southern and western Africa during the mid- such as concentrated solar and geothermal 1990s, and more recently in eastern and central energy, are deployed on a large scale. Africa. However, trade has yet to take off. Cross- The 10 largest potential power-importing border power trade accounts for only 16 percent countries could reduce their long-run marginal of the region's power consumption, more than cost of power by $0.02­$0.07 per kilowatt-hour 90 percent of it within the Southern African (figure 6.2). Those that stand to gain most tend Power Pool, and much of that between South to be smaller countries or those heavily reliant Africa and its immediate neighbors. Without on thermal power, such as Angola, Burundi, physical or regulatory impediments, about 40 Chad, Guinea-Bissau, Liberia, Niger, and percent of eastern and southern Africa's power Senegal. However, reaping the full benefits of consumption would be traded across national power trade would require a political willingness borders (Rosnes and Vennemo 2008). to depend heavily on power imports. As many as If pursued to its full economic potential, 16 African countries would be better-off (from regional trade would shave about $0.01 per a purely economic standpoint) importing more kilowatt-hour off the marginal cost of power than 50 percent of their power needs. generation in each of the power pools. The The future of power trade depends on the resulting overall potential savings of regional health of the power sector in a handful of key power trade would amount to about $2 billion exporting countries endowed with exception- a year in the costs of power system develop- ally large and low-cost hydropower resources. ment and operation (equivalent to about 5 per- In descending order of export potential, cent of total power system costs). The savings these include the Democratic Republic of come largely from substituting hydropower Congo, Ethiopia, Guinea, Sudan, Camer- for thermal power, even if higher immediate oon, and Mozambique (table 6.3). The first investments are required. three account for 74 percent of potential 150 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 6.2 Savings from Power Trade for Major investments needed to realize these export Potential Power-Importing Countries volumes is daunting. Each of the top three would need to invest more than $700 million Guinea-Bissau a year for the next decade to develop the gen- Liberia eration capacity for export, or more than 8 Niger percent of their GDP. Thus, supporting such Angola investments would be difficult without exten- Chad sive cross-border financing arrangements that Burundi allow importing beneficiaries to make up- front capital contributions. Senegal To make trade possible, countries would Mali need to develop some 22,000 megawatts of Congo, Dem. Rep. of interconnectors to allow the free flow of power Equatorial Guinea across national borders, at a cost of more than Mozambique $500 million a year over the next decade. The returns to the interconnectors can be as high Sierra Leone as 120 percent for the Southern African Power Lesotho Pool, and more typically 20­30 percent for Namibia the other power pools. For countries with the South Africa most to gain from power imports, investments Gabon in cross-border transmission have exception- Kenya ally high rates of return, typically paying for themselves in less than a year. 0 1 2 3 4 5 6 7 8 US cents per kilowatt­hour Transport Infrastructure--Facilitating Source: Derived from Rosnes and Vennemo 2008. Internal and External Trade Transport infrastructure is critical for linking Africa to the global economy and promoting Table 6.3 Profile of Top-Six Potential Power-Exporting Countries economic integration within the continent. Potential However, the infrastructure demands are quite Net revenues Required investment exports different in each case. (terrawatt $ millions Percentage $ millions Percentage Country hours per year) per year of GDP per year of GDP External Trade. For external trade, the conti- Congo, Dem. Rep. of 51.9 519 6.1 749 8.8 nent's economic geography makes transport Ethiopia 26.3 263 2.0 1,003 7.5 connections with the world something of a Guinea 17.4 174 5.2 786 23.7 public good. Major corridors to the sea con- nect the continent's 15 landlocked countries to Sudan 13.1 131 0.3 1,032 2.7 the major seaports, through a combination of Cameroon 6.8 68 0.4 267 1.5 road and rail infrastructure. The main ports Mozambique 5.9 59 0.8 216 2.8 include Douala (Cameroon) for central Africa; Source: Derived from Rosnes and Vennemo 2008. Durban (South Africa) and Maputo (Mozam- Notes: Net revenue is calculated as the estimated volume of exports multiplied by an illustrative profit margin of $0.01 per kilowatt-hour exported. Required investment represents the investment needed bique) for southern Africa; Dar es Salaam for the country to be able to exploit its full economic power-exporting potential. (Tanzania) and Mombasa (Kenya) for eastern Africa; and Abidjan (Côte d'Ivoire), Cotonou (Benin), and Dakar (Senegal) for western power exports, which could become big busi- Africa (table 6.4). About $200 billion a year in ness for them. Based on a purely illustrative imports and exports moves along these cor- profit margin of $0.01 per kilowatt-hour on ridors, barely 10,000 kilometers long. About power sales, the net export revenues for these 70 percent is in good or fair condition, with top three exporters could be 2­6 percent of donors channeling more resources to improve their respective GDP. However, the size of the infrastructure along the routes. Deepening Regional Integration 151 However, regulatory and administrative Table 6.4 Average Delivery Time for Containers from Vessel to Consignee hurdles continue to inflate costs and pro- Distance Transit time long delays for freight movements along Gateway Destination (kilometers) (days) these strategic arteries (table 6.4). Despite Mombasa, Kenya Kampala, Uganda 1,100 20 the reasonably good physical condition of the Mombasa, Kenya Kigali, Rwanda 1,700 27 roads, the implicit velocity of freight move- Dar es Salaam, Tanzania Bujumbura, Burundi 1,800 21 ment is no more than 10 kilometers per hour Abidjan, Côte d'Ivoire Ouagadougou, Burkino Faso 1,200 7 (roughly the speed of a horse and buggy). The Abidjan, Côte d'Ivoire Bamako, Mali 1,200 7 cause is the extensive delays of 10­30 hours Dakar, Senegal Bamako, Mali 1,200 10 at border crossings and ports. Transit times between major cities are very high by inter- Cotonou, Benin Niamey, Niger 1,000 11 national standards (table 6.4). Member coun- Douala, Cameroon Ndjamena, Chad 1,900 38 tries have organized corridor associations to Lagos, Nigeria Kano, Nigeria 1,100 21 address nonphysical barriers to transit, estab- Source: Arvis 2005, quoting an international logistics company. lishing one-stop integrated border posts and improving ports and custom administration. Tariffs for road freight can be several times Table 6.5 Key Transport Corridors for International Trade, Sub-Saharan Africa higher than in other parts of the developing Trade density Implicit Freight world (table 6.5), which is attributed not to Road in good ($ millions velocitya tariff Length condition per road (kilometers ($ per ton- higher road transport costs in Africa, but to Corridor (kilometers) (percent) kilometer) per hour) kilometer) exceptionally high profit margins in the truck- Western 2,050 72 8.2 6.0 0.08 ing industry (Teravaninthorn and Raballand Central 3,280 49 4.2 6.1 0.13 2008). These margins in turn reflect carteliza- tion and restrictive regulatory frameworks, such Eastern 2,845 82 5.7 8.1 0.07 as market entry barriers; technical regulations; Southern 5,000 100 27.9 11.6 0.05 and the tour de role system that allocates freight Source: Teravaninthorn and Raballand 2008. a. Includes time spent stationary at ports, border crossings, and other stops. business based on queuing, particularly in cen- tral and western Africa. This system favors large fleets with mostly older trucks in poor condi- its own airport facility and jet fleet. Instead, tion. Moreover, it fosters corruption, because a regional hubs serving multiple countries are transport operator can increase its volume of needed, with fleets of smaller commuter jets cargo only by bribing the freight bureaus, the to move passengers along spokes and into government entities that allocate freight among hubs. Liberalization begun under the Yamous- transport operators. soukro Decision in 1999 should allow carriers Upgrading the remainder of the corridors serving key routes to consolidate and a better to the sea to good condition is estimated to package of intraregional services to emerge. cost about $1.5 billion, while the annual cost However, implementation has been lagging, of maintenance is close to $1.0 billion. Simu- especially in harmonizing competition rules lations suggest that corridor rehabilitation and removing nonphysical barriers such as would yield an internal rate of return of 20­60 landing rights and tariffs. In eastern and percent in eastern Africa's northern corridor. southern Africa, this consolidation of carriers However, low traffic, poor truck use, and an and hubs has already occurred, with Ethio- aging fleet in central and western Africa would pian Airlines (Addis Ababa), Kenya Airways undermine the economic viability of corridor (Nairobi), and South African Airways (Johan- upgrades. Investments in these regions would nesburg) emerging as the main ones. Yet, in likely become attractive only when more fun- central and western Africa, hubs are conspicu- damental regulatory and institutional reforms ous by their absence (figure 6.3). The collapse improve trucking productivity. of key regional carriers, notably Air Afrique, For air transport, the size of the market is partly caused this gap. Particularly striking simply too small to support a proliferation is the failure of Lagos to emerge as a hub for of national carriers with each centered on western Africa. 152 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 6.3 Uneven Distribution of Airport Hubs across Africa: Traffic Flows between Emirates; Jeddah, Saudi Arabia; and Salalah, Top-60 Intraregional City Pairs Oman). For southern Africa, the government of South Africa has decided to develop a sizable hub port in Richards Bay, which will likely cap- ture a significant portion of the seaborne trade between Asia and subequatorial Africa. On the west coast, despite the growth of Tangiers, Morocco, room may still exist for one or two regional hubs. Dakar, with its recent container- terminal concession and port expansion plans, is a strong candidate. Apapa (Lagos), though more centrally positioned, is already strug- gling with its local market and facing heavy congestion. Intraregional Trade. Intraregional trade depends on the internal network linking Afri- can countries to each other. Except in southern Africa, the rail network does not typically pro- vide such intraregional connectivity, because of the incompatibility of gauges and the isolated parallel corridors to the sea. Even along the more frequented sea corridors, most African railways struggle to reach economic viability because of very low traffic volumes. With traf- fic volume on intraregional rail routes much Source: Bofinger 2008. Note: Thickness of lines reflects traffic volumes. lower than even that on corridors to the sea, feasible further intraregional integration of rail networks in the near future is difficult to imagine. For seaports, large vessels (more than It follows that the road network has the 200,000 20-foot equivalent units per year that greatest potential to support intraregional provide scale economies to seaborne trade on trade. In the 1970s, the Trans-African High- east-west ports) are today able to call at only way system was conceived as a network of a handful of Sub-Saharan Africa's larger ports all-weather roads to provide direct routes (Luanda, Angola; Abidjan, Côte d'Ivoire; Tema, between the region's capitals; to contribute Ghana; Mombasa, Kenya; Maputo, Mozam- to the region's political, economic, and social bique; Apapa [Lagos], Nigeria; Dakar, Senegal; integration and cohesion; and to ensure road Durban and Cape Town, South Africa; Dar es transport between important areas of produc- Salaam, Tanzania; and Lomé, Togo). Several tion and consumption (AfDB and UNECA of these ports act as regional hubs, but with 2000). However, national governments have relatively small transshipment volumes. In had difficulty supporting needed invest- theory, coordinating the choice of hub ports ments. The official Trans-African Highway on Africa's different seaboards to reach greater system comprises nine main corridors and economies of scale is desirable, but in prac- reaches just over 50,000 kilometers. In mid- tice, it is difficult because of rivalry between 2008, almost half of the network was in poor national ports. condition, with about 70 percent of roads For the east coast ports, the major regional paved but 25 percent with either an earth sur- hubs are already being developed in the Middle face or no formed road. Most of these miss- East (Djibouti, Djibouti; Jebel Ali, United Arab ing links--those with the greatest potential Deepening Regional Integration 153 to knit together the continent's economies-- coordination based on established interna- are in central Africa (Buys, Deichmann, and tional public law governing water sharing can Wheeler 2006). ensure equitable distribution of the benefits of Extending the network to connect all Sub- common-pool water resources. For upstream Saharan African cities with more than 500,000 neighbors, apart from greater regional stabil- people would add an estimated 50,000 kilo- ity, the benefits include sharing of the large meters of road. The costs of completing an investments in hydropower or irrigation intra-African road network of this kind are infrastructure. of a larger order of magnitude than those Transboundary water resource management associated with the corridors to the sea. Some requires strong institutional commitment. estimate a one-time cost of $20 billion and Between the 1960s and 1980s, many countries an ongoing maintenance cost of $1 billion a created river basin arrangements such as the year (Buys, Deichmann, and Wheeler 2006). Senegal River Development Organization in The associated benefits are more speculative. 1972; the Gambia River Development Organi- Well-established relationships between trade zation in 1978; and the Niger River Commission volumes and transport costs mean that a fully in 1964, later transformed into the Niger Basin operational Trans-African Highway network Authority. External support, through media- could nearly triple intra-African trade, from tion and financial backing, encouraged initial $10 billion a year to almost $30 billion (Buys, enthusiasm. However, nearly three decades later, Deichmann, and Wheeler 2006). Even if one with few exceptions, these transboundary orga- assumes rehabilitation costs as high as $20 bil- nizations are still in the emergent stages. Some lion, the benefit-cost ratio over 15 years would of their challenges included waning political be as high as 5 to 1. However, even these higher commitment, poor cooperation, management projected volumes of regional trade look small and technical difficulties, armed conflict and alongside Africa's existing volumes of interna- political instability, and poorly defined goals tional trade, which stand close to $200 billion or insufficient capacity for proposed plans. As per year. donor support dwindled, basin organizations also lacked the financial backing to carry out Water Resources--Minimizing their programs. Conflicts, Maximizing Benefits Coordination costs are high, given the Africa has more than 60 transboundary river sensitive nature of water resources, especially basins, almost half shared by three or more for countries in arid environments. Technical countries with riparian rights. The region's sur- assistance and capacity building can strengthen face water resources benefit economic devel- basin organizations. One coordination tool is opment in several ways. Well-managed water a management system that measures the prog- resources can create low-cost hydropower, ress of water resource management in river abundant irrigation, and cost-effective surface basins (see UNESCO 2006). Such a system transport. However, hydrologic variability and sets benchmarks and defines the monitoring limited storage leave economies vulnerable to framework to track water discharge, quality, floods and droughts. and development effects. The transboundary nature of most water- The Senegal River Development Organiza- sheds in Africa implies that regional coordi- tion is generally considered successful in trans- nation in water management is important boundary cooperation for water management. (UNESCO 2003). What happens in an upstream It built the Manantali and Diama dams, which country can benefit or harm its downstream brought irrigation to some 375,000 hectares neighbor. Hydropower and water storage infra- of land, enabled 800 gigawatt-hours a year structure can provide cheaper electricity and of hydropower generation, and added about balanced water flows, but excessive extraction 800 kilometers of navigation on the Senegal or pollution upstream can hurt agriculture and River from Senegal to Kayes, Mali. Close coor- drinking water supplies downstream. Regional dination between riparian neighbors also 154 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION allows them to address early any negative benefit especially from regional power trade effects of water management development on that reduces the costs of energy supply. As long agriculture and fisheries. as regional integration provides a substan- tial economic dividend to some of the par- ticipating countries, designing compensation Meeting the Challenges of mechanisms that benefit all of them should be Regional Integration of possible. The concept of benefit sharing was Infrastructure in Africa pioneered through international river basin treaties, such as that for the Senegal River, and The benefits of regional infrastructure develop- could be applied to regional infrastructure more ment are clear, but reaping those benefits poses broadly. political, institutional, economic, and financial A key prerequisite for any regional initiative challenges that are far from trivial. The starting is building political consensus both nationally point is building political consensus among and across borders. Although methods will neighboring states that may have diverging vary from country to country, some broad national agendas or even recent histories of principles apply. conflict. Thereafter, effective regional institu- tions are needed to take forward a collabora- Get High-Level Buy-In. Africa needs more tive cross-border infrastructure development high-level advocacy and leadership to promote program and to ensure an equitable distribu- regional integration for infrastructure develop- tion of benefits. Given the vast needs and lim- ment and beyond. Regional integration issues ited resources, some form of priority setting is remain only a small part of parliamentary necessary to guide efforts on the regional inte- debate in most countries. Between infrequent gration agenda. Even when priorities are clear, regional meetings of heads of state, a sense of however, funding and implementing extensive inertia and lack of follow-up frequently exist. project preparation studies and arranging Governments and international institutions cross-border finance for complex multibillion- must provide leadership. The African Union dollar projects is far from straightforward. has the mandate to coordinate the regional Furthermore, once the regional infrastructure integration program of Africa as spelled out is in place, its efficacy will ultimately depend in the Abuja Treaty (1991), which created the on harmonizing the associated regulatory and African Economic Community, with regional administrative procedures. economic communities as building blocks. The African Union identified infrastructure Building a Political Consensus and regional integration as major components Regional infrastructure is only one aspect of of economic growth and poverty reduction in broader regional integration. In contrast to Africa. The main vehicle is the New Partner- economic or political integration, however, ship for Africa's Development (NEPAD), which cooperation in infrastructure provision is eas- has not always received the requisite support ier to achieve, because benefits are more clearly from political leaders to build a consensus defined, and countries need to cede less sover- around financially and economically viable eignty. Regional infrastructure cooperation is projects. The NEPAD Heads of State Imple- therefore an effective initial step on the path to mentation Committee, set up to help remove broader integration. political blocks to projects, has not been effec- Some countries have more to gain from tive and now meets less regularly than origi- regional integration than others. Landlocked nally intended. A strong commitment from countries depend particularly on effective road regional leaders is essential to move projects and rail corridors to the sea, as well as on intra- forward. When the West African Gas Pipeline continental fiber-optic backbones that link ran into political differences, for example, it them to submarine cables. Coastal countries was the shuttle diplomacy of Nigeria's Presi- depend particularly on sound management dent Olusegun Obasanjo that kept the project of water resources upstream. Small countries on track. Deepening Regional Integration 155 Build Trust. Trust is important for regional agreement, contributing to broader political integration--especially when some countries and economic cooperation and stability. stand to benefit more than others. Coun- tries that do not trust each other may fail to Think Regionally, Even When Developing reach a cooperative solution. For example, National Policies. Regional interdependence implementation of the road-railway bridge is a fact of life in all parts of Africa. It is criti- project between Kinshasa and Brazzaville cal for not only landlocked countries but also and extension of the Kinshasa-Ilebo rail- larger and coastal countries, which deal with way are intended to accelerate trade between regional trade, labor migration, and expand- the Democratic Republic of Congo and the ing markets. National policy makers should Republic of Congo. Trust between the two therefore consider the regional consequences of countries will be a key factor in the decision national policy making. Donors can encourage to proceed with this project. Starting small, this approach. For instance, developing an assis- with relatively well-defined projects, is one tance strategy for Burkina Faso without consid- way to build that trust. Frequent interaction ering its place in the region in relation to Côte among policy makers at all levels of govern- d'Ivoire, Ghana, and Mali makes little sense. ment builds relationships that help overcome inevitable disagreements. Supranational orga- Establishing Effective Regional nizations can be honest brokers for sharing Institutions gains and resolving disputes. No shortage exists of regional institutions in Africa, but few are effective. The institutional Invest in Credible Information. Trust is easier architecture that supports African integration to build when the facts are available to all. Good comprises more than 30 executive continental evidence must be gathered and made acces- bodies, regional economic communities with sible to decision makers to gauge the full costs many overlapping memberships, sectoral tech- and benefits of regional infrastructure invest- nical bodies, and national planning bodies. The ments, many of which involve large allocations result is a high degree of complexity, unclear of funds and ceding of some sovereignty. The functional responsibilities for strategy and regional economic communities must commu- project development, and uncertain financing nicate the potential benefits to all stakeholders strategies. This lack of clarity has slowed prog- to help build consensus. Countries are unlikely ress on coherent regional strategies, realistic to bear the full cost of public goods if the ben- programs for integration priorities (such as efits are not clear. Because regional integration regional infrastructure and trade integration), can create winners and losers, assessing the and technical plans for specific projects. likely benefits and costs will help. Becoming more effective is easier for agen- cies with a narrow set of tasks and responsibili- Focus on Sharing Benefits, Not on Sharing ties than for those with a broader design. The Resources. Regional projects often fail because African Union Commission has struggled with of the perception of unequal access to a natu- its mandate because of a lack of human and ral or infrastructure resource. However, what financial resources and unclear responsibili- matters is how the economic benefits from ties. The regional economic communities have resources or infrastructure are shared. This limited capabilities and resources and, above philosophy is best illustrated in the manage- all, weak authority to enforce decisions. A dis- ment of transboundary water resources, where connect often exists between what is written in benefits include flood protection, hydropower, treaties and what happens on the ground. Insti- irrigation, fisheries, leisure, tourism, and peace tutions will be more effective if a greater willing- and security. One country may benefit most ness exists to cede some sovereignty in return from hydropower, while another requires steady for greater economic benefits. Greater use of access to water for irrigation. Successful benefit qualified majority rules in some areas of policy sharing includes the Lesotho Highlands Water making would lead to more streamlined deci- Project and the Incomati Basin water-sharing sion making. This approach has been debated 156 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION for some time in many regional economic com- implementation by high-level government munities, without resolution. Adequate financ- officials. ing is also a problem because member states Five actions to improve the effectiveness of often fail to pay assessed contributions in full, institutions can aid regional cooperation in if at all. Regional economic communities have infrastructure provision. First, the roles and multiple functions, with infrastructure not nec- responsibilities of regional bodies concerned essarily prominent (ICA 2008). As a result, they with regional integration must be clarified. often fail to attract and retain professional staff Second, increased legal authority is required with the experience to identify and promote for regional entities to improve and acceler- complex regional infrastructure projects. ate decision-making processes. Third, the key In 2006, the regional poverty reduction regional bodies must boost their professional strategy for West Africa by the Economic capacity. Fourth, national planning agencies Community of West African States and the must improve their ability to strengthen links West African Economic and Monetary Union between regional strategies and national devel- was a significant milestone. Other regions also opment plans. Fifth, delivery mechanisms for have completed strategic planning exercises: priority programs (for example, regional infra- for example, the Regional Indicative Strate- structure) should be strengthened to underpin gic Plan (developed by the Southern African confidence in integration by delivering tangi- Development Community) and the East Afri- ble results. can Community Master Plan. However, links Africa's efforts to strengthen regional between these regional strategic plans and integration have focused on the fifth action. country programs remain limited. Improved However, national priorities have limited the links are critical for coordinated implementa- support for regional programs overall. Poor tion of regional programs, which is essential reflection of regional priorities in national to leverage outcomes at the country level. For plans has slowed priority programs, sapping example, the Common Market for Eastern and government willingness to cede sovereignty to Southern Africa, the East African Community, other regional initiatives and creating a vicious and the Southern African Development Com- circle. Progress requires rebalancing efforts munity have been successfully coordinating among the five institutional challenges. programs through a tripartite task force. Regional special-purpose entities or sectoral Setting Priorities for Regional technical bodies have been more effective. A Infrastructure power pool, for example, has a clear mandate, With a very large backlog of infrastructure sufficient autonomy to execute its responsi- investments as well as limited fiscal space and bilities, a dedicated funding mechanism, and borrowing ability, countries must set effective career opportunities that attract and retain priorities for infrastructure investment in high-caliber staff. It also receives substantial Africa. Projects should be well justified to com- capacity building. The members of a power pete with investments in other sectors such as pool are national electricity utilities, which health or education. The long life of infrastruc- similarly have clear functions and roles within ture means that bad decisions are locked in for their national contexts and are less subject to decades. An unwise investment can saddle gov- immediate political pressures than less techni- ernments with an ineffective project that will cal public agencies. also require costly maintenance. How should The need to increase capacity and streamline priorities be set? Suitable criteria include pre- decision making extends to national agencies. dicted economic returns, spatial targeting, and For complex regional infrastructure projects, scope for private participation. several line ministries are often involved in each country. This practice complicates consensus Economic Returns. Projects with the highest building and assigning of clear responsibilities. returns may not always be new infrastruc- A further problem is the frequent lack of ture. Strategic investments that improve the follow-up on regional commitments to national performance of infrastructure systems, such Deepening Regional Integration 157 as addressing bottlenecks at ports or border to increase investment and facilitate business crossings or installing power interconnectors reforms. The Zambia-Malawi-Mozambique between countries with large cost differen- growth triangle--initiated in 2000 and cov- tials, are often the most effective. Investments ering northern Zambia, northern and central in maintenance and rehabilitation of existing Malawi, and some central-eastern parts of infrastructure, such as roads and rail links in Mozambique--seems to have facilitated trade a network, can often yield economic benefits and generated new economic activities.2 faster than new transport infrastructure. The Economic Community of West African States recently adopted cross-border initia- Spatial Targeting. Too often, political expedi- tives between bordering areas with economic ency encourages spreading out investments to complementarities, such as the Sikasso- all parts of a country or region, when a con- Korhogo-Bamako initiative (based on the centration of productive investments in high- cotton basin shared by Burkina Faso, Côte potential areas would yield greater economic d'Ivoire, and Mali) and the Kano-Katsina- benefits (World Bank 2009). The spatial devel- Maharadi initiative (based on agriculture and opment initiative supported by NEPAD aims cattle in the Nigeria-Niger border region). By to link trunk infrastructure with countries' identifying what is needed to facilitate cross- natural resource endowments. The initiative border production networks, bordering coun- was inspired by the Maputo Development tries can make regional infrastructure invest- Corridor, which bundled infrastructure invest- ments based on profitable joint projects. ments and used private finance as the catalyst to exploit natural resources along a corridor Scope for Private Participation. ICT, power between Mozambique and South Africa. plants, and ports and airports have significant South Africa's leadership was key in mov- potential for private provision and operation. ing the initiative forward. However, whether The prospect of a larger regional market can similar leadership will emerge in other parts attract more interest for private financing or of the region is unclear. Indeed, most of the public-private partnerships. Larger private spatial development initiatives are brownfield involvement can help overcome the large infra- initiatives, in which some regional infrastruc- structure financing gap, but governments must ture already exists. This fact raises the ques- then ratify and implement protocols to facili- tion of why private anchor investments have tate investment. Public control in many coun- not yet been seen. A concern is that corridor tries continues to stifle private investment. For development will simply facilitate exports of example, regarding ports, only two African raw materials, whereas the goal should be eco- countries (Ghana and Nigeria) have adopted nomic development and employment growth the internationally favored landlord model, through manufactured exports. which aims to strike a balance between public Another example of a spatial approach is the (port authority) and private (port industry) development of better links between the large interests. coastal agglomerations in the Gulf of Guinea (Abidjan, Accra, Cotonou, Lagos, and Lomé) Priority-setting exercises are under way and providing a competitive business environ- or planned. A new continental task force will ment sustained by policy harmonization along report on a broad set of criteria for helping the corridor. Such an initiative would allow all development institutions set investment pri- countries to benefit from access to major ports orities in each of the main infrastructure sec- in Abidjan and Lagos and, ultimately, reduce tors. The report will feed into a joint African international transport costs for the entire Union­NEPAD­African Development Bank subregion. study, the Program for Infrastructure Develop- Through shared infrastructure and bet- ment in Africa, which will elaborate a vision of ter trade facilitation, Africa can emulate East regional integration on the continent through Asia's economies, which took advantage of eco- infrastructure. The program will need to con- nomic complementarities in bordering regions sider other ongoing processes, such as the 158 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Africa-EU Energy Partnership, which aims to operators relying on private financing, a firm agree on an Electricity Master Plan for Africa, planning horizon is even more critical than for and the 2009 African Union Summit, which the public sector (ICA 2009). agreed on a short-list of flagship projects for Support for regional projects by the Infra- priority support and investment. In addi- structure Consortium for Africa grew sharply tion, many regional economic communities from about $430 million in 2005 to $2.8 billion and other technical regional institutions have in 2007 (ICA 2007).3 Although the bilateral share 10-year investment plans that present a large has increased over time, multilateral members investment menu for external financiers. accounted for 60 percent of total commitments A requirement for any priority setting is in 2007 (World Bank 2008). Multilateral institu- transparency in decision making and agree- tions have been developing specific mechanisms ment on selection criteria. Decision making for dealing with regional projects. must be well documented and motivated, The World Bank has four criteria for using sufficiently detailed data and a clear regional projects to qualify for concessional explanation of assumptions, all publicly acces- funding from the International Development sible. The investments in better information at Association (IDA): (a) at least three countries the country and regional levels will be small must participate, though they can enter the relative to the public and private funds at stake, project at different stages; (b) countries and but the benefits of better decision making will the relevant regional entity must demonstrate be large. a strong commitment; (c) economic and social benefits must spill over country boundaries; Facilitating Project Preparation and and (d) projects must provide a platform for Cross-Border Finance policy harmonization among countries and be Project design is complex. The appraisal phase priorities within a well-developed and broadly establishes social, economic, financial, techni- supported regional strategy. A recent evalua- cal, administrative, and environmental feasibil- tion of World Bank regional integration proj- ity (Leigland and Roberts 2007). For regional ects concluded that regional programs have projects, coordination among national agencies been effective, if still on a relatively small scale with different procedures, capacity, and admin- (World Bank 2007). istrative constraints adds to the complexity. The African Development Bank adopted Thus, the project preparation costs for regional similar principles in 2008, though it requires projects tend to be higher, and the process can that only two countries participate. To help take longer than that for national projects. with country ownership, both institutions Preparation costs are typically about 5 per- use a one-third, two-thirds principle, whereby cent of total financing, or approximately double participants are expected to use one IDA or the cost of preparing national projects. These African Development Fund (ADF) credit from outlays occur up front when the success of the their country allocation, supplemented by two project and the likelihood of a sufficient return credits from regionally dedicated resources. from the investment are still uncertain. Regional Currently, 17.5 percent of ADF and 15 percent institutions and donors have tried to address of IDA resources in Africa are dedicated to these issues, setting up more than 20 project regional programs. preparation facilities, many of which explicitly For the EU-Africa Infrastructure Trust support regional activities (ICA 2006). How- Fund, eligible projects must have African ever, the resources do not match the regional ownership and long-term project sustainabil- needs. African countries need to commit more ity. They must also be cross-border projects or funds and more people with the proper techni- national projects with a regional effect on two cal, legal, and financial skills for infrastructure or more countries. Regional projects funded planning and project implementation. Timely by the Development Bank of Southern Africa execution of project preparation activities and a must either involve a minimum of two coun- sufficient pipeline of new projects also encour- tries or be located in a single country with ben- age participation of the private sector. For efits to the region. Deepening Regional Integration 159 Some challenges remain. Although recipi- In the energy sector, African borders limit ents of funds from the ADF and the IDA can market size through political and regula- leverage their country allocations by partici- tory barriers to electricity trade and through pating in regional projects, those receiving a physical barriers. Regional power infrastruc- small allocation may be reluctant to use a large ture requires harmonized power pricing and percentage on one regional project for which third-party access regulations, effective cross- the benefits are unclear. How such concessional border trading contracts, and reliable and resources are allocated and whether enough of creditworthy national utilities. In much of the overall envelope is dedicated to regional Sub-Saharan Africa, bilateral arrangements projects remain issues of debate. In addition, between vertically integrated utilities guide limited financing instruments exist for middle- cross-border electricity exchanges, although, income countries, which is an issue in North increasingly, regional power pools are liberal- Africa (for connectivity with countries south of izing the electricity markets. the Sahara) and in southern Africa (for projects Worldwide experience in developing power that might involve Botswana or South Africa). pools has led to a consensus on three key Regional organizations may not always building blocks for success: a common legal qualify for grants or concessional finance from and regulatory framework, a durable frame- donor institutions because of their suprana- work for systems planning and operation, and tional character, limiting the availability of an equitable commercial framework for energy resources for capacity building. Furthermore, exchanges (USAID 2008). some projects with significant regional spill- The four power pools in Sub-Saharan Africa overs may not involve three or more countries are at different stages of development, but as and thus not qualify for regional financing, countries move from bilateral to multilateral such as the Ethiopia-Sudan interconnector or power exchanges, a commercially acceptable any national power-generation project that framework is essential. In 2006, the West African may have export potential. Developing Regional Regulatory BOX 6.2 Frameworks Building physical infrastructure by itself will not yield high economic returns in regional One-Stop Border Posts to Facilitate Trade growth and employment. Improving the legal, Trade logistics has three components: international shipping, opera- regulatory, and administrative environment is tions at the gateway (final clearance or transit clearance at customs necessary to ensure infrastructure's efficient and handling), and inland movement (often under a transit proce- use (box 6.2). dure). At the Chirundu crossing between Zambia and Zimbabwe, Air transport is profitable enough to allow the average transit times for northbound trucks range from 26 to the private sector or public-private partnerships 46 hours. The border has more than 15 agencies from both gov- to invest in and improve infrastructure, but ernments, each enforcing different pieces of legislation. A joint task the regulatory environment and government force formed by the Common Market for Eastern and Southern safety and security regimes are key to success. Africa, the Southern African Development Community, and the East The Yamoussoukro Declaration on free access African Community for harmonization of regional trade arrange- ments started operating in 2006. The one-stop border posts under to the African skies has improved intraregional this initiative illustrate what political will can achieve. The trade facili- and international connectivity. tation measures being addressed by the task force include using a A study of 73 African ports concluded that single document for customs clearance, harmonizing information capacity additions and institutional reform technology and electronic customs management systems, harmoniz- must be placed on a fast track to realize the ing axle loading and road transit charges, and instituting regional potential. Although some countries are under- driving license and insurance schemes. taking new master plans for national ports, Source: Based on interviews with staff from the World Bank's Africa Transport not all address the need to improve weak labor Department, 2008. skills and stifling bureaucracy and to provide independent regulation. 160 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Power Pool was granted special status to rein- Collier, Paul, and Anthony Venables. 2007. force its autonomy, and the 2007 ratification "Rethinking Trade Preferences: How Africa Can Diversify Its Exports." World Economy 30 (8): of an overarching legal framework (ECOWAS 1326­45. [Economic Community of West African States] ICA (Infrastructure Consortium for Africa). 2006. Energy Protocol) will help promote security for Infrastructure Project Preparation Facilities in investors and enshrine the principle of "open Africa: User Guide for Africa. Tunis: ICA. http:// access" to national transmission grids across the www.icafrica.org/en/documentation. region. In 2008, the ECOWAS Regional Elec- ------. 2007. Annual Report 2007. Tunis: ICA. tricity Regulatory Authority was established ------. 2008. "Mapping of Donor and Govern- to regulate cross-border electricity exchanges ment Capacity-Building Support to African between member states. RECs and Other Regional Bodies." Report of Economic Consulting Associates to the Infra- Notes structure Consortium for Africa, Tunis. The authors of this chapter are Souleymane ------. 2009. Attracting Investors to African Public- Coulibaly, Andrew Roberts, Vivien Foster, and Private Partnerships: A Project Preparation Guide. Uwe Deichmann, who drew on background Washington, DC: World Bank. http://www material and contributions from Alvaro Feder- .icafrica.org/fileadmin/documents/guides/ Attracting-investors-to-African-PPP.pdf. ico Barra, Pinki Chauduri, Siobhan Murray, and Alex Rugamba. Leamer, Edward E., and Michael Storper. 2001. 1. Both figures exclude South Africa. "The Economic Geography of the Internet Age." 2. See http://www.afrol.com/News/maw008 Journal of International Business Studies 32 (4): 641­65. _growth_triangle.htm. 3. Infrastructure Consortium of Africa members Leigland, James, and Andrew Roberts. 2007. "The are the Group of Eight countries, World Bank African Project Preparation Gap: Africans Group, African Development Bank, European Address a Critical Limiting Factor in Infrastruc- ture Investment." Gridlines, Note 18 (March), Community, European Investment Bank, and Public-Private Infrastructure Advisory Facility, Development Bank of Southern Africa. World Bank, Washington, DC. Limão, Nuno, and Anthony Venables. 1999. "Infra- References structure, Geographical Disadvantage, and AfDB (African Development Bank) and UNECA Transportation Costs." Policy Research Working (United Nations Economic Commission for Paper 2257, World Bank, Washington, DC. Africa). 2000. Review of the Implementation Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim Status of the Trans-African Highways and the Kelly, Richard Green, and Alvaro Federico Barra. Missing Links. Volume 2. Tunis and Addis Ababa: 2008. "Connecting the Continent: Costing the African Development Bank and United Nations Needs for Investment in ICT Infrastructure in Economic Commission for Africa. Africa." Background Paper 3, Africa Infrastruc- Arvis, Jean François. 2005. "Transit and the Special ture Country Diagnostic, World Bank, Case of Land-Locked Countries." In Customs Washington, DC. Modernization Handbook, ed. Luc De Wulf and Minges, Michael, Cecila Briceño-Garmendia, Mark José B. Sokol, 243­64. Washington, DC: World Williams, Mavis Ampah, Daniel Camos, and Bank. Maria Shkaratan. 2008. "Information and Com- Bofinger, Heinrich C. 2009. "Air Transport Sector munications Technology in Sub-Saharan Africa: Review." Background Paper 16, Africa Infra- A Sector Review." Background Paper 10, Africa structure Country Diagnostic, World Bank, Infrastructure Sector Diagnostic, World Bank, Washington, DC. Washington, DC. Buys, Piet, Uwe Deichmann, and David Wheeler. Rosnes, Orvika, and Haakon Vennemo. 2008. 2006. "Road Network Upgrading and Overland "Powering Up: Costing Power Infrastructure Trade Expansion in Sub-Saharan Africa." Policy Investment Needs in Sub-Saharan Africa." Back- Research Working Paper 4097, World Bank, ground Paper 5, Africa Infrastructure Country Washington, DC. Diagnostic, World Bank, Washington, DC. Collier, Paul. 2007. "Poverty Reduction in Africa." Teravaninthorn, Supee, and Gaël Raballand. 2008. Proceedings of the National Academy of Sciences Transport Prices and Costs in Africa: A Review of 104 (43): 16763­68. the Main International Corridors. Directions in Deepening Regional Integration 161 Development Series. Washington, DC: World Development Framework." White Paper, USAID, Bank. Washington, DC. UNESCO (United Nations Educational, Scientific World Bank. 2007. The Development Potential of and Cultural Organization). 2003. Conflict Pre- Regional Programs: An Evaluation of World Bank vention and Cooperation in International Water Support of Multicountry Operations. Washington, Resources. Paris: UNESCO. DC: World Bank, Independent Evaluation Group. ------. 2006. The 2nd United Nations World Water ------. 2008. "Africa Regional Integration Assis- Report: Water, a Shared Responsibility. Paris: tance Strategy." World Bank, Washington, DC. UNESCO. ------. 2009. World Development Report 2009: USAID (U.S. Agency for International Develop- Reshaping Economic Geography. Washington, ment). 2008. "Sub-Saharan Africa's Power Pools: DC: World Bank. Par t 2 Sectoral Snapshots Chapter 7 Information and Communication Technologies: A Boost for Growth I nformation and communication technolo- to address the specific challenges facing the gies (ICTs) have been a remarkable success ICT sector in Africa: in Africa. Sector reform, particularly in the mobile segment of the market, has trans- · Complete the reform agenda by estab- formed the availability, quality, and cost of lishing full competition throughout the connectivity across the continent. In less than sector. 10 years, mobile networks have covered 91 · Revise the licensing framework to accom- percent of the urban population, and cover- modate rapid technological change and age in rural areas is growing. However, these emerging competition. high overall levels of coverage hide significant variation between countries, and particularly · Reform the state-owned enterprises (SOEs) that hinder sector growth and in the proportion of their populations that development. have access to services. Some countries have been much more successful in providing basic · Ensure low-cost international access infra- structure by preventing monopoly control voice services than others, and some segments over bottleneck facilities. of the market, such as fixed-line telephone service and broadband Internet, have been less · Promote the development of high-bandwidth successful than the mobile segment. Fixed-line backbone infrastructure (the networks that penetration rates remain low and are falling carry communications traffic between fixed in most countries, while broadband Internet points in a network). is expensive and available to only a small pro- · Stimulate innovation in the use of wireless portion of the population. technologies by reforming the way the radio Although large parts of the ICT sector have spectrum is allocated and managed. been transformed, much remains to be done. · Promote universal access to ensure that ICT Policy makers need to take the following steps availability is as extensive as possible. 165 166 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION The African ICT Revolution Figure 7.2 Global System for Mobile Communications Coverage in Africa, 1998 to Third Quarter of 2006 In Africa, the greatest expansion in ICT has 100 94 been in voice services. Internet services, in 80 contrast, have grown only slowly. Overall, the 79 % of population ICT sector has had a strong, positive effect on 60 62 Africa's GDP. 40 43 43 Access to Basic Voice Services 20 16 22 Sub-Saharan Africa has witnessed dramatic 4 growth in the penetration of ICT services since 0 1 1998 1999 2000 2001 2002 2003 2004 2005 2006 the mid-1990s--mainly in mobile telecommu- rural urban total nications, where the number of mobile users grew from 10 million in 2000 to more than Source: Minges and others 2008. 180 million in 2007 (ITU 2008). During the mid-2000s, more than 25 million new mobile subscribers were added each year, and annual landlines after introducing competition), growth rates exceeded 30 percent (figure 7.1). growth rates for fixed lines have stagnated or The fixed-line market has grown much more turned negative. For example, the number of slowly, from 10 million fixed telephone lines fixed lines in South Africa declined by 300,000 in 2002 to 11.8 million in 2006. between 2000 and 2007. Competition among mobile operators has Access to new ICT services has been remark- created a race to increase the percentage of the ably broad. Across Africa, the rural mobile population covered by their networks. By 2006, penetration rate is 3 percent, while in middle- one or more of the mobile networks covered income countries it is as high as 13 percent. 62 percent of the Sub-Saharan population, In urban areas, the penetration rates range which was hence able to access a mobile signal, from 22 percent in low-income countries to whether they actually subscribed to the service 38 percent in middle-income countries. Even or not. This coverage continues to increase people among the lowest income groups have each year (figure 7.2). access to ICT through mobile networks; in the All countries in the region have seen growth bottom three income quintiles, access ranges in the use of mobile telephones, but with the from 1.6 percent to 5.5 percent. In middle- exception of Nigeria (which added 750,000 income countries, the penetration rate in the lowest-income quintile is 10 percent. The widespread use of prepaid telephone service has revolutionized access to mobile Figure 7.1 Growth of Mobile Subscribers in Africa, 1998­2006 networks for low-income households. An 100 estimated 97 percent of consumers in Sub- Saharan Africa are prepaid users. With pre- penetration per 100 inhabitants 80 paid charging systems, customers can purchase services in small increments and control their 60 expenditures. Operators have introduced other innovative price schemes, some targeted at 40 poor customers: low-cost on-net calling, caller ID (to facilitate callbacks within social and 20 business networks), low and sometimes free off-peak tariffs, and systems to transfer mobile 0 fixed telephone lines mobile subscribers internet users phone credit electronically between subscrib- high income East Asia & Pacific South Asia Sub-Saharan Africa ers. For operators, these schemes, particularly prepayment, dramatically reduce credit risk Source: Minges and others 2008. and the cost of revenue collection. Information and Communication Technologies: A Boost for Growth 167 This rapid growth in access to ICT in Africa Figure 7.3 Price of One-Minute, Peak-Rate Call to the has happened despite the relatively high price United States, 2006 of services. In 2007, a representative basket of Chad 4.9 prepaid mobile services cost $12.58 a month, Cape Verde 2.1 six times the $2 that it cost in Bangladesh, India, Zambia 1.8 and Pakistan. Prices in Africa are declining but Ethiopia 1.1 not as fast as in other world regions. In 2000, Malawi 1.0 each mobile subscriber paid about $39 a month Namibia 0.9 Madagascar 0.8 in African countries and in Bangladesh, India, Tanzania 0.7 and Pakistan. By 2005, that figure had fallen to Niger 0.7 $7 in Bangladesh, India, and Pakistan but only Lesotho 0.6 to $20 in Africa. If prices were to fall to the lev- Rwanda 0.5 els seen in South Asia, access to ICT in Africa Uganda 0.5 could be significantly higher. Sudan 0.4 The average price of international calls in Congo, Dem. Rep. of 0.4 Mozambique 0.4 Sub-Saharan Africa has fallen significantly Burkina Faso 0.4 since 2000, but prices for calls to countries Benin 0.4 outside the region remain much lower than Senegal 0.3 for calls within the region. The average peak Nigeria 0.3 price of a one-minute call from Africa to the Côte d'Ivoire 0.3 United States is $0.45, compared with $1.23 for Kenya 0.2 an international call within Africa. These aver- Cameroon 0.2 South Africa 0.2 ages mask significant variation among coun- Ghana 0.2 tries (figure 7.3). Price variation is much lower South Asia 0.3 for calls within Africa. 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 US$ per minute Access to the Internet Source: Minges and others 2008. Unlike the expanded access to basic voice ser- Note: Peak rate includes taxes. vices, rates of access to the Internet are low and growing only slowly in Africa. High prices and limited availability are key reasons, com- Figure 7.4 Price Basket for Internet Access, 2005 pounded by poor fixed-line access networks, 70 1.0 limited access to the broadband radio spec- 60 0.9 trum, poor domestic backbone networks, and 0.8 50 US$ per month 0.7 limited use of computers (figure 7.4). % of GDP 40 0.6 0.5 30 0.4 0.3 ICT Sector Developments 20 0.2 10 0.1 Growth in the ICT sector in Africa has taken 0 0 South Asia high income East Asia & Pacific Sub-Saharan Africa place primarily in mobile phones through global systems mobile (GSM) networks. The Internet access price (US$) Internet access price (% of GDP) economies of scale generated by the global Source: Minges and others 2008. standardization of GSM equipment have dra- matically reduced prices of handsets and net- work equipment, and international standards contentious issue in many parts of the world, allow customers to use networks in more than but multinational African operators such as one country at a low cost. As a result, several Celtel, MTN, Safaricom, and Vodacom have pan-African operators have emerged, and they international on-net tariffs offering savings to are highly innovative in their tariffs and ser- their customers of about 15 percent per min- vices. For example, international roaming is a ute in call charges. The major alternative global 168 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION standard for wireless voice services, Code- increasing investment in ICT services results Division Multiple Access, is also making slow in higher long-run rates of economic growth but steady progress in Africa. Operators in the (Roeller and Waverman 2001); according region have also pioneered innovative services to estimates, the ICT revolution in Africa is using the mobile telephone networks, such as responsible for about 1 percentage point of mobile banking and remittance payments. the improvement in Africa's per capita eco- The fixed-line market in Sub-Saharan Africa nomic growth rate between the mid-1990s and continues to be dominated by incumbents-- the mid-2000s (Calderón 2008). This positive operators that were either formerly or are effect will continue as investment in the sector currently owned by the state--and their perfor- continues and as the use of ICT raises produc- mance remains relatively poor (figure 7.5). The tivity in all types of businesses. productivity of these operators is low, and most Large-scale private investment, reaching of them have higher levels of personnel than a cumulative value of about $20 billion, has international benchmarks, as measured by the driven the expansion of access to ICT. Between number of lines per employee. In July 2008, the 1992 and 2005, the vast majority of the 82 pri- Sub-Saharan operators ranged from 20 to 346 vate sector transactions in the ICT sector were lines per employee, whereas the figure was 427 for new operations in mobile communica- in Latin America and the Caribbean region, and tions (World Bank 2009). SOE privatizations 700 in Organisation for Economic Co-operation and license fees generated a further $3.3 bil- and Development member countries. lion of revenues for the state. This investment The low productivity of African incumbent continues today, as new deals in the region are telecommunication companies creates hidden announced regularly. The current financial costs for the economy, through suboptimal crisis has adversely affected investment rates, allocations of resources to the sector and low however, limiting operators' access to finance. consumption of telecommunication services. Overall ICT employment has grown as the The cost of this excess labor can be on the mobile sector directly and indirectly added order of 0.4 percent of GDP, and even higher in jobs in African countries. Multiplier effects and some cases (figure 7.6). For Cameroon, Ghana, new lines of business (mobile airtime agents and Namibia, among other countries, the level and m-transactions) also add to employment of this inefficiency exceeds the cost of meeting growth and income generation. In East Africa, universal access targets. the mobile industry directly and indirectly pro- vides employment for close to 500,000 people Economic Impact of the ICT Industry (GSMA 2007). The ICT sector has positively affected eco- The new ICT infrastructure and related nomic growth in Africa. Research shows that reforms have increased government revenues through one-time fees for licenses and ongoing payments through licenses and taxes. The rev- enue generated by the ICT industry in African Figure 7.5 Net Change in Fixed-Line Market, 2001­05 countries ranges from 1.7 percent to 8.2 percent of GDP, with an average of 4.0 percent. The tax 600 and license revenues generated by the industry net change in no. of subscribers 500 have also had a significant positive fiscal effect 400 (figure 7.7). (thousands) 300 200 100 0 Institutional Reforms in the ­100 ICT Sector ­200 2001 2002 2003 2004 2005 South Africa Nigeria others Market liberalization has been the most important cause of the ICT sector's growth Source: Minges and others 2008. in Africa. Regulatory reforms and the Information and Communication Technologies: A Boost for Growth 169 privatization of SOEs have complemented Figure 7.6 Costs of Overstaffing for Fixed-Line Incumbents in Selected Countries these market reforms. 1.0 Market Reforms 0.8 The widespread liberalization of markets in % of GDP 0.6 Africa and the emergence of competition, par- ticularly among mobile operators, have been 0.4 the main drivers of the ICT sector's perfor- 0.2 mance (figure 7.8). Countries with more competitive markets 0 a pia d n a nia a nin e a cover, on average, 64 percent of their popu- u ny ibi ric an a oo biq Ch hio za Be Ke Af m Gh er n Na am m Et h Ta lation with mobile networks, compared with Ca ut oz So M 57 percent for the less competitive markets. investment needs overstaffing costs Among the low-income African countries, Source: Minges and others 2008. those with more competitive markets have 31 percent higher mobile penetration rates, 6 percent lower mobile prices, and 39 percent lower international call prices (as measured by the price of a call to the United States). Figure 7.7 Value Added and Excise Taxes on Mobile The benefits of market liberalization Telephone Services, 2006 increase as competition intensifies. In gen- eral, the annual increase in penetration rises Uganda as more firms enter the market. Relatively Tanzania little growth occurs in market penetration in Kenya the initial change from monopoly to duopoly, Niger but when a country issues its fourth mobile Cameroon license, penetration rates increase, on aver- Senegal age, by almost 3 percentage points every year. Madagascar A country's average income also affects the Côte d'Ivoire performance of the telecommunication sector. Chad In poorer countries, increased competition is Burkina Faso felt most strongly when a market reaches four Benin operators, whereas for middle-income coun- Zambia tries, the effect is strongest when a third opera- Malawi tor is introduced. Mozambique Some countries that have established a Rwanda legal framework for a liberalized market have Namibia nonetheless failed to establish effective com- Ghana petition. Few countries have legislation with Ethiopia outright prohibitions on competition in tele- Cape Verde communications, but many have restrictions South Africa on competition arising from exclusivity clauses Congo, Dem. Rep. of granted in licenses to existing operators. In Sudan 12 countries where data were available, a gap of Nigeria at least two years elapsed between ending the Lesotho legal restrictions on competition and granting 0 5 10 15 20 25 30 new licenses. Twelve Sub-Saharan countries % surcharge on mobile service have competition in the fixed-line and inter- value added tax excise tax national markets, but only a few of them have Source: Minges and others 2008. more than two operators in these segments. Note: In Kenya, Tanzania, and Uganda, the excise taxes shown are applicable to mobile calls. Rwanda is planning to implement an Even in the mobile segment, barely half of the excise tax on mobile airtime. 170 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 7.8 Status of Mobile Competition, 1993­2006 Figure 7.9 Tariff Rebalancing in African Countries with a Liberalized Telecommunication Sector, 24 1993­2006 20 number of countries 250 16 price index 2000 = 100 200 12 150 8 4 100 0 50 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 no network monopoly duopoly three or more operators 0 2000 2001 2002 2003 2004 2005 fixed basket mobile basket Source: Minges and others 2008. three-minute peak rate call to U.S. Source: Minges and others 2008. Note: Price index represents prices as a percentage of their level countries have more than three active opera- in the base year 2000 so that only relative changes over time are tors despite evidence that most markets in the highlighted. region can support more. Across the region, market reform is far from being complete. important. It can be challenging for regula- Regulatory Reform tors with limited technical capacity and legal Changes in the legal framework that governs powers to impose decisions on operators, but the sector have matched the reforms in mar- some regulators have recently been successful. ket structure. All African countries have laws In Tanzania, the national regulatory author- and regulations covering telecommunications. ity introduced a phased reduction in mobile Typically, a new law (supplemented by decrees termination rates, based on a calculation of and regulations) establishes a national regula- operators' costs. Nigeria's regulatory authority tory agency with general provisions for com- has established a target for mobile termination petition, licensing, interconnection, allocation rates, and in Kenya, the regulatory authority of scarce resources (for example, numbering recently established a ceiling on mobile termi- and spectrum), and pricing. Of the 24 coun- nation rates, as well as a cap on the retail price tries sampled, 23 have such a broad institu- of calls between networks. tional framework and independent regulatory authorities, up from 5 in 1996. Reform of State-Owned Enterprises Regulators still exert considerable control As governments have liberalized their markets over incumbent fixed-line operators' tariffs. and reformed the institutions for regulating They have allowed these operators to respond them, many have also reformed the opera- to competition by rebalancing their tariffs, tors they formerly owned. By the end of 2006, allowing monthly fixed charges to increase and 15 African countries had sold shares in their national and international call prices to fall. state-owned telecommunication operator to Regulators have less control over mobile oper- the private sector. These transactions largely ators' tariffs, which are forced down through involved equity and management partner- competition (figure 7.9). ships with strategic investors; only Nigeria and Some regulators have been successful in set- Sudan privatized by issuing shares on local or ting wholesale tariffs (the charges that opera- regional stock exchanges.1 From 1993 to 2006, tors pay each other for handling calls when they the total value of such privatizations was $3.5 pass from one network to another), which feed billion, half of which was accounted for by directly into the retail price that customers pay South Africa's Telkom. for their calls. As competition develops, par- The nature of strategic partnerships and ticularly among mobile operators, regulating their success have varied over time. Direct these interconnection charges becomes more investment by developed-country investors in Information and Communication Technologies: A Boost for Growth 171 the fixed-line business has been complemented to benefit from lower prices and better quality by sales to developing-country investors, par- of service. ticularly from the Middle East and South Asia. The performance of these privatizations and Domestic Backbone Infrastructure partnerships has been mixed. In some cases, Backbone network infrastructure to carry such as Uganda, the privatization of the state- communications traffic between fixed points in owned fixed-line incumbent was part of a the networks is limited, thus constraining the successful overall reform of the sector. In oth- development of broadband Internet. Mobile ers, private investors have withdrawn, resulting operators do not require high-capacity back- in the renationalization of ICT assets. Ghana bone networks to carry voice traffic and have and Rwanda resold the businesses after the typically developed their own using wireless first privatization failed, indicating a sustained technologies. Broadband Internet backbone commitment to reform. networks need much greater capacity, how- Despite the notable successes, the gov- ever, typically using fiber-optic cables. The ernments of many African countries retain limited extent of these networks is a constraint ownership of at least one telecommunication on the development of the broadband market operator, which distorts the market and creates in Africa. inefficiencies. Thus, the region has some dis- Considerable variation exists across the tance to go before it has a fully privately owned region in how markets for domestic backbone and competitive telecommunication market. infrastructure operate. In many countries, both implicit and explicit constraints limit Access to International Connectivity development of this type of infrastructure. For One of the main drivers of the high cost of example, mobile operators may be required to Internet and of international voice calls is use the incumbent's backbone network, or the price of international connectivity, deter- they might be allowed to build their own but mined by physical access to submarine fiber- not to sell backbone network services to other optic cables and the level of competition in the operators on a wholesale basis. These types of international market. Countries with access to regulations limit the development of back- submarine cables have lower international call bone networks and hinder the development of prices than those without access. Nevertheless, broadband. countries that have competitive access have Countries that have fully liberalized the significantly lower prices than those retaining market for backbone networks have seen a gateway monopoly (table 7.1). rapid growth in infrastructure competition. In Access to high-capacity submarine fiber- Nigeria, at least four of the major operators are optic infrastructure is therefore a necessary developing high-capacity fiber-optic cable net- but insufficient condition for low-price inter- works capable of supporting high-bandwidth national voice services. Countries also need to services, and a similar pattern is emerging in ensure that the international facilities segment Kenya. These networks are concentrating on of the market is competitive if customers are major urban areas and on interurban links Table 7.1 Prices for Access to International Voice and Internet Connectivity Price for a Price for a Price for 20 hours Price for ADSL call within call to the of dial-up broadband Share of Sub-Saharan Africa United States Internet access Internet access Access level countries (%) ($ per minute) ($ per minute) ($ per month) ($ per month) No access to submarine cable 67 1.34 0.86 67.95 282.97 Access to submarine cable 32 0.57 0.48 47.28 110.71 Monopoly on international gateway 16 0.70 0.72 37.36 119.88 Competitive international gateways 16 0.48 0.23 36.62 98.49 Source: Minges and others 2008. Note: ADSL = asymmetric digital subscriber line. 172 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION where the majority of customers are. If high- Nevertheless, the size of the coverage gap var- capacity backbone networks are to extend ies immensely across countries (figure 7.10), beyond these areas, some form of public sup- and in a handful of cases (the Democratic port will likely be needed, preferably in part- Republic of Congo, Madagascar, and Zambia) nership with the private sector. can exceed 15 percent of the population. These analytical results are robust; the size of the coverage gap increases only from 4.4 Completing the Remaining percent to 5.9 percent of the population if the Investment Agenda amount spent on telecommunication services drops from 4 percent of GDP per capita (the baseline assumption) to 3 percent. Similarly, Voice Services even if costs were three times greater than in the The cost of completing mobile network cover- base case, the coverage gap would increase from age for voice in Africa is relatively modest. By 4.4 percent of the population to 12.6 percent. adopting a spatial approach to modeling the cost of providing access to mobile phone net- works, reliable estimates have been developed of Internet Services the capital and operating expenditures required Despite the anticipated positive economic for completing the rollout of GSM voice signal effect that widespread use of broadband would throughout Africa. Potential revenues are esti- have on African economies, mass-market mated based on population density and income broadband Internet at speeds seen in other distribution. Potential costs are estimated based parts of the world is unlikely to be commer- on terrain characteristics and cell size and the cially viable in Africa for the near future. The resulting number of additional base stations broadband Internet available in most African needed to complete national GSM coverage. countries is typically limited to major urban These raw base station numbers drive estimates areas and to Internet cafés, businesses, and of capital and operating expenditures. high-income residential customers. Network Reaching all the unserved population coverage is limited, prices are high, and speeds would require investments of $0.8 billion a are lower than in other regions of the world. year over 10 years. Currently, 43.7 percent of This limited current level of service could be the population lives in areas not covered by expanded to national coverage using wireless wireless voice networks. If the right competi- network infrastructure with the same techni- tive environment is established, the private cal and economic advantages as GSM voice sector could fill most of this gap, reaching 39 networks (lower operating and security costs percent of the population--the vast majority than wired networks and the potential to use of the unserved--with a voice signal. Only $0.3 prepaid billing systems). The investment to billion per year of public investment would be cover the entire population using limited- needed to reach the remaining 4.7 percent of performance wireless broadband technology has the population in the coverage gap (table 7.2). been estimated at approximately $0.9 billion. Table 7.2 Investments Needed to Close Gaps in Voice and Broadband Coverage in Sub-Saharan Africa Voice coverage Broadband coverage Total Efficient Coverage Total Efficient Coverage Indicator investment market gap gap investment market gap gap Average annual investment ($ billions) 0.8 0.5 0.3 0.9 0.7 0.2 Percentage of population affected 43.7 39.0 4.7 100.0 89.0 11.0 Source: Mayer and others 2008. Note: Efficient market gap is the portion of total investment need that the private sector could meet under commercial terms if all regulatory barriers to entry were dismantled to allow the market to function efficiently. Coverage gap is the portion of the total investment need that the private sector could not meet even under efficient market conditions. This gap would require public subsidy because the service lacks commercial viability. Information and Communication Technologies: A Boost for Growth 173 Figure 7.10 Voice Coverage Gaps in 24 Sub-Saharan Figure 7.11 Broadband Coverage Gaps in 24 African Countries Countries Congo, Dem. Rep. of Congo, Dem. Rep. of Madagascar Madagascar Zambia Mozambique Niger Zambia Mozambique Niger Chad Namibia Namibia Chad Lesotho Malawi Tanzania Lesotho Ethiopia Tanzania Cameroon Cameroon Cape Verde Ethiopia Burkina Faso Cape Verde Malawi Burkina Faso Sudan Senegal Kenya Kenya Senegal Sudan Côte d'Ivoire Côte d'Ivoire Benin Benin Ghana Uganda Uganda Ghana Rwanda Rwanda Nigeria South Africa South Africa Nigeria 0 20 40 60 80 100 0 20 40 60 80 100 % of population % of population existing coverage efficient market gap efficient market gap coverage gap coverage gap Source: Mayer and others 2008. Source: Mayer and others 2008. Note: Efficient market gap is the portion of the unserved market Note: Efficient market gap is the portion of the unserved market that the private sector could serve under commercial terms if all that the private sector could serve under commercial terms if all regulatory barriers to entry were dismantled to allow the market regulatory barriers to entry were dismantled to allow the market to function efficiently. Coverage gap is the portion of the unserved to function efficiently. Coverage gap is the portion of the unserved market that the private sector could not serve even under efficient market that the private sector could not serve even under efficient market conditions. This gap would require public subsidy because market conditions. This gap would require public subsidy because the service lacks commercial viability. the service lacks commercial viability. As long as the right competitive environment Republic of Congo, Madagascar, Mozambique, is established, the private sector would cover Namibia, Niger, and Zambia) can exceed 20 most of that amount, which could reach 89 percent of the population (figure 7.11). percent of the population with this limited- reach broadband access. Only $0.2 billion of Backbone Networks public investment would be needed to reach Although the existing limited level of broad- the remaining 11 percent of the population in band service could be expanded at relatively low the coverage gap (table 7.2). However, the cov- cost, the shift toward higher-speed mass-market erage gap varies hugely across countries, and broadband Internet access in Africa at prices that in a handful of cases (Chad, the Democratic would be affordable for a significant proportion 174 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION of the population would involve major invest- as provider of communication infrastructure ments in backbone infrastructure. The revenue has been superseded in most countries by a generated from customers would be insufficient new role as establisher and regulator of market to make this investment commercially attractive. structure. Few countries in the region, if any, If governments wished to achieve this level of have completed the reform agenda, however. broadband Internet access, significant levels of Regulatory frameworks still contain restric- public subsidy would likely be required. tions on investment and competition, and the High-bandwidth backbone networks are a poor quality of regulation in many countries key part of the investment needed for broad- creates costly inefficiencies. Many incumbent band in Africa. These networks connect towns operators remain under state ownership, cre- and cities within countries and across borders. ating a burden on the public sector, inefficien- They also link to the international submarine cies in the market, and conflicts of interest for fiber-optic cable networks that carry commu- regulators. Major challenges therefore remain if nications traffic between continents. widespread ICT services at affordable prices are Cross-border and interregional connectiv- to be available. ity in Africa is currently underdeveloped. One- time investment needs range from $229 million Completing the Reform Agenda for a minimum set of links to $515 million for Establishing full and effective competition in an extensive interregional network connecting the ICT sector can deliver rapid and sustained all African capitals to one another with fiber- improvements in the availability of commu- optic cables. The private sector will provide nication services. The majority of countries much of that investment as regional operators have implemented some reforms, but they still connect their networks across borders. Private have a long way to go. Completing the reform investment is also driving international sub- agenda is therefore the primary challenge fac- marine cable infrastructure in Africa. Of the ing the ICT sector in Africa. five major submarine fiber-optic cables either The most important remaining reform is to already operating or under construction in the increase competition through further market region, only one has direct government involve- liberalization. In practice, that means issuing ment; four are owned and financed by private more licenses and reforming the licensing struc- operators on commercial terms. These two ture to allow operators more freedom to inno- types of backbone infrastructure are linked. vate and compete across a range of services. As submarine fiber-optic cables are developed, As mobile networks expand into marginal cross-border links to channel traffic to landing areas, reducing the cost of network rollout and points become more commercially viable. operation will become a more important aspect Aside from routes connecting major urban of the reform agenda. Some form of collabora- centers, high-bandwidth backbone networks tion among competitors (for example, in the are unlikely to be commercially viable. Back- civil engineering aspects of networks, such as bone network development in these areas may masts and towers) could reduce costs enough require some form of public support, either to allow companies to operate profitably in through financial support or through the pro- areas where they would not otherwise be able vision of easier access to existing infrastructure to do so. Regrettably, this type of collaboration (for example, transport and energy networks). can also enable collusion among operators, so it must be carefully regulated. The reform agenda will evolve as the market Policy Challenges changes. Competition regulation is increasingly becoming part of modern sector legislation in The liberalization of telecommunication mar- Africa, particularly regarding the behavior of kets since the mid-1990s has provided affordable dominant operators and controlling access to ICT services to the public. It has also radically essential facilities. Even where a country may reshaped the roles of the public and private not have competition legislation, ICT regu- sectors. The traditional role of the public sector lators are applying the tools of competition Information and Communication Technologies: A Boost for Growth 175 analysis in telecommunication regulation. They Internet Protocol (VoIP), limited mobility, and are also adapting their regulatory approach to Internet protocol television. Many licensing reflect the evolving marketplace: for example, regimes restrict either the VoIP technology or by relaxing controls on tariffs as competition its derived services. Direct consumer access to becomes more effective at controlling them. VoIP allows voice calls over Internet connections African countries will continue to see ben- instead of the public switched telephone net- efits as competition intensifies and access to works. Such services offer much lower prices for ICT increases. As prices fall, even as far as the long-distance and international calls; however, $0.01­$0.03 per minute range currently seen in restrictions are common because the widespread South Asia, mobile phone services will become use of VoIP could undermine the main sources affordable to much of the African population, of voice revenues for incumbent operators. bringing with it positive economic and social Licensing constraints on the mobility of benefits. Governments will also benefit from specific wireless telecommunication operators the expansion of telephone services. First, lower are common in Africa. Operators with limited- prices will fuel uptake and access to services, mobility licenses can provide wireless telecom- directly reducing the costs involved in deliver- munication services while allowing customers ing universal service. Second, greater competi- to move around within a limited area. No tion will expose the hidden costs of the incum- technological reason exists why these networks bent state-owned operators, which represent cannot offer full mobility; the restrictions a burden on government finances and a more are often imposed to protect existing mobile general effect on the economy. The expansion licensees. As competition in the full-mobility of the ICT sector resulting from market liber- market increases, these restrictions will seem alization will increase the tax and license-fee increasingly anachronistic. revenues earned by governments, and ICT ser- Finally, the use of Internet networks to vices themselves will become a more effective provide television services is increasing in platform for delivering public services. Africa as the number of broadband subscrib- ers increases. This raises many challenges for Revising the Licensing Framework regulatory systems that have traditionally dealt The traditional model of licensing is becoming with communication and broadcasting media obsolete. In the first wave of market liberaliza- through separate institutions and through sep- tion, licenses were linked to market segments arate legal and regulatory frameworks. These and technologies. GSM licenses granted the separations are creating obstacles to invest- right to provide mobile communications ment and competition as convergence blurs in specific spectrum bands using a specific the boundaries between the technologies. technology, and data licenses were granted to The initial response of policy makers to operate in specific value added markets. Two these trends has been to move toward unified factors are making this traditional approach to licenses that remove technological distinctions licensing obsolete. First, the growth of compet- and allow operators to provide a full range of itive ICT markets in Africa has demonstrated services to customers. The design and imple- that multiple players can compete successfully, mentation of a unified licensing system can be even in small markets. Managing liberaliza- complex, however, magnified by the need to tion through technology and service-specific adjust a wide range of existing rights and obli- licenses has therefore proved to be ineffective gations, annual fees, and acquisition costs. This as a policy tool. Second, technological conver- adjustment can be done in a transparent way gence allows networks to deliver multiple ICT through public consultation, but the migration services, thereby reducing costs and promoting process has to be managed carefully to avoid service innovation. The traditional approach undue destabilization of the market. to licensing often prevents operators from tak- In the medium term, licenses will have to ing advantage of this convergence. become simpler and less restrictive to facili- The negative effect of current licensing tate the development of new services at lower frameworks is especially evident in Voice over prices. Ultimately, the challenge for African 176 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION countries will be to migrate from the current Providing SOEs with monopoly control over licensing regime to one in which controls on specific segments of the market to make them market entry and the services delivered by more attractive to potential buyers will ulti- market parties would be largely abandoned. mately be unsuccessful, for it will distort the The countries of the European Union have market and constrain its development. taken this approach, moving from a system of individual licenses to a general authoriza- Ensuring Low-Cost Access to tion regime. International Infrastructure Creating the conditions for widespread broad- Reforming State-Owned Enterprises band access is a complex policy issue facing Reform of state-owned, fixed-line incumbents the ICT sector in Africa. The markets in the remains a major policy challenge for govern- region are so different from those in other ments in the region. The last decade has seen parts of the world that governments have no the fixed-line incumbent operators eclipsed. obvious models to draw on. Some lessons are Compared with the mobile operators, they beginning to emerge, however. One, in par- now play a minor role in telecommunication ticular, is the importance of access to high- service in most African countries. Incum- capacity, low-cost bandwidth via submarine bent operators can be a disruptive force in fiber-optic cable infrastructure. the economy through misallocation of public The private sector has demonstrated its resources, use of incumbents as social buf- capacity to develop, finance, and operate such fers, and the regulatory uncertainty created by cables in Africa. The challenge for govern- their presence in the market. In some cases, ments is to minimize the obstacles to this type preferential treatment of these operators-- of investment by readily issuing cable opera- exclusivity agreements (for example, in con- tors permits and licenses. The development of trol of international gateways and backbone infrastructure, on its own, will not guarantee capacity), banning of innovative services such better services for customers. The experience as VoIP, and distortion of prices--inhibits of the South Atlantic 3/West Africa Subma- innovation and investment and amplifies the rine Cable (SAT-3) cable on the west coast economic burden of SOEs on national econo- of Africa shows that physical access to a cable mies. This issue has emerged again as some is necessary but not sufficient for low-cost Sub-Saharan governments finance the devel- connectivity. A consortium of private opera- opment of fiber-optic backbone networks tors with little direct regulation controls access through their SOEs. to the SAT-3 cable. Because these operators At a minimum, SOEs should be brought are protected from competition on the cable, fully within the regulatory and licensing frame- customers have not received the full potential work so that they are treated in the same way benefit of the facility. The challenge for gov- as private operators. This move will stimulate ernments seeking to improve access to inter- competition and efficiency in resource alloca- national infrastructure is to avoid creating tion. Encouraging greater private participation infrastructure bottlenecks and to encourage in SOEs to transform and grow the businesses competition between submarine cables and may also be appropriate. Given the state of landing stations. Where they cannot do this, many incumbent operators' networks, that may regulators should ensure access to the facilities require some form of financial and manage- on equitable terms. ment incentives to attract partners and inves- Landlocked countries face a special chal- tors. The challenge for governments will be to lenge in ensuring that their operators have ensure that this transition is achieved without access to submarine fiber-optic infrastruc- distorting the market. It can be done by allocat- ture. If the private sector does not provide ing mobile and other wireless spectrum to these competitive infrastructure in the intervening operators, offering management control, and countries, the government may have a role to minimizing network coverage commitments. play through public-private partnerships. Information and Communication Technologies: A Boost for Growth 177 Promoting the Development of the radio spectrum and is challenging govern- High-Bandwidth Backbone Networks ments' traditional systems of spectrum allo- Domestic backbone networks will become cation and management. When one or two more important as governments focus their operators and the government dominated attention on delivering affordable broadband the mobile market, management of the radio Internet. Without these networks, countries spectrum did not present major challenges will have difficulty making broadband services to governments. Market liberalization and widely available at prices that significant num- technological innovation have increased the bers of people are willing to pay. number of players wishing to use the radio Private operators are investing considerable spectrum, particularly for new broadband resources in this infrastructure, and the pace is wireless Internet services. The way in which increasing as operators look at broadband as a access to the radio spectrum is organized is source of future market growth. Such network therefore an increasingly important issue for development is typically limited to urban areas the development of the ICT sector. and interurban routes where the private sector The traditional approach to organizing the is willing to invest in network development. spectrum's use is to constrain development of No single policy approach exists to backbone the ICT sector. Governments have tradition- network development. Some governments pro- ally accomplished this by deciding how each mote a competition-only policy, whereas oth- frequency band is used and who is entitled to ers invest public resources in publicly owned use it. This approach is unsuitable for markets networks. Regulatory frameworks often con- with multiple players and spectrum uses that strain investment through restrictions on fixed- are continually changing. Governments are network investment and on the services that ill suited to decide the best uses for the radio can be sold. Wholesale markets in backbone spectrum and are typically unable to move fast services are thus underdeveloped, contribut- enough in the allocation process. The effect is ing to the high price and limited availability of to constrain market development, particularly broadband Internet in the region. in new segments of the market, such as broad- Successful policy for domestic backbone band Internet. network development must encourage pri- The introduction of market forces will vate investment in commercially viable areas improve management of the radio spectrum. and provide public support for investment in Where demand for the right to use certain areas that are not commercially viable. Such a areas of the radio spectrum exceeds supply, policy should encourage infrastructure com- usage rights can be auctioned. Such spectrum petition by removing regulatory restrictions auctions are widely used in developed coun- and should reduce the cost of investment in tries, and similar systems are used in Africa fiber-optic infrastructure by providing access to allocate mobile network licenses (which to alternative transport and energy infra- usually include the right to use specified sec- structure. Public resources should focus on tions of the radio spectrum). Market forces areas of the country that are not commercially can also be introduced in spectrum manage- viable. To the extent that public investment is ment after initial allocations have been made needed, it should be made in partnership with by establishing formal property rights over the private sector to ensure that the design of the spectrum and allowing owners to trade the infrastructure meets the needs of market them. Establishing such primary and second- participants. ary markets in spectrum usage would free up the spectrum and would help ensure its most efficient use. Reforming Management of the Radio Further evolution in how the radio spec- Spectrum trum is managed is possible by establish- The rapid evolution of ICT markets in Africa ing a shared-use system for certain bands of has increased the number of potential users of spectrum, known as a commons approach. 178 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Recent developments in wireless technology these countries, a service target of 100 per- have allowed multiple users of the same radio cent coverage may be economically feasible. spectrum bands to operate without undue In countries where the gap left by the market interference. Allowing anyone to share the is larger, a more modest target will likely be radio spectrum, with little or no registration necessary. and usually without a fee, reduces the cost of When a universal service target is set, the entry into the market and therefore encour- major challenge is to establish a mechanism ages innovation in technology and service to achieve it. The majority of countries across delivery. the region currently apply a universal service Changing how the radio spectrum is man- levy on private operators, using the funds for aged requires political will. The establishment specific ICT projects. This approach has had of a property rights scheme can arbitrarily cre- very limited success, particularly when con- ate windfall gains and losses for current and trasted with the commercially driven network future users. Some users of the radio spectrum expansion into rural areas. Universal service may be difficult to incorporate into a pure funds often suffer from bureaucratic obstacles market-based system. For example, requiring and political interference in expenditure, and users in the military or emergency services to frequently they are not spent on the sector at participate in spectrum markets may be par- all. Universal service policy in Africa there- ticularly challenging (although not impos- fore requires new thinking. The challenge is sible) and would certainly have budgetary to meet the government's policy objectives of implications for those agencies. universal service at minimum cost to taxpay- Reforming the allocation and management ers while harnessing the beneficial effects of of the radio spectrum would change the role competition. An alternative to the traditional of government. Its primary role in spectrum fund-based approach is to provide direct management would no longer be to make incentives for operators to deliver services in technical and licensing decisions. Rather, the rural areas. For example, governments could government's role would be to design, oper- offer operators a reduction in license-fee pay- ate, and regulate the market in the radio spec- ments in exchange for providing services in trum. Such a change would require changes specified areas, or they could establish pay-or- in the legal framework governing the radio play schemes in which operators can choose spectrum and the capacity of the regulatory between building networks in specified areas institutions involved. and contributing to a universal service fund, which is then used to subsidize operators that Promoting Universal Access do provide services in unprofitable areas. The As more people in Africa gain access to ICT major advantage of these approaches is the services, those who remain outside the range reduction in transfers between operators and of networks are at a disadvantage. Several gov- the government, thereby lessening bureau- ernments in Africa have attempted to extend cratic delays or the diversion of funds. access to ICT beyond the perceived limits of the Once a coverage target is defined, gov- market. The quickest and most effective way ernments may also wish to address the of getting infrastructure to poor rural users issue of access for low-income groups. is through competition. Malawi and Uganda Call-by-call resale of services has signifi- have set up effectively competitive mobile mar- cantly reduced the costs of accessing the kets that already cover over 80 percent of the network, and these systems are widespread population and are continuing to expand. in Africa (for example, the VillagePhone For the majority of countries in Africa, only program of cell phone company MTN). Uni- competition will result in mobile networks versal service targets could potentially include that cover the whole population. With a few subsidies for prepayment directed at specific exceptions, such as the Democratic Republic target groups of the population. However, they of Congo, the additional cost to make voice would have to be carefully designed to avoid network coverage universal is modest. In mistargeting and leakage. Information and Communication Technologies: A Boost for Growth 179 Notes ITU (International Telecommunication Union). The authors of this chapter are Michael Minges, 2008. African Telecommunication/ICT Indicators Mark Williams, Rebecca Mayer, Cecilia Briceño- 2008: At a Crossroads. 8th ed. Geneva: ITU. Garmendia, and Howard Williams, who drew Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim on background material and contributions from Kelly, Richard Green, and Alvaro Federico Barra. Mavis Ampah, Alvaro Federico Barra, Daniel 2008. "Costing the Needs for Investment in ICT Camos-i-Daurella, Ken Figueredo, Richard Green, Infrastructure in Africa." Background Paper 3, Mike Jensen, Tim Nelly, Maria Shkaratan, Maria Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. Vagliasindi, and Bjorn Wellenius. 1. Initial public offerings of government stakes in Minges, Michael, Cecilia Briceño-Garmendia, state-owned enterprises in the telecommunica- Mark Williams, Mavis Ampah, Daniel Camos, tion sector have more recently been carried out and Maria Shkaratan. 2008. "Information and Communications Technology in Sub-Saharan in Kenya (2008) and Burkina Faso (2009). Africa: A Sector Review." Background Paper 10, Africa Infrastructure Country Diagnostic, World References Bank, Washington, DC. Calderón, César. 2008. "Infrastructure and Growth Roeller, Lars-Hendrik, and Leonard Waverman. in Africa." Working Paper 3, Africa Infrastruc- 2001. "Telecommunications Infrastructure ture Country Diagnostic, World Bank, and Economic Development: A Simultaneous Washington, DC. Approach." American Economic Review 91 (4): GSMA (GSM Association). 2007. "Taxation and 909­23. the Growth of Mobile Services in East Africa World Bank. 2009. Private Participation in Infra- 2007." GSMA, London. structure Database. http://ppi.worldbank.org/. Chapter 8 Power: Catching Up A frica's chronic power problems have importing countries that could potentially escalated in recent years into a crisis meet more than half their power demand affecting 30 countries, taking a heavy through trade. toll on economic growth and productivity. The The operational inefficiencies of power utili- region has inadequate generation capacity, lim- ties cost $3.3 billion a year, deterring investments ited electrification, low power consumption, in electrification and new capacity, while under- unreliable services, and high costs. It also faces pricing of power translates into losses of at least a power sector financing gap of approximately $2.2 billion a year. Full cost-recovery tariffs would $23 billion a year. It spends only about one- already be affordable in countries with efficient quarter of what it needs to spend on power, large-scale hydropower- or coal-based systems, much of which is on operating expenditures to but not in those relying on small-scale oil-based run the continent's high-cost power systems, plants. If regional power trade comes into play, thus leaving little for the huge investments generation costs will fall, and full cost-recovery needed to provide a long-term solution. tariffs could be affordable in much of Africa. Further development of the regional power The key policy challenges are to strengthen trade would allow Africa to harness larger- sector planning capabilities, too often over- scale, more cost-effective energy sources, looked in today's hybrid markets. A serious thereby reducing energy system costs by recommitment to reforming state-owned $2 billion a year and saving 70 million tons of enterprises (SOEs) should emphasize improve- carbon emissions annually. Economic returns ments in corporate governance more than to investments in cross-border transmission purely technical fixes. Improving cost recov- are particularly high, but reaping the prom- ery is essential for sustaining investments in ise of regional trade depends on a handful of electrification and regional power generation major exporting countries' raising the large projects. Closing the huge financing gap will volumes of finance needed to develop gen- require improving the creditworthiness of eration capacity for export. It would also utilities and sustaining the recent upswing in require political will in a large number of external finance to the sector. 181 182 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Africa's Chronic Power Problems scale of most national power systems and the widespread reliance on expensive oil-based Africa's generation capacity, stagnant since the generation make the average total historic 1980s, is woefully inadequate today. The entire cost of producing power in Africa exception- installed generation capacity of the 48 Sub- ally high: $0.18 per kilowatt-hour with an Saharan countries is 68 gigawatts, no more than average effective tariff of $0.14 per kilowatt- Spain's, and without South Africa, the total falls hour.2 Compare that with tariffs of $0.04 per to 28 gigawatts (EIA 2006). As much as one- kilowatt-hour in South Asia and $0.07 in East quarter of that capacity is unavailable because Asia. Rising oil prices, lower availability of of aging plants and poor maintenance. hydropower, and greater reliance on emer- The growth in generation capacity has been gency leases have put further upward pressure barely half that in other developing regions. In on costs and prices. 1980, Sub-Saharan Africa was at approximately Power consumption is tiny and falling. the same level as South Asia in generation capac- Given limited power generation and low ity per million people, but it has since fallen access, per capita electricity consumption in far behind. Sub-Saharan African countries lag Sub-Saharan Africa (excluding South Africa) even compared with others in the same income averages only 124 kilowatt-hours a year, bracket (Yepes, Pierce, and Foster 2008). barely 1 percent of the consumption typical Only about one-fifth of the Sub-Saharan in high-income countries. Even if that power population has access to electricity, compared were entirely allocated to household light- with about one-half in South Asia and more ing, it would hardly be enough to power than four-fifths in Latin America. Since 1990, one lightbulb per person for six hours a day. East Asia, Latin America, and the Middle East Sub-Saharan Africa is the only region in the have all added at least 20 percentage points to world where per capita consumption is falling their electrification rates, but access rates in (World Bank 2005). Sub-Saharan Africa are relatively stagnant, as Power shortages have made service even population growth and household formation less reliable. More than 30 African countries outstrip new connections. now experience power shortages and regular At current trends, less than 40 percent of interruptions in service (figure 8.1). From African countries will reach universal access to 2001 to 2005, half of the countries in Sub- electricity by 2050 (Banerjee and others 2008). Saharan Africa achieved solid GDP growth Overall, household access to electricity in urban rates in excess of 4.5 percent. Their demand areas is 71 percent, compared with only 12 per- for power grew at a similar pace, yet generation cent in rural areas. Moreover, access rates in the capacity expanded only 1.2 percent annually. upper half of the income distribution exceed South Africa shows what happens when gen- 50 percent, whereas they are less than 20 percent eration capacity fails to keep up with demand in the bottom half. Given that rural areas account (box 8.1). In some countries, supply shocks for about two-thirds of the population, extend- exacerbated the situation. Causes of the sup- ing access presents a major challenge. Only ply shocks include droughts in East Africa; 15 percent of the rural population lives within oil price inflation, which made it difficult for 10 kilometers of a substation (or within 5 kilo- many West African countries to afford diesel meters of the medium-voltage line) and could imports; and conflicts that destroyed the power thus be added to the electricity grid at relatively infrastructure in some fragile states. low cost. As much as 41 percent of the rural Inadequate power supplies take a heavy population lives in areas considered isolated or toll on the private sector. Many African enter- remote from the grid1 and is reachable in the prises experience frequent outages: in Sen- medium term only by off-grid technologies such egal 25 days a year, in Tanzania 63 days, and as solar photovoltaic panels, which typically cost in Burundi 144 days. Frequent power outages $0.50­$0.75 per kilowatt-hour (ESMAP 2007). mean big losses in forgone sales and damaged The cost of producing power in Africa equipment--6 percent of turnover on average is exceptionally high and rising. The small for formal enterprises, and as much as 16 percent Power: Catching Up 183 Figure 8.1 Underlying Causes of Africa's Power Supply Crisis MAURITANIA CAPE VERDE MALI NIGER SE CHAD SUDAN ERITREA NE GA GAMBIA L BURKINA DJIBOUTI GUINEA FASO BENIN GUINEA-BISSAU NIGERIA GHANA D'IVOIRE CÔTE SIERRA LEONE ETHIOPIA CENTRAL ON AFRICAN REPUBLIC RO LIBERIA IA ME AL TOGO DEMOCRATIC CA M SO GO EQUATORIAL GUINEA REPUBLIC OF UGANDA GABON CON CONGO KENYA SAO TOME AND PRINCIPE RWANDA . OF S LE REP BURUNDI HEL TANZANIA YC SE MALAWI COMOROS Main cause or trigger MAYOTTE ANGOLA natural causes (droughts) ZAMBIA E IQU oil price shock MB AR ZA ZIMBABWE ASC MO system disrupted by conflict NAMIBIA MAURITIUS DAG high growth, low investment/structural issues BOTSWANA MA SWAZILAND SOUTH AFRICA LESOTHO Source: Eberhard and others 2008. BOX 8.1 South Africa's Power Supply Crisis South Africa has long had a reliable and cheap supply of far below the marginal cost of new generation, private electricity. However, delays in investment by the state-owned investors had no way of entering the sector without spe- electricity provider Eskom (which provides 70 percent of cial contracting arrangements. After a four-year hiatus, the the electricity in Sub-Saharan Africa), breakdowns of power government abandoned the idea of a competitive market plants, and negligence in coal contracting have eroded spare and again charged Eskom with expanding capacity (while capacity in the system, leaving the country prone to periodic retaining the option of contracting with a few independent rounds of rolling power cuts. Many of South Africa's neigh- power producers in the future). These planning and invest- bors, dependent on imports, are also feeling the economic ment failures are typical of hybrid electricity markets. costs of power scarcities. To help finance investment and reduce demand, electric- The government had earlier imposed a moratorium on ity prices in South Africa will increase substantially over the Eskom's building new plants. It considered unbundling the next several years. But the supply-demand balance will likely utility and introducing private participation and competi- remain tight for at least the next seven years, up to 2015, tion in the market, similar to Nord Pool in Scandinavia or until new base-load generation capacity comes on line. PJM in the United States. But the new market arrange- Source: Based on interviews with World Bank staff from the Africa ments were never implemented, and with average prices Energy Department, 2008. 184 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION of turnover for informal enterprises unable to the private provider. At least an estimated to provide their own backstop generation 750 megawatts of emergency generation is (Foster and Steinbuks 2008). Therefore, many currently operating in Sub-Saharan Africa, enterprises invest in backup generators. In representing for some countries a large pro- many countries, backup generators repre- portion of their national installed capacity. sent a significant proportion of total installed Because of the preponderance of small diesel power capacity: 50 percent in the Democratic units, the costs have typically been $0.20­$0.30 Republic of Congo, Equatorial Guinea, and per kilowatt-hour, and for some countries, the Mauritania, and 17 percent in West Africa as price tag can be 4 percent of GDP (table 8.1). a whole. The cost of backup generation can easily run to $0.40 per kilowatt-hour or several times higher than the utility's costs of generat- Figure 8.2 Economic Cost of Outages in Selected ing power (Foster and Steinbuks 2008). Countries The economic costs of power outages are substantial. The immediate economic cost of Malawi power shortages can be gauged by looking at South Africa the cost of running backup generators and Uganda forgoing production during power shortages. Tanzania These costs typically range between 1 and 4 Kenya percent of GDP (figure 8.2). Over time, the lack of a reliable power supply is also a drag on eco- Senegal nomic growth. From the early 1990s to the early Cameroon 2000s in Cameroon, Côte d'Ivoire, the Demo- Madagascar cratic Republic of Congo, Ghana, and Senegal, Cape Verde inadequate power infrastructure shaved at least Niger one-quarter of a percentage point off annual Burkina Faso per capita GDP growth rates (Calderón 2008). A common response to the immediate cri- Benin sis is to tender short-term leases for emergency 0 1 2 3 4 5 6 7 power. Unlike traditional power generation % of GDP projects, this capacity can be put in place in a Source: Eberhard and others 2008, using World Bank 2007 data. few weeks, providing a rapid response to press- Note: Economic cost is estimated as the value of lost load ing shortages. Equipment is leased for up to multiplied by the volume of load shedding. Value of lost load is derived from country-specific estimates based on enterprise two years, sometimes longer, and then reverts survey data for sales lost due to power outages. Table 8.1 Economic Cost of Emergency Power Generation Emergency Total Emergency generation capacity generation capacity generation Cost of emergency Country (megawatts) (megawatts) capacity (% of total) generation (% of GDP) Angola 150 830 18.1 1.04 Gabon 14 414 3.4 0.45 Ghana 80 1,490 5.4 1.90 Kenya 100 1,211 8.3 1.45 Madagascar 50 140 35.7 2.79 Rwanda 15 31 48.4 1.84 Senegal 40 243 16.5 1.37 Sierra Leone 20 15 133.3 4.25 Tanzania 40 881 4.5 0.96 Uganda 100 240 41.7 3.29 Source: Eberhard and others 2008. Power: Catching Up 185 A Huge Investment Backlog presented earlier are based on growth pro- jections before the onset of the 2008 global Addressing Africa's chronic power problems financial crisis. The International Monetary will require major investments in the refur- Fund reduced its GDP growth projections bishment and expansion of power infra- for Africa from 5.1 percent a year to 3.5 percent structure. Of the 70.5 gigawatts of installed a year because of the global economic crisis. generation capacity, some 44.3 gigawatts need Sensitivity analysis suggests that even low- to be refurbished. An additional 7,000 mega- ering the original projected growth rates of watts of new generation capacity need to be 5.1 percent to half their levels would reduce built each year to meet suppressed demand, estimated power sector spending needs by keep pace with projected economic growth, only about 20 percent in absolute terms, and provide additional capacity to support the lowering required new generation capacity rollout of electrification. Compare that with from just over 7,000 megawatts to just under expansion of less than 1,000 megawatts a year 6,000 megawatts. The decrease in required over the period 1990­2005. The bulk of this spending would be somewhat larger in the new power generation capacity will be needed Southern and West African Power Pools and to meet nonresidential demands. In addi- somewhat smaller in the Central and East tion, raising electrification rates will require African Power Pools. Even so, when power extending distribution networks to reach an spending needs are expressed as a percentage additional 6 million households a year from of GDP, the effect of a slower-growth sce- 1996 to 2005. nario is much smaller. Because slower growth The total spending needs of the power reduces GDP as well as power spending needs, sector amount to $40.6 billion a year (Rosnes the overall economic burden of power sec- and Vennemo 2008), or 6.4 percent of the tor spending needs is only very slightly lower region's GDP, skewed toward capital expen- under a low-growth scenario. diture (table 8.2). The greatest absolute Existing spending on the power sector is spending requirements correspond to the $11.6 billion, or just over one-quarter of what middle-income countries, which need to is required. The adoption of high-cost genera- spend $14.2 billion a year, but the largest eco- tion solutions skews existing spending toward nomic burden is borne by the fragile states, operating expenditure, leaving only $4.6 bil- which would have to devote an implausible lion a year to fund the long-term investments 13.5 percent of GDP to meet this goal. to address the continent's power supply crisis, Economic growth is an important driver more than half of which comes from domes- of demand for power generation capacity. tic public finance. Existing spending repre- The estimates of power investment needs sents 1.8 percent of regional GDP, although Table 8.2 Power Sector Spending Needs $ billions annually Percentage of GDP Operation Operation Capital and Total Capital and Total Country type expenditure maintenance spending expenditure maintenance spending Sub-Saharan Africa 26.60 14.00 40.60 4.20 2.20 6.40 Middle-income countries 6.29 7.90 14.19 2.30 2.92 5.22 Low-income fragile countries 4.50 0.70 5.20 11.70 1.80 13.50 Low-income nonfragile countries 7.60 2.20 9.70 6.90 2.00 8.80 Resource-rich countries 8.40 3.35 11.77 3.79 1.50 5.29 Source: Briceño-Garmendia, Smits, and Foster 2008. Note: For a more detailed exposition of power sector spending needs, see chapter 2 in this volume. Totals may not add exactly because of rounding errors. 186 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION in the nonfragile low-income countries, this (IPPs). In recent years, 34 IPP contracts share increases to 2.9 percent of GDP. Of the in Africa have involved investments of external capital flows, finance from coun- $2.4 billion for the construction of 3,000 tries not belonging to the Organisation for megawatts of new power generation capacity. Economic Co-operation and Development Those projects have provided much-needed (OECD) is the most significant, accounting for generation capacity. An independent assess- $1.1 billion a year, primarily from the Export- ment concluded that they have also been Import Bank of China. Official development relatively costly because of technology choices, assistance follows at $0.7 billion a year and procurement problems, and currency devalu- then private capital flows of $0.5 billion a ations (calling for adjustments in dollar- or year (table 8.3). euro-denominated off-take agreements) Most of the private sector finance recorded (Gratwick and Eberhard 2008). relates to independent power producers The existing resource envelope would go significantly further if the sector oper- ated more efficiently. Addressing the operat- Table 8.3 Financing Flows to the Power Sector ing inefficiencies of the power utilities could $ billions annually reduce the funding gap by $3.3 billion a year, Operation improving cost recovery would bring an addi- and maintenance Capital spending tional $2.2 billion a year, and $0.3 billion a year could be recouped by improving execution of Public Public Non-OECD Total Country type sector sector ODA financiers PPI Total spending the capital budget. Sub-Saharan Even if all these inefficiencies could be Africa 7.00 2.40 0.70 1.10 0.50 4.60 11.60 eliminated, a sizable power sector financing Middle-income gap of $23 billion a year would remain (table countries 2.66 0.80 0.03 0 0.01 0.80 3.50 8.4). Three-quarters of this financing gap is Low-income a shortfall in capital expenditure, while the fragile countries 0.60 0 0.04 0.20 0.01 0.30 0.80 remaining quarter is a shortfall in operation Low-income and maintenance spending. The largest por- nonfragile countries 2.00 0.40 0.60 0.10 0.20 1.30 3.20 tion of the gap--nearly $11 billion per year-- Resource-rich corresponds to the middle-income countries. countries 1.60 1.20 0.10 0.70 0.30 2.30 3.90 However, the largest financing burden relates Source: Briceño-Garmendia, Smits, and Foster 2008. to the low-income fragile states, where the Note: Operation and maintenance includes other current expenditures. ODA = official development financing gap amounts to roughly 7 percent of assistance; OECD = Organisation for Economic Co-operation and Development; PPI = private participation in infrastructure. Totals may not add exactly because of rounding errors. their GDP. Table 8.4 Composition of Power Sector Funding Gap $ billions annually Percentage of GDP Operation Operation Capital and Capital and expenditure maintenance expenditure maintenance Country type gap gap Total gap gap gap Total gap Sub-Saharan Africa 17.6 5.6 23.2 2.7 0.9 3.6 Low-income fragile countries 2.6 0.1 2.8 6.9 0.2 7.1 Low-income nonfragile countries 4.5 0.1 4.7 4.1 0.1 4.2 Middle-income countries 5.5 5.2 10.7 2.0 1.9 3.9 Resource-rich countries 3.5 1.0 4.5 1.6 0.5 2.0 Sources: Briceño-Garmendia, Smits, and Foster 2008; Yepes, Pierce, and Foster 2008. Note: Totals do not add because efficiency gains cannot be carried across country groups. Power: Catching Up 187 The Promise of Regional importing more than 50 percent of their power Power Trade needs through regional trade. Savings range from $0.01 to $0.07 per kilowatt-hour. The Although Sub-Saharan Africa is well endowed largest beneficiaries tend to be smaller nations with both hydropower and thermal resources, without domestic hydropower resources. For only a small fraction of its power generation those countries, the cost of building cross- potential has been developed. Of the 48 Sub- border transmission would be paid back in Saharan countries, 21 have a generation capacity less than one year, once neighboring countries of less than 200 megawatts, well below the mini- have developed adequate generation capacity to mum efficiency scale, which means they pay a support trade. (For a more detailed analysis of heavy penalty: costs reach $0.25 per kilowatt- regional power trade potential, see chapter 6 in hour, twice the $0.13 per kilowatt-hour in the this volume on regional integration.) region's larger power systems. One reason is that some of the region's most cost-effective energy resources are too distant from major centers Improving Utility Performance of demand in countries too poor to raise the through Institutional Reform billions of dollars needed to develop them. For example, 61 percent of the region's hydropower The operational inefficiencies of power utilities potential is in just two countries: the Demo- cost the region $2.7 billion a year (0.8 percent cratic Republic of Congo and Ethiopia. of GDP on average; figure 8.3). They divide Pooling energy resources through regional roughly evenly between distribution losses power trade promises to reduce power costs. and revenue undercollections. Average distri- The Southern, West, East, and Central African bution losses in Africa are 23.3 percent, more Power Pools, created mainly to support power than twice the norm of 10 percent, affecting trade efforts, are at varying stages of maturity. all countries to some degree. Average collec- If pursued to their full economic potential, tion ratios are 88.4 percent, compared with regional trade could reduce the annual costs of the best practice of 100 percent. Burkina Faso, power system operation and development by Ghana, Niger, and Uganda face much greater $2 billion per year (about 5 percent of total undercollections than the rest, up to 1 percent power system costs). These savings are already of GDP. incorporated in the power sector spending needs Operational inefficiencies have been hold- previously presented. They come largely from ing back the pace of electrification and pre- substituting hydropower for thermal power, venting utilities from balancing supply and substantially reducing operating costs, even demand. They drain the public purse and though it entails higher up-front investment undermine the performance of the utilities. in capital-intensive hydropower and associ- One casualty of insufficient revenue is main- ated cross-border transmission. The returns tenance. Utility managers must often choose to cross-border transmission can be as high among paying salaries, buying fuel, or pur- as 120 percent for the Southern African Power chasing spares. They must frequently cannibal- Pool and more typically 20­30 percent for the ize parts from other working equipment. The other power pools. By increasing the share of investment program is another major casu- hydropower, regional trade would also save alty. Utilities with below-average efficiency 70 million tons of carbon emissions a year. electrify only 0.8 percent of the population Under regional power trade, a handful of in their service area each year, much lower large exporting countries would serve a sub- than the 1.4 percent electrified each year by stantial number of power importers. The utilities with above-average efficiency. Utilities Democratic Republic of Congo, Ethiopia, and with low efficiency also have greater difficulty Guinea would emerge as the major hydropower in keeping pace with demand. The suppressed exporters. As many as 16 countries would be or unmet power demand in those countries better-off (from a purely economic standpoint), exceeds 13 percent of total demand, twice the 188 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 8.3 Economic Burden Associated with Power 6 percent in countries with higher efficiency Utility Inefficiencies in Selected Countries (figure 8.4). Institutional reform measures hold the key South Africa to improving utility performance. Countries Zambia that have advanced the institutional reform Benin agenda for the power sector show substantially Chad lower hidden costs than those that have not, Kenya as do countries with more developed power Mozambique regulatory frameworks and better governance Ethiopia of their state-owned utilities (figure 8.5). Mea- Lesotho sures that seem to have a substantial effect on reducing hidden costs are private participation Rwanda in the power distribution sector and (among Madagascar state-owned utilities) performance contracts Burkina Faso that incorporate clear incentives. The case of Tanzania Kenya Power and Lighting Company is par- Senegal ticularly striking (box 8.2). Nigeria Labor redundancy is another source of Cameroon utility inefficiencies. Power utilities in Africa Malawi have overemployment of 88 percent relative to a developing country benchmark of 413 Cape Verde connections per employee. Overemployment Niger by utilities results in labor overspending Uganda in the range of 0.07 percent to 0.6 percent Ghana of GDP. Congo, Dem. Rep. of The application of management contracts 0 1 2 3 4 5 has been more complex than originally sup- % of GDP posed. More than 20 African countries have unaccounted losses undercollection experimented with private sector participation in power distribution, split evenly between Source: Briceño-Garmendia, Smits, and Foster 2008. concessions and management contracts. Man- Note: Power utility inefficiencies include undercollection of rev- enues and unaccounted-for distribution losses. agement contracts have attracted interest Figure 8.4 Inefficiency in Utility Performance a. Effect on pace of electrification b. Effect on magnitude of suppressed demand 1.6 16 suppressed demand as % of generation annual increase in access to power, % 1.4 14 1.2 12 1.0 10 0.8 8 0.6 6 0.4 4 0.2 2 0 0 high efficiency low efficiency high efficiency low efficiency Source: Derived from Eberhard and others 2008. Note: High efficiency refers to utilities with below-average levels of inefficiency caused by revenue undercollection and distribution losses. Low efficiency refers to utilities with above-average levels of inefficiency caused by revenue undercollection and distribution losses. Power: Catching Up 189 because they are a simpler way of addressing Figure 8.5 Effect of Reform Measures on Hidden Costs inefficiencies, but their application has proved complex and contentious, and they have not high reform always proved sustainable. Of 17 African man- agement contracts, 4 were canceled before the high regulation originally designated expiry date, and at least 5 more were not renewed after their initial term, high governance reverting to state operation. Only 3 manage- ment contracts remain in place. management contract or concession Problems with management contracts have included unrealistic expectations and limited performance contracts with incentives ability to address broader sector challenges. First, many management contracts were under- 0 50 100 150 200 250 300 350 400 average hidden cost of inefficiencies as a percentage of utility revenue taken with donor involvement. Donors saw the no yes contracts as an initial step on the road to more extensive sector reform that would be extended Source: Eberhard and others 2008. BOX 8.2 Kenya's Success with Private Participation in Power Kenya's Electric Power Act of 1997 introduced independent $0.07 in 2000 to $0.15 in 2006 and to $0.20 in 2008. As economic regulation, essential for private sector participa- a result of those measures, the hidden costs of the power tion. It has since become government policy to put all bids sector fell to 0.4 percent of GDP in 2006 and were elimi- for generation facilities out for competition, open to both nated by 2008 (see figure). This outcome put the sector on public and private firms, and to give no preferential treat- a firmer financial footing and has saved the economy more ment to the national generator. than 1 percent of GDP. The sector was unbundled in 1998 when Kenya Electric- ity Generating Company (KenGen; generation) and Kenya KPLC'S Success in Driving Down Hidden Costs, 2001­08 Power and Lighting Company (KPLC; transmission and dis- tribution) were established. KenGen is now 30 percent pri- 100 2.0 vately owned, and KPLC is 51 percent privately owned. Established in 1998, the Electricity Regulatory Board (the 75 1.5 Energy Regulatory Commission since 2007) maintains a sig- % of revenues nificant degree of autonomy. It has issued a grid code and % of GDP rules on complaints and disputes, supply rules, licenses, a 50 1.0 safety code, and a tariff policy. Four independent power producers supply about 12 per- cent of all power. Four more recently received licenses, and 25 0.5 another three are expected to apply for licenses. In the early 2000s, KPLC had substantial hidden costs in underpricing, collection losses, and distribution losses that 0 0 absorbed 1.4 percent of GDP. In the run-up to a manage- 2001 2002 2003 2004 2006 2008 ment contract, revenue collection improved from 81 per- distribution losses undercollections cent in 2004 to 100 percent in 2006. Distribution losses underpricing total as % of GDP also began to fall, though more gradually, reflecting the greater difficulty in resolving them. Power pricing reforms Source: Interviews with World Bank staff from the Africa Energy also allowed tariffs to rise in line with escalating costs, from Department, 2008. 190 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION long enough to allow parallel policy and Most utilities score better on internal gover- institutional changes to be enacted and to take nance criteria, such as board structure and root. In contrast, many African governments accountability, than on external governance saw them as costly reform measures needed to criteria, such as outsourcing and labor and secure donor finance and had no intention of capital market disciplines. taking the process any further. Second, although The acute need to improve the management management contracts can produce financial of utilities and the frameworks they operate and efficiency gains, they cannot overcome under has long been acknowledged. Over the broader policy and institutional weaknesses. years, substantial sums have been spent on Moreover, the efficiency gains do not always institutional reforms: training management, provide tangible improvements for customers, improving internal accounting and external even though they impose substantial adjust- auditing, strengthening boards of directors, ment costs on management, making political providing financial and operational informa- support for these measures hard to build. tion, building reporting systems, creating and Most African power utilities remain state reinforcing supervisory and regulatory agen- owned and operated. On average, Africa's cies, and much more. Some enduring suc- state-owned power utilities embody only cesses have been registered (box 8.3; further 40 percent of good governance practices for discussion of institutional issues can be found such enterprises (Vagliasindi and Nellis 2009). in chapter 4 of this volume). BOX 8.3 Botswana's Success with a State-Owned Power Utility The Botswana Power Corporation (BPC) is a BPC constantly weighs its options of government-owned monopoly that produces, importing against expanding its own gen- transmits, and distributes electrical power eration facilities, taking into account both in Botswana. It was formed by government economic and strategic factors. The national decree in 1970 with the objective of expand- system provides 132 megawatts, with the ing and developing electrical power poten- remaining 266 megawatts supplied by neigh- tial in the country. From its small beginnings boring countries through the Southern African with one power station in Gaborone and a Power Pool. Since the pool's inception in 1995, network that extended some 45 kilometers Botswana has been a major beneficiary, and its outside the city, the power utility's responsi- active trading position promoted multilateral bilities, along with the national network, have agreements among pool members, generally expanded enormously. The government has enhancing regional power cooperation. a regulatory role through the Energy Affairs Part of BPC's strong performance is thanks to Division of the Ministry of Minerals, Energy, cheap imported power from South Africa (now and Water Affairs. severely threatened by the power crisis). But BPC increased access to power to 22 percent analysts give institutional factors equal weight: a in 2006 and is set to reach 70 percent in 2009 strong, stable economy; cost-reflective tariffs; and 100 percent by 2016. Through govern- lack of government interference in managerial ment funding, BPC is extending the electricity decisions; good internal governance; and com- grid into rural areas and developing the reach petent, well-motivated staff and management. of the national transmission grid. Overall, the (For a more detailed discussion of institutional power system operates efficiently, with sys- reforms, see chapter 4 in this volume.) tem losses of no more than 10 percent and a decent return on assets. Sources: Molefhi and Grobler 2006; PPA 2005. Power: Catching Up 191 The Challenge of Cost Recovery where large strategic customers have purchased power at heavily discounted rates of just a few Underpricing power costs the sector at least cents per kilowatt-hour. These arrangements $2.2 billion a year in forgone revenues (0.9 per- were initially justified as locking in base-load cent of GDP on average). Underpricing power demand to support very large power proj- is widespread across Africa. In the worst cases ects that exceeded the country's immedi- (Malawi, Tanzania, and Zambia), underpricing ate demands, but they are now questionable can result in utilities' capturing less than half because competing demands have grown to of the revenues they need and creating an eco- absorb this capacity. nomic burden in excess of 2 percent of GDP Power prices have risen substantially in (figure 8.6). recent years, but they have nonetheless failed to These figures probably understate the keep pace with escalating costs. Because of ris- underpricing because of the difficulty of cap- ing oil prices, lower availability of hydropower, turing subsidies to large industrial and mining and greater reliance on emergency leases, the customers, which are usually contained in costs of power production in Africa rose sub- bilateral contracts and not reflected in the stantially in the early to mid-2000s (figure 8.7, general tariff structure. Key examples include panel a). In response, several countries have the aluminum-smelting sector in Cameroon increased power tariffs, so that the average and Ghana and the mining sector in Zambia, revenue of power utilities almost doubled over the same period (figure 8.7, panel b). Even so, because of historic pricing shortfalls, overall Figure 8.6 Underpricing of Power in Selected average revenues by the end of this period had Countries barely caught up with average operating costs at the beginning of the period. Uganda Most countries are achieving no more South Africa than operating-cost recovery. The correlation Ghana between average revenues and average oper- Congo, Dem. Rep. of ating costs across Sub-Saharan countries is as high as 90 percent, indicating that operating- Burkina Faso cost recovery is the driving principle behind Benin power pricing in most cases. Cameroon, Mozambique Cape Verde, Chad, Malawi, Niger, Rwanda, Kenya Senegal, and Tanzania (countries under the Nigeria 45-degree line in figure 8.8, panel a) fail to Cape Verde meet even operating-cost recovery, and sev- Chad eral of them face particularly high operating Lesotho costs (figure 8.8). Madagascar The longer-term cost-recovery situation is somewhat more hopeful. Comparing existing Senegal average revenues and average operating costs Ethiopia misrepresents long-term cost recovery for two Rwanda reasons. First, because of major inefficiencies Cameroon in revenue collection, the average revenue col- Niger lected per unit of electricity sold is substantially Tanzania lower than the average effective tariff charged Zambia today. Second, because of the major inefficien- Malawi cies in generation technology and the potential for regional trade, for more than two-thirds of 0 1 2 3 4 the countries the average incremental cost of % of GDP power looking forward is lower than the aver- Source: Briceño-Garmendia, Smits, and Foster 2008. age historical cost of power production looking 192 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 8.7 Electricity Costs and Revenues by Type of Power System, 2001­05 US$ per kilowatt-hour a. Average operating cost b. Average revenue 0.30 0.30 0.25 0.25 0.20 0.20 0.15 0.15 0.10 0.10 0.05 0.05 0 0 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 overall predominantly diesel predominantly hydropower Source: Eberhard and others 2008. Figure 8.8 Past and Future Cost-Recovery Situation US$ per kilowatt-hour a. Average revenues against average operating costs b. Average effective tariff against average incremental cost 0.8 0.4 0.7 average effective tariff 0.6 average revenue 0.3 0.5 0.4 0.2 0.3 0.2 0.1 0.1 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4 average operating cost average incremental cost Source: Eberhard and others 2008. backward and including both historic operat- In most cases, the state or donors have almost ing and capital costs. entirely subsidized the historic capital costs of A truer picture of long-term cost recovery power development. Although the residential comes from comparing today's average effec- sector accounts for 95 percent of power utility tive tariff with the average incremental cost customers in Africa, it contributes only around looking forward (figure 8.8). At least in some 50 percent of sales revenue. Thus, the pricing of countries, even the current tariff would be power to commercial and industrial consumers adequate for cost recovery, if only all revenues is just as important for cost recovery. Neither could be collected and the power system could commercial nor residential customers are close move toward a more efficient production to paying full cost-recovery prices. structure. In other countries, however, signifi- Subsidies to residential consumers are cant tariff adjustments would still be needed in highly regressive. Across the bottom half of the long term. the income distribution, barely 10 percent of Power: Catching Up 193 households have access to electricity (Wodon available (table 8.5). Such modest bills would 2008). Indeed, three-quarters of the house- be affordable to all but the poorest 25 percent holds with electricity come from the top of the population. In eastern and western two quintiles of the income distribution. Africa, the subsistence monthly bill would Because poorer households are almost entirely fall in the $7.00­$9.00 range. Although this excluded, they cannot benefit from subsidies amount would likely be affordable for exist- embedded in electricity prices. In many cases, ing customers, it would represent a problem targeting performance is further exacerbated as power access is expanded to lower-income by poor tariff design, with the widespread use populations. When a more efficient power sys- of increasing block tariffs that provide large tem develops, full cost-recovery tariffs would lifeline blocks of highly subsidized power to be affordable for the vast majority, except per- all consumers. haps in West Africa. With subsistence consumption of 50 If regional trade is pursued, the average costs kilowatt-hours a month, the cost of a monthly of power production could be expected to fall utility bill priced to recover full historic costs toward $0.07 in central and southern Africa, of production would be as much as $24.30 in $0.12 in eastern Africa, and $0.18 in western central Africa, which is manifestly unafford- Africa. Assuming, again, subsistence consump- able for the vast majority of the population tion of 50 kilowatt-hours a month, a monthly (table 8.5). Elsewhere in Africa, a subsistence utility bill under full cost-recovery pricing would monthly bill priced at full historic cost would be about $4 a month in central and southern range between $7.00 and $10.70 and would be Africa, $6 a month in eastern Africa, and $9 a affordable to the relatively affluent sections month in western Africa. Based on an affordabil- of the population that already enjoy access to ity threshold of 3 percent of household income, power, but not to the poorer segments of the full cost-recovery tariffs would prove affordable population that remain unconnected. Indeed, for the vast majority of the population of low- affordability of cost-recovery power bills for income countries in central, eastern, and south- existing customers is today really only a prob- ern Africa (see figure 8.9). In West Africa, about lem in low-income countries reliant on small- half the population of the low-income countries scale, oil-based generation. would face affordability problems. A number of Looking into the future, pricing at the lower West African countries--notably Côte d'Ivoire, long-run marginal cost of power would reduce the subsistence monthly bill to the $3.00­$4.00 range in central and southern Africa where Figure 8.9 Affordability of Subsistence Consumption of Power at Cost-Recovery abundant low-cost hydropower would become Pricing than 5% of their monthly budgets % of households spending more 1.2 1 Table 8.5 Cost and Affordability of Monthly Power Bills at Cost-Recovery Prices: Past and Future 0.8 $ per month 0.6 Long-run Location Historic cost marginal cost 0.4 Central African Power Pool 24.30 3.50 0.2 East African Power Pool 9.50 7.00 0 2 4 6 8 10 12 14 16 18 Southern Africa Power Pool 7.00 3.00 US$/month West Africa Power Pool 10.70 9.00 cost-recovery bill, low bound cost-recovery bill, medium bound Source: Derived from Rosnes and Vennemo 2008. cost-recovery bill, high bound low-income countries Note: Dark gray shading: power bill is unaffordable to the vast middle-income countries majority of the population; light gray shading: power bill is affordable to existing customers only, who are typically the richest 25 percent of the income distribution; no shading: Source: Banerjee and others 2008. power bill is affordable to all but the poorest 25 percent of the Note: A power bill for subsistence consumption of 50 kilowatt-hours per month is considered income distribution. affordable if it absorbs no more than 5 percent of household income. 194 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Ghana, Nigeria, and Senegal--already have are simply too small to support any meaning- power coverage of around 50 percent and would ful competition. The new reality is thus one of face affordability issues as coverage broadens. At "hybrid markets," with the state-owned utility any of these levels, power tariffs do not repre- remaining intact and occupying a dominant sent a significant affordability issue in the mid- market position. At the same time, because dle-income countries. (For a fuller discussion of many governments and utilities lack suffi- the social issues associated with utility pricing in cient investment resources, the private sector Africa, see chapter 3 in this volume.) participates, typically as IPPs. Africa's hybrid electricity markets pose new challenges in policy, regulation, planning, and procure- ment. The widespread power shortages across Policy Challenges the continent and the increasing reliance on emergency power indicate the seriousness of The depth and extent of Africa's power cri- those challenges. sis and its associated costs demand renewed Too often, the planning function has fallen efforts to tackle the policy and institutional between the cracks. Traditionally, planning and challenges needed to improve performance procurement of new power infrastructure were and financing. The key challenges can be char- the province of the state-owned utility. With acterized as follows: the advent of power sector reforms and the · Strengthening sector planning IPPs, those functions were often moved to the ministry of energy or electricity. A simultane- · Recommitting to the reform of SOEs ous transfer of skills did not always occur, how- · Increasing cost recovery ever, resulting in plans that were not adequately · Accelerating electrification informed by the complexities on the ground: a · Expanding regional trade in power new hybrid market of private and public actors. · Closing the financing gap. In many cases, planning has collapsed. Where still present, planning tends to take the form of These interdependent challenges must be outdated, rigid master plans. The lack of strate- dealt with simultaneously. Efforts to boost gic policy and planning for the electricity sector generation through regional power trade will at the central government level is a critical weak- stumble if the utilities, which will continue ness. Interventions have been piecemeal rather to be central actors, remain inefficient and than integrated; many countries have focused insolvent. Expanding electricity distribution on generation without investing in efficient systems without addressing the shortages in transmission and delivery of power. generation and improving transmission capac- This situation has led to very costly delays ity would clearly be futile. In addition, focusing in commissioning new plants. In the absence of exclusively on utility reform would be fruitless strong political leadership, good information, without a start on substantial, long-gestation and the requisite planning capability, incum- investments in both generation and access to bent state-owned utilities often undermine the improve the quality of service and make the entry of IPPs by arguing that they can supply utilities viable. In short, these strategic priori- power more cheaply or quickly than private ties must progress together. alternatives, even if they lack the resources to do so. Poor understanding of the hybrid market Strengthening Sector Planning deprives policy makers of clear and transpar- Most African power markets present an insti- ent criteria for allocating new plants between tutional "hybrid," with public and private the incumbent, state-owned utility and the actors operating in parallel. The 1990s reform IPPs. New plants are rarely ordered on a timely prescription of unbundling and privatization, basis, thereby opening power gaps that prompt leading to wholesale and retail competition, recourse to temporary power and discour- did not prove very relevant to Africa, not least age investors. When procurement is (finally) because most of the region's power systems undertaken, the authorities may not take the Power: Catching Up 195 trouble to conduct international competitive focused too heavily on technical issues to the bidding. This outcome is unfortunate, because exclusion of governance and accountability. a rigorous bidding process lends credibility Future SOE reforms seem justified as long and transparency to procurement and results as they focus on these deeper institutional in more competitively priced power. issues. Restoring and strengthening planning The starting point for SOE reform should capabilities are imperative. Hybrid power be to reform corporate governance. Key mea- markets will not disappear from the African sures include greater decision-making auton- landscape any time soon. To make the best of omy for the board of directors, more objective them, African governments and their develop- selection criteria for senior managers, and rig- ment partners must strive to develop a robust orous disclosure of conflicts of interest, as well institutional foundation for the single-buyer as more transparent and merit-based recruit- model, with clear criteria for power purchase ment processes. (off-take) agreements and dispatches of power Parallel efforts are needed to strengthen under those agreements. Governments must financial and operational monitoring of SOEs restore a strong sector planning capability at by their supervisory agencies, whether they are the line ministry level, establish clear policies line ministries or ministries of finance. Trans- and criteria for allocating new plant oppor- parency and accountability of SOEs depend on tunities between the state-owned utilities and solid financial management, procurement, and IPPs, and commit to competitive and timely management information systems. Today, basic bidding processes. A well-articulated plan for operational and financial data on firm perfor- the sector will allow governments to move mance are not produced, reported, or acted beyond the "firefighting" that has reduced their on. Without information or, perhaps worse, ability to anticipate exogenous shocks, such as without action based on whatever informa- drought or high oil prices. tion is produced, better outcomes cannot be Development partners need to tread care- expected. Key measures include auditing and fully in the hybrid marketplace. They can help publishing financial accounts and using com- by providing advice on transparent contract- prehensive cost-based accounting systems that ing frameworks and processes and by lend- allow functional unbundling of costs and a ing expertise to governments and utilities as clearer sense of cost centers. the latter seek to reach financial closure with In principle, regulation can be an important project sponsors and private investors. Lend- part of this process, but in practice, it proves ing to public utilities needs to be handled care- challenging to develop. Electricity regulators fully; if done without adequate attention to have been set up across Africa, precisely to insu- the peculiarities of the hybrid market, it may late utilities from political interference while have the unintended effect of deepening the closely monitoring enterprise performance. contradictions inherent in those markets and Some critics argue that regulatory agencies even crowding out private investment. What is have simply created additional risks because of needed above all is to strengthen public insti- their unpredictable decisions, resulting from tutions to enable them to engage effectively excessive discretion and overly broad objec- with the private sector. tives. Moreover, regulatory autonomy remains elusive; in some countries, turnover among Recommitting to the Reform of commissioners has been high, while the gap State-Owned Enterprises between law (or rule) and practice is often Renewed efforts on SOE reform should favor wide. The challenge of establishing new governance over technical fixes. State-owned public institutions in developing countries is utilities are still prevalent across Africa, and often underestimated. Independent regula- their performance is generally poorer than in tion requires a strong political commitment other regions. Fortunately, improving the gov- and competent institutions and people. Where ernance of SOEs can improve performance. some or all are lacking, it seems wise to consider Past efforts at improving utility management complementary or transitional options that 196 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION reduce discretion in regulatory decision making local management teams, each of which offers through more explicit rules and procedures, or advantages. Nonetheless, clarity about what outsourcing the regulatory functions to advisory they can and cannot achieve, particularly given regulators and expert panels (Eberhard 2007). their short time horizons, is important. At When this foundation is in place, con- best, a management contract can improve per- tracting mechanisms can be used to improve formance on a handful of manageable aspects performance. These mechanisms could be of efficiency, such as revenue collection and performance contracts in the public sector labor productivity. It cannot solve deficien- or management contracts with the private cies in the broader institutional framework, sector. which ideally should be addressed earlier. Nor Public sector performance contracts need can a management contract raise investment to incorporate strong performance incen- finance or significantly affect service quality tives. Initial attempts to improve African if substantial investments or long gestations SOEs through performance contracts with are required. the line ministry or other supervisory agency Utilities that have the institutional basics were minimally effective. Recent efforts in the in place would likely benefit from technical water sector (in Uganda, for example) have assistance (box 8.4). In particular, operational had a stronger and much more positive effect. efficiency programs are needed to reduce the The key feature of these contracts is to incor- high rates of technical, nontechnical (electric- porate incentives for good managerial (and ity theft), and collection losses. Such programs staff) performance and, more rarely, sanctions can include capacity building and technical for failure to reach targets. This approach to assistance to improve management, business more comprehensive performance contracts practices, and planning. The priorities are deserves further consideration. improved load management (to better match Creating effective performance incentives supply with priority customer needs), theft within a public sector context can be quite reduction initiatives, and increased revenue challenging. Management contracts with collection (through enhanced metering and the private sector are thus a relevant option. better-run customer service units). Capital They can be applied with either expatriate or spending can also be reduced by using low-cost BOX 8.4 CREST Spreading Good Practices The Commercial Reorientation of the Electric- highly reliable armored and aerial bunched ity Sector Toolkit (CREST) is an experiment cables on the low-tension consumer point to under way in several localities served by West reduce theft) and managerial changes (intro- African electricity providers. Based on good ducing "spot billing" and combining data practices from recent reforms in Indian, Euro- recording, data transfer, bill generation, and bill pean, and U.S. power corporations, CREST is distribution). Transaction times are reduced, and a "bottom-up" approach for attacking system cash flows improve. Early applications of CREST losses, low collection rates, and poor customer have reportedly produced positive changes in service. several neighborhoods in Guinea and Nigeria, To accomplish its objectives, CREST uses two difficult settings. technical means (replacing low-tension with Source: Based on interviews with World Bank staff high-tension lines, for example, and installing from the Africa Energy Department, 2008. Power: Catching Up 197 technology standards, as in Guinea and Mali. West Africa, where the costs of power will Innovations have included adjusting technical remain relatively high even with regional design standards to meet the reduced require- trade.) A case thus exists for these countries ments of low-load systems, maximizing the to start moving their tariffs toward longer- use of material provided by local communities term cost-recovery levels, accepting that the (such as locally sourced wooden poles), and sector will continue to register financial defi- recruiting employees and supervisors from the cits in the short term. local community. Cost recovery is particularly important Finally, institutional change is a long-term for emergency power leases, to avoid divert- matter, but well worth the wait. Victories on ing budgetary resources from long-term this front will be small and slow in coming. investments. Numerous African countries Donors may prefer the large and the quick, have responded to the power crisis by leasing but they must recognize that positive changes emergency power generation. This solution is in this field lie at the heart of African power rapid and effective but simultaneously costly sector reform. and temporary. Charges typically amount to $0.20­$0.30 per kilowatt-hour, without con- Increasing Cost Recovery sidering transmission and distribution costs The financial viability of incumbent utilities or associated losses. Given that the cost of is a key foundation of a healthy power sector. backup generation for the private sector is Financially viable utilities are more effective approximately $0.40 per kilowatt-hour, and operationally, because they are able to finance that the value of lost load may well be higher timely maintenance activities. They are also than that, the private sector should be will- more creditworthy and thus may begin to ing to pay the full cost of this emergency secure their own access to domestic or inter- power. Nevertheless, when emergency power national capital markets. Achieving this goal is provided without any adjustment to power demands power tariffs that are high enough to tariffs, the resulting fiscal drain can be very cover operating costs and to contribute as much large, diverting scarce resources from the as possible to covering capital costs as well. investments needed to provide a longer-term Cost recovery already looks feasible in coun- solution to the power problem. To avoid this tries with relatively low-cost domestic power fiscal drain, utilities must price emergency sources. In the continent's larger countries, power at cost-recovery levels for nonresiden- and in those that rely on hydropower and coal- tial customers. based generation, cost-recovery tariffs already Power subsidies will still be needed, but appear affordable for the majority of the popu- they should be well targeted and focus initially lation, and certainly for the affluent minority on expanding access. Existing power subsidies that enjoys access to power. A case thus exists are captured largely by higher-income groups for these countries seriously to consider mov- and do little to broaden access to electricity. ing closer to full cost recovery. Redesigning power subsidies would free scarce For countries with high-cost domestic fiscal resources that could be redirected to power, cost recovery may become feasible subsidize the expansion of power networks in the medium term as regional trade devel- to serve lower-income rural and periurban ops. In the continent's smaller countries, and communities, or for other poverty-alleviation those reliant primarily on oil-based genera- programs. In some of Africa's poorest coun- tion, cost-recovery tariffs are largely unaf- tries, even low-cost power will remain unaf- fordable. As regional trade develops and fordable for a significant minority of the access to more cost-effective sources of power population, so well-targeted subsidies would generation open up, however, the total cost be needed as part of the strategy for reaching of power production will fall, making cost universal access. What is clearly untenable, recovery a much more reasonable goal in the however, is the situation where power subsi- medium term. (The possible exception is dies that benefit only a privileged minority of 198 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION the population create a significant fiscal drag responsible for extending the grid, have been on the economy. more successful than those that followed decentralized approaches, where a rural elec- Accelerating Electrification trification agency attempted to recruit multiple From a social and political perspective, expand- utilities or private companies into the electrifi- ing access is imperative. Yet financing expan- cation campaign. Therefore, expecting special- sion to lower-income households will further ized agencies to solve the rural electrification strain the financial viability of the power sec- challenge on their own may be unrealistic. tor. Tackling this dilemma will require sig- They may be most productive in promoting nificantly higher concessional financing from minigrids and off-grid options as extensions of development partners for access programs, the national utility's efforts to extend the grid, as well as improved financial and operational as in Mali (box 8.5). performance from utilities. Rural electrification may need to follow Given the scale of investments needed, a urban electrification. In an African context, systematic approach to planning and financing one can legitimately ask how far rural electri- new investments is critical. The current ad hoc fication can progress when the urban electrifi- project-by-project approach in development cation process is still far from complete. Across partner financing has led to fragmented plan- countries, a strong correlation exists between ning, volatile and uncertain financial flows, urban and rural electrification rates, as well as and duplicated efforts. Engagement across a systematic lag between the two. Countries the sector in multiyear programs of access with seriously underdeveloped generation rollout, supported by multiple development capacity and tiny urban customer bases are partners as part of a coherent national strat- not well placed to tackle rural electrification, egy, will channel resources in a more sustained either technically because of power shortages and cost-effective way to the distribution or financially because of the lack of a basis for subsector. Coordinated action by develop- cross-subsidization. ment partners will also reduce the unit costs Finding ways of spreading the benefits of of increasing access, by achieving economies of electrification more widely is also impor- scale in implementation. tant. Because universal household electrifica- Completing the urban electrification pro- tion is still decades away in many countries, cess requires careful attention to the social sectorwide programmatic approaches need issues raised. Chapter 3 of this volume found to ensure that the benefits of electrification that approximately half of the nonelectrified touch the poorest households that may be too urban population lives in proximity to the far from the grid or just unable to pay for a grid. Densification is thus a key challenge. grid connection. Street lighting may be one Demand-side barriers, including high connec- way of doing that in urban areas. In rural tion charges and insecure tenure, need to be areas, solar-powered electrification of clinics addressed as part of this process. Expansion and schools that provide essential public ser- into periurban slums will need to face power vices to low-income communities is another theft, for which technical fixes are available way of allowing them to participate in the (see box 8.4). benefits of electrification. Another is appro- For rural electrification, emerging evidence priate technology, such as low-cost portable favors more centralized approaches (Mostert solar lanterns that are much more accessible 2008). Countries with dedicated rural electrifi- and affordable to the rural public. The Light- cation funds have higher rates of electrification ing Africa initiative is supporting the develop- than those without. Of greatest interest, how- ment of the market for such products. ever, are the differences among the countries with funds. Case studies indicate that the coun- Expanding Regional Trade tries that have taken a centralized approach A strategic priority is to tackle head-on the to electrification, with the national utility generation capacity deficit through major Power: Catching Up 199 BOX 8.5 Rural Electrification in Mali Among new African rural electrification agencies, AMA- minimum technical and commercial quality of service stan- DER (Agence Malienne pour le Développement de l'Énergie dards and maximum allowed tariffs for both metered and Domestique et l'Électrification Rurale, or Malian Agency for unmetered customers. the Development of Domestic Energy and Rural Electrifica- To ensure that the projects are financially sustainable, tion) has had considerable success. The starting point for AMADER permits operators to charge residential and com- AMADER is a country in which only about 3 percent of the mercial tariffs that are higher than the comparable tariffs rural population has access to electricity. Until they are con- charged to similar customers connected to the national nected, most rural households meet their lighting and small grid. For example, the energy charge for metered residential power needs with kerosene, dry cells, and car batteries, customers on isolated minigrids is about 50 percent higher averaging monthly household expenditure of $4­$10. than the comparable energy charge for grid-connected resi- Created by law in 2003, AMADER uses two major dential customers served by Énergie du Mali (the national approaches to rural electrification: (a) spontaneous, "bot- electric utility). Many of the minigrid operators also provide tom-up" electrification of specific communities and (b) planned, service to unmetered customers, who are usually billed a flat "top-down" electrification of large geographic areas. The monthly charge per lightbulb and outlet, combined with bottom-up approach, which typically consists of minigrids load-limiting devices to ensure that a customer does not managed by small local private operators, has been more connect lightbulbs and appliances beyond what he or she successful. By late 2008, about 41 bottom-up projects had has paid for. been financed, comprising 36,277 household connections Financing has been a problem for both AMADER and at an average cost per connection of $776. Typically, AMA- potential operators. AMADER has been hindered by insuf- DER provides grants for about 75 percent of the connection ficient and uncertain funding for providing capital cost capital costs. grants. Potential operators have had difficulty raising equity Because Mali has limited renewable resources, most of or obtaining loans for the 20­25 percent share of capital the minigrid systems are diesel fired. Customers on these costs not funded by AMADER. Promoting leasing arrange- isolated minigrids typically receive electricity for six to eight ments and instituting a loan guarantee program for Malian hours a day. In promoting these new projects, AMADER banks that would be willing to lend to potential operators performs three main functions: it acts as a (a) provider of have been discussed as methods of reducing financial barri- grants, (b) supplier of engineering and commercial techni- ers for operators. cal assistance, and (c) de facto regulator through its grant agreements with operators. The grant agreement can be Source: Interviews with World Bank staff from the Africa Energy viewed as a form of "regulation by contract" that establishes Department, 2008. regional projects. Africa's considerable off-take is needed, blending private partici- hydropower, gas, and coal resources remain pation and donor funding. underexploited. The best way to scale up gen- Power pool development must proceed in eration at the lowest unit cost is to develop parallel so that this new capacity can be trans- a new generation of large power generation mitted to users. Challenges common to all the projects. A substantial number of these trans- pools are rehabilitating and expanding the formational projects should be developed in cross-border transmission infrastructure to the near term to begin to make a material increase the potential for trade and harmonizing difference on the supply-demand balance. regulations and system operating agreements. However, individual countries do not have Equally important is formulating market trad- the necessary investment capital, or even ing mechanisms so that the additional energy the electricity demand, to move forward generated from large projects can be priced and with these large projects. A project finance thus allocated efficiently and fairly (for example, approach predicated on regional power through competitive pool arrangements). 200 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Although the economics of large regional Closing the Funding Gap generation projects are convincing, they may Africa's power funding gap is particularly give rise to significant political challenges. daunting, even more so in the global finan- Africa could potentially save $2 billion a year in cial crisis. At $23 billion, the funding gap in energy costs if trade were pursued to its fullest Africa's power sector is the largest of any infra- desirable extent, but the gains are much larger structure sector. The global financial crisis will for some countries than for others. Small ther- likely exacerbate the problem. As noted earlier, mal power­dependent countries and a hand- slower growth could reduce spending needs ful of major exporters are likely to benefit the by as much as 20 percent, but tighter global most. About one-third of African countries financial markets could similarly reduce avail- would end up importing more than half of able funding, widening the funding gap even their power needs, and self-sufficiency some- further. times has more political weight than access to Improving creditworthiness is an important low-cost power. first step that could eventually assist in access- Moreover, reaping the benefits of regional ing capital markets. The immediate subsidy power trade essentially depends on realiz- savings from addressing operational efficien- ing massive investments in three challenging cies and cost recovery, though substantial, do countries. The Democratic Republic of Congo, not come close to closing the gap. In principle, Ethiopia, and Guinea would be the major utilities achieving operational efficiency and power exporters under a regional trading sys- cost recovery (whether state owned or pri- tem. To become major exporters, however, all vately run) could aspire to raising their own three would need to invest massively in hydro- capital on domestic or international markets, power, which could easily absorb more than but that ability is still some way off. External 8 percent of their GDP for a decade. Even with finance to Africa's power sector had been very support from cross-border finance, the lim- low for some time but has picked up in recent ited financial capacity of these countries and years (figure 8.10). the numerous governance challenges faced by Official development assistance to public the fragile states (the Democratic Republic investment in power has risen substantially. of Congo and Guinea) make this quite a tall In response to the power crisis, donors have order. increased their emphasis on the power sec- These considerations call for an incremen- tor. Commitments averaged $1.5 billion a year tal approach to developing regional trade. The for 2005­07, reaching a peak of $2.3 billion initial emphasis needs to be on quick wins by in 2007. This is an important turnaround in building bilateral exchanges between neigh- funding, but more will be needed if any sub- bors where a particularly strong economic stantial inroads are to be made into Africa's case exists and where the political context is power sector challenges. supportive. This strategy will allow trading Non-OECD countries have emerged as experience to build up gradually, paving the major new power financiers in Africa (Foster way for adding more complexity over time. and others 2008). Commitments of non- Even if Africa's first-best generation options OECD countries, particularly the Chinese cannot always be developed or if the ultimate and Indian export-import banks, came from pattern of power production on the continent nowhere to average about $2 billion a year in turns out to be driven more by financial mus- 2005­07. Most of the Chinese financing has cle than by economic expansion, the benefits gone to 10 large hydropower projects with a of interconnection remain clear. Given the combined generating capacity of over 6,000 small scale and undiversified nature of most megawatts. Once completed, these projects countries' power systems, cross-border trans- will increase Sub-Saharan Africa's installed mission will always make sense as a means of hydropower capacity by 30 percent. China is boosting the efficiency and reliability of power also financing 2,500 megawatts of thermal production. power, and the Indian Bank has financed Power: Catching Up 201 significant thermal generation projects in Figure 8.10 External Financing Commitments for the African Power Sector, Nigeria and Sudan. 1994­2007 Private finance was also buoyant until 2007, 6,000 but significantly lower than official finance. Private commitments to Africa's power sector 5,000 averaged about $1 billion a year in 2005­07, putting it in third place behind non-OECD 4,000 US$ millions finance and traditional official development 3,000 assistance. The bulk of private resources has gone into 3,000 megawatts of independent 2,000 power projects. Although it will not be enough to close the financing gap, private finance is 1,000 very much needed. Successful private invest- ments in energy projects in Africa are still rare, 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 however, and increased private investment will official development assistance private participation in infrastructure not materialize simply because of large financ- non-OECD ing gaps. The lessons from past failures must be addressed because private investment will Source: Briceño-Garmendia, Smits, and Foster 2008. flow only where rewards demonstrably out- Note: OECD = Organisation for Economic Co-operation and Development. weigh risks. Some early but encouraging signs indicate that scaling up generation capacity through large private sector­led projects is Notes starting to gather momentum. A prominent The authors of this chapter are Anton Eberhard, example is the privately owned 250-megawatt Vivien Foster, Cecilia Briceño-Garmendia, and Bujagali hydropower plant in Uganda, sup- Maria Shkaratan, who drew on background ported by World Bank Group guarantees and material and contributions from Daniel Camos- funded by a private consortium. Ambitious i-Daurella, Gabriel Goddard, Jaakko Hellaranta, regional projects undoubtedly present techni- Rob Mills, Fatimata Ouedraogo, Timo Reiss, cal, financing, and political risks and will con- Orvika Rosnes, Jevgenijs Steinbuks, Prasad tinue to be underpinned by substantial public V. S. N. Tallapragada, Maria Vagliasindi, Tjaarda sector and donor contributions. Storm P. Van Leeuwen, and Haakon Vennemo. 1. Isolated areas are more than 50 kilometers from Shorter-term measures on energy effi- a substation and are either in the power plant ciency can ease the transition. Most of the buffer (within 10 kilometers for capacity below measures described here are medium term 10 megawatts, 20 kilometers for capacity below and cannot be implemented overnight. Many 100 megawatts, and 50 kilometers for capacity Sub-Saharan countries will continue to face below 100 megawatts) or within 10 kilometers a very tight demand-supply balance in the of a lit urban area or lit pixel. Remote areas coming years. Therefore, shorter-term mea- are more than 50 kilometers from a substation sures to soften the economic and social effects and are not in the power plant buffer or within of power scarcity must complement longer- 10 kilometers of a lit urban area or lit pixel. term efforts at addressing the underlying 2. These costs are calculated at the consumption structural causes of the power supply crisis. level of 100 kilowatt-hours a month. Recent experiences from countries such as Brazil show that well-designed demand-side References management measures (for example, a quota Banerjee, Sudeshna, Quentin Wodon, Amadou system with price signals, combined with a Diallo, Taras Pushak, Hellal Uddin, Clarence public energy-efficiency campaign) can go Tsimpo, and Vivien Foster. 2008. "Access, Afford- ability and Alternatives: Modern Infrastructure a long way toward trimming peak demand, Services in Africa." Background Paper 2, Africa substantially reducing power rationing at Infrastructure Country Diagnostic, World Bank, fairly low economic and social cost. Washington, DC. 202 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Briceño-Garmendia, Cecilia. 2008. "Quasi-Fiscal Gratwick, Katharine Nawaal, and Anton Eberhard. Costs: A Never Ending Concern." Internal note, 2008. "An Analysis of Independent Power Africa Infrastructure Country Diagnostic, World Projects in Africa: Understanding Development Bank, Washington, DC. and Investment Outcomes." Development Policy Briceño-Garmendia, Cecilia, Karlis Smits, and Review 26 (3): 309­38. Vivien Foster. 2008. "Financing Public Infra- Molefhi, Benjamin O. C., and L. J. Grobler. 2006. structure in Sub-Saharan Africa: Patterns, Issues, "Demand-Side Management: A Challenge and and Options." Background Paper 15, Africa Opportunity for Botswana Electric Energy Infrastructure Country Diagnostic, World Bank, Sector." North-West University, Potchefstroom, Washington, DC. South Africa. Calderón, César. 2008. "Infrastructure and Mostert, Wolfgang. 2008. Review of Experience with Growth in Africa." Working Paper 3, Africa Rural Electrification Agencies: Lesson for Africa. Infrastructure Country Diagnostic, World Eschborn, Germany: EU Energy Initiative Part- Bank, Washington, DC. nership Dialogue Facility. Eberhard, Anton. 2007. "Matching Regulatory PPA (Power Planning Associates). 2005. "Towards Design to Country Circumstances: The Potential Growth and Poverty Reduction: Lessons from of Hybrid and Transitional Models." Gridlines, Private Participation in Infrastructure in Sub- Note 23 (May). Public-Private Infrastructure Saharan Africa--Case Study of Botswana Power Advisory Facility, World Bank, Washington, DC. Corporation." Summary report, World Bank, Eberhard, Anton, Vivien Foster, Cecilia Briceño- Washington, DC. Garmendia, Fatimata Ouedraogo, Daniel Rosnes, Orvika, and Haakon Vennemo. 2008. Camos, and Maria Shkaratan. 2008. "Under- "Powering Up: Costing Power Infrastructure powered: The State of the Power Sector in Investment Needs in Southern and Eastern Sub-Saharan Africa." Background Paper 6, Africa." Background Paper 5, Africa Infrastruc- Africa Infrastructure Country Diagnostic, ture Sector Diagnostic, World Bank, World Bank, Washington, DC. Washington, DC. EIA (Energy Information Administration). 2006. Vagliasindi, Maria, and John Nellis. 2009. "Evalu- International Energy Annual. Washington, DC: ating Africa's Experience with Institutional U.S. Department of Energy. http://www.eia.doe Reform for the Infrastructure Sectors." Working .gov/emeu/international/contents.html. Paper 23, Africa Infrastructure Country Diag- ESMAP (Energy Sector Management Assistance nostic, World Bank, Washington, DC. Program). 2007. "Technical and Economic Wodon, Quentin, ed. 2008. "Electricity Tariffs Assessment of Off-Grid, Mini-Grid and and the Poor: Case Studies from Sub-Saharan Grid Electrification Technologies." ESMAP Africa." Working Paper 11, Africa Infrastructure Technical Paper 121/07, World Bank, Country Diagnostic, World Bank, Washington, Washington, DC. DC. Foster, Vivien, William Butterfield, Chuan Chen, World Bank. 2005. African Development Indicators. and Nataliya Pushak. 2008. Building Bridges: Washington, DC: World Bank. China's Growing Role as Infrastructure Finan- cier for Africa. Trends and Policy Options no. 5. ------. 2007. Enterprise Surveys database. http:// Washington, DC: Public-Private Infrastructure www.enterprisesurveys.org/. Advisory Facility, World Bank. Yepes, Tito, Justin Pierce, and Vivien Foster. Foster, Vivien, and Jevgenijs Steinbuks. 2008. 2008. "Making Sense of Sub-Saharan Africa's "Paying the Price for Unreliable Power Supplies: Infrastructure Endowment: A Benchmarking In-House Generation of Electricity by Firms in Approach." Working Paper 1, Africa Infrastruc- Africa." Policy Research Working Paper 4913, ture Country Diagnostic, World Bank, World Bank, Washington, DC. Washington, DC. Chapter 9 Transport: More Than the Sum of Its Parts S o much variety exists among the trans- Increasing the scope for competition among port modes and their infrastructure, in carriers is a challenge confronting most trans- both policy and technical matters, that port modes. In most parts of the world, com- the following four chapters deal with them petition has involved extending the role of the separately. Those modes interact and intercon- private sector. Franchises and concessions per- nect in complex ways, however, raising com- mit this extension without loss of government mon issues. influence over activities regarded as strategic. Many movements of passengers and freight Bus transport and trucking are already pre- involve more than one mode of transport, with dominantly private. Many African railways are impediments at the point of interchange, often now concessioned, and the role of the private caused by corrupt customs administration, or sector is increasing in ports and air transport. restrictions on entry to the transport market. In However, the regulation of these markets often freight, Sub-Saharan Africa suffers particularly remains obstructive rather than constructive, from such blockages, which delay shipments, with tour de role traffic allocation and dispatch- increase costs, and hinder the development ing1 reducing efficiency and increasing costs. of logistics systems so critical in sophisticated A common reason for restricting competi- global markets. Because Africa has so many tion is the belief that the government can ensure landlocked countries, this problem is multina- provision of socially desirable services only tional and must be confronted on a regional or when they are supplied by a public enterprise or a corridor basis. strongly protected private enterprises. However, One of the main sources of restrictive entry this belief is largely fallacious, particularly if the is the perceived need to protect national carri- protected suppliers have no incentive to be effi- ers or the existing carriers. Although this pro- cient or no segment of the market is profitable tection has largely withered away in interna- enough to support unremunerative "social" ser- tional shipping and inter-African air transport, vices. Publicly owned airlines, railways, shipping it still prevails in road freight. companies, and bus companies have all fallen 203 204 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION short. Thus, much work remains, both nationally can increase their profits through predatory and regionally, in developing regulatory regimes practices, as in Zambia's treatment of copper that reconcile public and private interests. exports by the Democratic Republic of Congo (see chapter 11 in this volume). In East Africa, joint concessioning of railways is part of a Integrating Multimodal Transport donor­funded corridor, where border-crossing arrangements have been reformed. Some coun- Inland transport costs often seriously inhibit tries are now trying to develop coordinated foreign trade, which passes through a seaport corridor systems, as in the Ghana Gateway and or airport and then one or more land modes. Maputo corridors. For example, inland transport accounts for an Other transport modes may also be involved. estimated 40 percent of the total cost from the Inland waterway transport historically carried point of origin to the port of destination for primary products from landlocked countries but West African cocoa and coffee exports. is now in decline. The three major lakes in East The transport chains are no stronger than and Central Africa--Victoria, Tanganyika, and their weakest links, usually the interchanges. Malawi--used to be important in transit and The weaknesses are partly physical, with no intraregional trade. Particularly on Lake Victo- physical connection between the modes and ria, the lake services were part of the railway sys- no infrastructure for transshipment; partly tems linking the railheads at the inland ports of institutional, with responsibility for the inter- Bell (Uganda), Kisumu (Kenya), and Mwanza changes not falling clearly to one modal agency (Tanzania). The Kenya and Uganda lake opera- or the other; and partly operational, with the tions were concessioned together with the rail- government collecting taxes and duties or staff ways in those countries, whereas in Tanzania, the collecting bribes, slowing movements, and lake services were separated from the railways. pushing up costs. Only one service now operates on Lake Victo- The port-rail link is the first major weakness. ria, and some of the railway track leading to the Although rail transport is advantageous for ports is in poor repair, especially in Kenya. long-distance, time-insensitive commodities, it The story is similar in West and Central depends heavily on internationally traded traf- Africa, where the Congo basin has a navigable fic. To hold this traffic, it must be linked effi- network of 12,000 kilometers, covering nearly ciently to good port connections, but conflicts 4 million square kilometers in nine countries. between rail and port jurisdictions over rail seg- In principle, this waterway could be a very ments in port areas often inhibit this link. Except valuable resource in a multimodal transport in South Africa, inland transport and facilities network serving the region, particularly given are poorly aligned with port development. The low associated transport costs of $0.05 per ton- stripping and stuffing of containers in port kilometer versus $0.15 per ton-kilometer for areas also increases congestion in many ports. It road or rail freight in Central Africa, albeit at is no accident that some of the most successful significantly lower speeds. In practice, however, rail lines in Africa operate in national corridors the river corridor suffers from an outdated and where specialized rail and port facilities are ver- insufficient infrastructure, inadequate chan- tically integrated (for example, Spoornet coal nel markings and maintenance, feeble regula- and ore lines and Gabonese manganese ore). tion, and numerous nonphysical barriers to Good links between complementary rail movement. Thus, it is ever more marginal in systems are also essential. Some railway orga- transport. Recognizing this untapped poten- nizations already contribute to this. The bina- tial, in October 2005, the Executive Secretary tional railways in Côte d'Ivoire­Burkina Faso of the Economic and Monetary Community of and Senegal-Mali provide good examples, as Central Africa encouraged the governments of does the operation of contiguous railways by Cameroon, the Central African Republic, the the same concessionaire (Central East African Democratic Republic of Congo, and the Repub- Railways in Mozambique and Malawi). How- lic of Congo to establish the Commission Inter- ever, they also create local monopolies that nationale du Bassin Congo-Oubangui-Sangha Transport: More Than the Sum of Its Parts 205 to improve the physical and regulatory arrange- developed in Africa. Whatever the main mode ments for inland navigation. A consultancy of transport, the most serious impediments are study is examining the current arrangements administrative. For example, in road transport, in the four participating countries and identi- regulation and market structures of the road fying steps to begin an effective redevelopment freight industry, rather than the quality of the of inland navigation (CICOS 2007). road infrastructure, are the binding constraints on international movements (Teravaininthorn and Raballand 2008). African governments must Developing Logistics Systems understand the importance of the qualitative aspects of logistic performance and act to remove The transport problems of doing business in administrative obstacles to that performance. Africa are not just a matter of poor infrastructure Corruption, through bribery of a range of or high transport costs. Modern logistics systems officials having a combination of discretion- also emphasize the efficiency of customs and ary power and a quasi-monopoly position in other border agencies, the ease and affordability the logistics chain, is a critical problem. For of arranging international shipments, the com- example, corruption can increase the total petence of the local logistics sector, the ability to shipping costs, including costs of overland track and trace shipments, and their timeliness. transport, port clearance, and ocean ship- These various aspects of logistic performance ping, of a standard 20-foot container trav- are encapsulated in a logistic performance eling between South Africa's economic hub index (figure 9.1). Except for South Africa, Sub- and eastern Africa or the Far East by up to Saharan African countries perform poorly not 14 percent and the total port costs by up only on infrastructure quality, but also on all the to 130 percent. Contrary to common belief, main aspects of logistics competence. no robust evidence links corruption with low Africa is thus viewed as logistics-unfriendly; wage levels or lack of job rotation. Rather, it third-party logistics systems, so important in is highly correlated with the extent to which increasing production and distribution effi- rules, regulations, and the organizational fea- ciency in the industrial countries, are poorly tures of bureaucracies give public officials the Figure 9.1 Africa Registers Low Scores on the Logistics Performance Index, 2007 1 <= LPI <= 2.29 2.29 <= LPI <= 2.53 2.53 <= LPI <= 3.14 3.14 <= LPI <= 5 no data Source: World Bank 2007. Note: LPI = Logistics Performance Index; 1 is the lowest score and 5 is the highest score. 206 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION bargaining rights to extort bribe payments Most of the critical trade facilitation imped- from shippers. In a comparison between the iments stem from the procedure and services ports of Maputo (Mozambique) and Durban in the transit country, which often sees no (South Africa), bribery of customs officials advantage for its own economy in addressing accounted for 80 percent of total bribery the needs of transit traffic. This view may, in in Maputo, but only 10 percent in Durban fact, be mistaken if transit traffic contributes (Sequeira and Macchi 2009). The explanation to scale economies in the transit country's was that, in Maputo, a low level of automa- own port and maritime or air service sectors, tion existed and both monitoring and sanc- and particularly if the goods transported com- tioning were weak, whereas in Durban, the plement, rather than compete with, its own opposite was the case. In contrast, bribery of products. Even in those circumstances, how- port officials was lower in the privately con- ever, the vested interests of the beneficiaries cessioned Maputo port, where a higher level of bribes and providers of unnecessary border of automation, monitoring, and sanctioning service may be difficult to overcome. Further- exists than in the publicly operated Durban more, governments of transit countries with port, where automation is low and monitoring potential benefits may give the wrong signals and sanctioning are weak. (see box 9.1). These problems with transit countries are increasingly recognized. In August 2003, Developing Transit Corridors for attendees at an intergovernmental conference Landlocked Countries at Almaty, Kazakhstan, agreed to a Program of Action calling for joint efforts by landlocked The weakness of logistics systems impinges and transit countries to revise their regulatory most on the landlocked countries. In southern frameworks and improve transit-related infra- Africa, operating costs are not excessively structure and procedures with the assistance of higher than in Europe ($0.08 per ton kilo- other countries and microfinance institutions. meter in 2008), but queues at borders and Some critical elements of a reform program restrictions on access to the market mean that can be identified. Reform of trucking regulation rates on international routes are higher by is essential and is discussed further in chapter 10­30 percent. Being landlocked adds four 10 of this volume. This must be supplemented days to exports and nine days to imports by a reengineering of the transit system. In the for land distribution of equivalent distances longer term, something similar to the European within a seaport country. However, this differ- Transports Internationaux Routiers (TIR), ence is usually less than the sea transit time. which provides a reliable system based on a Much more problematic is the high variability single manifest (the TIR carnet) and a chain of transit time, especially for imports. The of guarantees, could speed freight through unreliability of import and export chains international borders, though the immediate stems largely from inadequate transit proce- introduction of a system is doubtful. More dures, overregulation, multiple controls, and immediately, one-stop border posts, in which poor border service (World Bank 2008). In the two border countries share the informa- addition, bribes and unnecessary charges can tion provided, could eliminate duplication and add 50 percent to the transport costs between a speed transit. Customs must be automated, port and a landlocked country. Because of these limiting direct contact between officials and impediments, landlocked countries have a agents; monitoring and sanctioning must be much smaller involvement in world trade than strengthened; and port management must be the coastal countries, and they have remained privatized to reduce corruption. Inland clear- predominantly dependent on exports of raw ance centers or dry ports, which deal with traf- materials. Although increasing prices for raw fic under bond, can circumvent the traditional materials over the past five years have helped impediments of the seaports. Development of landlocked countries in Africa, they remain airfreight can also overcome the constraints for vulnerable to a change in that trend. some traffic. Transport: More Than the Sum of Its Parts 207 BOX 9.1 Impediments to Transit Traffic Growth in the Maputo Corridor Maputo is the closest port to the industrial center of South must face increased transport unpredictability because of Africa in Gauteng province (581 kilometers compared with the absence of direct calls by shipping lines (except Mitsui 750 kilometers between Durban port and Johannesburg) O.S.K. Lines) due to low volumes. Corruption is also greater and current congestion at Durban port gives it an apparent in Maputo, averaging 15 percent of a one-way overland ship- advantage for shippers with time-sensitive goods. If Maputo ping rate for a standard 40-foot container, compared with could capture 1 percent of the 700 million tons of traffic only 7 percent for Durban (Sequeira and Macchi 2009). from and to Gauteng and Mpumalanga provinces, its total Above all, the Mozambique government has not been throughput would be doubled. However, despite major road acting to overcome this lack of confidence, as exempli- investment in the Maputo-Gauteng corridor and a rather fied by its policies on traffic scanning. In April 2006, successful port concession process, transit traffic accounted the government granted the concession of a scanner to for less than 40 percent of Maputo's total traffic in 2007, Kudumba, a private sector operator. Initially, all exports and the port operated at less than 30 percent of its capacity. had to be scanned, including bulk cargo with a fee per With so much spare capacity at Maputo and so much con- ton and exported cars from South Africa. Although scan- gestion in Durban, both countries would seem to benefit if ning fees have decreased considerably since 2006 for cer- transit traffic were increased. tain terminal operators (for example, from $100 per car to Why has this not yet happened? Road conditions are not $15 per car), fees continue to be high compared to world a major obstacle, and transport prices are not abnormally practices. The absence of transparency and participation high along the international corridor. However, many large in the decision to introduce scanning technology and the South African shippers are reluctant to shift their transport determination of fees and the lack of a transparent bid- routes from Durban because they perceive the business cli- ding process in the award of the concession to Kudumba mate in Mozambique to be too unpredictable to invest in have further strengthened the perception of uncertainty in redirecting their logistics routes. Mozambique has a logistics the business environment. performance index below the Sub-Saharan Africa average and much below that of South Africa. Maputo port users Source: Raballand and Macchi 2008. Some progress has been made in this direc- the planned Madarail bonded container ter- tion. Trucking regulation is more liberal in minal near Antananarivo, Madagascar. southern Africa than in other parts of Sub- Much remains to be done to fill gaps in the Saharan Africa. Transit system reform is being implementation of the Almaty program. Gen- introduced to facilitate transit traffic through erally accepted corridor-monitoring indica- Cameroon to Chad and the Central African tors are required to check progress. Trucking Republic (see box 9.2). markets and the associated freight-forwarding One-stop border post initiatives apply sector reforms need to be extended. Agreement to the border between Kenya and Uganda at between neighboring countries on transit sys- Malaba; between Zambia and Zimbabwe at tem reforms must be encouraged by donor Chirundu; between Zimbabwe and Mozam- contributions and facilitated by the develop- bique at Forbes/Machipanda; along the ment of best practice standards. Trans-Kalahari Corridor; and in West Africa on selective borders of Burkina Faso, Ghana, Mali, and Togo. Some landlocked countries Increasing Competition already have bonded warehouses at the ports in West Africa, and rail concessionaires are Competition between suppliers of an individ- developing facilities to speed transits, such as ual transport mode improves service quality the Sitarail intermodal terminal proposal in and reduces costs, as shown in chapter 10 on Ouagadougou, Burkina Faso; the ZRS com- road transport and chapter 13 on air transport pany customs bond at the Victoria Falls border in this volume. Competition between modes crossing between Zambia and Zimbabwe; and may also result in the replacement of one by 208 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 9.2 A New Attempt to Reform the Transit System in the Cameroon­Central African Republic­Chad Corridor A major reform is being piloted as part of the Central A TIR (Transport Internationaux Routiers)-based interna- Africa Transport and Transit Facilitation Project financed by tional road transit convention (TIPAC, or Transit Inter-États des the African Development Bank, the Agence Française de Pays de l'Afrique Centrale), signed in 1991, was never fol- Développement, the European Commission, and the Inter- lowed. In 2008, despite the reluctance of the vested interests national Development Association. The object is to meet to lose the multiple rents in the existing system, and thanks the needs of two landlocked countries, the Central African mainly to strong leadership and pressure for reform from Republic and Chad, that suffer some of the highest inter- the Cameroonian government and customs, agreement was national transport costs and worst logistics conditions eventually reached on a substantially revised transit system. among developing countries. Freight bureaus enforcing The main elements are introduction of one common transit mandatory freight allocations and queuing manage inter- document, removal of all checkpoints on the roads, and use national transport between the two countries and the port of information technology based on ASYCUDA (Automated of Douala, Cameroon, under bilateral conventions. The SYstem for CUstoms Data) by the United Nations Conference main problems with the existing transit regime include on Trade and Development. The last adds a bar code to the very slow release of goods from the port of Douala, requir- transit document and container that is read optically at the ing seven documents, all of which must be cleared by start, destination, and border to simplify transit procedures for three separate offices. Multiple checkpoints and controls authorized freight forwarders who have obtained a standing also exist on the roads to the landlocked countries. Both customs guarantee from the banking system. If implemented, transport charges and the guarantees required from banks the changes are expected to yield substantial benefits of are significantly more costly than for comparable services in shorter delivery times, greater predictability, and lower prices. other countries. Source: World Bank 2008. the other. For example, road improvement in rail service is expected to cover its full costs, Mauritania effectively eliminated domestic air including track. The viability of privately transport, and all African railways have had concessioned railways can be hurt by under- difficulty retaining passenger traffic where charging for road maintenance and not enforc- a road alternative exists. However, various ing axle load limits on road traffic. modes may be complementing each other or National transport strategies that put competing with each other at the same time. infrastructure charging for road and rail on a Reconciling these tensions is very difficult. common basis would probably reduce budget Where modes compete, efficient alloca- burdens and improve road conditions. The net tion of traffic between them would normally effect on modal split is more difficult to predict depend on their relative prices reflecting their because of security, reliability, and other non- relative costs; however, that can be difficult to cost items considered in determining choices. interpret. Where traffic is heavily imbalanced For example, a comparison of road and rail (as in international trade) and cost structures tariffs on five concessioned railways showed differ between modes, both commercial pres- road tariffs exceeding rail tariffs by 44­213 per- sure and economic efficiency may call for cent. More competition between road and rail, widely varying price­marginal cost ratios, and abetted by increasing road costs, might reduce predatory pricing is difficult to define. trucker profit margins rather than increase the Attention to the fairness of competitive mode share or pricing power of rail. Given conditions is important. Roads are publicly the cartels and high profit rates in road haulage, provided at costs to users that do not even even greater benefits could likely be obtained cover maintenance, whereas privately provided by promoting competition in road haulage. Transport: More Than the Sum of Its Parts 209 Revisiting Attitudes toward important challenge facing the freight logistics Private Supply and Profit sector in South Africa" (Department of Trans- port 2005: 13). The dangers of exploiting pub- Private participation in supply, central to com- lic monopoly powers also arise in airports in petition, can increase the efficiency of operation several countries where the benefits of private and the mobilization of private capital. How- managerial efficiency are forgone by keeping ever, to activate these benefits, governments such activities in the public sector. must understand the commercial realities that Careful consideration must therefore be given motivate private business. Excess private profit to establishing appropriate oversight and regu- from a monopoly should be constrained. Yet latory institutions. Creation and exploitation of private firms will not participate if governments monopoly powers by cartelization need to be deny them a reasonable return on investment, a under continual review, covering both indus- necessity for a private operation and not a sin. trial structure and commercial behavior. Many Consistent policies to attract private sec- countries could develop a small but skilled unit tor finance and management should include to advise governments and specialized modal an explicit determination of the objectives of regulators on the principles of regulation. private participation. Policy makers should One further concern about private foreign recognize that private participation can bring involvement in supply merits consideration. efficiency benefits not only in cases of marginal Concerns about reliance on private foreign commercial viability, but also in sectors deemed control of services critical to national secu- highly profitable (ports) or deemed desirable rity have been widely used as an argument for albeit highly unprofitable (some railway conces- maintaining national air transport and ship- sions). That awareness calls for a wider range of ping fleets and for limiting foreign capital in forms of private participation, including nega- national infrastructure finance. In practice, tive concessions and affermage arrangements.2 uneconomically small airlines and shipping A specialist national agency dealing with priva- fleets protected by cargo reservation tend to tization could help in producing such policies push up costs and drain national resources (to and forms of participation. the detriment of national security). In addition, Government use of public sector monop- resistance to involving the global container oly powers to generate excess revenues can be port terminal operators denies countries the equally damaging, even if ostensibly to sup- efficiency and investment that such participa- port other unprofitable services. Internal cross tion could bring. subsidies have usually had adverse effects. Thus, countries need mechanisms to rec- For example, South Africa's Spoornet draws oncile private (and foreign) participation in cross subsidies for loss-producing freight and financing transport infrastructure with eco- passenger services both from the ostensibly nomic, social, and strategic objectives. This profitable ore and coal services and, through approach calls for more fundamental consid- Transnet, from ports and pipelines. eration of the nature and appropriateness of In practice, the effect has been to denude the social and strategic objectives and for analysis core services of investment. In Africa, the aver- of the relative effectiveness of different instru- age age of locomotives is 25 years and of freight ments in achieving those objectives. Develop- wagons, 25­30 years, both nearly twice the ing a range of contract designs appropriate to international best practice. In the ore services, a different objectives is part of this challenge. capacity gap exists, and Spoornet's safety record appears to be low and deteriorating. The port system is also being denied investment. Despite Meeting Social Obligations the high technical competence of the South African system, the recent National Freight Most African governments, wishing to be seen Logistics Study concluded, "Restoring rail reli- as keeping down costs of public passenger ability is fundamental and is the single most transport, have maintained low bus and rail 210 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION passenger fares. Where fares are commercially competition and usually results in low vehicle unviable, however, they are likely to under- use and high unit costs. mine the supply of the public services. Even 2. A negative concession is one offered for a loss- where the controls are enforced only on pub- making service, in which the competition is conducted in terms of the minimum public licly provided modes (notably rail or conven- subsidy required by the operator to take on the tional large bus services), they typically shift loss-making service. An affermage contract is business to an informal sector with higher similar to a management contract, but the pri- fares or lower service quality, failing to assist vate operator takes responsibility for all opera- poorer citizens. tion and maintenance functions, both technical Governments could pay compensation for and commercial. the imposed social obligations. However, only in rare cases, such as the support of the public References bus company in Addis Ababa, are public service CICOS (Commission Internationale du Bassin obligation payments adequate and timely. More Congo-Oubangui-Sangha). 2007. Plan d'action generally, the failure to achieve this timeliness strategique pour la promotion de la navigation dans le bassin Congo-Oubangui-Sangha. Hamburg: postpones, but does not avoid, the collapse HPC Hamburg Port Consulting GmbH. of public bus service. The same appears to be Department of Transport. 2005. "National Freight happening with the provisions for subsidizing Logistics Strategy." Department of Transport, rail passengers in some rail concessions. Where Republic of South Africa. subsidized rail services have economically Raballand, Gaël, and Patricia Macchi. 2008. The viable road alternatives, the subsidy (if paid) Critical Importance of Strengthened Regional drains resources from the economy. More com- Integration for Growth and Development in prehensive and well-thought-out strategies for Mozambique. Washington, DC: World Bank. fare controls and subsidies, probably by pro- Sequeira, Sandra, and Patricia Macchi. 2009. "The curing subsidized services through competitive Importance of Soft Transport Infrastructure: tendering, can ensure the most effective use of Customs Officials in Maputo versus the Port Operators in Durban." Afrique Contemporaine transport infrastructure. 230 (2). Teravaininthorn, Supee, and Gaël Raballand. 2008. Notes Transport Prices and Costs in Africa: A Review The author of this chapter is Kenneth Gwilliam, of the Main International Corridors. Directions who drew on background material and con- in Development Series. Washington, DC: tributions from Jean Francois Arvis, Rodrigo World Bank. Archondo-Callao, Jose L. Guasch, Alberto World Bank. 2007. Connecting to Compete: Trade Nogales, Gael Raballand, Sandra Sequeira, and Logistics in the Global Economy--the Logistics Kavita Sethi. Performance Index and Its Indicators. Washington, 1. Tour de role dispatching is a system in which DC: World Bank. service suppliers are allocated traffic strictly in ------. 2008. Improving Trade and Transport for turn. It is widely applied in the public transport Landlocked Developing Countries. Washington, and trucking markets. This system limits direct DC: World Bank. Chapter 10 Roads: Broadening the Agenda A frican governments have been address- and financial flows for the interurban roads, ing the low density and poor condition other challenges have surfaced that will require of their road networks. Institutional different types of solutions. reform since the mid-1990s has progressed well, First, the reforms to the interurban road with a remarkable consensus on the content. network have affected rural roads much Most countries have second-generation road less. Even though agriculture is viewed as funds supported by fuel levies, and many oth- an engine of growth, only one-third of rural ers have autonomous road agencies. Specialist inhabitants live within 2 kilometers of an maintenance management agencies have been all-season road. Doubling this percentage established, and new forms of contract-based would be very costly, absorbing more than maintenance are being introduced. Although 1 percent of GDP a year for a decade. The important funding gaps remain, results are dis- rural environment presents particular insti- cernible. On average, 80 percent of the main tutional challenges for road maintenance. road network is in good or fair condition, and Second, surface transportation is about more the current value of the national road net- than good roads. Africa continues to be handi- works is at least 70 percent of their potential. capped by very high road freight tariffs, driven The limited time series available also suggests primarily by high profit margins rather than that a number of countries have improved road high costs (or defective roads). In Central and conditions over time. West Africa particularly, trucking industry car- Despite this progress, the reform agenda tels and restrictive tour de role traffic allocation is incomplete. In many cases, fuel levies have and dispatching practices are responsible for been set too low to be effective, and road funds low vehicle mileage and poor fleet quality. The and agencies do not always meet all good- most urgent reforms are to liberalize trucking practice design criteria. Modern contracting while mitigating associated social effects. With- and contract management methods are far out such measures, further improvements in from universal. Furthermore, while policy mak- road quality will only translate into higher profit ers' attention has focused on the institutions margins for the trucking industry. 211 212 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Third, Africa's rapidly growing cities face the road conditions lag somewhat behind those major mobility problems. Urban road den- of other developing regions, but not so much sity is low by developing-country standards. for the main trunk road network as for other Moreover, following the demise of large buses roads (Gwilliam and others 2008). in many cities, myriad informal minibus With accelerating urbanization, Africa needs operators largely dominate urban transport to develop intraurban roads, but networks in services. Services are costly, and availability is 14 African cities were found to be substan- inadequate. Few countries capture sufficient dard (Kumar and Barrett 2007). Road den- financial resources to develop and main- sity (paved-road density in particular) lags far tain the urban road network. Overlapping behind that in other developing cities. Capacity national, metropolitan, and municipal juris- is generally limited. The majority of roads have dictions present serious institutional chal- one lane in each direction, and where roads lenges. Furthermore, the cross-sectoral links are wider, pedestrians and parked vehicles between urban transport and land use plan- often take up one lane. Intersections are close ning are unexploited. together and are ill designed for turning. Ser- vice lanes are absent, pavement is deteriorating, and street lighting is minimal. Because traffic Road Infrastructure--Lagging management is limited, accidents are frequent, Other Regions Somewhat with pedestrians accounting for two-thirds of fatalities. The region's trunk road network comprises For rural roads, beyond the classified ter- strategic trading corridors linking deep-sea tiary network, which is typically the responsi- ports to economic hinterlands. These corri- bility of local government, a vast unclassified dors, which carry about $200 billion of trade network of tracks providing service to rural a year, include no more than 10,000 kilome- areas is usually the responsibility of local com- ters of road. The concept of an intraregional munities. Nevertheless, African rural commu- trunk network, or Trans-African Highway, nities have by far the lowest accessibility to an remains a distant reality because of missing all-season road in the developing world. Evi- links and poor maintenance on key segments. dence indicates that physical isolation prevents Between 60,000 and 100,000 kilometers of large areas of the continent from reaching their road are required to provide such intraconti- agricultural potential. With low population nental connectivity. density, achieving good overall rural accessi- Africa's national road density is substantially bility would imply at least doubling the length lower than that in other developing regions: of the classified network for most countries only 204 kilometers of road per 1,000 square (Starkey and others 2002). kilometers of land area, with only one-quarter Traffic volumes remain low and heav- paved, compared with a world average of ily concentrated on the main road network 944 kilometers per 1,000 square kilometers, (table 10.1). In most countries, at least with more than half paved. That density is less 90 percent of reported traffic on the classified than 30 percent of the next-lowest region, South network is carried on the main networks.1 Asia. However, Sub-Saharan African road den- Except in Nigeria and South Africa, the traffic sity in relation to population is slightly higher on the main road network in Africa averages than South Asia's and only slightly lower than only about 500 vehicles a day. the Middle East's and North Africa's. Rural networks typically carry less than Relative to GDP, however, Sub-Saharan 10 percent of the classified network's traffic; Africa has a large road network. In Madagascar, however, in Ethiopia, Malawi, and Nigeria, Malawi, Mozambique, and Niger, the asset they carry more than 20 percent. Except in value of the road network exceeds 30 percent Nigeria, the absolute volumes of traffic on the of GDP, an indication of the consequently large rural network are very low, averaging about economic burden of maintenance. As a result, 30 vehicles a day. Roads: Broadening the Agenda 213 Table 10.1 Average Daily Traffic on the Main Road Network Classified network Classified Paved Unpaved Country type network Primary Secondary Tertiary network network Low income 236 934 182 28 1,054 50 Lower middle income 341 1,186 303 39 1,474 95 Upper middle income 1,066 5,469 117 24 2,883 5 Source: Gwilliam and others 2008. Road Infrastructure Institutions captured fewer than $1.6 billion in investment and Finance--Promising commitments, small in relation to the region's needs. Less than 10 percent of the region's road Developments network attracts the 15,000 vehicles a day that are the minimum traffic flow needed to make The initial thrust of institutional reform has concessions economically viable. Toll roads been to create an independent source of fund- have potential only in South Africa and to a ing for road maintenance, based on road user lesser extent, in Nigeria. charges, segregated from the general govern- ment budget, and administered by an autono- mous board. Implementation Agencies A second stage of reform has created road Funding Arrangements agencies, independent from line ministries, Donors have played a major role in promot- with responsibility for contracting out pub- ing this institutional framework. Most coun- lic works. About two-thirds of the 24 coun- tries already have such second-generation tries sampled have a road agency, and others road funds, and most others, except Nigeria are planning one, but only one-third of the and South Africa, are establishing them. Not agency boards have private representation. all funds have good-practice designs, how- Nigeria, Senegal, and South Africa have a ever, and their performance varies substan- road agency but not a road fund. Autonomy tially (figure 10.1). varies from full responsibility for road net- Despite widespread application of fuel lev- work management to limited responsibility ies to fund road maintenance, the level of the for road maintenance programs defined by fuel levy, hence its utility, varies enormously the road department or ministry. across countries. The range extends from sym- About half of the countries sampled con- bolic levels of about $0.03 a liter, nowhere near tract out more than 80 percent of maintenance high enough, to about $0.16 a liter, which cov- work. Some road agencies are adopting per- ers most road maintenance needs. Many coun- formance-based maintenance contracts, under tries have major difficulties in collecting the which a private contractor maintains a public levies, whether because of evasion (Tanzania) road to achieve and maintain specified condi- or delayed transfers of revenues (Rwanda), and tion standards for periods ranging from 3 to capture perhaps as little as half the planned 10 years in return for a fixed payment stream. resources. Therefore, the road funds in Benin, The advantages of such contracts are that they Côte d'Ivoire, Ethiopia, Gabon, and Zambia can provide a strong incentive for contractors still depend on budget allocations for more to undertake effective maintenance, and they than three-quarters of their resources rather can reduce expenditure uncertainties for the than being funded largely from fuel levies, as road fund. The contracts started in Canada is the intention of road funds. in the late 1980s, and industrialized coun- Toll roads affect barely 0.1 percent of the tries have now adopted them. In developing region's classified road network, almost entirely countries, they were first applied in Argen- in South Africa. Toll-road concessions have tina in mid-1990, but they rapidly spread 214 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 10.1 Progress with Road Fund Reforms a. Prevalence of second-generation b. Scores on overall performance index design characteristics Tanzania independent auditing Namibia Kenya Rwanda road user charges Malawi Madagascar Ghana revenue allocation rules Ethiopia Chad Zambia separation of functions Niger Mozambique Cameroon clear legal basis Lesotho Côte d'Ivoire direct transfer Benin South Africa Senegal user representation on board Cape Verde Burkina Faso 0 20 40 60 80 100 0 20 40 60 80 100 % of countries % of score Source: Gwilliam and others 2008, based on data collected by World Bank 2007. to neighboring countries so that more than responsible for maintaining a level of service 40,000 kilometers of Latin American roads are linked to the condition of the road surface. Under now being maintained under such contracts. a performance-based maintenance contract, the In Africa, Ethiopia, Ghana, and Zambia have condition of the road improves steadily, whereas begun to use them. under the traditional approach, the road condi- Cost savings from performance-based tion improves for a short period following the maintenance contracts on paved roads have work and then starts deteriorating quickly until ranged between 10 and 40 percent in industrial- new maintenance is carried out. ized countries and between 10 and 20 percent A parallel institutional development, par- for several developing countries (Stankevich, ticularly relevant to rural roads, involves Qureshi, and Queiroz 2005). Even where such delegating project management to special- cost savings were not achieved, the benefits have ist agencies. In many countries, these agen- been substantial. In Chad, the only example in cies, agences d'exécution des travaux d'intérêt francophone Africa, the cost per kilometer of public (AGETIPs; public works implementing routine maintenance under a performance-based agency), now manage private consultants and maintenance contract for a set of gravel roads contractors on behalf of the public authority was significantly higher ($5,000) than under and perform all the necessary functions for a traditional maintenance contract ($1,500) contract preparation, implementation, and but with the benefit that the contractor was supervision (box 10.1). Roads: Broadening the Agenda 215 BOX 10.1 The Role of AGETIPs Delegating the function of managing the their work. There are now 19 such agencies planning, procurement, and implementation in 17 countries, mostly in francophone West of public works to a specialist private agency Africa. The international association AFRICATIP is well established in French public administra- (Association Africaine des Agences d'Exécu- tion. In many African countries, an AGETIP tion des Travaux d'Intérêt Public) develops and was attractive to international financial insti- shares best practices among its members. tutions as a means of obtaining more effective The agencies play three roles: (a) compe- implementation of donor-funded works, par- tent technical agencies using private sector ticularly in rural sectors, where administrative recruitment and payment procedures, (b) man- skills were weakest. agers of special funds, and (c) directors of the Following the creation of the first AGETIP planning and programming of the investments in connection with a donor­financed sites and of local authorities. They also provide techni- services project in Senegal in 1989, the greatly cal manuals and contractor training. Although improved administration of public works and originally established to facilitate donor financ- timely payment enable a substantial increase ing, they now handle mostly national funds in the participation of small and medium-size and have become instruments for indigenous enterprises in public works programs, often business development. using labor-intensive techniques. Roads, par- Source: Diou, Henry, and Demy 2007. ticularly rural roads, are an important part of Road Expenditures--More in road spending, with investment account- Maintenance, Less Rehabilitation ing for two-thirds of total spending in the resource-rich and low-income countries, par- On average, countries spend $9,000 a kilometer ticularly those without adequate institutional for main road networks in Sub-Saharan Africa, mechanisms for funding road maintenance. just below 2 percent of GDP, compared with 1 Middle-income countries and those with high percent in typical industrial countries and 2­3 fuel levies tend to spend more on maintenance percent in fast-growing emerging economies without incurring higher road expenditure (figure 10.2). Although the effort is high rela- overall. This finding clearly shows that timely tive to the size of Africa's economies, it remains attention to maintenance reduces the expen- low in absolute terms, with low-income coun- diture needed to sustain the road system in tries spending no more than $7 per capita a the long term (Harral and Faiz 1988). Aid has year. For the main road network, maintenance partly fueled this capital bias. For example, aid spending ranges from barely $200 per kilome- financing covers just over 50 percent of road ter in Chad to more than $6,000 per kilometer investment in Senegal and almost 90 percent in Zambia. Maintenance spending per kilometer in Rwanda. of the main network tends to be about twice that The capital bias would be even more pro- of the rural network. nounced if capital budgets were fully executed. Paradoxically, low-income countries spend On average, countries have budgeted 50 percent 50 percent more per kilometer overall than more on road investment than they spend dur- do middle-income countries, while countries ing a given budget cycle. This underspending with road agencies and high fuel levies seem produces a capital budget execution ratio aver- to spend somewhat less than those without. aging about 70 percent. Deficiencies in project The explanation is a pronounced capital bias planning and delays in procurement processes 216 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 10.2 Average Annual Spending on Road Transport, by Country, 2001­05 4.5 35 4.0 30 3.5 25 3.0 US$ per capita % of GDP 2.5 20 2.0 15 1.5 10 1.0 5 0.5 0 0 n in ire Rw a da Ni r ria Ta a ia Ug l da ag d Za r bia am ia ue wi pe a Na e Le a o ga ge ca rd ny an Ca fric ibi oo M Cha th an p Cô Ben ala biq an an Ivo ge as ne m M hio Ni Ve so Ke m Gh er nz A M Se d' m Et h Ca ad ut te oz So as % GDP US$ per capita Source: Briceño-Garmendia, Smits, and Foster 2008. are the main culprits. Middle-income coun- Figure 10.3 Rehabilitation and Maintenance tries and those with road funds and fuel levies Spending Relative to Norms fare best in executing their capital budgets. High capital spending may be justified by South Africa large rehabilitation backlogs. Except for Chad Zambia and Ethiopia, capital spending for many sam- Cameroon pled countries indeed falls well below or is close Namibia to what is needed to clear rehabilitation back- Benin logs within (a reasonable) five years. However, Mozambique such high levels of spending on rehabilitation Tanzania make sense only when a broader policy is in Kenya place to ensure proper maintenance of these Ghana roads after they have undergone rehabilita- tion. In practice, half of the countries sampled Rwanda are not devoting adequate resources to main- Lesotho tain the main road network, and about half of Ethiopia this subset is not even spending enough for Madagascar routine maintenance (figure 10.3). In Chad, Malawi Niger, Nigeria, Senegal, and Uganda, mainte- Senegal nance spending is less than half the norm. Niger Unit costs of road construction have recently Uganda escalated, with cost overruns on multilateral agency projects rising from 30 percent in 2005 Nigeria to more than 60 percent in 2007, threatening to Chad dilute further the adequacy of current budget ­200 ­100 0 100 200 300 400 500 allocations. Inflation has been substantial for % deviation from norm the basket of road construction inputs linked maintenance rehabilitation to oil prices, but it does not tell the whole story. The lack of effective competition for civil works Sources: Briceño-Garmendia, Smits, and Foster 2008; Gwilliam contracts, with a small number of bidders and and others 2008. Roads: Broadening the Agenda 217 wide price spreads across bids, looks to be the substantially better road conditions than do main culprit (Van Zyl, Coetzer, and Lombard those that lack either. Moreover, both the road 2008). Substantial delays in project implemen- fund design and the level of the fuel levy appear tation also add to the costs. to significantly affect the quality of the main These higher costs double the importance road network, although again the effect on rural of ensuring that engineering standards are the road quality is much less pronounced. The main most cost-effective possible. Network analy- exception, South Africa, has very good primary sis reveals that on average, about 30 percent roads without a conventional second-generation of main road networks are overengineered road fund, but it does have a very effective road relative to observed traffic volumes, and only agency and consistent government commitment 10 percent are underengineered. Particularly to finance road maintenance adequately. where rapidly expanding the extent of the Surprisingly little variation exists in the passable road network is desirable, as in the road network's asset value as a percentage of its Democratic Republic of Congo, adopting potential value if it all were in good condition. lower standards may be sensible, upgrading All countries realize at least 70 percent of this only when traffic growth or local bottlenecks potential, suggesting that maintenance efforts make it necessary. About 15 percent of rural are fairly well concentrated on preserving the networks are underengineered. Greater efforts high-value paved-road network. Moreover, the are needed to adapt road design standards to limited time series available suggests that most local conditions and materials to avoid exces- countries have improved the quality of their sive costs in road construction, particularly for roads in recent years. low-volume sealed roads. Good governance is thus critical for safe- guarding road quality through good budget finance and a professionally competent public sector implementation agency. Countries with Road Conditions--Reflecting road funds and high fuel levies are substan- Quality of Sector Governance tially more successful at raising finance that translates into higher road maintenance expen- About half of the main network is in good con- ditures. Countries with road funds and quasi- dition, and an additional one-third is in fair independent road agencies show substantially condition--whereas only one-quarter of the higher quality on main road networks. rural network is in good condition and a further one-quarter in fair condition (figure 10.4). The large variation in road quality reflects several interacting factors. First is affordabil- Infrastructure Spending ity. GDP per capita is most strongly correlated Needs--an Average of with the percentage of the main road network 1.5 Percent of GDP a Year in good condition, suggesting that richer countries tend to spend more on maintenance. A modest set of connectivity objectives might However, no such clear relationship exists for include the following: (a) connecting large cities rural roads. Second are some fundamental top- and international frontiers with a good-quality, ographic and climatic influences. Mountain- two-lane paved road; (b) connecting intermedi- ous and wet countries generally have poorer ate cities and the provincial capital with a good- conditions, in both main and rural networks, quality, one-lane paved road; (c) increasing to associated with climate rather than traffic. 100 percent the proportion of the rural popula- Even controlling for income and climate, tion living within 2 kilometers of an all-season however, substantial variation still exists in road in the agricultural areas comprising the 80 observed road quality across countries, vary- percent highest production value within each ing with the quality of their road sector insti- country; and (d) putting the urban population tutions and financing framework. Countries within 500 meters of a road supporting all-sea- with both road funds and road agencies show son bus access. 218 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 10.4 Distribution of Road Network Length across Condition Classes, by Country a. Main network b. Rural network South Africa Burkina Faso Burkina Faso Ghana Kenya Kenya Namibia Malawi Ethiopia Chad Malawi Namibia Ghana Cameroon Zambia Ethiopia Mozambique Lesotho Niger Niger Cameroon Tanzania Nigeria South Africa Madagascar Côte d'Ivoire Tanzania Benin Chad Senegal Senegal Mozambique Lesotho Zambia Benin Nigeria Rwanda Madagascar Uganda Uganda Côte d'Ivoire Rwanda 0 25 50 75 100 0 25 50 75 100 % road network length % road network length good fair poor Source: Gwilliam and others 2008. On the basis of this package, spending Existing spending on the sector amounts to needs for the road sector amount to $9.6 $6.9 billion a year, significantly less than what billion a year, skewed toward capital expen- is needed (table 10.3). Spending in the fragile diture (table 10.2). Because of their mature states is particularly low, barely one-tenth of road network, the middle-income countries what is required. The public sector finances two- account for little more than 10 percent of thirds of road sector spending and more than this total. Except in middle-income coun- one-half of road sector investment. In the low- tries, about two-thirds of spending require- income countries--whether fragile or not-- ments relate to capital expenditure, with about half of road sector expenditure is donor the remainder attributable to operation financed. The contribution of the private sector and maintenance. Thus, overall, the region to road finance in Africa is almost negligible. needs to spend 1.5 percent of GDP on roads, Financiers from outside the Organisation for of which 0.6 percent of GDP is needed for Economic Co-operation and Development are road maintenance. However, the burden for not making a major contribution to this sector. low-income fragile states is very high--in Implementation of efficiency-oriented excess of 7 percent of GDP. reforms could raise a total of $3.8 billion a year, Roads: Broadening the Agenda 219 Table 10.2 Road Sector Spending Needs $ billions annually Percentage of GDP Operation Operation Capital and Total Capital and Total Country type expenditure maintenance spending expenditure maintenance spending Sub-Saharan Africa 5.98 3.65 9.63 0.93 0.57 1.50 Middle income 0.40 0.46 0.86 0.15 0.17 0.32 Low income, fragile 1.89 0.83 2.72 4.92 2.15 7.07 Low income, nonfragile 1.84 1.23 3.07 1.67 1.11 2.78 Resource rich 1.86 1.14 3.00 0.84 0.51 1.35 Source: Carruthers, Krishnamani, and Murray 2008. Note: Totals may not add exactly because of rounding errors. Table 10.3 Financing Flows to the Road Sector $ billions annually Operation and maintenance Capital spending Public Public Non-OECD Total Country type sector sector ODA financiers PPI Total spending Sub-Saharan Africa 1.45 3.22 1.80 0.37 0.05 5.44 6.88 Middle income 0.41 1.21 0.09 0.02 0.05 1.37 1.77 Low income, fragile 0.06 0.19 0.23 0.03 0 0.45 0.50 Low income, nonfragile 0.61 0.58 1.24 0.14 0 1.96 2.57 Resource rich 0.31 1.29 0.23 0.17 0 1.69 2.01 Source: Briceño-Garmendia, Smits, and Foster 2008. Note: Operation and maintenance includes other current expenditures. ODA = official development assistance; OECD = Organisation for Eco- nomic Co-operation and Development; PPI = private participation in infrastructure. Totals may not add exactly because of rounding errors. largely eliminating the funding gap, except in "hard" infrastructure. However, what ulti- fragile states. The greatest scope for efficiency mately matters from an economic perspective gains lies in practicing sound preventive main- is the extent to which roads support efficient, tenance, which in the medium term would reliable, and safe transport services for various substantially reduce the investments needed kinds of freight, as well as for urban and rural to clear the rehabilitation backlog, saving an populations. Although the private sector typi- estimated $1.9 billion per year. Low ratios of cally provides these services, the government capital budget execution are also holding back has a critical role to play as regulator and public investment in roads. Addressing this facilitator of service provision. These "soft" issue would capture a further $1.3 billion annu- transport issues deserve more attention. ally. Finally, some countries face difficulties in collecting revenues owed to their road funds; Road Freight solving this problem would capture another The regulation and market structures of the $0.6 billion a year of resources. road freight industry, rather than the quality of the road infrastructure, are the binding con- straints on performance in the international Transport Services--the Forgotten corridors (Teravaninthorn and Raballand 2008). Problem Although the associated road infrastructure is generally in good condition, the administrative Road sector interventions have traditionally bottlenecks at borders and ports keep the effec- focused on constructing and improving the tive velocity of transit along these routes very 220 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION low (typically less than 10 kilometers an hour). Out-of-village travel by motorized transport, Even for national traffic, the exceptionally high from villages to market towns and from towns road freight tariffs in parts of Africa--reaching to cities, is less common. Supply is typically $0.13 per tonne-kilometer in Central Africa-- fragmented and informal, and rural com- are attributable more to high profit margins munities are often captive markets for local (60­160 percent) than to any inherent cost dis- monopolists. advantage (table 10.4). Marked performance differences occur Urban Road Transport across subregions. Performance is worst in Buses are the common mode of public tran- Central and West Africa and best in southern sit in most cities (Kumar and Barrett 2007). Africa, with East Africa in between. The dif- Except in a handful of cases, however, mini- ference can be explained by industry cartels in buses are much more prevalent than large Central and West Africa, together with the tour buses. About twice as many trips are taken by de role regulatory framework. That framework minibuses and shared taxis than by large is based on market sharing and centralized buses. The use of motorcycles for commercial allocation of freight that limits vehicle mile- transport has also grown very rapidly in recent age and undermines incentives for improving years, mainly because of the poor state of the fleet quality. In southern Africa, by contrast, roads and the inability of bus companies to a much larger share of freight traffic is allo- meet growing demand. Small-scale suburban cated through competitive bilateral contracts rail networks exist in a few cities, but nowhere between clients and shippers. do they account for more than 2 percent of the market. Rural Road Transport Supply is inadequate and tariffs are high. Most rural transport takes place near vil- Most African cities have 30­60 public trans- lages. Trips generally involve short distances port vehicle seats per 1,000 residents, but only and small loads carried on paths, typically 6 large bus seats per 1,000 residents (compared for marketing, collecting water and firewood, with 30­40 in middle-income countries). Low and tending crops and animals. Most trips fleet capacity is exacerbated by poor use of the are walking trips. Nonmotorized transport, limited vehicle fleet, with vehicles achieving such as a bicycle, is often unaffordable, and fewer than 200 kilometers a day. The quality where mechanical transport is available to the of public transport is consequently poor, with household, it tends to be appropriated by a long walking and waiting times typically dou- male household head. Changing this practice bling the in-vehicle time. Extreme overcrowd- may be a demanding cultural task; in some ing is also common, particularly on large parts of Africa, women's use of bicycles may buses. The average cost of a one-way trip, at be considered unseemly. Nevertheless, when about $0.30, is high in relation to household enough women begin to use them and the budgets. Regulations that keep fares for large benefits to the whole household become buses below those of minibuses and inappro- apparent, the practice may rapidly gain social priate cost benchmarks have contributed to acceptance, as it has in parts of Burkina Faso. the demise of large buses. Table 10.4 Overview of Key Road Freight Parameters on Main International Corridors Trade density Implicit velocity Freight tariff Roads in good ($ millions per (kilometers ($ millions per Profit Corridor condition (%) kilometer) per hour) tonne-kilometer) margins (%) Western 72 8.2 6.0 0.08 80 Central 49 4.2 6.1 0.13 70­160 Eastern 82 5.7 8.1 0.07 70­90 Southern 100 27.9 11.6 0.05 20­60 Source: Teravaninthorn and Raballand 2008. Roads: Broadening the Agenda 221 Minibus ownership is generally highly Completing the Institutional Reforms fragmented, with most individual entre- Countries with road funds are more successful preneurs owning no more than one or two in safeguarding road maintenance expendi- vehicles, generally rented out to drivers. Pow- tures and spending on segments that maxi- erful unions, associations, or syndicates that mize network value, while those that also have organize the sector and provide a degree of road agencies see greater funding going to self-regulation, typically based on equitable better road quality. Although sector reforms sharing of the market in the tour de role sys- are widespread, closer inspection reveals that tem, offset the highly fragmented ownership. the quality and depth of those reforms are This approach contributes significantly to quite variable across countries. poor vehicle use and to long walking times Road funds need to be designed in line with and waiting times at terminals. accepted good-practice criteria. Key areas of Although not confined to the urban trans- deficiency include (a) a lack of user represen- port situation, road safety is also a very seri- tation on road fund boards, (b) the absence of ous issue in most African countries. In the direct transfer mechanisms to ensure that fuel early 2000s, nearly 3,000 people were killed levy revenues go directly to the fund (circum- on Kenyan roads annually, about 68 deaths venting national budgets), (c) a strong legal per 1,000 registered vehicles, 30 to 40 times foundation for road funds to safeguard their the rate in highly motorized countries. Traffic autonomy, (d) clear allocation rules for road accidents are the third-leading cause of death fund revenues, and (e) a lack of systematic after malaria and HIV/AIDS, presenting major economic analysis guiding resource alloca- public health problems in disability and health tion. Some evidence indicates that the benefits care costs. of road funds are larger where countries adopt the full set of good-practice design criteria. Fuel levies need to be set at an adequate level and supported by an effective revenue col- Moving Forward--Broadening the lection mechanism. Although many countries Reform Agenda have fuel levies, only a subset is high enough (over $0.10 a liter) to generate revenues com- The institutional reform agenda needs to be mensurate with road maintenance require- completed and broadened to encompass the ments. Even where levies are right, difficulties demands of urban and rural connectivity. in revenue collection can prevent their full cap- This connectivity is not just about physical ture. Effective administration is thus equally infrastructure; it is also about the regulatory important. framework governing transport services. The Efficient road implementation agencies are recent (and likely continuing) escalation of a necessary complement to road funds, ensur- unit road costs will strain already stretched ing that resources are well spent; however, transport budgets. Road safety also remains quasi-autonomous road agencies are not yet as a concern. To deliver on these challenges widespread as road funds. Moreover, many of will require continuing emphasis on creating them fall short of good practice, particularly in efficient agency structures to manage road autonomy and user representation on boards. programs and strengthen government capa- A key element for their success appears to be bilities for oversight. The agenda comprises the adoption of performance-based mainte- the following: nance contracts, which are still less developed · Completing institutional reforms in Africa than in other regions. Although agencies can overcome con- · Increasing rural accessibility straints of public sector salaries and processes, · Developing urban transport services government structures must still carry out key · Liberalizing road freight transport functions: (a) determining road standards, (b) · Dealing with escalating unit costs carrying out road classifications, and (c) set- · Improving road safety. ting long-range planning goals. Governments 222 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION will continue to formulate transport policy be achievable simply through spot improve- and to regulate and oversee the new road insti- ments at vulnerable points (such as creeks and tutions. If they do not perform these functions riverbeds) without the need to upgrade the effectively, the new institutions are bound to surface of the road along its entire length. suffer. For example, road funds will not get the Whatever the chosen objective for rural revenue increases they need, and road agen- road development, financing will likely remain cies will be unable to deliver realistic network a challenge. Local governments mobilize only improvement programs. modest revenues of their own, with market and business taxes as the main sources. Intergov- Increasing Rural Accessibility ernmental transfers are thus the main source Providing full road accessibility to 100 percent of domestic funding for local government of Africa's rural inhabitants would entail a vast spending in many countries. This situation expansion in the all-season road network, vir- poses three main problems. First, throughout tually tripling its length. For many countries, most of Sub-Saharan Africa, less than 5 percent this goal is unlikely to be affordable in the of aggregate public revenue is generally made medium term, highlighting the need to care- available to local governments managing rural fully select and prioritize rural road invest- networks. Second, general budgets rarely allocate ments. One way of doing so is to strategically adequate funds for maintaining main roads, align rural road investments with agricultural much less rural roads. Third, capital and recur- development programs at the national level, rent allocations to local governments are usually to prioritize those rural roads likely to have not fungible, and the allocation for recurrent the largest effect on agricultural productivity expenditures may barely cover the salaries of and market access. Recent analysis of Central the rural road unit. Moreover, the budget cycle Africa suggests that the most attractive rural dictates such transfers, so that central-to-local road investments may lie in areas that are at transfers are unlikely to be adequate and timely some distance from major urban markets but for maintaining local government roads. still within reasonable reach, because in these Adequate, steady funding for local govern- cases, rural road investments may be the criti- ment maintenance is more likely to be forth- cal intervention needed to open up market coming from a dedicated road fund, as long as accessibility (Briceño-Garmendia and others some formal commitment exists in the road 2009a, 2009b). By contrast, road investments fund law to ensure that it accepts responsi- in very isolated rural areas may be less attrac- bility and provides for local roads. Although tive because they would not make enough of 60 percent of road fund revenues is typically a difference in overall travel times to provide allocated to the main interurban road network, adequate market access. countries have, to varying degrees, channeled Given the vast scale of Africa's potential rural portions to the maintenance of rural road net- road network, the issue of keeping down unit works. This approach appears to be effective. costs becomes particularly critical. Policy mak- Countries that allocate at least $0.015 of their ers face a stark trade-off between the standard fuel duties to rural roads have 36 percent of to which rural roads are built and the length their rural networks in good condition, com- of the rural network that can be developed for pared with 21 percent for those that do not. any given budget envelope. This choice raises Building full capacity for all management questions about what kind of rural roads farm- functions in each local government and com- ers really need. In many rural communities, munity is unrealistic and inefficient. Individ- volumes of production may be well below the ual local government networks are small, and threshold needed to justify the use of a truck the management contract for an individual to collect produce, and simpler roads targeted local government may be too modest to attract more at ensuring accessibility for two-wheeled competent consulting firms. In Madagascar, vehicles or animal-drawn carts may be more the average network for a local government is suitable (Raballand and others 2009). In other 140 kilometers; in Cameroon and Nigeria, it is settings, ensuring all-season accessibility may 180 kilometers; and in Tanzania and Zambia, Roads: Broadening the Agenda 223 it is 280 kilometers. All fall far short of the Community infrastructure, including mainly 500­2,000 kilometers needed to justify unclassified roads and paths for which no level employing an engineer in a local unit. Joint of the formal government accepts responsi- service committees of local authorities can bility, faces particular problems. Community achieve economies of scale in procurement for contributions in cash and kind are suitable the authorities they represent, but they usually primarily for community roads and paths, but require substantial technical assistance from in-kind contributions may be inefficient, and central ministries or from the regional offices other sources of money are necessary. Cost of a main road authority. In countries with an sharing between local communities and other autonomous road authority responsible for government or external agencies can help raise main roads, local governments can contract the volume of resources mobilized and thereby with the road authority to manage the roads increase the proportion of the network that on their behalf or to assist with planning and receives regular maintenance. Well-structured procurement. donor financing through rural road projects or Some countries centralize the technical through social and community or rural infra- responsibility for rural roads. Relying on a road structure funds can support investment in com- ministry or another central ministry to man- munity infrastructure. Cost sharing may also age rural roads has the advantage of a formal be effective in maintaining community roads. channel for technical support. However, the Many local authorities in Africa have more disadvantage is that the ministry often operates roads to maintain than they can afford, so cost completely independently of the local govern- sharing with communities has merit. ment structure and thus is poorly connected Lack of technical know-how often impedes to local needs and developments. In principle, community management. Communities in a central coordinating unit for local govern- Sub-Saharan Africa need technical advice (on ment roads should perform as well as a central road design and standards, appropriate materi- government rural roads department. In prac- als, work planning) and managerial advice (on tice, however, coordinating units for local gov- financial accounting, contract management, ernment roads are not always as strong as they procurement) to perform the responsibilities need to be, as is illustrated by the experience of that come with ownership. Tanzania and Zambia in the late 1990s. For rural transport services, the main issues Delegation of planning, procurement, and are increasing service quantity and keeping management has already been improved in services affordable. The priority in transport many countries through the establishment of services in rural areas must be to maintain AGETIPs, and the national institutions have basic year-round vehicle access for the types of benefited from their association in AFRICA- vehicles likely to be operating. The quantity of TIP and from donor assistance. Considerable access is even more important than the quality. room for improvement exists in the work Better rural telecommunications can provide of the AGETIPs, however, particularly with the means of more effectively matching vehi- respect to their technical capacity, the quality cles to loads. Given the monopoly power of of preparatory studies and contract supervi- service providers, communities can organize sion, and delays in project implementation to increase their bargaining power through a (Diou, Henry, and Demy 2007). collective lobby. Operating subsidies are usu- The options are not mutually exclusive. ally infeasible, but providing credit for vehicle For example, a joint services committee can purchases, possibly through piggybacking on use private consultants, hired through a con- agricultural credit programs, is an option. tract management agency. The best option for managing local roads depends on many local Developing Urban Transport Services factors, including the size of the authorities, Urban public transport requires simultane- the nature of the network, and the compe- ous and integrated attention to planning tence of the private sector or higher public urban structures, building and maintain- authority units. ing infrastructure, and organizing transport 224 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION services. In practice, these three primary resources to reestablish a large-bus sector while functions are seldom housed in the same the private sector remains wary of investing in institution, and even where they remain with large vehicles. The short-term options are thus the central government, several ministries to rely on self-regulation of the sector, which are usually involved. Only a handful of has usually failed, or to devise competitive African cities have agencies with metropolitan structures, either "for" or "in" the market to responsibilities and overarching functions, build private confidence in a managed private and even those agencies lack the executive market. Attempts can also be made to stimu- powers to implement their vision and must late increased vehicle size, as in South Africa. work through other units of government. This sort of strategy has achieved limited The institutional arrangements for urban success in some central Asian countries. roads are frequently complex. Legislation In the very largest cities, exclusive road- pertaining to roads is usually separate from based track systems, such as bus rapid transit that governing transport services, and several or (more costly) light rail, may have a role. Such national and local bodies often share juris- developments are being considered in Dar es diction. In Conakry, Guinea, several institu- Salaam in Tanzania and Lagos in Nigeria, but tions are responsible for segments of the road they are still experimental. network. In Accra, Ghana, responsibility for urban transport has been devolved from cen- Liberalizing Road Freight Transport tral to local government--at least in principle. Freight tariffs in much of Africa are unneces- However, local governments have neither the sarily high because of restrictive regulation resources nor the technical know-how to carry and weak competition among truckers. The out their assigned functions, so the ministry most damaging aspect of trucking organi- of transportation (through the department of zation is the combination of self-regulation urban roads) is effectively responsible for road with national protection. Both militate for the maintenance and development. interests of the incumbent national operators In the passenger transportation market, at the expense of their customers. They cre- the self-regulation of operators associations ate scope for corruption while leaving socially and cartels has ossified fragmentation of costly problems (vehicle overloading) relatively informal operations and wasteful institution- untouched. Both areas offer alternatives. alized procedures. Supply is inadequate and Self-regulation is a means of maintaining expensive. Two main options exist to remedy on-the-road discipline in an excessively frag- this situation: small and medium-size vehi- mented market. It fills a vacuum created by the cles and, in the largest cities, light rail. absence of effective public regulation. By its In principle, traditional, disciplined large- nature, it concentrates on ensuring an equita- vehicle services could be reintroduced, but ble distribution of traffic between members of such attempts--in Dakar (Senegal), Accra, the association, typically through the wasteful and other cities--have failed. The association operating procedure of tour de role dispatching. of large vehicles with traditional large public The alternative is a combination of freedom of monopoly companies made them vulnerable entry and market pricing, with independent to political intervention and to the failures enforcement of rules on quality and operating in cost control that destroyed them in the behavior, as in efficient road-haulage markets first place. Moreover, although subsidies may in Europe and the United States. be required to sustain service levels and fare National protection appears to secure a aspirations, open-ended subsidies of a public "fair" share of traffic for the haulers of each operator will almost certainly pass the benefit of the national partners in a transit market. It of subsidies to managers and employees rather operates through quotas that reduce the use of than to passengers. vehicles and thus increase the costs. It is often For some time, the small and medium-size supported by enforcing regulated rates, which vehicles must be part of the structure, not deny the shipper the opportunity to shop least because few governments have the fiscal around for a better deal. The alternative is to Roads: Broadening the Agenda 225 combine free entry to the market with rigorous Dealing with Escalating Unit Costs enforcement of national safety and operational The recent escalating costs of roadwork can behavior rules in all countries. Regulatory sys- be attributed to rising input costs against a tems would combine strict quality control backdrop of growing demand for contracting, with liberal approaches to pricing and market which appears to have been exploited in an entry. Moving in this direction would include environment of generally low competition for developing internationally agreed strategies to contracts. No one solution exists. Inflation in improve the range of elements on the main input costs lies beyond the control of policy transit corridors and to strengthen enforce- makers, but they can take other measures. ment on overloading. A key issue is to ensure effective compe- Breaking the regulatory status quo in Cen- tition for contracts. Road agencies should tral and West Africa is difficult because of a actively market contracts to obtain a set of coalition of interest groups opposing change. good bidders. If at any stage in the bidding Truckers have strong leverage on high-level a competitive choice set of bidders does not authorities because they have enough monop- surface (say, at prequalification), something oly power to block trade. Governance issues is seriously wrong, and the agency should also intrude because some high-level authori- consider postponing the process until it has ties own or indirectly control trucks or truck- identified and corrected the underlying issues. ing companies and thus benefit from the status Continuing the bid without a proper choice quo and current market-sharing schemes. set in the hope of achieving an acceptable bid Deregulating the trucking industry in Cen- price is an unnecessary gamble. tral and West Africa is thus more of a politi- A better understanding of the underlying cal and social challenge than a technical one. cost trends and their links to contract pric- The main concern is the potential reduction ing is also critical. Although cost inflation in the number of trucks to match demand in lies beyond the control of sector authorities, road transport. That reduction could lead to they can increase the accuracy of the design a drop in trucking employment and profits, cost estimates, improve the allowance for because some companies (or owner-operators) cost fluctuations, and monitor cost increases would disappear and others would shrink, and through the procurement period. To this end, these social effects would need to be mitigated. agencies need to understand the cost struc- Some chance exists that the coalition of inter- ture underlying road contracts more clearly est groups opposing change in the transport and to track international price trends for key market in most countries in Central and West inputs over time. Africa might not resist reforms if compensation The capacity of project-executing agencies schemes pay, at least partly, for the social costs. also needs to be strengthened to support the The southern African international trans- timely implementation of contracts. Delays port market is a good model for the rest of often result from deficiencies in the planning the continent because it combines liberalizing and procurement of sector agencies, making entry with enforcing quality and load-control this a third area for attention. rules applicable to all operators. Operations Whatever the improvements to road agency to and from South Africa are governed by procurement processes, the unit costs of road bilateral agreements that provide for sharing infrastructure are likely to remain on an information on traffic development and define upward path, straining already limited sector the types of permits that can be issued. This budgets. Beyond measures to improve procure- system restricts the carriage of bilateral trade ment, considering how to design roads to keep to operators from the two countries concerned costs down is important. Overengineering of and prohibits cabotage.2 It does not establish roads--beyond the surface type needed for the quotas, however, and it allows rates to be deter- anticipated traffic volume--is an issue in parts mined by the market to enable direct contract- of Africa and represents a waste of resources ing between shippers and transporters and that should be avoided. Careful economic giving incentives to efficient operators. analysis of road investments can avoid the 226 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION overengineering of networks observed in some Above all, enforcement will have to be countries. In addition, experimentation with drastically improved. Eliminating corruption innovative technologies that keep costs down, in licensing, enforcing on-road behavior, and for example, by making greater use of locally inspecting and controlling vehicle conditions available materials, deserves consideration. are essential. Using technology to eliminate arbitrariness in implementation, together Improving Road Safety with carefully designed market incentives, has Governments recognize the seriousness of the worked well in privatizing vehicle inspections road safety problem. The February 2007 Pan- in Mexico. African Road Safety Conference, held in Accra, Ghana, resolved to set road safety as a national Notes health and transport priority. Areas for fund- The authors of this chapter are Kenneth Gwil- ing include (a) strengthening prehospital liam, Kavita Sethi, Alberto Nogales, and Vivien emergency services; (b) mainstreaming safety Foster, who drew on background material and design issues in road investment programs; contributions from Rodrigo Archondo-Callao, (c) collecting reliable road accident statistics; Fanny Barrett, Cecilia Briceño-Garmendia, and (d) enacting national legislation to deal Robin Carruthers, Arnaud Desmarchelier, Ranga Krishnamani, Ajay Kumar, Gael Raballand, Karlis with speeding, driving unroadworthy vehicles, Smits, and Supee Teravaninthorn. failing to use safety helmets, using mobile 1. These networks typically comprise a centrally phones when driving, and driving under the administered primary network plus secondary influence of alcohol. networks, but in Malawi, Nigeria, South Africa, For institutional arrangements, the choice and Uganda, only the primary network is cen- lies between establishing a special agency and trally administered and included here. broadly injecting safety skills and procedures 2. Cabotage is the provision of transport within a (such as safety audits on projects and policies) country by a foreign operator. in all relevant agencies. Involving transport, education, and health agencies is a minimum, References which probably requires at least a national Briceño-Garmendia, Cecilia, Vivien Foster, coordination agency, such as the National Hyoung Wang, Alvaro Federico Barra, and Transport Safety Committee in Ghana. To Ranga Rajan Krishnamani. 2009a. "Prioritizing Infrastructure Investments in the Democratic have authority, the agency needs to be directly Republic of Congo: A Spatial Approach." Report, responsible to the chief minister or the cabinet. Sustainable Development Department, Africa The urban equivalent would be a special unit Region, World Bank, Washington, DC. in the mayor's office. ------. 2009b. "Prioritizing Infrastructure For program composition, the choice lies Investments in the Republic of Congo: A Spa- between a sequence of consistent measures tial Approach." Report, Sustainable Develop- and a comprehensive "big bang" approach. ment Department, Africa Region, World Bank, Washington, DC. Evidence from various parts of the world suggests that the greatest success is through Briceño-Garmendia, Cecilia, Karlis Smits, and Vivien Foster. 2008. "Financing Public Infra- concentrated, multidimensional programs of structure in Sub-Saharan Africa: Patterns, action. For example, Japan turned a situation Issues, and Options." Background Paper 15, from disastrous to exemplary over a fairly short Africa Infrastructure Country Diagnostic, period by combining more stringent rules on World Bank, Washington, DC. vehicle condition, speeding, and drunk driv- Carruthers, Robin, Ranga R. Krishnamani, and ing with very high-profile publicity campaigns Siobhan Murray. 2008. "Improving Connec- and strict enforcement by traffic police. One tivity: Investing in Transport Infrastructure in Sub-Saharan Africa." Background Paper 7, of the most successful initiatives in Africa Africa Infrastructure Country Diagnostic, has been the comprehensive program intro- World Bank, Washington, DC. duced in the Richards Bay area of KwaZulu- Diou, Christian, Michel Henry, and Babaly Deme. Natal, based on a model already introduced in 2007. La Délégation de Maîtrise d'Ouvrage Victoria, Australia. en Afrique en 2007. Republic of Senegal, Roads: Broadening the Agenda 227 Public-Private Infrastructure Advisory Facility, Stankevich, Natalya, Navaid Qureshi, and Cesar and AFRICATIP. http://www.africatip.net/fr/ Queiroz. 2005. "Performance-Based Contract- publications/downloads/2008-11-11%2006: ing for Preservation and Improvement of Road 28:29/Rapport_AGETIP_MOD_vfinale.pdf. Assets." Transport Note TN-27, World Bank, Gwilliam, Ken, Vivien Foster, Rodrigo Archondo- Washington, DC. Callao, Cecilia Briceño-Garmendia, Alberto Starkey, Paul, John Hine, Simon Ellis, and Anna Nogales, and Kavita Sethi. 2008. "The Burden Terrell. 2002. "Improving Rural Mobility: of Maintenance: Roads in Sub-Saharan Africa." Options for Developing Motorized and Non- Background Paper 14, Africa Infrastructure Motorized Transport in Rural Areas." Technical Country Diagnostic, World Bank, Washington, DC. Paper 525, World Bank, Washington, DC. Harral, Clell, and Asif Faiz. 1988. Road Deteriora- Teravaninthorn, Supee, and Gael Raballand. 2008. tion in Developing Countries. Washington, DC: "Transport Prices and Costs in Africa: A Review World Bank. of the Main International Corridors." Working Kumar, Ajay, and Fanny Barrett. 2007. "Stuck in Paper 14, Africa Infrastructure Country Diag- Traffic: Urban Transport in Africa." Background nostic, World Bank, Washington, DC. Paper 1, Africa Infrastructure Country Diagnos- Van Zyl, Willem, Lynette Coetzer, and Chris tic, World Bank, Washington, DC. Lombard. 2008. "Unit Costs of Infrastructure Raballand, Gaël, Somik Lall, Arnaud Desmarchelier, Projects in Sub-Saharan Africa." Background and Patricia Macchi. 2009. "Economic Paper 11, Africa Infrastructure Country Diag- Geography and Aid Effectiveness in Transport nostic, World Bank, Washington, DC. in Sub-Saharan Africa." Report, Transport World Bank. 2007. "Road Maintenance Initiative Department, Africa Region, World Bank, Matrix." Sub-Saharan Africa Transport Program, Washington, DC. World Bank, Washington, DC. Chapter 11 Railways: Looking for Traffic A frican railroads have changed greatly stock. Moreover, various conflicts and wars in the past 30 years. Back in the 1980s, have rendered several rail sections unusable. many railway systems carried a large As a result, some networks have closed and share of their country's traffic because road many others are in relatively poor condition, transport was poor or faced restrictive regu- with investment backlogs stretching back over lations, and rail customers were established many years. businesses locked into rail either through Few railways are able to generate signifi- physical connections or (if they were para- cant funds for investment. Other than for statals) through policies requiring them to use purely mineral lines, investment has usually a fellow parastatal. Since then, most national come from bilateral and multilateral donors. economies and national railways have been Almost all remaining passenger services fail liberalized. Coupled with the general improve- to cover their costs, and freight service tariffs ment in road infrastructure, liberalization has are constrained by road competition. More- led to strong intermodal competition. Today, over, as long as the railways are government few railways outside South Africa, other than operated, bureaucratic constraints and lack dedicated mineral lines, are essential to the of commercial incentives will prevent them functioning of the economy. from competing successfully. Since 1993, sev- Rail networks in Africa are disconnected, eral governments in Africa have responded by and many are in poor condition. Although concessioning their systems, often accompa- an extensive system based in southern Africa nied by a rehabilitation program funded by reaches as far as the Democratic Republic of international financial institutions. Congo and East Africa, most other railways For the most part, concessions have are disconnected lines reaching inland from improved operational performance. Although the ports, serving small markets by mod- results have been mixed, many concessionaires ern railway standards. Most were built rela- have increased traffic volumes and have gener- tively lightly, and few, other than Spoornet ally performed more efficiently, and there has in South Africa, have invested in rehabilitat- been little evidence of monopolistic behavior. ing and renewing infrastructure and rolling Relations with governments have often been 229 230 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION uneasy, however, especially concerning ade- followed a similar pattern in almost all Afri- quate compensation for loss-making passen- can countries. Typically, isolated lines headed ger service obligations, and many governments inland from a port to reach a trading center clearly had unrealistic expectations about the or a mine, and over time, a few branch lines private sector's ability to improve operations were built. Many of the lines were state owned, and generate investment. but some were constructed as concessions or, Concessionaires appear willing to spend their in the case of some mineral developments, own funds only on day-to-day maintenance, not as part of a mining company's operation. on infrastructure. Financing asset renewal and Although continental rail master plans have upgrading remains an open question for most existed for over a century, most of the Afri- of the African rail network. Without infrastruc- can network remains disconnected, operating ture investments, the competition from road within a single country or linking a port and networks will thwart railway survival except to its immediate regional hinterland. The only carry large-scale mineral traffic. Although con- significant international network is centered cessioning has generally improved service and in South Africa and stretches north to Zim- reduced the financial burden on governments, babwe, Zambia, and the Democratic Republic it does not appear to be a full solution to financ- of Congo (figure 11.1). Trade between African ing the investment needs of African railways. countries (other than to and from South Africa) has always been minimal, largely because of the similarity in the products exported, which sug- Africa's Rail Networks gests that interregional links would be lightly used even if they existed. At the end of 2008, 47 railways were operating in 32 countries in Africa. Railway development has Low Rail and Traffic Density African railway networks' spatial density, a Figure 11.1 Map of African Rail Networks metric that compares track mileage with the size of a country, is low (UIC 2008).1 The highest measurement of spatial density is 16 in South Africa, but most other countries fall in the range of 1 to 6, and 13 countries have no operating railway at all. Too much should not be read into this indicator, however; network PORT SUDAN density is strongly affected by the pattern of population. Australia, Canada, China, and the DAKAR DJIBOUTI Russian Federation, all with vast undeveloped CONAKRY and sparsely populated areas, have densities LAGOS of between 5 and 7, whereas most European LOME PORT ABIDJAN TEMA HARCOURT DOUALA countries range from 20 to 100. A complementary indicator is the net- POINTE NOIRE MOMBASA work density per million inhabitants, which is LUANDA DAR ES SALAAM highest in Gabon (520) and Botswana (480), followed by South Africa (460). Most other African countries range from 30 to 50. Euro- TOAMASINA BEIRA pean countries range from 200 to 1,000, and WALVIS BAY Australia and Canada exceed 1,500. China is MAPUTO much lower, at 50. railway DURBAN These metrics alone cannot justify network major port CAPE TOWN EAST LONDON expansion in Africa. To be an economical PORT ELIZABETH investment, a new line needs a minimum level of traffic, and the geographical distribution of Source: Bullock 2009. potential customers within a country and the Railways: Looking for Traffic 231 Figure 11.2 Rail Network Size and Traffic by Region a. Regional share of network and traffic c. Passenger transport 100 14 90 12 80 passenger-km (billions) 70 10 60 percent 8 50 40 6 30 4 20 2 10 0 0 route-km passenger-km net ton-km southern Africa Central Africa East Africa West Africa southern Africa Central Africa East Africa West Africa nonurban suburban b. Route-kilometers in operation d. Freight transport 40 140 35 120 route-km (thousands) 30 net ton-km (billions) 100 25 80 20 60 15 40 10 5 20 0 0 southern Africa Central Africa East Africa West Africa southern Africa Central Africa East Africa West Africa general mineral general mineral Source: Bullock 2009. Note: Southern Africa = Angola, Botswana, Madagascar, Malawi, Mozambique, Namibia, South Africa, Swaziland, Zambia, and Zimbabwe; Central Africa = Cameroon, the Democratic Republic of Congo, Gabon, and the Republic of Congo; East Africa = Djibouti, Eritrea, Ethiopia, Kenya, Sudan, Tanzania, and Uganda; West Africa = Benin, Ghana, Guinea, Mali, Mauritania, Nigeria, Senegal, and Togo. level of usage that can be expected are more Traffic density on African railways is gen- important than these national averages. erally low.2 The highest average network South Africa has the most important net- traffic density outside Spoornet is in Gabon work (figure 11.2). Specialized mineral lines in (2.7 million traffic units), with Cameroon and western and southern Africa carry over half of Swaziland having the only other railways over the region's freight, most of it on the Spoor- 1 million; many railways average fewer than net coal and ore export lines. Southern Africa 300,000 units (figure 11.3). By comparison, the dominates general rail freight, handling over average traffic density of the Maghreb systems 80 percent of the freight traffic on the non- (Algeria, Morocco, and Tunisia) is nearly 2 mil- mineral lines. Southern Africa also dominates lion units, and the Arab Republic of Egypt, with the passenger business, with over 70 percent of its heavy passenger traffic, exceeds 8 million. passenger traffic, largely because of its heavy Most European systems average 2 million to commuter passenger business in cities. Some 5 million, with densities under 1 million found other African cities also operate commuter ser- only in Albania and Montenegro. With such vices, but with the exception of Dakar, Senegal, light usage, many networks struggle to gener- they mostly provide one or two trains at peak ate enough funds just to maintain, much less hours along a short line. renew, their infrastructure. 232 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 11.3 Average Railway Network Traffic Density, 2001­05 low-speed, small-scale, undercapitalized net- works ill suited to modern requirements. Many Madarail, Madagascar structures and some of the track work are now CEAR, Malawi over 100 years old. Many sections of track have CDN, Mozambique deteriorated almost beyond repair. Although CCFB, Mozambique this situation can be tolerated on low-volume RVRC-URC, Uganda feeder lines, and indeed may be the only way concessioned Transrail, Mali some can be viably operated, it is a major RSZ, Zambia handicap when competing against the modern Sitarail, Burkina Faso and Côte d'Ivoire roads being constructed in major corridors. Most rail systems have considerable sections Camrail, Cameroon of track in need of repair or replacement. Some TRC, Tanzania have major sections that are not in operation RVRC-KRC, Kenya and will require rehabilitation before operations SETRAG, Gabon can resume. Even where service exists, poor track condition forces speed restrictions, resulting in FCE, Madagascar lower railway competitiveness and rolling-stock NRC, Nigeria productivity. CFMK, Dem. Rep. of Congo In some countries, parts of the network are OCBN, Benin not operated because of war damage, natu- ral disaster, or general neglect. Much of the CDE, Ethiopia Mozambican central and northern networks GRC, Ghana and railways in Angola, Côte d'Ivoire, Eritrea, state owned CFCO, Rep. of Congo Ethiopia, and the Republic of Congo either CFM, Mozambique have been damaged or have had to suspend SNCC, Dem. Rep. of Congo operations for as long as 20 years. The total SR, Swaziland African rail network is about 69,000 kilometers, BR, Botswana of which some 55,000 kilometers is currently TAZARA, Tanzania being operated (see figures 11.1 and 11.2). SRC, Sudan Almost all the network is single track, except for sections of the Spoornet network. Much of Transnamib, Namibia the South African network is electrified, but the 0 500 1,000 1,500 2,000 2,500 3,000 3,500 only other electrified sections in Sub-Saharan traffic units per route-km (millions) Africa are in the mining region of the Demo- passenger freight cratic Republic of Congo and a short section in Zimbabwe (the latter is not in use). Source: Bullock 2009. Note: The overall traffic units carried by a railway are the sum of the passenger-kilometers and the net Signaling on many networks still relies on tonne-kilometers of freight carried. This simple standard measure is widely used as a means of aggregat- ing freight and passenger traffic. The relative weighting of passenger and freight is conventionally taken manual systems. On lines with low train den- as 1:1. BR = Botswana Railways; Camrail = Cameroon Railway Corporation; CCFB = Companhia dos sity, mechanical signals are adequate from Caminhos de Ferro da Beira (Mozambique); CDE = Chemin de Fer Djibouto-Ethiopien; CDN = Corre- dor de Desenvolvimento do Norte (Mozambique); CEAR = Central East African Railways Corporation a capacity viewpoint, but significant safety (Malawi); CFCO = Chemin de Fer Congo-Océan (Republic of Congo); CFMK = Chemin de Fer Matadi- problems can result from human error. Where Kinshasa (Democratic Republic of Congo); CFM = Caminhos de Ferro do Mocambique; FCE = Fianarant- soa Côte Est (Madagascar); GRC = Ghana Railways Corporation; NRC = Nigeria Railways Corporation; power signaling has been installed, it often OCBN = Organisation Commune Bénin-Niger; RSZ = Railway Systems of Zambia Ltd; RVRC-KRC = Rift Valley Rail Corporation-Kenya Railways Corporation; RVRC-URC = Rift Valley Rail Corporation-Uganda does not operate because of short circuits, Railways Corporation; SETRAG = Société Transgabonnaise (Gabon); SNCC = Société Nationale des lack of electrical power, and dilapidated cable Chemins de Fer du Congo (Democratic Republic of Congo); SR = Swaziland Railways; SRC = Sudan Railways Corporation; TAZARA = Tanzania-Zambia Railway; TRC = Tanzania Railways Corporation. networks. Telephone exchanges in many com- panies are similarly obsolete, with limited Dilapidated Infrastructure capacity and the need for spare parts that are Most networks outside South Africa still virtually impossible to find. operate with their original facilities. Limited Most African railways use either the Cape upgrading has occurred, but the lines can still gauge (1.067 meters or 3 feet, 6 inches) or the be characterized as relatively low axle-load, meter gauge. The main network in southern Railways: Looking for Traffic 233 and central Africa uses the Cape gauge, which Traffic--Low and Growing Slowly is also used in some anglophone countries far- Outside South Africa, the traffic volumes ther north. The meter gauge is used in most of serviced by African railways are very small; francophone Africa and much of East Africa. about half of the 26 railway operators sur- A number of isolated standard-gauge lines are veyed carried traffic of less than 500,000 used primarily for mineral traffic, although traffic units annually, while only 5 of them Nigeria is developing a new standard-gauge exceeded 1 million traffic units annually--a network to serve its capital, Abuja. Narrow- volume comparable to a moderately busy gauge lines have operated at various times, branch line on other railways (figure 11.4). but most are now derelict. Apart from the By comparison, Spoornet in South Africa network in East Africa and the one extending carries 1 million traffic units every three days north from South Africa, few railways cross (Thompson 2007). In some cases, the light international borders. Instead, they reach rail- traffic is caused by a lack of demand; in oth- heads from which traffic can be carried farther ers, it is caused by shortages of rolling stock, by road. particularly locomotives. Despite the multiplicity of gauges, interop- Although the average haul on African net- erability is not a major problem in Africa. Two works is relatively long with regard to their gauges exist in the same location in only three size, it is not especially so vis-à-vis road trans- places--two in Tanzania and one in Guinea. port. Some railways carry mostly end-to-end However, mixed gauges will become a problem traffic; Tanzania Railways Corporation, Tazara if some of the proposed connecting lines are (Tanzania-Zambia Railway Authority), and constructed. Transrail (Dakar-Bamako Railway) all haul In summary, most African railways are con- freight an average distance of 1,000 kilometers, fronting major infrastructure problems pri- and some smaller railways, such as Uganda marily associated with aging track: insufficient Railway or CEAR (Central East African Rail- ballast, rail wear, deteriorating earthworks and ways), act as feeders to other systems, which formation, decrepit structures, and rail signal- carry the traffic a few hundred kilometers ing and telecommunications with obsolete farther. These systems have a good chance of equipment and lack of spare parts. The cost of competing for general freight traffic, even as a rehabilitating the networks is large compared road network improves, as long as satisfactory with the existing traffic volumes and revenues. service levels can be achieved, but the shorter The means by which rehabilitation can be systems that require transshipment to road at done on a sustainable basis is the central ques- railheads will generally find they can compete tion faced by most African railways. effectively only for bulk traffic. Most systems operate only limited passen- ger commuter services, if any, and the aver- The African Rail Market age distance of passenger trips is the distance between the capital of a country and major Typically railways in Africa are small, carrying provincial centers. The only significant cross- no more traffic than a moderately busy branch border flows are on the Sitarail (Côte d'Ivoire), line in other parts of the world. African rail- Tazara, and Transrail networks. ways carry far more freight than passengers, Since the mid-1990s, most African coun- with freight averaging about 80 percent of tries experienced steady economic growth. traffic between 1995 and 2005. Almost all rail- Average annual GDP grew 4 percent, with cor- ways carry passenger traffic; only Swaziland responding increases in trade. Per capita GDP and Uganda have freight-only railways. The grew by about 1.5 percent a year. Countries passenger business is steadily shrinking, how- such as Mali, Mozambique, and Tanzania that ever, and several of the railways still retaining avoided political upheaval grew as much as a reasonable passenger business do so only 50 percent faster. Despite the generally favor- because competing road networks are in poor able economic background, only four African condition or do not exist. railways increased both their passenger and 234 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 11.4 Average Railway Traffic Volumes, 2001­05 have been concessioned, freight traffic has gen- erally increased, whereas passenger traffic has Madarail, Madagascar generally stagnated or declined. CEAR, Malawi The growth or decline of traffic on many CDN, Mozambique systems over the last decade often had little to CCFB, Mozambique do with changes in the underlying demand. RVRC-URC, Uganda War or natural disaster has had a major effect concessioned Transrail, Mali in some cases; on other railways, the volume RSZ, Zambia carried reflects the availability of rolling stock, Sitarail, Burkina Faso and Côte d'Ivoire particularly locomotives. Many railways are short of locomotives. When this situation Camrail, Cameroon improves with new or secondhand locomo- TRC, Tanzania tives or through a locomotive rehabilitation RVRC-KRC, Kenya project, traffic will increase accordingly. SETRAG, Gabon Passenger Services--in Decline FCE, Madagascar Several African cities have announced plans to NRC, Nigeria introduce modern heavy-rail suburban com- CFMK, Dem. Rep. of Congo muter networks. Such services are currently OCBN, Benin limited to South Africa and Dakar, Senegal. CDE, Ethiopia Experiences elsewhere in the world suggest that any new services will need substantial GRC, Ghana external financial support for both capital and state owned CFCO, Rep. of Congo recurrent operating costs and should be oper- CFM, Mozambique ated by new independent transport authori- SNCC, Dem. Rep. of Congo ties. Almost all other passenger services face SR, Swaziland strong competition from buses and shared BR, Botswana taxis in both price and service frequency, and TAZARA, Tanzania few corridors remain in which rail passenger SRC, Sudan services are the only means of transport. Bus Transnamib, Namibia fares are typically about 30­50 percent higher than the economy rail fare, but on most routes 0 0 0 0 0 0 0 0 0 0 0 buses are faster (sometimes twice as fast) and 20 40 60 80 00 20 40 60 80 00 1, 1, 1, 1, 1, 2, traffic units (millions) more frequent. Buses have the lion's share of net tonne-km passenger-km the market, although they suffer from the same problems as rail: unreliable departures, delays Source: Bullock 2009. and breakdowns, and overcrowding. Note: Traffic units are passenger-kilometers in the case of passenger traffic and net tonne-kilometers in the case of freight traffic. BR = Botswana Railways; Camrail = Cameroon Railway Corporation; CCFB = The long-term prospects for nonurban Companhia dos Caminhos de Ferro da Beira (Mozambique); CDE = Chemin de Fer Djibouto-Ethiopien; rail services are generally poor (Amos and CDN = Corredor de Desenvolvimento do Norte (Mozambique); CEAR = Central East African Railways Corporation (Malawi); CFCO = Chemin de Fer Congo-Océan (Republic of Congo); CFMK = Chemin de Bullock 2007). Rail services start competing Fer Matadi-Kinshasa (Democratic Republic of Congo); CFM = Caminhos de Ferro do Mocambique; FCE = Fianarantsoa Côte Est (Madagascar); GRC = Ghana Railways Corporation; NRC = Nigeria Railways with roads at speeds higher than 70 kilometers Corporation; OCBN = Organisation Commune Bénin-Niger; RSZ = Railway Systems of Zambia Ltd; RVRC- per hour. However, the cost of maintain- KRC = Rift Valley Rail Corporation-Kenya Railways Corporation; RVRC-URC = Rift Valley Rail Corporation- Uganda Railways Corporation; SETRAG = Société Transgabonnaise (Gabon); SNCC = Société Nationale ing track and signaling systems that would des Chemins de Fer du Congo (Democratic Republic of Congo); SR = Swaziland Railways; SRC = Sudan enable these commercial speeds is signifi- Railways Corporation; TAZARA = Tanzania-Zambia Railway; TRC = Tanzania Railways Corporation. cantly more than the cost of maintaining the 30- to 40-kilometer-per-hour commercial freight traffic over the period, two of which speed needed for a freight railway. In addi- had been concessioned. One other railway saw tion, a very large capital investment would an increase in average passenger traffic, and be required to construct new medium-speed all others saw a reduction. Fifteen railways (for example, 200 kilometers per hour) inter- increased their freight traffic. Where railways urban railways. Such investment is justified Railways: Looking for Traffic 235 only on the basis of substantial demand (sev- differences in the commodity mix, with many eral million passengers a year) and relatively requiring specialized cars, mean freight trains high-income passengers who can afford to are rarely fully loaded in both directions. In cover at least operational costs. Few, if any, some cases, this natural imbalance in traffic is corridors in Africa could justify such invest- accentuated for rail because road vehicles deliv- ments, at least for the medium term. ering imports tend to backload freight at mar- Formal compensation schemes, such as pub- ginal cost, leaving rail to transport the remaining lic service obligations, have been introduced in freight without a compensating return load. a few cases to support passenger rail services, Average freight tariffs range from $0.03 to but they rarely provide timely compensation $0.05 per net tonne-kilometer, similar to tariffs for service operations. Payment may be delayed on other general freight railways in compara- several years or may otherwise take the form ble countries. Tariffs are generally constrained of a subsidy calculated to break even, limiting by competition, either from road or alternate the ability of railways to increase their main- routes (particularly in the Great Lakes region, tenance and negating any attempts to improve Malawi, West Africa, and Zambia) and are also the financial performance of the freight ser- influenced by the traditional value-based tar- vices. As a result, most long-distance passenger iff structures, the relative cost of carrying dif- services in Africa are trapped in a cycle of mini- ferent commodities (as reflected in net tons mal investment, deteriorating services, declin- per railcar round-trip), direction of travel, ing patronage, and financial losses. and volume. Although most rail rates are well The few instances in which local trains serve below comparable road rates, especially for villages with no road connection pose a differ- containers, rail typically carries only 20­50 ent problem. These trains are used by traders percent of the traffic in a corridor, and some bringing goods to and from regional centers, of the smaller state-owned railways have an and although heavily loaded with passengers, even smaller share. they nonetheless incur major losses. Although Line-haul tariffs are only part of the cost such services can be funded through govern- equation for freight traffic. Much is often ment subsidy, the long-term solution is to cre- made of the inherent lower cost of rail com- ate feeder roads for motorized access, enabling pared to road. This is true where minerals more cost-effective means of transporting must be transported from a rail-connected goods and greatly improving accessibility to mine to a rail-connected port but is not so such locations. clear for medium-distance general freight that also must be transported by road to and from Freight--Needs Improving railheads. Haulage between the railway and Freight traffic on railways is mostly bulk and the ultimate origin and destination can be semibulk commodities, principally to and from surprisingly expensive, often as much as the ports. The actual commodities transported by equivalent of 200­300 kilometers of line-haul rail reflect the economic structure of countries transport, negating any advantage rail may served by the railway, with mining products have in pure line-haul tariffs. New sidings are important in several countries and timber and sometimes constructed, but they need a certain export crops important in West Africa. Imports amount of traffic to be economical. Traffic that are mostly manufactures, such as cement and needs to be collected at a central depot before petroleum products, and general freight. On being dispatched by rail is more vulnerable to some systems, much of the general freight is road competition, and even bulk traffic is not containerized (cash crops with high value are immune if distances are not too long. In many increasingly traveling this way), particularly countries, collection and distribution chains when the trip involves crossing an intermedi- are being streamlined, often eliminating up- ate border before reaching the port. Unlike pas- country depots and distribution centers, and senger services, significant imbalances between marketing channels have become more diver- traffic in the two directions are common. Even sified. The railways have often been slow to where tonnage is approximately balanced, the respond, steadily losing market share. 236 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Level of service is a key factor in the freight to merely a line-haul operation, can reduce the business. For rail to play a significant role in price discount between rail and road, increas- the general freight transport system, it must ing the contribution that freight can make to improve its service (specifically, overall tran- the maintenance and renewal of infrastructure. sit time, reliability, security, and service fre- This improvement is one of the major benefits a quency) and ensure that it is addressing the concessionaire can offer a state-owned railway. needs of customers. Too often, what rail has Moreover, because of the lack of inter- offered as transport has been quite different connection services and cross-border service from what the competing road hauler can contracts, rail freight suffers huge delays in offer, and road carriers can charge a significant crossing national borders. For example, a premium. In general, freight markets in Africa rail freight journey of 3,000 kilometers from require reliable services (a commercial speed Kolwezi on the Democratic Republic of Congo of 40 kilometers per hour is usually sufficient) border to the port of Durban in South Africa rather than high-speed services, with (a) rail takes 38 days to complete, an effective speed infrastructure and rolling stock maintained for of only 4 kilometers per hour. Only 9 of these service, (b) operating discipline to ensure that days are spent traveling, with the remainder schedules are maintained, and (c) commercial (a staggering 29 days) taken up primarily with arrangements that ensure that customers fulfill loading and interchanging freight, as well as their contractual responsibilities. some time for customs clearance. Each day of Most railways can win bulk mineral traffic delay costs $200 per railcar. The main cause of when it is offered, but general freight requires the problems in the rail sector is the absence a reasonable level of service from rail if it is of reliable interconnection services when to compete with road without offering a sig- trains cross borders. Locomotives from one nificant price discount. By 2025, any remain- country are currently not allowed to travel on ing monopolies for general freight will have another country's network, mainly because of run their course, and the only traffic on which the inability to provide breakdown assistance African railways will have an undisputed grip to foreign operators. As a result, rail freight will be minerals (although mining compa- crossing borders must wait to be picked up nies are increasingly running even this traf- by a different locomotive. The delays are often fic directly, either as third-party operators or extensive, partly because of the lack of reli- on their own private networks). Experience in able, well-maintained locomotives. Delays also many countries has shown that general freight reflect the lack of clear contractual incentives transport requires operators to be flexible, to service traffic from a neighboring country's responsive, and adaptable. Fewer custom- network. Reducing such delays would there- ers are fellow parastatals under order to use fore require totally rethinking the contractual a state-owned railway, and few government- relationships and access rights linking the rail- owned organizations, no matter how cor- ways along the corridor. It would also likely poratized they may be, have the commercial require the establishment of a regional clear- freedom to operate effectively in a fully com- inghouse to ensure transparency and fairness petitive environment. in reciprocal track access rights. Rail in Africa must become a transport busi- ness in the broadest sense and must be able to adapt to new markets. The predicaments of the How Much Investment Can Be remaining government-owned railways, how- Justified? ever, show that rail cannot compete effectively while it is handicapped by the bureaucratic con- Providing an estimate of the investment needed straints and lack of commercial incentives and by African railways is a daunting task (Carruthers, accountability of a government organization. Krishnamani, and Murray 2009). In addition to Achieving an acceptable level of service, com- building detailed inventories and assessments bined with flexible pricing policies and a strat- of infrastructure and determining how much egy of providing a transport service as opposed needs to be repaired or replaced, the question of Railways: Looking for Traffic 237 how much investment is economically justified by using assumed average asset lives. Excluding must be asked. Lines that have been superseded South Africa, the Sub-Saharan network carries by road developments and those with low traffic about 15 billion net tonne-kilometers a year, levels will rarely merit reconstruction and invest- excluding the mineral lines, and about 4 billion ment, and funds should instead be directed to passenger-kilometers. That level of traffic will those parts of the network with long-term value. require, on average, replacing 500 freight cars, Although a government's desire to reinstate 20 passenger cars, and about 20 locomotives such links is understandable, doing so is often a year. As with infrastructure, much of that extremely expensive. stock will be secondhand (from India or South Investment has historically been used Africa), but the estimated cost will still average for new construction and rolling stock, for about $80 million a year, equivalent to about replacement of rolling stock, and sometimes $0.04 per net tonne-kilometer or passenger- for rehabilitation and replacement of track. kilometer. The steady-state investment in the Long-term maintenance neglect has caused a African network north of South Africa should huge backlog investment of up to $3 billion thus be about $200 million a year (allowing for Africa's railways. In practice, this one-time $20 million for facilities, maintenance, equip- expenditure needed to eliminate the rehabili- ment, and other costs). tation backlog could be spread over a 10-year That amounts to a combined annual pro- period at an annual rate of $300 million. gram of about $500 million for 10 years, after After the network is restored to good con- which investment would drop to the steady- dition, the annual bill would fall substantially state level of $200 million (Bullock 2009). The to cover only what was needed for ongoing $500 million a year requirement refers to the track rehabilitation and renewal. Excluding period during which the rehabilitation back- South Africa, the Sub-Saharan network con- log is being cleared. These calculations are only sists of about 44,000 kilometers of track, of broad order-of-magnitude estimates. However, which about 34,000 kilometers is operational. the amount needed to overcome these prob- The infrastructure on this network will have lems is large, equal to the annual revenues of a life of at least 40­50 years, given the gener- some of the railways and well beyond their ally low traffic volumes; the cost of periodic capacity to self-finance. The only option in reconstruction (about $350,000 per kilome- most cases is to seek large concessional loans ter) is thus equivalent to an annual cost of or grants from third parties. about $8,000 per kilometer. Few lines with In addition to reinvestment in the cur- an average density of fewer than 1 million rent network, investment in new projects is net tons a year are likely to warrant this kind a possibility. For years, proposals have been of major rehabilitation expenditure, because floated to create new routes for landlocked traffic would need to earn $0.08 per net countries and to integrate the isolated net- tonne-kilometer to fund the reconstruction, works. The most comprehensive proposal was whereas typical rail freight tariffs are no more the 1976 master plan of the Union of African than $0.05 per net tonne-kilometer. Lines with Railways for a pan-African rail network that a density under 250,000 tons a year probably included 26,000 kilometers of new construc- cannot support anything more than routine tion. Designed to create a grid to support maintenance. Even if low-volume lines are intra-African trade development and regional reconstructed using cheaper, secondhand economic integration, the plan was approved materials, this level of expenditure is unlikely by the Organization of African Unity in 1979, to be justified for more than 20,000 kilometers but few, if any, of the proposed links have of the network. Overall, the ongoing annual gone beyond the drawing board. The Union cost of track reconstruction would thus aver- of African Railways is now concentrating on age approximately $100 million a year. a revised plan containing a subset of 10 cor- Sustaining an adequate fleet of rolling stock ridors, some of which are already partially will cost an additional $80 million a year. The constructed, and the proposal has generated a cost of replacing rolling stock can be estimated number of regional studies and action plans. 238 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Several proposals for individual segments have countries suffering or recovering from civil been made, and mining companies have pro- disruption (Angola, the Democratic Republic posed a number of dedicated mineral lines. of Congo, and Zimbabwe), most countries are Few of these projects will be financially or at various stages of reform. Of the 30 African economically viable. The cost of new construc- countries with publicly owned railways, 14 have tion of a single-track, nonelectrified railway on opted for a concession arrangement and 1 oper- relatively flat terrain is at least $1.5 million per ates under a management contract (figure 11.5). kilometer, increasing to about $5 million in Four others have begun the process. more rugged country. In many cases, the pro- posed new routes would compete with existing Concessions--Becoming the Norm road and rail routes, which would constrain The introduction of concessions has required the rates that typically could be charged to at substantial changes in the legal and regulatory most $0.05 per net tonne-kilometer. In the case framework in many countries. In the franco- of export mineral traffic, the potential rate is phone countries, concessions can generally be generally constrained to about $0.02­$0.03 per done within the existing legal system, but most net tonne-kilometer by the long-term delivered anglophone countries have had to amend their market price. Because a serviceable two-lane railway acts. Arrangements have also been road can generally be constructed for approxi- made for the economic and safety regulation mately $1 million per kilometer, the additional of concessions, and new government bodies rail investment would be economically justified have been established to own the assets leased only if expected traffic was at least 2 million­ to the concessionaires. 4 million tons a year. If the capital costs of the Those railways that have not been conces- infrastructure do not have to be recovered, the sioned remain subject to significant political lines can probably be operated successfully at and governmental influence. Arrangements 0.5 million­1.0 million ton. vary across countries, but the sectoral minis- try (normally transport) exercises political and administrative control, while the ministry of Institutional Arrangements and finance exercises financial control. Board direc- Performance tors are generally a combination of ministry officials and internal senior management, who Until the 1980s, almost all African railway are often appointed by the government. Over- companies were publicly owned corporations, sight is nominally assigned to the parliament, with varying degrees of financial and manage- but in practice such control may be limited to ment autonomy. Attempts at commercializa- an audit of the company accounts in its annual tion while retaining public ownership were report (often several years in arrears). Although generally unsuccessful, and concessions were the governing regulatory frameworks nominally introduced in the 1990s. Under concessional provide financial and management autonomy, arrangements, the state remains the owner in practice this arrangement is considerably of all or some of the existing assets, typically limited by the many opportunities for state the infrastructure, and transfers the other intervention permitted under the legal and assets (normally the rolling stock) and the regulatory frameworks at both the institutional responsibility to operate and maintain the rail- and jurisdictional levels. This conflict between way to a concessionaire. the control and decision functions, as well as Most countries in Central, East, and West frequent reviews by political authorities of ini- Africa have moved all or part of the way to con- tiatives taken by the government's authorized cessioning, often under the pressure of multi- representatives in the corporation, discourages lateral and bilateral organizations that have management initiative and effectiveness. until recently been the only source of large The first railways to be concessioned were loans for asset rehabilitation and renewal. With in West Africa, beginning in 1995 with the the exception of southern Africa (Botswana, Sitarail concession linking Burkina Faso and Namibia, South Africa, and Swaziland) and Côte d'Ivoire and followed in the late 1990s by Railways: Looking for Traffic 239 Cameroon, Gabon, and Malawi. The reform Figure 11.5 Private Participation in African Railways since 1990 momentum accelerated in the 2000s, but Sizarail implementation has often been a slow process, Transrail Dem. Rep. of Congo Senegal/Mali 1995­97 typically taking three to five years, sometimes 2003 2008 much longer. RVRC Most African networks leave little room Sitarail Kenya/Uganda Côte d'Ivoire/Burkina 2006 for competition, and few governments have Faso seriously considered the European model of 1995 full vertical separation. However, third-party CANAC/WACEM TRC Togo Tanzania operators run on government lines in Kenya 1995/2002 2007 Camrail and Senegal, and a through freight service has Cameroon Transgabonaise CEAR Gabon Malawi operated for some years from South Africa to 1999 1999 1999 Tanzania. Concessions do not always include railway operated by state RSZ Madarail Madagascar the entire network, with lightly used branch railway company Zambia 2003 private sector participation 2003 lines sometimes excluded. projects planned or under way The initial duration of concessions varies part of rail network now BBR CDN under private management Zimbabwe Ressano Garcia Mozambique from 15 to 30 years, and the concessionaire 1997 Mozambique 2005 railway now under private canceled CCFB is free to operate its activity as a business, with management Mozambique freight tariffs generally determined by supply 2005 and demand, and passenger fares subject to Source: Bullock 2009. some form of indexation. Formal regulatory Note: BBR = Beitbridge Bulawayo Railway; CANAC/WACEM = CANAC Railway Services Inc./ West African Cement; CCFB = Companhia dos Caminhos de Ferro da Beira; CDN = Corredor de Desenvolvi- structures with real teeth are rare in Africa, mento do Norte; CEAR = Central East African Railways Corporation; RSZ = Railway Systems of Zambia and many rail concessions are potentially Ltd; RVRC = Rift Valley Rail Corporation; TRC = Tanzania Railways Corporation. open to market abuse, even though conces- sion agreements generally include some pro- tection, at least on paper. For example, the have been terminated (for example, Ressano Zambian rail concessionaire flagrantly price- Garcia in Mozambique), and two conces- discriminates by charging freight tariffs of sions (Transrail and Rift Valley) changed the $2.00 per tonne-kilometer on transit traf- operator. fic from the Democratic Republic of Congo Rail concessions in Africa have attracted a to Dar es Salaam, Tanzania, while charging limited pool of mostly foreign private opera- only $0.05 per tonne-kilometer on other tors. These operators fall into two distinct freight. The reason is to divert the Democratic groups: (a) those seeking vertical integration Republic of Congo traffic southward toward of the distribution chain by acquiring domi- the port of Durban in South Africa and over nant positions in specific production and the Beit Bridge, which the same concessionaire transport sectors, and (b) those specializing in operates. As a result, most of the Democratic a single transport activity (such as railways or Republic of Congo's copper exports end up ports). The business cases for these rail invest- going to Durban by road. ments often appear weak, however, suggesting A number of consumer protection devices that the companies that seek these conces- exist, but they are rarely invoked. The two most sions focus on the financial benefits that can common protections are (a) the power to refer be extracted from managing large investment rail tariffs to either the government or an inde- plans (financed for the most part by govern- pendent authority and (b) the power to allow ments) rather than concentrating on business third-party operators onto the railway to com- cash flows. pete with the concessionaire. Where a conces- Private companies are the majority share- sionaire fails to comply with the terms of the holders in all concessions to date. State partici- concession, whether by design or by force of pation is highest in Mozambique, which holds circumstance, procedures exist for terminating 49 percent of both CCFB (Companhia dos the concession. These procedures have rarely Caminhos de Ferro da Beira­Mozambique) been applied. Only one or two concessions and CDN (Corredor de Desenvolvimento do 240 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Norte­Mozambique) and is also a significant productivity is similarly low, with the source shareholder in the adjacent CEAR conces- generally being low availability caused by a sion. In Madagascar, the government holds lack of spare parts. 25 percent of Madarail, while governments Labor and asset productivity have improved own 10­20 percent in Abidjan-Ouagadougou steadily in most concessions, typically doubling Railway (Sitarail), Dakar-Bamako Railway because of workforce reductions either before (Transrail), and Cameroon Railway Corpo- or at the time of concessioning, the scrapping ration (Camrail). Local private participation of obsolete rolling stock, and increased traffic in concessions has generally been relatively volumes (figure 11.7). low and is often fraught with problems Safety is also an important aspect of opera- during the bidding process. Employee share- tional performance. Rail travel is still safer holding remains under 5 percent where it than road travel, but rail's record in Africa is exists at all. much worse than that of comparable railways elsewhere, caused by obsolete track infrastruc- Operational Performance-- ture, poorly maintained rolling stock, and lack Concessioning Helps of operational discipline. As with productivity, Both labor productivity and asset productivity however, safety has generally improved follow- (locomotive and railcar use) are low in most ing concessioning. African networks, compared with railways elsewhere, because of the poor condition of Financial Performance--Generally the infrastructure and rolling stock, low traf- Unsustainable fic levels, and government ownership. Under Most state-owned railways in Africa just about concessions, however, these indicators have break even cashwise after receiving govern- improved sharply, partly because of growth in ment support. Often, this balance occurs only traffic but mostly from major reductions in because a significant amount of maintenance the workforce. has been deferred; when the maintenance back- Since about 1990, almost all railway com- log becomes too great, it is typically addressed panies have streamlined their workforces. This by a loan that is treated as investment. The measure has often been the prelude to conces- two companies that have been concessioned sioning, but in some cases, it has also been a the longest (Camrail and Sitarail) make mod- general policy to improve efficiency. Still, labor est operating profits. The performance of RSZ productivity on most African systems is rela- (Railway Systems of Zambia) is unknown, and tively low by world standards, with few railways the cases of Kenya and Tanzania are too early achieving over 500,000 traffic units per staff a to judge. year, compared with an average 3.3 million Passenger services generally do not con- traffic units per staff a year for the South Afri- tribute significantly to the cost of maintaining can operator Spoornet (figure 11.6). This low infrastructure or to covering corporate over- productivity not only reflects the continuing head. In a few cases, they cover their marginal use of labor-intensive methods with relatively costs (train crew, rolling-stock maintenance, little outsourcing, but it is also the consequence fuel or traction electricity, and passenger- of a decline in traffic without adjustments to handling costs). Passenger tariffs on many staff levels. With low wages, the direct financial railways are essentially regulated, often within impact is not always catastrophic, but having a a framework that includes only a subset of large number of underemployed staff members total costs. However, many of the poorer per- corrodes morale and is a strong disincentive forming systems in Africa would be unable to for those who wish to improve efficiency. An cover above-rail working expenses on a sys- important effect is that railways have difficulty temwide level even if they could set their own recruiting and retaining technically competent tariffs. staff or introducing the technology required to Freight services normally cover their avoid- improve service levels, for which a better-paid able operating costs. Some also earn enough and more skilled workforce is essential. Asset to cover infrastructure costs and even capital Railways: Looking for Traffic 241 costs for rolling stock. Earnings are a function Figure 11.6 Labor Productivity on African Rail Systems of the tariff rate and the average carload on Madarail, Madagascar the revenue side, and factors such as train size, commercial speed, and rolling-stock use and CEAR, Malawi availability on the cost side. In general, freight CDN, Mozambique can earn enough to make operating services CCFB, Mozambique worthwhile, but only in some cases can it fund RVRC-URC, Uganda concessioned replacement of rolling stock, and very rarely Transrail, Mali can it earn enough to finance infrastructure RSZ, Zambia renewal. Sitarail, Burkina Faso and Côte d'Ivoire Where railways have been concessioned, Camrail, Cameroon low-interest sovereign loans to concessionaires have usually made a substantial contribution TRC, Tanzania to the financing of investments. Concession- RVRC-KRC, Kenya aires provide a relatively low proportion of SETRAG, Gabon the equity. Most plan to finance over 80 percent of their investment with debt, and the share FCE, Madagascar of the privately financed investments is in NRC, Nigeria many cases well below 50 percent. Conces- CFMK, Dem. Rep. of Congo sions that planned a substantial contribu- OCBN, Benin tion from commercial borrowing have faced CDE, Ethiopia consistent criticism for their lack of invest- ment in practice. Because the value of the GRC, Ghana state owned rolling stock transferred to the concessionaire CFCO, Rep. of Congo more than compensates for the equity put into CFM, Mozambique the concessions in most cases, the result is a SNCC, Dem. Rep. of Congo significant transfer of the financial risks asso- SR, Swaziland ciated with infrastructure investment from the BR, Botswana private sector to the public sector. The busi- TAZARA, Tanzania ness fundamentals of many concessions are SRC, Sudan insufficient to support major investment on a Transnamib, Namibia commercial basis, and they are all too prone to significant liquidity problems. Major asset 0 500 1,000 1,500 2,000 2,500 maintenance and reinvestment are thus likely traffic units/employee (thousands) to be problems. Source: Bullock 2009. Concessions normally pay the government Note: Common averages have been used for Kenya and Uganda, which are included in a single concession, and for Nacala (Mozambique) and Malawi, which share common resources. concession fees as well as a series of taxes (for The overall traffic units carried by a railway are the sum of the passenger-kilometers and the net example, value added tax, personnel social tonne-kilometers of freight carried. This simple standard measure is widely used as a means of aggre- gating freight and passenger traffic. The relative weighting of passenger and freight is conventionally taxes, income tax), often of the same order taken as 1:1. BR = Botswana Railways; Camrail = Cameroon Railway Corporation; CCFB = Companhia of magnitude. Given the relative size of taxes dos Caminhos de Ferro da Beira (Mozambique); CDE = Chemin de Fer Djibouto-Ethiopien; CDN = Corredor de Desenvolvimento do Norte (Mozambique); CEAR = Central East African Railways Corpo- (largely income tax) and concession fees, ration (Malawi); CFCO = Chemin de Fer Congo-Océan (Republic of Congo); CFMK = Chemin de Fer Matadi-Kinshasa (Democratic Republic of Congo); CFM = Caminhos de Ferro do Mocambique; governments should consider the combined FCE = Fianarantsoa Côte Est (Madagascar); GRC = Ghana Railways Corporation; NRC = Nigeria effect of both revenue streams when negoti- Railways Corporation; OCBN = Organisation Commune Bénin-Niger; RSZ = Railway Systems of Zambia Ltd; RVRC-KRC = Rift Valley Rail Corporation-Kenya Railways Corporation; RVRC-URC = Rift Valley Rail ating a concession. Regardless of the mix of Corporation-Uganda Railways Corporation; SETRAG = Société Transgabonnaise (Gabon); fees and taxes and of any promises made dur- SNCC = Société Nationale des Chemins de Fer du Congo (Democratic Republic of Congo); SR = Swaziland Railways; SRC = Sudan Railways Corporation; TAZARA = Tanzania-Zambia Railway; ing the bidding process, a concessioned rail- TRC = Tanzania Railways Corporation. way's strategy will always be constrained by the business fundamentals of the proposed borrowing costs, or rolling-stock acquisition railway privatization deal. A concession- costs, and concessions with high levels of aire will be able to bear only a finite level of both debt and concession fees will be prime charges, whether they are concession fees, candidates for renegotiation. 242 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 11.7 Rail Concession Labor Productivity have generally been unenthusiastic about run- ning passenger services, which do not generate labor productivity/(traffic units [thousands]/staff) 700 the same revenues as freight; this situation has 600 not been helped by delays and disputes about date of the payment of government compensation for 500 concessioning unprofitable services. Further problems have 400 arisen over the level of concession fees, the length of the concession, and arrangements for 300 Rivi-Rivi bridge redundant staff. In some cases, these issues have collapse led to renegotiation of the concession contract. 200 Despite these vicissitudes, the results to date 100 civil war in are encouraging. Even if not all expectations Côte d'Ivoire have been met, most of the concessioned rail- 0 from late-2003 ­5 ­4 ­3 ­2 ­1 0 1 2 3 4 5 6 7 8 9 ways have improved their traffic levels and their years since concessioning productivity and are providing better service Camrail CEAR RSZ Sitarail to users, albeit after a solid injection of invest- ment by donors and international financial Source: Bullock 2009. institutions. Arguably, some of this improve- Note: The overall traffic units carried by a railway are the sum of the passenger-kilometers and the net ment might have occurred anyway. In addition, tonne-kilometers of freight carried. This simple standard measure is widely used as a means of aggre- gating freight and passenger traffic. The relative weighting of passenger and freight is conventionally responsibility for the ongoing rehabilitation taken as 1:1. Rivi-Rivi bridge refers to the Rivi-Rivi River bridge in Balaka, Malawi. Camrail = Cameroon Railway Corporation; CEAR = Central East African Railways Corporation (Malawi); RSZ = Railway and maintenance of track is rapidly emerg- Systems of Zambia Ltd; Sitarail = railway operator for Burkina Faso and Côte d'Ivoire. ing as a key issue between concessionaires and governments. A key government objective in many railway concessions is to obtain finance The Verdict on Concessions--Generally (whether private or through international Beneficial but Not the Full Answer financial institutions) to rehabilitate track infra- Since 1992, there have been 16 rail concessions structure. For most private operators, however, in Africa. Two of the 16 have been canceled, track rehabilitation, especially track renewal, is 1 has been badly affected by war, and 1 has a major expense that drains available funds, but suffered from natural disasters and procedural it is also one that can be easily deferred. delays. Six have operated for five years or more The greatest effect of concessionaires has but only 2 of those without a significant dislo- been improved operations. Given the weak cation of some sort.3 investment and regulatory climate in many Except for the railways immediately adja- African countries, investment flows have been cent to South Africa, those that have not been limited. Under concessioning, operations concessioned have deteriorated continuously have been positive, and efficiency has clearly since the mid-1990s. In a number of cases, improved. Labor productivity has increased these declines will prove to be terminal. Many steadily in all the concessions in operation for governments in Africa will consider conces- over five years, and similar figures will likely sions only as a last-ditch solution, but in many come from recent concessions. Asset produc- cases, the railways have been left to deteriorate tivity has also generally increased. Although for too long, and rectifying the situation will concessionaires in Africa typically have a more be a struggle. appropriate cost structure than their predeces- The concessions have not been without sors, it is rarely the ideal cost structure. Operat- their problems. In many cases, finding more ing costs on railways are a function of capital than a few bidders has been difficult, and in invested, as well as operating efficiency, and several cases, bidders' financial resources have many African railways have been starved of been insufficient to finance the major invest- capital, substantially increasing overall operat- ments required. As a result, the state has had to ing costs. guarantee investments; even then, mobilizing Allocative efficiency is difficult to measure the financing has been slow. Concessionaires directly, but the evidence is generally positive. Railways: Looking for Traffic 243 Improved productivity, an active search for the low-volume operators, the sensible choice new traffic by concessionaires, and better inter- is to find secondhand equipment. Much of nal business practices have all improved rail- the investment to date has been for mainte- way cost and pricing structures and lifted the nance and renewal backlogs, without which level of service, thus helping attract traffic to the railway often would not function, and can the mode that can carry it most efficiently and be characterized as one-time investment to get improve intermodal competition. the systems running. Even that investment has Most concessionaires have fulfilled the pas- been slow, more than four years in Cameroon senger service requirements in their concession and five years on the Nacala line--a long time to agreements, even where it has been operation- wait when a business is barely breaking even. ally difficult or where agreed public service Are concessions a long-term answer? Or obligation payments have not been forthcom- are they merely quick fixes that are living off ing. Many of these services were inherited, investment by third parties and will prove and passenger service would often be more unsustainable in the long term? What more economical with a road-based system. must be done to ensure a sustainable sector? A recent review of four concessions found Many of the answers to those questions must little evidence of monopolistic behavior by con- come from governments. cessionaires (Pozzo di Borgo and others 2006; World Bank 2006). This review examined freight rates and whether services were being reduced Key Issues for Governments so resources could be redeployed to favored users, beyond changes in services that any com- Classic concession schemes4 in Africa are mercialized railway undertakes in response to unlikely to be financially attractive to bidders changing traffic patterns. Few concessions are other than those who can secure financial ben- immune from road competition, except in the efits not directly linked to the railway opera- few cases where roads still must be constructed tions.5 Consequently, unless the structure of or where heavy mineral movements occur. No African rail concessions changes or the mar- evidence exists that personal travel has been ket environment in which they operate alters made more expensive for the poor. favorably, private operators will continue to The greatest disappointment for govern- show limited interest in African railway con- ments has been the lack of infrastructure cessions. Two key areas need to be addressed: funding from sources other than international the financing of passenger services and major financial institutions. Concession agreements track renewals and rehabilitation, both requir- clearly put the responsibility of financing track ing substantial public funding in most conces- maintenance and renewal on private opera- sions. If this funding is provided, governments tors. Likewise, rolling-stock financing has been will also need to strengthen their regulatory left to concessionaires under their contracts. capacity to ensure that the conditions are met However, most concessionaires initially rely and that the effect on the rail sector in general, on loans from international institutions, with and concessionaires in particular, is properly below-market borrowing costs, lengthy loan considered when policies in other sectors of terms, and grace periods to finance infrastruc- the economy are developed. ture. (The exceptions are the Beitbridge Railway [Zimbabwe to South Africa], which relies on Passenger Services take-or-pay clauses that guarantee minimum If governments want the concessionaire to revenues; the Nacala Railway in Mozambique, operate passenger services, they should make which is being funded at semicommercial rates; clear compensation arrangements that can be and Zambia and the Rift Valley Railways [cover- monitored. Few passenger train services will ing Kenya and Uganda], where the investment likely cover even their above-rail costs. Their program is modest and is funded directly by financial contribution to infrastructure costs the concessionaire.) Loans have been provided is minimal, and few services would justify for rolling stock in some cases, but for many of investment in rolling stock, whether hauled by 244 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION locomotive or self-propelled. If these services cash, on a small network when the expenditure are to operate for more than the initial years may not occur for 5 or 10 years, a concessionaire of a concession, governments need to develop is unlikely to reserve funds annually and hold a simple compensation scheme with timely them in reserve that long. Furthermore, raising payments. Any scheme should enable the con- debt financing for rail repair will generally be cessionaire to keep all the revenue, which will possible only through a general corporate loan, encourage maximum operation, and should which is almost impossible for a small stand- include a public contribution, possibly per alone railway. carriage-kilometer, toward the cost of run- Profits to the concessionaire need to be ning unprofitable passenger services. The boosted, or supplementary funding sources scheme should be easily audited and should be need to be developed, or both. Today, Afri- reviewed periodically, perhaps every five years. can railway concessions offer two models for If such schemes are not introduced, pas- financing infrastructure. In the first, govern- senger services will be a constant source of ments finance initial track rehabilitation and conflict between the government and the renewal costs, generally by securing loans from operator. Moreover, the issue will divert the international financial institutions. These loans focus of the concessionaire from the freight are then made to private operators and tend to services, where improvement is far more cover only the initial five-year investment plan important economically for the country. in the hope that they will propel each conces- sionaire's traffic to a level that will then enable Capacity or Willingness of Private it to self-finance future track investments. This Operators to Finance Track Renewal approach is commonly used for railways with a Few, if any, concessions are generating signifi- high ratio of initial track investment compared cant profits for their operators and certainly not with revenues and that are thus unlikely to be enough to fund long-term renewals. Although able to mobilize sufficient private financing. In most concessionaires pay fees into general gov- the second model, governments do not finance ernment revenue, none can afford to do so and initial track renewal but commit to compen- accrue funds for future renewals at the same sating concessionaires for their investment by time. Whether a purely privately financed rail the end of the concession (for example, Kenya, concession model is sustainable in much of Uganda, and Zambia railways). In such cases, Africa remains doubtful. Track structures have the initial amount to be invested is relatively (or should have) lives of several decades, given small in relation to expected revenues, and pri- the traffic volumes typically carried on an Afri- vate operators are assumed to be able to secure can railway. On a small system, track renewal private financing on the merits of their busi- is needed somewhere on the network only ness case. Under both models, governments about every 20 years. It is almost always pos- usually agree to purchase at the end of their sible to defer renewals for several years, albeit concessions the nonamortized portion of any at the cost of deteriorating track conditions infrastructure investment concessionaires have and reduced operating speeds. For any conces- financed. However, the ability of many govern- sionaire who is uncertain about the future, the ments to make such a payment is uncertain, safest decision is to do as little track renewal which often affects infrastructure investment as possible. in the later stages of a concession, although a Even if they do want to renew track, private partial risk guarantee can strengthen the gov- operators will often struggle to generate suffi- ernment's reimbursement commitment. cient cash flow for it. Few concessions are strong Three conditions must be met to secure financially. If a government makes the level of privately financed track investment: (a) gov- the concession fee or rolling-stock purchase ernments ensure that the concession (and thus price the ultimate measure of a successful deal, it the proposed track investment) is financially will limit the successful bidder's ability to renew sound, (b) the nonamortized value of the infrastructure. Even if an operator has sufficient assets owed to the concessionaire at the end Railways: Looking for Traffic 245 of the concession period remains reasonable, the performance of a number of conces- and (c) the concession agreement allows for a sions. Examples range from administratively possible extension of the concession period. imposed salary levels to restrictions on access Often, however, governments will still need to container facilities and unfunded public to assist. Notwithstanding the likely improve- service requirements. Most of these actions ments in efficiency from concessioning, many could be avoided by establishing a properly agreements will probably fail the first hurdle staffed and funded oversight body (the con- of financial soundness. If the government still cession counterparty is generally the obvious wishes to pursue a concession because of the choice for this). A government should ensure benefits of rail transport, it will need to con- that such a body has the necessary politi- tribute grant funds regularly. One option is to cal and technical powers to coordinate and partially finance infrastructure renewal inde- control government actions toward pri- pendently of the concessionaire through a vate rail operators. In practice, that means land transport renewal fund, which could be the agency should meet regularly to discuss an extension of a road fund, created as a com- pending issues with the concessionaire. The mon pool of funds by both the road and rail oversight body should include, or have ready sectors. For example, concession payments access to, a railway technical expert and a could be paid into the fund rather than into railway financial expert, and someone should general revenue. A rationale for this option can head it whose sole responsibility is to moni- be developed from the external costs avoided by tor the railway concession and who reports the carriage of passengers and freight by road directly to the transport and finance minis- rather than rail. ters at least. Effective and Efficient Regulation of Consistent Government Approach to Private Rail Operators Infrastructure Cost Recovery In practice, many concessions ignore many Governments should also develop a coherent or all of their reporting obligations under the and realistic policy regarding infrastructure concession agreements. In some cases, this cost recovery. The road sector has an articulate situation obtains because of operator intransi- and organized lobby. Advocates for government gence, in others because of a lack of expertise or railways, where they exist, have generally been initiative. Not surprisingly, both politicians and ineffectual and poorly prepared, although con- bureaucracies are often ill informed about the cessionaires are generally able to make aggres- problems facing a concessionaire and the rem- sive representations. The lower the road costs edies being attempted. Most concessions have a are and the greater the degree of overloading long list of requirements for the concessionaire permitted, the lower the freight rates by both to meet, and allowing reporting to be ignored road and rail will be--and less money will be inevitably creates plenty of scope for later dis- available from a concessionaire to maintain putes. Regulatory bodies must strengthen their and upgrade the railway infrastructure. capacity and impose annual independent finan- Road competition is strongest in south- cial and operational audits as part of concession ern Africa, which has the most liberal market contracts. One solution for funding the regu- structure, the largest trucks, and the best roads. latory bodies is to use the concession fees, but In addition, the level of road user charges and funding from a land transport fund, if one can the prevalence of overloading heavily affect be established, may be preferable. rail. Requiring rail to fund all its long-term maintenance and upgrades, while tolerating Consistent Government Behavior road cost underrecovery and overloading on toward Railway Concessionaires arterial routes, may help government budgets Aligned with Good Business Practice in the short run, but it is an almost impossible Uncoordinated actions from ministries handicap for most general freight railways to within governments have negatively affected overcome. 246 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION The Way Ahead must be devised to ensure that while govern- ments continue to reap the substantial poten- A wide gap often exists in the minds of tial economic benefits of concessions, private government officials between their expecta- operators' financial returns are high enough tions of what concessioning can achieve and to attract broad and competitive investor what actually happens after they award the participation. concession. Service volumes on most African railways are low, often about that of a mod- Notes erately busy branch line in many countries. The authors of this chapter are Dick Bullock These low volumes can commercially justify and Kenneth Gwilliam, who drew on back- no more than the minimum infrastructure ground material and contributions from Pierre maintenance, which allows operation at a Pozzo di Borgo. 1. Spatial density is measured in route-kilometers speed of 40­60 kilometers per hour. That per 1,000 square kilometers. speed does not permit an attractive passenger 2. Traffic density is expressed as traffic units per service except where no practical alternative route-kilometer. The traffic units carried by a exists--an increasingly rare situation. Govern- railway are the sum of the passenger-kilometers ments that are unprepared to invest substan- and the net tonne-kilometers of freight carried. It tial sums of their own funds in upgrading and is a simple standard measure that is widely used, maintaining infrastructure should therefore although it has some limitations as an indicator expect only a "fit for purpose" freight railway (for example, a first-class passenger-kilometer in operating at moderate speeds but doing so a commercial high-speed TGV train is treated reliably and safely. This type of railway can be identically with a passenger-kilometer in a operated successfully under concession at typi- crowded suburban train). The relative weighting of passenger and freight is conventionally taken cal African traffic densities. If traffic volumes as 1:1, although alternative weightings have been are very low (250,000 tons a year or less) or if a used on some railways from time to time, usu- high standard of passenger service is expected, ally trying to reflect relative costs. continuing financial support from the govern- 3. For more detailed discussions, see Bullock ment will be necessary. 2005. After a concession is awarded, the govern- 4. Classic concession schemes require the private ment must monitor concessionaire behavior operator to take on a significant debt burden in and ensure that the government's interests relation to revenues. are fulfilled. Most important, a government 5. That is, by controlling the entire distribution must ensure that the infrastructure does not chain or through the supply of rail equipment deteriorate over the life of the concession, as is and services. often the case. Deterioration generally occurs when concessionaires have short- or medium- References term financial objectives that do not align with Amos, Paul, and Richard Bullock. 2007. "The Financial Performance of Non-Urban Passen- the longer-term economic objectives of the ger Services." Transport Paper 14, World Bank, government. A concession agreement should Washington, DC. try to reconcile these two objectives as much Bullock, Richard. 2005. "Results of Railway Priva- as possible, and compliance should then be tization in Africa." Transport Paper 8, World monitored regularly. Bank, Washington, DC. Despite these problems, well-run railways ------. 2009. "Taking Stock of Railway Companies should still offer the most economical solu- in Sub-Saharan Africa." Background Paper 17, tion to transporting general freight that is not Africa Infrastructure Country Diagnostic, World time sensitive in major corridors for distances Bank, Washington, DC. over 500­800 kilometers and bulk commodi- Carruthers, Robin, Ranga R. Krishnamani, and ties over shorter distances. The revival of a Siobhan Murray. 2008. "Improving Connectiv- ity: Investing in Transport Infrastructure in Sub- railway through concessioning is warranted Saharan Africa." Background Paper 7, Africa when the business fundamentals supporting it Infrastructure Country Diagnostic, World Bank, are sound. At the same time, better solutions Washington, DC. Railways: Looking for Traffic 247 Pozzo di Borgo, Pierre, Alain Labeau, Raphael UIC (International Union of Railways)-Statistics Eskinazi, Julien Dehornoy, Alan Parte, and Centre. 2008. Railway Time-Series Data 2007. Marouane Ameziane. 2006. "Review of Selected Paris: Railway Technical Publications. Railway Concessions in Sub-Saharan Africa." World Bank. 2006. "Sub-Saharan Africa: Review World Bank, Washington, DC. of Selected Railway Concessions." Africa Thompson, Louis. 2007. "Spoornet and Transnet Transport Sector Report 36491, World Bank, Sectoral Reference Paper." World Bank, Washington, DC. Washington, DC. Chapter 12 Ports and Shipping: Landlords Needed A frican shipping has been largely dereg- Kenya) are not major hubs on the main inter- ulated. However, many African coun- national itineraries, and they appear unlikely tries are trapped in a vicious circle of to become so. Several ports suffer from low high tariffs discouraging traffic and further capacity, particularly in terminal storage, increasing costs. Poor inland links and waste- maintenance, and dredging capability. Overall, ful and costly port administration accentuate however, total use of African port capacity is this problematic situation. The lack of an inte- estimated at 80 percent and likely to remain at grated land distribution system, particularly this level in the near future. for transit, impedes container traffic. Many ports are poorly equipped and inef- Since the mid-1990s, both general cargo and ficiently operated. Container handling rates containerized cargo passing through African fall well below international norms. Port public ports have trebled. Southern Africa has charges for both containers and general cargo had the fastest growth in general cargo traffic are substantially higher than in other regions. and West Africa in container traffic, albeit from Security standards are still extremely variable, a low base. Dry bulk traffic (coal, grain, and and few ports are prepared for the dramatic some chemicals) and liquid bulk traffic (mostly changes in trade and shipping patterns now oil) have also been growing rapidly. By inter- occurring. national standards, however, these traffic cat- The main requirements are organizational. egories are unbalanced, increasing the costs for Many capacity constraints could be overcome African trade. Export volumes greatly exceed simply by making the existing ports more effi- import volumes for dry and liquid bulks, while cient. Regional port planning is required to imports dominate exports for general cargo counter the costs of fragmentation. Port pric- and container trades. ing and regulatory policies need to be more Many ports handle the traffic, few of them commercial and to better respond to the inter- large by world standards. The main transship- national shipping market. Comprehensive ment points for regional traffic (Abidjan, Côte policies are required for modal integration d'Ivoire; Dar es Salaam, Tanzania; Djibouti, and administrative simplification, and modern Djibouti; Durban, South Africa; and Mombasa, port management structures are essential. 249 250 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION The landlord port system has been more accounts for less than 1 percent of total world successful than the service port in Africa (as container traffic and for little more than elsewhere) and is best suited to introduce the 2 percent of all African traffic. East Africa has private sector. Within a landlord port structure, a heavy concentration in Mombasa (6 percent attracting container line operators and major of the Sub-Saharan African total, according to international terminal operators could produce the United Nations Conference on Trade and efficiency improvements. Ghana and Nigeria Development) while West Africa has five ports have moved toward the landlord port, and sev- handling more than 350,000 TEUs each. eral francophone countries operate a hybrid The lack of an integrated land distribution model. However, development is slow, and system, particularly for transit traffic, impedes the involvement of the efficient private global container traffic. Handling of dry and liquid operators is very low. bulk exports is making the most progress, with many port facilities privately owned and inte- grated in a comprehensive logistic system. The African Shipping Market From 1995 to 2005, general cargo has grown at an average annual rate of 6.6 percent and Africa's maritime traffic has been growing rap- at a rate as high as 15.7 percent in southern idly across all cargo types, although container Africa (table 12.2), rates higher than in the rest traffic is highly imbalanced and faces major of the world because of later containerization. challenges because of the lack of efficient General cargo has traditionally been the major transportation links back to the hinterland. type of cargo moved to landlocked countries. Shipping markets are small and thin, contrib- Little congestion occurs in the ports, but han- uting to relatively high costs. dling is inefficient, and the transfer of some of this traffic to containers is contingent on Maritime Traffic--On the Rise but Out inland distribution systems. of Balance Dry bulk traffic is sometimes handled at Except in South Africa, container transport in common-user general cargo facilities, but the Sub-Saharan Africa is still at an early stage of major flows (grain from Mombasa, ferro- development; however, it is growing rapidly chrome from Maputo, and coal from Richards from a very low base, with an average annual Bay [South Africa]) pass through privately growth rate of 7.2 percent and as high as owned and operated dry and liquid bulk termi- 13.8 percent in West Africa (table 12.1). Of the nals, for which the traffic volumes are generally 7.6 million 20-foot equivalent units (TEUs) not well reported. In 2007, the total throughput handled by all Sub-Saharan African ports in of Richards Bay was 66 million tons, making it 2005, Durban handled nearly 2 million TEUs, the world's ninth-largest bulk exporting port. and the three main South African ports together Because major global interests control these handled more than 3 million TEUs. West Africa businesses, the port and shipping arrangements likely conform to the best international stan- dards. Some dry bulk traffic is still handled on general cargo quays, suggesting the possibility Table 12.1 Traffic Trends for Container Trade, Sub-Saharan Ports, by Region, for further specialization, though that depends 1995­2005 on having a large enough basic flow. Average Liquid bulk traffic is predominantly oil, TEUs Overall annual percentage percentage with 11 countries (dominated by Nigeria and Region 1995 2005 change change Angola) supplying 12 percent of world demand East Africa­Indian Ocean 505,100 1,394,956 176 5.8 and 19 percent of U.S. demand. In 2006, oil Southern Africa 1,356,000 3,091,846 128 2.5 made up 85 percent of exports by value from West Africa 673,400 3,126,901 364 13.8 West and Central Africa. It is a growing sector, Total 2,534,500 7,613,703 200 7.2 with Asian countries (China, in particular) and Source: Mundy and Penfold 2008. the United States making significant invest- Note: TEU = 20-foot equivalent unit. ments in Africa, including placement of the Ports and Shipping: Landlords Needed 251 proper export platforms. Again, international Table 12.2 Traffic Trends for General Cargo, 1995­2005 standards are generally met. Average For the most part, African countries are Thousand tons Overall annual exporters of minerals (including oil) and agri- percentage percentage Region 1995 2005 change change cultural products, handled either by specialized East Africa­Indian Ocean 13.84 38.42 177 5.9 or dedicated dry or liquid bulk terminals or by general cargo facilities. For example, agricul- Southern Africa 2.73 14.52 431 15.7 tural products are often handled over the quay West Africa 19.57 51.68 164 5.1 at general cargo facilities by grabs or mobile Total 36.14 104.62 189 6.6 conveyors. Export volumes (loadings) greatly Source: Mundy and Penfold 2008. exceed import volumes (unloadings) for dry and liquid bulks, while imports dominate Figure 12.1 Balance of Sub-Saharan African Container exports for general cargo. Trade, 2005 The dominance of imports is most pro- nounced in the container trades, increasing the 3,500,000 20-foot equivalent units costs. Of Africa's outgoing containers, 23 per- 3,000,000 cent are full, and for West Africa, 12 percent. 2,500,000 Only the southern African ports approach 2,000,000 trade levels of most other world regions, with 1,500,000 containers 30­40 percent full on the backhaul 1,000,000 to Asia (figure 12.1). 500,000 0 East Africa West Africa southern Africa Traffic Patterns--Low Volumes, imported loaded exported loaded High Costs exported empty African shipping has been largely deregulated, and Africa has been integrated into the global Source: Mundy and Penfold 2008. liner network through global players' acquisi- tion of regional operators and replacement of by introducing "congestion charges," rang- direct calls by transshipments from elsewhere. ing from $35 per day for a 20-foot container For example, Maersk uses Salalah (Oman) as in Dakar to $425 per day in Tema (Ghana) the hub for its East African trade and Algeciras in 2006. Delays are often caused by long pro- (Spain) and Tangier (Morocco) as the hubs for cessing and administration times and by poor its West African trade. As a result, the number handling in congested port areas, rather than of direct calls is falling in some areas, and con- by lack of basic quay capacity. Where customs tainer vessel capacity serving African ports is allow the transport of boxes under bond, some relatively small, mostly under 2,000 TEUs. ports have developed off-dock terminals to The proliferation of ports and the limita- move container yards to a less congested area. tions on traffic volumes add to the high costs Most landlocked countries have alternative of shipping to Africa. Greater port efficiency outlets to the sea. For example, the five land- and regional integration to provide better locked countries in West Africa have 15 transit links between the port and its hinterland are possibilities, and Zambia alone has five com- the only solutions for small ports to increase peting corridors. The total cost (including bor- traffic. Without greater port and distribution der and port delays) determines the choice of efficiency, several maritime countries will shippers. In southern Africa, much traffic takes continue to be served by feeder services (par- the longer route to the congested and fairly ticularly in East Africa) and by regional liner inefficient Durban port because of the more services (in West Africa). liberal and efficient land transport and border Delays at the ports are very costly. In 2006, arrangements on that route, as well as the more one extra day in port cost more than $35,000 for frequent sailings. More competition among a 2,200-TEU vessel, and proportionately more corridors could lower the administrative block- for larger ships. Shipping lines have responded ages to free flows of goods on the corridors. 252 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION African Ports are constraints. Maintenance dredging is often inadequate (and much more costly than a few Africa has many small and medium-size ports, years ago) because of the reliance on ad hoc with a low concentration by world standards. contracting rather than long-term performance West Africa has about 25 significant ports, but contracts. In addition, many ports have poor none is among the 70 largest world ports. In navigational aids. addition, new port developments are increas- Despite these circumstances, one source ing the proliferation (figure 12.2). considers total use of African port capacity to be 80 percent overall and likely to remain Port Configuration--Need for a Better there through 2010 (Drewry Shipping Con- Hub-and-Feeder System sultants Ltd. 2006, 2008). Making ports more All three regions, eastern, western, and southern, efficient could overcome many of the capacity claim to suffer from specific capacity problems. constraints, because handling rates are below Mombasa and Dar es Salaam have reached their international standards. For example, the Dur- storage limits for containers in their terminals. ban Container Terminal manages only about Durban is struggling to bring in new capacity to 17 moves per crane hour, short of an interna- meet container handling and storage demand. tional norm of 25 to 30 moves. In West Africa, Luanda and Tema are short of On the East African coastline, Mombasa and container capacity, and Luanda, Douala (Cam- Dar es Salaam are competing as regional trans- eroon), and Tema are under pressure for general shipment points, but both face severe capac- cargo. Numerous factors contribute. Location ity constraints in the short term. Feeders serve in a major urban area limits the capacity of both, mainly Salalah (Oman) and Dubai (the some ports (Apapa [Lagos], Nigeria). For many United Arab Emirates). For example, Europe's ports, equipment availability and maintenance main port, Rotterdam (the Netherlands), has no direct container flows to either port. Traffic Figure 12.2 African Ports, by Size through Dar es Salaam has increased greatly since Hutchison Port Holdings took over the manage- ment contract for the container terminal held by International Container Terminal Services, Inc.; however, Hutchison did not undertake infrastructure investment responsibility. A com- bination of many factors resulted in terminal MAURITANIA congestion, leading to terminal dwell times for MALI NIGER containers of up to 30 days and increased wait- CAPE VERDE SENEGAL GAMBIA CHAD SUDAN ERITREA ing times for vessels. The present contract termi- GUINEA-BISSAU GUINEA BURKINA FASO DJIBOUTI nates in 2010, and negotiations are in progress BENIN NIGERIA SOMALIA SIERRA LEONE COTE D'IVOIREGHANA LIBERIA TOGO CENTRAL AFRICAN REPUBLIC ETHIOPIA to extend the contract for a much longer period CAMEROON with further equipment purchases planned. In EQUATORIAL GUINEA SAO TOME AND PRINCIPE GABON UGANDA KENYA Mombasa, a contract is being awarded to deepen CONGO CONGO, DEM REP RWANDA BURUNDI the port, and a new container terminal to com- SEYCHELLES Cargo (million tons) TANZANIA pete with Dar es Salaam is being discussed. In <3 the near future, however, both ports are likely 3­10 ANGOLA MALAWI COMOROS MAYOTTE to remain relatively poor in facilities, perfor- ZAMBIA 10­50 MOZAMBIQUE mance, and hinterland connections. Meanwhile, ZIMBABWE MADAGASCAR REUNION Djibouti may soon provide competition for > 50 NAMIBIA BOTSWANA Salalah and Dubai, with DP World scheduled to railway SWAZILAND bring on stream a new container terminal facility major road SOUTH AFRICA LESOTHO at Doraleh, targeted specifically at offering sig- nificant transshipment capacity for East Africa and the Indian Ocean. In addition, container Source: Mundy and Penfold 2008. terminal facilities in Jeddah (Saudi Arabia) are Ports and Shipping: Landlords Needed 253 being extended, and in 2008, DP World signed as its main hubs for West African container a concession for the operation and further trade, relaying West African cargo moving development of the Aden Container Terminal to and from Europe and Asia. Nevertheless, in Yemen. Since 2008, a consultant has been the number of public-private partnerships is finalizing the National Port Strategy project increasing, and competition between the pri- for Tanzania. vate port operators in the area is fierce. Con- In South Africa, Durban struggles to han- sequently, some of the big global operators dle its own national traffic and experiences have become willing to look at medium-size recurring berth congestion crises during the and even small terminal projects, which they peak season. Shipping lines are threatening to previously snubbed (Harding, Pálsson, and reintroduce a surcharge for berthing delays, as Raballand 2007). existed between 2002 and 2005. Durban also has problems of environment, security, hinter- Port Ownership and Management-- land connections, and space. Although plans Still Mainly a Public Service Model exist to bring on stream major new capacity, Port planning and management are generally such as the new Pier One scheme, demand is outdated, though seven sampled countries are very strong, and over the short to medium developing new port master plans, several with term, it may well outstrip the new capacity. a focus on institutional reform. Port regula- The number of carriers seeking alternative tion is normally undertaken by a ministry of locations for transshipment in the Indian transport, rather than by a quasi-independent Ocean islands (notably Mauritius) reflects agency; thus, it tends to be highly politicized. Durban's problems. Although superstruc- With its independent regulator, South Africa is ture and infrastructure usually are separated the exception. in South Africa, the common ownership of The dominant port management model in both within the publicly owned Transnet has Africa is still the public service port: the state clearly failed to deliver the necessary improve- enterprise owns the port infrastructure and ments required of a great world port ideally undertakes all port operations. This model is located to act as a transshipment center for beginning to change. Some statutory incorpo- southern Africa. Currently, South Africa is rated port agencies are being reestablished as an end-of-the-line country, and unlike other limited liability commercial companies. Ghana major or global hubs located on one of the and Nigeria have moved toward the landlord very large east-west routes that make econo- port, where the state owns and operates major mies of scale possible, its problems arguably port infrastructure but allows the private sector lie at least partly in the organization and the to provide basic services. In addition, several provision and management of equipment francophone countries have a hybrid model, and handling space, as much as in basic quay called amodiation, in which the port author- capacity. The significance of that distinction ity rents on-dock storage space to privately is that the solution lies in institutional reform owned, licensed stevedoring companies hired and the mobilization of private sector capa- by shipping lines for cargo handling. bilities in port service management, as well as Since 2000, some major container terminals in public sector investment. have been concessioned to the major interna- On the West African coast, Abidjan has tional terminal operators (table 12.3). How- enjoyed some success as a container trans- ever, involvement of the efficient private global shipment center, but it has suffered because operators is still low; in 2007, the top 20 global of the country's internal strife and the specific terminal operators handled only 16 percent of problems relating to ownership of operating throughput in Africa, compared with about rights for the container terminal. The need 70 percent in other regions of the world. Con- for an alternative to Abidjan is indicated by cessioning has proved controversial in some the Maersk Line's (and its affiliate Safmarine) cases, with the results contested in both Luanda using the Spanish port of Algeciras and the and Dakar. No generally accepted "clean" model new container terminal at Tangier (Morocco) exists, and influence and corruption remain. 254 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 12.3 Private Transactions for All Port Sectors, 2000­07 Royalty Number payments to Investment in Number of of canceled government facilities Transaction Countries Ports transactions transactions ($ millions) ($ millions) Management or lease contract Cameroon, Kenya, Mozambique Douala, Mombasa, Maputo 4 1 0 0 Concession contract Angola, Comoros, Equatorial Luanda, Mutsamudu, Luba, 32 0 1,366 1,052 Guinea, Gabon, Ghana, Owendo, Tema, Toamasina, Madagascar, Mozambique, Beira, Maputo, Quelimane, Nigeria, Sudan, Tanzania Apapa (Lagos), Calibar, Harcourt, Lilypond, Onne, Warri, Tin Can, Juba Greenfield projects Côte d'Ivoire, Equatorial Guinea, Abidjan, Luba, Tema, 6 0 316 236 Ghana, Kenya, Mauritius Mombasa, Freeport Total 42 1 1,683 1,288 Source: Mundy and Penfold 2008. Port Performance--Room for Figure 12.3 Average Moves per Hour by Category of Port Improvement Container handling falls below international 20 standards in most ports. Even when container 16 moves per hour gantry cranes are available, the number of 12 container moves per crane hour is usually 10 to 20, compared with 25 to 30 moves in the 8 world ports (figure 12.3).1 When ships' gear 4 is used, the performance is even worse, with 0 only 7 to 10 moves per hour in Dakar, East s r d d ea rie ne ne 'g nt sio sio London (South Africa), Matadi (the Democratic ga ips es es sh nc nc Republic of Congo), and Walvis Bay (Namibia). co co un The low performance is partly explained by the Source: Mundy and Penfold 2008. lower number of containers handled per call with smaller vessels. However, management is even more important. Higher handling rates bulk port facilities, less documentation exists are generally achieved in locations where pri- about their efficiency, but a recent study of vate operators have been in residence for some South African ports showed the bulk ports time; although the hybrid Mozambique model, performing well on international benchmarks in which the government retains a major share, (Bell and Bichou 2007). has not been so successful. The growing interest in Africa as a source of Rates for general cargo handling are also energy products, agricultural products, timber, lower in Africa at 7 to 25 tons per crane hour, and minerals might aid in creating the proper compared with more than 30 tons in other maritime export capacity. The funding for this world ports.2 Almost all handling is through new capacity is invariably provided as part of public ports. a turnkey project, not under a traditional port Specialized oil and coal terminals usually authority budget. Thus, financial impediments do not fall under public port management. are not envisaged. Traditionally, state-owned organizations, The quality of container handling inland is private interests, or a combination have indicated by the cycle times of trucks dropping developed the facilities, which fall outside off and picking up containers at the terminal the mainstream of port operations and fol- and the average container dwell times in a ter- low an integrated supply-chain logic. Because minal. The typical target for an efficient truck of the private involvement in dry and liquid cycle is 1 hour. Average cycle times are estimated Ports and Shipping: Landlords Needed 255 at 5­6 hours in East Africa, 4 hours in south- Table 12.4 Average Port Delays ern Africa, and 10 hours or more in West Africa Range of truck Range of container (table 12.4). Average container dwell times are Region cycle times dwell times 6 days in southern Africa, 12 days in East Africa, East Africa 3.5 hours to 1 day 5 to 28 days and 15 days or more in West Africa, more than Southern Africa 2 to 12 hours 4 to 8 days an accepted international standard of 7 days West Africa 6 hours to 1 + day 11 to 30 days or less. The range of dwell times in southern Source: Mundy and Penfold 2008. Africa (4­8 days) is a much tighter band than in East and West Africa, thanks largely to better organization and control of container storage Table 12.5 Typical Gateway Container and General Cargo Handling Charges in at the terminal. World Markets dollars Like all ports in countries that are signatories to the International Convention for the Safety Container handling General cargo over Region from ship to gate the quay per ton of Life at Sea, African ports have been required, West Africa 100­320 8.00­15.00 since 2004, to comply with the International Ship and Port Facility Security Code. Although East Africa 135­275 6.00­15.00 standards of security are still extremely vari- Southern Africa 110­243 11.00­15.00 able, the estimated costs of compliance for Southeast Asia 80 8.00 African ports--averaging about $2 per TEU Far East 144 8.00 and about $.04­$.05 per ton of general cargo-- Middle East/South Asia 96 7.00 have not increased overall costs significantly United Kingdom 100 8.50 and may have generated compensating bene- Northern Europe 110 7.50 fits, including reduced losses through pilferage and higher customs yields (Kruk and Donner Southern Europe 95 7.00 2008). The long-term concern is a sensible bal- Latin America 154 9.00 ance between security and costs. Australasia 130 9.00 Port charges for both containers and gen- Source: Mundy and Penfold 2008. eral cargo are substantially higher in African ports than in other regions (table 12.5). For container handling, the charges applied in with finding appropriate responses to changes Sub-Saharan Africa can be more than twice in international trade and shipping markets. those typically applied for the same service in other parts of the world, with at least 50 per- Responses to Changes in the cent more as the norm. Normal charges for International Shipping Market general cargo handling per ton offloaded from The problem is not just port capacity. East a vessel in Sub-Saharan Africa are also about Asian ports use vessels in the 8,000­11,000 40 percent above world rates. TEU range, but most African ports cannot efficiently handle container vessels above 2,000 TEUs. Moreover, an upper limit exists Policy Issues and Implementation to optimal vessel capacity because of the low Challenges total volume of freight to African ports. Serv- ing multiple African ports directly with vessels World trade and shipping patterns are chang- of 8,000 TEUs or more is therefore unlikely in ing. Expanding containerization, in ever-larger the near future. Thus, a tendency will exist to vessels, requires port facilities to handle large transship through a small number of African vessels quickly and efficiently. The financial regional hubs with container transshipment crisis of 2008­09 adds to the turmoil. By inter- facilities to distribute traffic along the coasts. national standards, African public port capac- The direction of the trade may also be ity is low, and its performance is poor, bringing changing as some lines consider liner services higher costs and further losses in world trade from Asia by way of southern Africa to the east shares. African governments are thus faced coast of Latin America and the Caribbean. 256 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION For this trade, however, average vessel size is Governments will need to choose how unlikely to increase because ships are cascaded best to develop state-of-the-art ports, with down from longer-distance trade. appropriate technologies and management As the shipping industry has grown more skills. This determination will almost certainly capital intensive, more technically demanding, involve the international private sector, partic- and more subject to global regulatory change, ularly in the container terminal business. Stra- the number of active African shipping lines has tegic port planning must set the roles of the severely decreased. Liberalizing the shipping public and private sectors and identify the pro- market has already brought down deep-sea cesses to attract and select private partners. shipping costs; it should also facilitate the devel- Countries with congested city ports or with opment of less-costly feeder services for con- draft limitations will need to consider whether tainer shipping. As the major traders attract the to rehabilitate existing ports or develop new global operators, they may also develop a niche ones. Developments in the deep-sea shipping market in African feeder services, reestablishing markets may also trigger the need to change African-owned shipping companies. For exam- location. For example, an east-west axis between ple, the establishment of Togo-based Ecoma- Asia and Latin America would be economical rine in the West African feeder market in 2003 only for vessels of 6,000 TEUs or larger. Any was the first indigenous development since the such service would necessitate a port of call in decline of West African national companies South Africa, unlikely to be satisfied by Durban. in the 1990s.3 Where collusion or barriers to Cape Town is developing a new container termi- market entry remain in the shipping market, nal, but it is too far from the industrial heart of governments will need to assess the level and the country in Gauteng province to be a strong distribution of benefits from the restraints and South African hub port. Richards Bay, which compare them to the widely distributed benefits has deep water and a spacious environment, of lower shipping costs in a deregulated market. might be better. It recently launched plans for the staged development of a megaport, includ- Strategic Port Planning ing a container terminal. The expected changes have implications for port planning. Africa can support only a few Port Pricing and Regulation regional hubs and possibly one major hub (in Having the economy benefit from lower South Africa). Competition already exists for transport costs typically requires regulating the hub in East Africa (between Dar es Salaam port tariffs to obtain the most efficient sup- and Mombasa) and will intensify as facilities ply and the lowest real costs, thereby prevent- are upgraded in Djibouti, and regional collab- ing any monopoly, whether state owned or oration--though desirable--seems unlikely. private, from exploiting its advantage in the Simply investing in port capacity will not turn a market. In many countries, however, a single port into a hub unless it has a strategic location, port is a natural monopoly, tempting gov- adequate water depth, and the facilities and ernments to maintain direct ownership and performance to ensure low handling costs. operation and, thus, to use the port as a "cash A strong corridor for transit traffic also cow" to support other government activities. helps. This requires facilitating traffic on the For example, in South Africa, all major ports main trade corridors from the port to the are owned and operated by the National Ports landlocked hinterland. A common problem is Authority (responsible for infrastructure) and the failure to address international, intermodal South African Port Operations (responsible transport holistically. Inland movement, par- for port operations), both part of the state- ticularly across green borders, has been slow owned Transnet, a wider monopoly cover- and expensive, thereby stifling trade. Although ing rail, pipelines, and ports. Transnet has logistics chains are a commercial matter, they presided over an extensive but opaque struc- require the facilitation of integrated port, cus- ture of cross subsidies, allowing the whole toms, and inland transport arrangements--a operation to exist without government sub- matter for government. sidy. Thus, South African ports suffer from Ports and Shipping: Landlords Needed 257 underinvestment. This type of port strategy in terminal operations and with the benefits of inevitably reduces the broader benefit to the an efficient hub-and-feeder structure in the economy of having lower transport costs. deep-sea trade. The need is thus to mobilize private capital Port Management and management skills to improve efficiency African ports do not necessarily lack basic and develop a logistics system. In Africa, as else- quay capacity (though some ports appear to be where, the landlord port system has generally straining their limits). However, they are ineffi- enjoyed greater success than the public service cient in using the basic infrastructure. The lack port and is the best way to attract the private of modern superstructure, particularly cranes, sector. Attracting major international container inhibits fast vessel turnarounds and imposes lines and terminal operators can increase effi- costs on customers. Continuing reliance on ciency. However, mobilization of that potential the public service port structure accentuates still requires appropriate public sector actions overmanning and forgoes the advantages of in the administration of customs and the regu- the modern technologies and management lation of inland transport, as recent experience practices that have revolutionized world mar- in Lagos demonstrates (see box 12.1). kets for shipping and cargo handling (see table Private participation by the most efficient 12.4). The prevalence of state-owned service international port operators must be stimu- ports is also associated with the low concentra- lated by a landlord port philosophy conducive tion of global operators in African ports. Hav- to their participation and by transparent ten- ing global terminal operators in this business dering. Such port reforms are likely to involve would almost certainly improve matters: they retrenchment and compensation. Govern- are well acquainted with the advantages of scale ments should develop advance strategies to BOX 12.1 Private Participation and Port Efficiency: The Case of Apapa Container Terminal, Lagos, Nigeria Lagos port has long been notorious for inadequate facili- mid-April, to enable terminals to clear "alarming" backlogs. ties and congestion. As part of a broader program of port The controller of the Nigeria Customs Service for Apapa reform in early 2006, the Nigerian Ports Authority awarded blamed the low clearance volume on the need to physically a concession to APM Terminals to manage, operate, and examine every container because of the high incidence of develop the Apapa container terminal, increasing capacity concealment and false declaration by importers. However, from 220,000 TEUs per year to 1.6 million TEUs. Within even cleared containers were not being collected. At the end months of the award of that concession, delays for berth- of January, of the reported 9,741 containers in the port for ing space dwindled significantly, and shipping lines reduced delivery to the importers, only 851 had been cleared by cus- their congestion surcharge from $740 to $105 per TEU, toms, with all charges paid and documentation completed saving the Nigerian economy $200 million a year. By early but not picked up by agents. The Nigerian Ports Authority 2009, new gantry cranes had been acquired to triple the consequently proposed introducing demurrage charges of original capacity. $4 per TEU in a bid to force owners to move their contain- However, that was not the end of the story. Although ers out of the ports. In their turn, however, the containers' the port's equipment is able to handle more than 500 con- agents blamed a lack of trucks, arguing that many had been tainers per day for customs examinations, the majority of booked to empty containers. Although the moratorium on the containers are returned to stacking by the end of each entry of new vessels was lifted in early March, some back- day. By January 2009, the port was clogged by uncollected logs and delays remained and significant organizational and containers, and at the end of February, the head of the Nige- regulatory problems still remain. rian Ports Authority announced a temporary suspension of ship entry with immediate effect, lasting until sometime in Source: Press reports assembled by C. Bert Kruk, World Bank. 258 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION create employment alternatives and to handle efficiency (and thereby reduce costs), but also the administration and finance of the adjust- to respond to the growing importance of and ment in the terms of the concession contracts. future obligation in supply chain security. Comprehensive Policies for Modal Notes Integration The authors of this chapter are Mike Mundy Governments must decide how best to fos- and Kenneth Gwilliam, who drew on back- ter and finance integrated port and transport ground material and contributions from Michel Luc Donner, Bradley Julian, Cornelis Kruk, and facilities and associated land uses. A national Andrew Penfold. port plan covering modal integration and port- 1. Some concession contracts specify required per- specific issues is the key. Allocating enough land formance in TEUs per crane hour. Moves per for integrated development should be consid- hour is preferred here as an indicator because ered in the early stages of port planning, partic- moving a 20-foot box requires the same time as ularly for new ports. Links between rail and port moving a 40-foot box. concessions, while having some risk of exploita- 2. This comparison must be viewed cautiously tion in downstream markets, may provide the because the productivity depends on the type best incentive to good modal integration. of cargo handled, which is not allowed for in Governments of both coastal and land- this crude statistic. locked countries must decide which transit 3. Those national companies include Sitram of Côte d'Ivoire, NNSL (Nigerian National Ship- corridors to support and develop. The key to ping Line) of Nigeria, Black Star Line of Ghana, exploiting the major scope for traffic growth Sotonam of Togo, and Camship of Cameroon. is coordinated system development similar to that for the Ghana Gateway and for the Maputo Corridor between the port of Maputo References Bell, Michael G. H., and Khalid Bichou. 2007. and South Africa, Swaziland, and Zimbabwe. "The Port Sector in South Africa: Towards an Landlocked countries will likely want more Integrated Policy and Institutional Reform." than one alternative. Where the bottlenecks are World Bank, Washington, DC. at the seaports, planning and developing inland Drewry Shipping Consultants Ltd. 2006. Global ports (dry ports) deserves consideration, par- Capacity Issues: Annual Review of Global ticularly for landlocked transit countries. Container Terminal Operators. London: Drewry Shipping Consultants Ltd. ------. 2008. Annual Review of Global Container Investment for Quality--the Operators. London: Drewry Shipping Consul- Communications Needs tants Ltd. To reduce dwell times and handling costs, coun- Harding, Alan, Gylfi Pálsson, and Gaël Raballand. tries need to invest in information systems, com- 2007. "Port and Maritime Transport Challenges munications technology, and modern customs in West and Central Africa." Working Paper 84, practices. Customs procedures, in particular, act Sub-Saharan Africa Transport Policy Program, as a bottleneck to port efficiency when they are World Bank, Washington, DC. outdated or open to corruption. As an extreme Kruk, C. Bert, and Michel Luc Donner. 2008. example, one port had to close for an extended "Review of Cost of Compliance with the New International Freight Transport Security period because of customs problems. Modern Requirements: Consolidated Report of the customs procedures and other soft infrastruc- Investigations Carried Out in Ports in the Africa, ture have a major role in delivering efficient Europe and Central Asia, and Latin America and port and freight transportation systems. Caribbean Regions." Transport Paper 16, Striving for efficient ports must be comple- Transport Sector Board, World Bank, Washington, DC. mented by associated measures to increase transparency and reduce corruption in cus- Mundy, Michael, and Andrew Penfold. 2008. "Beyond the Bottlenecks: Ports in Sub-Saharan toms administration. The African ports, like Africa." Background Paper 8, Africa Infrastruc- all world ports, must create port community ture Country Diagnostic, World Bank, systems not only to improve productivity and Washington, DC. Chapter 13 Airports and Air Transport: The Sky's the Limit A ir transport can stimulate regional eco- cases, the protection of small national carriers, nomic development. In Africa, 120,000 as an instrument for this cross subsidy, adds to people are employed directly in air the budget burden and hinders efficient service. transport, and 20 percent of tourism jobs Generally, infrastructure capacity is not a are associated with travelers by air. Air cargo serious problem. The number of airports is sta- is also important in some export trades (such ble, and enough runways exist to handle traffic as flowers from Kenya and fish from Tanzania). with better scheduling and modest investment Overall, traffic has been growing at about in parallel taxiways and some terminal facili- 6 percent a year from 1997 to 2006. In south- ties. Aircraft fleets are being modernized, but ern and eastern Africa, the market growth is air traffic control facilities need substantial strongest, with three vigorous hubs and three improvement. Revenues from airports and air major African carriers dominating interna- traffic are probably high enough to finance the tional and domestic markets. In central and necessary investments, but the sector does not western Africa, however, the market is stag- capture them. The problem is both political nating, with the vacuum created by the con- and organizational. flict in Côte d'Ivoire and the demise of several Two other serious challenges remain. First, regional airlines still unfilled. liberalization of the international regime Air transport in Sub-Saharan Africa is still within Africa must be completed, as commit- expensive by international standards. Landing ted to in the Yamoussoukro Decision of 1999. charges are high, partly caused by the absence of Many international agreements within Africa the support from concessions revenue enjoyed have been liberalized, resulting in routes and by many airports in the world. Because of rela- aircraft sizes that are better adapted to the tively low volumes of traffic on many routes in market and some large, viable indigenous Africa, airfares are also high, despite the efforts carriers that are expanding. However, the of some governments to subsidize domestic fares domestic and intercontinental markets often from protected intercontinental routes. In many remain protected, and many small, nonviable 259 260 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION state-owned operations continue, particularly A notable acceleration has occurred in all in southern Africa, that are protected at the main traffic categories (measured in number expense of potential users of air transport. of seats) since 2004, including domestic, inter- Second, air safety must be addressed. Sub- national, and intercontinental (table 13.1). Saharan African carriers have the world's worst Intercontinental traffic, which relies heavily accident record. Contrary to many accounts, on the three major gateways of Johannesburg, this unenviable record is largely attributable to Nairobi, and Addis Ababa, grew at an average poor pilot capabilities and weak safety adminis- of 6.2 percent a year between 2001 and 2007. tration rather than the age of aircraft. Supervi- Although the South African routes to Germany sory oversight of operators is particularly lax. and the United Kingdom are still the most heavily trafficked, the most notable feature of this growth has been the significant rise in ser- The African Air Transport Market vice through the Middle East from all of the main gateways. Air traffic in Africa has been growing, but at International traffic within Sub-Saharan the same time concentrating, so that fewer Africa grew slightly faster, at an average of routes are served. Lack of competition keeps 6.5 percent a year between 2004 and 2007, with costs relatively high, and the safety record the same three major hubs handling 36 per- remains very worrisome. However, fleet cent of such traffic (figure 13.2). The national renewal has been substantial in recent years, airline dominates the interregional traffic of downsizing aircraft toward a city-jet size. each hub: South African Airways (33 percent of the international traffic at the hub), Kenya Traffic Trends--on the Rise Airways (70 percent), and Ethiopian Airlines All segments of the air passenger transport (83 percent). Both Kenya Airways and Ethio- market in Sub-Saharan Africa have been grow- pian Airlines have been developing new routes ing steadily since 1997, with a small hesitation as the sole carrier, while most of South Afri- following September 11, 2001, and a larger can Airways' international routes have more downturn in West and Central Africa associ- than one carrier in competition. East Africa ated with the collapse of Air Afrique in 2004 has the more developed network. In West and (figure 13.1). Central Africa, only Nigeria has a significant number of intercontinental and international connections. Figure 13.1 Growth of Air Traffic, 1997­2006 Domestic traffic grew fastest, at more than 200 12 percent a year between 2004 and 2007. But that growth--as high as 67 percent in Nigeria-- 180 varied greatly and actually declined in about half 160 of the countries in the region between 2001 and number of seats (thousands) 140 2007. Overall, the number of domestic city-pairs 120 served dropped by 229 in that same period. In addition, excluding Mozambique, Nigeria, and 100 South Africa, domestic traffic showed an aver- 80 age annual decline of 1 percent a year and a loss 60 of 137 routes between 2004 and 2007. 40 The steady growth of traffic overall dis- 20 guises some severe problems at the subregional level. The Banjul Accord Group of countries, 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 including Nigeria, has shown the fastest all Europe Africa Middle East growth, followed by the more developed, yet North America Southeast Asia healthily growing regions of East and southern Africa (figure 13.3). Because of regional airline Source: Bofinger 2009. collapses, a swath of nations surrounding the Airports and Air Transport: The Sky's the Limit 261 Table 13.1 Air Traffic Growth in Sub-Saharan Africa, 2001­07 Estimated number of seats Percentage of growth Market 2001 2004 2007 2001­04 2004­07 2001­07 All Sub-Saharan Africa 50,410,448 54,544,861 72,338,706 2.66 9.87 6.20 Domestic Sub-Saharan Africa 18,184,071 19,356,818 27,477,027 2.11 12.39 7.12 International Sub-Saharan Africa 11,758,107 11,868,280 14,327,728 0.31 6.48 3.35 Intercontinental Sub-Saharan Africa 19,544,122 22,051,174 28,068,536 4.11 8.38 6.22 Between North Africa and Sub-Saharan Africa 924,148 1,268,588 2,465,415 11.14 24.79 17.77 Other 1,036,932 1,076,010 794,621 1.24 -9.61 -4.34 Source: Bofinger 2009. Banjul Accord Group of countries has experi- Figure 13.2 Top-60 International Routes within Sub-Saharan Africa, 2007 enced negative growth. Of 19 countries that lost international connections since 2004, 16 are in these areas of West and Central Africa. Most worrisome are the Central African Republic (only one flight a week in November 2007), Chad, Eritrea, Mauritania, and the Seychelles. Not only are they minimally connected, but also their connectivity plunged between 2004 and 2007. Air Transport Supply--the Process of Concentration In the early 1960s, many of the newly indepen- dent African states created their own, mostly government-owned, national air carriers. Most pursued a business strategy designed to protect profits on international routes through restrictive use of the bilateral permission sys- tem in order to cross subsidize costly yet extensive domestic route networks. Until 1991, nearly all African carriers were state owned. Some very small carriers failed early or were assimilated into Air Afrique. Since 2001, how- ever, several medium-size airlines have ceased Source: Bofinger 2009. operations, including Air Afrique, Air Gabon, Note: The thickness of the lines is on a continuous scale, with the thickest line (South Africa­Zambia) Ghana Airways Corporation, and Nigeria representing nearly 1 million seats for 2007 and the thinnest (Ethiopia­Ghana) representing 62,000 seats. The size of the end points results from the thickness of the lines. Airways. Countries now fall into one of four distinct groups. Countries in the first group (Ethiopia, Kenya, and South Africa) have suc- withdrawn from state-owned carriers and left ceeded in establishing an efficient state-owned their markets to private operators only. The carrier. The 17 countries in the second group three countries (the Central African Republic, continue to operate weak, highly subsidized, Lesotho, and Niger) in the final group have no state-owned carriers mainly operating in very known operators and rely on services from small, protected markets and often in domes- other countries. tic and shorter-distance international sec- The combined effect of airline failures and tors. The 25 countries in the third group have regulatory restrictions on competition has 262 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 13.3 Regional Growth Zones in Seats Offered, All Travel Categories negative growth (­20% or so less since 2001) good growth (20% or so more since 2001) very high growth (almost 40% or more since 2001) highest growth (60% or more since 2001) data not available Source: Bofinger 2009. been to increase concentration, both in the by the national flag carrier and is highly con- total market and on individual routes. By early centrated. Of the 286 routes with service 2008, 15 airlines accounted for 59 percent of in 2007, only 54 had more than one service all seat capacity offered in Africa. The top-20 provider--generally, the national flag carrier. intercontinental routes each average 3.45 com- Standing out among the larger countries for peting airlines. International service within allowing competition are South Africa, where Sub-Saharan Africa is less competitive, with competition occurs only on the heaviest routes, just 15 airlines providing more than 82 percent and Tanzania, where more than one service of capacity, and the big 3 (Ethiopian Airlines, provider exists on all of its 17 domestic routes. Kenya Airways, and South African Airways) providing 57 percent. The Equipment Operated Sixteen of the top 60 routes, and 66 of the The two most significant trends for Sub- 206 total routes, have only one carrier. Ethi- Saharan African air transport have been the opian Airlines and, to a lesser extent, Kenya downsizing of aircraft toward the city-jet size Airways have contributed to this concentra- (Boeing 737 or Airbus 319) and, contrary to tion by developing new routes as sole carri- many accounts, the renewal of the fleet. The ers. In most cases, domestic travel is serviced proportion of seat miles flown on older Western Airports and Air Transport: The Sky's the Limit 263 aircraft has increased to accommodate growth Figure 13.4 Airfares by Distance on African Routes, Including North Africa, 2008 quickly and inexpensively only in some domes- 3.00 tic markets, and even that proportion is only 4 percent of seat miles overall. A slight increase 2.50 has already occurred in the use of commuter propeller aircraft on the low-volume routes 2.00 in Tanzania and in West and Central Africa, US$/mile and such aircraft may have an important role 1.50 in the recovery of the air transport market in that subregion. 1.00 Fares--More Expensive Within Africa 0.50 than Outside 0 Air travel within Africa is considerably more 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 expensive per mile flown than intercontinental distance (miles) travel, especially on routes of less than 2,000 international intercontinental domestic nautical miles (figure 13.4). This cost reflects between Sub-Saharan Africa and North Africa the fact that domestic and international mar- kets are less dense and less competitive than Source: Bofinger 2009. the longer distance intercontinental markets. Moreover, aircraft landing charges are gener- ally high by international standards, partly Some commentators have ascribed this bad because of the absence of nonflight revenues record primarily to the use of Eastern-built air- from airport concessions. In some cases, the craft. Certainly, the hull loss rate per million skewed cost is limited by subsidized or fixed departures of Eastern-built aircraft reached pricing on domestic routes operated by a 54.35 in 2006 in Africa, 10 times the world national flag carrier. average for such aircraft (5.61). However, a study by the Interstate Aviation Committee Flying in Africa--a Dangerous Business concluded that the flight safety of most Soviet- The most notable problem of the African air made types of aircraft is no worse and, in some transport industry is safety. In 2004, Africa cases, is better than that of their Western equiv- experienced 22.0 percent of all accidents world- alents. Much more significant is that mainly wide, despite accounting for only 4.5 percent of small, poorly regulated fringe carriers fly these all sectors flown globally. In 2006, African car- aircraft. The high accident rate in Sub-Saharan riers lost 4.31 aircraft per million departures, Africa is thus primarily a result of poor safety compared with 0.65 aircraft worldwide. standards and lax supervision, not operation of The major intercontinental carriers, most of Eastern-built or older aircraft. them Asian, European, or U.S. registered, have an excellent safety record. In contrast, compa- nies operating Western-built aircraft that are Air Transport Policy in Africa still in use in most developed countries but reg- istered in a Sub-Saharan African country have Much of the world has moved from a strictly had 15 fatal accidents since the mid-1990s, regulated air transport industry to a more lib- involving 1,080 deaths. Even worse, African eralized one. In the United States, deregulation carriers operating older Western- or Eastern- resulted in the failure of weaker carriers, emer- built aircraft have reported at least 29 accidents gence of new low-cost carriers, rearrangement over same period (and suspicions exist that of routes, and evolution of the hub-and-spoke many others are not reported). In addition, system. Fares fell on average, and load factors many accidents involve flights conducted by generally increased through differentiated the air force, which in many African countries pricing. In Europe, too, the rise of low-cost car- transports civilian passengers and cargo. riers was a highly visible effect of deregulation. 264 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION On October 17, 1988, the ministers in liberalizing fares. Signatory states were obliged charge of civil aviation in African states met in to ensure the fair opportunity to compete on Yamoussoukro, Côte d'Ivoire, and expounded a nondiscriminatory basis. A monitoring body a new African Air Transport Policy, later called would supervise and implement the decision, the Yamoussoukro Declaration. Although the and an African air transport executing agency ministers foresaw the gradual elimination would ensure fair competition. The decision of traffic restrictions, the declaration was paid special attention to improving air trans- aimed primarily at cooperation among Afri- port safety. However, even though the decision can air carriers to better compete with non- is a pan-African agreement to which most African carriers. Nevertheless, it did stimulate African states are bound, the parties decided the United Nations Economic Commission that separate regional economic organizations for Africa to initiate a further conference in should implement it. Yamoussoukro, which resulted in the historic The monitoring body has met only a few agreement on pan-African liberalization of times. Competition rules and arbitration air services, the 1999 Yamoussoukro Decision procedures remain pending. Although an (Schlumberger 2008). executing agency was finally created in 2007 The decision's main objective was the grad- by assigning the responsibilities and duties to ual liberalization of scheduled and nonsched- the African Civil Aviation Commission, a spe- uled intra-African air services, abolishing limits cialized institution of the African Union, the on the capacity and frequency of international commission has yet to prove its effectiveness. air services within Africa, universally grant- In contrast, operational implementation has ing traffic rights up to the fifth freedom,1 and been much more productive (table 13.2), with Table 13.2 Air Service Liberalization in African Regional Groupings General status of Percentage of flights Yamoussoukro Decision Status of air services under fifth and seventh Community Members implementation liberalization freedoms Banjul Accord Group Cape Verde; Gambia, The; Principles of YD have been First through fifth freedoms 43 Ghana; Guinea; Liberia; Nigeria; agreed upon in a multilateral have been granted, tariffs are and Sierra Leone air service agreement. free, and capacity and frequency are open. Economic and Monetary Cameroon; Central African Principles of YD have been First through fifth freedoms have 28 Community of Central Africa Republic; Gabon; Equatorial agreed upon in an air transport been granted, tariffs are free, Guinea; Congo, Rep. of; Chad program. Some minor restric- and capacity and frequency are tions remain. open. A maximum of two carri- ers per state may participate. Common Market for Eastern and Most East and southern Afri- Full liberalization has been Liberalization is pending. When 14 Southern Africa can states except Botswana, agreed upon, but implementa- it is applied, operators will be Lesotho, South Africa, and tion is pending until a joint able to serve any destination Tanzania competition authority is (all freedoms), and tariffs and established. capacity and/or frequency will be unregulated. East African Community Kenya, Tanzania, Uganda EAC Council issued a directive Air services are not liberalized 16 to amend bilaterals among EAC because the amendments of states to conform to YD. bilaterals remain pending. Southern African Development Most countries south of No steps were taken toward No liberalization within SADC 6 Community Tanzania implementation, even though has been initiated. civil aviation policy includes gradual liberalization of air services within SADC. West African Economic and Benin, Burkina Faso, Côte Within WAEMU, YD is fully All freedoms, including cabo- 44 Monetary Union d'Ivoire, Guinea-Bissau, Mali, implemented. tage, have been granted. Tariffs Niger, Senegal, Togo are liberalized. Source: Bofinger 2009. Note: EAC = East African Community; SADC = Southern Africa Development Community; WAEMU = West African Economic and Monetary Union; YD = Yamoussoukro Decision. Airports and Air Transport: The Sky's the Limit 265 greater freedom to negotiate bilateral agree- Johannesburg, and Nairobi) act as gateways to ments in all of the subregions. the continent for intercontinental traffic and In West Africa, the Economic Community as hubs for its distribution. They currently of West African States was unable to take any handle 36 percent of all international traffic significant steps toward liberalizing air services, in Africa. Lagos could perform a similar func- but the smaller West African Economic and tion for West and Central Africa but is lag- Monetary Union went even beyond the ging. Second, about 40 medium-size airports Yamoussoukro principles with a regime that are connected to the hubs and primarily serve includes cabotage rights. The Banjul Accord international and domestic traffic. Third, more Group, also in West Africa, agreed to a multi- than 200 small and often nonviable airports lateral air service agreement, fully compatible act as the distribution points for frequently with the Yamoussoukro Decision. In Central declining domestic air traffic. Except for those Africa, the Economic and Monetary Commu- in the Banjul Accord countries (Cape Verde, nity of Central Africa has implemented all the the Gambia, Ghana, Guinea, Liberia, Nigeria, necessary legislative and regulatory elements to and Sierra Leone), the number of airports with comply with the provisions of the Yamoussoukro scheduled service dropped by 20­40 percent Decision. In East and southern Africa, the Com- between 2001 and 2007. mon Market for Eastern and Southern Africa Nearly all airports that had scheduled achieved the most progress, but full application advertised service in November 2007 have at of the decision principles remains suspended. least one paved major runway. Only a dozen The East African Community has chosen the or so airstrips are unpaved, most of them in effective strategy of directing the bilaterals to countries either in or just out of military con- conform to the decision, but signing of the flict. An exception is Tanzania, which has five agreement remains pending. The Southern airports with scheduled service and with alter- African Development Community has pro- natively surfaced runways, though projects for gressed the least. Apparently, the dominant posi- resurfacing are in progress. tion of South Africa remains the main obstacle Airports in Africa do not often have "pure" toward implementing the decision. Overall, runway capacity constraints. With a five- about two-thirds of air transport service within minute separation between flights, a single- Africa is now liberalized. runway airport could accommodate 144 flights in 12 hours, or more than 1,000 flights a week. With an average load of 120 passengers, more African Air Transport than 17,000 passengers a day could be handled. Infrastructure Even with a 20-minute separation, passenger numbers would not exceed 4,300 a day. Very The existing air transport infrastructure can few airports in Africa handle more passengers fairly well accommodate Africa's current air than this. Capacity constraints may appear, traffic and foreseeable growth; only a hand- however, on taxiways, aprons, and jetways. ful of cases warrant investments in taxiways or In many African airports, aircraft must taxi terminal upgrades. Instead, the main invest- to the turning bay, turn around, and taxi back ment need lies in air traffic control and surveil- toward the access to the apron, usually in the lance equipment, which with few exceptions is center of the runway. This procedure is accept- largely inadequate. able in most airports where enough time elapses between departing and arriving aircraft to Airports--a Declining Number in do so. Only a relatively few high-volume air- Service, but Adequate Runway Capacity ports require parallel taxiways with multiple At least 2,900 airports exist in Africa, of which turn-off ramps from the runway. Despite this an estimated 261 in Sub-Saharan Africa generally adequate flight capacity, policy makers received scheduled services in 2007. These in several countries are urging the construction airports fall into three groups. First, three of entirely new international airports for which major international airports (Addis Ababa, no economic justification exists. 266 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION The airside infrastructure in such major hubs capacity issues are already being addressed. as Johannesburg and Nairobi meets high inter- For example, Nairobi's passenger terminal is national standards in runway length, instru- going through an extensive upgrade to allow ment landing systems, and so on. For airports more than 9 million passengers a year. In other with lower traffic volumes, however, significant cases, the declared capacity needs to be exam- differences in the quality of the infrastructure ined. For example, Malawi's airport in Lilongwe, are apparent. Fewer than 50 percent of airport though clearly in need of some upgrades, does runways are in excellent or very good condi- not appear on the ground to be as deficient as tion, compared with 96 percent in North Africa. the figures suggest. Rescheduling to prevent Although traffic to airports without paved run- the simultaneous arrival of too many flights ways is low, the number of airports with poor can be of great help. runways is relatively high in some countries (table 13.3). Fortunately, 87 percent of the seats Airport Management--Limited landed are on excellent or very good runways, Privatization and only about 4 percent of the traffic is to air- Airports usually have some quasi-independent ports with marginal or poor ratings. operating agency, whether government owned Instrument landing systems can be found at or not. Even the company that owns South Afri- nearly all airports with an estimated capacity ca's nine most important airports, including of 1 million seats or more, but their presence Johannesburg, is only partially privatized, drops off rapidly below this traffic volume. In with majority ownership still held by the state. many smaller, older airports, nondirectional Cameroon, Côte d'Ivoire, and Madagascar beacon systems--now very outdated--still pre- have concessioned their major airport groups, vail. However, this circumstance does not nec- although in Madagascar, the government has essarily imply that new investment is needed in a majority shareholding in the concessionaire. ground-based navigation infrastructure; today's Kenya has concessioned the development of satellite technology can easily replace many of the cargo terminal at Nairobi's international the ground-based systems at a much lower cost. airport. South Africa has divested some smaller Nevertheless, in practice, few plans have been airports completely. Even without full airport made to replace obsolete technologies. concessions, the range of airport service pro- viders is wide. In Tanzania's Dar es Salaam Airport Terminals--Few Capacity International Airport, passenger services are Constraints performed by Swissport, and at Nairobi's Jomo Some evidence exists of inadequate capacity Kenyatta International Airport, a broad range of passenger terminals, though data are not of competition exists for landside services. readily available. Many Sub-Saharan African Navigation and air traffic control still typically terminals report traffic volumes at or above fall directly under governmental agencies, with their declared capacity, and in some cases, some services subcontracted. Table 13.3 Runway Quality in Sub-Saharan Africa, 2007 Percentage of total Percentage of total Rating Airports (number) airports Seats (millions) seats Excellent 31 18 67.75 68 Very good 50 29 18.49 19 Fair 46 27 8.51 9 Marginal 10 6 2.29 2 Poor 36 21 2.42 2 Total 173 100 99.50 100 Source: Bofinger 2009. Airports and Air Transport: The Sky's the Limit 267 Air Traffic Control--the Critical to maintain unremunerative services without Infrastructure Deficiency direct subsidies, often performing politically Air traffic control infrastructure in Africa is determined services for no return at all. Even wanting, with the exception of airports in with protection of the home market, they South Africa and Kenya.2 Addis Ababa uses no lacked the potential to be commercially viable. civilian radar, forcing extra distance and time Consequently, many of them have failed, as did separations between aircraft. In Malawi, as the multinational Air Afrique. equipment has aged and become too expen- Despite the failures, Africa still has many sive to maintain, surveillance has fallen into small, nationally owned flag carriers that were disrepair. Even when the equipment exists, established decades ago. They survived ini- radar procedures (and radar separations) are tially through protection of both the domestic not always implemented. In Kenya, only the market and the national share of international airport in Nairobi has full-time radar vector- markets. Even when losses mounted, gov- ing, while that in Mombasa switches to radar ernments often argued that the national flag only if weather so demands. In Tanzania, Dar carrier should be retained in a restructured es Salaam's airport has a good radar installa- form, because subsidized domestic routes tion, with a secondary radar range in excess of would be dropped without a national carrier, 200 miles, but has no radar vectoring because causing regional isolation, and revenues from of a lack of radar-certified controllers. foreigners traveling within the country would Closely related to traffic surveillance is the be lost. However, restructuring has often capability for aircraft communication to and simply meant strengthening protection and from the ground. In certain areas of Africa, an investing in new aircraft, without improving airliner could fly for more than an hour and the basic economics of the operation. be unable to make contact with the ground. Africa has some successful nationally owned The lack of adequate surveillance also raises flag carriers, usually the dominant carrier in an concerns about search and rescue operations.3 intercontinental hub. However, they are rare. In Weather installations are also sparse, often most cases, instead of protecting a flag carrier, relying on physical observation using manual countries should pursue a better alternative of techniques now commonly automated in the opening the market to allow a successful oper- West. Moreover, broadband infrastructure is ator to provide service. This approach could not available in most airports. include a successful flag carrier from another country. In principle, the merger of several regional airlines should create some strength, Policy Challenges but the experience of Air Afrique shows the dangers of political intervention and noncom- Countries in Africa face five main challenges in mercial operation. A possible compromise is developing their airports and air transport: to paint one of the successful operator's air- craft in the flag carrier's colors and hire a crew · Deciding what to do with national flag for passenger services in the country. carriers · Improving air safety Improvement of Air Safety According to the African Airlines Association, · Liberalizing air transport markets the age of the fleet is the greatest concern: · Financing infrastructure nearly one-third of the total fleet of 750 aircraft · Developing and maintaining skills. is more than 20 years old, with a prevalence of ex-Soviet types in certain countries. That view A Strategy for National Flag Carriers is not widely accepted. Although most acci- Most national flag carriers in Sub-Saharan dents in 2006 involved old, Soviet-built turbo- Africa have been small by international stan- prop aircraft, more recent, devastating crashes dards, with a very weak base of demand. have involved mainly Western-built aircraft. They usually experienced political pressures Worldwide averages suggest that, vintage for 268 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION vintage, Russian-built aircraft are as safe as record. The U.S. Federal Aviation Adminis- Western-built aircraft if properly maintained tration and the European Union also rate air and operated. The problem is that many small safety in many African countries as poor. carriers acquire one or more old aircraft on Because of the interaction between national the nontransparent aircraft supply market and systems, air safety is a regional problem that operate them without supervision by the civil needs to be addressed regionally. ASECNA aviation authorities. Their pilots work long (Agence pour la Sécurité de la Navigation Aéri- hours and regularly operate aircraft in a dan- enne en Afrique et à Madagascar), founded in gerous environment, resulting in crashes. Even 1959, has 15 member states aiming to pool air for U.S.-built aircraft, U.S. National Transpor- navigation services and other infrastructure. In tation Board inquiries highlighted several cases addition to navigation infrastructure, the orga- of pilot error, in which poor pilot training and nization manages eight airports. Further steps assessment contributed to an accident. are under way. In East Africa, a centralized East In general, the International Air Trans- African civil aviation authority has just been port Association identifies poor regulatory formed with support from the U.S. Depart- oversight as the top threat to safety in Africa, ment of Transportation's Safe Skies for Africa followed by inadequate safety management program. Though not yet fully implemented, systems. Similarly, the International Civil Avia- the organization, now headquartered at the tion Organization's Universal Safety Oversight East African Community in Arusha, Tanzania, Program shows that safety implementation in would supplement the existing civil aviation Africa is very deficient. For example, West and authorities in the member countries by pro- Central Africa, and East and southern Africa viding resources in a central pool available to perform below the world average in all the community countries. Also, two projects for critical elements of safety implementation-- the Cooperative Development of Operational in most cases, by a factor of two (figure 13.5). Safety and Continuing Airworthiness Program These deficiencies are highly correlated with are being planned for the Southern African accident rates, suggesting that institutional Development Community and the Economic failings explain much of Africa's poor accident and Monetary Community of Central Africa. Thus, regional pooling of resources is address- ing Africa's shortcomings in oversight. Figure 13.5 International Civil Aviation Organization Analysis of Safety Implementation in Africa, 2004 Liberalization of Air Transport Markets Formal implementation of the Yamoussoukro 70 Decision remains very slow, particularly in 60 southern Africa. Many national airline offi- % of nonimplementation 50 cials have advised their governments that the airlines are not ready for a free market and still 40 need protection. Many are concerned that an 30 open skies policy for intercontinental trans- 20 port, particularly to Europe, would drive Afri- 10 can carriers from the intercontinental market, with adverse secondary effects for international 0 and domestic services in Africa. at n ion g at ion at nce tio ty ue f er l sin son ied isl atio lat tin at nica iss o nc fe ion s lig at s s ns This attitude almost certainly hinders lig lla ty on gu ra fu sa r if s ion ion ial l ne leg avi ob fic pe al e m ch re pe ob rvei fe ti ht & al qu sa solu rti nc te co y sig re su ce ar regional development. Ample evidence already er ctu ifi re im ed & ec ov u pr g inu sp str exists of the benefits for southern Africa from nic ida nt A en ch gu CA co lic te critical element wider liberalization. For example, liberalizing the domestic market in South Africa in 1990 global East and southern Africa West and Central Africa fueled passenger growth of 80 percent between Source: ICAO 2004. 1994 and 2004, and eventually led to the Note: CAA = Civil Aviation Authority. establishment of domestic low-cost carriers. Airports and Air Transport: The Sky's the Limit 269 Similarly, liberalizing the Nairobi-Johannesburg does the allocation of many airport revenues route in 2000 increased flights from 4 a day to (such as those for overflight) to the state trea- 14 and increased passenger volumes by 69 per- sury, with only a small part left to the sector. cent. Competition is important here. On the The sensible investment strategy would seem Johannesburg-Lusaka route (for which South to be investment in existing infrastructure, not African Airways had been the only carrier for new airports, and use of newer, cheaper tech- over 10 years), designating the South African nologies for air traffic surveillance and naviga- low-cost carrier, Kulula, as the Zambian carrier tion systems. reduced fares by 33­38 percent and increased passengers by 38 percent. Development and Maintenance of Skills Recent modeling suggests that full liberal- The International Air Transport Association ization in the Southern African Development has identified the lack of effective flight crew Community would reduce airfares by 18­40 training and proficiency as a major source of percent with a low-cost carrier entering the safety problems. Even where training is ade- market and would increase traffic volume quate, highly trained flight crews in the poorer by 20 percent (Genesis Analytics and others countries can usually command higher salaries 2006). Another half a million foreign tourists when working for a larger foreign airline. A would arrive by air each year, spending more similar drain of skills affects regulatory staff. than $500 million. The multiplier effect would Safety inspectors trained with donor funding increase the Southern African Development abandon oversight almost immediately to earn Community's GDP by about $1.5 billion, or much more working for an airline. 0.5 percent growth. Another study on liber- Many issues are interconnected. Poor coun- alizing routes on 20 city-pairs to and from tries trying to maintain a domestic flag carrier Addis Ababa came to a similar conclusion do not pay enough to train and maintain good (Abate 2007). flight crews. Because funds are usually insuffi- cient to provide competitive salaries for safety Financing of Infrastructure inspectors, oversight standards also fall. Main- Two agencies are typically concerned with air taining staff is thus linked to the policies for transport infrastructure. Air traffic control is flag carriers, market liberalization, and sector usually the function of a national civil aviation finance. Only a coherent and fiscally affordable authority, whereas airport infrastructure (both set of policies in these challenging areas can airside and landside) is usually the responsi- produce a more sustainable air transport sec- bility of a separate airports agency. Both have tor in Sub-Saharan Africa. financing problems. Civil aviation authorities survive on fees. Where the land mass is large and the geographic Notes location important, significant air navigation The authors of this chapter are Heinrich charges (exceeding many of the other service Bofinger and Kenneth Gwilliam, who drew charges relied on by authorities) can be gained on background material and contributions from overflights. However, reallocating those from Michel Iches, Pierre Pozzo di Borgo, and charges can be politically contentious. A truly Charles Schlumberger. independent regulatory body could improve 1. The eight "freedoms of the air" are the focus of services, but in many cases, the revenues end international regulation of air transport. The up in the national treasury. first and second freedoms are technical free- doms to overfly a foreign country or to land The same situation applies to airports. In for refueling. The third and fourth freedoms a study by the International Civil Aviation are commercial freedoms to carry passengers Organization using 2005 data (ICAO 2008), from a carrier's home country to another or the main access airports in Africa were con- vice versa. The fifth to seventh freedoms con- sidered inherently profitable. However, the cern the rights to carry passengers between two nonpayment of tariffs by weak flag carriers in foreign countries, either as an extension of a small countries undermines airport finance, as flight from the home country (fifth), through 270 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION a stop in the home country (sixth), or without Bofinger, Heinrich C. 2009. "Air Transport: Chal- ongoing service to the home base (seventh). lenges to Growth." Background Paper 16, Africa The eighth freedom, pure cabotage, is the right Infrastructure Country Diagnostic, World Bank, to carry traffic between two points in a foreign Washington, DC. country. Genesis Analytics, Andrew Myburgh, Fathima 2. The lack of radar installations should be discussed Sheik, Fatima Fiandeiro, and James Hodge. as the lack of surveillance infrastructure, because 2006. Clear Skies over Southern Africa. Wood- radar is now an obsolete technology. Newer, much mead, South Africa: ComMark Trust. more accurate, and much less costly technologies ICAO (International Civil Aviation Organization). are now being installed, as in the United States. 2004. "An Update of the ICAO Universal Safety Similarly, navigation aids are being supplanted by Oversight Audit Programme." Presentation to technologies based on global positioning systems the ICAO TCB Seminar, Singapore, January in modern aircraft. 12­14, pp. 9 and 11. http://www.icao.int/icao/ en/tcb/TCB-Singapore-2004/Attachements/ 3. In a recent accident in Cameroon involving a Presentations/Acrobat/ICAO%20USOAP%20 new Boeing 737, the aircraft could not immedi- &%20Follow-up%20Programme.pdf. ately be located because the last known position was the departure end of the runway. ------. 2008. "Financial Situation of Airports and Air Navigation Service Providers 2005." ICAO, Montreal, Canada. References Schlumberger, Charles E. 2008. "Air Transport Abate, Megersa A. 2007. "The Economic Effects Policy Research: The Implementation of the of Progressive Air Transport Liberalization Yamoussoukro Decision." Working Paper 20, in Africa." School of Graduate Studies, Addis Africa Infrastructure Country Diagnostic, World Ababa University, Ethiopia. Bank, Washington, DC. Chapter 14 Water Resources: A Common Interest W ater management is critical for Building water infrastructure will fuel meeting Africa's development chal- growth, reduce weather-induced risk, and lenges. Though water is vital for alleviate water-related conflicts. Both large agriculture, only 5 percent of Africa's cultivated and small infrastructure projects need to be land is irrigated. Hydropower is also largely part of a balanced water investment program undeveloped in Africa; less than 10 percent of that provides reliable water supplies for human its potential has been tapped. Water for people health and economic activities and that pro- and animals is vital for health and livelihoods, tects natural water and environmental assets. yet only 58 percent of Africans have access to Development of large multipurpose storage safe drinking water. facilities (often combined with hydropower African economies depend on a reliable generation) is necessary for mitigating the and adequate supply of water, but high rain- economic effects of hydroclimatic variabil- fall and hydrological variability result in fre- ity, for ensuring reliable water supply, and for quent droughts and floods that stifle economic using available water. Small-scale approaches growth. Moreover, water resources shared by to water management improve the ability of countries pose complex political and manage- the rural poor to cope with water shocks by ment challenges. increasing agricultural productivity and pro- Achieving water security to support growth viding cost-effective water supply and drought and to build climate resilience is at the heart of mitigation. Sound water management institu- water resource management in Africa. Water tions are necessary to ensure sustained returns security reflects a country's ability to function on infrastructure investments and to optimize productively in the face of water vulnerability. the use of the water by multiple users and It is a precondition for sustaining and increas- across administrative and political borders. ing investment returns and achieving dynamic The estimated annual capital cost of water economic growth. A minimum capacity of resource infrastructure is approximately $10 infrastructure and institutions, backed by billion, of which almost 80 percent is for robust water information systems, is needed development of large multipurpose hydro- to ensure basic national water security. power storage, and about 10 percent each is 271 272 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION for development of large storage capacity for of about 3­4 degrees Celsius by the end of urban water supply and investment in devel- the 21st century, compared with the period oping small-scale infrastructure projects. As from 1980 to 1990 (IPCC 2007). The semiarid a complement to these physical investments, margins of the Sahara and the central part of Africa will need an additional $1.0 billion a southern Africa will be most affected, whereas year to develop hydrological networks, meet equatorial latitudes and coastal areas will be gaps in water information, and develop water least affected. The Intergovernmental Panel management institutions. on Climate Change projects that the mean annual precipitation and runoff will decrease for northern and southern Africa and increase Water Resources and Economic for eastern Africa by midcentury. Precipitation intensity will also likely increase for the entire Development: Challenges continent; the benefit of increased rainfall in for Africa the wetter areas may be negated by the rainfall being concentrated in more extreme weather Africa faces difficult water legacies in the form events and thus less usable. In the drier areas, of high hydrological variability and a mul- the spread between high and low runoff will tiplicity of transboundary river basins. Both likely increase, substantially complicating the challenges can be impediments to the conti- challenge of water resource management. nent's economic growth. The second challenge is Africa's political and geographic legacy in which several countries Africa's Water Legacies share the same river basins. Africa has more Africa faces a complex challenge in water than 60 transboundary rivers, with many coun- resource management because of two legacies. tries sharing the same basin. International river The first is its natural legacy of high hydro- basins cover more than 60 percent of the con- climatic variability. The amount of water in tinent, and virtually all the region's rivers cross Africa is comparable to that in other regions several borders: the Nile crosses 10, the Niger 9, of the world; the continent has 9 percent of the Senegal 4, and the Zambezi 8 (figure 14.2). the world's water resources and 11 percent of Shared water resources present a manage- its population. However, Africa's water endow- ment challenge and require investment in ment conceals the fact that rainfall across much transboundary water management capacity of the continent is variable and unpredictable, and institutions, even if they also offer oppor- both between and within years. Interannual tunities for joint action and cooperation. rainfall variability in Africa, especially in east- Cross-border rivers have further implications ern and southern Africa, is high. These regions for regional security and development, partic- experience year-to-year variations exceeding ularly as Africa tries to develop and manage its 30 percent around the mean, a rate much water resources for economic development. greater than the temperate climates in Europe These legacies, compounded by underde- and North America (figure 14.1). High seasonal velopment of water infrastructure, present variability compounds these effects, causing significant social, economic, and political risks. droughts and floods. Runoff in Africa is extraor- The region's weak capacity to buffer the effects dinarily low, only half that in Asia, Australia, of hydrological variability creates uncertainty Europe, and North America, despite having the and risk for economic activity. The expectation same average precipitation. Low runoff coupled of variability and unpredictability in rainfall with high rainfall variability explains the unpre- and runoff can encourage risk-averse behav- dictable, and relatively low, seasonal and annual ior at all levels of the economy. It discourages flows in many African rivers. investment in land, advanced technologies, or Climate change is expected to increase agriculture. An unreliable water supply is also this variability. The Intergovernmental Panel a significant disincentive for investments in on Climate Change forecasts that Africa will industry and services. Growing demands for experience a significant rise in temperature water generate competition over water use, Water Resources: A Common Interest 273 Figure 14.1 Interannual Hydroclimatic Variability in Africa, by Selected Regions and Countries a. The Sahel a. Sudan 80 150 mean monthly variability (mm) annual rainfall anomaly (%) 60 130 40 110 20 90 0 70 ­20 50 ­40 30 ­60 10 ­80 ­10 1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12 b. East Africa b. Uganda 80 200 mean monthly variability (mm) annual rainfall anomaly (%) 60 180 160 40 140 20 120 0 100 ­20 80 60 ­40 40 ­60 20 ­80 0 1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12 c. Southeast Africa c. Malawi 80 350 mean monthly variability (mm) annual rainfall anomaly (%) 60 300 40 250 20 200 0 150 ­20 ­40 100 ­60 50 ­80 0 1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12 mean + standard deviation mean mean ­ standard deviation Source: Hulme and others 2001. posing social risks to poorer communities. those years (World Bank 2004). Mozambique's Weak international relations and institutions GDP growth is reduced by over 1 percentage for managing the international river basins point annually because of water shocks (World may cause regional tensions and possible con- Bank 2007). In Zambia, a study of how hydro- flicts among riparian countries. logic variability affects the economy found that rainfall variability will cost the country $4.3 Effects on Economic Development billion in lost GDP over 10 years, and it lowers Hydrological variability causes significant eco- the country's agricultural growth by 1 percent- nomic loss and constrains growth (Grey and age point each year (World Bank 2008). Sadoff 2006b). Africa lacks the capacity to buf- Underdevelopment of water resources leads fer the shocks of frequent droughts and floods. to underuse of economic potential. Water The abundance or shortage of rainfall affects resources in Sub-Saharan Africa compare well national agricultural outputs. In Kenya, losses in absolute terms with other countries in the from flooding caused by El Niño in 1997­98 world. The region has 9 percent of the world's and drought caused by La Niña in 1998­2000 water resources and about 6,000 cubic meters ranged from 10 to 16 percent of GDP during of annual renewable water resources per capita 274 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 14.2 Africa's International River Basins Weak institutional capacity in river basin planning and management creates potential conflicts and lost benefits. Conflicts are emerg- ing over water allocation and use in different parts of the region. Competing claims over water have been asserted over many of the lakes (for example, Lake Victoria and Lake Tana) MAURITANIA and river basins for economic and environ- CAPE VERDE MALI NIGER CHAD ERITREA mental uses. The growing demand for water SENEGAL GAMBIA GUINEA-BISSAU BURKINA FASO SUDAN DJIBOUTI from the major sectors of African economies, GUINEA SIERRA LEONE BENIN TOGO NIGERIA ETHIOPIA SOMALIA especially agriculture, imposes a serious con- CÔTE D'IVOIRE GHANA LIBERIA CENTRAL AFRICAN REPUBLIC CAMEROON straint on the medium- and long-term growth EQUATORIAL GUINEA in water availability in some river basins. The SAO TOME AND PRINCIPE GABON UGANDA RWANDA KENYA expected growth in hydropower production CONGO, DEM REP CONGO BURUNDI SEYCHELLES will likely require an increase in peak capac- TANZANIA ity that will add to the competition. However, COMOROS the water management institutions needed to ANGOLA MALAWI MAYOTTE SAINT HELENA ZAMBIA address these conflicts in many African coun- ZIMBABWE MOZAMBIQUE MADAGASCAR tries are weak and fragmented. Agencies with NAMIBIA BOTSWANA REUNION authority over a particular economic sector often make uncoordinated decisions about SWAZILAND SOUTH AFRICA water allocation and use, which lead to inef- LESOTHO international river basins ficiency and degradation of the resource. The absence of water rights regimes and incentives for efficient water allocation and conservation Source: Oregon State University 2005. Reprinted with permission. contributes to the problem. Cooperative management of water resources compared with Asia's 4,000 cubic meters and in international river basins is necessary to the Middle East and North Africa's 1,500 cubic increase the basins' yield of food, power, and meters. However, Africa has the lowest water economic opportunities while strengthening withdrawal per capita in the world (about environmental sustainability and mitigat- 170 cubic meters) because of hydrologic ing the effects of droughts and floods. The variability, underdeveloped water infrastruc- cost of noncooperation is high, including ture, and current water resource manage- the economic cost of negative environmental ment. Asia and Europe use about three times impacts, suboptimal water resource develop- the water per person; the water-scarce Middle ment, political tensions over shared resources, East withdraws more than four times as much and the forgone benefits of joint water and North America more than eight times as resource development (Sadoff, Whittington, much (table 14.1). and Grey 2003). The low level of water mobilization leads Lack of water infrastructure and inadequate to underdevelopment of Africa's economic water management mostly affect the poor. potential: less than 5 percent of cultivated land Africa's poverty is closely linked to its depen- is irrigated, and less than 10 percent of hydro- dence on rain-fed subsistence farming. About power potential is developed (figure 14.3). Irri- 28 percent of Africa's working population is gated land contributes only about 10 percent of engaged in agricultural production, ranging the agricultural production in Africa, with only from 4 percent in South Africa to 47 percent about 8.5 percent of cultivated land irrigated in Rwanda (You 2008). Because subsistence (figure 14.4). Less than 58 percent of Africa's agriculture is the dominant livelihood, rain- population has access to drinking water, fall, droughts, and floods, combined with the and 31 percent to sanitation services (WHO/ weak marketing network and difficult physical UNICEF 2006). access to many areas, affect food security across Water Resources: A Common Interest 275 Table 14.1 Water Availability and Withdrawal cubic meters per person Per capita actual renewal Per capita annual total Region water resources water withdrawals Asia (excluding the Middle East) 4,079.0 631 Central America and the Caribbean 6,924.4 603 Europe 10,655.1 581 Middle East and North Africa 1,505.0 807 North America 19,992.5 1,663 South America 47,044.0 474 Oceania 54,636.8 900 Sub-Saharan Africa 6,322.5 173 Source: FAO 2003. the region. These factors, along with limited Figure 14.3 Africa's Hydropower Potential irrigation and underdevelopment of water infrastructure, increase the rural economy's vulnerability to water shocks. Annually, some 220 million Africans are exposed to drought, and more than 1.5 million were affected by floods in 2007. Degradation of water catchments under- MAURITANIA mines investments already made in water CAPE VERDE MALI NIGER CHAD ERITREA resources. Loss of vegetation, erosion, and GAMBIA SENEGAL GUINEA-BISSAU BURKINA FASO SUDAN DJIBOUTI sedimentation are major threats to surface- GUINEA SIERRA LEONE BENIN TOGO NIGERIA SOMALIA ETHIOPIA water resources, because they cause lower base CÔTE D'IVOIREGHANA LIBERIA CAMEROON flows and higher flood peaks. Poor manage- EQUATORIAL GUINEA CENTRAL AFRICAN REPUBLIC ment of Africa's water catchments has led to SAO TOME AND PRINCIPE GABON UGANDA RWANDA KENYA excessive soil erosion, increased costs of water REPUBLIC OF CONGO CONGO, DEM REP BURUNDI SEYCHELLES treatment, rapid siltation of reservoirs, decline TANZANIA in economic life, and disruption of water sup- COMOROS ANGOLA MAYOTTE plies. In Malawi, any new dam is expected to ZAMBIA MALAWI fill with sediment within a few years of com- ZIMBABWE MOZAMBIQUE MADAGASCAR missioning. The most important dams in the BOTSWANA REUNION NAMIBIA country are the hydropower facilities on the Shire River, which are badly affected by sedi- SOUTH AFRICA SWAZILAND LESOTHO mentation. In Kenya, the rate of sediment out- flow from the Athi-Galana-Sabaki River into the Indian Ocean increased from about 50,000 Source: International Energy Agency 2006 (EI-Gazzar and others 2007, p.7). tons a year in the 1950s to 8.4 million tons a year by 1992 (World Bank 2004). In countries with degraded water catchments, develop- Regular monitoring of the hydrological sys- ment of water storage infrastructure needs to tem (such as meteorological stations, rain be accompanied by improved protection of gauges, and river flows) is steadily declin- watersheds to sustain the investment. ing, and most African countries have not Reliable hydrological and water quality updated their assessments. Africa also lags data are needed for effective water resource behind the rest of the world in the number management and informed decision making. of meteorological stations where data can be Hydrographic networks are outdated or in systematically collected for dissemination to need of rehabilitation in many countries. users. According to the World Meteorological 276 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 14.4 Sub-Saharan Africa's Irrigation Potential water vulnerability. It is a precondition for sus- 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° taining and increasing investment returns and 0 500 1,000 1,500 KILOMETERS achieving dynamic economic growth. Mediterranean MOROCCO TUNISIA Sea 0 500 1,000 MILES The more vulnerable an economy is to 30° Canary Is. (Sp) 30° water variability, the greater is the required ALGERIA FORMER SPANISH LIBYA ARAB REP. OF investment to achieve water security. If a EGYPT SAHARA country cannot provide water security, it will Re d 20° 20° not be resilient to water shocks, and it will not Se CAPE MAURITANIA a VERDE SENEGAL MALI NIGER CHAD ERITREA have reliable water supplies (figure 14.5, sce- THE GAMBIA GUINEA-BISSAU BURKINA FASO SUDAN DJIBOUTI nario 1; Subramanian, Yu, and Dankova 2008). GUINEA 10° 10° When an acceptable level of water security is BENIN NIGERIA TOGO GHANA ETHIOPIA SIERRA LEONE CÔTE D'IVOIRE CENTRAL LIBERIA AFRICAN REPUBLIC SOMALIA achieved along with basic climate resilience CAMEROON EQUATORIAL GUINEA SÃO TOMÉ AND PRÍNCIPE UGANDA and risk mitigation, vulnerability is no longer KENYA CONGO 0° ATLANTIC OCEAN GABON DEMOCRATIC REPUBLIC RWANDA 0° a severe constraint on growth. Beyond this INDIAN Cabinda OF CONGO BURUNDI OCEAN point, further investment in water infrastruc- (ANGOLA) TANZANIA SEYCHELLES ture contributes to economic growth (figure 10° 10° ANGOLA COMOROS Mayotte 14.5, scenario 2). Climate change will likely MALAWI Water-Managed Areas ZAMBIA (Fr) impose additional costs for achieving and sus- as Percentage of Irrigation Potentials ZIMBABWE MOZAMBIQUE MADAGASCAR taining water security through its effect on the 20° 20° 76%­100% NAMIBIA BOTSWANA MAURITIUS spatial and temporal pattern of water demand 51%­75% SWAZILAND and availability, as well as by increasing hydro- 26%­50% 10%­25% SOUTH LESOTHO logical variability in certain areas. 30° 30° ­10% AFRICA Water security is a dynamic state. Its defi- INTERNATIONAL BOUNDARIES This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information nition varies in different parts of the world, Source: Aquastat 2005 shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. reflecting geographic, social, and political 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° factors and the stage of economic develop- ment. In Africa, hydrologic variability and Source: ECA 2006. extremes are at the heart of water vulnerability because they weaken growth and retard devel- opment. African countries must achieve water Organization, Africa has only 1,150 observation security to keep the risks of droughts, floods, stations--one-eighth the recommended num- and unreliable water supplies at a socially and ber (UNFCCC 2006). economically acceptable level. Several studies have pointed to the strong correlation between rainfall variability and national GDP in coun- Addressing the Challenges tries as diverse as Ethiopia, Kenya, Lesotho, Mozambique, and Zambia (World Bank 2004, Achieving water security is a prerequisite for 2006, 2007, 2008, 2009). Emerging evidence attracting investment and promoting eco- nomic growth in Africa. Figure 14.5 Water Security and Growth Water Security, Investments, and GDP Growth Minimum water infrastructure and institu- Scenario 2: water enhances growth... tional capacity are required to ensure basic ... basic water security is achieved national water security (Grey and Sadoff 2006b). Water security is the ability of a coun- Scenario 1: water- vulnerable economy... try to function productively despite its inherent ...with climate change water vulnerability. A country's water security years depends on both its inherent water supply-and- demand patterns and its capacity to confront Source: Authors' elaboration. Water Resources: A Common Interest 277 also indicates a correlation between availability BOX 14.1 of water storage and road infrastructure and inflows of foreign direct investment (Brown and others 2008). The Metolong Dam and Water Supply Program in Lesotho Achieving Water Security: Priority Development of Lesotho's water sector is part of the government's Areas for Action effort to diversify the economy and improve the provision of essential Balanced investments in water resource infra- services. structure and institutions are needed to increase The greater Maseru area is the center of the country's garment productive uses of water, to mitigate the effect industry. It includes more than 50 firms employing approximately of recurrent floods and droughts, and to 50,000 people. This industry has resulted in an almost fourfold achieve basic water security as a platform for increase in exports since 2000, and it contributed $567 million in Africa's economic growth. Priority should be foreign exchange earnings in 2006 (38 percent of GDP). Water and wastewater services are essential to the continued economic contribu- given to investments that (a) focus on growth, tion of these companies. They currently account for half of all water (b) reduce rural poverty, (c) build climate consumed in Maseru, and the lack of water and wastewater infra- resilience and adaptation, and (d) foster coop- structure presents a major constraint to continued economic growth. eration in international river basins. The industrial growth has also stimulated an increase in urban migra- tion. Maseru, the capital of Lesotho, experienced population growth Focusing on Growth of 5.5 percent a year between 1996 and 2006, expanding to more Governments should link their water interven- than 350,000 people. Currently, the only source of raw water for treatment and supply to Maseru comes from a single intake on the tions with their development objectives and Mohokare (Caledon) River, which is unreliable, inadequate, and of poverty-reduction targets. To achieve higher inferior quality. economic returns and promote national To address the need for long-term, secure water supplies to the growth, focusing infrastructure investments lowlands, the government commissioned a feasibility study of the around main growth centers and along pri- Lesotho Lowlands Water Supply Scheme in 2004 with support from mary development corridors where produc- the European Development Fund. That study, as well as a separate tion, industrial development, and trade are study financed by the Arab Bank for Development in Africa in 2003, concentrated makes sense. The availability identified the construction and implementation of the Metolong of reliable water supplies in growth cen- Dam and Water Supply Program as the least-cost long-term solution ters protects investments against the risks for bulk water supply to Maseru and the surrounding lowlands. With of hydrologic variability and improves the the ability to provide 75,000 cubic meters of treated water a day, the dam, in conjunction with existing supplies, will ensure a secure performance of the manufacturing and ser- supply of 115,000 cubic meters of water per day, enabling Maseru vice sectors. It attracts new investments and to meet domestic and industrial requirements for at least the next enables industries to adopt water-saving tech- 40 years. The construction of the Metolong Dam is scheduled to nologies when the economic incentives are in begin in 2012. place. For municipal utilities, a mix of hydro- logical risk mitigation measures and rising Source: World Bank 2009. industrial and domestic water demands in the growth areas allows them to take advantage of economies of scale in production and dis- agriculture are major constraints to poverty tribution, to extend coverage, and to improve reduction in rural Africa. the systems' operation and maintenance. An A recent Food and Agriculture Organiza- example of a growth-oriented policy is the tion report (Faurès and Santini 2008) describes Metolong Dam and Water Supply Program in potential reductions in rural poverty through Lesotho (box 14.1) water resource interventions in Africa, specifi- cally in southern and eastern Africa and along Reducing Rural Poverty the east-west central belt (figure 14.6). These Water is an important asset for the rural poor areas have high levels of rural poverty, broad in Africa. However, high rainfall variability and opportunities for agricultural growth, and insecure access to water for consumption and sufficient water in absolute physical terms. 278 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 14.6 Water Interventions and Poverty Tunisia Morocco Algeria Libya A.R. of Egypt Former Spanish Sahara Mauritania Mali Niger Chad Sudan Eritrea Senegal Gambia Guinea-Bissau Djibouti Guinea Burkina Faso Nigeria Somalia Sierra Leone Togo Central African Republic Liberia Ethiopia Côte d´Ivoire Cameroon Equatorial Guinea Uganda Congo Kenya Gabon Rwanda DR Congo Burundi legend Tanzania potential for poverty reduction high Angola moderate Zambia low Zimbabwe country boundaries Namibia Botswana Mozambique rivers waterbodies Swaziland Madagascar Lesotho South Africa N Kilometers W E 0 250 510 1,000 1,500 2,045 S Source: Faurès and Santini 2008. Water Resources: A Common Interest 279 However, at the same time, water is a limiting environmental sustainability and mitigating factor for these rural livelihoods because of the effects of droughts and floods. Countries significant seasonal and interannual variabil- around the Senegal and Niger Rivers have ity and lack of water control. Such zones offer started to reap the benefits of a cooperative the greatest opportunities for expanding food and joint approach to managing their shared production through irrigation, rain-fed agri- waters (Andersen and others 2005; Yu 2008). culture, rainfall harvesting, and conservation Countries surrounding Lake Victoria are gain- of soil moisture. ing environmental benefits from their joint Community-based management and devel- efforts to eliminate encroaching water hya- opment of local watersheds and groundwater cinth and other weeds. The systemwide yield resources in the poorest areas are crucial for of water in the Nile could likely be increased by sustaining local livelihoods against the effects several percentage points a year if cooperation of climate variability. These measures include led to water storage upstream and coordinated constructing small hydraulic structures, intro- reservoir operation in the arid plains down- ducing local-scale hydropower units, harvesting stream (Sadoff, Whittington, and Grey 2003). water, developing smallholder irrigation, and The countries will also likely benefit from joint installing flood protection measures. investments in water infrastructure, thereby reducing infrastructure costs and maximizing Building Climate Resilience returns. Unilateral and uncoordinated deci- Managing rainfall variability is a critical task sions on water resources by riparian countries, for African countries. Expanding water storage in contrast, may lead to increased political capacity (as well as improving management tensions and conflicts, reducing development of the existing hydraulic structures) would opportunities in the basins. mitigate the effects of water shocks and build climate resilience. Along with large storage and hydraulic structures, small multiobjective Investing in Africa's Water water development is an important opportu- Security nity for new investment. These investments should take into account the potential effects Because most African countries have low of climate change. Appropriately designed irri- stocks of hydraulic infrastructure, emphasiz- gation investments would increase agricultural ing investments in infrastructure is appropri- productivity and significantly mitigate the dan- ate for them. However, institution building ger of rainfall variability. Investments to reduce and reform, improvements in water manage- climate vulnerability should also include ment and operations, and strengthening of hydrometeorological services and monitoring, water information systems must complement catchment protection and management, and growth in infrastructure. Development of risk assessment and mitigation. institutions is a lengthy and costly process, and adequately sequenced and balanced, it should Fostering Transboundary Cooperation be advanced in parallel with infrastructure Optimizing economic productivity and envi- investments, paying particular attention to the ronmental sustainability means managing riv- development of river basin organizations. ers as hydrological units at the basin level. The great hydrologic challenges that countries face Investing in Storage Infrastructure: at the national level of water resource manage- Both Large and Small ment (such as rainfall and runoff variability, Development of water infrastructure is a pre- degrading water quality, and flood protection) requisite for water security in Africa and for create opportunities to gain significant bene- meeting the targets of the eight Millennium fits from cooperation over shared river basins. Development Goals. Artificial water storage of Cooperative management of water resources adequate capacity is needed to ensure reliable increases a basin's yields of food, power, and water supply during droughts and to retain other economic goods while strengthening excessive water during periods of flooding. 280 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Despite Africa's vulnerability to frequent The direct and indirect long-term eco- droughts and floods, storage capacity remains nomic benefits of investing in large storage underdeveloped. Average per capita storage capacity are many, but they require consider- capacity in Africa is about 200 cubic meters a able initial capital investments. The aggregate year, much less than that of countries in other spending needs for African infrastructure regions (figure 14.7). were reported in chapter 1 of this volume, but Irrigation and hydropower have been the because water resource investments are typi- main drivers for dam construction in Africa, cally buried within investment programs for but a storage facility designed for a single other sectors, such as irrigation, power, and purpose has limited capacity to serve other water supply, the specific water resource com- economic and social needs, and its invest- ponents are explicitly highlighted here. ments often have higher opportunity costs. Table 14.2 details the component of power For example, the Cahora Bassa Dam on the sector investment needs explicitly attribut- Zambezi River was constructed for the sole able to water storage in large dams. If regional purpose of generating hydropower. The dev- power trade could be effectively harnessed, astating 2000 floods in Mozambique showed some 50,000 megawatts of new hydropower that the dam could play an important flood capacity would need to be built from 2006 to mitigation role, but its operational rules do not 2015, but without expanding regional trade, permit use of its storage capacity to mitigate only 33,000 megawatts of hydropower capac- water shocks. To convert this dam into a multi- ity could be developed. The trade scenario purpose reservoir would be too costly. Despite would translate into an annual average invest- the need for power, other considerations, such ment requirement of $7.8 billion in large- as flood control, salinity repulsion, irrigation scale water storage to support generation of development, and environmental require- electricity over the 2006­15 period (Rosnes ments, call for multipurpose storage develop- and Vennemo 2008, 2009). ment. Multipurpose water projects generally Not only would expanded regional trade result in optimal water development, maxi- lead to the development of more water storage, mize economic returns on investments, and but it would also improve the cost-effectiveness need to be implemented in the basinwide con- of water storage. The unit capital cost of text. Two important principles for developing hydropower investments would fall from large water infrastructure are equitable shar- $5.9 million to $5.4 million per megawatt ing of the benefits with the people affected and (table 14.2) because cross-border collabora- mitigation of possible negative environmental tion allows larger and more efficient storage effects. Thus, stakeholder participation is nec- sites to be developed. essary at all stages of decision making, project The key challenge is how to finance the design, and implementation. multibillion-dollar large-scale water storage projects needed to make this savings a real- Figure 14.7 Water Reservoir Storage per Capita in ity. Often the countries with the best storage Selected Countries, 2003 sites are those with the least financial capacity 7,000 to develop them. Regional collaboration offers 5,961 6,000 the possibility of cost-sharing arrangements 5,000 4,717 among countries for large water infrastructure, cubic meters 4,000 3,386 allowing downstream beneficiary countries 3,000 2,486 with greater solvency to provide up-front capi- 2,000 tal contributions. For example, a cooperative 1,104 1,277 1,000 687 effort by Burundi, Rwanda, and Tanzania to 38 0 design and build the Rusumo Falls hydroelec- ia a ico d ina il a ica tric project on the Kagera River could bring az ric ali an iop er ex Ch Br Af str ail Am h M Th Au Et h 60 megawatts of renewable power to an area ut rth So No where only 2 percent of households have access Source: Grey and Sadoff 2006a. to electricity. Water Resources: A Common Interest 281 Table 14.2 Capital Investment Needs in Large Multipurpose Hydropower Storage by 2015 Large storage-based hydropower Capital costs per unit Total capital costs production (megawatts) ($ millions per megawatts) ($ billions per year) Regional power No power trade Regional power No power trade Regional power No power trade Region trade expansion expansion trade expansion expansion trade expansion expansion Southern Africa 16,764 10,797 0.96 1.21 1.95 1.59 Eastern Africa 10,675 4,170 1.81 1.96 2.35 0.99 Western Africa 17,260 14,845 1.32 1.37 2.77 2.48 Central Africa 4,847 3,567 1.28 1.37 0.75 0.59 Total 49,546 33,379 5.37 5.91 7.82 5.65 Source: Authors' calculations based on Rosnes and Vennemo 2008. Note: These investment cost estimates are based on actual planned large storage-based (multipurpose) hydropower projects presented for each region in Rosnes and Vennemo 2008 and 2009. Totals do not add exactly because of averaging. Governments have traditionally shouldered Studies in Botswana and Zimbabwe show that the responsibility for financing large hydrau- 80­85 percent of all measurable rain can be lic structures, including most multipurpose collected from outside catchment areas and dams. Private sector participation in such stored (Dixit and Patil 1996). investments is possible, however, for mul- The estimated cost for small reservoir facili- tipurpose projects with revenue-creating ties is based on an assessment of small-scale components (such as hydropower produc- irrigation development potential from 2006 tion, commercial irrigation, and urban water to 2015 (You 2008). This analysis used a spa- supply functions) combined with nonrevenue tial allocation model on a 10-kilometer global and public-good functions (such as flood grid, considering economic profitability, crop control, fish breeding, recreation, small-scale pattern, prices, crop water productivity, water irrigation, and so forth). Promoting public- balance and availability, and distance to mar- private partnerships to finance multipurpose ket for each pixel. A macroscale hydrology water resource projects is a key challenge of model used climatic data to calculate runoff decision makers. for each pixel. The total small-scale irrigation A continuum of options from large to area was estimated at about 10 million hect- small infrastructure projects needs to be part ares, and the total water storage requirement of a balanced investment approach. Small was estimated at about 35 billion cubic meters water storage facilities can increase climate for the countries sampled. The total cost for resilience and improve food security. Small- small irrigation storage facilities was based on scale approaches to water management help the irrigation development area, required stor- the rural poor by providing cost-effective age volume, and their average costs. solutions to water supply and drought miti- A certain amount of storage is also needed gation. They improve the ability of the rural to keep pace with growing demands for the poor to address food insecurity by increas- urban water supply. The associated invest- ing agricultural productivity. Small storage ment can be estimated based on demographic options include (a) off-stream reservoirs, (b) growth, Millennium Development Goal tar- on-farm ponds and networks of multipurpose gets, trends in water consumption, as well as small reservoirs, (c) groundwater storage, and unaccounted for water and the availability of (d) water storage through a root zone with a supply from various sources. The storage need variety of water-harvesting techniques and is put at 5.4 billion cubic meters. conservation of soil moisture. Structures to Total capital investment needs for the devel- harvest rain require little space and are not opment of water resource infrastructure for labor intensive, but they must be designed 2006­15 are estimated at approximately $10 bil- in accordance with intended usage and local lion a year, included within estimates for power circumstances to ensure cost-effectiveness. ($7.8 billion), water supply ($1.3 billion), and 282 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION irrigation ($0.8 billion) presented in chapter 1 arrangement envisages stakeholder forums to of this volume. address multiple uses and optimize benefits across sectors and administrative jurisdictions. Investing in Institutions and South Africa has enacted far-reaching water Information: Managing Water legislation, and Botswana is restructuring its Resources across Sectors and water-related institutions. African countries Jurisdictions have relatively rich experience in managing Sound water management institutions in water supply and sanitation through utilities. Africa will ensure sustained returns on This is not the case with water resource man- infrastructure investments and optimize the agement, where for the majority of African allocation and use of water by multiple eco- countries, robust and sustainable institutions nomic sectors and across administrative and have yet to be developed. political borders. Africa has a longer history of developing Despite low water use for productive pur- regional institutions to manage transboundary poses in Africa, conflicts are emerging over waters. The first regional river basin organiza- water in areas of concentrated economic tions were established in the mid-20th century activity. These conflicts often intensify during (the Niger River Commission in 1964, later periods of water shortages. The drop in Lake transformed into the Niger Basin Authority Victoria's level because of increased withdraw- in 1980; the Senegal River Basin Organization als and the drought of 2003­06 affected other in 1972; and the Gambia River Development users of water around the lake. In Ethiopia's Organization in 1978). Burundi, Rwanda, and Lake Tana, reduced inflows and increased Tanzania established the Kagera Basin Orga- water abstractions in 2003­04 highlighted the nization in 1977. The Lake Chad Basin Devel- need for coordinated water planning and man- opment Fund was set up in 1973 to support agement. The need for improved water alloca- the Lake Chad Basin Commission's activities. tion regimes for multisectoral use can be seen More recently, the riparian countries of the in the conflicting water demands in the Inner Nile formed the Nile Basin Initiative in 1999 Niger Delta in Mali and in the emerging com- as a prelude to a more permanent institution. peting claims of irrigation expansion plans, In 2005, the countries around Lake Victoria hydropower production, and environmental established the Lake Victoria Basin Commis- water demands in the Kafue basin in Zambia. sion under the auspices of the East African Establishing priorities for water investments Community. and clear policy rules to govern optimal water Despite these early starts, with only a few allocation across economic sectors will be exceptions, these transboundary river basin key to enabling or constraining their relative organizations are still in their emerging stages growth. All of this should result from sound and remain relatively weak (UN-Water/Africa medium-term river basin planning. 2006). They suffer from waning political com- Full realization of Africa's water potential mitment, poor cooperation, management and the optimal allocation of water among and technical difficulties, armed conflict and various sectors require the right institutional political instability in member states, lack of arrangements at the national level, includ- defined goals or weak incentives for regional ing (a) capable water management organiza- cooperation, and insufficient capacity to carry tions, (b) provisions for public participation out their plans. As donor support dwindled, in water management decisions, (c) water basin organizations had insufficient financial rights regimes, and (d) tailored incentive sys- backing to carry out their programs. Today, the tems. Many African countries are beginning to organizations are at various stages of develop- develop national institutions for water man- ment (figure 14.8). agement. Tanzania, for example, has identified In a few cases, however, river basin organiza- nine river basins for which it will develop sus- tions have been backed by strong government tainable development plans. This institutional ownership and commitment and have enjoyed Water Resources: A Common Interest 283 support from multiple donors, enabling them Figure 14.8 Degree of Regional Water Cooperation to successfully complete several years of insti- Toward regional cooperation on shared waters tutional development, confidence-building Nile Basin Initiative Niger Basin Authority Senegal River Basin Authority measures, and cooperative investment pro- · 9 countries + 1 observer · 9 countries · 4 countries gramming. The Niger Basin Authority, for · established 1999 · established 1963, revitalized 2001 · established 1972 example, undertook an institutional assessment · no formal treaty framework · formal treaty framework · formal treaty framework and · confidence building and and formal river basin formal river basin organization to prepare for reorienting its efforts in early investment programs in place organization · sustainable development 2000. Subsequently, it introduced an organi- · sustainable development plan in place zational structure approved by the Niger Basin plan in place Council of Ministers, competitively recruited a team of professionals, and strengthened its fiduciary systems. The authority has also agreed on a water charter that spells out the rules of National Consultative Cooperative Cooperative Joint engagement for the riparian countries, includ- demand, planning, planning, planning and ownership national exploratory regional management, of assets, ing procedures for sharing information and market regional demand and regional joint formulating investment programs. The nine focus dialogue market optimization management considerations countries on the Niger have agreed to a 20-year sustainable development action plan and a Source: Adapted from Grey and Sadoff 2006a. related investment program. The experience of these organizations indi- cates that the process of river basin development such as sediments and fisheries. The pro- requires political leadership, government com- cess should include computers to analyze mitment, confidence building among countries, and store information. It will also include and realization of concrete benefits. Substantial support services for periodic calibration of investments are needed for assessments, project instruments. preparation, and feasibility studies. Invest- ments in facilitating greater regional coopera- · Spatial data: Satellite imageries, aerial sur- veys, ground surveys, bathymetry, and tion within the river basin organizations will other data sets from geographic informa- also be critical. These investments can support tion systems. a regional forum for dialogue, conflict resolu- tion, and cooperation around a shared resource · Information management: Hydrologi- (Sadoff and Grey 2005). cal design aids (such as maps of prob- Relevant, adequate, and reliable informa- able maximum precipitation and regional tion enhances institutional capacity for deci- flood frequency studies) for use by project sion making. An effective water information designers. system requires action on both the demand · Knowledge management: A model and deci- and supply sides. Planners and decision sion support system to interface with the makers must be aware of the importance of water information system. information in decision making. Information · Dissemination: Protocols to supply end managers must be able to develop the appro- users. priate mix of formal and informal knowledge and communication systems to support deci- Recent experience with water resource sion makers. Thus, a water information system projects indicates that investments in infor- should include the following components: mation and institutions will require close to 10 percent of the investment in water · Hydrological information: Collection system infrastructure. Given the estimated annual (including instrumentation, quality con- $10 billion investment for water infrastruc- trol, coding) for data capture for surface ture, close to $1 billion a year of additional water, groundwater, and water quality resources will be needed to advance the insti- parameters, and environmental information, tutional and information agendas. 284 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Note IPCC (Intergovernmental Panel on Climate The authors of this chapter are Rimma Dankova, Change). 2007. Climate Change 2007: Synthesis Satoru Ueda, Ashok Subramanian, Winston Yu, Report. Geneva: IPCC. and Jyothsna Mody, who drew on background Oregon State University. 2005. Transboundary material and contributions from Vahid Alavian. Freshwater Dispute database. http://www .transboundarywaters.orst.edu/database/. References Rosnes, Orvika, and Haakon Vennemo. 2008. Andersen, Inger, Ousmane Dione, Martha "Powering Up: Costing Power Infrastructure Jarosewich-Holder, and Jean-Claude Olivry. Investment Needs in Sub-Saharan Africa." 2005. The Niger River Basin: A Vision for Sustain- Background Paper 5, Africa Infrastructure Sec- able Management. Katherin George Golitzen, ed. tor Diagnostic, World Bank, Washington, DC. Washington, DC: World Bank. Sadoff, Claudia W., and David Grey. 2005. "Coop- Brown, Casey, Daniela Domeisen, Robyn Meeks, eration in International Rivers: A Continuum of Kenneth Hunu, and Winston Yu. 2008. "Exploring Securing and Sharing Benefits." Water Interna- the Effect of Hydro-Climatic Variability on Eco- tional 30 (4): 1­8. nomic Growth in Sub-Saharan Africa: A Water Sadoff, Claudia W., Dale Whittington, and David Security Index." Technical Report 08-04, Water Grey. 2003. Africa's International Rivers: An Eco- and Growth Report 2, International Research nomic Perspective. Washington, DC: World Bank. Institute for Climate and Society, New York. Subramanian, Ashok, Winston Yu, and Rimma Dixit, Medha J., and Subhash M. Patil. 1996. "Rain Dankova. 2008. "Sharing Water, Sharing Bene- Water Harvesting." Discussion paper presented fits, Sharing Costs." Paper presented at the Expo at the 22nd Water, Engineering and Develop- Zaragoza 2008 Water Tribune, Zaragoza, Spain, ment Centre International Conference, June 16­September 10. New Delhi, India, September 9­13. UNFCCC (United Nations Framework Conven- ECA (Economic Commission for Africa). 2006. tion on Climate Change). 2006. "Background "Water in Africa: Management Options to Paper on Impacts, Vulnerability and Adapta- Enhance Survival and Growth." Addis tion to Climate Change in Africa for the African Ababa, Ethiopia: ECA. Workshop on Adaptation Implementation of EI-Gazzar, Mohamed, Yassin Ibrahim, Mohamed Decision 1/CP.10 of the United Nations Climate Bedrous, and Aziz Maher. 2007. "Hydro Change Convention, Accra, Ghana, September and Nuclear Power for African Less-Carbon 21­23, 2006." http://unfccc.int/files/adaptation/ Development." Paper presented at the World adverse_effects_and_response_measures_art_ Energy Congress, Rome. 48/application/pdf/200609_background_afri- can_wkshp.pdf. FAO (Food and Agriculture Organization). 2003. FAO statistical database. http://www.fao. UN-Water/Africa. 2006. African Water Development org/corp/statistics/en/. Report, 2006. Addis Ababa, Ethiopia: UN-Water/ Africa. Faurès, Jean-Marc, and Guido Santini, eds. 2008. Water and the Rural Poor: Interventions for WHO/UNICEF (World Health Organization/ Improving Livelihoods in Sub-Saharan Africa. United Nations Children's Fund). 2006. MDG Rome: Food and Agriculture Organization of Assessment Report. Geneva and New York: the United Nations and International Fund for WHO/UNICEF Joint Monitoring Programme Agricultural Development. for Water Supply and Sanitation. Grey, David, and Claudia Sadoff. 2006a. "The World Bank. 2004. "Towards a Water-Secure Global Water Challenge: Poverty Growth and Kenya: Kenya Water Resources Sector Memoran- International Relations." Paper presented at dum." World Bank, Washington, DC. Global Issues Seminar Series, World Bank, ------. 2006. "Ethiopia Country Water Resources Washington, DC, January 25. Assistance Strategy." Agriculture and Rural ------. 2006b. "Water for Growth and Develop- Development Department, World Bank, ment: A Framework for Analysis." Theme docu- Washington, DC. ment of the 4th World Water Forum, Mexico ------.2007. "Mozambique Country Water City, March. Resources Assistance Strategy." Africa Hulme, Mike, Ruth Doherty, Todd Ngara, Mark Region Water Resources Unit, World Bank, New, and David Lister. 2001. "African Climate Washington, DC. Change: 1900­2100." Climate Research 17: ------. 2008. "Zambia Country Water Resources 145­68. Assistance Strategy." Draft, Africa Region Water Resources: A Common Interest 285 Water Resources Unit, World Bank, Africa Infrastructure Country Diagnostic, World Washington, DC. Bank, Washington, DC. ------. 2009. "Lesotho Water Sector Improvement Yu, Winston. 2008. "Benefit Sharing in Interna- Program (Second Phase): Metolong Dam and tional Rivers: Findings from the Senegal River Water Supply Project." Report 46272-LS, World Basin, the Columbia River Basin, and the Bank, Washington, DC. Lesotho Highlands Water Project." Working You, Liang Zhi. 2008. "Irrigation Investment Needs Paper 1, Africa Region Water Resources Unit, in Sub-Saharan Africa." Background Paper 9, World Bank, Washington, DC. Chapter Chapter 15 Irrigation: Tapping Potential A large segment of Sub-Saharan Africa's irrigation, involving an additional one-time population lives in rural areas and investment of $17.8 billion. In general, economic depends heavily on agriculture. Agricul- returns on small-scale schemes (on average tural growth is clearly key to poverty reduction 26 percent) are substantially higher than those and to achieving the Millennium Development on large-scale schemes (on average 17 percent). Goal of halving poverty by 2015. Agricultural These results, however, are critically depen- performance has significantly improved since dent on keeping investment costs down to 2000; nevertheless, agricultural productiv- best-practice levels of $3,000 per hectare for ity remains the lowest in the world. Climate the water distribution component of large- change and the global food price crisis provide scale irrigation and $2,000 per hectare for additional challenges. A comprehensive effort small-scale irrigation, rather than the signifi- is required to increase investment in agricul- cantly higher levels often observed at the proj- tural intensification. Water for agriculture is a ect level in the recent past. Another key finding critical ingredient of such programs. is that irrigation is in most cases viable only Africa's agricultural water remains com- for cash crops or high-value food crops (such paratively underdeveloped, despite economi- as horticulture) that raise revenues in excess of cally viable potential to expand irrigated areas. $2,000 per hectare; relatively few hectares are Today, only 3.5 percent of Africa's agricultural viable for irrigation of staple food crops. land is equipped for irrigation, some 7 million What would be the effect on development hectares concentrated in a handful of countries. of an aggressive expansion of Africa's irrigated However, further expansion of the irrigated agricultural area? If Africa does not increase its area would be profitable. At least 1.4 million slow growth in irrigated area, the food supply on hectares could be developed using existing or the continent will gradually diminish because planned dams associated with hydropower of climate change, leading to a huge surge in development, at a total additional one-time cereal imports and a significant increase in child investment of $2.6 billion in distribution of malnutrition. In contrast, if Africa's irrigated agricultural water. In addition, at least 5.4 mil- area could be tripled by 2050, the food supply lion hectares would be viable for small-scale would increase markedly, with a huge decline in 287 288 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION cereal imports. There would be 2 million fewer Despite these encouraging developments, malnourished children than under the lower agricultural productivity is the lowest in the irrigation scenario, or about the same level that world, with per capita output only 56 percent would be expected in the absence of climate of the world average. Output has not kept pace change. Thus, aggressive agricultural water with population increases, and growth has development could reverse the adverse effects occurred largely through expansion of har- of moderate global warming on food security. vested area (rather than through more intensive Aggressive scaling up of investments in use of existing cropland): more than 80 percent agricultural water raises issues associated with of output growth since 1980 has come from the performance and sustainability of irri- expansion of cropped area, compared with less gated agriculture. Considerations of economic than 20 percent for all other regions. viability, farm-level profitability, and sustain- Climate change and the global food price ability should guide future investment deci- crisis, developments likely to hit Africa dis- sions. In particular, investments in agricultural proportionately, further challenge agricultural water should be considered as part of a com- performance. Water sources will become more prehensive package, including (a) empowered variable. Droughts and floods will stress agri- farmer organizations; (b) sustainable, effi- cultural systems. The seas will inundate some cient, and accountable agricultural support coastal food-producing areas, and food pro- services; and (c) accessible, profitable markets. duction will fall in some places in the interior. Furthermore, efforts to expand irrigation Nevertheless, substantial uncertainty remains should be made in the context of national about where the effects will be greatest. agricultural water development strategies that A comprehensive effort is required to emphasize the importance of a more condu- advance agricultural productivity. Invest- cive institutional environment and that form ments in more reliable access to agricultural the foundation for sectoral programs that water1 are critical in support of that objective combine investment in infrastructure with (World Bank 2008). More reliable access to investment in institutional reforms. agricultural water increases the opportunity to use productivity-enhancing inputs and thus supports intensification and diversifica- Agriculture and Poverty Reduction tion, as well as the scope for agricultural wage employment. In addition, it reduces local food Agricultural growth is a key to reducing poverty. prices, improving real net incomes. It can also More than half of Sub-Saharan Africa's popula- reduce poverty indirectly through increased tion lives in rural areas, and agriculture accounts rural nonfarm and urban employment. More- for a significant percentage of GDP. Of Africa's over, investments in irrigation improve access poor, 85 percent live in rural areas and depend to markets. A more reliable, year-round supply largely on agriculture for their livelihoods. of products, a higher and more uniform qual- GDP growth originating in agriculture is about ity of products, and the option to manipulate four times more effective in raising incomes of harvest dates to capture higher seasonal prices extremely poor people than GDP growth origi- put a high premium on irrigation. nating outside the sector--and the potential African leaders have identified agricultural multipliers from agricultural water investment water development as a key area for investment. are even higher (World Bank 2008). The Comprehensive Africa Agricultural Devel- Agricultural performance in the region has opment Program prepared under the New improved significantly since 2000. Growth in Partnership for Africa's Development in 2002 agricultural GDP in Sub-Saharan Africa has adopted land and water management as the accelerated from 2.3 percent a year in the 1980s first of its four pillars for priority investment. It to 3.8 percent a year from 2000 to 2005. Aver- proposed extending the area under sustainable age incomes in Sub-Saharan Africa have been land management and reliable water control rising in tandem with those in other regions systems to 20 million hectares (more than twice since the mid-1990s. the area currently under water management in Irrigation: Tapping Potential 289 Sub-Saharan Africa) by 2015 (NEPAD 2003). In irrigation in Sub-Saharan Africa varies consid- response to this call for action, the Partnership erably by country but is generally very low, with for Agricultural Water in Africa was recently only a few countries reaching the 20 percent launched to scale up investments in agricultural mark (figure 15.1). In absolute terms, more than water and to harmonize donor programs. 60 percent of the total area is concentrated in just three countries--Madagascar, South Africa, and Sudan--each with over a million hectares Current State of Irrigation of irrigated area. Donor investments in agricultural water Irrigation carries significant potential to have declined sharply. From 1994 to 1996, the increase agricultural productivity. Across Sub- total value of projects funded by all donors for Saharan Africa, irrigated agriculture accounts irrigation and drainage was less than 10 percent for about 25 percent of the value of agricultural of the levels 20 years earlier--just $127 million output (table 15.1). This share is produced on from all sources (World Bank 2007). Significant just 3.5 percent of the cultivated land, confirm- scope exists for scaling up investments. Average ing the potential of irrigation to improve live- agricultural water withdrawals are 1.3 percent lihoods in Sub-Saharan Africa and suggesting of renewable water resources; groundwater use that more investment in irrigation would yield is less than 20 percent of renewable supplies, substantial benefits. indicating significant scope for further surface However, Sub-Saharan Africa's agricultural water and groundwater development. Expan- water remains underdeveloped. Of a cultivated sion of irrigated area has been slow in Sub- area of 197 million hectares, only 7 million hect- Saharan Africa. Over the last 40 years, only ares is equipped for irrigation, with a further 4 million hectares of new irrigation has been 2 million hectares under some other form of developed, by far the smallest expansion of any water management. Overall, this area amounts region. Over the same period, China added to only 23 percent of the 39 million hectares that 25 million hectares, and India 32 million. is believed to be physically suitable (though not Annual growth in irrigation development in the necessarily economically viable) for irrigation. region was 2.3 percent from 1973 to 2000 and The share of cultivated area equipped for slowed further from 2000 to 2003, but it has Table 15.1 Selected Irrigation Investment Indicators for Sub-Saharan Africa percentage Value of irri- Agricultural water Groundwater gated output withdrawals as Dam capacity as pumped as Average annual as percentage Cultivated area Irrigation percentage of percentage of percentage of expansion of of total value equipped for potential total renewable total available total renewable irrigated area, of agricultural Region irrigation realized water resources surface water groundwater 1973­2000 output Sudano-Sahelian 6.9 50 21.8 9.7 38.1 2.7 58.3 Eastern 2.6 11 4.9 5.5 3.1 2.4 5.0 Gulf of Guinea 1.5 7 1.2 47.1 0 2.2 6.3 Central 0.7 1 0.1 1.7 0 0.5 7.3 Southern 4.2 36 6.2 99.0 17.8 3.2 6.6 Indian Ocean Islands 30.4 71 4.2 0.1 8.7 3.5 0 Sub-Saharan Africa average 3.5 18 1.3 11.2 17.5 2.3 24.5 Asia average 33.6 67 15.8 12.0 -- 2.6 -- Source: Svendsen, Ewing, and Msangi 2009. Note: The regions shown are those adopted in Frenken (2005). The grouping of countries within these regions is based on geographical and climatic homogeneity, which directly influences irrigation. Sudano-Sahelian: Burkina Faso, Cape Verde, Chad, Niger, Senegal, Sudan; Eastern: Ethiopia, Kenya, Rwanda, Tanzania, Uganda; Gulf of Guinea: Benin, Côte d'Ivoire, Ghana, Nigeria; Central: Cameroon, the Democratic Republic of Congo; Southern: Lesotho, Malawi, Mozambique, Namibia, South Africa, Zambia; Indian Ocean Islands: Madagascar. -- Not available. 290 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 15.1 Percentage of Cultivated Area Equipped picked up in the last several years, particularly for Irrigation, by Country for the Central African Republic, Kenya, Mau- Madagascar ritius, Nigeria, Senegal, and Zambia. An appro- Swaziland priate institutional framework is required to Mauritius manage water for growth. Sub-Saharan Africa's São Tomé and Principe framework of irrigation-related institutions is Somalia undeveloped. These institutions are important Sudan for managing and sharing water over agroeco- South Africa logically and hydrologically diverse areas and Mauritania over transboundary basins. The establishment Cape Verde and sound functioning of water management Zimbabwe bodies will provide the institutional framework Guinea Mali and will specify the location of investment Senegal planning and implementation responsibilities Sierra Leone for irrigation infrastructure projects. Seychelles Guinea-Bissau Tanzania Economic Investment Potential Eritrea and Needs Zambia Mozambique Some 39 million hectares of agricultural land Ethiopia in Africa is deemed physically suitable for irri- Malawi Angola gation; however, physical suitability does not Kenya necessarily entail economic viability (box 15.1). Niger Economic potential for irrigation is highly sen- Burundi sitive to initial investment costs and requires Côte d'Ivoire market access, complementary inputs, extension Nigeria of credit, and a supportive enabling environ- Namibia ment. Both large-scale irrigation schemes (dis- Gabon tributing water collected in major dams) and Lesotho small-scale irrigation schemes (collecting water Chad locally at the farm level) are relevant for Africa. Rwanda Gambia, The Profitable dam-based, large-scale irrigation Burkina Faso appears feasible on 1.35 million hectares (table Ghana 15.2), if the attention is confined only to those Congo, Rep. of projects that pass a threshold of 12 percent for Cameroon internal rate of return. The associated one-time, Botswana on-farm investment is $2.6 billion, which would Benin nonetheless need to be spread out over at least a Togo decade. The countries with the greatest potential Liberia for dam-associated large-scale investments are Uganda Ethiopia, Nigeria, Sudan, and Zimbabwe, all with Congo, Dem. Rep. of Comoros more than 100,000 hectares of potential. These Central African Republic estimates are based on the assumption that dam 0 5 10 15 20 25 30 35 construction would be deemed profitable purely percent from a hydropower perspective and that the only Source: Svendsen, Ewing, and Msangi 2009. costs that would need to be covered by related Note: Djibouti, with 100 percent of its cultivated area equipped irrigation schemes would relate to water distri- for irrigation, is not shown here for easier visibility of other coun- tries. The percentage of area equipped for irrigation that is actu- bution infrastructure. ally irrigated spans a wide range in Sub-Saharan Africa. Overall, The small-scale irrigation potential is much the average use rate is 71 percent in the region, compared with a similar but slightly lower 67 percent in Asia. greater because large, existing rain-fed areas Irrigation: Tapping Potential 291 BOX 15.1 Assumptions for Irrigation Investment Potential Study For this study, a spatial analysis was conducted and $80 per hectare for operation and main- that combined hydrogeographic and economic tenance. Three hours' travel time to the near- parameters to estimate investment potential. est market was set as the cutoff value for Two categories of irrigation development were market access, excluding all pixels that fell assessed: (a) dam-based, large-scale irrigation beyond this range. Runoff use efficiency for associated with hydropower reservoirs (both small systems was set at 30 percent; that is, existing and planned) identified by a com- only 30 percent of the captured runoff can be panion study for hydropower; and (b) small- used for small-scale irrigation. scale irrigation based on small reservoirs, farm Large-scale irrigation assumed a medium ponds, treadle pumps, and water-harvesting investment cost of $3,000 per hectare for on- structures collecting local runoff. The main farm development, $0.25 per cubic meter for data sets were (a) spatially explicit current crop water delivery and conveyance, a proxy for distribution, (b) spatially explicit crop-specific canal operations and maintenance, and $10 biophysical potential, and (c) potential runoff per hectare for on-farm operations and mainte- and effective rainfall from a hydrological model nance. A 50-year investment horizon was used. for small-scale irrigation. Crop prices, based on Dam costs were not included because they are commodity-specific world prices for 2004­06, assumed to be fully justified and fully covered were adjusted for country differences in price by the hydropower schemes associated with the policy and market transaction costs. relevant dams. Overall irrigation efficiency for Small-scale irrigation assumed a five-year large systems was assumed to be 40 percent. cycle of investment, a medium investment cost Source: You 2009. of $2,000 per hectare for on-farm investment, Table 15.2 Potential Investment Needs for Large-Scale, Dam-Based, and Complementary Small-Scale Irrigation in Sub-Saharan Africa Large-scale irrigation Small-scale irrigation Increase in Increase in irrigated area Investment cost irrigated area Investment cost Region (million hectares) ($ millions) Average IRR (%) (million hectares) ($ millions) Average IRR (%) Sudano-Sahelian 0.26 508 14 1.26 4,391 33 Eastern 0.25 482 18 1.08 3,873 28 Gulf of Guinea 0.61 1,188 18 2.61 8,233 22 Central 0.00 4 12 0.30 881 29 Southern 0.23 458 16 0.19 413 13 Indian Ocean Islands 0.00 0 -- 0.00 0 -- Total 1.35 2,640 17 5.44 17,790 26 Source: You 2009. Note: See table 15.1 for definitions of regional groupings. The average value for IRR was weighted by the increase in irrigated area. Benin, Chad, and Madagascar have no profitable large-scale irrigation. IRR = internal rate of return. -- Not available. a. Unlike cost estimates presented elsewhere in this report, these are one-time investment costs rather than annualized figures. could be profitably converted to small-scale development of small-scale irrigation, restrict- irrigation. Costs would be only slightly lower ing it to about 30 percent of the cultivated because for small-scale irrigation, on-farm land. Again, restricting attention to those water storage would need to be built in addi- projects that pass a threshold of 12 percent for tion to water distribution infrastructure. internal rate of return, profitable small-scale Only land within three hours' travel time to irrigation could take place on 5.44 million a significant town is deemed suitable for the hectares at a one-time cost of $17.8 billion that 292 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION would nonetheless need to be spread out over under the Comprehensive Africa Agricultural at least a decade (see table 15.2 and figure 15.2). Development Program of the New Partnership In all regions except southern Africa, small-scale for Africa's Development, African countries have irrigation has a higher internal rate of return already committed themselves to raising alloca- than does large-scale irrigation. By far the great- tions of national budgetary resources to agricul- est potential is found in Nigeria, which accounts tural and rural development up to 10 percent for more than 2.5 million (or almost half) of the of the total by 2015. This planned increase in suitable hectares. Countries such as Cameroon, spending could go some way toward meeting Chad, Ethiopia, Mali, Niger, South Africa, Sudan, the costs of an expanded irrigation program. Tanzania, Togo, and Uganda each have at least Most of the hectares found to be viable for 100,000 hectares of potential. irrigation would be dedicated to higher-value Full development of economic irrigation crops (table 15.3). In most cases, irrigation is potential doubles the share of cultivated land found to be viable only when high-revenue- under irrigation, raising the share of cultivated yielding crops are cultivated, be they traditional land under irrigation from 3.5 percent to 7.0 per- cash crops (such as coffee) or higher-value food cent. Annualized investment costs over 10 years crops (such as horticulture). More than half of would constitute 1.8 percent of 2000 GDP and the viable hectares identified are associated account for 88 percent of agricultural spending. with crops that can yield in excess of $2,000 For many countries, that would imply a substan- per hectare annually. Relatively few hectares tial increase in agricultural spending. However, are found where irrigation investments can be justified simply to grow staple food crops. The investment cost estimates used here Figure 15.2 Investment Potential for Dam-Based and Small-Scale Irrigation reflect best-practice experience, but actual costs may often be higher. Studies suggest that well-designed and well-implemented irrigation projects in Africa can lead to costs of no more than $2,000 per hectare for small-scale irrigation schemes and $3,000 per hectare (the distribution component) for large-scale irrigation schemes. Therefore, they are the central parameters used in this modeling exercise. Nevertheless, in practice, irrigation projects in Africa may incur invest- ment costs well in excess of $4,000 per hectare (Inocencio and others 2005). Therefore, consid- ering the sensitivity of these results to possible changes in unit investment costs is important. Estimates for both large- and (particu- larly) small-scale irrigation potential are highly sensitive to assumptions about invest- ment costs (table 15.4). Results are sensi- tive to the assumptions about investment costs per hectare and other parameters (see Profitable area in small scale Profitable area in large scale box 15.1). For large-scale irrigation, the number of viable hectares would decline to 54 percent of the base case if investment costs rose from $3,000 to $6,000 per hectare. This situation might be the case, for example, if irrigation were required to contribute to water storage costs and not simply to water Source: You 2009. distribution infrastructure. For small-scale Note: Dark gray areas indicate positive internal rate of return for dam-based irrigation. Light gray areas indicate profitable areas for small-scale irrigation. Countries left blank were not covered in the sample. irrigation, the story is much more dramatic. Irrigation: Tapping Potential 293 Table 15.3 Share of Crops under Irrigation, Irrigation Investment Needs Assessment Crops Average revenue ($/hectare/year) Percentage total viable hectares Sugarcane, barley, soybeans, other pulses < 100 16 Bananas, beans, potatoes, sorghum, sugar beets, ground nuts, cassava, maize, cotton lint 100­500 14 Coffee, rice, sweet potatoes, millet 500­1,000 16 Horticulture and other high-value crops > 2,000 54 Source: Derived from You 2009. Table 15.4 Sensitivity of Irrigation Potential to Assumed Investment Cost Large-scale irrigation Small-scale irrigation Initial investment cost/ Hectares Initial investment cost/ Hectares hectare (US$) (percentage baseline) hectare (US$) (percentage baseline) 1,000 112 600 226 3,000 100 2,000 100 6,000 54 5,000 5 Source: You 2009. Note: The base case is in boldface and considers all projects that have a positive net present value including those whose internal rate of return may be below the 12 percent threshold. Raising the investment cost from the baseline the food economy looking ahead to 2020 and case of $2,000 per hectare to $5,000 per hect- beyond. are would all but eliminate the economic case A high but feasible 3.6 percent annual for small-scale irrigation. This finding under- increase in investments in irrigation would scores the fact that the economic viability of triple the irrigated harvested area to 22 mil- much of Africa's potentially irrigable land lion hectares in 2050. The irrigated yield would depends critically on containing investment grow by 10 percent in parallel with a gradual costs to best-practice levels. 10 percent increase in rain-fed crop productiv- ity (by increasing effective rainfall by 5 percent by 2020 and 10 percent by 2050). The effects of these investments are discernible in three Effect of Expanding Agricultural areas: food prices, food imports, and nutri- Water Development tional outcomes. Greater availability of food would help con- What would be the development effect of an tain major projected increases in food prices aggressive expansion of Africa's irrigated agri- (table 15.5). Food prices in Africa are projected cultural area? It has already been established to increase from the 2000 baseline level of $117 that an economic case exists for developing an per ton of cereals to $205 by 2050. The greater additional 6.8 million hectares of irrigation food production brought about by irrigation with respectable returns of at least 12 percent. scale-up would help limit this inflationary This conclusion comes with two important pressure, keeping the price of a ton of cereals caveats. First, investment costs would need to down to $177 by 2050. be contained to the $2,000­$3,000 per hectare Irrigation scale-up would also reduce range. Second, to ensure viability, most irri- Africa's reliance on food imports (table 15.5). gation development would need to focus on As of 2000, Africa was importing more than higher-value crops. Assuming such conditions 23,000 tons of cereals annually. Africa's food were met, what would be the broader develop- trade deficit is projected to grow dramatically ment effect? An illustrative modeling exercise to more than 98,000 tons in 2050, reflecting a is used to see how the higher agricultural yields substantial rise in food demand from growing resulting from irrigation scale-up would affect and increasingly urban populations combined 294 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 15.5 Food Price Changes for Various Indicators, 2020 and 2050 Rain-fed Irrigated Average Net cereal Number of Caloric Rain-fed Irrigated production, production, world price imports malnourished availability area, cereals area, cereals cereals cereals for cereals (thousand children (kilocalories/ (thousand (thousand (thousand (thousand Year ($/metric ton) tons) (thousands) person/day) hectares) hectares) metric tons) metric tons) Projection of current investment levels with climate change 2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829 2020 187 4,370 44,041 2,241 87,109 4,847 132,184 12,851 2050 205 75,417 33,756 2,761 92,908 6,294 203,680 26,011 Projection of current investment levels without climate change 2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829 2020 179 6,398 43,646 2,263 86,908 4,858 132,125 12,891 2050 159 98,963 31,894 2,886 92,441 6,441 204,427 26,454 Increased investments with climate change 2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829 2020 182 -7,331 42,507 2,235 85,793 7,666 138,904 18,625 2050 177 11,134 31,640 2,852 89,560 21,722 220,820 86,003 Source: Estimates provided by International Food Policy Research Institute, Washington, D.C., 2009. with relatively slow expansion of output. By Implementation Challenges expanding homegrown food production, irri- gation investments could reduce food imports Considerably increasing investments in irriga- to only 11,000 tons by 2050. tion raises issues about the performance and Some discernible (though far from dra- sustainability of these investments. A recent matic) effects on malnutrition would occur multidonor irrigation performance diagnostic (table 15.5). Because of the adverse market identified challenges that need to be addressed trends noted, child malnutrition is projected to improve the performance of irrigation to increase slightly in the coming decades, investments. This section summarizes the from 32.7 million cases in 2000 to 33.8 mil- findings and recommendations of that report lion in 2050. The greater availability of food (World Bank 2007). associated with irrigation investments would help relieve this problem, keeping the number Adopt a Strategic Vision of cases down to 31.6 million in 2050, albeit a National agricultural water development strate- rather modest reduction. gies need to be promoted. They should recognize Overall, irrigation investments help offset (a) the potential contribution of agricultural some of the adverse effects anticipated from water to poverty reduction and growth; (b) the climate change (table 15.5). Climate change imperatives of farm-level profitability and eco- is an important factor driving the projected nomic viability; and (c) the need for policies, deterioration of the food supply situation in legal frameworks, and organizations that foster Africa. Irrigation investments can be viewed as profitable, sustainable water-managed farming an adaptation measure insofar as they help off- by smallholders. The strategies would analyze set the negative supply-side effects of climate trade-offs and capture synergies of the various change. Thus, irrigation offsets the effect of investment options. Key areas to be covered by climate change on child malnutrition, and it such strategies include (a) increasing the pro- more than offsets the effect on the food trade ductivity and profitability of existing irrigation balance. Food prices would remain somewhat schemes; (b) expanding or developing new higher than they would have been without cli- irrigation (including systems based on water mate change, but still considerably lower than harvesting); (c) testing and disseminating rain- without irrigation. water harvesting technologies; (d) developing Irrigation: Tapping Potential 295 sustainable supply chains for agricultural water BOX 15.2 equipment; and (e) investing in research on agricultural water management. Agricultural water strategies need to be An Enabling Environment for Reform: Office incorporated into wider sectoral strategies for du Niger agriculture, rural development, and water. Water Initiated by the French in 1932, the Office du Niger is one of the strategies should be based on integrated water oldest and largest irrigation schemes in Sub-Saharan Africa. Located resource management principles that promote in Mali, the scheme was originally developed to supply the French an economically efficient allocation of water to textile industry with cotton and to increase food security for the the agricultural sector, ensure that water allo- Sahel region. Despite a disappointing performance in the first sev- cation and management take into account the eral decades, including limited area development, poor infrastructure needs of the poor, and provide for effective par- maintenance, and low yields, the project was rehabilitated in the early ticipation by smallholders in basin planning. 1980s with assistance from the European Union; the World Bank; Agricultural water needs to be more clearly and the governments of France, Germany, the Netherlands, and reflected in Poverty Reduction Strategy Papers the United States. Comprehensive reforms and rehabilitation tripled or similar national development strategies. average paddy yields to 5 tons per hectare, increased the area under cultivation to about 80,000 hectares, boosted settler population by over 220 percent, and increased paddy production per capita from Invest in Institutional Reforms 0.9 ton to 1.6 tons, reducing poverty and increasing food security. The new agricultural water strategies should The project's success is attributed to technical, institutional, and form the basis for sectoral programs that economic factors. Technical factors include water management combine investment in infrastructure with through physical rehabilitation of irrigation and drainage networks, investment in institutional reforms. A start- a comprehensive package of improved technologies, and appropri- ing point for reforms is to improve coordi- ate agricultural mechanization. Institutional and economic factors nation among the government organizations include liberalized paddy marketing and processing, land-tenure secu- responsible for infrastructure development rity, infrastructure improvements, institutional reforms, and stronger (a ministry of water) and those responsible for partnerships with farmers. Also important were donor coordination, irrigated farming (a ministry of agriculture) government commitment, and the right macroeconomic and policy and to build capacity and incentives for pub- environment. lic agencies to adopt a new agricultural water Source: Based on interviews with World Bank staff from the Africa Water development paradigm (box 15.2). It is also Resources Department, 2008. desirable to develop the instruments needed for private sector involvement through public- private partnerships. Responsibility for development should a minimum requirement, monitoring and be decentralized as much as possible, based evaluation systems should measure inputs; on the principle of subsidiarity. In almost costs; and changes in production, incomes, all cases, reforms will focus on empower- employment, health, and the environment. ing potential users of agricultural water to cope with their new roles and responsibilities Undertake Viable and Sustainable and to deal effectively with service providers, Projects including irrigation agencies (which should Future designs and investment decisions should become accountable to their clients), credit be based solely on economic viability, farm- organizations, and input supply and output level profitability, and sustainability. Nonviable markets. This measure should be accompanied investments for "social" or "strategic" purposes by investment in capacity building for farmer should be avoided. Subsidies (if any) should organizations. More generally, the role of be limited to (a) items having a medium- to farmers in cost sharing and in operation and long-term economic life (headworks and main maintenance should be transparent. canals on larger schemes), the cost of which is Monitoring and evaluation should be an beyond the financial capacity of most farmers, essential management tool for farmers, imple- rather than for lower-cost investments with a menting agencies, and financing partners. As short economic life (treadle pumps or on-farm 296 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION development for improved infield rainwater BOX 15.3 management); and (b) technology development and promotion. Subsidies for support services and operation and maintenance should pre- Nigeria's Fadama Water ferably be avoided or otherwise carefully tar- User Association: Expanding geted and provided only in the short term to Irrigation kick-start commercial production. Reducing per hectare development costs is Community-driven development can critical for success. The cost of public irrigation expand irrigation in Sub-Saharan Africa, particularly among homogeneous groups development in Sub-Saharan Africa has been with high social equality. The Fadama II excessively high. Many irrigation schemes failed agricultural development project in Nige- to capture significantly higher yield levels and ria supported water user associations. cropping intensities and failed to transition to It increased access to productive assets production of higher-value crops. Under these and infrastructure, including agricultural circumstances, high development costs rapidly inputs, irrigation infrastructure, and erode returns on investment. Costs have come postharvesting equipment. Community down in recent years because of competition groups were organized into user groups among contractors; the emergence of new con- based on their agricultural sector (live- tractors, particularly from low-income countries; stock, crops, forestry, and the like) and and the introduction of affordable irrigation were paid 10 percent of the asset costs. As a result, the value of productive assets technologies. Keeping down development costs for water and irrigation, which included is important, because projects with low returns water pumps, boreholes, and tube wells, have per hectare development costs four times increased nearly 3,000 percent. In addi- those of projects with good returns. Design- tion, the irrigation investment raised ing for maximum profitability is thus critical crop productivity in dry regions, increas- and involves both cost-effective design and an ing incomes in the dry savannahs nearly effective strategy to increase output. A new gen- 80 percent. eration of well-designed and well-implemented Source: Based on interviews with World Bank irrigation projects has proved only marginally staff from the Africa Water Resources Depart- more costly than those of other regions. ment, 2008. Provide Agricultural Water as Part of a Comprehensive Package Investments in agricultural water are part of livelihoods. This understanding will make the a comprehensive package to increase outputs, investment more pro-poor by selecting tech- including empowered farmer organizations; nology options that are low risk and afford- sustainable, efficient, and accountable agricul- able to the poor, and by seeking to maximize tural support services; and accessible, profitable farm-level profitability and agricultural wage markets. Therefore, investments in agricultural employment, as well as other indirect employ- water not only should focus on infrastructure ment opportunities. In addition, institutional delivery but also should address agricultural design should ensure that the role of women intensification in a holistic way (box 15.3). in production systems and their management Indeed, without complementary efforts to is taken into account and built on. improve agricultural productivity through other channels, irrigation is unlikely to deliver Notes the sizable increases in yield necessary to justify The authors of this chapter are Mark Rosegrant, the original investment. Claudia Ringler, and IJsbrand de Jong, who drew Investments need to be pro-poor. Project on background material and contributions from preparation studies should provide an under- Salah Darghouth, Mandy Ewing, Stephen Mink, standing of how investments in agricultural Siwa Msangi, Siobhan Murray, Mark Svendsen, water can assist beneficiaries to improve their and Liang Zhi You. Irrigation: Tapping Potential 297 1. Agricultural water and irrigation are used NEPAD (New Partnership for Africa's Develop- interchangeably, and both refer to the supply of ment). 2003. Comprehensive Africa Agricul- additional water to augment rainwater (if any) tural Development Programme. http://www for crops and livestock. In this report, agricul- .nepad-caadp.net/. tural water and irrigation include drainage, Svendsen, Mark, Mandy Ewing, and Siwa Msangi. where appropriate. 2008. "Watermarks: Indicators of Irrigation Sector Performance in Sub-Saharan Africa." Background Paper 4, Africa Infrastructure Country Diagnos- References tic, World Bank, Washington, DC. Frenken, Karen, ed. 2005. Irrigation in Africa in Figures: AQUASTAT Survey--2005. FAO Water World Bank. 2007. Investment in Agricultural Water Report 29. Rome: Food and Agriculture for Poverty Reduction and Economic Growth in Organization of the United Nations. Sub-Saharan Africa: Synthesis Report. Washing- ton, DC: World Bank. Inocencio, Arlene, Masao Kikuchi, Manabu Tonosaki, Atsushi Maruyama, and Hilmy Sally. ------. 2008. World Development Report 2008: 2005. "Costs of Irrigation Projects: A Compari- Agriculture for Development. Washington, DC: son of Sub-Saharan Africa and Other Developing World Bank. Regions and Finding Options to Reduce Costs." You, Liang Zhi. 2008. "Irrigation Investment Needs Report of component study for Collaborative in Sub-Saharan Africa." Background Paper 9, Programme, International Water Management Africa Infrastructure Country Diagnostic, World Institute, Pretoria, South Africa. Bank, Washington, DC. Chapter Chapter 16 Water Supply: Hitting the Target? S ub-Saharan Africa as a whole is unlikely to forgoes at least $1.8 billion a year in revenues. meet the Millennium Development Goal Typically, capital costs have been subsidized, (MDG) for water supply. Coverage in but the subsidies are highly regressive. Full urban areas has been declining as utilities have capital cost recovery should be affordable for struggled to keep pace with population growth. half of the population, including the bulk of In rural areas, more than 40 percent of the those that enjoy piped water access today, but population continues to rely on surface water. would not be affordable to the remainder. Overall, wells and boreholes are the fastest- Furthermore, the operational inefficiencies growing sources of supply. of water utilities cost the region $0.9 billion The price tag for reaching the MDG for a year and impede service expansion. Insti- access to an improved water source is estimated tutional reforms of legal and regulatory at $16.5 billion a year (roughly 2.6 percent of frameworks hold the key to improving per- Africa's GDP). For many countries, these costs formance. Private participation, particularly look prohibitive. By emphasizing lower-cost lease contracts, has significantly affected technologies, such as standposts and bore- utility performance, but state-owned utilities holes, those countries could reduce the cost of will remain the central actors, and greater meeting the MDG. However, standpost use is efforts are needed to improve their gover- affected by institutional challenges that remain nance frameworks. to be addressed. Even if all these inefficiencies could be Spending on the water sector today is eliminated, the overall financing gap for the $3.6 billion, one-fourth of what is required. water sector would still be $7.8 billion a year However, some $2.7 billion available to the sector (1.2 percent of GDP). is currently being wasted due to inefficiency. Looking ahead, the institutional reform An important example of inefficiency is agenda remains as relevant as before, even underpricing of services. Average water tar- if the focus has shifted toward a more plu- iffs are about $0.67 per cubic meter, below the ralistic view of public and private sector cost-recovery threshold of just over $1.00 per roles. The reform agenda also needs to move cubic meter. By underpricing water, the sector beyond utilities to encompass line ministries 299 300 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION and the whole public expenditure framework Guinea, Mali, Mauritania, Senegal, Sudan, that underpins, and too often hinders, sec- Uganda, and Zimbabwe. At the other end of tor investment programs. Room also exists the spectrum, some of Africa's most populous for more cost recovery, so that scarce subsidy countries, such as the Democratic Republic of resources are redirected to promote access Congo and Nigeria, are a long way from meet- among the poorest. For the majority that ing the target. The proximity of 2015 and the does not enjoy access to a piped-water con- daunting challenges highlight the importance nection, greater thought needs to be given to of understanding the performance of the water how standposts can become a more effective sector in the region, its achievements and part of urban water supplies. The burgeoning shortcomings, and the factors most critical in use of wells and boreholes for supply in urban expanding coverage. areas demands urgent attention from policy Service options for water supply can be makers both to improve their understanding organized on a hierarchical ladder accord- and to develop suitable regulatory tools. In ing to the delivery method and quality of the rural areas, the big challenge, in addition to associated service. At the top of the ladder is continuing to expand access, is the high break- piped water, which is both potable and conve- down rate from lack of maintenance, which nient. Standposts offer the same potability, but threatens the sustainability of what has already through a less convenient channel and with been achieved. some risk of the water becoming contaminated during collection. Next come wells and bore- holes, which, depending on their location, can The Millennium Development be more or less convenient than standposts. Goal for Water--Elusive for Many The water delivered can be of good qual- ity, though that depends on the local aquifer Whereas the rest of the world is on track to and protection from contamination. Surface achieve the MDG for water supply,1 Sub- water is at the bottom of the ladder, because Saharan Africa reports that only 58 percent of its quality is in most cases questionable, and its population enjoys access to safe drinking it is seldom convenient. Although the objec- water vis-à-vis a target rate of 75 percent to tive may be to provide universal piped-water be reached by 2015 (WHO/UNICEF 2006).2 access to the population, it may not be feasible Progress has been modest, with access increas- or affordable in the short run. An important ing by only 9 percentage points between 1990 first step is to move people away from surface and 2006, or less than 1 percentage point a water to one of the lower rungs of the water year. To meet the target, growth should stand supply ladder. at over 2 percentage points a year. As a result, Sub-Saharan Africa lags all other regions, including South Asia, whose performance was Differing Patterns of Urban and broadly comparable to that of Sub-Saharan Rural Access Africa in the past but which has moved at a much faster pace in recent years. In rural areas, reliance on surface water remains Some countries are closer to meeting the prevalent. The share of the population relying MDG targets than others. According to the on surface water fell quite steeply in the 1990s, most recent WHO/UNICEF Joint Monitoring from 50 percent to just over 40 percent, where Programme data for 2006, five African coun- it has remained through 2005 (table 16.1). tries have already met the MDG target: Burkina Boreholes are the main source of improved Faso, Ghana, Malawi, Namibia, and South water, accounting for a further 40 percent of Africa. Moreover, an additional 12 countries the population. Access to piped water and had reasonable prospects of meeting the tar- standposts is very low, barely increasing over get by 2015 if they continued to make steady the last 15 years. Indeed, in many countries, progress: Cameroon, the Central African less than 1 percent of the rural population Republic, the Comoros, Côte d'Ivoire, Eritrea, receives piped water. Strikingly, in countries Water Supply: Hitting the Target? 301 Table 16.1 Evolution of Water Supply Coverage in Africa, by Source percentage of population Well and Piped supply Standposts boreholesa Surface water Period Urban Rural Urban Rural Urban Rural Urban Rural 1990­95 50 4 29 9 20 41 6 50 1995­2000 43 4 25 9 21 41 5 41 2001­05 39 4 24 11 24 43 7 42 Source: Banerjee, Wodon, and others 2008. Note: The figures are based on household surveys. a. The WHO/UNICEF Joint Monitoring Programme (2006) considers protected wells and boreholes to be improved water sources. However, disaggregating the data into protected and unprotected wells and boreholes from the household surveys used in the Africa Infrastructure Country Diagnostic study is not possible. with higher levels of urbanization, access to Table 16.2 Services Provided by Utilities in Their Service Areas piped water and standposts in rural areas is percentage substantially higher. Population already enjoying Population gaining access access to utility water to utility water each year In urban areas, rapid population growth has caused piped-water coverage to fall mark- Resale of Country Piped neighbor's Piped edly over the last 15 years. However, at close to type water Standposts water Total water Standposts Total 40 percent, it is still the single-largest source of Low water supply. Standpost coverage has similarly income 30 19 15 68 1.9 1.0 2.9 declined, whereas that of boreholes has risen, Middle so that each of these types of service reaches income 89 10 0 99 4.5 -0.2 4.5 about 24 percent of the urban population. Averagea 44 14 6 86 2.5 0.9 3.2 The lower coverage of standposts compared Source: Banerjee, Skilling, and others 2008. Note: The figures are based on utility data for utility service areas. The coverage is higher than that with piped water is particularly striking, given obtained from household surveys because utility service areas do not cover the entirety of urban areas their relatively low cost and the pressure to and because for some countries, data were available only for utilities in larger cities. a Average is population weighted. expand services rapidly. Reliance on surface water, at 7 percent of the urban population, has hardly changed. Utilities are the central actors respon- of the unconnected households. Household sible for water supply in urban areas. Over- resale, while prevalent, is often illegal, although all, about two-thirds of the urban populace Côte d'Ivoire illustrates the potential benefits depends on utility water. In the middle- of legalization (box 16.1). income countries, utilities are essentially the Utilities report providing around 20 hours only players, reaching about 99 percent of the of service daily, and just over 80 percent of their urban population, the vast majority through samples pass chlorine tests (table 16.3). They private piped-water connections. In low- typically produce just over 200 liters per capita income countries, only 49 percent of urban served, though the amount for middle-income residents benefit from utility water, fewer countries is about twice that for low-income than half through private piped connections countries. If the utilities' total water produc- (table 16.2). For the rest, informal sharing of tion could be evenly distributed to the entire connections through resale between neigh- population residing in the utility service area, it bors (15 percent of the urban population) is would amount to 74 liters per capita a day, just almost as prevalent as formal sharing through about adequate to meet basic human needs. standposts (19 percent of the urban popula- Urban households that do not benefit tion). In Maputo, Mozambique, one-third from utility water rely on several alternatives. of unconnected households purchase water The rapid expansion of boreholes in urban from their neighbors, and in Maseru, Lesotho, areas has already been noted. Water vendors, household resellers provide water to 31 per- another alternative, may retail water from cent of the population, including almost half utilities, boreholes, or surface sources, either 302 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 16.1 Legalizing Household Water Resellers in Côte d'Ivoire Côte d'Ivoire legalized household resellers in informal settle- Nevertheless, the scheme faces implementation chal- ments to help the poor receive safe water. This legislation lenges. Household resellers pay SODECI twice: their reseller enables the Société de Distribution d'Eau de la Côte d'Ivoire payments and a price markup for network extensions in (SODECI) to indirectly influence the price and quality of the these areas. Furthermore, no special tariff applies to house- water in these areas. The utility has issued about 1,000 hold resellers; they pay the high consumer tariff. Therefore, licenses to water resellers, who can invest in last-mile net- the motivation to become a household reseller is limited, and work extensions to cater to demand in poor neighborhoods. most households pay the regular domestic consumer price. SODECI reduces the risk of nonpayment by requiring a sizable deposit (about $300) and invoicing resellers monthly. Source: Collignon and Vézina 2000. Table 16.3 Quality of Services Provided by Utilities in Their Service Areas coverage much more rapidly than all the Availability of utility water Quality of water supply utility-based alternatives together. Within the Liters per Liters per Percentage purview of the utility, access to standposts capita available capita available Hours/day of of samples seems to be growing faster than access to for residents in for utility continuous passing Country type service area customers service chlorine test piped water. But the combined growth rates of the various improved forms of water sup- Low income 74 149 19 83 ply in urban areas (less than 1 percent a year) Middle income 272 277 24 99 still fall short of population growth (more a Average 167 224 21 83 than 4 percent a year). Source: Banerjee, Skilling, and others 2008. Access to improved water sources is highly a Average is population weighted. inequitable across the income distribution (figure 16.2). Piped water and standposts are from trucks and carts, or sometimes through heavily concentrated among the more affluent their own private distribution networks. segments of the population, typically in urban Water vendors account for only 3 percent of areas. The poorest 40 percent of the population, the African urban market, rising to 7 percent by contrast, depends on surface water and on for West Africa. In some countries, how- wells and boreholes in almost equal measure. ever, their contribution to the urban water Piped supply covers only 10 percent of African supply is much larger: Nigeria (10 percent), households in the bottom 60 percent of the Chad (16 percent), Niger (21 percent), and population. For the middle-income countries, Mauritania (32 percent). In 15 large cities in access to piped water and standposts among Africa, the cost of vendor water, particularly the poorest quintiles is substantially higher when transported directly to the household, than in the low-income countries. can be 2 to 11 times more expensive than hav- ing a household connection (table 16.4). This strong willingness to pay for vendor water is a Financing the MDG potential revenue source that the utilities are typically unable to capture. The overall price tag for reaching the MDG The dynamics of service expansion reveal a target for access to water is estimated at $16.5 similar overall pattern in both urban and rural billion (roughly 2.6 percent of Africa's GDP), areas: the absolute number of people depend- which is somewhat higher than previously ing on surface water continues to grow, a grim thought (Mehta, Fugelsnes, and Virjee 2005). statistic in its own right (figure 16.1). Across Capital investment needs based on minimum the board, wells and boreholes are expanding acceptable asset standards and accounting for Water Supply: Hitting the Target? 303 Table 16.4 Average Price for Water Service in 15 Largest Cities, by Type of Provider House Small piped Stand Household Water Water connection network post reseller tanker vendor Average price ($/cubic meter) 0.49 1.04 1.93 1.63 4.67 4.00 Markup over house connection (%) 100 214 336 402 1,103 811 Source: Keener, Luengo, and Banerjee 2008. Figure 16.1 Increases in Access to Water by Source, 1990s to Early 2000s a. Urban b. Rural 2.0 2.0 access (% of population) annualized increase in access (% of population) annualized increase in 1.5 1.5 1.0 1.0 0.5 0.5 0 0 piped supply standposts wells and surface water boreholes ­0.5 piped supply standposts wells and surface water boreholes Source: Banerjee, Skilling, and others 2008. Figure 16.2 Coverage of Water Services, by Table 16.5 Estimated Annual Financing Needed to Meet the Water MDG Budget Quintile $ billions annually % of GDP 60 Country type O&M Capital Total O&M Capital Total 50 Sub-Saharan % of households 40 Africa 5.53 11.01 16.54 0.86 1.72 2.58 30 Low-income fragile 0.98 2.41 3.39 2.55 6.27 8.81 20 Low-income 10 nonfragile 1.91 4.36 6.27 1.73 3.95 5.68 0 poorest second third fourth richest Middle income 1.19 1.19 2.38 0.44 0.44 0.88 quintile Resource rich 1.47 3.12 4.59 0.66 1.40 2.06 piped water public standpost Source: Briceño-Garmendia, Smits, and Foster 2008. surface water well water Note: O&M = operation and maintenance. Row totals may not add exactly because of rounding errors. Source: Banerjee, Skilling, and others 2008. Note: In terms of household spending, poorest quintile = poorest 20 percent of the population; second quintile = second-poorest Although donors have been financing the 20 percent of the population; third quintile = middle 20 percent of associated capital costs, low-income coun- the population; fourth quintile = second-richest 20 percent of the population; richest quintile = richest 20 percent of the population. tries face a maintenance bill of approximately 2 percent of GDP, perhaps even more of a both new infrastructure and rehabilitation challenge given weaknesses in sector institu- of existing assets can be conservatively esti- tions and budgeting processes. The challenge mated at $11.0 billion a year (1.7 percent of to resource-rich countries, though not insig- the region's GDP). The maintenance require- nificant, appears much more manageable. ments are $5.5 billion a year (0.9 percent of the Middle-income countries should comfortably region's GDP) (table 16.5). meet their financing needs. The cost of achieving the water MDG is Existing spending on water and sanitation is very high for nonfragile low-income coun- $7.6 billion, or less than half of what is required. tries (5.7 percent of GDP), and particularly (Because of the difficulty of accurately separat- high for fragile states (8.8 percent of GDP). ing water and sanitation, spending for both 304 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION services is presented together here. The more addressed. Improving cost recovery of water detailed discussion of sanitation can be found utilities could reduce the gap by $1.8 billion a in chapter 17 of this volume.) African countries year, addressing operating inefficiencies would are already devoting significant resources to bring an additional $0.9 billion a year, and meeting the water and sanitation MDG targets. raising capital budget execution could recoup Domestically funded spending accounted for some $0.2 billion a year. A larger part of these over half of total spending. For capital invest- gains is for the low-income countries, whose ment, donors have had a predominant role, par- utilities should focus equally on improving ticularly in low-income countries, where they cost recovery and on reducing operating inef- finance most of the investment (table 16.6). ficiencies. For middle-income countries, much Financiers outside the Organisation for Eco- of the gain would come from reducing operat- nomic Co-operation and Development have ing inefficiencies. also played a role in the low-income coun- Even if all of these efficiencies could be tries, whereas private finance has been neg- addressed, the water sector would still face ligible. However, household self-finance of a sizable financing gap of $9.3 billion a year on-site sanitation facilities--such as latrines-- (1.5 percent of GDP; table 16.7). The gap for is estimated to be substantial. capital requirements is more than twice that Public investment in water and sanitation for operation and maintenance, suggesting that is almost exclusively the domain of the central the MDG challenge is mainly about expanding government. Within the public sector, general access to improved water sources and rehabili- governments3 carry out most capital spend- tating existing assets in poor condition. ing, whereas public enterprises tend to execute Although the costs of investment appear most spending for operation and mainte- high, the health dividends of meeting the MDG nance. Therefore, the institutional effective- target are substantial. Every dollar invested in ness of the line ministries is just as important water supply generates economic benefits of at as the institutional effectiveness of the utilities least $1.50 (Hutton and Haller 2004). Access in ensuring that resources are well used. Thus, to improved water brings a variety of benefits, better public expenditure management, good particularly improved health and reduced time selection of projects, and clear strategic guid- spent collecting water. Serious illnesses trans- ance for investment should all be considered mitted through unsafe water, such as infectious integral aspects of the sector reform agenda. diarrhea, are a leading cause of infant mortal- The existing resource envelope for Africa's ity. Waterborne illnesses can be a substantial water supply sector would go considerably economic burden, causing adult deaths and further if various serious inefficiencies-- workdays lost and affecting children's health amounting to $2.7 billion a year--could be and education. Table 16.6 Existing Financial Flows to Water Supply and Sanitation $ billions annually O&M Capital expenditure Public Public Non-OECD Household Country type sector sector ODA financiers PPI self-finance Total Total Sub-Saharan Africa 3.06 1.06 1.23 0.16 0.01 2.13 4.58 7.64 Low-income fragile 0.13 0.03 0.11 0.02 0.00 0.16 0.32 0.45 Low-income nonfragile 0.30 0.25 0.78 0.05 0.00 0.45 1.54 1.83 Middle income 2.17 0.15 0.10 0.01 0.00 0.21 0.47 2.64 Resource rich 0.15 0.72 0.24 0.08 0.01 0.52 1.57 1.72 Source: Briceño-Garmendia, Smits, and Foster 2008. Note: ODA = official development assistance; OECD = Organisation for Economic Co-operation and Development; O&M = operation and maintenance; PPI = private participation in infrastructure. Water Supply: Hitting the Target? 305 Table 16.7 Composition of Water Sector Funding Gap $ billions annually Percentage of GDP Country type O&M Capital Total O&M Capital Total Sub-Saharan Africa 2.74 6.60 9.34 0.43 1.03 1.45 Low-income fragile 0.75 2.00 2.76 1.96 5.22 7.17 Low-income nonfragile 1.43 2.92 4.35 1.30 2.64 3.94 Middle income 0.00 0.00 0.00 0.00 0.00 0.00 Resource rich 1.06 1.74 2.80 0.48 0.78 1.26 Source: Briceño-Garmendia, Smits, and Foster 2008. Note: O&M = operation and maintenance. Household members, primarily women and the majority of those gaining access to children, face a substantial opportunity cost in improved water in utility service areas each time to fetch water. More than 20 percent of year. A survey of 51 water utilities shows that the population in Cameroon, Ghana, Maurita- about 2 percent of the population in utility nia, Niger, and Tanzania must travel more than service areas gains access to a formal utility 2 kilometers to the primary water supply. Rural water service each year--1.5 percent to piped dwellers tend to travel farther than urban dwell- water and 0.5 percent to standposts. Given the ers. Therefore, the time savings from accessing higher unit cost of piped water, utilities are not water from a nearby source are enormous even maximizing the effect on coverage from their when valued at a discounted wage rate. limited investment budgets. Indeed, utilities could double the value for money of their investment programs (dollars per capita gain- Using Appropriate Technologies ing access to improved water) if they weighted their investment programs toward standposts For many countries, the costs of meeting the rather than private connections. water MDG look prohibitive. Although the cost Standposts provide safe water in urban of meeting the MDG amounts to 2.3 percent areas at about one-third the per capita cost of of GDP on average, for a handful of countries a private tap. Therefore, it is striking that the (including Benin, the Democratic Republic coverage of public standposts in urban Africa of Congo, Kenya, and Madagascar), the cost lags so far behind that of private taps and that would be in excess of 7 percent of GDP, well their distributional incidence is skewed toward beyond what could be feasibly attained. So the better-off households. these countries must choose between achieving Several institutional challenges prevent the MDG at a lower rung of the water supply standposts from being more widely adopted ladder with lower-cost technologies and sub- (Keener, Luengo, and Banerjee 2008). Stand- stantially postponing their achievement of the posts describe a range of public water supply goal. Concentrating all coverage expansion in arrangements, from an unattended public lower-cost technologies, such as standposts and access tap to a kiosk with a human operator. boreholes, would lower the cost of meeting the Despite their low investment cost, standposts MDG to 1.6 percent of GDP on average. For are in practice often beset by poor main- Benin, Kenya, and Madagascar, the cost could tenance and high retail prices. The service be reduced to less than 4 percent of GDP, which provided by the standpost may be very low would be more affordable. The Democratic because the utility's claims about the number Republic of Congo would be the only country served by a single standpost are highly vari- still experiencing an overall cost above 7 per- able and often unrealistic, ranging from just a cent of GDP. few customers to 5,000, with an average of This strategy, however, runs contrary to cur- about 700. Official utility data indicate that rent practice, which sees piped water serving about three-quarters of standposts are in good 306 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION working order, but more detailed surveys sug- 20 times in some extreme cases (for example, gest that fewer than half may be functioning in Kinshasa in the Democratic Republic of normally at any one time. Congo). Community engagement can increase The utility manages about half the stand- the accountability of kiosk operators, but it posts, but a growing number are being can also lead to corruption and mismanage- delegated to private or community manage- ment in countries that lack social cohesion or ment. Although this arrangement helps ensure oversight by supporting institutions. sustainability, price markups to the final Even so, a handful of countries are making consumer can be large, and local elites can headway with the expansion of standposts. capture the service. Many utilities originally Each year in Rwanda, an additional 3 percent offered standpost services free of charge and of the population resident in the service area of later moved toward charging preferential rates the national utility Electrogaz gains access to for standpost water, typically half the price standposts (box 16.2). Similarly, Lusaka Water charged for a private connection. When del- and Sewerage Company in Zambia adds 3 per- egated management is introduced, however, cent of its resident population to this category markups are added to cover the salaries of the every year. Another intermediate approach standpost operators, often with significant sometimes adopted is to install yard taps, profit markups. As a result, standpost prices shared by four or five households. They are rise to 3 times the utility rate, and as much as lower in cost than private taps, and they avoid BOX 16.2 Standposts in Kigali, Rwanda post costs $0.55 per cubic meter to supply when distribution losses are fully accounted The water production capacity of Electrogaz, for but brings in only $0.42 in revenue. the main utility in Kigali, is inadequate to meet Kigali has roughly 240 standposts, of network demand. The lack of bulk supply which an estimated 193 (80 percent) were causes rolling shortages throughout the city, operating in December 2008. Utility offi- often forcing residents with private connec- cials estimate that about 60,000 people use tions to seek water from public sources, such standposts, though this figure includes con- as standposts. sumers who use them only when their pri- The financial sustainability of standposts in mary source is unavailable. According to the Kigali can be estimated from the tariff paid by total water volume recorded by public stand- standpost operators ($0.42 per cubic meter), post meters, the standposts could supply only the total cost of Electrogaz's production ($0.36 48,500 people with 20 liters per capita a day. per cubic meter), the rate of unaccounted-for This figure would be more indicative of the water in distribution and selling (35 percent and upper bound of the population that primarily 5 percent, respectively), and the volume and depends on standposts (about 6 percent of price of water sold at the standposts. Three the city's population). hypothetical operators selling 100 jerricans The utility's limited production capacity has each per day at $0.02, $0.03, and $0.05 per affected both the level of peak demand at jerrican would earn estimated monthly net standposts and the cost of production. Obser- incomes of about $314, $949, and $1,584, vations and interviews with consumers indicate respectively (the 2007 GDP per capita was that prices have often been higher in areas $341), making this an attractive business prop- when and where water service has been cut osition for them. However, from the utility's and have been lower after periods of precipita- perspective, the combination of a low tariff tion that increase the availability of other supply and the 35 percent rate of unaccounted-for options, such as rainwater and natural springs. water in distribution creates a loss-producing scenario: each cubic meter supplied to a stand- Source: Keener and others forthcoming. Water Supply: Hitting the Target? 307 some of the issues associated with standposts, a set of complex financial arrangements with which serve hundreds of customers. the central government that prevents optimal resource allocation, financial sustainability, and economically efficient use of water resources The Challenge of Cost Recovery (Briceño-Garmendia, Smits, and Foster 2008). As a result, the utility management postpones By underpricing water, the sector forgoes at basic investment and rehabilitation decisions least $1.8 billion a year in revenues (0.3 percent to "make it through the day" financially. Thus, of GDP). Underpricing of water--a common although these policies appear to be socially problem throughout the world--is also wide- benign, by debilitating the financial position spread across Africa. In the worst cases, such of the utility, they ultimately lead to delayed as Côte d'Ivoire, Madagascar, and Senegal, investment and hence hold back service expan- underpricing can result in utilities' capturing sion to reach the unserved population. less than 40 percent of the revenues they need Average water tariffs in Sub-Saharan Africa and can lead to an economic burden of around are about $0.67 per cubic meter, two-thirds of 0.7­0.9 percent of GDP (figure 16.3). The asso- a cost-recovery threshold of just over $1.00 per ciated utility deficit is often concealed within cubic meter. Tariffs are already relatively high by international (even developing-country) standards (box 16.3). In 2004, average water Figure 16.3 Economic Burden of Water Underpricing, tariffs were about $1.00 per cubic meter in by Country members of the Organisation for Economic Co-operation and Development, around Senegal $0.35 in middle-income countries, and as low Côte d'Ivoire as $0.10 in South Asia (GWI 2005). Tariffs of Madagascar about $0.40 per cubic meter are considered Malawi more than adequate to cover operating costs Ghana in most developing-country contexts. In low- Lesotho income African countries, however, operating Niger costs are as high as $0.60 per cubic meter on South Africa average, reflecting inadequate selection of tech- Congo, Dem. Rep. of nologies, low population density, country risk Zambia premiums, and the high cost of inputs. Current tariffs cover operating costs on average but do Mozambique not contribute much toward capital costs. Namibia In most cases, the state or donors have Ethiopia subsidized capital costs almost entirely. Subsi- Kenya dies to residential consumers are highly regres- Benin sive (Banerjee, Wodon, and others 2008). Across Rwanda the bottom half of the income distribution, Sudan barely 10 percent of households have access to Nigeria piped water. Indeed, more than 80 percent of Burkina Faso households with piped water come from the Tanzania top two quintiles of the income distribution. Uganda Because poorer households are almost entirely Cape Verde excluded, they cannot benefit from subsidies 0 0.2 0.4 0.6 0.8 1.0 embedded in prices for piped water. In many % of GDP cases, targeting performance is further exacer- bated by poor tariff design, with widespread Source: Briceño-Garmendia, Smits, and Foster 2008. Note: The economic burden of underpricing is defined as the use of minimum charges and rising block tar- difference between the average effective tariff and the full cost- iffs that provide large lifeline blocks of highly recovery tariff multiplied by the total volume billed. It is the total revenue shortfall attributable to underpricing. subsidized water to all consumers. 308 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 16.3 Cost Recovery, Equity, and Efficiency in Water Tariffs in Africa During a 2006­07 survey of 23 African countries, tariff important in Africa than elsewhere: most consumers already structures for 45 water utilities were collected and ana- survive on little more than subsistence quantities of water. lyzed. These tariff structures were assessed using criteria of cost recovery, efficiency, and equity. Many countries in Equity Africa have adopted a two-part tariff structure that incor- More generally, African water tariff structures do not perform porates both a fixed charge and a water-use charge. The all that well in equity terms. A number of factors contribute to block structure can add to the complexity of tariffs; the this inequity. First, the subsidy to the low block under the cur- number of blocks ranges between two and seven with an rent increasing block tariff structure does not all go to small average of three. Only four utilities practice a one-block or consumers (usually the poor); instead, a substantial amount linear tariff. of the subsidy leaks to large consumers. Second, because of the fixed and minimum consumption charges and the large Cost Recovery size of the low blocks, small consumers often end up pay- ing higher effective prices per unit than large consumers. The experience of recovering operating costs in Africa In almost three-fourths of the cases, consumers with water is positive, with many utilities setting tariffs at levels high intake at the survival level pay as much as or more than aver- enough to recoup operation and maintenance costs. The age consumers pay. Despite the prevalence of rising block performance of African utilities in this respect is superior to tariffs, the smallest consumers do not always pay the lowest that found elsewhere in the world. In Africa, the tariff struc- price in Africa. Third, the connection cost is high for many tures are designed in a way that is more conducive to meet- utilities compared with gross national income per capita, indi- ing operation and maintenance costs at the high or low cating significant affordability problems for expanding net- ends of consumption. Capital cost recovery is largely elusive works into unserved areas. Fourth, the retail standpost price for residential consumers; only four utilities in Cape Verde, is significantly higher than the utility-imposed price because Namibia, and South Africa charge more than $1. of rent-seeking behavior on the part of operators. African utilities operate in an environment of high-cost Efficiency service provision and high-cost recovery. What emerges is a tariff structure that is relatively efficient and that recovers at The relatively high levels of metering and tariffs in Africa least the operation and maintenance cost. However, several suggest that consumers receive a price signal to economize utilities have enforced high connection costs and inequita- on water use. A significant number of utilities apply fixed ble tariff structures that limit the ability of poor households charges with minimum consumption attached, however, to secure access to service or to be able to afford the service and in these cases, low-volume consumers may not receive even when they have access. any kind of price signal over their consumption range. How- ever, the necessity of water demand management is less Source: Banerjee, Foster, and others 2008. a. Tariff structure b. Effective prices at different levels of household consumption 20 19 100 9% 18 90 27% 20% 16 80 27% number of utilities % of utilities 14 13 70 12 60 24% 45% 10 50 8 7 40 20% 49% 6 30 4 3 20 25% 29% 2 1 1 1 10 16% 9% 0 0 4 10 40 + um T r ge r ed f rif ea ea IB ar ap ta T ion im lin lin + household consumption (cubic meters) IB ch sh ck pt min ge + ed blo U ar ge fix ch ar ing < $0.20 > $0.20 and < $0.40 ch ed um as fix ed re ns fix inc > $0.40 and < $0.80 > $0.80 co Note: IBT = increasing block tariff. Water Supply: Hitting the Target? 309 Full capital cost recovery is generally Figure 16.4 Affordability of Cost-Recovery Tariffs in Low-Income Countries affordable in the middle-income countries 100 % of households falling beyond the 5% but in the low-income countries would be 90 affordable to only 40 percent of the popula- 80 affordability threshold tion. Assuming household subsistence con- 70 sumption of 10 cubic meters per month (or 60 about 65 liters per capita a day), a monthly 50 utility bill under full cost-recovery pricing of 40 $1 would be around $10. Based on an afford- 30 ability threshold of 5 percent of household 20 income, full cost-recovery tariffs would prove 10 affordable for 40 percent of the population 0 2 3 4 6 8 10 12 14 16 in low-income countries (figure 16.4). With monthly utility bill (US$) about 10 percent of the national population cost of recovery bill, low bound already enjoying a direct water connection, an cost of recovery bill, medium bound additional 30 percent of the population could cost of recovery bill, upper bound be added to service and be able to afford it. percentage of low-income country population unable to afford to pay The majority of the remaining 60 percent of the population would be able to afford bills of Source: Adapted from Banerjee, Wodon, and others 2008. Note: Tariffs are described as affordable if a monthly bill based on subsistence consumption absorbs around $6 a month that would result either no more than 5 percent of the household's budget. from operating cost recovery or from full cost recovery at a more modest consumption of about 6 cubic meters per month (or 40 liters Congo, Ghana, and Zambia, operating inef- per capita a day). ficiencies can create an economic burden of 0.7­1.0 percent of GDP (figure 16.5). Another source of inefficiency is overem- Improving Utility Performance ployment. State-owned enterprises can be through Institutional Reform social buffers to (very inefficiently) transfer rents or resources to the population. African The operational inefficiencies of water util- utilities have an average of five employees per ities, including revenue undercollection, dis- 1,000 connections, more than twice the two tribution losses, and labor inefficiencies, cost employees per 1,000 connections frequently the region $0.9 billion a year (or as much as used as the international benchmark for devel- 0.15 percent of GDP). Operating inefficien- oping countries (Tynan and Kingdom 2002). cies are rampant among Africa's water utili- Operating inefficiencies have been imped- ties, divided roughly evenly between revenue ing expansion. Inefficiencies not only drain the undercollection and distribution losses. Aver- public purse but they also seriously undermine age collection ratios are relatively high in Africa the performance of utilities. at 90 percent but still short of the best practice One casualty of insufficient revenue is of 100 percent. Average distribution losses in maintenance. The rate of bursts per kilome- Africa are 35 percent, far above the common ter of water main provides some indication of norm of 20 percent, with all countries affected the condition of the underlying infrastructure, to some degree. The overall magnitude of these and hence the extent to which it is adequately inefficiencies can be quantified by compar- operated and maintained. Among African util- ing the revenues available to the utility with ities, a huge variation occurs between low- and the revenues available to an ideal utility that middle-income countries, with bursts ranging is able to charge cost-recovery tariffs, collect from five per kilometer in the latter to just all its revenues, and keep distribution losses over one per kilometer in the former. Utility to the technical minimum (Ebinger 2006; managers must often choose among paying Saavalainen and ten Berge 2006). In the worst salaries, buying fuel, or purchasing spares. cases, such as the Democratic Republic of Frequently, they must cannibalize parts from 310 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 16.5 Economic Burden of Water Utility Utilities with fewer employees per connection Operational Inefficiencies, by Country manage to have, on average, 85 percent of the water supplied with adequate chlorine, in con- Congo, Dem. Rep. of trast to 75 percent of the remaining utilities. Ghana Conversely, utilities with higher hidden costs Zambia tend to deliver slightly higher-quality water. Malawi Institutional reforms hold the key to improv- Cape Verde ing utility performance. Good institutional Uganda frameworks pay off in lowering the inefficiency Sudan of utilities. Utilities that have decentralized or adopted private sector management have Mozambique hidden costs that are substantially lower than Burkina Faso those that have not (figure 16.7). In addition, Rwanda unbundling has a considerable effect on utility Madagascar efficiency; however, unbundling is rare in Africa Kenya and is concentrated exclusively in middle- South Africa income countries, whose superior performance Nigeria can be explained by many other reasons. Con- versely, higher levels of regulation and gover- Lesotho nance, as well as corporatization, are associated Ethiopia with higher efficiency. Namibia Nevertheless, introducing reforms is easier Benin said than done. In recent years, many African Tanzania countries have initiated water sector reforms Niger to improve performance (figure 16.8). This Côte d'Ivoire reform agenda has had two major thrusts: encouraging private participation and improv- Senegal ing governance from within. 0 0.2 0.4 0.6 0.8 1.0 The first thrust has focused on experiment- % of GDP ing with private participation. Overall, 26 unaccounted-for losses collection inefficiencies private sector transactions have occurred in African water, affecting the majority of coun- Source: Briceño-Garmendia, Smits, and Foster 2008. tries in one way or another. Most have been Note: Utility inefficiency is defined as the revenue lost from under- collection of water billed, as well as the value of technical and lease contracts (or affermage). Experience nontechnical losses of water on the distribution network valued at with private sector participation is dispro- the full cost-recovery tariff. portionately concentrated in the francophone countries of West Africa (Côte d'Ivoire, Guinea, other working equipment. The investment Niger, and Senegal), with some exceptions program is another major casualty. (Mozambique and Uganda). Another distinc- Service expansion (measured as a percentage tive feature of the African experience has been of the population resident in the utility service the use of concessions for joint power and area that every year gains access to either piped water utilities, as in Gabon and Mali. water or standposts) is significantly higher for The rate of cancellation of private sector con- more efficient utilities. In particular, utilities tracts for water supply in Africa has been much with low hidden costs achieve an average annual higher than elsewhere. Some 29 percent of pri- increase in coverage of more than 3 percent of vate contracts for water supply in Africa have the resident population, essentially twice as been terminated prematurely, a much higher much as that registered by utilities with high rate than in any other developing region. As a hidden costs (figure 16.6). result, the number of active private operators For similar reasons, more efficient utili- has shrunk to just a handful, with four in South ties deliver better-quality water (figure 16.6). Africa and one each in Cameroon, Cape Verde, Water Supply: Hitting the Target? 311 Figure 16.6 Effect of Utility Inefficiency on Access Expansion and Water Quality a. Access expansion b. Water quality 4.0 90 average annualized increase in access to improved water 85 % of sample passing (% of population) 3.0 chlorine test 80 2.0 75 1.0 70 0 65 employees per connection hidden costs employees per connection hidden costs above average below average Source: Banerjee, Skilling, and others 2008. Note: Hidden cost is a measure of utility inefficiency that sums the revenues lost from underpricing, distribution losses, and undercollection of billed revenues. Employees per connection is a second measure of utility inefficiency that focuses on the extent of overstaffing. Figure 16.7 Hidden Costs and Institutional Figure 16.8 Overview of Reforms Affecting Urban Frameworks Utilities hidden costs (% of revenue) 200 100 80 % of countries 150 60 100 40 50 20 0 0 ion y e cts ion e g nc nc nc lin ion t g ce ion ion ra en at at ge lin rie na nd an nt at tiz liz lat at em ya nd er pe bu rn co liz tiz ra ra gu ov bu ag ex or un ve ra nt po ra ce Eg re lat an un go nt P ce po an or PS h ce gu rm SO de Ec or hig h rm de hig re Ec cto h SO rfo hig SO se pe te iva yes no pr Source: Banerjee, Skilling, and others 2008. Note: PSP = private sector participation; SOE = state-owned Source: Banerjee, Skilling, and others 2008. enterprise. Note: Hidden cost is a measure of utility inefficiency that sums the revenues lost from underpricing, distribution losses, and undercollection of billed revenues. High governance means the and service continuity, but they have had little utility scores above average on the Africa Infrastructure Country Diagnostic indicator of the quality of state-owned enterprise effect on more intractable issues, such as unac- governance (see box 4.1 in chapter 4 of this volume). High regulation means the country scores above average on the Africa counted-for water and access. Lease contracts Infrastructure Country Diagnostic indicator of the quality of have had a substantial effect on improving regulation (see box 4.1 in chapter 4 of this volume). SOE = state- owned enterprise. access and have generally improved operational efficiency. With the exception of Côte d'Ivoire, however, the associated investments have Côte d'Ivoire, Gabon, Ghana, Mozambique, been publicly financed. The lease contracts Niger, and Senegal. in Guinea and Maputo, Mozambique, have been At their best, private sector contracts have affected by problems of coordination between accelerated access while boosting efficiency. the private contractor and the government that Private sector participation has had favorable have held back their progress in some key areas, results in improving utility performance (table such as unaccounted-for water. Overall, private 16.8), with Senegal being particularly notewor- sector contracts accounted for (though did not thy (box 16.4). Management contracts, being necessarily finance) almost 20 percent of the relatively short-term instruments, have had a increase of household connections in the region, material effect on improving revenue collection twice the amount that would be expected given 312 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 16.8 Overview of Private Sector Participation's Effect on Utility Performance Unit change in performance before and after private participation Private sector Household Service Unaccounted-for Collection Labor Country or city participation connections Improved water continuity water ratio productivity Gabon Concession contract +20 -- -- -8 -- -- Mali +15 +29 -- -14 -- -- Côte d'Ivoire Lease contract or affermage +19 +22 -- -- -- +2.6 Guinea -- +27 -- 0 -- -- Maputo (Mozambique) -- +2 +10 -1 +24 -- Niger +9 +3 -- -5 -- +3.2 Senegal +18 +17 -- -15 -- +2.8 Johannesburg (South Management Africa) contract -- -- -- 0 +10 -- Kampala (Uganda) -- -- +6 -2 +12 -- Zambia -- -- +5 -28 +19 -- Source: Adapted from Marin 2009. Note: Household connections and improved water are measured as additional percentage points of households with access; service continuity is measured as additional hours per day of service; unaccounted-for water is measured as reduced percentage points of losses; collection ratio is measured as additional percentage points of collection; labor productivity is measured as additional thousands of connections served per employee. -- Not available. BOX 16.4 Senegal's Successful Experience with Private Sector Participation Senegal's experience with private sector participation in the amount of water actually produced, which was used to water supply is characterized by two remarkable results: determine the operator's remuneration in lieu of the actual (a) an impressive expansion of access and (b) a large increase water sold. Whenever the operator fell short of the non- in operational efficiency that mainly originated from a reduc- revenue water and bill collection targets, the notional sales tion of nonrevenue water. volume would be lower than the actual sales, penalizing the The first result was mainly attributable to a massive subsi- operator. dized connection program sponsored by donors and in part Another innovation in Senegal's public-private partner- to the cash-flow surplus generated by the private operators. ship was the responsibility of the private operator to finance In particular, the social connection program implemented part of the network's rehabilitation using cash flow. This with donor support provided some 129,000 connections approach provided the operator with more flexibility to (75 percent of all new connections installed), benefiting poor identify and reduce water losses, lessening its dependence households living in targeted neighborhoods. A portion of on the public asset-holding company. the new connections ended up losing service because of The effect of these innovations on efficiency has been nonpayment, however, despite tariffs that had been declin- remarkable, making Senegal's affermage a prominent exam- ing in real terms up to 2006 and a social tariff targeted at ple of private participation in Africa. Today, Senegal can poor households that covered the first monthly block of report a level of nonrevenue water comparable to the best 6 cubic meters of subsistence consumption. water utilities in western Europe, but these results also con- The second result owed much to contractual innovations firm that operational efficiency is perhaps the area in which geared toward increasing the private operator's incentives private operators can make the most positive and consis- to perform efficiently. In particular, the contract included tent effect, since the parallel progress in service expansion targets for nonrevenue water reduction and bill collection, required substantial public sector support. backed by financial penalties for noncompliance. These tar- gets were then applied to a notional sales volume based on Source: Adapted from Marin 2009. Water Supply: Hitting the Target? 313 their market share of only 9 percent. Half of to improve internal processes and corporate those gains were made in Côte d'Ivoire alone governance mechanisms, more so than in other (Marin 2009). infrastructure sectors. In particular, a growing Anglophone and francophone countries number of utilities are using performance con- have taken two distinct approaches to sector tracts, for instance, in Lesotho, Uganda, and regulation. About half of the countries (mainly Zambia, although not all of them incorporate anglophone) have established distinct regula- the penalties, performance-based remunera- tory agencies for the water sector, although a tion, and third-party monitoring that make significant number of them have not adopted these mechanisms truly effective. Uganda's use private sector participation. Conversely, a num- of performance contracts to underpin sub- ber of francophone countries with private par- stantial improvements in sector performance ticipation have adopted regulatory frameworks is particularly noteworthy (box 16.5). contractually, without establishing an indepen- dent regulatory agency. No evidence appears to support the superiority of either of these two Reforms in the Rural Space approaches to regulation. Neither does any evi- dence indicate that the creation of a regulatory Africa remains a predominantly rural conti- entity has led to discernible improvements in nent. About 400 million rural inhabitants have utility performance. Even where explicit regu- no form of utility-provided water. Rural cover- latory frameworks have been established, they age of piped-water supply and standposts has typically meet only about half of the corre- barely risen in the past 15 years, and most of the sponding good-practice criteria. gains have come from rural inhabitants mov- The second thrust of the reform agenda has ing up the water supply ladder from surface focused on improving the governance of the water to wells and boreholes. Until improved sector from within. This approach was based water sources serve more of the rural popula- on the recognition that the standard infra- tion, MDG goals will continue to be elusive. structure reform prescriptions were not always In rural areas, the central challenge is to as relevant or as easily applied to the water sec- reduce reliance on surface water through a sus- tor as to other areas of infrastructure, and that tainable network of water access points, which service provision would remain dominated by are most typically boreholes. About half of the public enterprises in the near future. The focus sampled countries are reducing the share of of these reforms has been on moving toward the rural population reliant on surface water corporatization of state-owned enterprises and and in the best cases (Lesotho, Mozambique, on decentralizing responsibilities to lower levels and Uganda) are managing to shift 2­3 per- of government. In addition, some measures cent of the rural population away from this have been taken to improve the governance option each year. In the other half of the coun- of state-owned enterprises, aimed at adopting tries sampled, however, the share of the rural commercial principles and modern manage- population reliant on surface water is actually ment methods. About 80 percent of Africa's increasing. In Chad, Kenya, and Rwanda, an larger water utilities have now been corpora- additional 1 percent of the rural population tized, thereby laying the foundation for a more is forced to rely on surface water, whereas commercial form of management. Close to in Burkina Faso that number rises to 3 percent, half of the countries sampled have decentral- and it is nearly 10 percent in the Democratic ized their water utilities since the mid-1990s to Republic of Congo. bring responsibility closer to local communities; Even in countries that are successful in however, all the francophone countries studied expanding rural access to improved water retain centralized organization of the sector. sources, sustainability is still a concern. A recur- Africa's state-owned water utilities typically ring problem in rural water systems is the lack fulfill only about half of the good-practice cri- of technical or financial capacity to maintain teria for enterprise governance. Since the mid- assets. Decentralization and the breakdown of 1990s, some serious efforts have been made community management arrangements have 314 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 16.5 Uganda's Successful Case of State-Owned Enterprise Reform The National Water and Sewerage Corporation (NWSC) is A review of 10 years of NWSC operations shows that an autonomous public corporation wholly owned by the gains in operational and financial efficiency and service government of Uganda. The NWSC is responsible for provid- expansion have been substantial and impressive relative to ing water and sanitation services in 23 towns to 2.2 million the performance of NWSC's peers in Africa. people, 75 percent of the population in Uganda's large urban centers. NWSC Efficiency Gains Before 1998, large inefficiencies posed the urgent need to revamp operations. They included poor service quality, Performance indicator 1998 2008 very low staff productivity, and high operating expenses, Service coverage (percent) 48.0 72.0 with a collection rate of only 60 percent and a monthly cash Total connections 50,826 202,559 deficit of $300,000. New connections per year 3,317 25,000 Turnaround strategies culminated in establishing area Metered connections 37,217 201,839 performance contracts between the NWSC head office and a number of area managers. The head office performs contract Staff per 1,000 connections 36 7 oversight, capital investment, and regulation of tariffs, rates, Collection efficiency (percent) 60.0 92.0 and charges; the area managers, acting as operators, are Nonrevenue water (percent) 60.0 32.5 therefore responsible for management, operation and main- Proportion of metered accounts (percent) 65.0 99.6 tenance services, revenue collection, and rehabilitation and extension of networks. The objective was to enhance each Annual turnover (U Sh billions) 21 84 area's performance by empowering managers and making Profit (after depreciation) (U Sh billions) -2.0 +3.8 them accountable for results. A comprehensive system of Note: U Sh = Uganda shillings. more focused and customer-oriented targets was designed. Typical performance indicators included working ratio, cash operating margin, nonrevenue water, collection efficiency, Key success factors are identified in the empowerment and connection ratio. Performance evaluation looked at both of staff; the devolution of power from central to regional processes and outputs and was conducted through regular operations; the increased customer focus; and the adop- as well as unannounced inspections. Incentives were both tion of private sector­like management practices, including monetary, including penalties for below-target performances, performance-based pay, "customer pays for a good service" and nonmonetary, including trophies to the best-performing principle, and so forth. Also, the emphasis on planning, areas and departments and publication of monthly, quarterly, systematic oversight and monitoring, information sharing and yearly best, as well as worst, performance. through benchmarking, and the continuous challenging of In fiscal year 2003/04, the area performance contracts management teams with new and clear performance tar- were changed to internally delegated area management con- gets has created a strong system of checks and balances and tracts, aimed at giving more autonomy to operating teams powerfully triggered involvement, engagement, and sense and based on clearer roles, better incentive plans, and a larger of pride on the side of staff, beyond what simple financial risk apportioned to operating teams. The contract framework incentives might obtain. was later consolidated by using competitive bidding as a basis for awarding contracts to the operating units. Sources: Adapted from Muhairwe 2009; NWSC 2006. put greater strain on local governments to at least one in two (the Democratic Republic manage services. This situation leads to rapid of Congo, Madagascar, Malawi, Nigeria, and deterioration of rural water points to where Tanzania). Nevertheless, in the best-case sce- they no longer provide the intended service, narios, the share needing rehabilitation can be and populations are forced to return to relying as low as 10­20 percent (Benin and Uganda). on surface water. On average, one in three rural Inadequate maintenance of rural water water points needs rehabilitation, and for a sig- systems reflects both institutional weaknesses nificant number of countries, the share rises to and inappropriate technology choice. Besides Water Supply: Hitting the Target? 315 weak institutional capacity, undermainte- A strong link exists between institutional nance is worsened by inadequate attention reforms and progress toward the MDG in to technology choice, low pump density, rural areas. The most successful countries restrictive maintenance systems, and lack of in reducing the rural population's reliance on a supply chain to adequately maintain com- surface water are all--without exception-- plex machinery (Harvey and Reed 2006; Oyo among the most aggressive reformers in 2006). In a number of countries, problems Africa (table 16.9). Benin, Côte d'Ivoire, have been caused by divorcing the supply of Mozambique, Namibia, Nigeria, Senegal, hand pumps from the supply of associated and Uganda are outstanding performers in spare parts. Suggested solutions include gov- reducing the share of the population consum- ernments' taking a leading role in initial sup- ing surface water and rank highest in rural ply chain management and coordinating with reform. Conversely, the Democratic Republic donors until the private sector is capable of of Congo, Kenya, Malawi, Niger, and Zambia taking over (Oyo 2006). The unavailability increasingly rely on surface water and score of private supply chains is a result of limited very low on the Rural Reform Index. Burkina population density and income levels of Afri- Faso and Tanzania are two countries that can economies. Several initiatives, including perform poorly on access expansion, which is market demand assessments (Mozambique in 2003), coordination among links in the chain, Figure 16.9 Overview of Rural Water Reforms and development of supply chain products, are under way in Africa. 100 Several countries have made progress with 80 % of countries institutional reforms of rural water. Under- 60 standing what factors drive these differences 40 in performance across countries is important, particularly whether the institutional reforms 20 implemented have made a difference. The 0 y cy y ts reform agenda for rural water comprises five nd lic lic oin en fu po po ag lp er er ry key components: (a) adopting an explicit rural ra at er at ve ru lw at lw co lw of ra re water policy to guide interventions in the sec- ra ru ra ap ru st ru co m tor; (b) developing a map of rural water points so that their functionality can be monitored; Source: Banerjee, Skilling, and others 2008. (c) adopting cost-recovery policies to improve the financial sustainability of systems; (d) Table 16.9 Relationship between Rural Reform Index and Success in Expanding establishing a dedicated central budget fund- Rural Service Coverage ing source for rural water; and (e) creating a Success in getting people off surface water water agency to spearhead the implementation Decreasing reliance Increasing reliance of rural water projects. Reform score on surface water on surface water However, progress on key rural reform mea- Aggressive reformers Benin Burkina Faso sures has been uneven. The extent of progress scoring more Côte d'Ivoire Tanzania than 80% on Rural Mozambique on each of these reform steps is used to create Reform Index Namibia a Rural Reform Index for water. Burkina Faso, Nigeria Côte d'Ivoire, and Uganda score 100 percent Senegal Uganda on this index, whereas Niger scores only 20 Moderate reformers Ethiopia Chad percent. The most widely adopted reform mea- scoring 40­80% Lesotho Rwanda sures are rural water policies and rural water on Rural Reform Index funds, which can be found in almost all the Slow reformers scoring Congo, Dem. Rep. of sampled countries. The least widely adopted less than 40% Kenya on Rural Reform Index Malawi are rural water agencies and mapping of rural Niger water points, which are found in fewer than Zambia half of the countries sampled (figure 16.9). Source: Banerjee, Skilling, and others 2008. 316 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION surprising given their strong track record on focus of the agenda has shifted toward a more institutional reforms. For moderate reform- pluralistic view of public and private sector ers, the results can go either way (box 16.6). roles. The reform agenda also needs to move The degree of reform also affects whether beyond utilities to encompass line ministries rural water points are maintained adequately. and the entire public expenditure framework The percentage of rural water points needing that underpins, and too often hinders, sector rehabilitation tends to be lower for countries investment programs. Room for improvement with more advanced rural reform processes.4 exists in cost recovery so that scarce subsidy Thus, Benin and Uganda score high on sector resources are redirected to promote access reform as well as on maintaining rural water among the poorest. For the majority that points. The converse is true for the Democratic do not enjoy access to piped water, greater Republic of Congo and Malawi. thought needs to be given to how standposts can become more effective for urban water supply and how to get the best out of small- Policy Recommendations scale independent providers. The burgeoning use of wells and boreholes for supply in urban On the basis of this diagnostic, the water sec- areas demands urgent attention from policy tor evidences a number of key areas for policy makers, both to improve their understanding attention. The institutional reform agenda of this phenomenon and to develop suitable remains as relevant as before, even if the regulatory tools. BOX 16.6 Issues Constraining Rural Water Supply in Cross River State, Nigeria Cross River State, one of Nigeria's 36 states, is located in the Water Supply and Sanitation Agency, which remains a section rain forest belt of Nigeria. About 75 percent of its popula- of the Rural Development Agency. Unlike in other states, no tion, 3.25 million people, lives in rural areas and is engaged dedicated ministry champions reforms and allocations. More in subsistence farming. More than 70 percent lives on less important, although a rural water policy exists nationally, than $1 a day. state policies do not necessarily reflect national policy, and Water supply in Cross River State is in crisis. Coverage effective cooperation is not pursued between national and stands at only 25 percent in urban areas and 31 percent in state governments. Responsibilities are decentralized locally, semi-urban and rural areas. Rural water is supplied mainly but the State Rural Water Supply and Sanitation Agency through boreholes with hand pumps and wells, 65 percent continues to be characterized by a weak and poorly funded of which are not functioning. Moreover, no water treat- mandate and loose connections to the national water sector. ment is provided. To meet the MDG for water, an additional Maintenance and rehabilitation of rural water schemes 10,098 boreholes with hand pumps and 2,525 motorized are jeopardized by the lack of skilled labor and the substan- boreholes will need to be built across the state by 2015, a tial underdevelopment of a local private sector. Technical daunting task given its current financial, institutional, and capacity for routine maintenance remains low; spare parts technical capacity. for boreholes are difficult to find and are very expensive Cross River State has recently been the subject of an where available. assessment of rural water supply based on public expenditure Finally, no effective strategy to promote community par- reviews covering the period from 2002 to 2007. The review ticipation has been put in place, resulting in local community shows that lack of adequate budgetary funding and low involvement in rural water provision that remains shallow disbursement efficiency are major constraints. Rural water at best. Absent any sense of ownership, rural communities captures only 0.5 percent of the state's capital budget, and do not take responsibility for preserving and repairing facili- execution ratios average less than 20 percent. Weak insti- ties. Neither would they have the capacity to do so without tutions and fragmented responsibilities translate into feeble adequate training. leadership and rural water falling behind in the political agenda. The sector is the responsibility of the State Rural Source: Iliyas, Eneh, and Oside 2009. Water Supply: Hitting the Target? 317 Continue to Pursue Institutional even if the private sector can sometimes help Reforms execute it. A key lesson from Africa's experience Institutional reforms are the key to improving with lease contracts is the difficulty of achiev- water sector performance. Countries pursuing ing seamless coordination on investment plans institutional reforms create more efficient and between the contractor and the public holding effective sector institutions and promote more company. In addition, incorporating clear con- rapid expansion of higher-quality services. The tractual incentives for efficiency improvements potential dividend is large because address- (for example, by basing the contractor's rev- ing utility inefficiencies alone could make a enues on ideal rather than actual performance substantial contribution to closing the sector parameters) is important (Marin 2009). funding gap in many countries. The new agenda places greater emphasis on Although the majority of African countries broader reforms within governing state-owned have embarked on the sector reform agenda, enterprises. Given the limited scope of private few have completed it. The glass is still half participation, state-owned utilities remain cen- full, but the experience of those countries that ter stage. Without addressing the typical defi- are furthest ahead provides some guidance for ciencies that afflict such enterprises--including the region. numerous and conflicting objectives, political In rural areas, a few critical interventions can interference, and lack of transparency--the make a difference. Establishing a clear sector sector will have difficulty exiting low-level policy, creating a strong central capability for equilibrium. Three key areas for attention are sector financing and project implementation, (a) internal process improvements, (b) increased moving to greater cost recovery, and develop- managerial autonomy, and (c) more stringent ing a system to monitor the condition of rural performance monitoring. Incorporating mea- water points are all measures that, when imple- sures to streamline corporate processes, such as mented as a package, can boost performance. procurement, financial management, and per- In urban areas, the story is more complex. formance management, is essential for strength- The traditional reform agenda of the 1990s has ening the application of commercial principles not fully proved its relevance to the sector. Per- and accountability mechanisms. Measures to haps surprisingly, no clear evidence indicates broaden the board of directors, to increase the that regulation has made a positive contribution use of external audit and independent audit to sector performance across the board. Private of accounts, and to incorporate independent sector participation, although controversial in members from beyond the public sector would implementation, has in many cases proved a help depoliticize decision making and consoli- useful tool for improving operational perfor- date the arm's-length relationship. Adopting mance and efficiency (Marin 2009). Expecta- performance-based monitoring arrangements tions that the private sector would finance that mimic private sector contracts is also of new infrastructure for water utilities have not interest, but only to the extent that they create been met, with negligible private capital flows credible incentives by incorporating meaning- that are dwarfed by public and donor finance. ful rewards and penalties at the personal and However, the private sector has contributed to corporate levels and are subject to third-party expanding access, though typically with public monitoring. funding (Marin 2009). The new reform agenda for water retains a Improve Efficacy of Public Expenditure role for private participation. Lease contracts The bulk of investment in the water sector is by may be the form of private participation best line ministries through the budgetary process, suited to African water utilities. By transferring often with external support. The existing pat- more responsibility to the private sector than terns of spending clearly show that although in a management contract, they have provided utilities are instrumental in delivering services, greater scope for operational improvements. In the general government (using either domes- contrast to concessions, recognition is explicit tic or external capital) continues to make that investment will need to be publicly funded, most of the investment decisions. Therefore, 318 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION a solid public investment appraisal system and appraising, prioritizing, planning, and pro- strong public spending management are pre- curing investment projects has an equally requisites for improving both urban and rural important role. Donors can support countries water supply. in developing good project identification and Major bottlenecks hold back the disburse- appraisal tools that systematically consider ment of public investment funds. Capital bud- the technological alternatives for expanding get execution ratios for public investment in access and that examine the importance of water are relatively low, 75 percent on average. spending on maintenance and rehabilitation In many instances, the capacity to disburse against new investment. budgetary resources in a timely fashion is the binding constraint, rather than their availability Experiment with Different Institutional (Prevost 2009). In Tanzania, steep increases Models for the Unconnected in budget allocations to the sector followed The modest role played by standposts in urban its identification as a priority in the country's water supply is striking. In most countries, gov- poverty reduction strategy; however, disburse- ernment and utility attention continues to focus ments increased at a much slower pace, so on expansion of piped-water connections. This there was no immediately discernible effect on battle is being lost, however, because of a com- access (Van den Berg 2009). bination of rapid urbanization and financially Key aspects of the public expenditure debilitated utilities. Standposts are very limited framework must be addressed. The budgeting in the African urban water scene, are expand- process needs to move toward a medium-term ing relatively slowly, and remain concentrated framework and make stronger links between among the more affluent segments of the sector objectives (such as MDG targets) and population. Simple simulations suggest that if resource allocations. Clear sector plans that utilities were to shift their existing investment detail specific activities and their associated budgets from piped connections to standposts, costs should underpin the budget process. The the rate of service expansion could potentially careful incorporation of maintenance needs double. However, as long as urban households into medium-term sector planning tools is crit- are inconvenienced by higher payments and ical to prevent asset rehabilitation. Administra- longer water collection times, standposts are tive processes that delay the release of budgeted not necessarily going to be a superior solution funds also need to be overhauled. At the same even if they are a cheaper alternative to private time, procedures for procurement, disburse- piped connections. In low-income countries, ment, financial management, and accountabil- resale of water by neighbors through informal ity should be modernized and streamlined. standpost arrangements is almost as prevalent Donor resources are best channeled pro- as formal standposts. grammatically as budgetary support or through The key to this paradox could lie in the sectorwide projects. Given the sector's strong problematic institutional arrangements associ- dependence on external funds, a solid public ated with standposts in African cities. Utilities expenditure management system for African charge little or nothing for standpost water, countries also requires that donors improve the and standpost revenues constitute a negligible predictability of their support and make prog- portion of the revenue base. Therefore, utili- ress on streamlining and harmonizing admin- ties have little financial incentive to expand the istrative procedures. In that sense, a focus on service. Standpost operators, where they exist, multidonor initiatives that pool funds to pro- often charge large markups that make the vide general budgetary support for a sector- service prohibitively expensive and that may wide program of interventions is preferable. generate significant revenues never captured Technical assistance to the sector has tra- by the utility. The quality of service provided ditionally been understood as improving the by standposts can be very low because of the management practices of utilities. However, high rates of malfunction and the very large technical assistance to support line ministries (sometimes implausibly large) numbers of in improving the framework for identifying, people expected to rely on each one. Water Supply: Hitting the Target? 319 Solving this conundrum demands seri- compared with the formal one. Increasing ous attention. The way forward is not yet water production capacity and improving clear, but it calls for intensive experimenta- the efficacy of the distribution network can tion with alternative network designs and significantly affect the welfare of the uncon- institutional setups. Standposts cover a wide nected as well as the connected, because it range of communal arrangements or del- drives down the premium on alternative egated management models, some of which sources of water supply (Keener, Luengo, and may be more promising than others. One Banerjee 2008). option would be to increase the density of standposts to increase competition, with an Increase Cost Recovery with Careful immediate effect on convenience and price of Social Policies water supply. Yard taps, which provide com- Underpricing water is contributing to the munal access to a smaller group of four or five financial weakness of utilities, slowing access contiguous households, lower costs but only expansion, and holding back the quality of partially address the problem of maintenance service. Given that utility customers are drawn and management. Whatever the approach, from the upper end of the income distribution, an important component of the solution will the result is a highly regressive incidence of be to ensure a fairer distribution of revenues subsidies to the sector. A large (and generally between utilities and standpost operators or poor) segment of the urban population is pay- other secondary water retailers. The experi- ing multiples of these prices to access utility ences of the handful of low-income countries water indirectly, and in many cases more than that have achieved more than 20 percent urban the utility cost-recovery price. coverage of standposts--notably Côte d'Ivoire, Countries need to make progress toward Rwanda, and Senegal--deserve some study. further cost recovery while considering the The popularity of the household resale economic circumstances of their populations. option could also be exploited by making it The key principle is to verify the affordability an explicit part of the utility's rollout strategy. of water tariffs with reference to household Household resale of water through yard taps budgets, rather than simply to assume that appears to be a widely used option in many they will be unaffordable. Although the pur- African cities. Survey evidence highlights a chasing power of African households is quite variety of reasons why residents may find this limited, the analysis confirms that operating approach preferable to official standposts. cost recovery is a perfectly feasible objective Neighbors can offer more convenient operating for just about all African countries. Tariffs that hours and better water pressure; because they recover full capital costs also look to be afford- are nearby, less time is required to collect water. able for the richest 40 percent of the popula- In addition, they offer more flexible payment tion in low-income countries, where today's mechanisms than either public standposts piped water coverage is concentrated. Thus, or a private connection (Keener, Luengo, and little economic justification exists for today's Banerjee 2008). Therefore, increasing recogni- subsidies. Countries would be better served tion should be given to this water supply modal- by recovering full costs from their existing ity, removing any legal barriers and potentially customer base and using the resulting cash considering the creation of such household- flow to accelerate access expansion in poor based water retail enterprises as an integral neighborhoods. In the longer term, however, component of utilities' expansion plans. as access to piped water increases, low-income Ultimately, investing in utility production countries will need social tariffs that provide and distribution of water is the best policy for water priced at operating cost recovery levels keeping down the costs of alternatives. Within for a minimum level of consumption to the cities, water markets are strongly connected in substantial share of their population that can- the final price offered to the consumer. The not afford full capital cost-recovery tariffs. more disrupted the formal piped system, the Government entities need to become bet- higher the price will be in the informal sector ter customers. Government entities can easily 320 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION account for 20­30 percent of total billings. Improve Understanding of They can be the worst offenders in paying bills Groundwater's Role in the as well, with a significant lag in payment time. Urban Supply Often, a large chunk of arrears is paid back to Although wells and boreholes have long been the utility books with little indication of future a dominant source of improved water in payment schedules. This uneven payment cul- rural areas, they have also become the fastest- ture has significant implications not only for growing source of improved water supply in the investment planning of utilities, but also African cities. Groundwater, from water wells for developing a broader payment culture (boreholes and hand-dug wells), now sup- across society as a whole. plies one-fourth of urban dwellers and is the The design of tariff structures for water util- fastest-growing source of improved water ities deserves serious rethinking. Most African supply in African cities by far. This is true in utilities are applying increasing block tariffs, in more than just those cities where groundwater the expectation that they will make water tariffs has long been a major source of utility supply more equitable. However, these expectations (such as Lusaka in Zambia and Abidjan in Côte are not always being met (Banerjee, Wodon, d'Ivoire). With utility coverage rates falling and others 2008). About half of the utilities in urban Africa, groundwater has essentially using increasing block tariffs incorporate fixed stepped into the breach, and the rapid growth charges or minimum consumption thresholds of boreholes shows the appetite for lower-cost that actually inflate the costs of water for poor solutions. Investments in boreholes provide the households with modest levels of consump- opportunity to reach a larger proportion of the tion, which becomes counterproductive. A sig- population with relatively modest resources. nificant share of utilities with increasing block One in four urban Africans relies on wells and tariffs also has very high subsistence blocks boreholes for improved supply; that ratio rises (in excess of 10 cubic meters), and as a result to one in two urban Africans in the low-income they end up providing subsidized water to the countries. In Burkina Faso, Malawi, Mali, vast majority of consumers, rather than to a Mozambique, Uganda, and Zimbabwe, the targeted group of low-volume users. share rises as high as three in four. In Malawi, Connection charges should be kept as low Nigeria, and Rwanda reliance on urban wells as possible, and subsidies could be reoriented and boreholes is increasing particularly rapidly, toward connection. The majority of African with more than 3 percent of the population water utilities levy piped-water connection gaining access to this water source each year. charges in excess of $100, an insurmountable Not enough is known about the physical, barrier for low-income households. Utilities institutional, and financial characteristics of intent on universalizing access should explore groundwater use. Household surveys provide a ways to radically reduce connection charges good picture of overall reliance but leave many to levels that are more in line with household questions unanswered. The relative preva- affordability. A number of alternative ways exist lence of simple, shallow hand-dug wells versus to recover connection charges, including pay- professionally drilled boreholes is unknown ment plans that spread them out over time or and so then is the extent to which ground- sharing of connection costs across the whole water supplies are adequately protected from customer base through the general tariff. Con- direct wellhead contamination. The institu- nection costs may also be more suited to public tional arrangements associated with ground- subsidy than water usage tariffs. They have the water supplies are also unclear, particularly the advantage of being one-time payments linked to a extent to which they constitute stopgap services concrete and monitorable action that addresses provided by municipalities versus private or a real affordability constraint. Simulations sug- communal self-supply initiatives. Depending gest that connection subsidies can potentially on the conditions and arrangements, the capi- be much more pro-poor than general subsidies tal costs of such wells could be anywhere from to the water tariff, particularly if some simple $5,000 to $25,000 (or $10 to $20 per capita) targeting mechanisms are used (Wodon 2007). (Foster 2008). Water Supply: Hitting the Target? 321 In addition to growing groundwater reli- Sarah Keener, Manuel Luengo, Dennis Mwanza, ance, African cities are characterized by heavy Eustache Ouayoro, Heather Skilling, Caroline use of low-grade in situ sanitation, mainly in the van den Berg, Quentin Wodon, Guillermo form of unimproved latrines (see chapter 17 of Yepes, and Yvonne Ying, and received extensive support from the Water and Sanitation Pro- this volume). Deployment of latrine sanitation gram network in Africa. at excessive population densities or with lack of 1. The MDG for water supply commits countries proper latrine operation can lead to increased by 2015 to halving the percentage of their pop- groundwater contamination that can affect the ulations without access to an improved water entire urban aquifer providing the groundwater source relative to the baseline situation in 1990. supplies (Xu and Usher 2006). For the purposes of the MDG, "improved water" Furthermore, extensive unregulated use of includes access to piped water or standposts, as groundwater by private actors may prevent the well as some types of wells and boreholes that most rational and efficient exploitation of the are adequately protected. resource for public water supply. In particular, 2. The Joint Monitoring Programme of the World it prevents cities from reaching economies of Health Organization and the United Nations Children's Fund systematically tracks access to scale in groundwater exploitation and from improved water supply and sanitation, but the following the principle of conjunctive surface data constraints are immense. Systematic infor- water and groundwater use that allows ground- mation and data about suppliers' characteristics water to play its natural role as a backup supply and institutional environments are poor when in times of drought (Foster 2008). they exist. Often, even the well-performing ser- Developing an improved understand- vice providers are unrecognized outside their ing of the benefits and risks of groundwater immediate environments, and lessons learned use in fast-growing African cities and towns are not widely shared. Under the auspices of and of how those benefits and risks vary with the Africa Infrastructure Country Diagnostic, the hydrogeological setting is urgent (Foster, a limited effort has been made to use a spe- Tuinhof, and Garduño 2008). This objective cially designed questionnaire on institutional environment, governance structure, and tech- should begin with a city-level appraisal of (a) nical and financial performance to collect data the quantity and quality of available urban covering 51 utilities in 23 countries. This novel groundwater resources; (b) the drivers, dynam- database covering the period from 1995 to 2005 ics, and patterns of usage; and (c) an assess- is paired with household survey data of various ment of the vulnerability of urban aquifers years from 1990 to 2005. to pollution from the land surface. Creating a 3. General government includes central and local groundwater-monitoring framework and pro- governments and special funds when off the mulgating appropriate construction and oper- budget. ation protocols for wells and in situ sanitation 4. Rural water indexes are negatively--if admit- facilities (mainly latrines) would help safeguard tedly weakly--correlated. groundwater quality, but guidelines for safe use of groundwater sources should accompany this References framework. Appropriate governance arrange- Banerjee, Sudeshna G., Vivien Foster, Yvonne Ying, Heather Skilling, and Quentin Wodon. 2008. ments also need to be established, recognizing "Cost Recovery, Equity, and Efficiency in Water the broad reach of groundwater resources, and Tariffs: Evidence from African Utilities." Working should involve water utilities, public health Paper 7, Africa Infrastructure Country Diagnos- authorities, and municipal agencies, including tic, World Bank, Washington, DC. a suitable channel for public consultation. Banerjee, Sudeshna G., Heather Skilling, Vivien Foster, Cecilia Briceño-Garmendia, Elvira Morella, and Tarik Chfadi. 2008. "Ebbing Water, Notes Surging Deficits: Urban Water Supply in The authors of this chapter are Sudeshna Ghosh Sub-Saharan Africa." Background Paper 12, Banerjee, Elvira Morella, Cecilia Briceño- Africa Infrastructure Country Diagnostic, Garmendia, and Vivien Foster, who drew on World Bank, Washington, DC. background material and contributions from Banerjee, Sudeshna G., Quentin Wodon, Amadou Tarik Chfadi, Piers Cross, Alexander Danilenko, Diallo, Taras Pushak, Helal Uddin, Clarence 322 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Tsimpo, and Vivien Foster. 2008. "Access, Afford- Service Providers and Public Stand Posts in ability, and Alternatives: Modern Infrastructure Africa." World Bank, Washington, DC. Services in Africa." Background Paper 2, Africa Keener, Sarah, Manuel Luengo, and Sudeshna Infrastructure Country Diagnostic, World Bank, Banerjee. 2008. "Provision of Water to the Poor Washington, DC. in Africa: Informal Water Markets and Experi- Briceño-Garmendia, Cecilia, Karlis Smits, and ence with Water Standposts." Working Paper 13, Vivien Foster. 2008. "Financing Public Infra- Africa Infrastructure Country Diagnostic, World structure in Sub-Saharan Africa: Patterns, Issues, Bank, Washington, DC. and Options." Background Paper 15, Africa Marin, Philippe. 2009. Public-Private Partnerships Infrastructure Country Diagnostic, World Bank, for Urban Water Utilities: A Review of Experi- Washington, DC. ences in Developing Countries. Washington, DC: Cardone, Rachel, and Catarina Fonseca. 2003. Public-Private Infrastructure Advisory Facility "Financing and Cost Recovery." Thematic Over- and World Bank. view Paper 7, IRC International Water and Sani- Mehta, Meera, Thomas Fugelsnes, and Kameel tation Centre, Delft, the Netherlands. Virjee. 2005. "Financing the Millennium Devel- Collignon, Bernard, and Marc Vézina. 2000. Inde- opment Goals for Water Supply and Sanitation: pendent Water and Sanitation Providers in Afri- What Will It Take?" International Journal of can Cities: Full Report of a Ten-Country Study. Water Resources Development 21 (2): 239­52. Washington, DC: UNDP­World Bank Water Muhairwe, William T. 2009. "Fostering Improved and Sanitation Program. Performance through Internal Contractualisa- tion." Paper presented at World Bank Water Ebinger, Jane. 2006. "Measuring Financial Week, Washington, DC, February 17­20. Performance in Infrastructure: An Application to Europe and Central Asia." Policy Research NWSC (National Water and Sewerage Corpo- Working Paper 3992, World Bank, ration). 2006. "Corporate Plan 2006­2009." Washington, DC. NWSC, Kampala, Uganda. Foster, Stephen S. D. 2008. "Urban Water Supply Oyo, Anthony. 2006. "Spare Parts Supplies for Security in Sub-Saharan Africa: Making Best Handpumps in Africa: Success Factors for Sus- Use of Groundwater." Paper presented at the tainability." Field Note 15, Water and Sanitation Africa Groundwater and Climate Conference, Program, Africa Region, Nairobi. Kampala, Uganda, June 24­28. Prevost, Christophe. 2009. "Benin Rural Water Pub- lic Expenditure Review: Findings, Impacts and Foster, Stephen S. D., Albert Tuinhof, and Hector Lesson Learned." Paper presented at World Bank Garduño. 2008. "Groundwater in Sub-Saharan Water Week, Washington, DC, February 17­20. Africa: A Strategic Overview of Development Issues." In Applied Groundwater Studies in Saavalainen, Tapio, and Joy ten Berge. 2006. "Quasi- Africa: IAH Selected Papers in Hydrogeology, Fiscal Deficit and Energy Conditionality in vol. 13, ed. Segun Adelana and Alan MacDonald, Selected CIS Countries." Working Paper 06/43, 9­21. London: Taylor & Francis. International Monetary Fund, Washington, DC. GWI (Global Water Intelligence). 2005. "The GWI Tynan, Nicola, and Bill Kingdom. 2002. "A Water 2005 Water Tariff Survey." Global Water Intel- Scorecard: Setting Performance Targets for ligence 6 (9). Water Utilities." Viewpoint Note 242, World Bank, Washington, DC. Harvey, Peter A., and Robert A. Reed. 2007. Van den Berg, Caroline. 2009. "Public Expen- "Community-Managed Water Supplies in diture Review in the Water Sector: The Case Africa: Sustainable or Dispensable?" Commu- of Tanzania." Paper presented at Water Week, nity Development Journal 42 (3): 365­78. World Bank, Washington, DC, February 17­20. Hutton, Guy, and Laurence Haller. 2004. Evaluation WHO/UNICEF (World Health Organization/ of the Costs and Benefits of Water and Sanitation United Nations Children's Fund). 2006. Improvements at the Global Level: Water, Sanita- MDG Assessment Report. Geneva and tion and Health Protection of the Human Environ- New York: WHO/UNICEF Joint Monitoring ment. Geneva: World Health Organization. Programme for Water Supply and Sanitation. Iliyas, Mohammed, Dozie Eneh, and Igiri Oside. Wodon, Quentin. 2007. "Water Tariffs, Alternative 2009. "Public Expenditure Review in the Service Providers and the Poor: Case Studies Rural Water and Sanitation Sector for Cross River from Africa." World Bank, Washington, DC. State--Nigeria." World Bank, Washington, DC. Xu, Yongxin, and Brent Usher, eds. 2006. Ground- Keener, Sarah, Sudeshna G. Banerjee, Nils Junge, water Pollution in Africa. London: Taylor & and Geoff Revell. Forthcoming. "Informal Water Francis/Balkerma. Chapter 17 Sanitation: Moving Up the Ladder T arget 7 of the Millennium Development but coverage of unimproved latrines is grow- Goals (MDGs) for sanitation access calls ing much faster than coverage of any of the for halving by 2015 the percentage of the improved alternatives. population in 1990 without improved sanita- Although a bleak picture overall, some tion. At the present pace, Africa will unlikely countries have expanded or upgraded sanita- meet the target either at the regional or (with tion, each year moving as much as 3 percent few exceptions) at the country level. The Joint of their population up the "ladder" to better Monitoring Programme (JMP) of the World forms of sanitation. Ethiopia has done so with Health Organization and the United Nations unimproved latrines; Burkina Faso, Madagas- Children's Fund is charged with assessing the car, and Rwanda with improved latrines; and state of sanitation and progress toward the Senegal with septic tanks. MDG target. The JMP's latest data show only To meet the MDG target for sanitation, coun- modest improvement, from 26 percent of the tries need to spend an estimated 0.9 percent of population with access in 1990 to 31 percent GDP a year, of which 0.7 percent is for invest- in 2006 (United Nations 2008; box 17.1), and ment and 0.2 percent for operation and main- many countries face difficulty in making prog- tenance. A few countries already invest in new ress (Water and Sanitation Program 2006). sanitation facilities up to the recommended Today, about 30 percent of the population level, but many do not. Households pay for practices open defecation (40 percent in rural most of the investment bill, but whether they areas) and about half of the population, urban are spending enough on operation and main- and rural, rich and poor alike, relies on unim- tenance is doubtful. Based on limited evidence, proved latrines, a heterogeneous collection of governments contribute only a small fraction facilities with poorly understood health effects. of investments. The health benefits attaching Flush toilets, mostly connected to septic tanks to these investments are considerable, includ- rather than sewers, remain a luxury, as do ing significant reductions in the incidence of improved latrines, which have made headway diarrhea, intestinal worms, and trachoma. in only a handful of countries. The prevalence The challenges and policy options differ of open defecation has finally started to fall, substantially across and within countries. 323 324 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION BOX 17.1 supply problem and the need for training and local market development. Both responses are relevant, but by starting from the supply side, What Is Improved policy makers can minimize the need for a sub- Sanitation? sidy and promote cost-reducing innovations. Where septic tanks predominate, the chal- The improved sanitation category in lenge is providing access to improved sanita- the data from the Joint Monitoring Pro- tion to a larger, lower-income population. As gramme of the World Health Organiza- tion and the United Nations Children's population densities increase and water con- Fund includes both flush toilets and sumption rises, Africa's burgeoning cities will improved latrines. It also includes half eventually need to develop more extensive of the traditional latrines, an adjustment sewerage networks. Thus, reducing the cost that is made because this modality can- of networks through technological innova- not be disaggregated exactly between tion is critical. improved and unimproved sanitation. The household analysis presented here, based on Demographic and Health Surveys and Multiple Indicator Cluster The State of Sanitation in Africa Surveys, does not adopt this kind of adjustment. Instead, the analysis disag- gregates the different types of improved Rungs on a Ladder and unimproved sanitation to allow a One can think of the different types of sanita- richer discussion of sanitation options tion as rungs on a ladder, with each rung hav- and issues. Notwithstanding these ing a higher investment cost and greater health methodological differences, findings are benefits than the one below (figure 17.1). The broadly consistent with those reported bottom of the ladder is open defecation, a by the JMP. practice harmful to public health. The first rung is unimproved latrines, which comprise Source: Banerjee and others 2008. various kinds of pits that vary greatly in their efficacy but provide at best only basic sani- tary protection.1 The next rung is improved latrines, including a variety of engineered facil- ities such as SanPlat and Ventilated Improved Individual countries or urban and rural Pit (VIP) latrines, and basic pits with slabs.2 regions thereof may face three patterns of When appropriately used, these provide ade- access to sanitation; each requires appropriate quate sanitary protection at reasonable cost. policy responses. The final rung of the ladder is the flush toilet, Where open defecation remains prevalent, which may be connected to either a septic tank people must be encouraged to use latrines, if or (where it exists) the sewerage network. From available. Key factors in changing behavior a health perspective, the most critical move- are a community's commitment to cultural ment is from no service (open defecation) or change and peer pressure. Relatively modest unimproved service (unimproved latrine) to government expenditure to promote hygiene an improved or sanitary service. Once the basic education can raise awareness. level of sanitary protection is reached, returns Where unimproved latrines are prevalent, in health benefits diminish with each higher they should be upgraded to improved models. rung on the sanitation ladder. However, upgrading faces constraints on both Unimproved latrines are the most preva- the demand and supply sides. The significant lent sanitation option in Africa, but under- cost of improved facilities suggests a demand standing the health benefits they can deliver problem and the need for capital subsidy. Lack is difficult. Classifying unimproved latrines is of domestic construction capacity suggests a complicated by the variety of installations under Sanitation: Moving Up the Ladder 325 Figure 17.1 The Sanitation Ladder Figure 17.2 Percentage of the Population Sharing Toilet Facilities open cost fixed-point defecation defecation 50 flush toilet 45 improved 40 not latrine unimproved % of countries 35 acceptable latrine 30 25 20 health benefits 15 10 Source: Morella, Foster, and Banerjee 2008. 5 0 0­20 20­40 40­60 60­80 % of population sharing facilities this basic label. Sometimes, an unimproved Source: Banerjee and others 2008. latrine can, with some modification, provide enough sanitary protection to be regarded as improved. The extent to which latrines deliver to access the facilities but also they may have the intended health benefits depends on how to pay significant surcharges to facility own- they are used. Even basic latrines can provide ers. More important, maintenance of shared protection if they are covered and emptied in facilities is often poor, which poses health a timely fashion, and if hands are washed after risks and may discourage use. use. Conversely, improved latrines will not pro- vide sanitary protection if people do not use Differing Patterns of Access them properly or do not use them at all, for More than one-third of the population-- example, if their installation is not accompa- mostly in rural areas--must still defecate in nied by sufficient efforts in hygiene promotion the open (table 17.1). Unimproved pit latrines and social marketing. are by far the most prevalent facilities in both Waterborne sewerage systems are rare in urban and rural areas. Improved sanitation Africa. Only half of the large cities operate reaches less than 20 percent of the national a sewerage network at all. Only in Namibia, population, less than 10 percent in rural areas. South Africa, and the exceptional case of Coverage of improved latrines is no greater Senegal do some of the utilities covering the than that of septic tanks, despite the signifi- largest cities provide universal sewerage cover- cant cost difference between them. Only 10 age. More typical (as in Côte d'Ivoire, Kenya, percent of the national population has the Lesotho, Madagascar, Malawi, and Uganda) advantage of a septic tank; coverage in rural are sewer networks that reach barely 10 per- areas is practically negligible. In urban areas, cent of the population within the service area septic tanks are much more common than of an urban utility. Little more than half of the improved latrines. households with piped water also have flush Access to sanitation varies dramatically toilets, which are often connected to septic across income groups (figure 17.3). Unim- tanks rather than to sewers. proved latrines are by far the most egalitarian Typical urban sanitation includes the form of sanitation, accounting for about sharing of facilities by multiple families 50 percent of households across income ranges. (figure 17.2). More than 40 percent of all Open defecation is widely practiced in the low- urban households share their toilet facilities est income quintile and not practiced at all in with other households; in Benin, Burkina the highest. Conversely, improved latrines and Faso, Ghana, Guinea, Madagascar, and the septic tanks, virtually nonexistent among the Republic of Congo, the figure is more than poorest quintiles, are used by the richest 20­40 50 percent. Sharing sanitation facilities implies percent of the population. Access to improved that not only must household members wait latrines parallels that of septic tanks, suggesting 326 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION that despite their lower cost, improved latrines groups highlights a crucial issue: the most vul- remain something of a luxury, with little suc- nerable populations are failing to benefit from cess in penetrating the middle of the income efforts to improve sanitation. distribution. More important, the minimal Not only are unimproved latrines the most presence of improved sanitation across poorer prevalent facilities in Africa, but also their use is the fastest growing. In recent years, they have been used by an additional 2.8 percent Table 17.1 Patterns of Access to Sanitation in Africa percentage of population of the population each year in urban areas and an additional 1.8 percent in rural areas, Open Traditional Improved Septic Area defecation latrine latrine tank which is more than twice the rate of expan- Urban 8 51 14 25 sion of septic tanks and improved latrines combined (figure 17.4). Growth in the use of Rural 41 51 5 2 unimproved latrines is concentrated among National 34 52 9 10 the poorer quintiles, whereas growth in the Source: Morella, Foster, and Banerjee 2008. use of improved latrines and septic tanks is concentrated among the richer quintiles. Figure 17.3 Coverage of Sanitation by Budget Because the MDG target focuses on the Quintile two most improved sanitation options, the 100 expanding use of unimproved latrines does not always fully register in policy discussions. 80 Meanwhile, the prevalence of open defecation % of households 60 in Africa has finally begun to decline, albeit at a very modest pace. 40 Notwithstanding the overall dismal pic- 20 ture of sanitation in Africa, a number of countries have made major strides in recent 0 poorest second third fourth richest years, moving more than 3 percent of their quintile populations up the sanitation ladder annu- septic tank improved latrine ally. Côte d'Ivoire and Ethiopia have achieved unimproved latrine open defecation these results with unimproved latrines (figure 17.5), Madagascar and Rwanda with Source: Morella, Foster, and Banerjee 2008. Note: In terms of household spending, poorest quintile = poorest improved latrines (figure 17.6), and Senegal 20 percent of the population; second quintile = second-poorest with septic tanks (figure 17.7). Ethiopia is 20 percent of the population; third quintile = middle 20 percent of the population; fourth quintile = second-richest 20 percent making the most rapid progress in reduc- of the population; richest quintile = richest 20 percent of the ing open defecation, each year moving more population. Figure 17.4 Annual Growth in Coverage of Sanitation Types, 1990­2005 a. Urban areas b. Rural areas 3.0 3.0 annual change in coverage annual change in coverage 2.0 2.0 (% of population) (% of population) 1.0 1.0 0 0 ­1.0 ­1.0 septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation Source: Morella, Foster, and Banerjee 2008. Sanitation: Moving Up the Ladder 327 Figure 17.5 Moving Up to the Bottom Rung of the Sanitation Ladder: Côte d'Ivoire and Ethiopia, 1990­2005 a. Côte d'Ivoire b. Ethiopia 8.0 8.0 annual change in coverage annual change in coverage 6.0 6.0 (% of population) (% of population) 4.0 4.0 2.0 2.0 0 0 ­2.0 ­2.0 ­4.0 ­4.0 ­6.0 ­6.0 septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation Source: Morella, Foster, and Banerjee 2008. Figure 17.6 Upgrading Latrines: Madagascar and Rwanda, 1990­2005 a. Madagascar b. Rwanda 8.0 8.0 annual change in coverage annual change in coverage 6.0 6.0 (% of population) (% of population) 4.0 4.0 2.0 2.0 0 0 ­2.0 ­2.0 ­4.0 ­4.0 ­6.0 ­6.0 septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation Source: Morella, Foster, and Banerjee 2008. Figure 17.7 Mainstreaming Septic Tanks: Senegal, across urban and rural areas within countries 1990­2005 that can help structure policy alternatives 8.0 (figure 17.8). Urban areas tend to follow one annual change in coverage 6.0 of three possible patterns. The most common (% of population) 4.0 is where unimproved latrines are the domi- 2.0 nant mode of sanitation; the second is where 0 improved latrines prevail, but unimproved ­2.0 latrines still constitute a large share; the third ­4.0 is where half of the population uses septic ­6.0 tanks and half uses unimproved latrines, but septic improved unimproved open tank latrine latrine defecation where coverage of improved latrines is nearly nonexistent. Rural areas similarly tend to fol- Source: Morella, Foster, and Banerjee 2008. low one of three possible patterns. As in urban areas, the most common is where unimproved latrines dominate; the second is where open than 2 percent of its population away from defecation is most prevalent; and the third is the practice. where coverage of improved latrines is increas- These overall trends conceal contrast- ing, though most people rely on unimproved ing patterns across groups of countries and latrines or open defecation. 328 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Figure 17.8 Characterizing Patterns of Access to Sanitation across Urban and Rural Areas a. Urban: prevalence of unimproved latrinesa d. Rural: prevalence of open defecationd septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation b. Urban: prevalence of improved latrinesb e. Rural: prevalence of unimproved latrinese septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation c. Urban: bimodal patternc f. Rural: prevalence of improved latrinesf septic improved unimproved open septic improved unimproved open tank latrine latrine defecation tank latrine latrine defecation Source: Morella, Foster, and Banerjee 2008. a. Data include the Central African Republic, Chad, Comoros, the Democratic Republic of Congo, Ethiopia, Guinea, Lesotho, Malawi, Mali, Mauritania, Mozambique, Nigeria, the Republic of Congo, Sudan, Tanzania, and Uganda. b. Data include Benin, Burkina Faso, Cameroon, Ghana, Madagascar, Niger, and Rwanda. c. Data include Côte d'Ivoire, Gabon, Kenya, Namibia, Senegal, South Africa, Zambia, and Zimbabwe. d. Data include Benin, Burkina Faso, Chad, Côte d'Ivoire, Ethiopia, Mauritania, Mozambique, Namibia, Niger, and Sudan. e. Data include Cameroon, Comoros, the Democratic Republic of Congo, Gabon, Ghana, Guinea, Kenya, Malawi, Mali, Nigeria, the Republic of Congo, South Africa, Tanzania, Uganda, and Zambia. f. Data include the Central African Republic, Lesotho, Madagascar, Rwanda, Senegal, and Zimbabwe. Dispersed Institutional Effort example) are in the hands of different public As shown by an institutional survey of sector and private players, which prevents one agency institutions in 24 countries, complexity, a mul- from championing the sector and contributes tiplicity of actors, and lack of accountability for to sanitation's falling between the cracks. The sector leadership are the three salient features of recent trend toward government decentraliza- the institutional framework governing the san- tion has complicated the capture of adequate itation sector. Unlike water, many parts of the public resources for sanitation and allocated supply chain for sanitation (hygiene promotion, responsibilities to entities that lack technical latrine construction, and latrine emptying, for capacity. Fifteen countries have adopted formal Sanitation: Moving Up the Ladder 329 national sanitation policies, and most countries available evidence report negligible public have an accepted definition of sanitation and investment on sanitation of 0.02 percent of a hygiene promotion program. But only seven GDP, on average, although serious measure- countries have policies that include cost recov- ment problems mean a large portion of public ery, and only eight have a sanitation fund or a investment in sanitation is likely not separately dedicated budget line (in some cases, funded coded from water supply (Briceño-Garmendia, exclusively by donors, as in Chad and Ethiopia, Smits, and Foster 2008). Nevertheless, house- or by a combination of the government, sector holds appear to be footing the bulk of the levies, and donors). Côte d'Ivoire has the only investment bill. In countries with very low fund financed entirely by sector levies. current spending, whether households alone will be able to increase investment up to the level needed is uncertain. In addition, noth- Households Foot the Bill ing is known about their spending on opera- Building on earlier work (Mehta, Fugelsnes, tion and maintenance, which is estimated to and Virjee 2005), one can estimate the overall require an additional 0.2 percent of GDP in price tag for reaching the sanitation MDG at the future. Public spending on operation and $6 billion a year, or roughly 0.9 percent of maintenance appears to have already reached Africa's GDP (Morella, Foster, and Banerjee this level, but evidence is limited, and spending 2008). Capital investment needs based on refers mostly to sewer networks. Operation minimum acceptable asset standards and and maintenance of on-site sanitation remain accounting for both new infrastructure and a household responsibility, and facilities are rehabilitation of existing assets can be con- notoriously poorly maintained. servatively estimated at $4.5 billion a year Although the costs of meeting the MDG (0.7 percent of the region's GDP). The main- sanitation target are high, so is the associ- tenance requirements are $1.5 billion a year ated health dividend (Hutton and others (0.2 percent of the region's GDP). 2007). Sanitation reduces the risk of intestinal No reliable data exist on sanitation worms, diarrhea, and trachoma, and it is very expenditures because individual households important--more than access to safe water-- undertake so much of the expense. However, in fighting hookworm infection (Esrey and recent investment can be estimated from others 1991). Access to adequate sanitation household surveys, using the number of house- reduces diarrhea incidence by an estimated holds acquiring access in successive years and a 36 percent. Trachoma incidence was reduced standard unit cost. Because this method treats by 75 percent in Gambian villages solely by all new or improved facilities as newly built, it controlling flies that breed when excreta are may overestimate the cost of increasing access not safely disposed of (Emerson and others mainly by upgrading unimproved latrines. 2000). One study estimates that reaching the This approach suggests that, on average, Afri- MDG for both water and sanitation in Africa can countries are investing about 0.5 percent of would prevent 172 million diarrhea cases a GDP in new sanitation facilities, quite close to year, saving $1.8 billion in treatment costs the recommended investment level. Half of the (Hutton 2000).3 countries appear to invest less than 0.7 percent of GDP, which is the level needed to meet the sanitation access MDG. Some countries, par- Challenges and Policy Options ticularly Madagascar and Rwanda, appear to have made rapid progress, investing as much as One of the strongest findings emerging from 1.0 percent of GDP. At the other end of the this review is how much the sanitation chal- spectrum, Kenya, Lesotho, Namibia, and lenge differs across countries and between Zambia spend less than 0.2 percent of GDP. urban and rural areas within the same country. How much of the estimated total spend- Decisions about where to focus policy efforts ing on sanitation comes from the public purse along the sanitation ladder should be informed is hard to pin down. The few countries with by access patterns. Recommendations will 330 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION therefore distinguish the different groups in latrines without accompanying hygiene edu- identified in figure 17.8. Judicious and low- cation led to only 37 percent of men using the cost public interventions can leverage house- facilities despite 100 percent coverage (World hold spending for construction of latrines. Bank 2002). Hygiene education is critical The ultimate objective should be to pro- regardless of the type of sanitation challenge a vide universal access by expanding service and country faces; safe disposal of feces and hand reducing open defecation as much as possible. washing with soap protect health in all sani- Policy makers are often tempted to focus on tation settings. Promoting hygiene can start a rungs of the sanitation ladder well above the virtuous cycle that builds demand for better realities of their societies: for example, chan- sanitation, raising awareness of the benefits of neling limited public resources into sewer net- sanitation and establishing codes of conduct works that serve only a few thousand people and new life standards. while overlooking the urgent need to lift mil- Incorrect use of latrines can dramatically lions more people away from open defecation. reduce or even reverse their health benefits. A Policy attention needs to focus on moving facility is sanitary and safe not only because people up from the lowest rungs of the ladder. of the technology and material used but also More expensive options should be left to house- because of good practices and behaviors, holds with the resources to take them up. such as keeping the facility contained and African countries may face high preva- clean. An improved latrine that is not cor- lence of open defecation, especially in rural rectly used and emptied still poses high risks areas; dominance of unimproved latrines; or of environmental contamination and disease. significant development of septic tanks that Thus, rolling out a physical investment pro- reach a small share of the population, mainly gram without accompanying promotion of wealthier urban residents. The policy options hygiene makes little sense. Moreover, effec- for each issue are presented as separate cases tive hygiene promotion alone may stimulate below, and countries may need to use different self-financed household investment in bet- combinations of these approaches. The first ter facilities. Too often, these "soft" aspects priority is to stimulate demand for sanitation of sanitation are overlooked, and priority is and behavior change where open defecation given to the "hard" aspects, such as installing prevails. The second is to ensure an adequate and upgrading infrastructure. supply of improved sanitation options in Changing behavior requires sustained settings dominated by unimproved latrines, communication and public education at the before evaluating the need for policy interven- community level. Understanding the moti- tions on the demand side of the market. The vations that interest people in hygiene and third is to expand access to improved sanitation sanitation is important. Health is one consid- across larger shares of the population, which in eration, but not necessarily the foremost in high-density settlements requires making sew- people's minds: convenience, dignity, and social erage more affordable. status may be more important. Adapting hygiene and sanitation promotion programs to cultural Stimulate Demand for Sanitation and and institutional norms and intensely market- Behavior Change Where Open ing them to stimulate communitywide involve- Defecation Prevails ment are critical. Peer pressure--to improve Unlike other infrastructure services, demand one's status--can also help. When a commu- for sanitation cannot be assumed. Populations nity recognizes preferred behaviors, pressure to accustomed to open defecation may require conform arises, and social structures and lead- a substantial change in cultural values and ers begin to contribute. A successful example is behavior to use a fixed-point facility. Without the Southern Nations Regional Health Bureau's such change, people may not use latrines at all sanitation advocacy campaign launched in or they may use them in a way that undermines 2003 in southern Ethiopia. It increased latrine the potential health benefits. A study in south- coverage from 13 percent of the population to ern India showed that large public investment 78 percent in just two years (box 17.2). Once Sanitation: Moving Up the Ladder 331 BOX 17.2 Ethiopia's Success with a Community-Led Program The southern region of Ethiopia--home to diverse cultures including zero subsidies but allowing the community to devise and scores of ethnic groups--has a population of 15 million, its own innovative and affordable models. much larger than many African countries. Population density With a modest but dedicated sum of money, a mass com- varies, peaking at 1,100 people per square kilometer in the munication campaign was launched using the slogan "Sani- Wanago district. tation is everyone's problem and everyone's responsibility." It In early 2003, access to on-site sanitation was lower than promoted sustainable and affordable sanitation by creating 13 percent, below the national average of 15 percent (see awareness and encouraging self-financing across all income figure). Traditional latrines were most prevalent but scarcely quintiles. Close collaboration with all stakeholders helped to used, poorly maintained, smelly, and dangerous to children build consensus and capacity to facilitate community involve- and animals. Meanwhile, population expansion, growing ment in hygiene promotion and supervision of construction. household densities, and deforestation were combining to At the household level, women were identified as the reduce private options for open defecation. main drivers of latrine construction. At public consensus- building meetings, they complained about how open def- Latrine Construction, 2002/03­2005/06 ecation directly affects their lives, highlighting the health risks of contact with feces in the banana plantations and 100 88.8% in the fields where they collect fodder for cattle. They also 78.8% complained about the bad smell and the embarrassment 80 of seeing people defecate openly. Featured stories cited % increase 60 51.7% shame as an important factor in consensus building and a strong motivator for latrine construction. Volunteer com- 40 munity health promoters went house to house across vil- 20 12.8% lages with health extension workers and members of the subdistrict health committee to persuade households to 0 2002/03 2003/04 2004/05 2005/06 build latrines, and then they supervised construction. annual increase in latrines with 2005/06 projection Alongside other gains in public health, pit latrine own- ership rose from less than 13 percent in September 2003 Source: Southern Nations Regional Health Bureau. to more than 50 percent in August 2004. By August 2005, it had reached 78 percent, and a year later was on track The Southern Nations Regional Health Bureau, charged by toward reaching 88 percent. the national Ministry of Health with promoting sanitation and hygiene, applied a community-led total sanitation approach, Source: Reproduced from Water and Sanitation Program 2008. a culturally appropriate formula is found, dra- of moving people away from open defecation. matic change can be achieved with modest The problem is rather about improving facili- public spending focused on promoting sanita- ties. The debate centers on whether the main tion rather than on financing hardware. impediment to upgrading latrines comes from the supply side or the demand side. Ensure Adequate Supply before Using standardized unit costs from Senegal's Addressing Demand in Settings sanitation sector, one can estimate the percent- Dominated by Unimproved Latrines age of a household's monthly budget that would Where unimproved latrines prevail, the cen- be absorbed by the up-front investment in dif- tral problem is how to upgrade them to more ferent types of sanitation facilities (table 17.2). hygienic facilities so that the full health benefits Although unimproved latrines are affordable of fixed-point defecation can be achieved. Coun- across all income levels, improved latrines tries where unimproved latrines are widely used cost much more than a month's household have already overcome the behavioral challenge income in Senegal--even for households 332 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Table 17.2 Cost of Sanitation Facilities in Senegal percentage of monthly household budget Poorest Second Third Fourth Richest Facility National Rural Urban quintile quintile quintile quintile quintile Septic tank 289 427 209 641 491 396 292 167 Improved latrine 194 286 140 430 330 266 196 112 Unimproved latrine 22 32 16 48 37 30 22 13 Monthly household budget (2002 $) 227 154 315 102 134 166 225 394 Source: Morella, Foster, and Banerjee 2008. in the highest income group. These findings prevalence of improved latrines is low, even in are consistent with the patterns of access to middle-income countries, except in a handful sanitation observed across the socioeconomic of cases. Second, 40­50 percent of the popula- spectrum in Africa. The fact that half of Afri- tion use unimproved latrines, even among the can households have invested in unimproved highest-income groups who may be able to pay latrines in the absence of any subsidy suggests for more advanced facilities. that these modest investment costs are afford- A weak private sector dominated by small able across the income spectrum. The fact that entrepreneurs at the local level compounds the improved latrines are found only among the supply problem. Latrine construction demands wealthiest households indicates that affordabil- skills not widely available, and small enterprises ity may well be an issue in this case. In addition, often do not have the resources to develop new in urban areas, poor dwellers in slum settings skills or adopt new technologies. often do not own their land or house and so Supply bottlenecks should be tackled first. have fewer incentives to invest in improving Otherwise, subsidy resources may be wasted their living conditions. on households that could have financed the The appropriate policy response to these facilities on their own. Allowing the local mar- demand-side constraints likely entails a ket to develop also provides space for inno- public subsidy for the additional capital vation that can lower the cost of improved costs associated with a standard package of latrines. Technological innovation is needed improved facilities. However, a subsidy may to secure greater health benefits with cheaper have drawbacks, including distorting demand variants that are tailored to a locality. This and markets. Subsidies can reduce the demand innovation should be grounded in a better of households with the ability to pay. More- understanding of the most prevalent designs over, suggesting a standard package may make for unimproved latrines in any given local- poor households feel entitled to such a facility, ity and should explore how relatively minor regardless of whether it is the most appropriate changes in these designs could help achieve for their circumstances and geographic loca- health benefits. tion. Widespread adoption of a standard could Policies need to address supply-side limita- also discourage innovations that may lower tions. Government support is best channeled costs. Therefore, in the African context, many toward (a) conducting research and devel- other policy measures likely need to be taken oping products, (b) marketing latrines, and before subsidies become relevant. (c) opening supply channels for key inputs. From the supply side, poor knowledge in Training small service providers and provid- the construction sector about required designs, ing access to credit can also help. The National a lack of skilled construction workers, and a Sanitation Program in Lesotho, established shortage of materials can explain the low prev- 20 years ago, is dedicated to sanitation pro- alence of improved latrines. Access patterns motion and private sector training. House- already provide some clues that supply-side holds directly employ private latrine builders issues are a real constraint in Africa. First, the trained under the program. The program has Sanitation: Moving Up the Ladder 333 increased national sanitation coverage from in Latin America reveal savings of up to 20 percent of the population to 53 percent. 65 percent (Melo 2005). Pilot condominial systems are being implemented in several Make Sewerage More Affordable in African countries, most notably in the periur- High-Density Settlements ban areas of Dakar, Senegal. By 2009, the Dakar In much of Africa, on-site sanitation is the system is expected to furnish 60,000 house- most cost-effective and only practical way to holds (270,000 people) with on-site sanita- secure the health benefits of hygienic disposal tion and to support 160 condominial schemes of excreta. Nevertheless, on-site sanitation serving 130,000. has its limits. As the urban population grows, water consumption also increases, creating the challenge of safely returning large vol- Several Common Challenges umes of wastewater. In addition, given the Remain for All Countries growing urban population densities, limited land constrains the use of latrines (par- Irrespective of a country's position on the ticularly the simpler types), which require sanitation ladder, several common challenges rotation of sites. At high population densi- cut across all sanitation settings: (a) secur- ties, sewerage systems are both more suitable ing fiscal space for sanitation, (b) coordinat- and more cost-effective. ing the numerous players in the sector, and Whereas the annual population growth in (c) developing a more refined approach to Africa averages 2.5 percent, the urban popula- measuring progress. tion is growing at 3.9 percent. By 2020, nearly 60 percent of Africa's population will be in urban areas, and within 20 years, the popula- Securing Additional Resources tion of most African cities will have doubled. The unglamorous nature of sanitation puts Eventually, Africa's burgeoning cities will need it at a disadvantage in the competition for to develop more extensive sewerage networks. fiscal resources. Government decentraliza- The statistics on affordability in table 17.2 are tion and poor accounting for sector expen- particularly worrisome. If households are ditures impede understanding of the exact struggling to afford improved latrines, they amount of public funds allocated. Fewer than are much less likely to be able to afford water- half of the countries surveyed reported any borne sewerage, and the public subsidies to spending on sanitation, and those that did support such sewerage networks are equally averaged no more than 0.23 percent of GDP, unaffordable. Reducing the cost of sewer net- including both investment and operation and works through technological innovation is maintenance (Briceño-Garmendia, Smits, therefore critical. and Foster 2008). One lower-cost alternative that was devel- At the 2008 African Conference on Sani- oped in Latin America but that could be tation and Hygiene in Durban, South Africa, explored in Africa is condominial sewerage. governments committed themselves to raising These networks are designed to keep costs public expenditure on sanitation to 0.5 percent down by having the public collection network of GDP by 2010. This commitment would just touch each housing block (or condo- require spending close to the levels needed to minium) instead of surrounding it, with the reach the MDG target, but reaching the target pipes serving each household then laid within will still be difficult because of the need to the block itself at the residents' initiative. make up for lagging past performance. Better Decentralized microsystems of treatment and accounting of public expenditure on sanita- disposal can also replace the conventional cen- tion will also be needed to monitor progress tralized treatment system. Construction costs toward the target. are reduced by using small-diameter pipes, Although governments are called on to buried at relatively shallow depth, with work provide more resources, innovative financing partially carried out by residents. Experiences approaches that help providers and operators 334 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION are also needed. Cost recovery has proved to of sanitation functions, even within the public be a limited incentive because the only tariffs sector, prevent one entity from leading, and in sanitation are those on sewerage, and they sanitation issues fall between the cracks. apply only to the minority of the popula- A key policy issue is therefore to identify and tion served by that network. Moreover, most empower a clear sanitation champion within African utilities responsible for providing the public sector. In Senegal, the decision to wastewater services also supply water, and the take sanitation seriously was expressed through lack of accounting separation between these the creation of a dedicated sanitation utility. services makes it likely that water pays for Senegal was also the first country to establish sanitation. Burkina Faso has taken an innova- a government body at the national level-- tive approach by levying a sanitation tax on the Ministry for Prevention, Public Hygiene the water bill, which is then used to subsidize and Sanitation (recently reorganized as the access to improved sanitation facilities in Ministry of Urban Affairs, Housing, Urban Ouagadougou (box 17.3). Water, Public Hygiene and Sanitation)--to coordinate sector activity. Although creat- Needed--a Champion for the ing a ministry in the central government may Sanitation Sector not always be necessary, Senegal provides an Given that on-site sanitation, rather than water- important lesson in singling out one entity borne sewerage, will likely continue to domi- with a clear mandate to lead. nate sanitation in Africa, households rather than governments will remain center stage. Measuring Progress Even so, the government's role in promoting Although the JMP has made strides in moni- demand and addressing supply bottlenecks toring progress toward the MDG target for remains. Too often, dispersion and duplication sanitation, no commensurate effort has been made to create detailed and frequent country-level monitoring and evaluation sys- tems critical to guiding policy interventions. BOX 17.3 Most countries have no system, and in the countries that are developing a system, pro- viding a clear picture of the sector is not yet Burkina Faso's Sanitation Tax possible. Moreover, monitoring and evalu- The on-site sanitation problems in Ouagadougou are specifically ation systems rarely measure the effect of addressed in the Sanitation Strategic Plan being implemented by the improved sanitation on health. national public utility in charge of water supply and sanitation. At the country level, better monitoring and A sanitation marketing approach has enhanced construction ser- evaluation systems could be built by ensuring vices offered to households by small providers and stimulated house- more coordination at the ministerial level, for hold demand for improved sanitation facilities. Some 700 masons and social workers have been trained since the beginning of the instance, between the ministry in charge of sani- program. tation and the ministry in charge of health. In Burkina Faso's national utility offers to provide part of the mate- addition, a larger role should be played at the rial free to households--equivalent to about a 30 percent subsidy-- local level, especially by the decentralized techni- with the households financing the rest. The utility finances the subsidy cal departments, in collecting data and monitor- through a small sanitation tax on the water bill. ing progress. This would require more capacity This example shows that on-site sanitation corresponds to a and resources from the central government. strong demand from urban dwellers, with more than 60,000 pieces A limitation of the JMP's framework is of sanitation equipment subsidized so far--latrines as well as gray the inability to discriminate among the levels water­removal systems. It also demonstrates the importance of a of sanitary protection provided by different local financing mechanism. Donors have contributed to the mecha- variations within the large class of unim- nism, but only modestly. Most of the funds come from the tax on the water bill. proved latrines, which will continue to domi- nate African sanitation. Unimproved latrines Source: Reproduced from Water and Sanitation Program 2008. include a heterogeneous collection of instal- lations, only some of which can be regarded Sanitation: Moving Up the Ladder 335 as improved sanitation. Unfortunately, the are estimated at $0.50 or less. As for economic JMP's household survey instruments that losses from lost time at work and school and track progress toward the MDG target cannot from death, Hutton relies on the concept of discriminate among the qualities of installa- minimum wage rates for his estimates, adjusted to reflect the varying productivity of the dif- tions within the unimproved latrine category. ferent countries (for each country, the value As a result, the data on progress in sanitation of the minimum wage rate must be no larger in Africa are blurriest precisely where most than the local gross national product per capita of the progress is taking place. The precision and no smaller than the manufacture added of household survey instruments should be value). Both health and economic benefits improved in this respect. are presented, assuming that all interventions Tracking the intermediate goal of increas- were implemented within 2000. To account for ing the share of households making use of population growth, the projected population some kind of sanitation facility, even if it is figures for 2015 are used. an unimproved latrine, may also be relevant, given that this is where Africa has been making References the greatest progress. Banerjee, Sudeshna G., Quentin Wodon, Amadou Diallo, Taras Pushak, Helal Uddin, Clarence Tsimpo, and Vivien Foster. 2008. "Access, Afford- Notes ability and Alternatives: Modern Infrastructure The authors of this chapter are Elvira Morella, Services in Sub-Saharan Africa." Background Sudeshna Ghosh Banerjee, and Vivien Foster, who Paper 2, Africa Infrastructure Country Diagnos- drew on background material and contributions tic, World Bank, Washington, DC. from Piers Cross, Pete Kolsky, Marianne Leblanc, Briceño-Garmendia, Cecilia, Karlis Smits, and Eustache Ouayoro, and Ede Perez. Vivien Foster. 2008. "Financing Public Infra- 1. Unimproved latrines refer to various kinds of structure in Sub-Saharan Africa: Patterns, Issues, pits for excreta disposal. Well known in Africa, and Options." Background Paper 15, Africa Asia, and Latin America, they normally consist Infrastructure Country Diagnostic, World Bank, of a simple pit covered with logs, not usually Washington, DC. roofed, and sometimes with no walls. Emerson, Paul M., Sandy Cairncross, Robin L. 2. Improved latrines--comprising SanPlat, VIP Bailey, and David C. Mabey. 2000. "Review of latrines, and basic pits with slabs--ensure more the Evidence Base for the `F' and `E' Components hygienic separation of excreta from the immedi- of the SAFE Strategy for Trachoma Control." Tropical Medicine and International Health 5 (8): ate living environment. Improved versions have 515­27. walls and a roof and may include a ventilation pipe or a cover plate for the squat hole. The col- Esrey, Steven A., James B. Potash, Leslie Roberts, lection chamber may vary from an unlined pit and Clive Shiff. 1991. "Effects of Improved Water Supply and Sanitation on Ascariasis, to a composting chamber. The superstructure Diarrhea, Dracunculiasis, Hookworm Infection, may be a crude shelter or an attractive brick Schistosomiasis and Trachoma." Bulletin of the or thatch construction with or without a vent World Health Organization 69 (5): 609­21. pipe and with or without a seat. SanPlat latrines Hutton, Guy. 2000. Considerations in Evaluat- are slightly elevated for ease of use in the dark. ing the Cost-Effectiveness of Environmental They can be located close to the house and have Health Interventions. Geneva: World Health a fitted lid to prevent odors and keep away flies. Organization. VIP latrines consist of the normal pit but are Hutton, Guy, U-Primo E. Rodriguez, Lydia fitted with a screened vent pipe. Napitupulu, Pham Ngoc Thang, and Phyrum 3. Hutton (2000) uses health care unit costs from Kov. 2007. Economic Impacts of Sanitation in the World Health Organization to estimate Southeast Asia: A Four-Country Study Conducted the cost of treating diarrhea, to which he adds in Cambodia, Indonesia, the Philippines and other expenses (such as transport) incurred by Vietnam under the Economics of Sanitation the patient. A number of assumptions are made Initiative (ESI). Jakarta, Indonesia: World Bank, regarding treatment (such as number of visits Water and Sanitation Program. or length of hospitalization). As a result, the Mehta, Meera, Thomas Fugelsnes, and Kameel mean cost per case of diarrhea for the patient Virjee. 2005. "Financing the Millennium is $10­$23, and the additional costs per visit Development Goals for Water Supply and 336 AFRICA'S INFRASTRUCTURE: A TIME FOR TRANSFORMATION Sanitation: What Will It Take?" Water Resources Countries. Joint Report of the African Ministers' Development 21 (2): 239­52. Council on Water, African Development Bank, Melo, Jose Carlos. 2005. The Experience of Condo- the European Union Water Initiative, the United minial Water and Sewerage Systems in Brazil: Case Nations Development Programme, the Water Studies from Brasilia, Salvador and Parauapebas. and Sanitation Program­Africa, and the World Lima, Peru: Water and Sanitation Program-- Bank. Nairobi, Kenya: Water and Sanitation Latin America, World Bank. Program­Africa. Morella, Elvira, Vivien Foster, and Sudeshna G. ------. 2008. "Can Africa Afford to Miss the Banerjee. 2008. "Climbing the Ladder: The State Sanitation MDG Target? A Review of the of Sanitation in Sub-Saharan Africa." Back- Sanitation and Hygiene Status in 32 Coun- ground Paper 13, Africa Infrastructure Country tries." Joint Initiative of the African Ministers' Diagnostic, World Bank, Washington, DC. Council on Water, African Development Bank, World Bank, and Water and Sanitation United Nations. 2008. The Millennium Program, Washington, DC. Development Goals Report 2008. New York: United Nations. World Bank. 2002. "Urban Environmental Strategic Sanitation Planning: Lessons from Water and Sanitation Program. 2006. Getting Bharatpur, Rajasthan, India." Field Note 23771. Africa on Track to Meet the MDGs on Water and Water and Sanitation Program­South Asia, Sanitation: A Status Overview of Sixteen African New Delhi. Index Boxes, figures, notes, and tables are indicated with b, f, n, or t following the page number. Abuja Treaty (1991), 154 Agence Française de Développement, 32 access Agence Malienne pour le Développement de l'Énergie to basic infrastructure services, 95b Domestique et l'Électrification Rurale to electricity, 52, 88, 88f, 89, 198, 199b (AMADER), 199b inequality of, 88­90 Agence pour la Sécurité de la Navigation Aérienne en Afrique recommendation for, 24 et à Madagascar (ASECNA), 268 to sanitation, 88, 88­89f, 323, 329 Agences d'Exécution des Travaux d'Intérèt Public (AGETIPs), to water resources, 129, 129­130f, 300­302, 301­302t, 214, 215b, 223 313­316, 315f, 315t, 316b agriculture Accra District Rehabilitation Project (Ghana), 139 economic development and, 125­126, 287, 288­289 ADF (African Development Fund), 158, 159 exports, 251 affermage arrangements, 209, 210n2, 310, 312b food price changes and, 293­294, 294t affordability irrigation. See irrigation of modern infrastructure services, 90­94 urbanization impact on, 126, 127t, 140n2 service expansion challenges, 98­100 AICD. See Africa Infrastructure Country Diagnostic Africa Energy Commission, 40 Air Afrique, 151, 260, 261, 267 Africa­EU Energy Partnership, 158 Air Gabon, 261 Africa Infrastructure Country Diagnostic (AICD), 31­42 airports and air transport, 259­270 background papers, 33t, 40 air traffic control infrastructure, 267, 269, 270n2 genesis of project, 32­33 existing infrastructure, 262­263, 265­267 GIS platform for, 35b financing, 133, 269 scope of project, 33­41 flight crew training, 269 country coverage of, 33, 34f market, 260­263, 261­262f, 268­269 existing spending patterns, 36­38 national flag carriers, 267 future spending needs estimates, 34­36 policy, 263­265, 264t sector performance, 38­41 pricing, 263, 263f working papers, 40, 41t private sector participation, 117t, 266 African Air Transport Policy, 264 reforms, 267­269 African Civil Aviation Commission, 264 regional connectivity and, 56, 144, 151, 152f African Conference on Sanitation and Hygiene (2008), 333 runway capacity and quality, 265­266, 266t African Development Bank, 32, 34, 40­41, 78, 148b, 158 safety issues, 263, 267­268, 268f African Development Fund (ADF), 158, 159 supply, 261­262, 265­266 African Economic Community, 154 terminal capacity, 266 African Union, 154, 158, 264 traffic trends, 260­261, 260f, 261t African Union Commission, 32, 155 AMADER (Agence Malienne pour le Développement AFRICATIP (Association Africaine des Agences d'Exécution des de l'Énergie Domestique et l'Électrification Travaux d'Intérèt Public), 215b, 223 Rurale), 199b 337 338 Index amodiation model, 253 sanitation, 95, 323, 325, 334b Angola telecommunications industry privatization, 179n1 non-OECD financing, 78, 79 transit corridor development, 207 power trade, 149, 150f undermaintenance, 73 railways, 232, 238 urban growth, 134 Apapa Container Terminal, Lagos, Nigeria, 257b urban infrastructure financing, 137 APM Terminals, 257b urban public transportation, 96b Arab Bank for Development in Africa, 277b water supply, 313, 315, 320 ASECNA (Agence pour la Sécurité de la Navigation Aérienne en Burundi Afrique et à Madagascar), 268 affordability issues, 92 Association Africaine des Agences d'Exécution des Travaux electricity outages, 182 d'Intérèt Public (AFRICATIP), 215b, 223 ODA, 79 Automated System for Customs Data (ASYCUDA), 208b power trade, 149, 150f water resource management, 280, 282 Backbone networks, 171­172 bus services, 96b, 220­221 development of high-bandwidth, 177 investment needs, 173­174 Cameroon backup generators, 5, 184, 184t, 197 affordability issues, 92 Banjul Accord Group, 260­261, 264t airports and air transport, 266 bank financing, 81­82 drinking water access, 300 Bardasi, Elena, 95b infrastructure as constraint on economic growth, 45 basic infrastructure services, 95b multimodal transport network, 204­205 benchmarking, 114­115 power outages, 184 Benin power sector underpricing, 191 governance, 108 power trade, 149, 150f infrastructure as constraint on economic growth, 45 railways, 231, 239, 243 road funding, 213 road network, 222 sanitation, 325 sanitation, 95 transport infrastructure, 150, 151t telecommunications sector inefficiency, 168 urban infrastructure financing, 137 transit corridor development, 207, 208b water MDG and, 305 transport infrastructure, 150, 151t water supply, 88, 315, 316 urban growth, 134 bond markets, 81­82, 85n8, 110, 200 urban public transportation, 96b boreholes, 299, 300­301, 301t, 303f, 303t water resources as human development constraint, 46, 305 Botswana water utilities, 310 infrastructure as constraint on economic growth, 45 capacity. See institutional competence and capacity railways, 230, 238 Cape Verde state-owned enterprises, 118, 118b power sector underpricing, 191 utility tariffs, 99 water tariffs, 308b water resource management, 281, 282 water utilities, 310 Botswana Power Corporation, 118b capital budget execution, 68­70 Bourse Régionale de Valeurs Mobilières (BRVM), 81 capital markets bribery, 205­206. See also corruption costs of capital, 82, 83f broadband service, 57, 171­173, 171­172t, 173f local, 81­82, 81t budget surveys, 39 power sector funding gap, 200 building regulations, 138 carbon emissions, 149 Burkina Faso catchments, 275 drinking water access, 300 CDMA (Code-Division Multiple Access), 168 informal settlements, 136 cell phones. See mobile phone coverage infrastructure as constraint on economic growth, 45 Center for International Earth Science Information Network, 35b power sector inefficiencies, 187 Central Africa. See also specific countries private sector participation, 112 airports and air transport, 263, 268, 268f property rights, 135 air transport hubs, 151, 260 railways, 204, 238 economic growth, 44 rural road transport, 220 freight services profit margins, 50 Index 339 freight tariffs, 220 Commission for Africa, 7, 31, 43, 44, 58, 77 ICT investment, 147, 148b, 148t Commission Internationale du Bassin Congo­Oubangui­ infrastructure, 48 Sangha, 204­205 multimodal transport network, 204­205 Common Market for Eastern and Southern Africa, 32, 156, roads, 211 159b, 264­265, 264t rural road accessibility, 222 commons approach to radio spectrum use, 177 Central African Power Pool, 52, 53­54t, 185, 187 community-driven development Central African Republic irrigation, 296, 296b airports and air transport, 261 sanitation improvement, 330, 331b drinking water access, 300 service expansion, 97­98 irrigation, 290, 290f water resource management, 279 multimodal transport network, 204­205 Comoros transit corridor development, 207, 208b drinking water access, 300 Central Africa Transport and Transit Facilitation ICT investments, 58 Project, 208b competition Central East African Railways, 204 mobile phone market, 113, 169, 170, 170f Chad ports and shipping, 256 airports and air transport, 261 railways, 239 broadband access, 173 Comprehensive Africa Agricultural Development Program, power sector underpricing, 191 288, 292 power trade, 149, 150f concessions private sector participation, 115 airports and air transport, 266 road maintenance, 214, 215, 216 port sytems, 253, 254t sanitation, 329 railways, 238­240, 242­243, 242f, 245 transit corridor development, 207, 208b water utilities, 310 water supply, 313 condominial sewerage, 333 children Congo. See Democratic Republic of Congo; Republic of Congo infant mortality, 45 connection charges, 90, 98­99, 98b, 308b piped water and, 94 connection subsidies, 98b water resources and, 45­46, 304­305 construction costs, 136, 141n9, 216, 225 China Cooperative Development of Operational Safety and Continuing financing from, 79, 79t, 80, 82, 83f, 85n3 Airworthiness Program, 268 infrastructure expansion, 48 corporatization of state-owned water utilities, 313 irrigation, 289 corridor associations, 151 land use planning, 139 corruption, 108, 151, 205­206, 207b manufacturing, 145 cost recovery national infrastructure programs, 146 improvement of, 70­71, 72t power sector investment by, 200 power sector, 191­192f, 191­194, 197­198 public sector expenditures, 67 railways, 245 urban infrastructure financing, 137 sanitation, 334 urban population density, 134 for subsistence consumption, 91­92, 91f China Export-Import Bank, 85n3 tariffs and, 99­100, 99f civil aviation authorities, 268, 269 water supply, 307­309, 307f, 308b, 319­320 Clean Development Mechanism, 149 Côte d'Ivoire climate change airports and air transport, 266 agriculture and, 287, 288 electricity access, 193 food security and, 293­294, 294t infrastructure as constraint on economic growth, 2 water resource management and, 53­54, 272, 279 irrigation, 290 coastal countries local capital markets and, 81 population density, 134­135, 135t municipal budgets, 134 urbanization, 127, 127t piped water access, 88 urban vs. rural income level differentials, 128, 128t power outages, 184 Code-Division Multiple Access (CDMA), 168 private sector participation, 112, 115 Commercial Reorientation of the Electricity Sector Toolkit railways, 204, 232, 233, 238 (CREST), 196b road funding, 213 340 Index sanitation, 325, 326, 327f, 329 disease transport infrastructure, 150, 151t sanitation access and, 329 utility connection charges, 90 water resources and, 45­46, 304­305 water supply, 300, 301, 302b, 307, 315, 319, 320 distribution losses water utilities, 310, 311, 313 power sector, 187 Country Policy and Institutional Performance Assessment, 51 utilities, 11, 11f, 66, 72­73, 73t country typology of AICD, 48, 51b, 66 water resources, 309 CREST (Commercial Reorientation of the Electricity Sector domestic finance sources, 76­77 Toolkit), 196b DP World, 252­253 drinking water. See water resources DAWASA lease contract, 116, 116b debt service expenditures, 76­77 East Africa. See also specific countries deficit-financed public investment, 77b airports and air transport, 268, 268f Democratic Republic of Congo hydroclimatic variability, 272, 273f affordability issues, 92 ICT investment, 4, 147, 148b, 148t backup generators, 184 ICT sector impact on, 168 broadband access, 173 infrastructure, 48 drinking water access, 300 maritime traffic growth, 250­251t, 251f external financing, 79 port system, 252, 255 hydropower projects, 53, 149, 187 railways, 204 infrastructure as constraint on economic growth, 2 water resource management, 277 local capital markets and, 81 The East Africa Marine System (TEAMS, Kenya), 148b, 148t mobile phone coverage, 172, 178 East African Community multimodal transport network, 204­205 AICD and, 32 paved road density, 96b airports and air transport, 264t ports and shipping, 254 Lake Victoria Basin Commission, 282 power outages, 184 Master Plan, 156 power trade, 149, 150f, 187, 200 regional integration, 156, 159b railways, 204, 230, 232, 238, 239 East African Power Pool, 52, 53, 53­54t, 185, 187 regional integration, 155 East Asia road standards, 217 basic infrastructure access, 88 spending estimates, 59 capital budgets for infrastructure, 77 standposts, 95 economic growth and infrastructure, 44­45 undermaintenance, 73 electricity access, 182 urban public transportation, 96b infrastructure, 47­48, 49f utility inefficiencies, 309 public sector expenditures, 67 water MDG and, 305 Eastern Africa Submarine Cable System (EASSy), 146, 148b, 148t water supply, 313, 314, 315, 316 Economic and Monetary Community of Central Africa demographic and health surveys (DHSs), 39, 324b AICD and, 32 demographic trends airports and air transport, 264­265, 264t basic infrastructure access and, 88 multimodal transport network, 204 spending needs estimates and, 36 Economic Community of West African States (ECOWAS), 32, urbanization and, 127­128 156, 157 density Regional Electricity Regulatory Authority, 160 infrastructure costs and, 130­132 economic development paved road, 96b agriculture and, 125­126, 288­289 railway traffic, 230­231 ICT sector impact, 43, 45f, 52t, 168 sewerage affordability and, 333 infrastructure linked to, 2, 44­47, 44t, 45­46f telecommunications networks, 60n2 irrigation, 52t transportation networks and, 56, 60n2, 129 power sector and, 45f, 52­53, 52t, 141n5, 183­186, 184f, 184t urban growth and, 4, 134, 135t, 141n3 regional integration and, 156­157 urban infrastructure needs and, 132, 133f transportation sector, 52t Department for International Development (UK), 32 urban areas as engine for, 126 Development Assistance Committee, 37 water resources and, 52t, 272­276, 276f, 277, 277b Index 341 education Export-Import Bank of China, 186 electricity access impact on, 46 external financing on hygiene, 330 funding gap, 66, 67 transport network impact on, 46 infrastructure investment, 8­9 water resources and, 304­305 power sector, 186, 186t, 200­201, 201f efficiency gap. See operating inefficiencies road network spending, 218 Egypt and urban infrastructure financing, 137 external governance, 110 electricity. See also power sector access to, 52, 88, 88f, 89, 198, 199b Fadama Water User Association (Nigeria), 296b human development constrained by access to, 46 farming. See agriculture institutional competence and capacity Federal Aviation Administration (U.S.), 268 private sector, 114­115 Federal Energy Regulatory Commission (U.S.), 121 Electricity Master Plan for Africa, 158 fiber-optic networks, 146, 148b, 171­172 Electricity Regulatory Board (Kenya), 189b finance sources Electric Power Act of 1997 (Kenya), 189b airports and air transport, 269 emergency power generation, 5, 184, 184t, 197 for funding gap, 75­82 energy industry. See power sector railways investment needs, 244­245 Energy Regulatory Commission (Kenya), 189b regional integration projects, 158­159 engineering cost studies, 34, 217 urban infrastructure, 133­138 equity of water tariffs, 308b water supply, 302­305 Eritrea financial markets airports and air transport, 261 as external governance, 110 drinking water access, 300 infrastructure financing, 81­82, 81t railways, 232 power sector financing, 200 ESKOM, 100, 110, 183­184b fiscal costs, defined, 33 Ethiopia flight crew training, 269 affordability issues, 92 flood mitigation, 280 airports and air transport, 261 flush toilets hydropower projects, 53, 149, 187 access to, 88, 88­89f informal settlements, 139 cost of, 94, 96f, 97t infrastructure as constraint on economic growth, 45 sanitation ladder, 321­327, 324b land use policies, 138 Food and Agriculture Organization, 35b, 40, 277 power trade, 149, 150f, 187, 200 food security, 274­275, 281, 287­288, 293­294, 294t railways, 232 fragile low-income states road funding, 213 access issues, 89­90 road maintenance, 214, 216 in AICD typology, 48, 51b sanitation, 323, 326, 327f, 329­330, 330b capital budget for infrastructure, 69t, 76, 76t spending estimates, 59 economic growth constrained by infrastructure, 45 traffic volumes, 212 economic return on infrastructure investment, 70, 71t urban development, 126­127 external financing and, 67 urban public transportation, 96b financing needed to meet water MDG, 302­305, 303t water resource management, 276, 282 funding gap, 8, 66­67, 67t, 75, 75t Ethiopian Airlines, 151, 260, 262 infrastructure, 2­3, 45, 48 EU­Africa Infrastructure Trust Fund, 158 institutional competence and capacity, 107 European Commission, 32, 78 local capital markets and, 81, 81t European Development Fund, 277b mobile phone coverage, 166 European Investment Bank, 148b non-OECD financing, 68f, 79t European Union ODA to, 68f, 78t on air safety in Africa, 268 power sector spending needs, 185, 185t, 186t energy partnership, 158 PPI investment, 68f, 80, 80t freight transit regulation, 206 public sector expenditures, 67, 68­69f, 70­71t ICT licensing framework, 176 road network spending, 218, 219t irrigation investment by, 295b spending allocation, 43, 59, 65 water resource management, 274, 275t urbanization, 127, 127t 342 Index utility water supply, 301­302, 302t port system, 250, 253 water resources funding gap, 302­305, 305t, 309, 309f power outages, 184 France and irrigation investment, 295b power sector inefficiencies, 187 France Telecom, 113 private sector participation, 112, 157 freedoms of the air, 269n1 property rights, 135 freight services road maintenance, 214 air transport, 259 road network, 224 railways, 231f, 235­236 sanitation, 95, 325 regional integration, 150­151, 151t telecommunications sector roads, 219­220, 220t, 224­225 inefficiency, 168 shipping, 249, 250 privatization, 171 tariffs, 5, 50, 235 urban growth, 134 French Development Agency, 148b utility inefficiencies, 309 fuel levies, 73, 221 utility payment collection, 100 funding gap, 65­86 water resources as human development constraint, 46, 305 annual, 75 water utilities, 311 capital budget execution, improvement of, 68­70 Ghana Airways Corporation, 261 cost recovery improvement, 70­71, 72t Ghana Gateway corridor, 204, 258 costs of capital, 82, 83f GIS (geographic information systems), 35, 35b deferral of infrastructure targets, 83, 84f Global Rural-Urban Mapping Project (GRUMP) efficiency reforms, 9­12, 11f, 73t, 74, 74t urban extents, 140n3 finance sources, 75­82 global systems mobile (GSM) networks, 167­168, 172, 175 deficit-financed public investment, 77b governance. See also institutional competence and capacity domestic, 76­77, 76t Infrastructure Institutional Scorecard, 106­107b local capital markets, 81­82, 81t institutional competence and capacity, 109­110, 109t, 110f non-OECD, 78­80, 79t power sector, 188, 190, 195 official development assistance, 77­78, 78t inefficiencies and, 188­189, 189f private investment, 80 railways, 243­245 lower-cost technologies and, 83­84 road conditions and, 217 power sector, 200­201 road freight transport, 225 recommendation for, 22­23, 25­26 state-owned enterprises and, 118, 119f, 120t spending allocation, 66­67, 67t, 70 water utilities, 313 user charges, 70­71 Government Financial Statistics (IMF), 36 utilities, operating inefficiencies of, 71­74, 73f, 73t gray water removal systems, 334b greenfield transactions Gabon port systems, 253, 254t affordability issues, 92 private investors, 80 non-OECD financing, 79 telecommunications, 112 power trade, 150f groundwater, 281, 320­321 private sector participation, 115 Group of Eight, 32 railways, 230, 231, 239 GRUMP (Global Rural-Urban Mapping Project) road funding, 213 urban extents, 140n3 water utilities, 310, 311 Guinea Gambia River Development Organization, 153, 282 access to basic infrastructure, 95b Gas and Electricity Markets, Office of (U.K.), 121 affordability issues, 92 geographic information systems (GIS), 35, 35b backup generators, 184 German Development Bank, 148b drinking water access, 300 Germany and irrigation investment, 295b external financing, 79 Ghana hydropower projects, 53 drinking water access, 300 non-OECD financing, 79 electricity access, 194 power sector governance, 196b, 197 housing shortage, 136 power trade, 149, 150f, 187, 200 informal settlements, 139 railways, 233 land use policies, 138 road network, 224 municipal budgets, 134 sanitation, 325 Index 343 urban public transportation, 96b urban infrastructure financing, 137 water utilities, 310, 311 urban population density, 134 Guinea-Bissau Indonesia, capital budgets for infrastructure, 77 affordability issues, 92 infant mortality, 45 power trade, 149, 150f informal enterprises, 184 Gulf States, financing from, 79, 79t, 80, 82, 83f informal settlements electrical access and, 89 health urban areas, 136­137, 139 electricity access and, 46 information and communications technologies (ICTs), 165­179. informal settlements and, 137 See also specific subsectors sanitation and, 329, 335n3 backbone networks, 171­172, 173­174, 177 transport network impact on, 46 cost recovery, 5 water resources and, 45­46, 94, 304­305 developments, 167­168 Herfindahl-Hirschman Index, 113, 122n2 economic impact of, 43, 168 housing and urban planning, 134­136 economies of scale and, 144 hub-and-feeder port system, 252­253 financing, 10, 133 human development human development and, 47 informal settlements and, 137 infrastructure as constraint on economic growth, 45 infrastructure as constraint on, 45­47, 52t institutional competence and capacity, 108, 109f, 110f recommendation for, 23­24 international connectivity, 171, 176 sanitation access and, 329 Internet access, 167, 171­172t, 172­173, 173f water resources and, 274­275, 304­305 investment needs, 172­174 Hutchison Port Holdings, 252 licensing framework, 175­176 Hutton, Guy, 335n3 local capital markets and, 81 hybrid power markets, 194­195 market reforms, 169­170, 169­170f hydroclimatic variability, 272, 273f non-OECD financing, 68f risk mitigation, 277, 279 ODA for, 68f hydropower operating inefficiencies, 72, 74t, 168, 169f cost, 53 policy challenges, 174­178 economies of scale and, 144 ports and shipping needs, 258 local-scale units, 279 PPI investment, 67, 68f non-OECD financing, 78 private sector participation, 57­58, 80, 111­113, 111f, 113f, regional integration, 149, 153, 187 116, 117t, 146, 148b, 168 water resource management, 275, 275f, 280, 281t public expenditures on infrastructure, 68f, 69f, 70­71t hygiene education, 330 radio spectrum management reforms, 177­178 reform agenda, 168­172, 174­175 IBNET (International Benchmarking Network), 39 regional integration, 143, 146­148, 147f, 148b, 148t IBTs (increasing block tariffs), 101, 308b regulatory reforms, 170 ICTs. See information and communications technologies sector performance data collection, 39 IDA. See International Development Association spending needs, 6­7, 36, 56­58, 58t IMF. See International Monetary Fund state-owned enterprise reforms, 170­171, 176 improved latrines, 94­95, 96f, 97t, 324­327, 324b, 328f universal access, 178 Incomati Basin water-sharing agreement, 155 urban vs. rural access, 129, 129f, 130f income levels, urban vs. rural, 128, 128t, 140n1 voice services, 166­167, 166f, 171­172t, 172, 173f increasing block tariffs (IBTs), 101, 308b infrastructure, 43­63. See also specific types independent power producers (IPPs), 114­115, 186, 194 current deficit, 2­3, 3t, 47­49, 48t, 49f, 50t independent regulators, 120­121, 121b, 121t defined, 32 India as key to growth, 44­47, 44t, 45­46f financing from, 79, 79t, 80, 82, 83f price premium of, 4­5, 5t, 49­52, 50t infrastructure expansion, 48 regional integration, 156­158 irrigation, 289 spending needs, 7, 7t, 22­23, 52­58, 52t, 58t, 59f manufacturing, 145 Commission for Africa estimate, 58 mobile phone costs, 167 country distribution, 58­59 national infrastructure programs, 146 ICT, 56­58, 58t power sector investment by, 200 investment vs. maintenance, 60 344 Index irrigated areas expansion, 54, 55t IPPs. See independent power producers power sector, 52­53, 53­54t irrigation, 287­297 transport network, 55­56, 57t challenges, 294­296 water and sanitation, 54­55, 55t community-driven development, 296, 296b water security, 53­54 current state of, 289­290, 289t, 290f urban areas, 128­130 economic impact of, 288­289 costs of, 130­132 impact of, 293­294 financing, 133­138 institutional reforms, 295, 295b Infrastructure Consortium for Africa, 32, 78, 158, 160n3 investment needs, 36, 54, 55t, 290­293, 291b, 291t, Infrastructure Institutional Scorecard, 106­107b, 108­109f, 109t 292f, 293t inland waterways, 204 operating inefficiencies, 74t institutional competence and capacity, 105­123. See also specific project analysis, 295­296 infrastructure sectors public expenditures on infrastructure, 70­71t independent regulators, 120­121, 121b, 121t sector performance data collection, 40 private sector participation, 110­117, 111f, 111t, 116b, 117t strategic planning, 294­295 recommendation for, 16­19 underdevelopment of, 274, 276f, 289­290 reform efforts, 12­14, 13f, 106­110 water resource management and, 280 regional integration, 143, 155­156 road infrastructure, 214f Japan, ODA from, 78 scorecard, 106­107b, 108­110f, 109t Joint Monitoring Programme (WHO/UNICEF), 300, 321n2, state-owned enterprises, 117­120, 118b, 119b, 190b 323, 324b, 334­335 water and sanitation, 274 Intergovernmental Panel on Climate Change, 272 Kager Basin Organization, 282 International Air Transport Association, 268, 269 Kano­Katsina­Maharadi initiative, 157 International Benchmarking Network (IBNET), 39 Kenya International Civil Aviation Organization, 268, 269 airports and air transport, 261, 266, 267 International Container Terminal Services, 252 external governance and, 110 International Convention for the Safety of Life at Sea, 255 fiber-optic cable network, 171­172 International Development Association (IDA), 78, 82, 83f, governance, 108 158, 159 hydrologic variability, 273 International Energy Agency, 40 informal settlements, 136­137, 139 International Finance Corporation, 148b irrigation, 290, 290f International Food Policy Research Institute, 35b, 40 multimodal transport network, 204 international ICT connectivity, 171, 176 power sector, 189b International Monetary Fund (IMF), 36, 37, 77b, 185 power trade, 150f international phone service, 167, 167f, 171t PPI investments, 79 International Ship and Port Facility Security Code, 255 private sector participation, 113, 189b international shipping market, 255­256 railways, 244 International Telecommunication Union, 39 sanitation, 325, 329 International Union for Conservation of Nature, 35b state-owned enterprises, 118 Internet access submarine cable project, 148b ICT sector and, 165, 167f telecommunications, 170, 179n1 investment needs, 57, 172­173, 172t transport infrastructure, 150, 151t price premium on, 50­52 urban infrastructure, 96b, 137 pricing for access, 167, 167f, 171, 171t water MDG and, 305 regional cooperation opportunities, 146­148 water resource management, 275, 276 Interstate Aviation Committee, 263 water supply, 313, 315 intraregional trade, 152­153 Kenya Airways, 151, 260, 262 investment climate surveys, 45 Kenya Electricity Generating Company, 189b investment spending Kenya Power and Lighting Company, 188, 189b irrigation needs, 290­293, 291b, 291t, 292f, 293t kerosene lamps, 94, 97t maintenance vs., 8, 8f, 60 Korea. See Republic of Korea power sector backlog, 185­186 Kreditanstalt für Wiederaufbau (Germany), 32 railways, 236­238 Krugman, Paul, 144 water security needs, 279­283 KwaZuluNatal program, 226 Index 345 labor inefficiencies low-income states. See fragile low-income states; nonfragile power sector, 188 low-income states railways, 240, 241f, 242, 242f Lusaka Water and Sewerage Company, 306 telecommunications, 168, 169f utilities, 11, 11f, 72, 73t Madagascar water utilities, 309, 312t airports and air transport, 266 ladder of sanitation, 324­325, 325f broadband access, 173 Lake Chad Basin Development Fund, 282 ICT investments, 58 Lake Victoria, 279, 282 infrastructure as constraint on economic growth, 45 Lake Victoria Basin Commission, 282 irrigation, 289, 292 land leases, 137 mobile phone coverage, 172 landline telephones, 88, 88­89f, 146, 165, 166­167, 168f railways, 240 landlocked countries road network, 212, 222 construction costs, 136 sanitation, 95, 323, 325, 326, 327f, 329 ICT access, 176 spending estimates, 59 outlets to shipping, 251 water MDG and, 305 population density, 134­135, 135t water supply, 314 regional integration, 143, 144, 150 water supply pricing, 307 transit corridor development for, 206­207 maintenance urbanization, 127, 127t backlog, 10, 10f urban vs. rural income level differentials, 128, 128t economic impact of, 73, 85n1 landlord port system, 250, 253, 254t, 257 investment vs., 8, 8f, 60 land ownership and regulations, 134­136, 136b, recommendation for, 16 138­139, 140 road infrastructure, 56, 70, 213­214, 215­217, 216f large-scale irrigation, 290­292, 291b, 291t, 293t sanitation, 329 Latin America utilities, operating inefficiencies, 73­74, 73f basic infrastructure access, 88 water supply, 314­315, 316b capital budgets for infrastructure, 77, 85n2 water utilities, 309 condominial sewerage, 333 Malawi electricity access, 182 affordability issues, 92 telecommunications sector performance, 168 airports and air transport, 266 lease contracts, 115­116, 116b, 310, 311, 317 drinking water access, 300 Lesotho hydroclimatic variability, 272, 273f airports and air transport, 261 infrastructure as constraint on economic growth, 45 drinking water access, 301 mobile phone coverage, 178 governance, 108 power sector underpricing, 191 power trade, 150f railways, 204, 235, 239 sanitation, 325, 329, 332 road network, 212 utility payment collection, 100 sanitation, 325 water resource management, 276, 277b traffic volumes, 212 water supply, 313 utility payment collection, 100 water utilities, 313 utility tariffs, 99, 99f Lesotho Highlands Water Project, 155 water resources, 275 Lesotho Lowlands Water Supply Scheme, 277b water supply, 314, 315, 316, 320 Liberia and power trade, 149, 150f Mali licensing framework for ICT sector, 112, 175­176 affordability issues, 92 Lighting Africa initiative, 198 drinking water access, 300 literacy and electricity access, 46 independent regulators, 120 local capital markets, 81­82, 81t informal settlements, 136 as external governance, 110 infrastructure as constraint on economic growth, 45 power sector funding gap, 200 irrigation, 295b logistics systems, 207b non-OECD financing, 79 for transport network, 205­206, 205f ODA, 79 lower-cost technologies as funding gap solution, power sector governance, 197 83­84 power trade, 150f 346 Index property rights, 135 water resources funding gap, 302­305, 305t railways, 204 migration to urban areas, 126, 127 rural electrification, 199b Millennium Development Goals (MDGs) water resource management, 282 funding gap and, 75 water supply, 320 infrastructure linked to, 45, 47t water utilities, 310 poverty reduction, 287 Malian Agency for the Development of Domestic Energy and sanitation, 323, 329, 334­335 Rural Electrification, 199b water resource management, 43, 54­55, 58, 84, 279, 281 malnutrition, 45, 94, 294, 294t water supply, 299, 300, 302­305, 303t, 321n1 management contracts Minerals, Energy, and Water Affairs Ministry (Botswana), 190b institutional competence and capacity, 112­116, 119b, 122n3 minerals exports, 251, 254 port systems, 253, 254t minibus services, 96b, 220­221 power sector, 188­190, 189f, 196 mobile phone coverage, 56­57, 166­167, 166f water utilities, 311 access to, 88, 89f manufacturing sector and regional integration, 144­145 economic impact of, 168, 169f Maputo Development Corridor, 157, 204, 207b, 258 market competition, 170, 170f maritime traffic trends, 250­251 private sector participation, 113 Mauritania modern infrastructure services airports and air transport, 261 affordability of, 90­94 backup generators, 184 alternatives to, 94­95 drinking water access, 300 inequitable access to, 88­90 multimodal transport network, 208 Morocco ports and shipping, 253 non-OECD financing, 79 mortgage lending, 141n10 urban infrastructure financing, 137 motorcycles, 220 water resources as human development constraint, 46, 305 Mozambique Mauritius broadband access, 173 infrastructure as constraint on economic growth, 2, 45 drinking water access, 301 irrigation, 290, 290f hydrologic variability, 273 ports and shipping, 253 informal settlements, 139 MDGs. See Millennium Development Goals land use policies, 138 meteorological services, 275­276, 279 ports and shipping, 254 Metolong Dam and Water Supply Program, 277, 277b power trade, 149, 150f middle-income countries private sector participation, 112 access issues, 89­90 railways, 204, 232, 239, 244 in AICD typology, 51b road network, 212 capital budget for infrastructure, 76t, 216 standposts, 95 economic return on infrastructure investment, 70, 71t transit traffic growth, 207b financing needed to meet water MDG, 302­305, 303t transport infrastructure, 150 funding gap, 8, 66­67, 67t, 75, 75t urban planning, 138 infrastructure, 2­3, 48 utility connection charges, 90 institutional competence and capacity, 107, 108f water resource management, 276, 280 local capital markets and, 81, 81t water supply, 313, 315, 320 mobile phone coverage, 166 water utilities, 310, 311 non-OECD financing, 68f, 79t multi-indicator cluster surveys, 39 ODA to, 68f, 78t multimodal transport, 204­205 population density, 134­135, 135t Multiple Indicator Cluster Surveys, 324b power sector spending needs, 185, 185t, 186t municipal budgets, 133­137, 134f PPI investment, 67, 68f, 80, 80t public sector expenditures, 66­68, 68­69f, 69t, 70­71t, 76 Nairobi Stock Exchange, 110 road network spending, 218, 219t Namibia spending allocation, 43, 58­59, 65 broadband access, 173 urbanization, 127, 127t drinking water access, 300 urban vs. rural income level differentials, 128, 128t irrigation, 292 utility water supply, 301­302, 302t ports and shipping, 254 water resources cost recovery, 309 power trade, 150f Index 347 railways, 238 private sector participation, 113, 114b, 115, 157, 257b sanitation, 325, 329 railways, 233 telecommunications sector inefficiency, 168 road network, 213, 216, 222 utility payment collection, 100 telecommunications industry privatization, 170 utility tariffs, 99 traffic volumes, 212 water supply, 315 transport infrastructure, 151t water tariffs, 308b undermaintenance, 73 national flag carriers, 267 urban planning, 138 National Freight Logistics Study (South Africa Department of water supply, 314, 315, 316b, 320 Transport), 209 Nigeria Airways, 261 National Geospatial-Intelligence Agency (U.S.), 35b Nigerian Ports Authority, 257b National Ports Authority (South Africa), 256 Niger River basin, 272, 279 National Sanitation Program (Lesotho), 332 Niger River Commission, 153, 282 National Transportation Safety Board (NTSB, U.S.), 268 Nile Basin Initiative, 282, 283f National Transport Safety Commission (Ghana), 226 Nile River basin, 272, 279 National Water and Sewerage Commission (NWSC, nonfragile low-income states Uganda), 119b access issues, 89­90 negative concessions, 209, 210n2 in AICD typology, 48, 51b neighborhood effect of urban areas, 126 capital budget for infrastructure, 69t, 76, 76t Netherlands and irrigation investment, 295b economic return on infrastructure investment, 70, 71t New Partnership for Africa's Development (NEPAD), 32, 154, financing needed to meet water MDG, 302­305, 303t 288, 292 funding gap, 8, 66­67, 67t, 75, 75t Niger infrastructure, 2­3, 45, 48 affordability issues, 92 institutional competence and capacity, 107 airports and air transport, 261 local capital markets and, 81t broadband access, 173 mobile phone coverage, 166 connection subsidies, 90, 98b non-OECD financing, 68f, 79t governance, 108 ODA to, 67, 68f, 78t irrigation, 290, 292 power sector spending needs, 185, 185t, 186, 186t ODA, 79 PPI investment, 68f, 80, 80t power sector public sector expenditures, 67, 68f, 69f, 70­71t inefficiencies, 187 road network spending needs, 219t trade, 149, 150f spending estimates, 43, 58­59 underpricing, 191 urbanization, 127, 127t private sector participation, 112, 115 utility water supply, 301­302, 302t road network, 212, 216 water resources cost recovery, 309, 309f spending estimates, 59 water resources funding gap, 302­305, 305t utility tariffs, 90, 99, 99f non-OECD finance sources water resources as human development constraint, 46, 305 cost of capital, 82, 83f water supply, 315 funding gap, 66, 67, 76, 78­80, 79t water utilities, 310, 311 infrastructure investment, 8­9, 9t Niger Basin Authority, 153, 282, 283, 283f power sector, 186, 200, 201f Nigeria water resources, 304 airports and air transport, 260 North America's water resource management, 274, 275t capital budget for infrastructure, 76 drinking water access, 300 Oak Ridge National Laboratory, 35b electricity access, 194 Obasanjo, Olusegun, 154 fiber-optic cable network, 171­172 OECD countries informal settlements, 139 telecommunications sector performance, 168 irrigation, 290, 290f, 292, 296b water tariffs, 307 land use policies, 138 off-grid power technologies, 182 non-OECD financing, 78, 79 Office du Niger, 295b ports and shipping, 250, 253, 257b official development assistance (ODA) power sector governance, 196b cost of capital, 82, 83f PPI investments, 79 funding gap, 66, 67, 75­78, 78t 348 Index infrastructure investment, 8­9, 9t infrastructure as constraint on economic growth, 45 power sector, 200, 201f institutional competence and capacity, 108, 109f, 110f road network spending, 218, 219t international shipping market and, 255­256 water resources, 304, 304t maritime traffic trends, 250­251, 250­251t, 251f open defecation, 323­325, 328f, 330 market, 250­251 operating inefficiencies modal integration policies, 258 power sector, 181, 186, 187, 188f in multimodal transport network, 204 railways, 240 ownership and management, 253 recommendation for, 15­16 performance, 254­255, 254f, 255t state-owned enterprises, 118 policy issues, 255­258 utilities, 9­13, 11­12f, 12t, 71­74, 73t pricing, 256­257 water utilities, 309­313, 310­311f, 312t private sector participation, 111­113, 111f, 114b, 114f, 116, optical-fiber technology, 146, 148b, 171­172 117t, 157, 257­258, 257b Organisation for Economic Co-operation and Development regional connectivity and, 56, 144, 150, 152 (OECD). See also official development assistance regulatory framework, 256­257 (ODA) strategic planning, 256 AICD spending estimates and, 37 poverty reduction, 87­103 Principles of Corporate Governance for SOEs, 119 agriculture and, 125­126, 287, 288­289, 296, 296b Organization of African Unity, 237 basic infrastructure access, 95b Our Common Interest (Commission for Africa), 31 economic growth and, 44 overemployment modern infrastructure services power sector, 188 affordability of, 90­94, 91t telecommunications, 168, 169f alternatives to, 94­95 utilities, 11, 11f, 66, 72, 73t inequitable access to, 88­89f, 88­90 water utilities, 309, 312t recommendation for, 23­24 regional integration and, 156 Pakistan and mobile phone costs, 167 service expansion challenges, 97­102 Pan-African Infrastructure Development Fund, 85n7 affordability analysis, 98­100 Pan-African Road Safety Conference (2007), 226 demand side, 90t, 97­98 Partnership for Agricultural Water in Africa, 289 second-best solutions for, 87, 94­95, 96f, 96t, 102 passenger services, 229, 231f subsidies, 98b, 100­102 railways, 234­235, 243­244 urban vs. rural income level differentials, 128, 128t paved road density, 96b water resources and, 274, 277, 278f performance contracts, 118, 119b, 120f, 122n3 Poverty Reduction Strategy Papers, 295 power sector, 196 power sector, 181­202 road maintenance, 214 capacity issues, 5, 6f, 182, 183­184b, 183f water utilities, 313, 314b cost recovery, 50, 191­193f, 191­194, 193t, 197­198 Peru and joint household utility access, 130 economic impact of, 45, 45f, 52­53, 53­54t, 141n5 Pier One scheme, 253 economies of scale and, 144 piped water electricity grid access, 198, 199b access to, 88­89, 88­89f, 96f financing, 133 costs of, 95, 97t funding gap, 66, 75, 75t, 186, 186t, 200­201 human development benefits of, 94 independent regulators and, 121t urban vs. rural access to, 300­301, 301t, 303f, 303t infrastructure as constraint on economic growth, 45 policy challenges institutional competence and capacity, 107, 108, 109, airports and air transport, 263­265, 264t, 267­269 109f, 110f power sector, 194­201 institutional reforms, 187­190 water supply, 319­320 investment backlog, 185­186 political consensus for regional integration, 154­155 non-OECD financing for, 67, 68f, 78 polity executive constraint scores, 108 ODA for, 68f ports and shipping, 249­258 operating inefficiencies, 10­12, 11­12f, 12t, 66, 72, 73, 181, configuration of, 252­253, 252f 186, 188f financing, 133 performance data collection, 40 ICT needs, 258 planning, 194­195 independent regulators and, 120 policy challenges, 194­201 Index 349 PPI investment, 68f, 80 water utilities, 312b pricing, 4­5, 5t, 191­194, 191­192f public service port model, 253 private sector participation, 111­112, 111f, 115f, 116, 117t, 189b radio spectrum management reforms, 177­178 public expenditures on infrastructure, 68­69f, 70­71t railways, 229­247 regional integration, 143, 147f, 148­150, 150f, 187, 198­200 concessions, 238­240, 242­243, 242f, 245 spending needs, 36, 43, 52­53, 58, 185­186, 185t condition of, 232­233 state-owned enterprise reforms, 120f, 120t, 195­197 cost recovery, 245 subsidies, 92­93, 93­94f density of, 230­231 tariffs, 92­93 financial performance, 240­242 urban vs. rural access, 129, 129f, 130f financing, 133 water resource management and, 280 freight services, 235­236 PPI. See private participation in infrastructure government role, 243­245 prepaid telephone service, 166 independent regulators, 120 prepayment meters, 100, 100f institutional competence and capacity, 109f, 110f, 238­243 Prevention, Public Hygiene and Sanitation Ministry private sector, 113­114 (Senegal), 334 intraregional trade and, 152 pricing investment needs and justification, 236­238 airports and air transport, 263, 263f market for, 233­236 infrastructure cost premium, 4­5, 5t, 49­52, 50t in multimodal transport network, 204 ports and shipping, 256­257 networks, 230­233, 230­231f power sector, 4­5, 5t, 191­192f, 191­194 non-OECD financing for, 67 water resources, 299, 307­309, 307f operational performance, 240 primary school completion and electricity access, 46 passenger services, 234­235, 243­244 Principles of Corporate Governance for SOEs (OECD), 119 private sector participation, 111­112, 111f, 115f, 117t, principles of efficient urbanization, 139­140 239­240, 239f, 244­245 private participation in infrastructure (PPI), 66­67, 76, 79­80, regional connectivity and, 56 80t, 304, 304t regulatory framework, 245 Private Participation in Infrastructure database, 37 traffic private sector participation (PSP). See also specific infrastructure density, 230­231, 231­232f, 246n2 sectors trends, 233­234, 234f electricity access, economic impact of, 182 rainfall. See hydroclimatic variability institutional competence and capacity, 12­14, 14t, 110­117, rainwater harvesting, 294 111f, 111t, 113­115f, 116b reform. See specific infrastructure sectors power sector, 189b Regional Indicative Strategic Plan (Southern African railways, 239­240, 239f Development Community), 156 regional integration, 157­158 regional integration, 143­161 water utilities, 310­313, 311f, 312b, 312t ICT sector, 146­148, 147f, 148b, 148t, 174 privatization, 171, 179n1 importance of, 144­146 Program for Infrastructure Development in Africa, 157 infrastructure priorities, 156­158 property rights, 134­136, 136b, 138­139, 140 economic returns, 156­157 PSP. See private sector participation private participation, 157­158 public expenditures on infrastructure, 8­9, 9f, 9t spatial targeting, 157 documentation of, 37­38 institutional capacity, 155­156 funding gap, 66­67, 67­69f, 69­71t opportunities for, 146­154 road network spending, 218, 219t political consensus for, 154­155 sanitation, 329 power sector, 53, 148­150, 150f, 187, 198­200 water resources, 304, 304t project preparation and finance, 158­159 Public-Private Infrastructure Advisory Facility, 32 recommendation for, 20­22 public-private partnerships. See also private participation regulatory frameworks for, 159­160 in infrastructure (PPI); private sector transport and trade networks, 56, 150­153, 151t, 152f, 159b participation (PSP) water resources, 153­154, 279, 282­283, 283f agricultural water management, 295 regulatory framework hydropower projects, 281 benchmarking and, 114­115 ports and shipping, 253 ICT sector reforms, 170 350 Index on Infrastructure Institutional Scorecard, 106­107b private sector participation, 116, 117t institutional competence and capacity, 109, 110f public expenditures on infrastructure, 70­71t, 218­219, 219t market reforms and, 169, 170f reform agenda, 221­226 ports and shipping, 256­257 regional integration, 147f power sector inefficiencies and, 188­189, 189f, 195 rural transport and, 220, 222­223 property rights, 134­136, 136b, 138­139, 140 safety improvements, 221, 226 radio spectrum, 178 sector performance data collection, 39 railways, 245 spending allocation, 215­217, 216f recommendation for, 19­20 spending needs, 36, 217­219, 219t regional integration, 144, 159­160 transport services and, 219­221 road freight transport, 224­225 urban transport and, 220­221, 223­224 water utilities, 313 RONET (road sector modeling tool), 39 Republic of Congo runway capacity and quality, 265­266, 266t multimodal transport network, 204­205 rural areas railways, 232 electrification, 198, 199b regional integration, 155 financing, 138 sanitation, 325 ICT sector impact on, 47 Republic of Korea infrastructure, 2­3, 3f, 128­130 infrastructure as constraint on economic growth, 2 land use planning, 139 urban population density, 134 mobile phone coverage, 166, 166f reservoir water storage, 53­54, 58, 277, 279­282, 280f roads, 211, 212, 220, 222­223 resource-rich countries sanitation, 325­326, 326f, 326t, 327, 328f in AICD typology, 48, 51b transportation networks and, 46, 56 capital budget for infrastructure, 68­69, 69t, 76, 76t urban links to, 126­130 economic growth, 44 water resource management, 46, 277, 316b economic return on infrastructure investment, 70, 71t water resources as human development constraint, 305 financing needed to meet water MDG, 302­305, 303t water supply access, 300­302, 313­316, 315f, 315t funding gap, 8, 66­67, 67t, 75, 75t, 83 Rural Development Agency (Nigeria), 316b infrastructure, 48 Rural Reform Index for water, 315 institutional competence and capacity, 107, 108f Rwanda local capital markets and, 81t agricultural employment, 274 non-OECD financing, 67, 68f, 78­79, 79t, 85n3 ODA, 79 ODA to, 68f, 78t paved road density, 96b power sector spending needs, 185, 185t, 186t power sector underpricing, 191 PPI investment, 67, 68f, 80, 80t private sector participation, 112 public sector expenditures, 67, 68f, 69f, 70­71t road funding, 213 road network spending needs, 219t road maintenance, 215 spending estimates, 43, 58­59 sanitation, 95, 323, 326, 327f, 329 water resources funding gap, 302­305, 305t telecommunications sector privatization, 171 risk management undermaintenance, 73 hydroclimatic variability, 277 utility payment collection, 100 river basins water resource management, 280, 282, 306, 306b transboundary nature of, 153­154, 272­273, 274f, 279 water supply, 313, 319, 320 roads, 211­227 alternative technologies for, 84 Safe Skies for Africa program, 268 condition of, 217, 218f safety costs of, 225­226 airports and air transport, 263, 267­268, 268f economic impact of, 45, 45f ports and shipping, 255 freight services and, 219­220, 220t, 224­225 railways, 240 funding arrangements for, 213 road improvements, 221, 226 infrastructure, 212 sanitation, 323­336 institutional capacity and, 213­215, 214f, 215b, 221­222 access to, 88, 88­89f, 323, 329 intraregional trade and, 152 capital budget for, 68 maintenance, 56, 70, 74, 215­217, 216f challenges, 329­333 ODA and, 67, 218, 219t cost recovery, 329, 333, 334, 334b Index 351 defining improvement, 324b demand side, 97­98 demand stimulation for, 330­331, 331b second-best solutions for, 102 evaluation of, 324­329 subsidies, 98b, 100­102 financing, 75, 75t, 333­334 sewerage system, 325 human development constrained by, 45­46 Seychelles institutional competence and capacity, 328­329 airports and air transport, 261 ladder, 324­325, 325f ICT investments, 58 MDG for, 54­55, 323, 329, 334­335 shipping. See ports and shipping non-OECD financing, 67, 68f Sierra Leone ODA for, 67, 68f access to basic infrastructure, 95b operating inefficiencies, 74t power trade, 150f patterns of access, 325­327, 326f, 326t, 328f Sikasso­Korhogo­Bamako initiative, 157 PPI investment, 67, 68f slums. See informal settlements progress measurement, 334­335 small-scale irrigation, 290­292, 291b, 291t, 293t public expenditures on infrastructure, 67, 68f, 70­71t, 329 Société de Distribution d'Eau de la Côte d'Ivoire sewerage affordability, 333 (SODECI), 302b sharing of facilities, 325, 325f SODECI (Société de Distribution d'Eau de la Côte spending needs, 36, 55t, 58 d'Ivoire), 302b supply adequacy, 331­333 SOEs. See state-owned enterprises urban vs. rural access, 129, 129f, 130f solar power, 182, 198 SanPlat latrines, 324, 335n2 South Africa scale economies, 143, 144, 148­150 agricultural employment, 274 SEACOM (South Africa­East Africa­South Asia­Fiber Optic airports and air transport, 260, 261, 266, 268­269 Cable), 148b drinking water access, 300 seaports. See ports and shipping governance, 110 sector performance. See also specific sectors irrigation, 289, 292 AICD project scope, 38­41 maritime traffic growth, 250, 250­251t, 251f defined, 33 multimodal transport network, 209 Senegal municipal budgets, 133 affordability issues, 92 ports and shipping, 253, 254, 256 deficit-financed public investment, 77b power supply issues, 183­184b drinking water access, 300 power trade, 150f electricity access, 194 private sector participation, 112 electricity outages, 182 railways, 230, 231, 233, 234, 238, 244 independent regulators, 120, 121b regional integration, 157 informal settlements, 136 road network, 213, 217 infrastructure as constraint on economic growth, 2, 45 sanitation, 325 irrigation, 290, 290f spending estimates, 59 municipal budgets, 134 transport infrastructure, 150, 152 non-OECD financing, 79 undermaintenance, 73 power sector, 149, 150f, 184, 191 utility charges, 90, 99, 100 private sector participation, 112, 115 water resource management, 282 railways, 204, 234 water tariffs, 308b road network, 213, 215, 216 water utilities, 310 sanitation, 323, 325, 326, 327f, 331­333, 332t South Africa­East Africa­South Asia­Fiber Optic Cable transport infrastructure, 150, 151t (SEACOM), 148b urban planning, 138 South African Airways, 151, 260, 262, 269 water supply, 88, 307, 312b, 315, 319 South African Port Operations, 256 water utilities, 310, 311 South Asia Senegal River basin, 272, 279 basic infrastructure access, 88 Senegal River Basin Organization, 282, 283f drinking water access, 300 Senegal River Development Organization, 153 electricity access, 182 septic tanks, 323­325 infrastructure, 2­3, 47­48, 49f service expansion challenges, 97­102 water tariffs, 307 affordability analysis, 98­100 South Atlantic 3/West Africa Submarine Cable (SAT-3), 146, 176 352 Index Southern Africa street lighting, 94­95, 97t airports and air transport, 268, 268f submarine communication cables, 146, 148b, 149t, 174 water resource management, 277 Sub-Saharan Africa. See also specific countries Southern African Development Community agricultural growth, 288 AICD and, 32 AICD focus on, 33, 34f airports and air transport, 264­265, 264t, 269 airports and air traffic, 261­262, 262f infrastructure, 48 drinking water access, 300 regional integration, 156, 159b economic growth, 76 Southern African Power Pool, 52, 53­54t, 150, 185, 187, 190b economic return on infrastructure investment, 70, 71t Southern Nations Regional Health Bureau, 330, 331b electricity access, 182 spatial modeling geography as infrastructure challenge, 3­4 regional integration, 157 ICT investment, 147, 148b, 148t spending needs estimates and, 35, 35b infrastructure financing needs, 31­32 urban areas, 134­136 irrigation, 40, 289­290, 289t, 296 spending needs logistics system needs, 205­206 AICD estimates, 34­36 population density, 134­135, 135t Commission for Africa estimate, 58 water resource management, 272­275, 275t country distribution, 58­59 Sub-Saharan Africa Transport Policy Program (SSATP), 39 defined, 33 subsidies ICT, 56­58 power sector, 192­193, 197 infrastructure, 52­58 service expansion, 98b, 100­102 investment vs. maintenance, 60 utility connections, 87, 98­99, 98b, 100­101, 101t irrigated areas expansion, 54 of water and power tariffs, 92­93, 93­94f power sector, 52­53 Sudan roads, 217­219 drinking water access, 300 transport network, 55­56 hydroclimatic variability, 272, 273f urban areas, 132­133 irrigation, 289, 290 water and sanitation, 54­55 non-OECD financing, 78, 79 water security, 53­54 power trade, 149, 150f Spoornet (South Africa), 209, 233 PPI investment, 80 standposts private sector participation, 112 as second-best solution, 94­95, 96f, 97t, 102 telecommunications industry privatization, 170 as water supply, 299, 300­301, 301t, 303f, 303t, 305­306, 306b, utility connection charges, 90 318­319 surface water, 300, 301, 313 state-owned enterprises (SOEs) urban vs. rural access to, 301t, 303f, 303t airports and air transport, 260, 267 Swaziland, railways in, 231, 233, 238 capital budget for infrastructure, 68­69 corporatization of, 313 Tanzania ICT sector reforms, 170­171, 176 affordability issues, 92 institutional competence and capacity, 117­120, 119f airports and air transport, 262­263, 266, 267 lessons from, 118b, 119b budget allocation increases, 318 IPPs and, 194 electricity outages, 182 management contracts, 112­116, 119b, 122n3 independent regulators, 120 operating inefficiencies, 66, 71­74, 309 informal settlements, 136 port ownership and management, 253 irrigation, 290 power sector, 190, 190b ODA, 79 reforms, 195­197 power sector underpricing, 191 spending needs estimates and, 36 private sector participation, 113, 114b, 116, 116b water supply, 314b property rights, 136b State Rural Water Supply and Sanitation Agency railways, 233, 239 (Nigeria), 316b road network, 213, 222­223 storage infrastructure for water security, 53­54, 58, 277, standposts, 95 279­282, 280f telecommunications regulation, 170 strategic planning transport infrastructure, 150, 151t irrigation, 294­295 urban infrastructure, 128, 137 ports and shipping, 256 utility tariffs, 99 Index 353 water resource management, 46, 280, 282, 305 transportation sector, 203­210 water supply, 314, 315 alternative technologies for, 84 Tanzania Investment Centre, 136b capital budget for, 68 Tanzania Railways Corporation, 233 competition, 207­208 tariffs financing, 133 cost recovery and, 70­71, 99­100, 99f human development constrained by, 46 increasing block tariffs, 101, 308b infrastructure as constraint on economic growth, 45 power sector, 191 institutional competence and capacity, 108, 109 price premium of, 49­52 local capital markets and, 81 rail freight, 235 logistics systems for, 205­206, 205f, 207b road freight, 5, 50, 151 multimodal transport, 204­205 subsidies of, 92­93, 93­94f non-OECD financing, 68f telecommunications, 170, 170f ODA for, 68f water, 299, 307­309, 308b, 319­320 operating inefficiencies, 74t TEAMS (The East Africa Marine System, Kenya), 148b, 148t PPI investment, 68f telecommunications. See also information and communications private sector participation, 116, 117t, 209 technologies public expenditures on infrastructure, 68f, 69f, 70­71t access to, 88­89f regional integration, 145, 147f cost recovery, 5 roads. See roads economic impact of, 45, 45f, 47 sector performance data collection, 39 independent regulators and, 120, 121t social obligations and, 209­210 institutional competence and capacity, 107, 108, 109f, 110f spending needs, 36, 55­56, 57t landline. See landline telephones transit corridor development, 206­207, 207b, 208b operating inefficiencies, 66, 72 urban vs. rural network quality, 129, 130f price premium on, 50­52 Transports Internationaux Routiers (TIR, EU), 206, 208b private sector participation, 111­113, 111f, 113f trucking regulation, 206­207, 224­225 regional cooperation opportunities, 146­148, 148b sector performance data collection, 39 Uganda state-owned enterprises, 118 affordability issues, 92 Tema Development Corporation, 138 deficit-financed public investment, 77b thermal power, 149, 200 drinking water access, 300 time allocation and basic infrastructure access, 95b hydroclimatic variability, 272, 273f TIPAC (Transit Inter-États des Pays de l'Afrique Centrale), 208b hydropower project, 201 TIR (Transports Internationaux Routiers, EU), 206, 208b informal settlements, 139 Togo, urban infrastructure financing in, 137 mobile phone coverage, 178 toll roads, 213 multimodal transport network, 204 tour de role dispatching, 203, 210n1, 220, 224 paved road density, 96b trade power sector inefficiencies, 187 by landlocked countries, 206 PPI investment, 80 power sector, 53, 187, 198­200 private sector participation, 112, 113 regional integration and, 149, 150­153, 150f, 151t, 159b railways, 233, 244 transportation network and, 153, 204­205 road maintenance, 216 traffic volumes sanitation, 325 air transport, 260­261, 260f, 261t standposts, 95 maritime, 250­251, 250­251t, 251f state-owned enterprises, 118, 119b, 314b railways, 233­234, 234f telecommunications industry privatization, 171 roads, 212 undermaintenance, 73 Trans-Africa Highway Network, 34, 152­153, 212 water supply, 313, 315, 316, 320 transboundary river basins, 153­154, 272­273, 274f, 279 state-owned enterprise reforms, 314b transit corridor development, 206­207, 207b, 208b water utilities, 310, 313 Transit Inter-États des Pays de l'Afrique Centrale (TIPAC), 208b underwater communication cables, 146, 149t, 174 Transnet, 256 unimproved latrines, 94, 96f, 97t, 321­327, 328f transparency Union of African Railways, 237 private sector participation, 114­115, 117 Union of Producers, Transporters and Distributors of state-owned enterprises, 195 Electric Power in Africa, 40 Transportation Department (U.S.), 268 United Arab Emirates, submarine cable project, 148b 354 Index United Kingdom and development assistance, 32 economic development and, 272­276, 277 United Nations Children's Fund (UNICEF), 300, 321n2, financing, 133, 302­305, 303t 323, 324b funding gap, 66, 75, 75t, 302­305, 305t United States and development assistance, 78, 295b groundwater, 320­321 Universal Safety Oversight Program (International Civil Aviation human development constrained by, 45­46, 274­275 Organization), 268 independent regulators and, 120, 121t Urban Affairs, Housing, Urban Water, Public Hygiene and institutional competence and capacity, 108­109, 109­110f, Sanitation Ministry (Senegal), 334 115­116, 116b, 282­283 urban areas, 125­142 management of, 4, 272­275, 274­275f, 275t agricultural productivity near, 126­127, 127t, 140n2 MDG for, 299, 300, 302­305, 303t, 321n1 capital budgets, 133­137, 134f non-OECD financing, 68f, 304 electrification, 198 ODA for, 67, 68f, 304, 304t as growth engine, 126 operating inefficiencies, 10­12, 11­12f, 12t, 66, 72, 73, 74t, growth of, 127­128, 127t, 136­137 310­311f, 310­313, 312t housing planning, 134­136 poverty reduction and, 274, 277, 278f informal settlements, 136­137, 139 PPI investment, 68f, 304, 304t infrastructure, 2­3, 3f, 128­130, 129­130f, 131t private sector participation, 111­112, 111f, 115­116, 116b, costs of, 130­132, 131t, 133­134f 117t, 310­313, 311f, 312b financing, 133­138, 135t public sector expenditures, 68f, 69f, 70­71t, 304, 304t investment needs, 132­133 reforms, 309­313, 311f, 315t, 317 land regulations, 134­136, 136b, 138­139 regional integration of resources, 153­154, 279, mobile phone coverage, 166, 166f 282­283, 283f principles of efficient urbanization, 139­140 resellers, 301, 302b public transportation, 94, 96b rural access, 129, 129­130f, 300­302, 301­302t, 313­316, roads, 220­221, 223­224 315f, 315t, 316b rural links to, 126­130, 140n1 sector performance data collection, 39­40 sanitation, 325­326, 326f, 326t, 327, 328f security. See water security spatial planning, 134­136 spending needs, 36, 54­55, 55t, 58, 280­281, 281t territorial planning and, 138­139 standposts for, 305­306, 306b transportation networks and, 56 state-owned enterprises, 119b, 120f, 120t, 314b water supply access, 300­302, 311f, 320­321 subsidies, 92­93, 93­94f user charges, 68, 70­71, 72t tariffs, 92­93, 299, 307­309, 308b, 319­320 U.S. Geological Survey, 35b technologies for, 305­307, 318­319 utilities. See also specific sectors transboundary river basins, 153­154, 272­273, 274f, 279 connection subsidies, 90, 98­99, 98b, 100­101, 101t, 308b urban access, 129, 129­130f, 300­302, 301­303t, 320­321 institutional competence and capacity, 108 water security nonpayment issues, 91, 91f challenges, 276­277 operating inefficiencies, 10­12, 11­12f, 11f, 12t, 66, 71­74, economic development and, 276f, 277, 277b 73t, 310­313, 311f, 312t investment needs, 279­283 distribution losses, 72­73 spending needs, 53­54 overemployment, 72 storage infrastructure, 279­282 undercollection of bills, 73 Water Utilities Partnership, 39 undermaintenance, 73­74, 73f weather forecasting, 275­276, 279 prepayment meters, 100, 100f wells, 94, 96f, 97t, 300 water supply and, 301, 301t, 302­303t, 306b, 318­319 urban vs. rural access to, 300­301, 301t, 303f, 303t, 320 West Africa. See also specific countries Ventilated Improved Pit (VIP) latrines, 324, 335n2 airports and air transport, 263, 268, 268f Voice over Internet Protocol (VoIP), 175 air transport hubs, 151, 260 voice services, 166­167, 172, 172t backup generators, 184 capital markets, 81 water resources, 271­285, 299­322. See also irrigation economic growth, 44 alternative technologies for, 84, 96f, 97t electricity access, 193 capital budget for, 68 freight services profit margins, 50, 220 climate resilience, 279 ICT investment, 147, 148b, 148t cost recovery, 307­309, 307f, 308b, 319­320 infrastructure, 48 Index 355 maritime traffic growth, 250, 250­251t, 251f Yemen ports and shipping, 253 multimodal transport network, 204 Ying, Yvonne, 95b port system, 252, 255 railway concessions, 238­239 Zambezi River basin, 272, 280 regional integration, 156 Zambia roads, 211 border crossing delays, 159b water utilities, 310 broadband access, 173 West African Economic and Monetary Union, 156, 264­265, 264t governance, 108 West African Gas Pipeline, 154 hydrologic variability, 273 West African Power Pool, 52, 53, 53­54t, 159­160, 185, 187 irrigation, 290, 290f Master Plan, 34 land use policies, 138 WHO/UNICEF Joint Monitoring Programme, 300 mobile phone coverage, 172 WiMAX technology, 57, 58 power sector underpricing, 191 wireless technologies, 178 railways, 204, 230, 235, 239, 244 Wodon, Quentin, 95b road network, 213, 214, 215, 222­223 women and water resources, 45­46, 94, 304­305 sanitation, 329 World Bank shipping, 251 AICD and, 32, 37, 40 transit corridor development, 207 AICD GIS data from, 35b utility inefficiencies, 309 Country Policy and Institutional Performance Assessment, 51 water resource management, 276, 282, 306 irrigation investment by, 295b water supply, 315, 320 ODA from, 78 water utilities, 313 regional project criteria, 158 Zambia­Malawi­Mozambique growth triangle, 157 World Development Report 2009 (World Bank), 127, 141n4, 145 Zimbabwe World Health Organization (WHO), 300, 321n2, 323, border crossing delays, 159b 324b, 335n3 drinking water access, 300 World Meteorological Organization, 275­276 railways, 230, 232, 238, 244 transit corridor development, 207 Yamoussoukro Decision (1999), 151, 159, 259, 264, 268 water resource management, 281 yard taps, 319 water supply, 320 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The · 31 trees Office of the Publisher has chosen to print · 10 million Btu of Africa's Infrastructure: A Time for Trans- total energy formation on recycled paper with 30 percent · 2,957 lb. of net postconsumer fiber in accordance with the greenhouse gases recommended standards for paper usage set · 14,242 gal. of waste by the Green Press Initiative, a nonprofit water program supporting publishers in using fiber · 865 lb. of solid waste that is not sourced from endangered forests. For more information, visit www.greenpress initiative.org. Sustainable infrastructure development is vital for Africa's prosperity. And now is the time to begin the transformation. This volume is the culmination of an unprecedented effort to document, analyze, and interpret the full extent of the challenge in developing Sub- Saharan Africa's infrastructure sectors. As a result, it represents the most comprehensive reference currently available on infrastructure in the region. The book covers the five main economic infrastructure sectors--information and communication technology, irrigation, power, transport, and water and sanitation. Africa's Infrastructure: A Time for Transformation reflects the collaboration of a wide array of African regional institutions and development partners under the auspices of the Infrastruc- ture Consortium for Africa. It presents the findings of the Africa Infrastructure Country Diagnostic (AICD), a project launched following a commitment in 2005 by the international community (after the G8 summit at Gleneagles, Scotland) to scale up financial support for infrastructure development in Africa. The lack of reliable information in this area made it difficult to evaluate the success of past interventions, prioritize current allocations, and provide benchmarks for measuring future progress, hence the need for the AICD. Africa's infrastructure sectors lag well behind those of the rest of the world, and the gap is widening. Some of the main--policy-relevant--findings highlighted in the book include the following: infrastructure in the region is exceptionally expensive, with tariffs being many times higher than those found elsewhere. Inadequate and expensive infrastructure is retarding growth by 2 percentage points each year. Solving the problem will cost over US$90 billion per year, which is more than twice what is being spent in Africa today. However, money alone is not the answer. Prudent policies, wise management, and sound maintenance can improve efficiency, thereby stretching the infrastructure dollar. There is the potential to recover an additional US$17 billion a year from within the existing infrastructure resource envelope--simply by improving efficiency. For example, improved revenue collection and utility management could generate US$3.3 billion per year. Regional power trade could reduce annual costs by US$2 billion. And deregulating the trucking industry could reduce freight costs by one-half. So, raising more funds without also tackling inefficiencies would be like pouring water into a leaking bucket. Finally, the power sector and fragile states represent particular challenges. Even if every efficiency in every infrastructure sector could be captured, a substantial funding gap of $31 billion a year would remain. Nevertheless, the African people and economies cannot wait any longer. Now is the time to begin the transformation to sustainable development. ISBN 978-0-8213-8041-3 SKU 18041