4~~C&~\ C- ~~~~~~~- 4~~~~ 13S, Sion lit- 'NN'fr4 ;j~~~~~~~~~~~~~~~~~~~' - S - -~~~~Ad3 iaI S - - C~~~~~~-77777i 7-77 MOROCC ALGERIA A' ~~~~~~~~LIBYA ARAB VERIDE THE1 GAMBIA UKN SAG TOME AND PRENCI CCTAN COMOROS~~~~ILA SIJB-SAHARAN AFRICA OTSWAFRIAA Adjustment in Africa A World Bank Policy Research Report Reforms, Results, and the Road Ahead Published for the World Bank OXFORD UNIVERSITY PRESS Oxford University Press OXFORD NEW YORK TORONTO DELHI BOMBAY CALCUTTA MADRAS KARACHI KUALA LUMPUR SINGAPORE HONG KONG TOKYO NAIROBI DAR ES SALAAM CAPE TOWN MELBOURNE AUCKLAND and associated companies in BERLIN IBADAN C 1994 The International Bank for Reconstruction and Development I THE WORLD BANK 1818 H Street, N. W Washington, D. C. 20433, U.S.A. Published by Oxford University Press, Inc. 200 Madison Avenue, New York, N.Y 10016 Oxford is a registered trademark of Oxford University Press. All rights reserved. No part ofthis publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States ofAmerica First printing February 1994 The boundaries, colors, denominations, and other information shown on the map in this vol- ume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. Library of Congoress Cataloging-in-Publication Data Adjustment in Africa: reforms, results, and the road ahead. p. cm. - (A World Bank policy research report) Includes bibliographical references. ISBN 0-19-520994-X 1. Structural adjustment (Economic policy)-Africa, Sub-Saharan. I. World Bank. 11 Series Hc8OO.A55258 1993 338.967-dc2O 93-47887 CIP IssNv 1020-0851 ® Text printed on paper that conforms to the American National Standardfor Permanence ofPaper for Printed Library Materials, Z39.48-1984 Contents Foreword xi The Report Team xiii Acknowledgments xv Definitions and Data Notes xvii Overview 1 Policies Are Getting Better 3 Better Policies Pay Off 4 Policies Are Not Good-Yet 8 The Road Ahead for Adjustment 9 Notes 16 1 Why Africa Had to Adjust 17 Understanding the Stagnation and Decline 20 The Needed Switch in Policies 34 Notes 41 2 Moving Toward Sound Macroeconomic Policies 43 Fiscal Deficits Down-But Not Out 45 Monetary Policy Mostly on Track 48 Mixed Progress in Exchange Rate Policy 51 Further Progress Needed in Macroeconomic Policies 57 Notes 59 3 Unleashing Markets 61 Trade Reforms 62 Agricultural Reforms 76 Letting Other Markets Work 88 Establishing a Sound Incentive Framework: Mixed Results 95 Notes 98 v A0J E IAFRICA 4 Reforming the Public Sector 99 Public Enterprise Reform: Little Evidence of Significant Progress 101 Financial Reforms: Limited Signs of Sustainable Progress 110 Improving Public Sector Management: A Long-Term Challenge 120 Notes 129 5 The Payoffs to Reforms 131 How to Assess the Payoffs to Reforms 132 Better Macroeconomic Policies Boost Growth 133 Agriculture Is Growing Faster 143 Industry Is Expanding 149 Exports Are Growing 152 No Quick Response in Investment or Savings 153 Notes 158 6 Poverty and the Environment 161 Adjustment, Growth, and Poverty 162 The Environment 174 Notes 179 7 The Road Ahead for Adjustment 181 The Guiding Principles for Adjustment 182 Moving Forward Where There Is Consensus 184 Rethinking Adjustment Where There Is Less Success- And Less Consensus 196 Policy Reform and the Poor 209 Adjustment Policies and the Environment 210 Debt, Aid, and Adjustment 213 A Broader Commitment to Reform 217 Conclusion 219 Notes 220 Appendix A: Statistical Tables 221 Appendix B: The Indexes of Macroeconomic Policy Change and Stance 259 Appendix C: Agricultural Policy Indicators 271 References 275 Vi CONTENTS Boxes 1.1 Policies Matter for Growth 22 1.2 Nigeria's Missed Opportunity 32 1.3 Country Coverage and Time Frame of the Study 36 1.4 Adjustment for Sustained Development 37 1.5 Moving from Adjustment to Development in Ghana 40 2.1 Differences in Monetary and Fiscal Policies for Fixed and Flexible Exchange Rate Regimes 52 2.2 Devaluations Need Not Be Inflationary 54 3.1 What Is Trade Reform, and How Would We Know If There H!as Been Some? 64 3.2 Trade Reform: Mauritius and Ghana 68 3.3 Own-Funds Importing 70 3.4 Reversing Tanzania's Cashew Decline 82 3.5 Maize Marketing Reform 86 3.6 What Removing Fertilizer Subsidies Would Mean 89 3.7 Africa's Harsh Climate for Business 90 3.8 Getting More Mileage Out of Petroleum Reform 92 4.1 Restructuring Public Enterprises: Zimbabwe's Encouraging Results 109 4.2 Lease Contracts and Concessions in Public Utilities 111 4.3 Public Finance Reform in Ghana 127 4.4 Changes in Public Spending in Kenya 128 5.1 Explaining the Increase in GDP per Capita Growth 140 5.2 The Payoffs to Fixed and Flexible Exchange Rate Regimes 144 5.3 Why Agricultural Is Growing Faster in Nigeria Than in Ghana 147 5.4 Aggregate Supply Tends to Respond Slowly in the Short Term 148 5.5 How to Know Whether Deindustrialization Is Occurring 150 6.1 The Costs of Recession and an Inadequate Adjustment in C6te d'][voire 166 7.1 Pushing Exports in Today's Global Marketplace 189 7.2 Africa's Adding-Up Problems: Not Serious in Most Cases 190 7.3 Using Vouchers to Privatize: Mongolia's Innovative Approach 202 Text figures 1 Median Changes in Average Annual Growth Rates of Adjusting African Countries between 1981-86 and 1987-91 5 2 Policy Stance and Median GDP per Capita Growth in Adjusting African Countries 6 1.1 Average ANnnual GDP per Capita Growth, 1965-85 18 1.2 GDP per Capita 19 1.3 Average Annual Growth in Exports 20 vii ADJUSTMETF IN AFRICA 1.4 Premium on the Parallel Market for Foreign Exchange 23 1.5 Government Consumption as a Share of GDP 24 1.6 Outward Orientation of Selected Country Groups, 1965-85 25 1.7 Barter Terms of Trade 27 1.8 Changes in GDP per Capita Growth and External Income in Sub-Saharan Africa, Various Periods 30 1.9 Education and GDP per Capita Growth 39 2.1 Fiscal Indicators for African Adjusting Countries 46 2.2 Barter Terms of Trade and the Real Effective Exchange Rate 56 2.3 Change in Macroeconomic Policies, 1981-86 to 1987-91 58 3.1 Evolution of Trade Policy in Selected Countries during the Adjustment Period 67 3.2 Evolution of Import Restrictiveness in Ghana 71 3.3 Change in the Real Producer Price of Agricultural Exports, 1981-83 to 1989-91 78 3.4 Change in Overall Taxation of the Agricultural Sector, 1981-83 to 1989-91 79 3.5 Changes in Explicit and Implicit Taxation of the Agricultural Sector, 1981-83 to 1989-91 80 4.1 Public Sector Wage Bill 122 5.1 Change in Average Annual GDP per Capita Growth, 1981-86 to 1987-91 134 5.2 Average Annual GDP per Capita Growth: Actual Change between 1981-86 and 1987-91 and Predicted Change Based on External Transfers 136 5.3 Relation between the External Environment and GDP per Capita Growth, 1981-86 to 1987-91 137 5.4 Changes in Macroeconomic Policies and GDP per Capita Growth, 1981-86 to 1987-91 141 5.5 Relation between Agricultural Growth and Two Measures of Agricultural Policy, 1981-86 to 1987-91 146 5.6 Changes in the Growth of Industry and Manufacturing, 1981-86 to 1987-91 151 5.7 Median Change in Real Export Growth, 1981-86 to 1987-91 153 5.8 Median Change in Gross Domestic Investment as a Share of GDP, 1981-86 to 1987-91 155 5.9 Median Change in Gross Domestic Savings as a Share of GDP, 1981-86 to 1987-91 157 6.1 Allocation of the Education Budget in Selected African Countries 173 7.1 Debt Burdens of Severely Indebted, Low-Income African Countries under Alternative Rescheduling Terms, 1991 215 vii' CONTENTS Text tables 1.1 External Capital Transfers for Low-Income Countries 28 1.2 Effects of Changes in External Income on GDP, 1971-73 to 1981-86 29 1.3 Sub-Saharan Africa in 1990 versus Southeast Asia in 1965: Means for Selected Indicators 38 2.1 Fiscal Policy Stance, 1990-91 48 2.2 Monetary Policy Stance, 1990-91 49 2.3 Exchange Rate Policy Stance for Countries with Flexible Exchange Rates, 1990-91 55 2.4 Exchange Rate Policy Stance for Countries with Fixed Exchange Rates, 1990-91 57 2.5 Countries Ranked by Overall Macroeconomic Policy Stance, 1990-91 58 3.1 Phases in Liberalizing Import Regimes 66 3.2 Countries Classified by Agricultural Policy Environment 84 3.3 Government Intervention in Marketing Major Food Crops 85 3.4 Price Controls and Subsidies for Fertilizer 88 3.5 Price Controls on Goods 91 3.6 Government Intervention in Selected Sectors 96 3.7 Countries Classified by Macroeconomic and Market Intervention Policies 97 4.1 Financial Auditing of Major Public Enterprises, Late 1992 103 4.2 Divestiture of Public Enterprises, 1986-92 104 4.3 Financial Flows between the Government and Public Enterprises in Burundi 107 4.4 Financial Sector Reforms Undertaken during the Adjustment Period 112 4.5 Governnment Intervention in the Financial Market 115 4.6 Government Participation in the Capital of Commercial Banks 119 4.7 Reductions of Civil Service Personnel in Selected Countries, 1981-90 123 4.8 Change in the Number of Civil Service Personnel, 1985-92 124 4.9 Public Sector Management of the Payroll System, Late 1992 125 4.10 Public Investment Program, Late 1992 126 5.1 GDP per Capita Growth 138 6.1 Changes in Real Health and Education Expenditures in Selected Countries, 1980-83 to 1987-89 171 6.2 Social Spending in Selected Countries 172 Appendix tables A. 1 Share of Sub-Saharan Africa's Agricultural Export Earnings, by Major Crop 221 A.2 Fiscal Indicators 222 Ix ADJUSTMENT IN AFRICA A.3 Seigniorage and Inflation 226 A.4 Real Interest Rate for Deposits 227 A.5 Parallel Market Exchange Rate Premium 228 A.6 Change in the Real Effective Exchange Rate, 1980 to 1990-91 229 A.7 Foreign Exchange Allocation and Import Controls for Selected Countries 230 A.8 Import Items Subject to Nontariff Barriers 231 A.9 Marketing Controls on Major Agricultural Exports 232 A.10 Monopoly Activities before Reforms 234 A. 11 Monopoly Activities, Late 1992 236 A.12 Government Intervention in Selected Markets before Reforms 238 A. 13 Government Intervention in Selected Markets, Late 1992 239 A.14 Tax Revenue in Selected Countries 240 A.15 Changes in External Income and in GDP per Capita Growth between 1981-86 and 1987-91 241 A. 16 Changes in the Real Effective Exchange Rate (REER) and in GDP per Capita Growth between 1981-86 and 1987-91 242 A.17 Average Annual GDP per Capita Growth for Countries Classified by Macroeconomic and Market Intervention Policies 243 A.18 Changes in Producer Prices of Export Crops and in Agricultural Growth 244 A.19 Changes in Agricultural Taxation and Agricultural Growth 245 A.20 Agricultural Growth 246 A.21 Growth in Industry and Manufacturing 247 A.22 Growth in Exports 249 A.23 Investment 250 A.24 Savings 252 A.25 Sources of Income of Poor, Rural Smallholders in Selected Countries 254 A.26 Sources of Agricultural Income of Poor, Rural Smallholders in Selected Countries 255 A.27 Expenditures of Poor, Rural Smallholders in Selected Countries 256 A.28 Allocation of the Education Budget in Selected Countries 257 A.29 Change in Net External Transfers as a Share of GDP, 1981-86 to 1987-91 258 B.1 Change in Macroeconomic Policies, 1981-86 to 1987-91 260 B.2 Alternative Approaches for Calculating Overall Change in Macroeconomic Policies 263 B.3 Economic Outcomes Using Alternative Approaches for Calculating Overall Change in Macroeconomic Policies 265 B.4 Economic Outcomes Using Alternative Time Periods for Calculating Overall Change in Macroeconomic Policies 266 B.5 Components of Macroeconomic Policy Stance, 1990-9 1 268 x Foreword A BROAD-BASED PATTERN OF RAPID ECONOMIC GROWTH IS vital to reducing poverty in Sub-Saharan Africa. Many African .ucountries have undertaken structural adjustment programs to reverse the economic decline of the 1980s and accelerate growth. GDP per capita growth remains low, however, raising troubling questions about the extent and efficacy of the policy reform efforts. For this rea- son, the Development Economics Vice Presidency conducted a study to assess how much policy reform has taken place in Africa, how successful it has been, and how much more remains to be done. This report, Ad- justment in Africa: Reforms, Results, and the Road Ahead, summarizes the findings of that research. A companion report, Adjustment in Africa: Lessons from Country Case Studies (Husain and Faruqee forthcoming), documents reform efforts in seven countries. Adjustment programs are necessary but not enough to raise eco- nomic growth. As discussed at length in Sub-Saharan Africa: From Cri- sis to Sustainable Growth (World Bank 1989a), investments in human capital and infrastructure, efforts to build the economic institutions necessary to a well-functioning market economy, and initiatives to in- crease technical capacity must also continue apace. This report, with its focus on adjustment, is intended to complement other World Bank publications dealing with the various facets of Africa's long-term devel- opment strategy. Adjustment in Africa reviews the policy reforms typically included in African adjustment programs during the second half of the 1980s and analyzes their relation to economic performance. The evidence shows that progress has been mixed, and that in every African country, key re- forms are still incomplete. There are rewards to adjustment, however, as countries that have come the furthest in implementing good policies-particularly good macroeconomic policies-have enjoyed a resurgence of growth. But the level of per capita growth, even among the countries that have adjusted the most, is still below what is needed for rapid poverty reduction. xi ADJUSTMENT IN AFRICA Where do adjustment programs go from here? The report concludes that in the macroeconomic, trade, and agricultural sectors, the major task is to move forward with the current approach to policy reform. In the financial and public enterprise sectors, some rethinking of strategy is called for. This report highlights the role that adjustment needs to play in improving the policy environment for the provision of basic social services and protecting the environment. Government ownership of an economic reform program is a prereq- uisite for its success. But ownership must not stop with the government. Political leaders must build a broad-based consensus on the need for re- form so that adjustment programs are not derailed by powerful interest groups. One of the major challenges for the next generation of adjust- ment programs is for governments and donors alike to find ways of widening ownership and building consensus. This study is the second in a series of Policy Research Reports, which are intended to bring to a broad audience the results of World Bank re- search on development policy issues. As reports on policy issues, these books should help us take stock of what we know-and what we do not know. While remaining accessible to nonspecialists, they should con- tribute to the debate among academics and policymakers about appro- priate public policy objectives and instruments for developing economies. And as research documents, these books may also provoke debate, both within the Bank and outside, concerning the analytic methods used and the conclusions drawn. Adjustment in Africa is a product of the staff of the World Bank, and the judgments made herein do not necessarily reflect the view of the Board of Directors or the governments they represent. Michael Bruno Vice President Development Economics and Chief Economist The World Bank xii The Report Team C HRISTINE W. JONES AND MIGUEL A. KIGUEL WERE THE PRIN- cipal atthors ofthe report. Lant Pritchett and Michael Walton col- laborated closely and made important contributions. Takamasa Akiyama, Elliot Berg, Tyler Biggs, Lawrence Bouton, Gerald Caprio, Yoon Je Cho, Wilfrido Cruz, William Easterly, Ibrahim Elbadawi, Faezeh Foroutan, Ronald Johannes, Ross Levine, John Nash, Jo Ann Paulson, David Sahn, Hafeez Shaikh, and the Fiscal Affairs Department of the International Monetary Fund contributed to particular sections. Francisca Castro and Heidi Zia provided research assistance. Lawrence H. Summers played a leading role in the initial stages of the report's preparation. The work was carried out under the general direction of Nancy Birdsall and completed under Michael Bruno. Bruce Ross-Larson was the principal editor. The editorial-production team for the report was led by Kathryn Kline Dahl. Lawrence MacDonald provided additional editorial input. The support staff was headed by Cecilia Guido-Spano and included Julia Baca, Milagros Divino, Raquel, Luz, Anna Marafion, Rebecca Martin, and Christopher Rollison. xliii Acknlowledgments T HIS REP'ORT WAS PREPARED IN CLOSE COLLABORATION WITH the World Bank's Africa Region. Special thanks are due to the Africa Region country economists, whose input was essential to assessing the extent of policy reforms. Under the direction of Ishrat Husain, the lead economists of the Africa Region-Lawrence Hinkle, Peter Miovic, lJlrich Thumm, Ajay Chhibber, Fran,ois Laporte, and Gene Tidrick--made valuable comments and contributions, as did the Africa Technical Department under the guidance of Kevin Cleaver and Michel 'Wormser. Philip Birnbaum provided liaison with the donors participating in the Bank's Special Program of Assistance for Africa (SPA). The strong support of Edward V. K. Jaycox is much appreciated. Preparation of this report drew on the seven country case studies of adjustment coordinated by Ishrar Husain (Husain and Faruqee forth- coming). The report also benefited from research from the African Economies in Transition Project managed by Jo Ann Paulson and from the Trade Policy Expansion Project directed by John Nash. Takamasa Akiyama, Mark Blackden, Ajay Chhibber, Graeme Donovan, Alan Gelb, Alfred Gulstone, Rebecca Hanson, Brendan Horton, Steven Jaf- fee, Chad Leechor, John Nash, John Page, Paul Popiel, Mustapha Rouis, Shala Torabi, and Rogier van den Brink contributed material for boxes and tables in the report. Many individuals inside and outside the World Bank provided valu- able comments. Particular thanks are due to those Bank staff outside the Africa Region who commented on various drafts: Alan Gelb, Magdi Iskander, Pierre Landell-Mills, Johannes Linn, Gobind Nankani, John Nellis, John Page, and Lyn Squire. External reviewers included Paul Collier, Gerald Helleiner, Tony Killick, Benno Ndulu, Ademola Oyejide, Jorn Rattsoe, and representatives of the SPA donors, the Eco- nomic Commission for Africa, and the Organization of African Unity; the usual disclaimer absolving them of responsibility for the contents of the report applies. xv Definitions and Data Notes Country Groups Analytical groiups. This book focuses on twenty-nine Sub-Saharan countries with a population of 500,000 or more in mid-1991, reason- able social stability, and adjustment programs in place during 1987-91. If no other group is specified, the text and data refer to these countries: Benin, Burkina Faso, Burundi, Cameroon, the Central African Republic, Chad, Congo, Cote d'Ivoire, Gabon, The Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Madagascar, Malawi, Mali, Mauritania, Mozambique, Niger, Nigeria, Rwanda, Senegal, Sierra Leone, Tanzania, Togo, Uganda, Zambia, and Zimbabwe. These African adjusters are sometimes classified into subgroups: *Countries with flexible exchange rates: Burundi, The Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Madagascar, Malawi, Mauritania, Mozambique, Nigeria, Rwanda, Sierra Leone, Tanza- nia, Uganda, Zambia, and Zimbabwe. * Countries with fixed exchange rates: Benin, Burkina Faso, Cameroon, the Central African Republic, Chad, Congo, C6te d'Ivoire, Gabon, Mali, Niger, Senegal, and Togo. * Low-income countries are those with a gross national product (GNP) per capita of $610 or less in 1990. They are Benin, Burkina Faso, Burundi, the Central African Republic, Chad, The Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Madagascar, Malawi, Mali, Mauritania, Mozambique, Niger, Nigeria, Rwanda, Sierra Leone, Tanzania, Togo, Uganda, and Zambia. * Middle-income countries are those with a GNP per capita of more than $610 but less than $7,620 in 1990. They are Cameroon, Congo, Cote d'Ivoire, Gabon, Senegal, and Zimbabwe. * Oil exporters Cameroon, Congo, Gabon, and Nigeria. Occasionally, this book also presents data for the following analytical groups: xvii AJTME tN AFRICA * Other developing countries: For convenience, low- and middle- income countries are sometimes referred to as developing coun- tries. Classification by income does not, however, imply a judgment about development status. The composition of "other developing countries" varies depending on the availability of data. * Other adjusting countries refers to developing countries outside Sub-Saharan Africa with adjustment programs during 1987-91. * The high-performing Asian economies are Hong Kong; Indonesia; Japan; the Republic of Korea; Malaysia; Singapore; Taiwan, China; and Thailand. Geographic groups. The geographic groupings used in this report are not intended to be comprehensive lists of all countries in a particular region. Furthermore, complete data may not be consistently available for all countries in each group. Where coverage differs significantly from the standard definitions that follow, those differences are indicated. * Sub-Saharan Africa comprises World Bank borrowers south of the Sa- hara: Angola, Benin, Botswana, Burkina Faso, Burundi, Cameroon, Cape Verde, the Central African Republic, Chad, Comoros, Congo, C6te d'Ivoire, Djibouti, Equatorial Guinea, Ethiopia, Gabon, The Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Liberia, Madagascar, Malawi, Mali, Mauritania, Mauritius, Mozambique, Namibia, Niger, Nigeria, Rwanda, Sao Tome and Principe, Senegal, Seychelles, Sierra Leone, Somalia, Sudan, Swaziland, Tanzania, Togo, Uganda, Zaire, Zambia, and Zimbabwe. In this report, the term "Africa" refers to Sub-Saharan Africa. * East Asia and the Pacific comprises China, Fiji, Indonesia, the Re- public of Korea, Malaysia, Papua New Guinea, the Philippines, the Solomon Islands, Thailand, and Vanuatu. * Latin America and the Caribbean comprises Argentina, the Ba- hamas, Barbados, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, El Salvador, Grenada, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago, Uruguay, and Venezuela. * The Middle East comprises the Islamic Republic of Iran, Iraq, Israel, Jordan, Oman, the Syrian Arab Republic, and the Republic of Yemen. * North Africa comprises Algeria, Egypt, Morocco, and Tunisia. xvii. DEFINITIONS AND DATA NOTES * South Asia comprises Bangladesh, India, Myanmar, Nepal, Pak- istan, and Sri Lanka. Data Notes HI ISTORICAL DATA IN THlS BOOK MAY DIFFER FROM THOSE IN other World Bank publications if more reliable data have be- come available, if a different base year has been used for con- stant price data, or if countries have been classified differently. Additionally, in this book: * Billion is 1,000 million. * Trillion is 1,000 billion. * Dollars are current U.S. dollars unless otherwise specified. * Growth rates are based on constant price data. * Net transfers equal net transfers on debt (loan disbursements minus arnortization minus interest) plus grants (excluding techni- cal assistance) plus International Monetary Fund net transfers. * Ratios reported over a multiyear period are annual averages unless otherwise specified. Acronyms and Abbreviations CFA franc Currency known in West Africa as the "franc de la Commu- naute financiere d'Afrique" and known in Central Africa as the "franc de la Coop&ation financiere en Afrique centrale" EPZ Export processing zone FAO Fooct and Agriculture Organization of the United Nations GDP Gross domestic product GNP Gross national product HPAEs Higlh-performing Asian economies IMF International Monetary Fund NTB Nontariff barrier OGL Open general license REER Real effective exchange rate RIPP Real producer price of agricultural exports SPA Special Program of Assistance for Africa VAT Value-added tax xix Oveirview N THE AFRICAN COUNTRIES THAT HAVE UNDERTAKEN AND sustained major policy reforms, adjustment is working. But a number- of countries have yet to implement the reforms needed to restore growth. And even among the strongest adjusters, no country has gone the full distance in restructuring its economy. Of the twenty-nine countries studied in this report, the six with the mosiL improvement in macroeconomic policies between 1981-86 and '1987-91 enjoyed the strongest resurgence in economic performance. 1 TIhey experienced a median increase of almost 2 percent- age points in the growth rate of gross domestic product (GDP) per capita, bringing their median rate of growth up from a negative level to an average 1.1I percent a year during 1987-9 1. The increase in their indus- trial and export growth rates was even more striking. And agricultural growth also accelerated in the countries that taxed their farmers less. By contrast, countries that did not improve their policies saw their median GDP growth fall to a level of -2 percent a year, in all likelihood increas- ing the numbe=r of the POOr Policv reforns have been uneven across sectors and across countries. The countries studied here have generally been more successful in im- proving their macroeconomic, trade, and agricultural policies than their public and financial sectors. Almost two-thirds of the countries man- aged to put better macroeconomic and agricultural policies in place by the end of the 1980s. Improvements in the macroeconomic framework also enabled countries to adopt more market-based systems of foreign exchange alloCation and fewer administrative controls over imports. However, reforms remain incomplete. No African country has achieved a sound macroeconomic policy stance-which in broad terms means inflation. under 10 percent, a very low budget deficit, aind a con- trial and export growth rates was even more striking. And agricultura ADJUS ENTJ IN AFRICA petitive exchange rate. In a third of the countries, macroeconomic poli- cies actually deteriorated over the decade. Furthermore, countries are still taxing their farmers heavily, through marketing boards and/or over- valued exchange rates. Most countries have further to go in eliminating nontariff barriers and adopting a moderate, tariff-based level of protec- tion. Social spending, while not showing an overall decline during the adjustment period, is misallocated within the health and education sec- tors. And the politically difficult reform of the public enterprise and fi- nancial sectors lags well behind. Moreover, there is considerable concern that the reforms undertaken to date are fragile and that they are merely returning Africa to the slow- growth path of the 1960s and 1970s. At the same time, there is hope that Africa, like East Asia thirty years ago, will move onto a faster develop- ment track. For that to happen, more progress will be required in macro- economic reform-to provide a stable environment in which economic activity can flourish. Much more progress in trade, agricultural, and reg- ulatory reform will also be needed-to create a favorable climate for busi- ness so that Africa can join the world economy. And growth with equity will call for strong political resolve to tackle money-losing public enter- prises and bloated bureaucracies-to free up the resources needed to im- prove basic health and education services for the poor. Adjustment alone will not put countries on a sustained, poverty- reducing growth path. That is the challenge of long-term development, which requires better economic policies and more investment in human capital, infrastructure, and institution-building, along with better gov- ernance. But development cannot proceed when inflation is high, the exchange rate overvalued, farmers overtaxed, vital imports in short sup- ply, prices and production heavily regulated, key public services in dis- repair, and basic financial services unavailable. In such cases, fundamen- tal restructuring of the economy is needed to make development possible. The objective of structural adjustment programs thus is to es- tablish a market-friendly set of incentives that can encourage the accu- mulation of capital and more efficient allocation of resources. This report addresses three questions: How much did adjusting African countries change their policies? Did their policy reforms restore growth? And what is the road ahead for adjustment? In answering these questions, the report advances the debate on adjustment by providing the most comprehensive data so far on policy changes in Sub-Saharan Africa. It takes a careful look at whether reforms are paying off, and it 2 OVERVIEW identifies the areas where the adjustment strategy needs to be redirected. The report shows that African countries have made great strides in im- proving policies and restoring growth, but that they still have a long way to go in adopting the policies needed to move onto a faster growth path and reduce poverty.2 Policies Are Getting Better T HE TWENTY-NINE AFRICAN COUNTRIES EXAMINED HERE drew Up adjustment programs in the 1980s-programs intended to improve the poor policies that were the primary cause of the 15 percent fall in Africa's GDP per capita between 1977 and 1985. The outcomes? Macroeconomic reforms have spurred exter- nal competitiveness while keeping inflation low. Trade reforms have increased access to the imports needed for growth. And the reduced taxation of agriculture has helped the poor while encouraging produc- tion and exports. * On the macroeconomic front, six of the adjusting countries had a large improvement in policies, nine a small improvement, and eleven a deterioration.3 As a whole, they cut their budget deficits (by a median of 1.9 percent of GDP between 1981-86 and 1990-9 1) and reduced inflation to moderate levels. And the coun- tries with flexible exchange rates (those outside the CFA franc zone) depreciated the real effective exchange rate by 50 percent and re- duced the premium on the parallel market for foreign exchange (from a nmedian of 60 percent during 1981-86 to 25 percent dur- ing 1990-9 1). * In trade, many countries have substantially reduced the number of imports subject to nontariff barriers and begun to rationalize the tar- iff structure. Most of the flexible exchange rate countries have moved to more automatic systems of granting foreign exchange licenses. * In agriculture, two-thirds of the adjusting countries are taxing their farmers less. Despite huge declines in real export prices, pol- icy changes have increased real producer prices for agricultural ex- porters in ten countries. Of the fifteen governments that had major restrictions on the private purchase, distribution, and sale of 3 ADUTENT IN AFRICA major food crops before adjustment, thirteen have withdrawn from marketing almost completely. For public enterprises and financial enterprises, however, there have been few policy changes. * African governments have sold off only a small share of their as- sets. The value of privatizations in Nigeria between 1988 and 1992 was less than 1 percent of that in Argentina, Malaysia, or Mexico, even after adjusting for Nigeria's smaller GDP.4 Explicit and implicit financial flows to public enterprises are still high. But one encouraging trend is that governments have stopped expand- ing their public enterprise sectors. * In most African countries, the financial sector, despite reform ef- forts, is still heavily burdened by public sector demands for credit-with the central government alone (excluding public en- terprises) absorbing more than 30 percent of domestic credit. Better Policies Pay Off T HERE HAS BEEN MUCH TALK ABOUT THE COSTS OF ADJUST- ment, less about the substantial benefits. Most countries that improved their policies have returned to positive rates of GDP per capita growth. This turnaround shows that adjustment policies work when implemented properly. And although GDP per capita growth rates remain low, it is unreasonable to expect that African countries would quickly match the rapid rise of the best performers in Asia and elsewhere. Even before the macroeconomic crisis of the early 1980s, Sub-Saharan Africa was growing more slowly than other regions. As we have noted, the six adjusting countries wvith the most improved macroeconomic policies had a median increase in GDP per capita growth of almost 2 percentage points between 1981-86 and 1987-91 (figure 1). That compares with an increase of 1.5 percentage points for those countries with less improved policies and a decline of 2.6 percentage points for those with a deterioration in policies. The median increase in export growth was almost 8 percentage points for countries with the most improved macroeconomic policies, while in countries with policy 4 OVERVIEW 7 ,rd? 1% -' Change in GDP per Capita Growth Change in Real Export Growth Percentage points Percentage points 4 10 3 9 8 2 _ _ _7 1 K- ~~~~~~~~~~~~~~~~6 -12 0 4~~~~~~~~~~~~~~~~ -1 -4 -2 Countries with large Countries with Countries with Countries with large Countries with Countries with improvement in small improvement deterioration in improvement in small improvement deterioration in macroeconomic in macroeconomic macroeconomic macroeconomic in macroeconomic macroeconomic policies policies policies policies policies policies Change in Industrial Growth Change in Agricultural Growth Percentage points Percentage points 8 3 7 2 6 51 s <-11 ~~~~~~~~~~~~~~1X 43 . i0 0 E --- -2 Countries with Iarge Countries with Countries with Countries with Countries with Countries with improvement in small improvement deterioration in large decrease in small decrease in increase in taxation macroeconomic in macroeconomic macroeconomic taxation of export taxation of export of export crop policies policies policies crop producers crop producers producers Note: See source tables for a listing of countries in each group. Sources: Table 5.1 and appendix tables A. 19, A.2 1, and A.22. Policy reforms paid off in higher growth rates in income, exports, industry, and agriculture. deteriorations, export growth declined 0.7 percentage points. For the best performers, industrial growth accelerated by more than 6 percent- age points, compared with 1.7 percentage points in countries with dete- riorating policies. And countries that substantially reduced the taxation of export crop farmers increased median agricultural growth by 2 per- centage points, while countries that taxed farmers more saw growth fall by 1.6 percentage points. 5 ADJUSTMENT IN AFRICA Policy packages to address the adverse external shocks and severely overvalued real exchange rates of the early 1980s had high payoffs. Countries that brought about a real depreciation of 40 percent or more between 1981-86 and 1987-91-all of them with flexible exchange rates-had a median increase in GDP per capita growth of 2.3 percent- age points. Countries that had appreciations-all of them with fixed ex- change rates-suffered a median decline of 1.7 percentage points. These results demonstrate the payoffs to improving policies. What about the payoffs to good policies? Countries that maintained or ended up with fair or adequate macroeconomic policies during 1987-91 did better than countries with poor or very poor policies (figure 2). The me- dian rate of GDP per capita growth in countries with the better macro- economic policy stance was 0.4 percent a year between 1987 and 1991-low but at least positive, and a turnaround from annual declines of about 1 percent a year in the early 1980s. By contrast, in countries with poor or very poor macroeconomic policies, median GDP per capita growth fell 2.1 percent a year on average. The extent of government in- tervention in markets also made a difference in growth. Countries that limited their intervention in markets had median GDP per capita growth of almost 2 percent during 1987-91, compared with declines of more than 1 percent for the countries that intervened more extensively. Countries with better policy stances had faster GDP per capita growth. .-'g'Zn 2 BiFoJ,y aaouSae andf NfJeiffan GOP per Cap'a Growth in A djusting Africani Countiies Macroeconomic Policies and Growth Market Intervention Policies and Growth Average annual growth, Average annual growth, 1987-91 (percent) 1987-91 (percent) 3 3 2 2 1 1 0 0 -1 -1-. c -2 -2 -3 -3 Countries with fair or Countries with poor Countries with Countries with Countries with adequate policies or very poor policies limited intervention medium intervention heavy intervention in 1990-91 in 1990-91 in 1992 in 1992 in 1992 Note: See source tables for a listing of countries in each group. Sources: Table 5.1 and appendix table A. 1 3. 6 OVERVIEW External Transfers Helped Increases in external transfers (a median rise of 2.4 percent of GDP be- tween 1981-86 and 1987-91) also contributed to faster growth. Six- teen countries benefited from higher external transfers. Countries with increases in external transfers (a median increase of 0.6 percent of GDP) had a median increase in GDP per capita growth of 1.2 percentage points. Those with reductions (a median decrease of 0.6 percent of GDP) suffered a small slowdown in GDP per capita growth. External transfers relieved import constraints, financed investment, and smoothed con- sumption-just what they are intended to do. But overall, policy re- forms were more strongly associated with increases in growth rates than external transfers were. The Impact of Adjustment on the Poor and the Environment In African countries that have undertaken some reforms and achieved some increase in growth, the majority of the poor are probably better off and almost certainly no worse off. The poor are mostly rural, and as producers, they tend to benefit from agricultural, trade, and ex- change rate reforms and from the demonopolization of important com- mercial activities. As consumers, both the urban and the rural poor tend to be hurt by rising food prices. But adjustment measures have seldom had a major impact on food prices in either the open market or the par- allel market, which supplies most of the poor. Where rationing was widespread, as in Tanzania, real prices for key consumer goods have even fallen. Similarly, the layoffs of public sector employees, who are among those hardest hit by adjustment, have not generally added to the num- ber of poor people. Many of those who lost their jobs were able to find other work, often by returning to rural areas. The absence of empirical studies makes it difficult to document any clear and speci.fic link between adjustment reforms and environmental changes in Sub-Saharan Africa. To the extent that policy reforms have encouraged souand pricing of energy, fertilizer, and water resources, they have reduced wasteful distribution and consumption. Not all distor- tions have been eliminated, however, and there is still much room for progress in instituting appropriate systems of natural resource pricing and taxation. 7 ADJSiTMENT IN AFRICA Policies Are Not Good-Yet D ESPITE THE EFFORTS TO IMPROVE THE MACROECONOMIC environment, open up markets, and strengthen the public and financial sectors, most African countries still lack policies that are sound by international standards. Even Africa's best performers have worse macroeconomic policies than the newly industrializing economies in Asia. Few besides Ghana come close to having adequate monetary, fiscal, and exchange rate policies. And Ghana lags behind other adjusting countries elsewhere-Chile and Mexico, for exam- ple-in trade and public enterprise reform. In trade, many African countries have, by eliminating extensive im- port controls, returned to the regimes they had before the crisis-helped in many cases by successful exchange rate depreciations that restored competitiveness. Other countries that never experienced a severe macro- economic crisis, such as Kenya and Zimbabwe, have moved slowly to- ward import liberalization. The current policy stance in countries with flexible exchange rates is free of the heavy administrative controls that characterized the period before adjustment, but most African countries still have some nontariff barriers and high and dispersed tariffs. The policy stance for agricultural pricing and other price controls is more difficult to quantify. Most countries have eliminated price controls and restrictions on the marketing and pricing of food staples, and many have eliminated costly subsidies for fertilizer (with no apparent reduc- tion of fertilizer use) and liberalized its distribution. But governments continue to intervene heavily in the marketing of export crops. The scarce evidence on public enterprise reform suggests that there has been no significant reduction in financial flows to public enterprises or in the volume of assets held by the government. Nor has there been a sustainable improvement in the efficiency of enterprises remaining pub- lic. The paucity of data partly reflects institutional weaknesses, but it probably also reflects the lack of government commitment to results. Financial reform lags behind as well. The financial position of the bank- ing sector is weak because of poor macroeconomic management, which in- duces the monetization of fiscal deficits through the banks. It is also weak because of the slow pace of reform in the public enterprise sector. And it re- flects continuing government interference in the management of the fi- nancial sector. A large share of bank lending still goes to the public enter- prise sector, making it more difficult for the private sector to borrow. 8 OVERVIEW Although public spending on health and education did not decline in the adjustment period-an achievement given the fiscal problems of African countries-there is little evidence of an increase in that spend- ing. Nor is there much evidence that public spending within those sec- tors is being reallocated away from costly tertiary programs and toward the basic services most likely to reach the poor. The Road Ahead for Adjustment D RAWING ON SUCCESSFUL EXPERIENCES ELSEWHERE AND taking Sub-Saharan Africa's circumstances into account, three principles can guide African governments undertaking reform programs. * Get macroieconomic policies right. Keeping budget deficits small helps in controlling inflation and avoiding balance-of-payments prob- lems. Keeping a realistic exchange rate pays off in greater interna- tional coimpetitiveness and in supporting convertible currencies. * Encourage competition. Competition means higher productivity, and firms forced to compete are more efficient than those with privi- leged access to credit or foreign exchange. A top priority for reform in Africa is to increase competition through domestic deregulation, trade refcorm, and the privatization of public enterprises. * Use scarce institutional capacity wisely. Because most African countries have limited capacity to govern well, high priority should be given to reforms that minimize unnecessary government involvement in markets. For example, marketing boards should be abolished, public enterprises privatized, and import restrictions replaced by tariffs. Many African countries are moving in the right direction with their macroeconomic, agricultural, and trade policies, and most policymakers agree on what still needs to be done. But there has been little progress in reforming public enterprises and the financial sector, and there is much less consensus on how to proceed. Reform in these sectors is particularly difficult because of the powerful vested interests that have been created through government intervention. A strong social consensus on the need to improve governance is thus a prerequisite for progress. 9 ADJUSTMENT IN AFRICA Moving Forward Where There Is Consensus Getting macroeconomic policies right. Countries should continue with the current strategy: avoiding overvalued exchange rates and keeping inflation and budget deficits low. Good macroeconomic policies have paid off in East Asia, and they will pay off in Africa, too-indeed they are already starting to do so. Most countries in the region still need to cut budget deficits and in- direct fiscal losses (those covered by the banking system) in order to lessen the need for inflationary financing or additional external financ- ing. There is litde scope for cutting overall public spending in many countries, although the composition of spending can and should be im- proved. Increasing tax revenues is thus the best avenue for reducing deficits, but the increases should come by levying broad-based taxes that do not unduly penalize businesses and by granting fewer exemptions that favor the politically well-connected. Domestic savings, which are low in Africa relative to other develop- ing regions, must increase to finance investment. Eliminating large neg- ative real interest rates is a crucial first step. But given the complexity of devising additional policies to encourage private savings, raising public savings is the best option in the short run. The surest way to increase savings in the long term is to boost growth, because growth and savings reinforce each other in a virtuous circle, with high growth leading to high saving and to higher growth. Taxing agriculture less. In agriculture the main task is to continue reducing the taxation of farmers by liberalizing pricing and marketing and by reducing the protection of industry. Progress has been made, but countries need to do more to help farmers, and the elimination of agricultural marketing parastatals, particularly for export crops, must be high on the agenda. Liberalizing markets so that private agents can compete with parastatals and linking producer prices to world market prices may be useful transitional mechanisms in the near term. These reforms can help farmers reap the full benefit of the exchange rate depreciations, which might otherwise merely shore up the financial profitability of parastatals. Care must be taken not to undermine market liberalization efforts with restrictive licensing procedures and other interventions that give marketing parastatals an undue competitive advantage. Traders often face a thicket of regulations for licensing, transportation, the movement IO OVERVIEW of goods, trading hours and locations, and weights and measures. Elim- inating these burdensome obstacles is essential for increasing profitabil- ity and production in agriculture. Simultaneous progress in the devel- opment agenda is also important. Improving the quality of public spending for transport networks, rural infrastructure, and agricultural research and extension will enhance the payoffs to improving agricul- tural policies. Putting exporters first. Because exports are so beneficial for growth, countries should consider the needs of exporters carefully and apply an "exporters first" rule. One easy way for government to help exporters is to remove unnecessary policy impediments-by providing automatic access to foreign exchange, eliminating export monopolies, and facilitating access to intermediate inputs and capital goods. Governments also need to welcome foreign participation, because for- eign firms can bring the contacts and production knowledge needed for penetrating global markets. But governments and international agencies should abandon the practice of trying to pick "winners"- that is, pushing particular exports-because they have consistently made poor choices in the past. Export processing zones have seldom been more effective than simple free-trade zones and bonded produc- tion areas, so it is important to find other mechanisms to help exporters avoid administrative, regulatory, and tariff impediments. A high priority is developing workable schemes to provide exporters access to duty-free inputs. The potential for export growth is great, because African countries are starting from a very low base. Even modest success in increasing their share of world markets will translate into tremendous growth. The fu- ture is in nontraditional exports, but traditional exports still need to be part of an outward-oriented strategy. Gaining just a very small foothold in the world market for such traditional, labor-intensive goods as cloth- ing and footwear would substantially increase the region's exports. But this does not rnean that Africa should neglect its traditional export of primary commodities, even those that face limited world demand. Al- though the region already has a large market share in a handful of agri- cultural commodities, notably cocoa, it is possible to expand that share further. Good policies and investments in infrastructure and research and extension activities can help to raise the productivity of African pro- ducers and displace higher-cost producers elsewhere (as Indonesia and Malaysia have demonstrated). II ADJUS NET IN AFRICA Rationalizing import barriers. There has been progress in liberalizing imports, but most countries have gone only halfway. African countries should continue to eliminate nontariff barriers (NTBs) to rationalize the trade regime and increase transparency. The focus should be not on fine-tuning tariff levels but on establishing a credible schedule for sub- stituting tariffs for NTBs. Even very high tariffs, if imposed only for a clearly limited period, can support the objectives of adjustment. The next steps on the agenda are to simplify the tariff structure, reduce the highest rates to more moderate levels, and institute a minimum tax- so long as effective systems are in place to provide exporters duty-free access to imports. These reforms can often generate enough revenue to offset a fairly substantial overall lowering of tariffs, while leading to a more competitive environment and productivity gains. Beyond that, further progress toward a low and completely uniform tariff structure should not sacrifice fiscal revenues. Rethinking Adjustment Where There Is Less Success- and Less Consensus Privatizing public enterprises. The efforts to privatize state corporations and to improve their performance have yielded meager results so far. African governments have resisted privatization, especially of the most important public enterprises. But the alternatives-imposing hard budget constraints, granting the enterprises greater autonomy, and putting them on a commercial footing-seldom work. Countries elsewhere are getting around the obstacles to privatization, and their experience might be useful in Africa. Some of these countries have fostered broad-based ownership by giving private citizens vouchers for shares in public enterprises, or reserving shares for employees. Others are using various types of private investment and holding companies to improve corporate management. Nonasset divestiture-through leas- ing, concessions, and incentive-based performance contracts-can in- crease private sector management of the public utilities and other nat- ural monopolies and improve their productivity. Prudent financial reform. The overall approach to financial develop- ment is on target, but reforms have suffered from too much faith in quick fixes. African countries need to continue with a three-part strat- egy of reducing financial repression, restoring bank solvency, and improving financial infrastructure. But adjustment programs have I2 OVERVIEW been overly hasty in cleaning balance sheets and recapitalizing banks in an environrnent where institutional capacity is weak and the main borrowers (the government and public enterprises) are financially dis- tressed. Many programs were based on the assumption that banks could improve their performance simply by removing the bad loans from their balance sheets, replacing managers, and injecting new capi- tal to bring assets up to international standards. This usually was insufficient for several reasons: reforms were not accompanied by needed macroeconomic and structural changes, bank managers con- tinued to be exposed to political interference, and regulatory and supervisory capacities were inadequate and could only be developed over time. A more prudent strategy to restore bank solvency involves downsiz- ing publicly owned banks, privatizing them where possible, and encour- aging new entrants. Because most African countries lack the capacity to regulate and supervise, the challenge is to devise a financial system that offers extra cushions against risk-by setting higher-than-normal capital- adequacy ratios, relying more on foreign banks, and limiting entry to reputable banks with a solid capital base. Countries must strike a bal- ance between the need to increase competition and the need to ensure the solvency of financial institutions. Improving public sector management remains a major challenge for the road ahead-but one that probably extends beyond what ad- justment-related policy reforms alone can accomplish. Perhaps the biggest challenge is to build a more effective civil service to provide the elements necessary for a well-functioning market economy, in- cluding a sound macroeconomic and legal framework and a system for providing basic social services consistent with the development objective of growth with equity. There is increasing recognition that adjustment programs, with their focus on containing civil service costs, have had limited success in tackling the more fundamental problems of the public sector, such as the lack of accountability and transparency, civil service employment and pay practices that are un- related to technical competence and productivity, regressive patterns of resource mobilization, expenditures that conflict with develop- ment priorities, and the limited capacity for policy analysis. Broader approaches that address the difficult tasks of strengthening the ad- ministrative structure and creating the conditions for improved gov- ernance are thus called for. 13 AJbUSTMENT IN AFRICA More Adjustment-Not Less-Would Help the Poor and the Environment Findings from Brazil, C6te d'lvoire, and Peru show that the lack of adjustment is what most hurts the poor and most increases their num- ber. Addressing the fundamental policy distortions that inhibit growth is thus an essential part of a strategy to reduce poverty. The poor will benefit more from an increase in growth if spending programs to develop human resources are protected during the adjust- ment process, and if the policy package eliminates the distortions in labor, land, and output markets that disadvantage the poor. More could have been done, and should have been done, to reduce poverty in the context of adjustment programs. This has been changing in the past few years, as adjustment programs strive to improve public expenditure in the social sectors. But the fundamental development challenge of im- proving Africa's human resource base requires more than policy change-it also requires sustained investment and institution-building. In addition to reducing poverty, adjustment programs in Sub- Saharan Africa can promote judicious use of natural resources by insti- tuting policy reforms that affect the pricing of agricultural and forest outputs, petroleum products, energy, and so forth. But macroeconomic and broad sectoral policies are very general and cannot substitute for specific environmental interventions. Designing effective systems for environmental protection when institutional capacity is limited is no simple task. It may be preferable to give firms and communities incen- tives to protect the environment rather than to depend on governmen- tal regulatory and enforcement capacity. As with poverty, many envi- ronmental problems require a combination of policy reform, investment, and institution-building. Aid and Growth Aid to African countries must be structured in ways that speed, rather than impede, growth. Higher income generates greater domestic savings and, in time, reduces the dependence on foreign savings. But today's large volume of aid poses dangers: it could soften budget constraints and thus finance the postponement of public sector reforms. Expanded aid flows should therefore be linked to strong reform programs and better governance. In financing country-specific adjustment programs that 14 OVERVIEW have a good probability of yielding substantial reforms, a key issue is to design transfer mechanisms and to allocate aid across countries and sec- tors so that it supports a policy and investment framework for high ac- cumuLation of capital and rising public savings. Another key issue is to design aid so th[at it supports reforms without adding distortions in for- eign exchange or labor markets and so that it builds institutions up in- stead of wearing them down. One of the major challenges on the road ahead is finding ways to help governments promote widespread owner- ship of adjustment programs and muster support among the interest groups that have the most to gain from reforms. Efforts by donors to bring Africa's stock of debt down to sustainable levels can, when linked to strong adjustment efforts, help countries real- ize the benefits of policy reforms. The debt burden of many African countries is huge, and many will have too much debt even under the very favorable cLebt relief proposals under consideration. So far, aid flows and concessional lending have more than offset debt service payments. But in the medium and long term, as countries adopt better policies, the debt overhang is likely to deter private investment. And the debt service burden threatens to eat away at increased export earnings and domestic savings that might otherwise be used in pursuit of long-term develop- ment objectives. For countries undertaking comprehensive and sus- tained policy reform, reducing the debt stock burden to a manageable level would improve their development prospects. This means rethink- ing the current debt relief strategy, which still leaves many countries with debt service requirements beyond their capacity to pay. The focus should be on reducing the stock of debt to sustainable levels, even if that means differences in treatment across countries. Even with transformed policies, higher savings, and better invest- ments, Africa will still require exceptional external assistance for at least another decade. But countries cannot expect an increased flow of foreign resources without undertaking the economic reforms necessary for growth and poverty reduction. And such economic reforms will probably not take place until the conditions for good governance are established. MEMO Adjustment is the necessary first step on the road to sustainable, poverty-reducing growth. But adjustment programs in Sub-Saharan Africa have been burdened with unrealistically high hopes, driven in part by awareness of the real poverty that economic growth can help al- '5 AJST NT IN AFRICA leviate. Some proponents of adjustment thought that it could quickly put African countries on a much higher growth path than before. Too often there has been little effort to determine whether Africa's disap- pointing economic performance in the aggregate represents a failure to adjust or a failure of adjustment. Opponents have wrongly cast and crit- icized adjustment as an alternative to measures supporting long-term development. The resulting confusion has sometimes led to sterile de- bate about the efficacy of adjustment policies. More important, it has risked creating undue pessimism among African countries and donors. That pessimism is unwarranted, for there has been progress. The turn- around in growth shows that adjustment-even incomplete adjust- ment-can put African countries back on the road to development. Notes 1. See box 1.3 in chapter 1 for a listing of the countries 3. Complete macroeconomic data were available for included in the study sample. only twenty-six countries. 2. Schadler and others (1993) examined similar issues 4. Data on the value of privatizations come from for the group of countries benefiting from the Interna- Schwartz and Lopes (1993). tional Monetary Fund's Enhanced Structural Adjustment Facility. They used a different methodology but reached broadly similar conclusions, I6 CHAPTER 1 Why Africa Had to Adjust UB-SAHARAN AFRICA'S ECONOMIC GROWTH, NEVER spectacular, has been the weakest among developing regions (figure 1.1). The region most in need of growth to reduce poverty has had the least. Between 1965 and 1985, its GDP per capita increased less than 1 percent a year on average. More worrisome, its economic performance actu- ally began deteriorating in the mid 1970s. After growing an average of 2.6 percent a year between 1965 and 1974, GDP per capita stagnated or turned down thereafter in most Sub-Saharan countries (figure 1.2). By the early 19 80s, few Sub-Saharan countries had kept their per capita GDP growth in line with the rest of the world's, and many had a lower GDP per capita than before independence some twenty years earlier. The economic situation worsened in the first half of the 1980s, with further deteriorations in the terms of trade and sharply reduced access to international finance. For more than two-thirds of all Africans, real incomes were lower in 1985 than in the mid-1970s. Southeast Asia presents a startling contrast. In 1965 Indonesia's GDP per capita was lower than Nigeria's, and Thailand's lower than Ghana's. Indonesia relied on oil as much as Nigeria did. Thailand, much like Ghana, was a poor agricultural country. Who could have predicted then that in 1990 Indonesia's GDP would be three times that of Nigeria? Or that Thailand would become one of the world's best performing economies, while Ghana would be struggling to regain its former in- come level? Few areas of economic activity were exempt from the stagnation and decline. After healthy increases in exports in 1965-73, the growth slowed-and then plunged to negative levels between 1981 and 1986 (figure 1.3). Exports of manufactures declined slightly between 1970 17 ADJUSTMENT IN AFRICA Figure 'Ll Avw?g.s. rnuma "UP pe C'aprmrlQa ne<}C.Alh9OS Percent 5 4 3 2 Long-term growth rates in Sub- Saharan Africa lagged behind those of other regions. o East Asia and North Africa Latin America South Asia Middle East Sub-Saharan the Pacific and Turkey and the Africa Caribbean N`ote: Growth rates were calculated using the ordinary least-squares method. Source: World Bank data. and 1986, while increasing fivefold for Latin America, sixfold for the Middle East and North Africa, and thirteenfold for Asia's newly indus- trializing countries. Even agricultural exports slipped, as the region's share in developing-country exports of food and other agricultural prod- ucts went from 17 percent in 1970 to 8 percent in the mid-1980s. The region also lost ground in its exports of ores and minerals. Only in oil did Africa improve its export share. Most of the region's economies failed to diversify their export base and continued to rely on only one or two commodities. In the mid- 1980s, primary products generated 80 percent of Africa's export rev- enue, roughly the same share as in the 1960s, and manufactured exports were significant in only a handful of countries. Meanwhile, exports of agricultural products became even more concentrated. Nine major com- modities accounted for 76 percent of the region's agricultural exports in the 1980s-up from 70 percent in the 1960s (appendix table A.1). Countries elsewhere, by contrast, diversified their export base and in- creased their shares in world exports of primary commodities. I8 WHY AFRICA HAD TO ADJUST Constant 1987 dollars 1,000 Thirty-five other developing countries , . . - , - ' ' 840 4 .*' * t 680 , 520 , <- 7 Sub-Saharan Africa 360 East Asia and the Pacific African economies stagnated while 200 others improved steadily. 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 Adjustment period Source: World Banik data. Agriculture, particularly important for Africa's growth, did worse than other sectors. Between 1965 and 1980, agricultural growth rose only 2 percent a year-less than the rate of population growth-and be- tween 1981 and 1985, it fell 0.6 percent a year. Compare that with agri- cultural growth of 3.2 percent a year in East Asia, 2.5 percent in South Asia, and 3.1 percent in Latin America. By the micL 1980s, symptoms of the malaise were evident almost everywhere. The returns on World Bank investment projects were much lower in Africa than in other regions, and more than a quarter of those projects failed to generate a positive rate of return. It was (and still is) al- most impossible to attract foreign private capital-either in investment or loans-and portfolio investment flows were negligible. The interna- tional prices for Africa's government debt in secondary markets were the lowest for developing countries, reflecting the markets' perception of African countries as uncreditworthy. The physical infrastructure, already poor, deteriorated from lack of maintenance, and the quality of govern- ment services suffered. Health and education indicators, though better than in the 1 960s, were still a long way below those for other developing countries. Clearly, it was time for African economies to begin to adjust. I9 ADUTMENT IN AFRICA Percent 12 1965-73 10 1974-80 1981-86 E 1987-91 8 6 4 Africa's exports took a sharp (live in the early 1980s. -2 Sub-Saharan Africa East Asia and the Pacific Thirty-six other developing countries Note: Exports valued in local currency, constant prices. Source: World Bank data Understanding the Stagnation and Decline THERE IS NO SINGLE EXPLANATION FOR AFRICA'S POOR PER- fo rmance before the adjustment period. The main factors behind the stagnation and decline were poor policies-both macroeconomic and sectoral-emanating from a development para- digm that gave the state a prominent role in production and in regu- lating economic activity. Overvalued exchange rates and large and pro- longed budget deficits undermined the macroeconomic stability needed for long-term growth. Protectionist trade policies and government monopolies reduced the competition so vital for increasing productivi- ty. In addition, the state increased its presence in the 1970s, nationaliz- ing enterprises and financial institutions and introducing a web of reg- ulations and licenses for most economic activities. More important, the development strategy had a clear bias against exports, heavily taxing agricultural exports, one of the largest suppliers of foreign exchange. 20 WHY AFRICA HAD TO ADJIJST- The choice of poor policies may be understandable in light of condi- tions in Africa after independence. Because of the lack of domestic cap- ital and entrepreneurs, the unwillingness to rely on foreign capital, and the underlying distrust of the market, almost all African countries chose (with the full support of aid donors) to rely on the state. They had the company of countries in Eastern Europe, Latin America, and South Asia-and the encouragement of many development economists. Industrialization was believed to be the key to rapid growth, because of declining prices for primary commodities and because of the benefits of reducing reliance on imported manufactures. Agriculture, rather than being stoked as the engine of growth, was taxed to provide the resources to build a modern industrial sector. Governments drew up five-year plans, created public enterprises, and enacted regulations to control prices, restrict trade, and allocate foreign exchange in pursuit of social goals. At the same time, countries were struggling to establish them- selves as nations and put new governmental structures in place. But gov- ernments became overextended, particularly relative to their weak insti- tutional capacities, as they tried to build national unity and deliver on the promises of independence. Factors outside Africa also contributed to the decline of the 1970s and 1980s, though their importance is too often exaggerated. Non-oil exporters suffered falls in the terms of trade, but the losses, which actu- ally began in the 1 960s, were no larger than those for other developing countries. But the volatility of export receipts-and thus of foreign ex- change earnings and fiscal revenues-complicated macroeconomic management. Depending on the availability of foreign exchange, im- ports went through cycles of compression and decompression, as did growth. Investment budgets and public sector employment expanded rapidly in boom years, hampering adjustment in bust years. Poor poli- cies failed to encourage export diversification and increase international reserves to smooth the impact of adverse shocks. Domestic Facitors: Poor Policies Largely to Blame There is ample evidence that sound macroeconomic and sectoral poli- cies are associated with higher growth (see box 1.1). Policies in Sub- Saharan Africa, however, have generally been worse than those elsewhere. Overvalued exchange rates. The presence of unofficial, parallel markets for foreign exchange-in countries with flexible exchange rates- 2I ADJUSTMENT IN AFRICA Box 1 I Picies Matter ftDr GroWhi RESEARCH ON THE DIETERMINANTS OF LOFNG- between 1965 and 1985, would bring Africa's long- term growth shows that better policies typically mean run GDP per capita growth rate to about 4 percent a faster growth (Barro 1991; Easterly 1992; Killick year-close to East Asia's. 1992). The most successful economies maintained good macroeconomic policies, as measured by low in- flation, prudent fiscal stances, and realistic exchange %x Table 1. i2f i!Yllluld C Ihages:n Seeede rates. Among the macroeconomic variables, the ab- iI! GOP bPes iiCCi t2P l i a rokE sence of a paraRlel market-an indication thatthe ex -Esti m ate d change rate is not overvalued-gealas the Increase changein strongest explanatorypower. opolicies, such as in variable growth those for nance and trade, also 0mat Cies (percentage (percentage with larger and rmore developed fa l rks riable points) points) generally grow faster. The same is te frtri tiof i ettoDP 1 +0.1 to 0.2 with more open trade regimes. Primary school enrollment Estimate;of how changes in indicators of cr10 +0.2 to 0.3 policies would affect growth sust thatpoliesm- Secondary school ensglo02t0 ter a great dleal (box table 1.1). For exampl, a reform 10 x+0.2hto0.3 package that in the long run raised enrollment ratios10 Pa.4rallelmaeexchange by 10 Percentage points, eliminat a prle mR o t 10 +0.2to0.4 premium of 20 percent, raised the ratio of ement tio of average producer investment to GDP by 3 percent point,andin- input price to world creased financial depth (the ratio of M2 to GDP') by 10 marke price 10 -0.4 percentage points would lily raise the annual per Ratio ofovernmentf capita growth rate tby 2.6 percentage points. And if consumption to GDP 10 -1.2 policy reforms were helped by an improvement in the Ratio ofequipment terms of trade of l percentage point of GDP a year, an- Ratinfvestrent toaGDP 1 +0.3 nual GDP per capita wouldl grow between 0.4 and 0 Ratio ofaverage change in percentage points. All these improvements, added to terms oftrade to GDp 1 +0.4to0.8 the 0.9 percent annual growth that Africa experiened Source: Easterly (1992), table 1. reveals overvaluation of the real effective exchange rate.1 Africa's paral- lel market premiums have been by far the largest in the developing world, and they increased significantly until the adjustment period (figure 1.4). During the economic crisis in the first half of the 1980s, the average premium was almost 300 percent. Rough estimates indi- cate that a 10 percent premium is likely to reduce GDP growth by 0.4 percentage points a year (box 1.1). But the impact wanes as the premi- um goes up, and a 100 percent premium cuts GDP growth by 2 per- centage points a year. For Ghana between 1974 and 1980, the parallel 22 WHY AFRICA HAD TO ADJUST rigre 2.4 irsmNv[ n Pfia iAaKk i0 r or [½FaEFT Lir!a Percent 300 SubSaharan adjusting 250 K countries with flexible exchange rates 200 Twenty-three other 150 100 50 L The parallel market premium has -been dramatically higher in Africa o than elsewhere. 1974-80 1981-86 1987-91 (Adjustment period) NVote: The premium is calculated as the percentage difference between the parallel market exchange rate and the official exchange rate (in domestic currency at the end of the period). Sources: International Currency Analysis, Inc. (various years); IMF data. market premium of more than 200 percent explains a slowdown in GDP growth of'about 3 percentage points a year. Heavy government spending. Among the more common indicators of good fiscal policy are a small budget deficit and a low ratio of govern- ment consumption (that is, current spending on goods and services) to GDP. Most Sub-Saharan countries had neither. However, while Africa's budget deficits were not much larger than those elsewhere, the ratio of government consumption to GDP was a different story. Consumption began to increase in the early 1960s and reached its peak-nearly 17 percent of GC P-in the early 1980s, exceeding the ratios of other regions by 5 to 6 percentage points (figure 1.5). That excess was important in deterring growth, for statistical analysis shows that over the typical range of government expenditure, each 10 percentage point increase in the ratio of government consumption to GDP typically reduces GDP per capita growth by 1.2 percentage points. Inward-looking trade policy. Africa's trade policies have also been poor. Most Sub-Saharan economies followed an inward-oriented, import- substitution strategy, supplemented by widespread use of tariff and 23 ADJSTMENT IN AFRICA Percent 18 Sub-Saharan African = 16 14 12 Twenty-six other developing countries 10 8 - African governments outspent others East Asia and the Pacific on purchases of goods and services. 6 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 Adjustment period Note: Data are weighted by 1980 GDP in U.S. dollars. a. Includes twenty-five countries only. Source: World Bank data. nontariff barriers to reduce external competition, mainly in manufac- turing. Dollar's (1992) index of outward orientation showed Sub- Saharan African countries as the least outward-oriented-and Asia's high-performing economies as the most (figure 1.6).2 Of the twenty countries with the highest nontariff barriers, eleven were in Sub- Saharan Africa (Pritchett 1991). This protectionism was another unfortunate policy choice, because competition increases productivity while trade restrictions increase input prices and the cost of capital, choking growth. Political instability. Political and social stability also are usually associ- ated with higher rates of growth (Barro 1991; Fosu 1992), and the instability in Sub-Saharan Africa partly explains its sluggish perfor- mance. More than half the region's countries have been rocked by civil war, uprisings against the government, and devastation from drought and famine, and their average per capita GDP growth was -0.5 percent a year for 1965-85. By contrast, the region's eleven stable countries had an average growth rate of 1.4 percent (Hodd 1991). But the strong 24 WHY AFRICA HAD TO ADJUST High-performing East Asian economies South Asia - Latin America and the Caribbean N _;7 Middle East and North Africa Sub-Saharan Africa .. :- In openness to foreign trade, Africa trailed the world. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Index scores Note: A high score corresponds to more outward orientation. Source: World Banik (1993a), figure 6.4. association berween stability and growth merely raises the question: which is the cause, and which the effect? Other domestic factors. What about other domestic factors, such as human capital? Of all endowments, human capital probably does most to fuel long-term economic gains. Simply put, countries with skilled people grow faster. The fact that Sub-Saharan Africa trails other regions in social indicators can thus help to explain its slow growth. But how does one reconcile Africa's protracted decline with the fact that human capital, measured by several indicators, improved in Sub- Saharan Africa after independence? Primary school enrollments increased from 41 percent of the eligible population to 69 percent between 1965 and the mid-1980s, with girls accounting for much of the increase. Secondary school enrollments increased from 2 to 14 per- cent, with almost equal increases for boys and girls. Health and nutri- tion indicators also showed progress. Infant mortality was cut in half, the number of doctors and nurses per capita increased, and life expectancy at birth rose by eight years. Financial strength is another important determinant of growth: countries with more developed financial sectors grow faster. And finan- cial strength, rneasured by various indicators, improved in Africa. For instance, M2 as a share of GDP, a measure that indicates how well-devel- oped the financial sector is, increased from about 15 percent in the early 25 ADJUSTMENT IN AFRICA 1970s to 20 percent in the mid-1980s. So financial development, like human development, was stemming the decline, not causing it. International Factors: Surmountable Obstacles Changes in the terms of trade do have an impact on long-term growth. Easterly and others (1993) found that over the 1980s, for ex- ample, a negative terms-of-trade shock averaging 1 percentage point of GDP a year lowered growth by 0.8 percentage points a year. Like other regions, Africa suffered a fall in the terms of trade between the early 1970s and the mid-1980s. But on an annual basis, the terms-of-trade decline was small and thus not a major factor in Africas poor growth record. Moreover, increased external transfers partially offset the falling terms of trade, limiting the downturn in growth. Of course, external income from terms-of-trade shocks and external income from transfers (grants or loans) differ in their economic effects and fungibility. * External loans have to be repaid, while grants and export earnings do not. * The windfall of additional foreign exchange from higher com- modity prices can be used freely, while many grants-in-kind, such as food and medicine, do not provide readily available foreign ex- change. Even the financial resources provided under grants and loans are often partially tied. * Loans and grants primarily affect public income, while changes in the terms of trade have a larger impact on private income. In practice, however, it is difficult to draw a sharp distinction between income from terms-of-trade gains and that from grants and loans. The reason: loans are often converted to grants or forgiven, and terms-of- trade changes affect the public sector because marketing boards and trade taxes have generally stabilized the real producer prices of exports. Because of the overlapping nature of these two sources of income, we ex- amine their joint evolution. Declining terms of trade. Most Sub-Saharan economies faced deteriorat- ing terms of trade between 1970 and 1986 (figure 1.7). Oil exporters were the main exceptions, with improvements of more than 100 per- cent. Non-oil exporters suffered losses of about 30 percent, while min- eral exporters were even harder hit, with losses of about 50 percent. 26 WHY AFRICA HAD TO ADJUST [UeL EF~,ajIS a] nnr s Du 2oMal Index: 1970-73 = 100 140 130 Sub-Saharan Africa 120 110 100 * \ , Sub-Saharan Africa excluding Nigeria 90 * ,, , Twenty-four other developing countries 60 Setting aside Nigeria, Africa's terms- 50 of-trade losses resembled those in 40 other developing countries. 1970 197:2 1974 1976 1978 1980 1982 1984 1986 1988 1990 l l Adjustment period Note: Data are means weighted by 1980 GDP in U.S. dollars. Source: World Barnk data. For Sub-Saharan Africa as a whole (excluding Nigeria because of its oil and sheer size), the declining terms of trade meant a drop in external income equivalent to 5.4 percent of GDP between 1971-73 and 1981-86.3 Though large, the loss was spread over twelve years, for an average annual loss of 0.4 percentage points. Adding Nigeria to the pic- ture reduces the twelve-year loss to just 3.3 percent of GDP, or 0.3 per- centage points a year. Again, the losses were biggest for mineral ex- porters (about 1.5 percentage points a year) and more moderate for the agricultural exporters (about 0.3 percentage points a year). Oil exporters gained from higher prices, increasing GDP about 0.3 percentage points a year. Although the worsening terms of trade hindered growth, they were not decisive in Africa's stagnation and decline. If an annual income loss of 1 percentage point from declining terms of trade reduces the annual rate of growth by no more than 0.8 percentage points (Easterly and others 1993), then the falling terms of trade can account for no more than 10 percen-t of the reduction in growth rates between the early 1970s and the imid-1980s. Even for Zambia, probably the region's hard- est-hit country; the terms-of-trade deterioration explains only 1 per- 27 ADJUTMETIN AFRICA centage point of the 3.4 percentage point reduction in annual GDP growth. Moreover, Sub-Saharan Africa was not alone. Most low-income countries elsewhere had similar deteriorations in their terms of trade- around 20 percent on average since the early 1970s-and yet they en- joyed faster growth as a result of better policies. Higher external transfers. Net external transfers increased during the 1970s, in part to compensate for the income losses from the weaken- ing terms of trade. Although the rise in net transfers did not fully offset those losses, it would have been difficult for external sources to have given Africa more assistance. Almost all indicators show that Sub- Saharan Africa already received more assistance than any other region (table 1.1). Net transfers to Sub-Saharan Africa (minus Nigeria) increased from 3.7 percent of GDP in the early 1970s to between 6 and 7 percent of GDP in the late 1970s and early 1980s. By contrast, they remained under 2.4 percent of GDP for the low- and lower-middle- income countries in other regions. Africa was receiving almost $20 per capita in net external transfers, four times the amount going to poor countries elsewhere. The higher external transfers, however, did not always coincide with African countries' needs for foreign exchange. The transfers in the sec- ond half of the 1970s rose with commodity prices, and they fell when commodity prices retreated in the early 1980s. But the average annual Table 1.1 External Capital Transfers for Low-income Countries Adjustment period, Indicator 1971-73 1974-80 1981-86 1987-91 Net transfers (percentage of GDP) Sub-Saharan Africa 2.5 4.3 3.6 4.7 Excluding Nigeria 3.7 7.0 6.4 7.0 Selected other countries' 1.1 2.3 1.7 1.1 Net transfers per capita (U.S. dollars) Sub-SaharanAfrica 5.14 19.18 18.23 16.30 Excluding Nigeria 6.63 22.92 23.11 24.65 Selected other countriesa 1.59 5.37 5.70 4.01 NVote: Net transfers equal net transfers on debt (loan disbursements minus amortization minus interest) plus grants (exduding technical assistance) plus IMF net transfers. a. Twventy-seven non-African countries with up to $1,200 GNP per capita in 1990. Source: World Bank data. 28 WHY AFRICA HAD TO ADJUST flow of net transfers in the late 1970s and early 1980s was about 3 per- centage points higher than in the early 1970s, and the economic decline and stagnation started in the mid-1970s when external transfers were high-about 6 percent of GDP. Moderate income losses overall. The external environment for Sub- Saharan Africa worsened only slightly between the early 1970s and the mid-1980s, because the increase in external transfers partly offset the income loss from deterioration in the terms of trade. If Nigeria is excluded, the decline in the terms of trade caused GDP to drop 5.4 per- centage points between 1971-73 and 1981-86, but the increase in external transfers meant an increase in GDP of 2.7 percentage points (table 1.2). Adding these two effects, the total income loss between 1971-73 and 1981-86 was 2.7 percent of GDP, or 0.3 percentage points a year. From this perspective, the external environment was not much different in the mid-1980s from what it had been fifteen years before (and for the oil exporters, the external environment was better). Meanwhile, Sub-Saharan Africa's GDP per capita growth rate sank 3.2 percentage points (from 2.5 to -0.7 percent). Only a small part of that decline, however, can be attributed to the reduced flows of external in- come (figure 1.8). Recall that an annual income loss of 1 percent of GDP is estimated to reduce GDP growth by 0.4 to 0.8 percentage points a year (box table 1.1). Between 1971-73 and 1981-86, the annual income Ta,'M e I.2 ZeF:s 4 'e h a Eng lauiE ThCOi¶ ea22E F, 197 i-73 to aEDK-se Change in GDP (percentage points) Due to Due to changes in changes the terms in net Item of trade transfersa Total Sub-Saharan Africa Period average -0.3 1.1 0.8 Annual average -0.0 0.1 0.1 Sub-Saharan Africa excluding Nigeria Period average -5.4 2.7 -2.7 Annual average -0.5 0.2 -0.3 a. Net transfers equal net transfers on debt (loan disbursements minus amortization minus interest) plus grants (excluding technical assistance) plus IMF net transfers. Source: World Bank estimates. 29 ADJUSTMNT I:N AFRICA Fi1gure 1.8 Changes in GDP per Capita Growth and External Income in Sub-Saharan Africa, Various Periods Change in growth (percentage points) 10 ES 5 El ~El ES I M~~~~~~~~~~~~~ a 0 a 5 X a 4wE n , la * ' ' 0sl2 a 0 ~ a a -1 -15 -10 -8 -6 -4 -2 0 2 4 6 Change in external income (percentage points) Nlote: This figure plots three changes (between 1965-73 and 1974-80, between 1974-80 and 1981-86, and between 1981-86 and 1987-91) for thirty-nine countries. Source: World Bank estimates. External income shocks had only loss from changes in net transfers and the terms of trade was only 0.3 a small effect on long-term GDP percent of GDP. The expected effect on GDP growth would thus be a de- growth. cline of 0.1 to 0.2 percentage points a year-far less than the 3.2 per- centage point drop Africa actually suffered. Failures to capitalize on better terms of trade. Long-term trends in the external environment played only a secondary role in Africa's long- term decline in growth. They may have had a much larger indirect influence on growth, however, to the extent that they contributed to poor policy choices. African countries' limited capacity to manage terms-of-trade fluctuations often led them to adopt exchange rate, fis- cal, and agricultural policies that undermined growth over the medium term. Paradoxically, positive terms-of-trade shocks were often as costly as negative ones. A handful of countries had faltering growth despite a better external environment. For example, GDP per capita stagnated in 30 WHY AFRICA HAD TO ADJUST Nigeria during its oil boom, when the country was flooded with for- eign exchange from oil revenues and external loans (box 1.2). Why? The windfall financed wasteful current expenditure, ill-advised invest- ment projects, and capital flight. The main problem with a windfall is that governments increase spend- ing as if the higher revenues were permanent. And once increased, gov- ernment spencling is difficult to reduce. Consider the coffee and cocoa booms in C6te d'Ivoire and Kenya during the late 1970s. Current and es- pecially capital spending increased, leveraged by external borrowings. But once the booms were over, revenue could not continue to match the spending increases, and governments were left with large deficits. The high expectations generated by the boom made trimming the public sec- tor politically painful. And increasing revenues to cover debt service obli- gations was difficult because export earnings were dropping. Many of C6te d'Ivoire's macroeconomic problems in the 1980s stemmed directly from expanding the public sector too much during the 1970s boom. Explaining What Econometric Analyses Don't Explain The evidence shows that poor policies clearly hurt Africa's long-term growth far more than a hostile external environment did. But neither policy choices nor external factors explain all of the low growth. Cross- country econometric analysis shows that Sub-Saharan Africa has grown more slowly than other regions even after differences in macroeconomic policies, endowments, political instability, and external shocks are taken into account (Easterly and Levine 1993). Other factors-the bias against agriculture, extensive government intervention in the economy, weak infrastructure, and the difficulties of the political and social transi- tion following independence-have been suggested to account for the residual difference in growth performance. Let us start with the bias against agriculture, perhaps the most im- portant "omitted" variable in econometric analysis. Most Africans live in rural areas, and the region relies heavily on agriculture for foreign ex- change. Yet agr iculture has been heavily taxed, much more than in other regions (Schiff and Valdes 1992). Producers of agricultural exports- typically forcecl to sell their crops to marketing boards that monopolized exports-recei ved real prices half those received by producers of similar crops in other countries (Akiyama and Larson 1989). This stifled the re- gion's agricultural growth. 3I ADJSTENT IN AFRICA Bo Ni §g-ili. `;g;.g,d d1 _1ir,, THE GOVERNMENTS OF INDONESIA AND NIGERIA c'- rici8a 8 2 n collected sizable windfalls (about 20 percent of GDP) ,. , r! M from the increases in oil prices between 1973 and 1981 (Gelb 1988). Spending judiciously, Indonesia 1965- 1973- 1981- 1987- grew steadily. Nigeria, however, saw its GDP per capita Indicator 1972 1980 1986 1990 growth drop considerably after 1973, and by the end Public consumption (percentage of GDP) of the 1980s its real GDP per capita was below that in Indonesia 7.3 9.5 10.9 9.2 1973 (box figure 1.1). Where did Nigeria go wrong? Nigeria 8.4 12.4 15.6 11.6 Investment (percentage of GDP) * In Nigeria large budget surpluses from higher Indonesia 12.8 23.3 28.1 33.9 oil prices gave way to large budget deficits as ex- Nigeria 16.6 26.5 16.5 15.4 penditure (particularly capital expenditure) in- R p 'n- ~~~~~~~~Real effective exchange rate (index: 1972 = 1. 0) creased. In Indonesia, fiscal policy was much Indonesia 1.0 0.8 0.9 1.3 more conservative, due in part to a balanced- Nigeria 1.2 0.8 0.5 1.3 budget law that prohibits government spending Source: World Bank data. from exceeding government revenues, includ- ing official foreign borrowing. * In Nigeria, government consumption almost doubled between 1965-72 and 1981-86 and increase and the subsequent decline in govern- then declined sharply after oil prices plum- ment consumption were not as dramatic. meted in 1986 (box table 1.2). In Indonesia the * Nigeria directed its spending to the cities rather than the countryside. Most agricultural spend- ~~~ox ~~~~~~~~~- ~~~ing went to large-scale, capital-intensive proj- ects with low rates of return. Nonagricultural aiid Higerna projects, such as a new capital city and an inte- Index: 1973=1.0X grated iron and steel industry, also had low re- 22 -.....turns. Indonesia, by contrast, balanced its in- 2.0 vestments among physical infrastructure 1.8 projects, education, agricultural development, 1.6 and capital-intensive industries (primarily fer- 1.4 Indonesia tilizer), directing a high proportion of its re- sources to rural areas. 1.2 * With the collapse of oil prices, Nigeria slashed 1.0 - -- - ,,- igeria investment dramatically, while Indonesia con- 0.8 tinued to devote a larger share of its GDP to in- 0.6:, vestment. The low incremental capital-output OA _____________________________________ _ ;ratios (IcoRs) in Indonesia (4.1 for 1973-81) 1967 1971 1975 1979 1983 1987 1991 suggest that it was much more successful than Oil booms Nigeria (with an ICOR of 30.2) at converting investment into sustained growth in current Source: World Bank data. output. 32 WHY AFRICA HAD TO ADJUST * Niger ia's real exchange rate had appreciated al- percent of Nigeria's exports, compared with most 50 percent by the end of the second oil 50 percent for Indonesia. Nigeria moved to- price increase in 1981, and not until the big ward greater dependence on hydrocarbons, devallaations of 1986 and 1987 did it return to with oil exports rising to more than 90 percent 1973 levels. Indonesia, by contrast, had greater of exports. Only after the big devaluations in real exchange rate appreciation than Nigeria the mid-i 980s did Nigeria have any growth in after the first oil price shock in 1973-74, but it its non-oil exports. Indonesia strengthened devalaed its currency much sooner, with a big and diversified non-hydrocarbon exports dur- devaluation in 1978 that led to further depre- ing and after the oil price booms. With sub- ciation of the exchange rate thereafter. stantial growth in non-oil exports throughout These different policy choices influenced the 1970s and 1980s, oil exports constituted the growth and structure of exports. Before the only about 20 percent of Indonesia's total ex- first oil price shock, oil made up more than 80 ports by the end of the 1980s. Another reason why econometric analyses do not fully explain Africa's low growth may be that easily accessible policy indicators fail to capture the intricacies of policy intervention in the region. African gov- ernments took an active stance in setting prices, nationalizing banks, es- tablishing price controls, rationing foreign exchange, creating public monopolies for agricultural exports, imposing licenses to restrict the ac- tivities the private sector could undertake, and creating many state en- terprises and giving them special access to scarce credit and foreign ex- change. None of these interventions nor their impact is easy to quantify, because the re(quired data are generally either unavailable or unreliable. Moreover, these interventions created a strong bias against the private sector, which also must have contributed to the economic decline. Ad- ministrative bottlenecks, rents from licensing requirements, and ineffi- cient public services imposed high costs on private business. The grad- ual breakdown in judicial systems and the expropriation of private property in many countries in the 1970s left many private enterprises, particularly foreign ones, doubtful about the wisdom of continuing to invest. Some businesses continued only under special contracts or spe- cial trade restrictions that offset the high risks. A third important factor is the extent and quality of infrastructure. Africa's slow growth may be attributable in part to its relatively ineffi- cient transporit system, telecommunications network, and public utili- ties. It is unlikely, however, that the infrastructure base actually deterio- 33 ADJUSMENTIN AFRICA rated during the mid-1970s to early 1980s, which might have ac- counted for some of the decline in GDP growth, because investment in infrastructure burgeoned during this period. A fourth possible factor that econometrics neglects is ideology, which has been important in setting the policy agenda in African countries. Classifying countries as African capitalists, African pop- ulist-socialists, and Afro-Marxists-according to the prevailing ideol- ogy-shows that the capitalists experienced the fastest growth (Young 1982). There were exceptions, of course. Some Afro-Marxist countries (especially Congo) did well thanks to the oil windfall, while some cap- italist countries (such as Zaire) did not, because of political and social instability. A fifth factor that should be considered is the widespread deteriora- tion in governance in the 1970s and 1980s. As countries tried to meet the aspirations for rapid development that were unleashed at indepen- dence, and as they tried to consolidate their political base, they ex- panded the role of the state. Highly authoritarian and highly centralized governments were inimical to the development of local organizations and associations that might have demanded better governance. The costs associated with poor governance extend beyond what is usually captured in policy variables. All these factors no doubt played some role in Africa's poor economic performance. But we can only speculate about their relevance, because they are difficult to quantify and test rigorously. The Needed Switch in Policies THE REFORM PROGRAMS THAT MANY AFRICAN COUNTRIES initiated in the mid-1980s-with the support of the Inter- national Monetary Fund, the World Bank, and other donors- reflected a new paradigm. The reforms attempted to reduce the state's role in production and in regulating private economic activity. They assigned more importance to exports, especially those from the much- neglected agricultural sector. And they placed more emphasis on main- taining macroeconomic stability and avoiding overvalued exchange rates. The process of revamping the policy framework in line with this new paradigm became known as structural adjustment. 34 WHY AFRICA HAD TO ADJUST One of the most fundamental shifts in the development strategy for Africa was to view agriculture not as a backward sector but as the engine of growth-an important source of export revenues and the primary means to reduce poverty. Improving the incentives and the infrastruc- ture services for farmers is now a key element in adjustment programs. On the macroeconomic side, the focus is on keeping inflation low, ex- change rates competitive, and budget deficits sustainable. Furthermore, to enable people to develop efficient and productive businesses, the state is pulling back from direct intervention in the economy and improving its capacity to provide basic services and a stable policy environment. In the next three chapters, we assess the extent of structural adjust- ment reforms in Sub-Saharan Africa. We report on twenty-nine coun- tries with adjustment programs in place during the latter part of the 1980s (see box 1.3). How far have those countries come in reforming their policies? And how much further do they have to go? Chapter 2 fo- cuses on changes in macroeconomic policies, usually tackled early in the adjustment process to eliminate major distortions. Chapter 3 examines progress in tra(le and agricultural reforms, vital to making more efficient and productive use of economic resources. Chapter 4 examines reforms in the areas crucial to a reorientation of the role of the state: the finan- cial sector, the public enterprise sector, and the public sector. These as- sessments provide a basis for examining the payoffs to the adjustment process, the subject of chapter 5. Chapter 6 reviews the evidence on the impact of adjustment on poverty and the environment. The book con- cludes with chapter 7, a look at the road ahead for African adjusting countries. Will the shift in policies enable Africa to catch up? Standard growth theory predicts that low-income countries will grow faster than high- income countries, because they can borrow technologies from the rest of the world and increase the marginal productivity of capital more rapidly than advanced countries. But this requires actively taking advantage of the technology, knowledge, and experience of other nations. And in Africa, the commitment to an outward-oriented development strategy is not yet strong. Even with good policies, catching up is not easy. So far, low-income countries have had less success than middle-income countries in grow- ing quickly. Poor policies explain some of the lag, but poor endowments might also limit the potential for rapid growth in the short term, under- lining the importance of investing in human capital and strengthening 35 ADJUSTMENT IN AFRICA Box t3 Coutity6 oagand Tmee Frame o fthe SU THIS STUDY FOCUSES ON TWENTY-NINE COUN- The text refers to 1987-91 as "the adjustment pe- tries in Sub-Saharan Africa that were undergoing riod," although individual country experience does structural adjustment sometime between 1987 and not always conform precisely to this definition. We 1991 (see box table 1.3). We excluded the very small used 1987 as the starting point because more than half economies, some of which had adjustment programs, the countries had initiated reform programs by then. because there is less information about them and be- We made 1991 the cutoff point because macro- cause external aid disproportionately affects their economic data for all countries were available only macroeconomic performance. We also excluded through that year. A few countries in the sarmple, such countries that did not have adjustment programs be- as Ghana, Kenya, and Malawi, actually launched their tween 1987 and 1991, either because they experi- adjustment programs in the early 1980s and had al- enced great social unrest or civil waXr during most of ready implemented some durable reforms by 1987. those years or because--mainy in the case o coun- Other countries, including C6te d'Ivoire and Zambia, tries in the South African Customs Unon- y had started adjustment in the early 1980s but later reversed a tradition of better policies and were less affected by important reforms; later still, they adopted new pro- the external problems of the early 1980s. Mauritius grams. Several other countries, such as Burkina Faso, was also excluded because it "graduated" from adjust- Rwanda, Sierra Leone, and Zimbabwe, did not initiate ment in the mid-1980s. reforms until very late in the adjustment period. Box Table 1.3 Classification m&` Comtrles Countries not in the study Countries in the study sample pCeunties with (adjusters during 1987-91) Smafllcountries;' civil:unrest Other countries Benin M/ladagascar Cape Verde Angola Botswana Burkina Faso Malawi Comoros Ethiopiab Lesotho Burundi Mvali Djibouti Liberia Mauritius Cameroon : :Mauritania Equatorial Guinea Somalia Namibia Central African Mozambique Sao Tome and Sudan Swaziland Republic Niger Principe Zaire Chad Nigeria Seychelles Congo Rwanda C6te d'Ivoire Senegal Gabon Sierra Leone The Gambia Tanzania Ghana Togo Guinea U-ganda Guinea-Bissau Zambia Kenya Zimbabwe a. With populations under 500,000 in 1991. b. Ethiopia recenty ended its civil war and embarkledon a wde-ranging eform progratn. 36 WHY AFRICA HAD TO ADJUJST institutional capacity. Africa's high rate of population growth-over 3 percent a year--puts additional stress on the limited resources available for improving human capital. Adjustment programs can establish a framework that encourages sound investment and efficient resource use. They are no substitute, however, for the long-term development efforts needed to buildl the capabilities of Africa's people. (See box 1.4 for a dis- cussion of the relationship between adjustment and development.) What is encouraging is that in the 1960s, before their takeoffs, Indonesia, Malaysia, and Thailand had conditions similar to those in Africa in 1990 (table 1.3). GDP per capita was higher on average in Southeast Asia., but Indonesia's income was close to that in Ghana, Malawi, and Tanzania today, and Thailand's was comparable to that of Box I4 Adjustment for Sustained Development ECONOMIC AND SOCIAL DEVELOPMENT IS A Africa's structural adjustment programs, and process of achieving sustainable increases in health, those elsewhere, were almost always mounted in re- education, material consumption, and environmen- sponse to a crisis triggered by external shocks but en- tal protection-in short, improving standards of liv- gendered by poor policies. No economy can function ing over the long term. Structural adjustment con- well for long if it has rampant inflation, an overval- tributes to development by establishing the ued exchange rate, excessive taxation of the agricul- market-friendly incentives needed to put economies tural sector, scarce supplies of needed imports, regu- on sustainable, poverty-reducing growth paths. But lations on prices and production, deficient public adjusting the incentive framework to foster efficient services, and limited financial services. To shore up production and private initiative will not alone bring such economies, the first order of business should be about development. As Sub-Saharan Africa: From fundamental policy reform-restoring macroeco- Crisis to Sustainable Development (World Bank nomic balances and reducing the large distortions in 1989a) points out, development also depends on the incentive framework. This sets the stage for the good governance, appropriate infrastructure and in- more efficient allocation of resources and capital ac- stitutions, and better-trained people. Sound policies cumulation needed for development to take place. and a strong infrastructure create the conditions for Often there is synergy between policy reform and economic growth, but growth will be raised and sus- a country's level of development. Greater technical tained only if human and institutional capabilities and institutional capacity makes it easier to adjust are built up. This requires investments in basic health policies to changing conditions, and stronger adjust- and nutrition, education, and technical skills. It also ment efforts increase the returns on investment and requires that public and private institutions be re- speed up the development process. Moving from a structured to create an environment in which capable vicious circle of bad policies and a low level of devel- people can work effectively, and that political leader- opment to a virtuous circle of good policies and a ship be coimmitted to nurturing those institutions, higher level of development is the challenge facing not to politicizing them for narrow objectives. Africa. 37 A AFRICA Table 1.3 Sub-Saharan Africa in 1990 versus Southeast Asia in 1965: Means for Selected Indicators Sub-Saharan Indonesia, Malaysia, Indicator Africa, 1990 and Thailand, 1965b GDP per capita (constant 1985 international dollars)c 1,030.0 1,320.0 Agriculture share of GDP (percent) 32.0 37.0 Manufacturing share of GD1P (percent) I od 10.3 Savings share of GDP (percent) 16.0 17.0 Investment share of GDP (percent) 16.0 16.0 Export share of GDP (percent)e 29.0 21.0 M2 share ofGDP (percent) 19.4 19.6f Urban share of total population (percent) 29.0 18.3 Primary school enrollment (percentage of eligible population) 68.0 80.0 Secondary school enrollment (percentage of eligible population) 17.0 18.0 Adult illiteracy rate (percentage of population over age 1 5)r 50.0 35.3 Population per physician (thousands)h 23.5 15.0 Infant mortality (number of deaths before age 1 per 1,000 live births) 107.0 90.3 Life expectancy (years) 51.0 52.5 Population growt rate (percent) 3.1' 2.7i a. Data are weighted means for selected countries. b. Data are unweighted means. c. Data are based on resulrs of the United Nations International Comparisons Program (ICP). For details on the calculations, see World Bank (1992d), pp. 299-301. d. Data are for 1989. e. Exports of goods and nonfactor services. f. Indonesia data are fromn 1969. g. Illiteracy is defined here as the inability to read and write a short, simple statement on one's everyday life. h. Physicians lhere include medical assistants who have less training than qualified physicians but who perform similar services, including operations. i. Average annual growth rate from 1980 to 1990. j. From 1964 to 1965. Sources: World Bank (1 992d and 1993d); World Bank data; IMF data. middle-income African countries today. One major difference between Asia then and Africa now: Africa has lower primary school enrollment. This is not inconsequential; emphasis on universal primary education was a key to achieving rapid and equitable economic growth in the best- performing Asian economies (World Bank 1993a). And African coun- tries with a higher percentage of children enrolled in primary school in 1960 grew more quickly over the next two decades than countries with lower enrollment rates (figure 1.9). Africa's runaway population growth also puts it at more of a disadvantage than Asia in the 1960s. With half the population under fifteen years old, Africans of working age face an enormous challenge-bringing about the high rate of economic growth needed to provide a better future for the next generation. 38 WHY AFRICA HAD TO ADJUST Figure 1.9 Education and GDP per Capita Growth Average annual GD1' per capita growth, 1960-85 (percent) 4 Botswana * Gabon Congo E U Swaziland OlLesotho 2 Tanzania U ± Cameroon u * Sierra Leone Uganda P 0 *.. Ethiopia Malawi Zimbabwe . Niger Sudan a African Republic Mozambique ThehGambia N Benin Angola Rwanda I ECte d Ivoire Mauritius E Malit Senegal @ Guinea-Bissau Kenya Zaire m Burundi Chad * M Togo El Zambia -2 Somalia B CaI Ghana -3 Madagascar I -4 Low -<* High Primary school enrollment, 1960 Note: This figure is a partial scattergram based on a regression of primary school enrollment and other factors on GDP per capita growth. Source: Barro (11991). Higher primary school enrollments pay off in higher growth. The importance of reforms for Africa's economic future cannot be overstated. With today's poor policies, it will be forty years before the re- gion returns to its per capita income of the mid-1970s. But a sound de- velopment strategy and a dose of good luck can change the picture. Ghana, for example, has the potential to become a middle-income country by the fiftieth anniversary of its independence, as box 1.5 shows. We should be mindful that the Asian countries that have been stunning the world with their strong GDP per capita growth since 1960 had almost no growth between 1930 and 1960 (Maddison 1989). This does not mean that success will come easily to Africa in the next three decades. It does mean that rapid growth is possible, and that Africa's poor record over the past twenty years is no reason for undue pessimism about its future. 39 U TEINT IN AFRICA THE GHANAIAN ECONOMY STANDS AT A CROSS- 7 7* -- roads. Its adjustment program is one of the most suc- cessful in Sub-Saharan Africa. Since 1983, a decade Human resources of stabilizing policies has yielded broad budget bal- (percentage of population that is literate) ance, strong export growth, a reasonable external po- 100 sition, and substantial structural reforms, including 80 privatization or closure of some loss-making publicly -\60\ owned companies. Even so, real growth has remained Global Q 40 \ Infrastructure 1 about ~~~ a year. And ~integration I ' 2 (number of at only about 5 percenit a year. And although per (ratio of 12010806040 20 8 10 telephones capita income rose abouIt 2 percent a year, ending the foreign trade per 100 1980s at $390, Ghana is still among the world's toGDP) '10 persons) 20 poorest countries. At this growth rate, the average \ poor Ghanaian will not cross the poverty line for an- \ 40 other fifty years. \ o What would it take for Ghana to become a middle- 60 income country by the year 2007, the fiftieth anniver- 70 sary of Ghanaian independence? What would ensure Private sector development sary (banking system credit to private that the average poor Ghanaian would no longer live sector as a percentage of GDP) in poverty by then? In recent years, Ghana's policy- Source: World Bank (1993b) makers have begun to refocus their attention from short-term adjustment issues to longer-term growth and development issues., education. Ghana's literacy rate is about 55 percent, The recent dynamic growth in East Asia shows and its functional literacy rate about 35 to 40 per- what can be achieved with pragmatic government cent-much lower than in many other African coun- policies and a disciplined, hard-working population tries. Ghana needs to focus the bulk of its education that responds to the right incentives. Malaysia and spending on literacy programs and primary educa- Thailand, which were poorer than Ghana in the tion, relying more heavily on private funding for sec- 1960s, managed to double per capita income and re- ondary and tertiary education. In the area of health, duce poverty dramatically in about ten years. They Ghana must redirect resources from hospitals and have now eclipsed Ghana in other measures of devel- curative care to equitable primary health care, pre- opment as well (box figure 1.2). Ghana can profit ventative medicine, and nutrition programs. from East Asia's experience by emphasizing three Openness. Strong export performance has been an areas: education and health, openness to international important factor in East Asia's rapid growth. If markets, and partnership between government and Ghana is to achieve similar success, it must export the private sector. more aggressively, focusing especially on agriculture, Human development. Most successful economies agro-processed products, and light manufacturing. achieve near-universal literacy as a precondition for At 15 percent of GDP, Ghana's export ratio is low rapid growth. While Ghana's secondary and tertiary because noncocoa, nonmineral exports are small. education systems compare favorably with those of What can the government do to push export the fast-growing Asian economies at the start of their growth? It can continue its appropriate exchange rate economic ascent, the same is not true in primary policies, which provide nondiscriminatory export in- 40 WHY AFRICA HAD TO ADJUST centives. It can change production and trade regula- funded liabilities, such as end-of-service benefits; re- tions to facilitate the entry of foreign companies and ducing costly subsidies for water, electricity, and improve tariff levels, export finance, and quality con- transport; and improving revenue mobilization trol. It can promote better export infrastructure (bet- through a value-added tax. At the same time, spend- ter telecommunications, warehousing, and refrigera- ing (in such areas as primary education and health, tion). And because labor costs are an important transport infrastructure, and research and extension) factor in determining the competitiveness of exports, will have to increase considerably to accelerate Ghana can work to link wage bargaining more growth. For this reason, Ghana must not only im- closely to labor productivity. prove financial management of the public sector but Working with business. Ideally, governments also encourage private investment. should have a clearly defined role and a complemen- Private sector confidence and investment have tary relationship with the private sector. East Asian been undermined by a web of old controls and regu- governments have geared their spending to promot- lations, a lack of transparency in enforcing those reg- ing and not competing with the private sector. ulations, and continued government ownership of Moreover, because of vigorous revenue mobiliza- production activities, despite the government's stated tion, the public sector has been a net saver since the intention of leaving those activities to the private sec- 1960s, rarely crowding out the private sector. tor. Ghana is, however, already making serious ef- Ghana's spending has been less prudent, and, al- forts to improve the climate for business-pushing though fiscal policy is improving, there is scope to do for clearer business procedures, due process, and far better. I'he public sector deficit could be reduced speedy divestiture. by cutting transfers to state enterprises; reducing un- Source: World Bank (1993b). Notes 1. The countries of the CFA franc zone, which have 3. Estimating how changes in the barter terms of trade fixed exchange rates pegged to the French franc, have no affect income is complex, and alternative methods may be parallel markets for foreign exchange. applied. We calculated the average terms-of-trade index (using the barter terms-of-trade index) for the periods 2. The index of outward orientation measures the ex- 1970-73 and 1981-86. The "income effect" of changes tent to which the real exchange rate is distorted away in the terms of trade between the two periods was then from its free-trade level by the trade regime. The high- calculated by multiplying the ratio of exports of goods performing Asian economies are Hong Kong; Indonesia; and nonfactor services to GDP in 1987 by the percentage Japan; the Republic of Korea; Malaysia; Singapore; change in the average terms-of-trade index between Taiwan, China; and Thailand. 1970-73 and 1981-86. The "income effect" thus is a measure of the share of GDP gained or lost as a result of terms-of-trade changes. 41 A CHAPTE]R 2 Movring Toward Sound Macroeconomic Policies I HE MACROECONOMIC SITUATION IN SUB-SAHARAN Africa reached crisis proportions in the early 1980s. External imbalances mounted as real export revenues plummeted and imports remained unchanged. Currencies were wildly overvalued, with the parallel market premiums for foreign exchange exceeding 100 percent. Budget deficits soared to more than 7 percent of GDP, and because of the debt crisis, many countries lost their access to com- mercial lending. African countries had to reestablish a balance between income and spending to improve the balance of payments. This required a tighten- ing of fiscal and credit policies and a depreciation of the real exchange rate. Tight fiscal and credit policies cut overall spending in the economy, while depreciation expanded production in the tradable sector and eased the recessionary impact of the tighter demand policies. Macroeconomic packages of this type worked for Costa Rica, Indonesia, the Republic of Korea, and other developing countries facing similar imbalances during the early 1980s. In Africa, Ghana and Mauritius were among the first to take bold steps along these lines. But in the first half of the decade, other countries did not do enough to reduce their budget deficits, and they were reluctant to depreciate their currencies to restore competitiveness. The failure to take drastic action early was coupled with a series of in- appropriate responses that merely deepened the crisis. Most countries with flexible exchange rates failed to devalue and to contract demand. Instead, they imposed foreign exchange controls and intensified the use of import licenses to allocate their increasingly scarce foreign exchange. The new restrictions deprived domestic firms of intermediate inputs and spare parts, strangling growth and worsening the allocation of re- 43 UTMENT IN AFRICA sources. Such moves locked many of these economies in a vicious circle. Increases in the parallel market premium for foreign exchange led to fur- ther deteriorations in the recorded trade balance, forcing the authorities to impose even tighter import restrictions. Shrinking exports and wors- ening export prices (in domestic currencies) reduced official export earnings, further increasing the parallel market premium. Growth suf- fered along the way, as producers lost access to the imported inputs they needed to keep production going. GDP per capita in countries with flex- ible exchange rates declined at 1.5 percent a year in the first half of the 1980s, after stagnating in the second half of the 1970s. Forced to look for better alternatives, most of them started reform programs in the mid-1980s-programs that combined better macroeconomic and sec- toral policies with large increases in external financing. The macroeconomic crisis was different for countries with fixed ex- change rates.2 They suffered balance-of-payments and fiscal problems similar to those in the countries with flexible exchange rates, but they never restricted access to foreign exchange.3 They continued to abide by the rules of the CFA franc zone and maintained fully convertible curren- cies. Their crises were less severe than those in other African countries because exporters benefited from the real depreciation of the French franc-the peg for their currencies-against the dollar in the first half of the 1980s. Nonetheless, the fixed exchange rate countries saw their GDP per capita growth plummet-about 3 percentage points between 1974-80 and 1981-86-and their real imports shrink by about 25 per- cent. C6te d'Ivoire's crisis was particularly severe, with GDP per capita growth collapsing from 3.5 percent a year in the late 1970s to -5 per- cent in the first half of the 1980s. The adjusting Sub-Saharan countries generally improved their macroeconomic policies in the second half of the 1980s. Tighter fiscal and credit policies, accompanied by increased foreign financing, helped improve the balance of payments and move economies out of the import-compression phase of adjustment. The policy packages have been most effective when they controlled inflation and brought about the much-needed depreciation of the real exchange rate. Despite the improvements, none of the countries in the region has yet achieved a good macroeconomic stance. In what follows we analyze the three key components of macroeconomic adjustment programs-monetary, fis- cal, and exchange rate policies-and we examine the overall policy packages.4 44 MOVING TOWARD SOUND MACROECONOMIC POLICIES Fiscal Deficits Down-But Not Out M OST AFRICAN COUNTRIES STARTED THE 1980S WITH large fiscal imbalances from high government spending- the legacy of commodity booms in the late 1970s-and declining trade tax revenues because of the collapse of commodity prices in the early 1980s. Budget deficits (including grants) in excess of 7 percent of GDP were the norm. Because high budget deficits usually mean rapid money growth, high inflation, and large current account deficits, this was hardly the basis for a sound macroeconomic climate (Easterly and Schmidt-Hebbel 1991; Kiguel and O'Connell 1993). Budget Deficits A low budget deficit is not a sufficient sign of good policy, but it is certainly a necessary component. More than half of the African adjust- ing countries reduced their overall budget deficits between 1981-86 and 1990-91. Other measures of fiscal health, such as the primary deficit, provide a similar picture. The overall fiscal deficit including grants indicates how much the government would have to borrow to achieve fiscal balance. Because most countries have limited access to domestic and foreign financing, the overall deficit measures the potential risks of resorting to inflation- ary finance or financing deficits domestically in other distortionary ways (such as incurring arrears with government suppliers or taxing the fi- nancial sector). Based on this indicator, the adjusting countries in the region improved their fiscal performance, reducing the median overall deficit including grants from 6.4 percent of GDP in 1981-86 to 5.2 per- cent of GDP in 1990-91 (figure 2.1). But the median improvement dis- guises differences among country groups (appendix table A.2). Five of the six countries with the largest reductions (more than 5 percentage points)-Burundi, The Gambia, Malawi, Tanzania, and Zambia-are low-income, flexible exchange rate countries. In contrast, the middle- income countries, those with fixed exchange rates, and the oil exporters had their overall fiscal balances deteriorate, with Cameroon and C6te d'Ivoire having the largest increases in deficits. Another measure of the fiscal balance is the primary deficit, calcu- lated by deductring interest payments from total expenditure. Change in the primary deficit is considered a better indicator of fiscal performance 45 ADUTMN:N AFRICA FG'gure 2.1 Flsczi M .t:s ic'' iC-`Ti 'ii, Ad>iSJrCg zntE'r fmedians) Fiscal Deficit Percentage of GDP 8 - ---- .. . .... . 1981-86 6 1990-91 4~~~~~~~~ Percentage of G 4 - :.................. 30 G G G G|;0tttt t40 U g1981-86 Overall deficit including grants Primary deficit including grants Total Revenue and Expenditure Percentage of GDP 30 1981-86 1990-91ti i ;X; 1990-91 20 1981-86 1990=91 0'44'~'t00 |18 Total revenue redutions were primai d A Composition of Total Expenditure Iercentage of GDP Sorcs ^--b stf esiats NahsibGlGggi audBazoni(194) 46~~~~~~~~~1 12 | Gi-p!000000000000. | lllll§ggggggggggggggggggg 1981 19 1-8 8 1 lESlElElElESEtEtEllEEti1990-91 1990-f91 4 1990-91 04 1 - Deficit reductions were primarily due o|-- Interest payments Current expenditure Public wages and Capital excluding interest salaries expenditure payments lVote: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. Sources: IMF staff estimates; Nashashibi and Bazzoni (I1994). 46 MOVING TOWARD SOUND MACROECONOMIC POLICIES than change in the overall deficit, because fluctuations in external inter- est payments are largely beyond government control in the short term. The primary fiscal deficit as a share of GDP improved in the adjusting countries, with a median decrease of almost 2 percentage points. This was more than the decrease in the overall fiscal deficit, because interest expenditures rose in most countries and partly offset efforts to reduce the overall deficit. Again, as with the overall fiscal balance, improve- ments in the primary fiscal balance were larger for the low-income countries and those with flexible exchange rates than for the middle- income countries and those with fixed exchange rates. Despite the deficit reductions, the fiscal situation in Sub-Saharan Africa is still fragile. Most countries still rely heaviLy on grants to avoid fiscal im- balances, and the median deficit excluding grants has remained large- about 8 percent of GDP-since the 1980s. Indeed, the 1.3 percentage point increase in external grants between the early 1980s and 1990-91 con- tributed heavily to reducing the overall budget deficit during this period. Expenditure Fell, But Revenue Increased Only Slightly The reduction in budget deficits was helped by lower spending, with a median decline of 1 percentage point of GDP, despite an increase in in- terest payments of about 0.9 percentage points of GDP. Most of the cuts were in capital spending; the median fell from 8.7 percent of GDP in the early 1980s to 6.1 percent in 1990-91. This was unfortunate, because the level of capital spending was not excessive in most countries, al- though its composition needed improvement. The cuts were deeper for the middle-income and the fixed exchange rate countries, and deepest for the oil exporters-not surprising because they started with bloated public investment programs. The overall public wage bill generally fell less than 1 percentage point, as inflation erocded public sector wages. But in the middle-income and oil- exporting countries, which expanded employment and kept real wages high, median spending on wages and salaries rose 3 to 4 percentage points. The five countries with the largest increases in the wage bill as a share of GDP (Crameroon, Congo, Cote d'Ivoire, Gabon, and Zimbabwe) are all middle-income countries, and three of them are oil exporters.5 Ideally, spending cuts by African governments would have been rein- forced by revenue increases. But in more than half the countries, gov- ernment revenue did not increase. For all the adjusting countries, the 47 ADJUTMET IN AFRICA median tax revenue as a share of GDP fell 0.5 percentage points. Revenue losses were particularly large-as much as 9.8 percent of GDP-in the oil-exporting countries with fixed exchange rates (Cameroon, Congo, Table 2.2 flsca Pcl$ciimz and Gabon). The decline in oil prices and the appreciation of the real 1990-91 exchange rate reduced their revenue from trade taxes, and fixed ex- Good or adequate change rate countries as a whole were hard-hit by deep recessions. Nige- The Gambia ria, the only oil exporter with a flexible exchange rate, avoided such rev- MaurGtania enue reductions by effecting a large real depreciation that more than Senegal offset the decline in oil prices. Low-income countries with flexible ex- Tanzania change rates did better as a group, because trade taxes increased and eco- Fair nomic growth picked up. Ghana and Tanzania did particularly well, but Burundi Madagascar, Uganda, and Zambia suffered revenue declines largely be- Burkina Faso cause of the erosion in the real value of tax revenues associated with Gabon higher inflation.6 Malawi highe Togo Poor Fiscal Stance Benin ** Central African Rep. ** The criterion for classifying countries by fiscal stance was the overall Kenya ** budget deficit (on an accrual basis including grants), which shows the Madagascar *amount of financing required.7 Countries with high budget deficits Mali *aon Nigeria ** after grants generally need to improve their fiscal accounts to avoid in- Rwanda i* flation or crowding out domestic investment. While high budget Uganda i* deficits are not necessarily inflationary in the short term-as they can be Very poor financed directly or indirectly, domestically or abroad-they are likely Cameroon to be inflationary in the longer term if they are sustained. C'te d'Ivoire * Based on the size of the total fiscal deficit including grants, all countries Mozambique * still need more fiscal adjustment-and in some instances considerably Niger * more. Only five countries had a good or adequate fiscal stance in 1990-91 Sierra Leone (G Zambia i (table 2.1), and two of them-The Gambia and Tanzania-rely heavily Zimbabwe i on grants. This shows how fragile public finances are in Africa, and how ,ote: A "good or adequate" rating quickly a fall in external assistance could destabilize them. indicates a budget surplus or an overall budget deficit (including grants) of less than 1.5 percent of GDP; "fair," a deficit of 1.5-3.5 Monetary Policy Mostly on Track percent; "poor," a deficit of 3.6-7.0 percent; and "very poor," a deficit of 7.1 percent or more. Chad, Guinea, and Guinea-Bissau are excluded HE MAIN GOALS OF MONETARY POLICY ARE TO MAINTAIN because of insufficient data. low rates of inflation and suitable levels of economic activity Source: Appendix table fS._ . through adequate real interest rates. In Africa during the 48 MOVING TOWARD SOUND MACROECONOMIC POLICIES adjustment period, there was no clear trend in the evolution of mone- tary policy. Median inflation-generally low to begin with-fell a cou- ple of percentage points, with wide variation across countries. There was also progress in reducing (or eliminating) the highly negative real interest rates that discourage savings. But in some countries, the choice Goouorkadequate of poor or inconsistent macroeconomic policies resulted in highly posi- Burundi tive real interest rates, which are not conducive to growth. Central African Rep. **** We analyzed the monetary stance in African adjusting countries C6te d'Ivoire .. . . ~~~~Gabon .** based on three indicators: the revenue from printing money (seignior- Malib age), the rate of inflation, and the real interest rate. With few exceptions, Fair monetary policies in 1990-91 were fair or better (table 2.2). Inflation Benin was not a major problem in most of the economies, and seigniorage was Cameroon low compared with other developing countries, but there was room for Congo improvement in interest rate policy. The Gambia Ghana Kenya Seigniorage and Inflation Madagascar Malawi Governments, through their central banks, obtain revenue from Mauritania their exclusive privilege of printing money. That revenue, expressed as Nigeria a share of GDP, is called seigniorage.8 The main advantage of using Rwanda seigniorage as an indicator of monetary policy-as opposed to other Senegal common indicators, such as money growth-is that it tells whether Uga central banks are printing money to finance budget deficits. The higher Zimbabwe the seigniorage, the larger the resources that governments gain from Poor printing money, and thus the more inflationary the monetary policy. Mozambique ** But higher seigniorage does not always mean a higher inflation rate, es- Tanzania ** pecially in the short run. It is not inflationary if it accommodates an in- Very poor crease in the demand for money, if it is mainly transitory, or if there are Sierra Leone * lags in the transmission of increases in the money supply to prices. In Zambia * the longer run, however, higher seigniorage means higher inflation. As Note: Ratings range from 'good a general rule, seigniorage of more than 2 percent of GDP is risky, be- low seigniorage (iess than 0.5 per- cause the economy will eventually have high inflation; seigniorage of cent of GDP), low inflation (less more than 3 percent for several years indicates large macroeconomic than 10 percent), and reasonable imbalances.9 interest rates (-3 to 3 percent), to imbalances. 'very poor," which means high Most Sub-Saharan African countries have had low or moderate seigniorage, inflation over 100 per- seigniorage, consistent with the fact that inflation has also been rela- cent, andhigh orextremely negative tively low (appendix table A.3). The median rate of inflation was just Guinea-Bissau are excluded because 10.6 percent before the adjustment period, certainly lower than in Latin of insufficient data. America and no higher than in other developing countries. By 1990-91, Source:Appendix table B.5. 49 ADJUSTMENT IN AFRICA it had dropped to 8 percent. Even among the flexible exchange rate countries, which have had the highest inflation rates in the region, the median has hovered around 20 percent-high by international stan- dards, but not a sign of major macroeconomic imbalance. The fixed exchange rate has been a useful discipline mechanism. Nine of the ten countries with seigniorage below 0.5 percent of GDP in 1990-91 had fixed exchange rates, and their median inflation was just 0.6 percent. Some countries with high seigniorage have not always had high inflation (though they might be headed in that direction if the high seigniorage persists).10 Only three countries-Sierra Leone, Tanzania, and Zambia-had high seigniorage and high inflation in 1990-91, showing that their monetary policy was poor and improvement clearly needed. Interest Rates High real interest rates usually mean that monetary (or credit) policy is tight and that the monetary authorities are trying to cool off the econ- omy. Low real interest rates, by contrast, generally indicate expansionary monetary policy. Real interest rates that for long periods remain high and negative-or high and positive-are a problem. The highly nega- tive ones indicate that inflation is high and depositors heavily taxed, hurting domestic savings. The highly positive ones suggest that credit policies are too tight and that there is excess demand for credit-or that a high rate is needed to stem capital flight because people expect a de- preciation. Highly positive rates are riskier for the financial soundness of banks; forced to lend at high real rates to remain profitable, they assume more risk than prudent financial practices would warrant. Interest rates in Africa have limited value as an indicator of monetary policy because they are not market-determined-financial markets are thin and the government sets the rates. Interest rate liberalization and a decline in inflation have helped eliminate extremely negative real inter- est rates (appendix table A.4), though interest rates continue to be neg- ative in a few countries, particularly Sierra Leone and Zimbabwe. In contrast, eleven countries-ten of them with fixed exchange rates-had high real interest rates in 1990-91 (above 3 percent). Those elevated rates were the result of tight monetary and credit policies needed to re- strain aggregate demand, to support the fixed exchange rate, and to avoid capital outflows. 50 MOVING TOWARD SOUND MACROECONOMIC POLICIES Mixed Progress in Exchange Rate Policy M OST SUB-SAHARAN AFRICAN COUNTRIES REQUIRED A real depreciation to compensate for the worsening terms of trade in the 1980s. In addition, many countries started with large premiums in the parallel foreign exchange market and needed massive devaluations of the official exchange rate. Countries with flexi- ble exchange rates-which either devalue from time to time or have a crawling peg or a managed float-can make real depreciations quickly. But countries with fixed exchange rates find rapid depreciations more difficult because prices are inflexible downward. By and large, coun- tries with flexible exchange rates have made significant progress in increasing their international competitiveness, while those with fixed rates are still struggling to make the much-needed real depreciation- in part because fiscal policies were not supportive (box 2.1). Different exchange rate regimes call for different methods of assessing exchange rate policy. For countries with flexible exchange rates and non- convertible currencies, the parallel market premium is a good indicator of exchange rate policy. For countries with fixed exchange rates and con- vertible currencies, by contrast, the parallel market premium is very small or nonexistent, even though the real exchange rate may be misaligned. We thus need other methods to assess the external competitiveness of these countries. Our strategy was to compare them with other countries that have similar exchange rate regimes and faced similar external shocks. Flexible Exchange Rate Regimes: The Parallel Market Premium Fell and External CDompetitiveness Quickly Improved Adjusting countries with flexible exchange rate regimes entered the adjustment period with markedly overvalued exchange rates and high parallel market premiums for foreign exchange. They defended their in- ternational reserves through foreign exchange rationing and trade re- strictions. RecLucing the high premiums, an important part of adjust- ment prograrns in these economies, reduces implicit taxation on exports, curbs rent-seeking activities, and lessens administrative discre- tion in the allocation of foreign exchange.1 ' Many African countries with flexible exchange rates traditionally had steeper premiums than other developing countries. Of the non-African adjusting countries covered in the World Bank's most recent report on 5S ADJUST NT:I TN AFRICA Box 2.1 Diffee6ceshn MnetayandFipApoici : for Fixed and Flexible E cfiage Aiiie es THE MACROECONOMIC ADJUSTMENT STRATEGY Sa y F igu 2.1 j( (2iE'3 i&aSJCy 4ShrsrS differs under fixed and flexible exchange rate StComarlrg Views 1s i trias matfh FfxeS and regimes. Countries with fixed exchange rates need to F$ xle xi chble his rely on tighter credit and fiscal policies as a way of reducing domestic demand and improving the bal- Percent ance of payments (sometimes called the internal ad- :10 justment" strategy). Thiese policies also put down- fixedexchange rateh ward pressure on domestic prices and real wages to bring about a real depreciation. The flexible ex- 0 change rate countries, by contrast, support tighter demand with devaluation. Countries with 0; 0 = iW, tc 0 0 0 po0g; EhV : : 0 ~~~~~~flexible exchange rates While both strategies result in a tightening of de- -10to ti ea mand, countries with fixed exchaLnge rates need to do more, especially in reducing budget deficits. But -15 10 198 9 9 1 1 while they adopted a policy of tight money, re- flected in high real interest rates andi low seignioragee PO a of GDP Seigniorage- (box figure 2. 1), they did not do enough in fiscal 4 policy. Their deficits irn 1990-91 were higher than Countries with those in the flexible exchange rate countries in the flexible exchange rates region. They reduced government sdinb ut 2- not enough to offset the falling revenus(boxfigure i 2.2). 0 Countries with flexible exchange rates also tight- C ened credit, but they genierally followed looser monie -fiedexhange rate tary policies and had higher seigniorage and inflation. -2 They also had better fiscal policies. Their median 1980 1982 1984 1986 1988 1990 overall budget deficit including grants fell from 7.3 Dofestc Credit percent of GDP in 1981--86 to 4.5 percent of GDP in Perentage of GDP 1990-91, while the median deficit for the fixed ex- trie34s wit change rate countries increased from 6.0 to 6.3 per- 32 flexible exchange rates cent of GDP. Revenues explain much of the difference: 30 median revenue rose from 18.4 to 19.7 percent of 28 GDP in the flexible exchange rate countries; it dropped 26 from 18.4 to 16.3 percent of GDP in the fixed ex- 24 Countre w : change rate economies. Both groups managed to re- 22 fixed exchangrates duce total spending, but the fixed exchange rate coun- 20 tries used less desirable mneans to do so: they increased 1980 1982 1984 1986 1988 1990 the wage bill slightly and cut capital spending, while Adjustment period the flexible exchange rate countries relied more heav- a. Based on Ml. t0fily on cuts 0in: the wage bAill.00000;E0 Source::00000Xi000 SaX0000World Bank data. 52 MOVING TOWARD SOUND MACROECONOMIC POLICIES Overall Deficit Including Grants Primary Deficit Including Grants Percentage of GDP Percentage of GDP 8 198B6 8 6 1981L46 1990-91 6 1990-91 4 L:: ----- 4 1981-86 1981-6 2 r 2., p | . r4l1990-91 1990-91 0 0 Countries with fixed Countries with flexible Countries with fixed Countries with flexible exchange rates exchange rates exchange rates exchange rates Total Revenue Total Expenditure Percentage of GDP Percentage of GDP 30 30 1L981--86 8-61909 25 25 . ..... 1990-91 20 1981-86r1914 1 990 91 20 15 ~~~1990-91 -- Countries with fixed Countries with flexible Countries with fixed Countries with flexible exchange rates exchange rates exchange rates exchange rates Capital Expenditure Public Wages and Salaries Percentage of GDP Percentage of GDP 15 . 15 1981-86 10 10 1981 896 1990-91 1990-91 198186 1 990-9 198146 5 - ---- 5 1. 990-91 0 ~ats0 ' Countries with fixed Countrles with flexible Countries with fixed Countries with flexible I exchange rates exchange rates exchange rates exchange rates Source: IMF data. 53 jADJUSTMENT IA FRICA adjustment lending (World Bank 1992a), few had the three-digit pre- miums of African countries, and for those few, the episodes were shorter than in Africa. After the mid- 1 980s, African adjusting countries made good progress in reducing the parallel market premium, and by 1990-91 none had the four-digit premiums of the early 1980s (appendix table A.5). Moreover, devaluations have not led to a permanent increase in inflation (box 2.2). @X$00k0~~isNed Not B Infationery WHY HAVE DEVALUATIONS IN THE FLEXILE EXCHANGE RATE countries changed relative prices without causing inflation in the medium term? There are at least four reasons. First, most countries needed a real depreciation to restore external balance, in part because they experienced a significant deterioration infthe terms of trade. Underthes circumsta nces, the devaluations heped move relativeprices toward their equi uand so were not Second,:ral depreciations generally had a faorable impact on the budget in domestic currency terms, increasing revenue more than spenineg. Thlisis party because devaluation increases the domestic urchasing power of external grants, an important source of revenue. VfAnother major source of government income, trade tax revenue, also rises as imprts become more expensive in domestic currency terms. And there are often gains in export revenue, especially in countries thaSt export mineral :resources-such as Nigeria and Zambia-and wher the state is in partnership with the mining companies or is able to taox dthewindfall gain fromthe real depreciation. Third, much of the potential impact of a devaluation was absorbed eary i th paalll mrkt by the depreciation of the parallel ex- change rate. Prices of mny tradable goods had increased as goods were smuggledinfromneighgcount res. So. :when the govern- ment finally devalued,domestic pricesf had already adjusted to the new"equilibrium"intheoffial exchange rate, andthere were no further inflationary pressures. ;?:: io: Fourth,d most countie in hei region d;o not have a tradition of high inflation,so, unKlkemany Latin American countries, they do not have n place the wage-indexing mechanisms that set in motion the vic£iousinflation-devaluation cycle. This pattern might explain why devaluations increased inflation where inflation was high to start with (Sierra :Leone andZaire)but not where it was low. 54 MOVING TOWARD SOUND MACROECONOMIC POLICIES Despite this progress, the median premium for African adjusting coun- tries was about 25 percent in 1990-91, four times that for adjusting countries in other regions. Not surprisingly, countries that reduced the premium significantly also had large real depreciations. The average ex- change rate premium fell from 300 percent in 1981-86 to 46 percent in 1990-91, with the median falling from 60 percent to 25 percent. Ghana and Mozambique made particularly impressive reductions (from 1,098 percent to 3 percent and from 2,111 percent to 63 percent, respec- tively), as did Guinea, Nigeria, Tanzania, and Uganda. However, be- cause these couantries had extremely high premiums in the first place, even these large reductions were not always enough to achieve good ex- change rate policies. Table 2.3 Exchange Rate Pekcy There were exceptions to the general progress in exchange rate policy. Stance Yor Countries Alth FS:eH' Zambia had a reversal when poor macroeconomic policy decisions led Exchange Rates, C990-S5 to an increase in the premium-from about 45 percent in the mid- Good or adequate 1980s to 600 percent in 1988.12 The premiums also rose in Mauritania Ghana and Sierra Leone after the mid-1980s, as macroeconomic instability in- Guinea creased. In Bu.rundi and Kenya, which had low or moderate premiums Guinea-Bissau at the outset, ithe changes were not as dramatic because exchange rate Madagascar policy remained fairly sound. Fair Reductions in the premium were accompanied by substantial in- Burundi creases in external competitiveness during the 1980s. The median real The Gambia effective exchange rate depreciated by 78 percent between 1981-86 and Malawi 1990-91. Even countries that entered the adjustment period with lower Nigeria Uganda premiums-such as Burundi, Kenya, and Madagascar-depreciated Zimbabwe their real exchange rates (by around 60 percent on average). Poor Most countries with flexible exchange rates had low or moderate pre- Rwanda ** miums in 1990-91, and thus fair to good exchange rate policies (table Very poor 2.3). This certainly is an improvement over the mid-1980s. But about a Mauritania * third of them still had a premium of more than 30 percent in 1990-9 1, Mozambique * indicating room to do better.13 Sierra Leone * Tanzania * Zambia * Fixed Exchange Rate Regimes: No Improvement in External Note: A "good or adequate" rating Competitiveness indicates a premium of 0-10 percent on the parallel market for foreign In the second half of the 1980s, adjusting countries with fixed ex- exchange; "fair," a premium of In ~~~~~~~~~~~~~~~~~~~~~~~11-30 percent; "poor," a premium change rates became less competitive internationally as their terms of of 31-50 percent; and "very poor," trade worsened. It is difficult to assess their policy stance because there a premium of 51 percent or more. is no parallel foreign exchange market to establish the misalignment of Source. Appendix table 1.5. 55 ADJUSTMENT IN AFRICA the real exchange rate. Looking at changes in the real exchange rate is no help without knowing what the new equilibrium rate would be, given changes in the terms of trade, the availability of external financing, and so on. Adjusting African countries with flexible exchange rates are not a useful benchmark, because they entered adjustment with grossly over- valued exchange rates. Instead, we looked at the behavior of the real effective exchange rate in a reference group of developing countries outside Africa-countries that exported primary products but did not enter the 1980s with high parallel market premiums.14 Their real exchange rates depreciated by 60 percent on average between 1980 and 1990, strongly suggesting that the African economies needed to depreciate by at least that much to remain competitive in international markets.15 But the real depreciation in Africa's fixed exchange rate economies-just 5 percent on average be- tween 1980 and 1990-91-fell well short of that mark (figure 2.2). A few countries even had a real appreciation (appendix table A.6). The depreciation of the dollar against the French franc that started in 1985 was the wrong thing at the wrong time for countries in the CFA Countries with fixed exchange ratesg gg achieved less real depreciation than franc zone. The lower inflation in France also exacerbated the problem, other developing countries, despite because the only way the franc zone countries could achieve a real depre- similar falls in the terms of trade. ciation in a reasonable time without devaluing was by reducing domestic African Adjusters witli Fixed Exchange Rates Thirty Other Developing Countries Index: 1980=100 Index: 1980=100 180 180. 160 160 Real effective exchange ratea 140 140 120 ^ ~ - Realeffective exchange ratea 120 100 -100 Terms of trade 8 T o 80 so__ > 8 _ _em ftrd 60 60 1980 1982 1984 1986 1988 1990 1980 1982 1984 1986 1988 1990 l l l Adjustment period Adjustment period a. An increase in the real effective exchange rate constitutes a depreciation; a decrease constitutes an appreciation. Source: World Bank data. 56 MOVING TOWARD SOUND MACROECONOMIC POLICIES prices. Their failure to depreciate as much as other countries exporting 2 . similar products meant a loss of competitiveness. Ten of the twelve fixed ,. L exchange rate countries with data available entered the 1990s with poor e-- or very poor stances on exchange rate policy (table 2.4). Good or adequate None Fair Further Progress Needed in Niger Macroeconomic Policies Poor Benin ** Burkina Faso ** OW MvIUCH HAVE MACROECONOMIC POLICIES CHANGED Central AfricanRep. ** overall? This is a difficult question to answer. There are clear Mali ** problems in constructing a single measure of all the changes Togo ** in fiscal, monetary, and exchange rate policies. Is a reduction in infla- Verypoor tion more or less important than one in the parallel market premium? Cameroon * Are there significant gains in the overall policy environment if the pre- Congo mium falls to low levels but inflation remains high (say, close to three Senegal * digits)? Are there net benefits to reducing inflation from 20 percent to Negal Note: A 'good or adequate" rating 5 percent if this causes an appreciation of a grossly overvalued domes- indicates a depreciation in the real tic currency? Does reducing the budget deficit by 3 percentage points effective exchange rate of more than have the same impact regardless of whether the initial deficit is 15 per- 40 percent between 1980 and 1990-9 1 (non-African countries cent of GDP or 4 percent of GDP? averaged 60 percent). A "fair" rating Although no method can address these complex questions satisfacto- indicates a depreciation of 21-40 percent; "poor," a depreciation of rily, we created an aggregate index that, while imperfect, summarizes 6-20 percent; and 'very poor," a changes in fiscal, monetary, and exchange rate policy. Details about the depreciation of 0-5 percent or an construction of the index and alternative ways of calculating it are pre- appreciation. Chad is excluded because of insufficient data. sented in apperndix B. Source: Appendix table B.5. By and large, the adjusting countries in Sub-Saharan Africa have im- proved their macroeconomic policies, increasing their international competitiveness and reducing inflation (figure 2.3). Of the twenty-six countries for which we could compute the index of change in overall macroeconomic policies, fifteen made advances. Ghana, Tanzania, and The Gambia improved the most, reflecting in part highly distorted ini- tial conditions. But eleven countries suffered deteriorations, with Cameroon, Gabon, Cote d'Ivoire, and Congo having the largest. Macroeconomic policies worsened slightly in Benin, the Central African Republic, Rwanda, Sierra Leone, and Togo. Even when different time periods or different policy variables are used to evaluate the change in macroeconomic policies, the evidence 57 ADJUSTMENT :N AFRICA More than half of the countries im- Figure 2.3 Change in Macraeconormh FoLice.s5, 103h-2 ;t '-o&7 proved their macroeconomic policies. Ghana Tanzania Large The Gambia Improvement tEnla 2.5 Couniries Ranked by Burkina Faso Nigeria uiverail MacroeconomhCOicY Zimbabwe sliance, ff 70-92 Madagascar Malawi Small Adequate Burundi Improvement 1 Ghana Kenya Mali Fair Mauritania 2 Burundi Senegal 3 The Gambia Uganda 4 Madagascar 5 Malawi Benin 6 Burkina Faso Central African Republic Rwanda Deterioration 7 Kenya Sierra Leone 8 Gabon Togo 9 Mauritania Zambia Mozambique 10 Nigeria Congo 11 Senegal *C6te d'lvoire 12 Togo Cameroon Gabon __ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 13 Mali 14 Uganda -2.0 -1.0 0.0 1.0 2.0 3.0 Index scores Poor 15 Central African Nlote: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. Republic ** Source: Appendix table B. 1. 16 Niger ** 17 Benin ** 18 Rwanda ** clearly points to progress in more than half of the adjusting countries 19 Tanzania ** (see appendix B). Burkina Faso, The Gambia, Ghana, Nigeria, Tanza- 20 Zimbabwe * Vegy poor nia, and Zimbabwe consistently show large macroeconomic improve- 21 C6te d'Ivoire * ments, while Cameroon, Congo, C6te d'Ivoire, Gabon, and Togo al- 22 Cameroon * ways appear among the group with deterioration in policies. And 23 Congo * although some countries shift groups, as should be expected, this does 24 Mozambique * not alter the finding (presented in chapter 5) that better policies are as- 25 Sierra Leone * 26 Zambia * sociated with better outcomes. None of the countries in the region, however, has yet achieved a good Note: Countries are ranked by the overall scores reported in appendix macroeconomic stance (table 2.5), with a four-star rating across the table B.5, which reflect their fiscal, board in fiscal, monetary, and exchange rate policies. Improvements in monetary, and exchange rate policy have moved several countries closer to a rating, but more stances. Chad, Guinea, and Guinea- polcies "good" Bissau are excluded because of insuf- reforms are still needed. Even in countries that have gone the furthest, ficient data,. table B.5. the fiscal balance is fragile, inflation is above international levels, and the Source: Appendix table B.5. parallel market premium for foreign exchange has not been eliminated. 58 MOVING TOWARD SOUND MACROECONOMIC POLICIES Ghana, which hias the best macroeconomic policies in Africa, still had an unacceptable inflation rate of about 28 percent in 1990-91, while Bu- rundi and The Gambia, which rank second and third, had parallel mar- ket premiums of about 20 percent. In other countries, even substantial steps forward leave them a long way from adequate policies. Tanzania, for example, cut the parallel market premium by two-thirds during the ad- justment periocd, but it was still more than 70 percent in 1990-91. It is worrisome that even the best performers in Africa do not come close to matching the good macroeconomic policies typified by Chile, Malaysia, Mexico, and Thailand, where inflation is close to 10 percent per year, budget balances range from deficits of 2 percent of GDP to small sur- pluses, and currencies are fully convertible (at least for current account transactions). Notes 1. See Rutihinda (1992) for a vivid description of Tan- fiscal indicators is ours, although it is based on data in that zania's response to the crisis. study measuring the size, quality, and sustainability of the fiscal adjustment. 2. The countries with fixed exchange rates are the thirteen membe.rs of the CFA franc zone. They are 5. The increase in Zimbabwe resulted from a shift of grouped into the West African Monetary Union (Benin, teachers' salaries to the central budget. Burkina Faso, C'ote d'Ivoire, Mali, Niger, Senegal, and Togo) and the Central African Monetary Area 6. This is known as the Olivera-Tanzi effect (see, for (Cameroon, the Central African Republic, Chad, example, Tanzi 1978). Congo, Equatorial Guinea, and Gabon). Each area is served by an independent regional central bank, the 7. Ideally, an evaluation of the fiscal policy stance Banque Centrale des Etats de l'Afrique de l'Ouest in would take into account such variables as the size of the West Africa and the Banque des Etats de l'Afrique Cen- current imbalance, the structure of government revenue, trale in Central Africa. Both areas have broadly similar the patterns of government expenditure, and the depen- monetary and exchange rate arrangements. The key fea- dence on external grants to close fiscal imbalances. But no tures are a fixed exchange rate, pegged to the French single fiscal indicator adequately summarizes fiscal stance, franc at 50:1 from 1948 until 1994; a fully convertible and there are great difficulties in evaluating fiscal policy- currency; open trading arrangements; and a relatively especially its sustainability, an issue discussed in more de- disciplined monetary policy. tail in Tanzi (1991) and Blejer and Cheasty (1993). 3. Cameroon, Congo, and Gabon benefited from 8. Our definition of seigniorage is a version of Fischer higher oil exports. (1982) and was calculated as (MI, - Ml_1)/GDP, -gt(M 1I /GDP) t, where M 1t is the stock of money at the 4. The section on fiscal policy is largely based on end of period t, GDP, is gross domestic product at time t, Nashashibi and 13azzoni (1994), which presents a detailed and g, is real GDP growth. In other words, seigniorage discussion of fiscal performance in Sub-Saharan Africa be- measures the inflationary impact of money creation (that tween 1980 and 1991. The classification of countries by is, the increase in monetary growth in excess of what is 59 ADUTMENT IN AFRICA needed to satisfy the transactions demand for money). icy. We view the parallel exchange rate premium primar- This concept of seigniorage corresponds to the so-called ily as a measure of overvaluation, but as an imperfect mea- inflation tax only in the long run (when there are no sure for at least two reasons. First, it indicates over- changes in money demand). We use Ml, as opposed to the valuation only over the long term; in the short term it is monetary base, because most African countries implicitly or driven by expectations and so should be viewed only as an explicitly collect this revenue on all non-interest-bearing indicator of the exchange rate that clears the market at a deposits (checking accounts), either because the banks are given time. Second, over the long term the premium re- publicly owned or because they impose high reserve re- flects not only overvaluation but also the risks of transac- quirements on these deposits. Other measures of seignior- tions in the parallel market. age (such as those based on the monetarv base or those based on total revenue from money creation-the first 12. There were signs in 1992 and early 1993 that the term in the equation above') do not change our results in premium was once again being reduced significantly. any significant way. 13. Mozambique and Tanzania have recently made 9. The relationship between seigniorage and inflation significant improvements. is discussed in Fischer (1982) and Dornbusch and Fischer (1993), among others. Seigniorage in countries with 14. The countries in the reference group faced a dete- single-digit inflation rates generally is less than 1 percent rioration in the external environment similar to (though of GDP (in the United States it is about 0.5 percent of probably smaller than) that faced by the African fixed ex- GDP). In Argentina in the early 1980s it was about 6 per- change rate economies between 1980 (the pre-crisis cent of GDP, with inflation in the range of 600 percent a period) and 1990 (when most developing countries out- year. There are cases, however, in which moderate side Africa had completed their adjustment of the real ex- seigniorage (about 2 percent of GDP) is associated with change rate). The reference group comprised thirty non- high and accelerating rates of inflation (Brazil and Israel African adjusters discussed in the World Bank's third in the 1970s and 1980s). report on adjustment lending (World Bank 1992a). The underlying assumption was that because the African 10. It is difficult to unde.rstand why Benin fell into this countries with fixed exchange rates experienced larger de- group, because it has traditionally been a country with clines in the terms of trade than did countries in the refer- low seigniorage and low inflation, and it abides by the ence group, they should have responded with roughly monetary rules of the CFA franc zone. High seigniorage in similar real depreciation. While changes in the terms of this period appears to be unusual and probably reflects a trade are usually the most important factor affecting the short-term increase in the money supply, following a cou- equilibrium real exchange rate, a more precise calculation ple of years in which the money supply declined. would also take into account changes in productivity, cap- ital flows, and so on. Readers interested in alternative 11. A premium for foreign exchange in the parallel methodologies should see Devarajan and de Melo (1991), market indicates that the official exchange rate is inconsis- Devarajan, Lewis, and Robinson (1993), and Elbadawi tent with domestic monetary and fiscal policies and that and Majd (1992). the central bank cannot satisfv the demand for foreign ex- change at the official exchange rate. To the extent that the 15. We take 1980 as our starting point because in that premium allows countries to keep the official exchange year, according to most observers, the real exchange rate rate out of equilibrium, it measures overvaluation. And in the countries of the CPA franc zone was at an appropri- because overvaluations indirectly tax exports, the pre- ate level to support internal and external balance under mium is also frequently used as an indicator of trade pol- the then-favorable terms of trade. 6o CHAPTER 3 Unleashing Markets S UST[AINED GROWTH REQUIRES MORE THAN A HIGH RATE of investment. It requires unleashing markets so that com- petition can help improve the allocation of economic resources. It also requires getting price signals right and creating a climate that allows businesses to respond to those signals in ways that increase the returns to invest- ments (World Bank 1991 c). Having the proper macroeconomic frame- work-the subject of chapter 2-provides the incentives needed for new investment and productive use of resources. But trade, agricultur- al, and other regulatory reforms-the subject of this chapter-are essential complements to reducing the government interventions that distort prices and tie up markets. In this chapter we assess the extent of reforms and come to the fol- lowing conclusions. The African countries undergoing structural ad- justment have moved a long way toward good trade policies. They have reduced administrative rationing of foreign exchange and eliminated many nontariff barriers. Yet they still have a long way to go. Much of the progress in import liberalization still depends on large flows of external assistance. Furthermore, adjusters have not done enough to reduce the commercially oriented elements of protectionism-tariffs, nontariff barriers, and industrial restrictions-that existed before the macroeco- nomic crisis of the early 1980s. Reforms of agricultural pricing and marketing systems are under way across the continent. Almost all countries have taken steps to ensure that producer prices for Africa's major agricultural exports track world prices more closely. In a few cases, they have done this by abolishing state mar- keting boards; rnore frequently, they have allowed the private sector to compete with the marketing board, or they have adopted pricing for- 6i AN o I AFRICA mulas with a clear link to world market prices. There has been a major retrenchment of government involvement in food crop marketing, par- ticularly where the evasion of controls was previously widespread. But reform of the maize marketing boards in eastern and southern Africa is proceeding more slowly. And much remains to be done in both the ex- port and food crop sectors to take advantage of entrepreneurial talent in agricultural marketing and input distribution and to create a level play- ing field for private traders. African governments have also moved to deregulate prices and mar- kets in other sectors. Widespread price controls are a thing of the past. There has been considerable progress in removing government monop- olies, particularly in commerce. In many countries, rice, sugar, tea, and other principal commodities can now be imported freely by the private sector. But monopolies linger in such key sectors as petroleum, wheat, and fertilizer and in certain financial and infrastructure services. Over- all, countries have a long way to go in creating a favorable business cli- mate for the private sector. Trade Reforms T * NHE COUNTRIES THAT ENTERED ADJUSTMENT WITH MAS- sively overvalued exchange rates and heavily rationed foreign exchange have made tremendous strides in liberalizing imports. The large depreciations of the exchange rate and the adoption of macroeconomic policies consistent with a sustainable balance of pay- ments have done away with foreign exchange rationing, eliminating the import scarcity premiums reflected in parallel market prices. The large inflow of external funds to support liberalized import schemes has also played a major role in abolishing foreign exchange rationing. While this constitutes enormous progress from the policies imposed during the crisis of the early 1980s, for many countries it is simply a return to their policy stance before the crisis and is dependent on con- tinuing external assistance. Any discussion of Sub-Saharan trade reform, particularly import lib- eralization, must make two distinctions. First, import controls driven by short-run balance-of-payments problems-usually implemented through foreign exchange rationing-differ from the longer-term, pro- 62 UNLEASHING MARKETS tectionist import restrictions-usually implemented through tariff and nontariff barriers. Much of the liberalization so far has attacked the short-run controls. Second, the import situation recorded in national sta- tistics under official exchange rates differs from the actual quantity of im- ports and their market prices because of widespread smuggling and the commonplace evasion of tariffs and import controls. Much import liber- alization in Afrlica has merely rationalized the situation in parallel mar- kets and reallocated the "rents" previously embodied in official controls. These two distinctions-between short-run import controls and long-run commLercial policy and between official statistics and real prac- tice-must be taken into account when evaluating the impact of trade reform. Normaliy, a devaluation would make domestic producers of im- port substitutes better off by raising the prices of competing imports, while a reduction in import barriers would make them worse off by in- creasing competition from goods at lower prices. The major reduction in import barriers in Africa has come from the lifting of import controls, not from tariff reform per se. But the price changes facing individual firms and industries have varied enormously, depending on the access to rationed imports and foreign exchange before the reforms. Many domestic manufacturers, particularly parastatals, had access to imported inputs and-perhaps more important-to external credit to fi- nance capital goods at the artificially cheap official exchange rate. Their products generally were protected from competition or faced competi- tion at the parallel market price. These firms, because they relied on for- eign financing to produce heavily protected goods, were hit extremely hard by devaluation and by a reduction in import barriers. First, the de- valuation increased the cost of servicing their external debt and the cost of their imported inputs. Second, the reduction in import barriers meant that they encourntered more competition, which dampened the potential rise in the prices of importable outputs from a devaluation. How much their profitabiliry suffered depended on how much they benefited from rationing before the reform and whether they continued to be protected by import monopolies or other trade barriers. This double shock to many large industrial parastatals is perhaps what lay behind their finan- cial difficulties in some countries. Firms operating in the informal sector, which had less access to rationed inputs, were hit less hard-and many even benefited from increased access to vital inputs. The reality is that actual import prices in many flexible exchange rate countries probably have not changed much on average, because they al- 63 ADJUSTMENT IN AFRICA ready reflected parallel market prices before the reform period. Liberal- izations probably have not substantially expanded the quantities of goods available on the open market, except in the few cases where goods were tightly rationed. But the liberalizations have eliminated rents ac- cruing to those who obtained rationed goods and enforced controls. Trade reform in the fixed exchange rate countries obviously differs from the foregoing storyline. The reason: the convertibility arrange- ments for these countries enabled them to avoid foreign exchange ra- tioning and many of the excesses of import restrictions. This partly ex- plains their superior economic performance before the adjustment period. But imports by these countries have recently plunged, despite the continuing real appreciation of the CFA franc, especially after 1985. This plunge reflects a compression of domestic demand-not an in- crease in direct barriers to imports. Moreover, the combination of an ap- preciating real exchange rate and severe recession has made it very diffi- cult for them to liberalize imports, as domestic producers of import substitutes already are under tremendous pressure from cheap foreign imports and declining domestic demand. Box 3.1 What IIs rade lrm- andHo id We Kno If There Has Been Some? UNDER ANY SET OF T RADE POLICIES, THE ECON- served monopolies, reference prices, and quality re- omy attaches a set of relative prices and profitaiities quirements) that do not intervene directly in trade to various activities. These prices act as incentives to flows but nevertheless affect trade. Assessing the dis- determine the structure of production and consump- tortions from each of these barriers is conceptually tion of goods, which in turn determines the amount straightforward but practically impossible in Sub- and composition of imports and exports. New trade Saharan Africa for three reasons. policies change relative prices, either implicitly or ex- First, in countries without fixed exchange rates plicitly, and thus affect production and consumption and convertible currencies, the policy restrictions decisions. Import liberalization consists of moving to- that have been most important in creating trade dis- ward market determination of prices and thus toward tortions are the foreign exchange allocations and ii- market determination of the level and composition of censing requirements for imports. The incidence of imports and domestically produced importables. barriers can be measured-for example, one can cal- To evaluate the extent of import liberalization in culate the fraction of all goods subject to import li- Africa, we must analyze the effect of government pol- censing requirements. But barriers generally are ad- icy instruments on relative prices. Four trade inter- ministered in a way that makes the distortionary ventions are common: rationing foreign exchange, impact far from transparent. In one country, all erecting nontariff barriers to imports, imposing tar- goods nimight require a license, but the licenses are iffs,and using other dlomestic policies (such as re- freely granted. In another, only half the goods might 64 UNLEASHING MARKETS Liberalizing Imports Because of overlapping, less-than-transparent trade restrictions in Sub-Saharan Afirica, no single measure of trade distortion can be used unassailably to assess how much any country has liberalized imports or to make fine comparisons across countries (box 3. 1). But Africa's trade regimes can be classified by the four phases of liberalization that Krueger (1978) described in her classic study: complete or nearly complete gov- ernment control or rationing of imports, heavy use of nontariff barriers for protectionist purposes, mostly or exclusively tariff-based protection, and free trade (table 3.1). The range of policies within each category is large, but the overall framework is useful in tracking reform efforts. There are two typical patterns of trade reform in Africa (figure 3.1): * Most adjusting countries with flexible exchange rates have moved away from complete control over imports and foreign exchange and toward the use of extensive trade barriers, especially nontariff barriers--and in some cases toward modest, mostly tariff-based require a license, but few receive them. The distor- justments for quality, durability, and fashion. These tion is thus greater in the country with fewer license direct price comparisons are difficult and expensive requirements. Moreover, the availability of foreign and hence are rarely possible to obtain for an econ- exchange can strongly determine the allocation of omy or even a sector (except in the case of commodi- imports and the severity of distortions, even for the ties). And they are never available over time in a way same incidence of trade barriers, because the admin- that would allow the evolution of import distortions istration of the system might change with the per- to be assessed. ceived constraints on foreign exchange availability. Third, the effects of various barriers often overlap, Second, measuring price distortions requires so examining the reform of any single aspect of the im- knowledge of what domestic prices are and what they port regime can mislead. The same import item might would have been without the barrier. Even using bor- be subject to a tariff of 30 percent, an import license der prices as a proxy for the latter requires accurate requirement, and general foreign exchange rationing. measurement of the difference between the world If tariffs on this good fall, have imports been liberal- price and the domestic price-extremely difficult for ized? Maybe, maybe not. If the premium implicit in all but the most standard commodities. To compare the licensing (the price increase due to the license) was the domestic price of, say, a man's shirt made in higher than 30 percent, import flows are unaffected by Kenya or Zambia with the "world" price of a man's a tariff reduction, so revenue is lost from the tariff re- shirt involves numerous complex and debatable ad- duction without any gain in import liberalization. 65 A :FST NAFRICA Table 3.1 Phases in Liberalizing Import Regimes Phase Comment Examples outside Afiica Complete or nearly complete Usually the result of severe Peru (1987), Brazil (1966) government control over imports macroeconomic disequilibria Extensive trade barriers, with Typical of highly protectionist Argentina (1970s), Brazil (1980s), widespread nontariff barriers import-substituting regimes India (1970s), Turkey (early 1980s) Modest, mostly tariff-based Typical of most industrialized Republic of Korea (late 1 980s) protection countries Free trade or low, uniforrr. Rare Hong Kong (ongoing), Chile (late 1980s), tariffs Argentina (early 1990s) Source: World Bank staff. protection. Some-Tanzania and Zambia, for example-have eliminated rationing but have not fully eliminated lists restricting allowable imports, or they have relied heavily on donor financing for systems of open general licenses (described later). But a few countries-notably Ghana-have begun to solidly commit them- selves to making foreign exchange available through realistic ex- change rates and to eliminating all but a few nontariff barriers. Even Ghana, however, lags behind Mauritius, which completed its adjustment program early on, including extensive trade reforms (box 3.2). * The second set of countries never resorted to widespread foreign exchange rationing and has moved slowly toward import liberal- ization. This group indudes the flexible exchange rates countries that never experienced a severe macroeconomic crisis, such as Kenya and Zimbabwe, as well as the fixed exchange rate countries that maintained convertible currencies throughout the period. A few of the latter have made progress in eliminating nontariff bar- riers and in tariff reform. To evaluate import liberalization in Africa, we look at four common interventions: foreign exchange rationing, nontariff barriers to imports, tariffs, and regulatory barriers that indirectly affect trade. Eliminating foreign exchange rationing. Devaluing extremely overvalued currencies and reducing the rationing of foreign exchange for imports are among adjustment's biggest achievements in the 1980s in flexible exchange rate economies. Some countries (Ghana and Uganda) have depreciated the exchange rate enough to eliminate rationing. Other 66 UNLEASHING MARKETS Figure 3.1 Evoluition of Trade Policy in Selected Countries during the Adjustment Period Complete/ nearly complete government Moderate Low tariff-based control over Heavy use of tariff-based protection/ Imports trade barriers protection free trade Ghana 1 | _ R = t;X : _~ Nigeria E l Tanzania El Zambia -, 0 = C6te dlvoire Kenya Mali Senegal MO Despite much progress, many Zimbabwe countries have not firmly committed to low or moderate tariffs. Soturce: World Ba nk staff. countries (Tanzania and Zambia) have undertaken substantial reforms, but the reforms have been halting and incomplete. Still others (Madagascar and Nigeria) launched substantial reforms now threat- ened with reversals. A fourth group (Kenya and Zimbabwe) has avoided an overvalued currency but retains import controls. Our assessment of progress in eliminating foreign exchange rationing is based on three indicators: nonmarket foreign exchange allocation mechanisms, paral- lel market premiums for foreign exchange, and the rationing of imports (appendix table A.7). 67 IN AFRICA Box 3.2 Trade Reform: Mauritius and Ghana SOME TWENTY-FIVE YEARS AGO, BRITISH ECONO During the 1:970s, administrative controls over mist James Meade pointed to Mauritius as an example foreign exchange allocation and imports were not of a country where the "outlook for peaceful develop- genera:ll used to correct balance-of-payments prob- ment is poor" (Meade 1967). Instead of fulfiling tlems, as they commonly were in other countries. Meade's bleak prognosis, Mauritius went on to be- Mauritius was not plagued by the distortions caused come the brightest star in the somewhatf dim firma- by large-scale parallel markets in foreign exchange and ment of Sub-Saharan African economies in the 1980s. multiple exchange rates. This changed to some extent Unemployment dropped from a high of 22 percent in ithe ate 1970s, and by 1981, quotas covered 65 1983 to aboutj3 percent in recent years, and per capita percent of imports. Elimination of these quotas was income increased by more than half between 1983 and one of the first steps in the country's structural adjust- 1990.Maitiusssuccessa ettbuted to many mentprogram. At the same time, tariffs on some of factors, particularly a propitious external environment, these products were increased, though overall protec- but surely muchn of its good fortune (including its abil- tion clearly declined. The most protected sectors were jrity to rake advantage of external events) was due to its a relatively small part of the manufacturing sector, so trade and macroeconomicpolicies, the potential for damage was limited. The adjustment Disciplined exchange rate management and sound program later reduced and rationalized tariffs, though macroeconomic management overall have kept the not to the low levels of Latin America or East Asia. real exchange rate remablyste. Throughout the' As is well known, Mauritius's success was reflected 1970s and early 1980s, it did not change fro ear to in the phenomenal growth of its export processing year by more than a few percentage points. About zones and (to a lesser extent) tourism. The foundation 1982 (the exact year depnsfor this growth was laid by the policies mentioned the real exchange rare didnei to apreciatein the above, in conjunction with unusually good ancillary wake of macroeconomic a olistrctural imbalances. policies. Unlike many other countries, traditional ex- Even in the early 1 980s, however, the appreciation ports were not heavily taxed. And most productive ac- was not severe (between 3 and 28 percent, ag e tiiry was in private hands (state-owned enterprises ac- pending on the series) and was soon revesd byde- counted for only 8.7 percent of GDP). This put the preciation to a level below4tat ofthe I197Os. private sector in a good position to take advantage of the Although nearly every country had almost universal discretionary control over imports in the crisis period, many had moved to nondis- cretionary access for imports by late 1992. In several countries, however, including Kenya, Nigeria, and Madagascar, renewed balance-of-pay- ments pressures have threatened, or resulted in, reversals. Despite mas- sive exchange rate reforms, no flexible exchange rate country has estab- lished complete convertibility of its current accounts. Even so, a large number permit untrammeled access to foreign exchange for imports, ei- ther through a foreign exchange auction or through less restrictive li- censing. In several countries, mechanisms other than eliminating or re- ducing foreign exchange licensing have helped to relax the foreign 68 UNLEASHING MARKETS export processing zone incentives, and indeed most of eliminated in 1987 and tariffs became the binding the early investment came from domestic sources. constraint on imports, protection fell precipitously. Ghana's advances in trade policy have been more Producers howled, and high surcharges were subse- recent and rnore fragile. They began slowly in 1983, quently imposed. These have since been reduced following a period in which per capita income had somewhat. fallen 30 percent since 1970. The most radical and Nontraditional exporters have benefited from im- important reform was the overhaul of the foreign ex- provements in the duty drawback system and from change sysrem, from one of grossly overvalued increased foreign exchange retention rights. In the exchange rates and administrative allocation to one very important cocoa sector, incentives to producers relying on an auction and interbank market. The re- have improved; producers' share of the free-on-board sult has been the virtual elimination ofa parallel mar- price increased from 35 percent to 54 percent be- ket premiurn that reached 2,100 percent in the early tween 1984 and 1990. While domestic marketing has 1980s, along with the corruption and rent-seeking been demonopolized, exports still are the exclusive caused by distortions of such magnitude. The ex- domain of the state marketing agency, and privatiza- change market is still not exactly free, since the auc- tion of state-owned plantations has gone slowly. tion is "Dutch"-meaning that different purchasers The adjustment program, though far from com- pay different prices for the same commodity-and plete, arguably qualifies Ghana as the most advanced exporters still do not have fill rights to retain their trade policy reformer in Sub-Saharan Africa after foreign exchange earnings. Further reforms are Mauritius. And the reforms have begun to show re- needed, but the progress since the pre-adjustment sults. Since the program's inception, GDP has grown period has been substantial. about 4 or 5 percent a year, compared with 0.1 per- Other trade reforms have been significant, though cent annually between 1960 and 1983. Cocoa pro- not as dramatic as in the exchange market. Tariff re- duction has shot up 77 percent. In spite of cocoa's form early in Ghana's adjustment program was im- growth, nontraditional exports have grown fast pressive (with the tariff rate structure now ranging enough to increase their share of total exports to 4.6 from 10 to 30 percent), but it preceded the elimina- percent. Manufactured exports grew by 52 percent tion of quantitative controls. Thus, when these were (though from a small base) between 1985 and 1990. exchange constraint. One, called own-funds imports, allows anybody possessing foreign exchange to use it for imports, either choosing from a specified list oiF allowed imports or freely importing any item (box 3.3). Another mechanism is an export proceeds retention scheme, which per- mits exporters to retain a part of their foreign exchange earnings. The paraLleL market exchange rate premium, the second indicator of the extent of foreign exchange rationing, has declined substantially since the initiation of reforms. This tells us that the value of a unit of foreign exchange is nearer the official price, and that rationing has lessened. A third usefal indicator is the gap between predicted (notional) im- ports and actual imports. This gap, attributable to an array of unob- 69 ADJUSTMENT IN AFRICA ALLOWING OWN-FUNDS IMPORTS ElIMIES ai be small. ut if exports are price-responsive and the penalties associate allowing own-funds imports gives actions and makes importi nglerisk.Atte sae xprtersan incetive to expand exports in order to time, it increases the demand for foreig nge suly iprts that are in demand. This aids trade and gives exporters an incentive to demincreasing imports and (unofficial) by smuggling more goods out of the coury (or, expors. where they are permitted to retain acertaert- B o-n imports are semiofficial flows, age of their foreign exchange earnin increasing their magnuis difficlt to assess. For Tanzania in legal exports). Analytically, these own nds sc s 198r rt he are equivalent to establishing multiple exa t o na d license eg the rates, because they implicily allow those with foreign amount offoreign exchange the government allocated exchange to sell it at the m t p a In Zb osiowere $100 ket rate to those who watto imsT em i ontt at regulation reduced the irmpact on trade liberal ip ontAlthough own-fudini- smuggling imports ard exports. Ifa arge liberalization, they are porous, liberalization will have no impa xprtstea au- are very difficult to smuggle out, the supply os im arfor eign exchange will not increase, and the ec wl m towa a liberalzedoficial economy. 0t served impediments to imports, is measured by the index of implied import restrictiveness (Narasimhan and Prichett 1993). A high index number-the result of notional demand for imports being much higher than actual imports-indicates that imports face substantial re- strictions, among them foreign exchange rationing.1 When a currency is overvalued, imports are cheap at the official exchange rate, and no- tional demand is accordingly high. As overvaluation increases and the foreign exchange available to purchase imports dwindles, imports often fall-sometimes dramatically. In Ghana before its devaluation in 1983, the demand for imports far exceeded the actual level (figure 3.2). After devaluation, the actual level of imports approached the predicted level of demand.2 Nigeria and Tanzania share a similar pattern. But some countries, like Kenya, appear to have tightened reductions on imports by raising nontariff barriers rather than by rationing foreign exchange. Eliminating nontariff barriers. The second strand of evidence on import liberalization is progress in eliminating nontariff barriers (NTBs). Such barriers in Sub-Saharan Africa are predominantly government licenses or approvals for imports. Import licenses are usually connected to for- eign exchange allocations, in the sense that a valid import license is 70 UNLEASHING MARKETS Figure 3.2 Evokil:ion of Import Restictiveness in Ghsan Volume of imports (logarithms) 15 Notional demand' 14 = In any year when demand exceeds imports, the percentage of deviation is the implied import 13 -restrictiveness 12 - ,'-. , Actual imports In Ghana, depreciation and trade reform brought imports in line with ____________________________________________________ dem and. 11- 1975 1978 1981 1984 1987 1990 a. Notional import demand is defined as a + ,BGDP + yREER, where a is a constant calcu- lated to equalize acaual and notional imports in the 1970s, / is an income elasticity equal to 1.25. y is an exchange rate elasticitv equal to i, and REER is the real effective exchange rate. Source: Narasimhan and Pritchett (1993). needed to obtain foreign exchange. But these licenses have the addi- tionaL function of protecting domestic producers from foreign compe- tition. NTBs are a particularly virulent form of protection because they are not transparent in the price distortions they create. Furthermore, because the licenses are often specific to a firm, individual, or end-use, they not only limit the quantity of imports but often reduce their eco- nomic benefit, because those imports do not flow to those who need them most. Progress has been made in reducing the number of goods requiring prior approval for import (appendix table A.8). Whereas many coun- tries routinely required licenses for all goods, they now increasingly use a "negative list"'-allowing what is not forbidden. Open general license (OGL) schemes have been one mechanism for liberalizing imports in this way. With an OGL scheme, a limited number of goods are placed on a list to be automaticaLly approved for import and for the allocation of foreign exchange. The 'ist is gradually expanded until the allowable items are so numerous that it becomes more efficient to move to a smaller, negative list. Many countries have an OGL scheme, and in some it covers a sub- stantial proportion of imports. 71 ADJUSTMENT IN AFRICA The OGL schemes have helped phase in import liberalization. The ex- pansion of unlicensed goods can move in step with the expansion of available foreign exchange, so as not to jeopardize the overall trade re- form. The OGL goods are not rationed according to user and end-use (as with foreign exchange and NTB licensing), and they are available with little delay, so they can at least reach the users who most need them. This alone reduces much of the distortion caused by NTBs in the efficiency of resource allocation. However useful they have been in reducing nontariff barriers, Africa's open general license schemes typically are neither open nor gen- eral. The schemes have enabled countries to move from a system of highly rationed imports to a more liberalized regime-but without committing to currency convertibility or an overall liberalization that might place central bank reserves at risk. Using an OGL scheme to lib- eralize nontariff barriers would not, then, constitute a rapid and com- plete opening of trade. The fraction of goods eligible for licenses often starts out small and increases gradually. In Zambia the scheme began in 1989 with only 10 percent of imports eligible, but 95 percent (exclud- ing petroleum and fertilizer) were covered by September 1992. Simi- larly, Tanzania established an OGL scheme in 1987 but did not move to a negative list until 1991; in April 1992, a fifth of non-oil items still were not eligible for import. There are two problems with open general licensing. First, in the early stages the move to an OGL scheme can increase, rather than de- crease, the protection of domestic production. The first goods placed on an OGL list are those that are not produced domestically (so no one lob- bies against them), but that are required as inputs for domestic produc- tion (so producers who use them lobby for them). By placing the input-but not the final product-on the list of goods that can be im- ported, the OGL scheme liberalizes the import of the input and lowers its price, but does not alter the price of the final product. Thus, it raises ef- fective protection (the protection of the value added in a given activity). For example, under an OGL scheme covering industrial equipment but not shoes, a domestic manufacturer of footwear could import its capital goods more cheaply while maintaining its established price for shoes, thereby reaping higher profits. A second problem with OGL schemes is that they often rely heavily on donor funds and are often linked rigidly to the official exchange rate set by the government. But the level of the official exchange rate, the goods 72 UNLEASHING MARKETS eligible for import, and the funds committed to an OGL scheme may not be mutually consistent. There may, for example, be excess demand for OGL funds at the official exchange rate. In such instances, OGL imports are being subsidized at the same time that there is a substantial premium for foreign exchange in the parallel or own-funds market. Because the transition from an OGL scheme to liberalized imports is difficult, the re- cent history of CJGLs has been rocky in many countries (such as Tanza- nia and Zambia). Finally, caution is needed in using the number of goods that require licenses as an indicator of NTBs. A very few products sometimes account for a large part of domestic production. If an African economy has only, say, seven items under NTBs, but those seven items are maize, fertilizer, sugar, beer, soft drinks, cement, and steel products, this small list could easily cover half or more of all domestic value added in tradable items (excluding services and nontraded agriculture and industry). Nearly every country, riot just those in Sub-Saharan Africa, retains a core group of commodities that it is reluctant to subject to import pressures, even with high tariffs. Reforming tariffs. Tariff reform typically encompasses three steps. The first, already taken by many African governments, is rationalization: reducing the number of tariff rates, ad hoc exceptions, and separate levies and then assigning rates to products according to systematic cri- teria. The second step, often undertaken simultaneously with rational- ization and cur:rently under way in much of Africa, is to reduce the dis- persion of tariff.s, generally by raising the minimum tariff and lowering the maximum. Lowering the maximum is typically symbolic, since in practice the highest rates are avoided or evaded. The third and final step is to lower the overall level of tariffs to reduce the average level of domestic protection. This becomes important only after all other bind- ing restrictions on imports have been eliminated-a point that most Sub-Saharan African countries have not yet reached. Progress in rationalizing tariff codes-often to increase revenue col- lections-has been substantial. Pre-reform tariff codes were enormously complex (invol ving more than 2,500 separate items) and were often de- termined by ad hoc (and increasingly arbitrary) tax levies, rates, exemp- tions, and conditions. This led to tariffs that were irrational-in the sense that no government, whatever its objectives, would have chosen them. Typically the tariff codes were so cumbersome, so complex, and so riddled with exemptions that much revenue was sacrificed. Most 73 ADJUSTMENT IN AFRICA African countries are still remedying these problems-reducing the number of tariff rates, rationalizing those that remain, and improving administration and collection. Burkina Faso, Ghana, Mali, Senegal, and Togo are among those that have reduced the number of rates, oftentimes to only four or five. Many countries have also taken the second step in tariff reform, by lowering the maximum rates-Kenya from 170 per- cent to 60 percent, Rwanda from 270 percent to 100 percent, Tanzania from 120 percent to 40 percent.3 Because of the importance of tariffs for government tax revenue (Sub-Saharan countries rely on import duties for more than 20 per- cent of all tax revenue) as well as the exigencies of fiscal austerity, few countries have moved to the third step of reducing average tariff rates substantially. Indeed, for many adjusting countries, the ratio of tariff collections to import value (the average collected rate) has in- creased, perhaps a beneficial start. Average collected rates increased in Madagascar, Malawi, and Uganda and did not fall much even in countries that significantly reduced rates. In Kenya the trade- weighted average tariff dropped only from 40 percent to 20 percent during the course of reform, while collected rates fell from 18 per- cent to 14 percent. Among the countries with fixed exchange rates, those in West Africa have made more progress in liberalizing trade than those in Central Africa. Five of the seven West African countries undertook important tariff reforms, although C6te d'Ivoire and, to lesser extent, Senegal later reversed their actions because of the combined pressure of the recession and import competition. Change has been slower in the Central African countries because of the need for a common approach (mandated by their participation in the Central African Customs Union), but a major reform effort is now under way. The Customs Union has announced plans for unionwide reform that will rationalize the tariff code by re- ducing the number of rates, minimizing dispersion, and lowering the average tariff-thus accomplishing most of the goals of tariff reform in one step. Regulatory barriers. Another barrier to imports-seldom considered a barrier because it does not operate directly on trade-arises when a firm or set of firms is granted monopoly privileges over all sales of a good, whether it is produced domestically or imported. If these monopolists are not private firms, they are likely to choose the level of imports needed to protect domestic production. Many key industries 74 UNLEASHING MARKETS (cement, sugar, ifertilizer, petroleum) were-and still are-subject to such parastatal monopolies. Limiting the power of these monopolies is a politically difficult but necessary step to increase competition and reduce consumer prices. Import liberalization also involves giving the private sector more free- dom to import. Burundi, for example, had very restrictive licensing pro- cedures for importers, limiting the number of importers. It has since re- laxed these restrictions, allowing more competitive pressure. Promoting Exports During adjustment, the need to increase exports was clear-to im- prove both growth and the balance of payments. An overvalued cur- rency has been the primary obstacle to exports, with devaluation a major part of the cure, But exports, especially traditional agricultural com- modities, have also been often hurt by three other mechanisms: * Export licensing and controls over export activity have been wide- spread. * Export taxes have traditionally been high and an important source of government revenue. Although many countries have lowered them as part of adjustment programs, they linger in other coun- tries because of fiscal concerns. * Marketing boards, with monopoly control over the sale of exports and over domestic purchases from farmers, have reduced the prof- itability of exporting in many countries by offering export pro- ducers low prices. In some cases, devaluation led to huge improve- ments in border prices, but output did not jump dramatically because the marketing boards, in an effort to boost profits or cut losses, did not pass all of the increase back to producers. For nontraditional exports as well, the major policy change has been the depreciation of the real exchange rate. Combined with reduced im- port protection, the devaluation increased the profitability of exports relative to the production of import substitutes or goods not traded. As an intermediate step, many countries (such as Nigeria, Tanzania, and Uganda) have given nontraditional exporters access to a more depreci- ated exchange rate through export proceeds retention schemes. Many countries are also doing more to promote nontraditional exports, using a variety of approaches: 75 TM0 NTIN A FRICA * Most countries have some legal provision for duty drawbacks-re- bating taxes paid on imported inputs used in export production. However, rebates are slow, documentation requirements onerous, and the programs administratively complex. Malawi, Senegal, Tanzania, Uganda, and Zambia have tried to simplify their duty drawback schemes, but when fiscal constraints are tight, the trea- sury tends to hold up customs tax rebates. * Export processing zones (EPZs) allow exporters access to inputs at world prices without cumbersome duty drawback schemes. Suc- cessful in Asia, the zones vary from large fenced areas with their own extensive infrastructure to bonded warehouses and factories where export inputs are processed duty-free. In Africa, Mauritius has the most successful EPZ, and Madagascar has made a promis- ing start with one. Kenya, Nigeria, Togo, and Zambia are also es- tablishing EPZs or in-bond production arrangements. * The relaxation of foreign investment codes, though not aimed ex- plicitly at export-oriented investments, has been another impor- tant policy shift. * Governments have also aided nontraditional exporters by stream- lining reporting and licensing requirements and strengthening infrastructure. Agricultural Reforms A* FRICAN FARMERS HAVE FACED THE WORLD'S HEAVIEST rates of agricultural taxation, perhaps partly because agricul- ture has been such a crucial source of revenue for African gov- ernments. African farmers were taxed explicitly through producer-price fixing, export taxes, and taxes on agricultural inputs. They were also taxed implicitly through overvalued exchange rates, which reduced the prices they obtained for their exports, and through high levels of industrial protection, which raised consumer prices. A study covering eighteen countries across the world showed that three broadly repre- sentative Sub-Saharan countries (C6te d'Ivoire, Ghana, and Zambia) taxed their farmers 70 percent more than the average for developing countries (Schiff and Valdes 1992). The high rates of taxation con- tributed to Sub-Saharan Africa's alarming decline in the average annual 76 UNLEASHING MARKETS rate of agricultural growth-from 2.2 percent in 1965-73 to 1.0 per- cent in 1974-80 to 0.6 percent in 1981-85 (Cleaver 1993). Reducing the taxation of the African farmer has been a top priority in agricultural reform. And agricultural reform is high on the adjustment agenda because agriculture accounts for roughly 35 percent of Africa's GDP, 40 percent of exports, and about 70 percent of employment. To meet food needs, raise incomes, and provide foreign exchange, the rate of growth in agricultural production over the next thirty years needs to be twice that of the past thirty years. Much of the increase will have to come from raising production in ways that conserve natural resources and thus ensure sustainable growth. Moreover, growth in the produc- tion of nontradable food staples is needed to enable production to keep pace with demand, limit the rise in consumer food prices, and help keep real wages competitive (Delgado 1992; Seckler 1993). Some Export Producers Are Better Off, Despite Declining World Prices Price trends for primary commodities over the past decade have been tough on African producers of agricultural exports. The real interna- tional prices of Africa's two major export crops-cocoa and coffee- dropped almost 70 percent between 1980 and 1990, and the real price of cotton, the third major export crop, 28 percent. Only two of the ad- justing countries, Benin and Chad, registered an increase in the border price for their major agricultural exports-most likely a statistical anomaly. In the remaining countries, border prices (in foreign currency) fell. Given these declines, African governments have found it very diffi- cult to improve real prices for producers of agricultural exports. But as figure 3.3 shows, ten of twenty-seven countries managed to do so by 1989-91, in pait because they had substantial scope for adopting better policies.4 Some gave producers a larger share of the border price at the official exchange rate (as measured by the nominal protection coeffi- cient) by lowering export taxes, raising administered producer prices, re- ducing marketing costs, or liberalizing marketing. Others depreciated the real exchange rate and passed part of the higher local currency earn- ings back to producers. (Appendix C explains how policy and external factors interact in determining the real producer price.) The remaining seventeen countries did not manage to offset the de- cline in world prices. In nine of these countries, governments nonethe- less took measures to help their farmers. Either they raised producer 77 ADJUSTMENT IN AFRICA 1!g Di=2.3 Chne A he6 A Ra - p,~0- Ghana Nigeria Burkina Faso Benin Mozambique Togo Tanzania Mali Madagascar Niger Central African Rep. Zimbabwe Malawi= Senegal Chad Burundi Rwanda The Gambia Kenya L=== Gabon Congo Uganda Zambia _____ Cameroon rr is Real producer prices of export crops Cte d'voire went up for ten countries, down for Guinea-Bissau seventeen. Sierra Leone _ _ _ _ _ _ _ _ _ -80 -60 -40 -20 0 20 40 60 80 100 Percent Note: Data for each country are based on the percentage change in producer price for each major crop, weighted by that crop's share in the total value of agricultural exports in 1985. Data on real prices were not available for Guinea. Mauritania was excluded because it has no major export crops. Source: Appendix table A. 18. prices more than enough to offset the effects of an appreciating ex- change rate, or they passed on to the producers part of the increase in the local currency proceeds from a depreciation. Although not enough to compensate entirely for the drop in world prices, these measures at least offset part of it. In the other eight countries, the combined effect of macroeconomic and agricultural policies worked against farmers and compounded the decline in world export prices.5 About two-thirds of the countries reduced the overall tax burden on agriculture during the 1980s (figure 3.4)6 Did countries reduce both explicit taxation (from low producer prices) and implicit taxation (from overvalued exchange rates)? In general, no. Most countries reduced one but not the other (figure 3.5). Ghana, Nigeria, and Uganda significantly 78 UNLEASHING MARKETS reduced the inmplicit taxation from the overvalued exchange rate through massive devaluations, but they taxed away some of the benefits by not passing all of the border price increase back to farmers. In con- trast, Cameroon, Cote d'Ivoire, and Senegal took steps to give farmers a larger share of the border price as the real exchange rate became increas- ingly overvalued.. Meanwhile, some countries did reduce levels of both explicit and implicit taxation: Guinea, Madagascar, Malawi, and, to a lesser extent, But undi. But a few others increased both explicit and im- plicit taxation, rnotably Zambia and Guinea-Bissau. Countries wlhose currencies became more overvalued in real terms Seventeen adjusting countries tried to offset the loss in competitiveness with a decrease in explicit tax- reduced overall taxation of ation-by reducing export taxes, squeezing marketing board profits, agriculture. Ghana Guinea Madagascar Malawi Uganda Central African Rep. Tanzania Burkina Faso Rwanda Burundi Togo Gabon Mali Kenya Congo Nigeria Niger Mozambique Zimbabwe The Gambia C6te dlvoire Chad Benin Senegal Sierra Leone Cameroon Guinea-Bissau Zambia . _ _:-__ __ -10( -50 0 50 100 150 200 250 300 350 Change in the real protection coefficient (percent)' Increase in Decrease in overall taxation overall taxation ATote: Mauritania was excluded because it has no major export crops. a. See appendix C for further discussion of .he real protection coefficient as a measure of overall taxation. Source: Appendix table A. 19. 79 ADUTENT IN AFRICA Change in the nominal protection coefficient (percent) 100 i Central African Republic El Congo 80 .I Burkina Faso n- 60 Togo rj 60 El~~~~~~~~~~ Gabon .. Mali C 40 3 Madagascar C Cameroon 20 C: Malawi 3 C Senegal Rwanda Cote d 'voire -Guinea Burundi C Niger F C Berin 0enya El Ch d Sierra Leone .M) j ~~~~Kenya C: Cad The Gamnbia 3 -20 Zimbabwe C Uganda C- C Nigeria Mozambique E C Tanzania -40 T Zamba -60 Ghana C Guinea Bissau -80 -100 -50 0 50 100 150 200 250 Change in the real effective exchange rate overvaluation proxy (percent) ,'eease in implicit iaxatlon increase in irnplicit taxation Note: Mauritania was excluded because it has no major export crops. Source: World Bank estimates. Few countries reduced both explicit and subsidizing producer prices by borrowing from the banking sector and implicit taxation. and not repaying. But in about half these countries, the increase in over- valuation more than offset the reduction in explicit taxation, leading to an overall rise in taxation. In contrast, countries that devalued their real exchange rates, such as Ghana and Tanzania, were more successful in re- ducing the overall burden of taxation, even though they did not pass the full benefit of the devaluation back to producers. Only in Zimbabwe did the increase in explicit taxation offset the reduction in overvaluation. Evidently, for countries whose real exchange rates became more over- 80 UNLEASHING MARKETS valued, it was difficult to reduce explicit taxation enough to counter the effects of the real exchange rate overvaluation. But in countries whose real exchange rates were devalued, it was easier for the government to re- duce the overall tax burden on export producers even while appropriat- ing some of the benefits of the real devaluation. Some Progress !in Liberalizing Export Crop Marketing Reform efforts for export crop marketing fall into three categories: eliminating marketing boards, linking producer prices to world market prices while reforming marketing boards to reduce costs, and allowing the private secto:r to compete with marketing boards in crop purchasing and exporting. M4ost countries have undertaken at least some reform of their export crop pricing/marketing systems, though few have abolished marketing boards (appendix table A.9). Eliminating marketing boards is generally the best option from an economic standpoint, but it is often politically difficult. Nigeria eliminated all its marketing boards in one step in 1986, with good results once the initial difficulties were resolved. Other countries have eliminated marketing boards for some but not all of their export crops-coffee in Madagascar, for example, and ground- nuts and cowpeas in Niger. Several adjusting countries are experimenting with efforts to link pro- ducer prices to world prices. West and Central African cotton-exporting countries, where many parastatals for cotton production and marketing are jointly managed by foreign shareholders, are trying a two-payment system: a fixed sum at the time of crop purchase and a variable payment based on the unit prices and marketing board profits. In theory, if world prices are higher than expected at the time of export, these parastatals pass the windfal[ back to the producer in a second payment. It is not yet clear whether such systems will work. Their success depends on efficient administration and on passing on the benefits of world prices as promised. In Kenya, coffee producers sometimes do not receive the ap- propriate secon(l payments, or they receive them only after long delays. In Malawi, shortcomings in a similar effort to link world prices to pro- ducer prices for smallholder producers of tobacco have led to a more lib- eralized system. Some smallholders now have the right to grow the lu- crative burley tobacco and sell their crop on the auction floor. The third type of reform-maintaining the marketing board but al- lowing private traders to compete-is meant to foster the benefits of 8i ADJUSTMENT IN AFRICA competition and gradually demonopolize the marketing board or trans- form it into a completely private operator. Burundi is trying this ap- proach with coffee; Tanzania, with cashew nuts. Most of these reforms are too new to evaluate, though some appear to be succeeding (box 3.4). But the barriers to private competition can remain formidable. Often, marketing boards retain a de facto monopoly through control over processing facilities or privileged access to bank fi- nancing. Similarly, boards can squeeze out private agents by setting un- profitably low price margins and relying on the government or state banks to subsidize their losses. This can happen even when processing and marketing boards are ostensibly privatized if there are substantial barriers to entry. In Tanzania, ownership of coffee processing facilities has recently been transferred to cooperatives that were formerly state en- tities but are now-in principle at least-private. Other private agents, however, lacking the means to invest in costly processing facilities and facing other restrictions, find it very difficult to compete. IN THE EARLY 1970S, TANZANIA HAD ONE OF THE traders were given access only to certain buying areas world's largest cashew nut industries, with marketed (to protect the cooperatives) and were restricted to production of 145,000 tons (about 30 percent of official buying prices. Private traders purchased only world production). Despite a buoyant international 17 percent of the marketed crop. Even so, producers market, Tanzania's production fell to less than got a higher share of the export price than they had 17,000 tons in the late 1980s. Several factors con- earlier. tributed to the downward spin: a cashew nut blight, The second season in 1992-93 featured some im- the country's prograrn of village development', a provements, but some barriers to private sector entry sharp and continuous (decline in real producer prices, lingered. Even so, private traders purchased 43 per- and major inefficiencies in the processing and cent of the crop and reestablished links to the trading marketing. firms operating in the 1950s and 1960s. They also With the industry on the brink of collapse, the expressed some interest in leasing long-dormant pro- Tanzanian government announced in 1991 that the cessing factories. cashew nut market woild be liberalized, with private Under the new system, farmers are having their traders allowed to purchase the crop and export it. crop collected early and are being paid promptly, The reforms have been gradual. leading them to harvest rather than neglect their In the first season of reform in 1991-92, the gov- cashews. Production rebounded from 29,000 tons in ernment issued only vague guidelines through the 1990-91 to 41,000 tons in 1992-93, and farmers media and sent no official instructions to the local are beginning to rehabilitate their cashew farms and authorities responsible for the policy changes. Private plant new trees. 82 UNLEASHING MARKETS A variant of the third type of marketing system is to allow full com- petition in domestic purchasing but to maintain government control over exporting, as C6te d'Ivoire is doing with cocoa and coffee. In such cases, producer prices would still be effectively linked to the export price set by the government, in contrast to other systems where private traders are allowed to export freely. Assessing the Current Export Policy Stance To compare countries' policy stances toward exporters today, the most appropriate method would be to compare the level of taxation across countries, corrected for reasonable transport and processing costs from the farmgate to the border. This is difficult to do with any accuracy. Instead, we used proxies for explicit and implicit taxation. As a proxy for explicit taxation, countries were ranked according to how much control the marketing boards or parastatals had over prices and marketing costs, taking into account significant overvaluation of the real exchange rate. Some marketing boards are more inefficient than others, of cours., so the degree of control is not always a good indica- tion of the degree of taxation. As a proxy of implicit taxation, the ex- change rate policy stance developed in chapter 2 was used. The results are presented in table 3.2. The conclusion: no country has good macro- economic policies and good agricultural policies. Most have exchange rate distortions (or did at the end of 1991) and government interven- uon in marketing. More Liberal Food Crop Marketing There has been good progress in liberalizing the marketing of the major staple food crops that were subject to extensive government con- trol before the reforms. In countries that once intervened heavily, the disastrous financial position of many food crop marketing boards was a strong impetus for rapid reform. Of the fifteen countries with highly in- terventionist policies before adjustment, only two still intervene in a major way (table 3.3). In countries with less intervention initially, gov- ernments controlled only a small part of the marketed surplus, and there was widespread evasion through parallel markets. Many of the market- ing reforms legitimized the existing situation, thus reducing barriers to private sector entry. 83 ADJUSTMENT IN AFRICA Real exchange rate Marketing policy, late 1992 policy, 1990-91 Mostfavorablea Mixedb Leastfavorable' Good Adequate Kenya Ghana Madagascar Fair Nigeria Burundi None The Gambia Malawi Niger Uganda Zimbabwe Poor or very poor Mozambique Benin Central African Burkina Faso Republic Cameroon Congo C6te d'Ivoire Gabon Mali Senegal Rwanda Tanrzania Sierra Leone Zambia Unclassified Guinea Chad Guinea-Bissau a. Parastatal marketing board eliminated for major export crops. b. Some government intervention in major export crops, but producer prices linked to world market prices. c. Extensive government control (de facto or de jure) over collecting and exporting major export crops; producer prices not directly linked to world market prices. Source: World Bank staff. There was only limited interference in roots and tuber crops in the forest zone before adjustment. For example, much of the domestic pro- duce sold through the government marketing boards in Congo and Guinea was grown on state farms that have since closed or been priva- tized. And although governments were involved in marketing coarse grains in many Sub-Saharan countries, they generally played a limited role. Government intervention was more widespread in domestic rice production and imports of rice and wheat, the major food staples of the urban sector. Almost all the countries that had public sector monopolies on rice imports have decontrolled these monopolies, though many still have monopolies on wheat. Government intervention was the strongest in maize marketing in eastern and southern Africa. Reforms to create efficient, private maize 84 UNLEASHING MARKETS la"Ae@ Z 3 s@Cvor`wm@Eu uz r,vJbml] h-L HL)a3b'oX0,g aEtm ZJaps30 Before Late Country Crop reforms 1992 Benin Tubers 0 G Burkina Faso Millet; sorghum * 0 Burundi Beans 0 0 Cameroon Cassava c 0 Central African Rep. Cassava 0 0 Chad Millet; sorghum 0 0 Congo Cassava C 0 C6te d'Ivoire Tubers 0 C Gabon Cassava 0 0 The Gambia Sorghum; millet * 0 Ghana Tubers 0 0 Guinea Rice 0 0 Guinea-Bissau Rice * 0 Kenya Maize 0 O Madagascar Rice 0 O Malawi Maize * c Mali Millet; sorghum * 0 Mauritania Millet C c Mozambique Maize * 0 Niger Millet 0 O Nigeria Yams 0 0 Rwanda Sorghum 0 0 Senegal Millet; sorghum c 0 Sierra Leone Millet; rice 0 0 Tanzania Maize *C Togo Maize C 0 Zambia Maize * C Zimbabwe Maize * S * Major restrictions on purchases and sales. * Limited intervention by government buying agency. O No intervention except in food security stocks. Source: World Bank staff. marketing have typically taken a long time to implement in these coun- tries and have in some cases been reversed (box 3.5). The common pack- age of reforms includes (a) moving farmer prices toward export or im- port prices, (b) announcing administered prices before planting times, (c) speeding payments to farmers and eventually liberalizing prices alto- gether, (d) relaxing maize movement controls and other restrictions on trade, and (e) restructuring parastatal maize marketing companies. The last of these has typically involved reducing the parastatals' commercial operations to convert them into buyers and sellers of last resort or 85 ADJUSTMENT IN AFRICA COUNTRIES IN EASTERN AND SOUTHERN AFRICA activities inhboth the grain and the milling markets. have reduced the scope of maize marketing boards, Tanzania. Since 1984, when some crop movement but the hrertnwere lifted, Tanzania has progressively mai0zemarkettns. XX0 00\ fuXX0 :i ;0 V0 0; . liberalized 0ithe marketing of food crops. Beginning in Kenya. ReformofKenya'sNationalCereals and0 1987, 00private:traders could purchase grain from tProduce Boardbegan inXL 1982. Theplanwasto cooperativeEs in competition with the National Vimprove farmer incentives by moving to importpan- Milling Corporation, although marketing outlets for :ty ;priciSng, havingthe treasurycompensatetheboard farmeSrs were still confined to the primary coopera- for f: its market development activities, an d relaxing tive societies. In1988the grain trade was fully liber- maize movement controls and trading restrictio n s, aired, andin 1990 the remnaining restrictions on pri- n0 +tat least two oasions,vthegovernment wroteoffvate grain purchase at ;theXf farm level were removed. thehoard'saccumlateddebts,whic at onepointThe corporationeffectierdied injthe 1991 season, had muntedto 5 ercen of CP. whn irsacces to cop fnance ended. But the con- After a:decade of slow progress on the reforms, the tinuing use of the strategiccgrainnreserveeassaamecha- government reversed several of them by retr introducing nisfor price support raises fears that it may evolve f 0maizemovement controals, haSltSing the efots to: lbrl- inato ianother maize board. 0 -ize prices and restructur-e the braovd Private trade st f many informal administra- equate perfrmance on ,and raising hea re tetarsd re tons. Local charges and levies servesi rather thanlowering them-thus falling short of are unevenlyapplied, and f regional authorities con- making thieboard a buyer and selleroflastF resort. tinuetoimpoSse: restrictions;;on private;traders that f000fKenyra has,; however, substantiallybard costs of maeting. In addition, the poor guardians of food security stocks. The easiest measures to implement have been moving producer prices toward export or import parity; the hardest have been restructuring parastatals and negotiating their with- drawal from a mainstream commercial role. Despite almost universal agreement (among donors at least) that a parastatal marketing agency should act only as a buyer and seller of last resort, it has proved difficult thus far to limit parastatals to food security functions, in light of con- cerns about protecting producers' access to markets in remote areas. Removing Fertilizer Subsidies and Liberalizing Distribution Fertilizer use is extremely low in Sub-Saharan Africa-9 kilograms of plant nutrients per hectare of arable land in 1990. This is substantially below the 69 kilograms used in South Asia and the 262 kilograms in China. The paucity of irrigated land in Africa is not the reason; India is estimated to use at least three times more fertilizer on rainfed land than 86 UNLEASHING MARKETS condition of feeder, regional, and trunk roads means agency's buying centers (though fewer than agreed); high transport costs. Besides being hampered by legal and some increase in the difference between retail restrictions, private traders have been unable to ob- prices in private markets and the agency's prices. tain credit, which is still confined to cooperative so- Private traders, who are heavily concentrated geo- cieties for grain purchases. Despite the obstacles, pri- graphically, confine themselves to short-haul consol- vate traders have, in the context of partial idation of loads for the agency. Regulatory and other liberalization, provided a market for farmers in acces- barriers keep them shut out of private export trade. sible locales, improved food supplies, and stabilized They face severe transportation problems and limited food prices in urban areas. The increased competi- access to credit. And they are still subject to official tion so far, though still restricted, has already reduced inspections on the roads, often requiring unofficial profit marginis in private trading. side payments that increase their cost of doing busi- Malawi. The government of Malawi undertook a ness. The marketing agency, by contrast, continues series of agricultural marketing reforms in 1986 to to operate a network of buying points only slightly address the financial crisis of the produce marketing less dense than before 1986. It cross-subsidizes its agency, increase the efficiency of agricultural mar- functions from lucrative export activities, to the keting, and allow prices to reflect more accurately detriment of private traders, and it is still far from the regional and seasonal demand for and supply of being a buyer and seller of last resort. Despite these agricultural commodities. Achievements so far barriers, liberalization measures appear to be taking include a significant expansion of private maize hold. In the 1992 season, private traders took a con- trade, with more traders operating outside their siderable share of the market and paid producers sub- home areas; the closure of some of the marketing stantially more than the marketing agency did. Sub-Saharan Africa does. More important deterrents to fertilizer use have been supply shortages and inefficient distribution systems, out- comes of extensive government intervention. Fertilizers are generally expensive in Africa because of small procure- ment lots, inefficient marketing by public agencies, and high shipping, handling, and dlomestic transport costs. Under such conditions, fertil- izer subsidies have had strong political appeal. Proponents argue that fertilizer subsiclies can speed the adoption of agricultural innovations. Opponents contend that large subsidies reduce fertilizer use, because with limited funds available to finance the subsidies, only a limited amount of fertilizer is available. Moreover, subsidies disproportionately benefit well-off farmers able to secure rationed supplies. They also argue that public funds would be better spent on activities that improve the re- turns to agriculture, such as rural infrastructure and research and exten- sion to develop and disseminate more fertilizer-responsive crop technologies. 87 A S T:[N AFRICA Efforts to eliminate subsidies have proceeded more slowly than ex- pected, and progress has frequently come only after extended negotia- 7?,e 3i eA r vnres and tions between governments and donors. Despite the problems, about .Fu ;rdis 'Drr rertifizer half of the adjusting countries have completely removed fertilizer subsi- Before Late dies and liberalized their distribution systems (table 3.4). Subsidies have Country reforms 1992 sometimes been lifted abruptly, usually when a devaluation suddenly Benin o eliminated indirect subsidies of imported fertilizer, and sometimes grad- Burkina Faso * O ually, as with the easing of price controls or the phased removal of specific Burundi 0 subventions. With subsidies reduced or eliminated, and most parastatal Cameroon * 0 . Central African or cooperative fertilizer credit programs having abysmal loan repayment Republic * Q records and being shut down, governments have allowed the private sec- Chad 0 0 tor to take a greater role in importing and distributing fertilizer. Congo - _ Fears that removing fertilizer subsidies would reduce fertilizer use and Cote d'Ivoire O 0 Gabon _ lower crop production have proved to be unfounded (box 3.6). Such The Gambia * o fears are often based on analysis that overlooks the potential for private Ghana o 0 marketing to develop and for marketing efficiency to be enhanced. Guinea 0 0 Often, the more widely available supply (as the rationing of subsidized Guinea-Bissau C 0 fr Kenya 0 fertilzer iS reduced) and the better pr0ce icentives for output counter Madagascar * O the disincentives of removing the subsidy. Furthermore, because subsi- Malawi * dies do not affect the marginal cost of production where fertilizer is ra- Mali * O0 . Mauritania * toned, removing the subsidies tends to have little effect on production, Mozambique - O at least in the short term. Niger 0 0 Nigeria * O Rwanda 0 Mr Senegal . 0 Letting Other Markets Work Sierra Leone 0 o Tanzania 0 o Togo * ° EFORE STARTING TO ADJUST, AFRICAN COUNTRIES HAD Uganda 0 1g)) Zambia * o extensive systems of price controls and other controls on goods, Zimbabwe 0 0 labor, and interest rates. Evidence is scarce, but widespread par- allel markets and rampant evasion suggest that controls on goods gen- * Marketing controlled and prices subsidized. eraliy dd not keep prices substantially below market levels. Indeed, O Marketing controlled, but at some controls, maintained by according monopoly privileges to world prices. O Marketing liberalized, but some importing or producing parastatals, kept prices substantially above fertilizers sold at below-market import prices (as with cement and wheat in Congo, for example). In prios or prices controlled many cases, controls provided rents and benefits for privileged elites C No controls on prices or markering. instead of protecting consumers. - Data not available. Reform programs have also included other regulatory and legal mea- Source: World Bank staff. sures to promote private activity, such as reforms of company and bank- 88 UNLEASHING MARKETS Sox , 3.6 INha i$Esm>X$ Ferive @Subsf id3ies Would3 Me!an THOSE WHO SUPPORT FERTILIZER SUBSIDIES the 1989 season. The average subsidy fell from 30 per- argue that cutting them would reduce fertilizer use cent in 1983-84 to 20 percent in 1987-88. But when and undermine food production. Others charge that rising transport costs and devaluation of the exchange subsidies result in widespread shortages and poor dis- rate led to substantial increases in imported fertilizer tribution. What does the evidence show? Tanzania prices, the attempt to reduce the subsidy was aban- and Malawi are finding that removing large subsidies doned, and it returned immediately to 30 percent. can ease supply constraints and result in greater fer- The government has since renewed its pledge to tilizer use, despite higher prices. In Nigeria, mean- phase out subsidies, this time by 1994-95. Fertilizer while, heavy subsidies have led to supply shortages sales have continued to rise, even in the face of sub- that are choking agricultural growth. stantial increases in their official prices and stagnat- Tanzania. The government of Tanzania aims to ing producer prices for smallholders. Between 1980 eliminate fertilizer subsidies by 1995. Already it has and 1991, fertilizer sales to smallholders more than reduced subsidies from about 80 percent of the doubled, from 49,000 to 107,000 tons. This meant a farmgate price in 1989 to about 40 percent in quadrupling of plant nutrients, as the government 1993. Critics of the subsidy cut have worried that it encouraged the use of higher-analysis fertilizers. It would undermine efforts to increase maize produc- also reflected an easing of supply, due in part to aid tion in Tanzania's Southern Highlands, an area and probably some leakage of fertilizer from the that receives a substantial portion of the subsidized smallholder sector to estates. fertilizer. So far, fertilizer use has not fallen, since Nigeria. Nigeria briefly reduced fertilizer subsi- supply previously was limited because of insuffi- dies-from 85 percent in 1983 to 28 percent in cient funds to cover the subsidy. Although the mar- 1986. However, when the currency was steeply ket for fertilizer supply has been liberalized in theo- devalued, the government failed to adjust fertilizer ry, the government agency that subsidizes fertilizer prices. Subsidies rocketed to about 90 percent of the has effectively become the sole supplier. Anecdotal price, and by 1989 fertilizer subsidies were consum- field evidence suggests that farmers have recently ing more than 70 percent of the central govern- paid prices about 30 to 40 percent higher than ment's entire agricultural budget. The public sector those reconmmended. Because of the large excess could not meet demand at the subsidized prices. demand at current subsidized prices, removing the Fertilizer supplies became inadequate and unpre- subsidy and getting the government out of the mar- dictable, and a large parallel market developed in ket is projecited to increase fertilizer use. which small farmers paid four times the official Malawi. Irn Malawi in 1983, the government com- price. The fertilizer shortages-the result of massive mitted to phasing out fertilizer subsidies completely by subsidies-now constrain agricultural growth. ruptcy laws an(i investment codes. Many countries have attempted to simplify investment procedures through one-stop investment shops, but numerous institutional problems remain to be ironed out. Several coun- tries (notably Cote d'Ivoire) have set up antitrust regulatory bodies, and other legal and judicial reforms are under way. But it is too early to as- sess how successful these reforms are and what impact they will have on private activity. Of even greater importance than these piecemeal re- 89 ADJUSTMENT IN AFRICA forms may be the overall business climate, which in many countries re- mains hostile to private business (box 3.7). Price Controls Largely Eliminated Most countries have removed almost all price controls, keeping them for only a few strategic goods (table 3.5). Nine countries have lifted price controls on all goods except refined petroleum. The countries re- taining substantial price controls include Burkina Faso, Cameroon, C6te d'Ivoire, Kenya, Madagascar, and Mozambique, but even these have sharply reduced the number of goods subject to control. In AFRICAN GOVERNMENTS SOMETIMES SHOOT THEMSELVES IN thefoot byta or openly encouraging the harassment of successful entrepreneurs-harassment that often stems from public antagonism toward the business class. Some businesspeople clearly do engage in corrupt acts harmful to the state and society, and campaigns to pros- ecute taxevaders and others who violate corruption statutes are prob- ably necessary, particularly If such prosecution recaptures revenue due the stte. But thejadvantages: must be weighed carefully, for overzeal- ous: administrative procedures can ruin profitable businesses and de- 0000stroy th inetv to invest.; In the worst instan bridled harassment of entrepreneurs is sanctioned by law. In Gaa,where it is illegal to carry out "any act with intent to sabotage the economy," the law has on occasion been applied to the government's political adversaries or anyone enjoying economic gains. Those accused of economic sabotage are subject to imprisonment without trial and the freezing of their assets. Cases are heard ouitside the regular jtudicial system, and proceedings are not published. Those found guilty may be executed; there is no external appeal mechanism. Large foreign companies, with ample capital and the option of going elsewhere, often have the economic clout to protect themselves from regulatory harassment. But local firms and smaller expatriate firmsusuLlly lack such leverage, making them the main victims. If the envi+ronment does norchange, many will find a way to pull up stakes-depriving Africancountiries of the capital, business acumen, andotechnical expertise essential for development. 90 UNLEASHING MARKETS Mozambique the proportion of GDP subject to price controls has hbize 3.5 lHos eq Ga-: k Az dropped from 70 percent to 10 percent since 1986, although controls Before Late are still extensive. Gounty reforms 1992 Because most controls were ineffective, their removal probably had Benin * 0 little impact on market prices, although there is not much evidence on Burkina Faso 0 O this. But they were probably still a source of rent, and removing them Burundi 0 O should have reduced transaction costs and facilitated new entry, thus Cameroon * i stimulating growth and greater competition. The largest benefits have Republic * 0 been the increased incentives to agricultural production and the greater Chad 0 0 availability of a few key imports for which controls had been more ef- Congo 0 ;3 fective, resulting in import rationing. Gabon * (D The one sector where prices have not been fully decontrolled is pe- The Gambia 0 0 troleum, and government intervention in that sector is very costly. Ad- Ghana * 0 justment programs have attempted to improve procurement, eliminate Guinea- a (0 panterritorial pricing, and reform pricing practices so that ex-depot Kenya 0 0 prices and pump prices are quickly adjusted to reflect changing interna- Madagascar * 0 tional price and exchange rates. Although procurement practices have Malawi 0 (0 improved in several countries, further improvements in procurement Mauritania * 0 could bring significant savings (box 3.8), exceeding annual disburse- Mozambique * 0 ments of World Bank adjustment loans to Africa. Furthermore, al- Niger 0 0 though many countries have reformed their pricing structure to reflect Nigeria 0 Rwanda S international prices, panterritorial pricing and fixed pump prices remain Senegal 0 the norm. The benefit of a system in which pump price changes are set Sierra Leone * 0 automatically according to some pre-agreed formula is that it prevents Tanzania * 0 the government from collecting windfall revenues when import prices Togo G 0 Uganda 0 drop. It also makes it easier to pass along price increases when import Zambia 0 O prices rise. Only Mali has moved to eliminate all price controls on pe- Zimbabwe *S 0 troleum products. Rwanda, Uganda, and more recently Madagascar * Extensive controls (on twenty- have automatic procedures for adjusting petroleum prices. Several other six goods or more). countries-Burundi, Malawi, Mozambique, Niger, and Togo-regu- 0 Limited controls (on ten to twenty-five goods). larly update their ex-depot prices even though they do not have auto- 0 Few conrrols (on fewer than matic procedures in place. ten goods). Reform is lagging in countries that refine imported or domestic crude 0 No controls. I ~~~~~~~~~~~~~~~~~~~~~Note: Price controls on refined petro- petroleum. In most, the ex-refinery price to wholesalers is not set ac- leum products were excluded from con- cording to an import parity formula but rather on a cost-plus basis. sideration. Source: World Bank staff. Consumers thus pay for the inefficiencies of refineries (or the state pays in forgone tax revenue). Distributors are required to buy all or part of their refined products from government parastatals, often paying more than they would otherwise. In Kenya, for example, the five privately 9I ADJUTME ENT IN AFRICA LARGE INEFFICIENCI[ES IN THE PROCUREMENT, their limited foreign exchange, poor credit standing, refining, and distribution of petroleum products in and inappropriate bidding procedures symptomatic of many Sub-Saharan countries make reform of the pe- monopoly control and government interference. The troleum sector urgent-and potentially fruitful. Ex- inefficiencies in procurement are greatest in the coun- cept for Angola, Cam-neroon, Congo, Gabon, and tries where governments are most heavily involved in Nigeria, the region's countries depend heavily on im- importing petroleum (box figure 3.1). ported crude and refined petroleum products. Petro- Another 40 percent of the savings would come leum imports absorb 20 to 35 percent of export earn- from improvements in refining. Refineries are poorly ings, while domestic wholesale and retail sales maintained and underused, the refined yields are in- : generate about 40 percent of government revenue sufficient to satisfy domestic demand, and the prod- (Schloss 1993). For rmany African countries, petro- uct mix is unsuitable for local markets. Locally refined leum imports are the single largest item in the balance products thus are not competitive with imported of payments, and taxes on domestic petroleum sales products. are the biggest source of revenue. For the region as a The rest of the cost savings would come from bet- whole, a rational system of oil procurement and dis- ter distribution. Poor storage, dilapidated infrastruc- tribution could save about $1.4 billion a year (at ture, inadequate market competition, and the exten- 1989-90 prices)-more than the World Bank's an- sive use of road transport all mean inefficiency. nual disbursements of adjustment policy loans to About two-thirds of the total savings would be Sub-Saharan Africa. achieved through new operating procedures, better Almost half the savings would come from improve- institutional arrangements, and the closing or retool- ments in procurement. African countries typically pay ing of inefficient refineries. The rest would require too much for their imported petroleum because of new investment in infrastructure. owned petroleum distributors are required to procure 30 percent of the crude to be refined domestically from the National Oil Company. Relaxing Labor Controls The rigidities in hiring and firing workers and the practices for set- ting wages in some countries have raised production costs, reduced pro- ductivity, and stifled investment and job creation. Labor regulations protect the small minority of workers fortunate enough to work in the formal sector. The excessive cost of formal sector jobs limits job growth in that sector and compels the remainder of the rapidly growing labor force to seek employment in the informal sector. To enhance competi- tiveness and increase growth, adjustment programs have aimed at in- creasing the flexibility in labor markets. There has been little systematic study of labor market regulations in Africa or of their impact on eco- nomic activity, so it is difficult to evaluate the costs of various restric- 92 UNLEASHING *MARKETS E2 7% g Roun . '- OY',,A D8MenleifoP -d th P>etrrienim Enduetry !Supply Onelffid^'ency, 2and PmentIMS nxng.zSJ;3iC Asg.9c.agnnpenatr Xlefanms, A;1g Zambia $79 Cameroon $74 Tanzania $73 l Ethiopia $72 ~ W6e d'lvoire $27 Burkina Faso $24 Niger $12 " Potential savings Central African Rep. $10 X (millions of 1989-90 dollars) Equatorial Guinea $3 Zimbabwe $35 . ~~~~ : - ~~SMalia S1 .0 Nigeria $4 Madagascar $3 Ghana $1 Angola Sl Uganda $9 Botswana $63 Kenya $ Namibia $4 Mauritania $ Swaziland $4 Djibouti $ Lesotho $: Togo $1 Mali. .1/ Rwanda $6 a. Supply inefficiency refers to procurement costs beyond those incurred in world mark-ets under competitive conditions. Low ineffi- ciency is defined as excess costs of $0-40 per ton; medium, $41-100 per ton; and high, more than $100 per ton. b. Little government intervention is defined as 0-25 percent of crude oil and/or finished petroleum products imported by government agencies; mediutn intervention, 26-75 percent; and heavy intervention, more than 75 percent. Source: Schloss (1993). 93 AJTE IN AFRICA tions and the benefits of the reforms. Most of the regulations are be- lieved to be more of an implicit tax than an absolute barrier. Firms usu- ally find ways-sometimes costly-of getting around regulations. The fixed exchange rate countries generally had more restrictive labor laws than the flexible exchange rate countries. This hampered their in- ternal adjustment strategy, which was predicated on reducing real wages through nominal wage declines. Programs to remove restrictions on hir- ing and collective layoffs and to reform wage-setting procedures met with varying success. Mali adopted a new labor code in 1992, substan- tially liberalizing the labor market. In C6te d'Ivoire and Senegal, new labor codes were drafted after a long process of government consultation with employers' associations and the labor unions on labor market re- form. Although the new codes have still not been officially adopted, both countries have already changed some provisions of the old codes. Removing restrictions on hiring. In 1987 Senegal abolished a govern- ment employment agency that had a legal monopoly over placement and hiring; C6te d'Ivoire did likewise in 1992. Previously in C6te d'Ivoire, daily or seasonal workers who worked for more than ninety days continuously or intermittently for the same employer had to be offered full-time positions. Under the reform program, seasonal work- ers are exempt from this provision, and the obligation to regularize daily workers now comes into effect only after they have worked for one year. Daily workers receive a small salary premium prorated to equal the benefits that full-time workers would receive. In Senegal, restrictions on renewing temporary labor contracts were progressively lifted, and temporary contracts can now be extended for five years. For firms in the industrial free zone and those benefiting from the invest- ment code, the period is unlimited. Removing restrictions on collective layoffs. Collective layoffs in Cote d'Ivoire are subject to priority authorization by the Ministry of Em- ployment. A new law was drawn up to change this, but it has been challenged by the unions. In Senegal the drafted labor code has sub- stantially liberalized firing practices for small and medium-sized enter- prises, with prior government authorization no longer needed. In Mali the right of an employer to lay off workers for economic causes is liber- ally defined to include economic distress, restructuring of the work process, and technological change. Reforming wage-setting practices. The draft labor code in Senegal calls for (a) delinking different industry-specific wage scales from the mini- 94 UNLEASHING MARKETS mum wage so that there is no automatic adjustment of the minimum wage increase, aLnd (b) encouraging collective bargaining at the firm level. C6te d'Ivoire has not yet tackled wage-setting practices. The Mali labor code sets minimum wages for different vocational cate- gories. It is not clear whether these minimums are binding, however, and thus what cheir impact is on wage rigidity and employment growth. Demonopolization Proceeding Apace Before reform, the private sector was prohibited-either by law or by practice-from competing with public enterprises in many key sectors (appendix table A. IO). Many of these state monopolies have since been eliminated (appendix table A. 11), either through the abolition of paras- tatals or through the liberalization of production or imports of key goods, both of which greatly increased the scope for private sector activ- ity. Reform has been particularly noteworthy in the import and distri- bution of fertilizer, the import of key consumer goods, the liberalization of labor hiring, arid commercial banking services. In addition: • More than two-thirds of the countries with monopolies on rice imports and a third of those with monopolies on wheat imports have eliminated them. * Eleven of rwenty-seven countries have eliminated the monopoly on export erops, and half the countries have ended the monopo- lies on the distribution of petroleum. _ TelecommLunications remains a public monopoly in almost all countries, although several have allowed private cellular phone services. Establishing a Sound Incentive Framework: Mixed Results A SSESSING GOVERNMENT INTERVENTION IN PRICES AND marketing is difficult because there is no standard indicator. We constructed a broad index to measure interventionist poli- cies based on three factors: (a) the government's interference in setting producer prices for the major agricultural export commodity, (b) the 95 AJU:STMENT IN AFRICA degree of competition in five major sectors (petroleum imports and wholesale supply, retail distribution of refined petroleum products, fer- tilizer imports and/or distribution, wheat and/or rice imports, and the domestic marketing and distribution of the major food staple), and (c) the extent of price controls. Appendix tables A.12 and A.13 present the classification in more detail. Twenty-five countries were rated as having heavy intervention in the pre-reform period. By late 1992, however, only four were in that category; most of those previously rated "heavy" had moved into the "medium" category (table 3.6). Almost all adjusting countries have more to do, however, to provide a stable set of incentives for private sector-led development. To evaluate the overall incentive framework in Sub-Saharan Africa, it is useful to consider market intervention policies in conjunction with Before Late Before Late Country reforms 1992 Country reforms 1992 Benin * O Madagascar * 0 Burkina Faso * * Malawi * C Burundi * C Mali 0 C Cameroon * C Mauritania * C Central African Mozambique * C Republic * * Niger 0 C Chad C 0 Nigeria 0 0 Congo 0 0 Rwanda C C C6te d'Ivoire 0 C Senegal 0 C Gabon 0 C Sierra Leone 0 0 The Gambia * 0 Tanzania 0 C Ghana 0 C Togo 0 C Guinea * 0 Uganda C 0 Guinea-Bissau 0 0 Zambia * C Kenya 0 C Zimbabwe C C * Heavy intervention (government control over producer prices of agricultural exports, private sector competition restricted in key sectors, and fairly extensive price controls). * Medium intervention (some government involvement in producer price setting, government monopolies in one or more key sectors, and some price controls). O Little intervention (no government involvement in producer price setting, private sector competition allowed in key sectors, and liberalized pricing of all goods other than petroleum products). Nlote: Sectors include petroleum importing and wholesale supply, retail distribution of refined petroleum products, fertilizer importing and/or distribution, wheat and/or rice importing, and domestic food-crop marketing. Sources: Appendix tables A. 12 and A. 13. 96 UNLEASHING MARKETS macroeconomic policies (fiscal, monetary, and exchange rate policies).7 As table 3.7 shows, no country in the region has both little market in- tervention and a good or even adequate macroeconomic framework (if one assesses the framework according to the criteria developed in chap- ter 2). Most countries are still clustered in the groups with medium in- tervention and fair, poor, or very poor macroeconomic policies. Still, it is encouraging to realize that if such a table had been con- structed in 1985, almost all the countries would have been classified as having heavy government controls. As for the macroeconomic frame- work, all the countries outside the CFA franc zone would have been judged as having unsound policies. Most of the countries in the franc zone would have garnered better macroeconomic ratings, but even in 1985 there were indications that some had an unsustainable fiscal situa- Macroeconomic Government intervention in markets, late 1992 policy stance, 1990-91 Limited Medium Heavy Good Adequate Ghana Fair The Gambia Burundi Burkina Faso Nigeria Gabon Madagascar Uganda Kenya Mali Malawi Mauritania Senegal Togo Poor or very poor Sierra Leone Benin Central African Cameroon Republic C6te d'Ivoire Congo Mozambique Niger Rwanda Tanzania Zambia Zimbabwe UnclassiFied Chad Guinea Guinea-Bissau Sources: Table 2.5 and appendix table A.13. 97 NADJSTEN I AFR ICA tion. Thus the majority of countries would have occupied spots near the bottom right of the table. That a number of countries have moved up and left testifies to their progress in creating a policy environment more conducive to growth and poverty reduction. Some countries' macroeconomic situations have also changed since 1990-91, the latest years for which standardized data are available. In Madagascar, for example, policies deteriorated significantly in 1992-93. In Sierra Leone, by contrast, the developments have been positive: infla- tion has plummeted, capital and social spending have increased, and real interest rates have been positive since mid-1992. No country appears to have reached the "good" macroeconomic classification, however. This underscores the fragility of the policy reform efforts. Notes 1. The estimates of notional demand are based on a nuts, cotton, and tea; Niger-cotton and cowpeas; country's GDP level and import prices at official exchange Nigeria-cocoa; Rwanda-coffee and tea; Senegal-cot- rates. See note "a" in figure 3.2 for more details. ton and groundnuts; Sierra Leone-cocoa, coffee, and palm kernels; Tanzania-coffee, cotton, and tea; Togo- 2. Interestingly, the predicted level was less than the cocoa, coffee, and cotton; Uganda-coffee and cotton; actual level after 1987, an indication that the method for Zambia-cotton and tobacco; and Zimbabwe-cotton, estimating notional demand is imperfect. sugar, and tobacco. The periods compared were 1981-83 and 1989-91. The intervening years were avoided be- 3. Burkina Faso, Mali, Mozambique, Senegal, and cause the mini-booms in coffee, cocoa, and tea in 1984 Zambia have also reduced their maximum tariff rates and 1986 would have skewed the results. significantly. 5. The nine countries that took measures to benefit 4. Because of the difficulty of establishing the price farmers were Burundi, the Central African Republic, that producers actually received for their crops, the data in Congo, Gabon, Kenya, Malawi, Sierra Leone, Uganda, figure 3.3 are indicative only of broad trends. The crops and Zimbabwe. The other eight countries whose policies used to calculate price data for each country were as fol- worked against farmers were Cameroon, Chad, C6te lows: Benin-cotton; Burkina Faso-cotton; Burundi- d'Ivoire, The Gambia, Guinea-Bissau, Rwanda, Senegal, coffee, cotton, and tea; Cameroon-cocoa, coffee (ara- and Zambia. bica and robusta), and cotton; the Central African Republic-coffee and cotton; Chad-cotton; Congo- 6. For a discussion of the measure of overall taxation, cocoa and coffee; C6te d'Ivoire-cocoa and coffee; see appendix C. Gabon-cocoa and coffee; The Gambia-cotton and groundnuts; Ghana-cocoa; Guinea-Bissau-cashew 7. Because the index of macroeconomic policy stance is nuts, groundnuts, and palm kernels; Kenya-coffee and based in part on the parallel market exchange rate premium, tea; Madagascar-cloves, coffee, and vanilla; Malawi- it also captures certain elements of trade policy stance. tea and tobacco; Mali-cotton; Mozambique-cashew 98 CHAPTER, 4 Refcorming the Pub]lic Sector _ HE PUBLIC SECTOR LIES AT THE CORE OF THE STAG- nation and decline in growth in Africa. By the mid- 1980s, the public sectors in most Sub-Saharan countries had taken on too much. They were inter- vening-with poor results-in activities where mar- kets work reasonably well, such as allocating foreign exchange and directing credit. They were also doing a poor job of providing such essential services as roads and primary schools. Several factors account for the overextension and poor performance of the public sector. The economic crisis of the early 1980s strained Africa's already-limited technical and institutional capacities. Declining growth reducecL the tax base, depriving governments of the resources to pay competentw staff, improve infrastructure, and deliver social services. The economic and political conditions prevailing when African govern- ments gained power also played a part. The indigenous private sector was very weak in Africa, and nonindigenous elites and foreigners rela- tively dominant. Given those circumstances, it was understandable that governments took an interventionist stance. The then-dominant devel- opment paradi gm supported this: the prevalent view was that develop- ing countries could grow faster by taking an active role in production and by directing the allocation of resources to bring about faster industrializatic) n. The failures of this strategy are manifest everywhere, but conditions have not changed radically. Public sector institutions remain weak, per- haps even weaker now that the economic crisis has taken its toll. The in- digenous private sector continues to be economically weaker than other, more established private sector groups. And reform has become even more difficult, as the rents and other side-benefits of interventionist 99 ADJUSTMENT IN AFRICA policies have created powerful public sector constituencies with a strong interest in their perpetuation. Now more than ever, fundamental reform of the public sector is needed to reverse the economic decline, yet it has become more difficult to achieve. World Development Report 1991 (World Bank 199 1c) discussed the emerging consensus on the need for a market-friendly approach to de- velopment. That approach calls for governments to do less where mar- kets work reasonably well and to do more where they don't. Four types of government action are especially vital: investing in human capital, providing a competitive climate for enterprises, opening markets to in- ternational trade, and ensuring stable macroeconomic management. Chapters 2 and 3 assessed the progress of countries in undertaking the policy reforms needed to provide a stable macroeconomic founda- tion, unleash markets, and participate more fully in international trade. This chapter continues the analysis of reform efforts by examining the closely linked public enterprise and financial sectors, where reform is important for ensuring macroeconomic stability and for spurring pri- vate sector-led growth. We also look at efforts to generate and reallocate the resources needed for the state to carry out its essential functions. We do not, however, discuss building institutional capacity within the pub- lic sector-a long-term development challenge beyond the scope of ad- justment programs.1 Reform efforts generally have not been too successful in the areas where the state has intervened most heavily. Privatizing and reforming state-owned enterprises and creating sound financial systems have proved to be among the most difficult of adjustment reforms. Public en- terprise reforms have not yet leveled the playing field for the private sec- tor because they have not significantly curtailed the public enterprises' privileged access to the budget, to the credit system, to tariff and non- tariff protection, to special tax status, and to regulatory protection. With financial reforms, there has been movement toward reasonable interest rates and the restructuring and privatization of banks. Nonetheless, large fiscal deficits, poorly performing public enterprises, and continu- ing political interference still threaten to undermine development of the financial system. Some countries have made progress rationalizing public expenditure, trimming the wage bill, and increasing resource mobilization. But there is little evidence that the allocation of resources within sectors or be- tween sectors has improved substantially in favor of basic services. The 100 REFORMING THE PUBLIC SECTOR long-run development tasks of creating the control structures and oper- ating procedures to manage a modern and efficient civil service have barely begun. A. precondition for this may be strengthening the institu- tions of civil society to create demand for better governance. Public Enterprise Reform: Little Evidence of Significant Progress T HE PUBLIC ENTERPRISE SECTOR PLAYS TOO BIG A ROLE IN economic activity and employment in all twenty-nine of the adjusting countries in our study. In the 1970s, as African gov- ernments tried to accelerate development, the number of public sector firms mushroomed, and parastatal enterprises came to dominate many key areas of the economy. Statistics on the number of public firms, their share of formal sector employment, and their share of value added are limited and unreliable, but they provide some insights into the public enterprise sector before the reform period. At least fifteen of the adjusting countries had seventy-five or more public enterprises by the early 1980s.2 Ghana, Mozambique, Nigeria, and Tanzania had more than 300. The public enterprises accounted for over 20 percent of formal sector wage employment in a number of countries.3 In Congo, C6te d'Ivoire, Guinea, Kenya, Mozambique, Nigeria, Tanza_ria, and Zambia, and possibly other countries as well, public enterprises accounted for more than 10 percent of GDP. They also accounted for a large share of public investment, domestic credit, and external loans. Quantifying these shares is difficult. First, domestic credit statistics for many countries include net domestic credit to public enterprises as part of private sector credit, making it impossible to distinguish the share of credit to the public enterprise sector. In Burundi, Chad, Ghana, Guinea-Bissau, and Zimbabwe (about half the countries for which dis- aggregated data were available), the stock of credit to the public enter- prises was as bi . as or bigger than credit to the private sector.4 In Tanza- nia in 1988, fewer than twenty public enterprises held more than three-quarters of the government-owned National Bank of Commerce's portfolio of loans over $100,000. Second, external debt statistics under- estimate the share of lending to public enterprises, because many gov- IOI ADJUSTMENTR IN ARICA ernments on-lent external funds to the public enterprises. Third, public enterprise investment is extremely difficult to estimate, because it has not been systematically monitored and is financed from a variety of sources. Public enterprises have been a major drag on the fiscal budget, the banking sector, and other quasi-fiscal sources of revenue (stabilization funds and social security funds). The returns on public capital invested in these enterprises were very low, and in many cases probably negative. Lacking proper accounting systems, most countries had little idea of the poor returns, although the opportunity cost of subsidies and equity con- tributions to public enterprises was frequently enormous. Financial sup- port for public enterprises reduced the funds available for basic social services, crowded out private sector borrowing, and required higher tax rates, undermining the profitability of the private sector. Moreover, to improve the profits of public enterprises, governments adopted high trade tariffs, imposed nontariff trade barriers, and gave public enter- prises preferential access to foreign exchange, special tax exemptions, and other regulatory advantages (such as monopolies over domestic pro- duction and imports). All in all, the special advantages accorded to pub- lic enterprises produced an environment unfriendly to private enter- prises and made it difficult for them to compete. Most African governments began public enterprise reform primarily to ease the fiscal burden and, usually secondarily, to increase efficiency. The first step in the reform process, difficult in itself, was to establish an inventory of public shareholdings. Countries then classified enterprises as strategic or nonstrategic, and the nonstrategic ones as commercially vi- able or nonviable. The viable nonstrategic enterprises were slated for pri- vatization; those judged nonviable, for liquidation. Strategic firms- generally utilities, telecommunications, major transport parastatals, heavy industries, and agricultural marketing boards-were exempt from divestiture because of their economic importance. This was unfortunate, because they were the major cause of losses and economic distortions. Many governments commissioned studies of the strategic enterprises to identify ways of improving their efficiency and profitability. A number signed performance contracts with key enterprises. Countries also adopted institutional reforms to clarify their regulatory and legal frame- works and increase enterprise autonomy. And some adjustment programs tried to limit fiscal transfers and reduce public enterprise borrowing. 102 REFORMING THE PUBLIC SECTOR Assessing Public Enterprise ReformFeIE YaWietZ a' iMzgr Padb3]V. rrnt:rpores3 The available data on the public enterprise sector are sparse and dis- Late 2992 appointing, showing no significant reduction in the number of enter- Timeliness prises, little improvement in their financial performance, unacceptable Country of audits returns on government investment, and inability to meet the demand Benin c for cost-effective, efficient provision of public utilities. Divestiture is Burkina Faso 0 proceeding slowly among small and medium-sized firms and scarcely at Burundi C all among large enterprises-not surprising since most big enterprises Central African Rep. 0 were classified as strategic. Spotty evidence suggests that financial flows Chad C from governments to public enterprises remain high, and there is some Congo c danger that direct fiscal transfers are being replaced by less obvious sub- CGte abvoire C sidies, such as bank financing. With a few notable exceptions, perfor- The Gambia 0 mance contracts and other attempts to boost the efficiency of enterprises Ghana 0 remaining under state ownership have failed. Parastatal accounts, par- Guinea 0 Guinea-Bissau o ticularly those of the strategic enterprises, are rarely audited by indepen- Kenya C dent accountants, and in most countries there is a long lag in producing Madagascar 0 any accounts a- all (table 4.1). In Ghana in 1990, information was avail- Malawi 0 able for only 70 of some 300 public enterprises. Few countries monitor Mali 0 Mauritania 0 credit to the public enterprise sector. Often there is no systematic ac- Mozambique 0 counting of external funds on-lent to the public enterprise sector, nor Niger C any systematic tracking of repayments and arrears. And public enter- Nigeria C prises continue to enjoy advantages provided by a regulating framework Senegal 0 that protects them from competition. Sierra Leone C The limited progress thus far suggests that there is an urgent need to Tanzania 0 rethink the approach to reforming the public enterprise sector in Africa, Togo O Uganda - particularly the major enterprises. Opposition to more far-reaching di- Zambia 0 vestiture progiams appears to be subsiding, and more ambitious pro- Zimbabwe 0 grams are on the drawing board, but the changes are too recent to have o No regular annual audits. yielded concrere results. C Some firms audited annually. Slow progress in reducing the size of the public sector. At first glance, priva- o Many firms audited annuafy. tization efforts appear to have been moderately successful. Almost all countries have managed to halt the increase in public enterprises, and several have begun to reduce the number. Throughout the region, reform has led, to the privatization of hundreds of public enterprises, mostly small and medium-sized but also a few large ones. Among the twenty-nine acdjusting countries, governments have divested themselves of about 550 firms-less than one-fifth of the total number of public enterprises-either by selling their holdings or by liquidating nonvi- 103 ADJUSTMENT IN AFRICA able enterprises, with slightly more privatizations than liquidations (Shaikh, Kikeri, and Swanson 1993). But progress has been uneven (table 4.2). Six countries-Benin, Ghana, Guinea, Mozambique, Nigeria, and Senegal-account for two- thirds of the divestitures. Only a handful of countries have divested more than 40 percent of their enterprises. And half the countries have been extremely slow to privatize any enterprises. In Kenya there have been almost no sales in ten years. A few countries have even expanded their public enterprise sectors. In Burundi the public enterprise sector grew during the adjustment period, with five firms divested but twelve new ones created. In Cote d'lvoire the number of public enterprises rose from 113 in 1977 to 140 in 1990, despite the privatization of some 30 enterprises in the mid-1980s. More important, the number of divestitures overstates the extent of privatization. Large enterprises with the bulk of public assets-airlines, railroads, mining, and utilities-have generally not been privatized. This is beginning to change, however. Ghana has recently privatized Percentage Number of enterprises before divestiture of enterprises More than divested 0-50 51-100 101-200 200 0-10 The Gambia Burkina Faso Cameroon Kenya Mauritania Congo C6te d'Ivoirea Tanzania Rwanda Uganda Malawib Sierra Leone Zambia Zimbabwe 11-25 Chad Burundi Madagascar Ghana Central African Mozambique Republic 26-40 Niger Guinea Nigeria 41-60 Guinea-Bissau Benin Mali Senegal Togo Note: Divestitures include partial sales, but not management contracts or leases. a. Data are for 1989-92. Some thirty transactions in the 1980s are excluded. b. Total numiber of enterprises inldudes 121 statutorv bodies and 18 commercial paras- tatals. Sources. Shaikh, Kikeri, arnd Swanson (1993); World Bank staff estimates. 104 REFORMING THE PUBLIC SECTOR gold and diamond mines, a brewery, and several medium-sized manu- facturing units, and Nigeria has sold several hotels and a gasoline distri- bution company to private investors. In most countries, including most of those where the government has sold many firms, the focus on small and medium-sized enterprises means that the size of the public enter- prise sector has hardly changed. The four public enterprises sold in Cote d'Ivoire in 199-1 accounted for a mere 0.1 percent of the government's holdings. Possible exceptions are Benin, Guinea, and Nigeria, where some 5 to 15 percent of government shareholdings have reportedly been privatized (Shaikh, Kikeri, and Swanson 1993). Further complicating the picture are problems with the data, as Berg (1993) documents. Sometimes enterprises are reported as sold when they are merely up for sale. Sometimes sales are recorded when the pri- vatization agency and a buyer reach a preliminary agreement, even though the deal later falls apart. Of twenty-one enterprises listed as pri- vatized in the Glhana divestiture agency's 1991 report, eleven returned to the state's portfolio in 1992. Another problem is that government data do not make nmuch distinction between various types of transactions. The sale of a 3 percent government holding in a company largely under private ownership is treated just like the full privarization of a company that was completely government-owned. Moreover, a sale might turn out to involve the transfer of assets from one state enterprise to another. Such transactiotis include Mali's sale of a publicly owned vegetable oil mill to another state-owned enterprise, and Ghana's purported privati- zation of a brewery when in fact 45 percent of the equity was transferred from the government to publicly owned banks and insurance compa- nies. Some liquidations are mere fiction, in that the firms had long be- fore ceased to operate. It is no surprise, then, that such limited privatization has had little impact on efficiency and economic growth. Berg (1993) cites several reasons. In some cases, a government has continued to hold a major share-and to intervene-in a partly privatized firm. In other cases, sev- erance benefits to dismissed employees were so large that governments had to siphon public resources away from productive uses. Many priva- tizations did not lead to greater efficiency because the new owners re- ceived favors-tax benefits, duty-free imports, tariff protection, and pri- ority access to credit and other scarce inputs-that reduced the social benefits. Where governments allowed buyers to defer payments, specu- lative buyers unqualified to run the enterprises sometimes bought them, 105 ADJUSTMENT IN AFRICA hoping to resell for a quick profit. Pressure on governments to sell may also have led to ill-considered transactions. And cronyism and cor- ruption have undercut the benefits of privatization. Transactions in Guinea and Nigeria, for example, have been criticized for their lack of transparency. High financial flows to public enterprises. Many countries have improved the financial profitability of some key enterprises, but placing this progress in a broader context is difficult. Data are scarce, and there has been no systematic analysis of changes in financial flows to public enterprises, whether in loans directly on-lent or guaranteed by the gov- ernment, in equity contributions from the government or other sources, or even in indirect grant assistance from donors. Nor are there estimates of revenue forgone from unpaid taxes, dividends, debt service payments, and tax exemptions. But a preliminary assessment based on scattered evidence suggests that financial flows to public enterprises remain substantial, and that the public enterprises remain a large drag on the government budget and financial system. In some countries, the relatively low direct budgetary subsidies may not be indicative of the real fiscal cost of public enterprises. For exam- ple, an examination of financial flows between the Burundi government and fourteen core public enterprises during 1987-90 reveals large im- plicit subsidies. The government accumulated payment arrears to the public enterprises, primarily because of underbudgeting, but the enter- prises themselves had considerable tax and debt service arrears to the government (table 4.3). The net outflow from the government to the public enterprises in 1989 was an estimated 19 percent of government expenditure. It is also instructive to examine the gross flows as costs to the government, under the assumption that the public enterprises, if privatized and operated like other commercial enterprises, would have received no government subsidies and would have paid taxes and divi- dends on their earnings. On this basis, public enterprises cost Burundi (in transfers and forgone revenue) an amount equal to 25 percent of total expenditure in 1989, or 1.5 times the education budget. Burundi is probably not atypical. Ghana's public enterprises are thought to command a similar share of total expenditure, though reli- able data are not available. In Tanzania, direct support (grants, subsidies, and transfers) was 7.3 percent of total expenditures in 1989. In addition to direct support, the central government engaged in short-term lending to public enterprises, which by 1991 had grown enormously. During io6 REFORMING THE PUBLIC SECTOR Table 4.3 Financial Flows between the Government and Public Enterprises in Burundi (millions of FBu, current prices) Item 1987 1989 1990 (estimated) A. Flows from government to public enterprises 3,441 8,947 9,613 Budgeted subsidies 2,342 5,974 7,545a Public enterprise debt service paid by government (net of reimbursement) 433 2,400 1,500b Customs exemptions 666 573 568 B. Flows from public enterprises to government 1,142 2,425 3,241 Profit taxes 594 940 1,860 Dividends 548 1,485 1,381 C. Net flows from government (line A - line B) 2,299 6,522 6,372 D. Net arrears owed to (owed by) government (735) 2,824 1,575 Owed to government by public enterprises 563 3,436c 2,01 Id Owed by government to public enterprises (1,298) (612) (436) E. Total net outflow from government to public enterprises (line C + line D) 1,564 9,346 7,947 As a share of total government expenditure 3.7 percent 19.2 percent 14.8 percent a. Includes taxes due and arrears. b. Includes FBu 1,292 million to the coffee sector to offset losses caused by price declines. c. Net debt service includes interest plus amortization. d. Includes FBu 1,011 million for debt service and an estimated FBu 1,000 million for tax arrears. Source: World Bank estimates. the first six months of 1991 alone, the treasury lent-at about half the prevailing commercial interest rate-an amount equal to more than 9 percent of the previous year's development expenditure and net lending. Furthermore, the government allocated foreign exchange to public en- terprises without requiring payment of the local currency counterpart- an implicit subsidy of about 3 percent of total government revenue in 1991. But the drain on public finances is not high for all countries: for Kenya, it was only about 1 percent of total expenditure in 1991, not counting implicit subsidies due to tax and other exemptions. No clear trend in quasi-fiscal transfers. In some countries, a decline in direct fiscal subsidies can mask a shift to quasi-fiscal sources of financ- ing, such as bank loans, and to other extrabudgetary sources of rev- enue. In Senegal, while direct operational subsidies from the budget declined between 1985-86 and 1988-89, there was a large rise in pub- lic expenditure overdrafts-a shift particularly difficult to evaluate. The International Monetary Fund's International Financial Statistics I07 ADJUSTMENT IN AFRICA distinguishes the claims on nonfinancial public enterprises from the total claims on the private sector for only eleven of the twenty-nine adjusting countries. For those eleven countries, the picture is mixed, with about half showing a decline in the stock of credit to the public enterprise sector. Moreover, data on other sources of quasi-fiscal fund- ing, such as social security funds, are particularly obscure, so there is no way of knowing whether the public enterprise sector has reduced its claim on the financial resources of the entire system. Ineffective reform of nonprivatized enterprises. Performance contracts- also called contract plans-have frequently been used to improve the performance of public enterprises not slated for privatization. Spelling out the rights and duties of the enterprise and the government, they were devised to attack vague objectives, insufficient autonomy, and weak incentives. Senegal was the first Sub-Saharan country to imple- ment performance contracts in the early 1980s. By 1988, fourteen countries either had signed or were drafting twenty-eight contracts, absorbing millions of dollars worth of technical assistance from donors. Although contracts can be useful in identifying problems facing an enterprise, they generally have not been effective in improving the per- formance of key enterprises (Nellis 1990; Sherif 1993). A few countries report some success. In Mauritania and The Gambia, the governments agreed to make appropriate budget allocations for water and electricity, helping to put ailing utilities on their feet. More frequently, however, contracts have not produced results. In Cameroon, where the govern- ment signed contracts with major utilities, neither the government nor the utilities have lived up to their obligations and the contracts have had little impact. In Ghana only four of the eleven firms with performance contracts reached the negotiated targets, because of the lack of financial discipline and performance accountability (Sherif 1993). Making a performance contract work may require conditions that seldom exist. Both parties must be committed. Some recapitalization is usually part of a restructuring program. And there must be enforceable targets, incentives for success, and censure or financial punishments for failure. Lacking these essentials, governments and enterprises have often disregarded carefully negotiated contract provisions. Governments have failed to make necessary management changes in the enterprises and to deliver promised financial resources. Enterprises have avoided difficult financial and personnel restructuring measures. So the considerable I08 REFORMING THE PUBLIC SECTOR time, effort, and resources spent developing performance contracts have achieved relatively little. Perhaps more disturbing, attention may have been diverted from more fundamental reforms, such as divestiture. It is difficult--in the absence of financial audits, information on quasi- fiscal transfers and tax exemptions, and productivity indexes-to judge whether the performance of public enterprises remaining under state ownership has inmproved. The general assessment, despite a few encourag- ing stories (box 4.1), is that the efforts at reform have been disappointing overall. Few "rehabilitated" public enterprises have made significant and lasting improvements, and many are being rehabilitated for a second and third time. Another worrisome trend is the ongoing, large annual invest- ment in these public enterprises (often financed in part by donor funds), even though they are generating low or negative returns. Most public en- terprise accounts are deficient, and where performance seems to be better, there is little evidence to suggest that sustainable improvements in man- agement are responsible. Instead, the better financial performance could be the result of inflation, major tariff increases to accommodate inefficient management, or the government's assumption of public enterprise debts. E,4. a. 'b7e whbrg Public Enterprises:. ,2bakbv''Ls Encouraginig Results IN ZIMBABWE, RESTRUCTURING IS AT WORKAND HAS PRODUCED particularly encouraging results in the national railways. After inde- pendence in the 1980s, the National Railways of Zimbabwe grew into a major loss-maker, operating in a noncompetitive environment and supported by ever-increasing subsidies. The National Railways' deficit grew from $8 million in 1982 to $230 million in 1990, at which time the government attempted reform. As part of the coun- try's adjustment plan, management of the railway was restructured, tariffs were increased by 50 percent over a year and a half to reflect economic costs, and the work force was cut about 3 percent. Other measures taken to improve efficiency included tightening staff super- vision, improving management information systems, filling critical job vacancies, and increasing security to decrease theft. With higher tariffs and relatively stable traffic, the company increased revenue by 75 percent irI the first year of adjustment and cut its deficit by half. The NationaL Railways of Zimbabwe is being called a model to be ap- plied to other public enterprises. lO9 ADJUSTMNT IN AFRICA To the extent that observed improvements in financial performance are due to these factors, there is little reason to think that they reflect a sus- tainable improvement in the enterprise's productivity. A few countries have experimented with nonasset divestiture for pub- lic utilities (leasing and concessions), with more success. Under leasing arrangements, a private contractor pays the public owner for exdusive rights to operate facilities and assumes full commercial risks. Conces- sions are similar to a lease, but the contractor has additional responsibil- ity for certain investments. C6te d'Ivoire and Guinea have had generally favorable experience with lease contracts and concessions in the water supply and power sectors (box 4.2). Financial Reforms: Limited Signs of Sustainable Progress E STABLISHING A SOUND, EFFICIENT FINANCIAL SYSTEM IS important to sustained economic growth, as a growing body of literature demonstrates.5 King and Levine (1992) show that financial development is correlated with growth, because of higher investment and greater efficiency. A country's financial sector should perform several critical functions: operating the payment mechanism, mobilizing savings, allocating financial resources, offering a means of diversifying risk, and providing services to facilitate trade. Financial systems in Sub-Saharan Africa have traditionally been char- acterized by weak resource mobilization, high credit losses, high inter- mediation costs, and excessive political interference. So Africa's financial depth is shallower than that in other developing regions. Bank deposits are just 15 percent of GDP. Furthermore, financial systems have not grown in real terms during the past decade. The low rate of financial savings in African countries is related to negative real interest rates, lack of confidence in the banking system, and political and macroeconomic instability, leading people to maintain savings in more tangible assets or to send their capital abroad. Part of the difficulty stems from structural, long-term problems: limited human resources and undiversified, small- scale economies. But government policy-reflected in ownership, inter- est rates, directed credit, and heavy taxation-has also weakened the fi- nancial systems. IIO REFORMING THE PUBLIC SECTOR Box 4.2 L@emonQel I' Public VUSIfte$ Water ini Guinea. IN 1989, OWNERSHIP OF Ivoirian and other investors (with a 49 percent Guinea's urban water supply authority and respon- share). CIE has a leasing arrangement to generate, sibility for sector planning and investment were transmit, and distribute power, and it receives con- transferred to a new autonomous water authority, siderable technical assistance from its foreign own- the Socite Nationale des Eaux de Guinee (SONEG). ers. The government, through EECI, continues to A new company, the Societe d'Exploitation des own the facilities and bears responsibility for invest- Eaux de Guin&e (SEEG), was created to operate and ments, sector policy, and planning. In the first eigh- maintain the facilities as a joint venture, with the teen months of the new arrangement, efficiency and government owning a 49 percent share and 51 per- service improved markedly. CIE raised the collection cent owned by a private, foreign consortium. The ratio from 60 to 90 percent, increased maintenance, strength of the Guinean arrangement lies in the reduced power outages, computerized business op- clarity of responsibilities and incentives. Under the erations, and eliminated operating subsidies. An ten-year lease contract with SONEG, SEEG operates important reason for CIE's success was its status as a and maintains the system at its own risk, with its re- newly created operating entity rather than a muneration based on user charges actually collected holdover from the past, even though it retained and on fees for new connections. SEEG benefits if it some of the staff of EECI. improves collections and reduces operating costs Water in Cote d'lIvoire. A concession for water and unaccounted-for water. SONEG has incentives supply services in Cote d'Ivoire was arranged in to seek adequate tariffs and to invest prudently, 1987, following twenty-five years of a lease con- based on realistic demand forecasts, because it has tract. The lease had improved the service and inter- responsibiliry for capital financing. So far the col- nal efficiency of the operating company, the Societe lection ratio has increased dramatically, from 20 des Eaux de C6te d'lvoire (SODECI), a mixed French percent to 70 percent, and technical efficiency and and Ivoirian enterprise. However, financial troubles coverage have improved. Tariffs are up sevenfold mounted in the 1980s because of policies enforced and are expected to reach the full cost-recovery level by the government regarding investment and tar- by 1998. In the interim, donor assistance is financ- iffs, for which it retained responsibility. Under the ing the foreign exchange costs of the operation, and new concession arrangement, SODECI became both the government has assumed debt service. The operator and investor, with responsibility for all arrangement has been hampered by delays in equip- new urban water supply investments in the country. ment procurement by SONEG, which has affected The company receives no operating subsidies, and SEEG's ability to improve the quality of service and all new investments are self-financed. Although the its financial performance. concession contract was not subject to competition, Power in C6te d'lvoire. The performance of the SODECI's operating fees were reduced substantially Enterprise d'Electricit6 de C6te d'Ivoire (EEci), the during the negotiations. The company's operating parastatal responsible for the electric power sector costs are now comparable to those of many water in C6te d'Ivoire, deteriorated during the 1980s. In utilities in West Africa, while the quality of its ser- 1990 a major restructuring transferred responsibil- vice is far better than most. Private Ivoirians own a ity for plant operation and maintenance to the majority of SODECI's shares, and the company has Compagnie Ivoirienne d'Electricite (CIE), a new succeeded in reducing expatriate staffwhile expand- joint venture owned by a consortium of two French ing operations. companies (with a 51 percent share), and by Source: Kessides (1993). III ADJUSTMENT IN AFRICA With so many problems, Sub-Saharan Africa's financial systems were ill-prepared to cope with the economic crisis that came to a head in the early 1980s. The overall worsening of the economy further weakened the financial soundness of borrowers, especially the central governments and the parastatals. Agricultural parastatals and cooperatives often ac- counted for a large share of unpaid loans. Despite the general absence of competition, bank profits were generally low, and margins were inade- quate to offset the growing deterioration of the portfolios. Problem loans mounted, and many private and public institutions became tech- nically insolvent. Loan-loss ratios regionwide approached 40 to 60 per- cent, with banks in some countries showing bad loans for more than 90 percent of their portfolios. Overstaffing and branch expansion under government pressures further reduced bank profitability. Large and neg- ative real interest rates and significant uncertainty about future policies led to capital flight, reduced financial intermediation, and problems with bank solvency and liquidity. Many countries in Africa are undertaking financial reform programs focused largely on banking (table 4.4). The first objective is to reduce fi- h ab!s 4.,^S FEr-d"ML sacv.E, Ra"e-sE ES;9''Ihs E9EEg> ."f9arERXE par$tc Liberalization and/or rationalization of interest Restructuring Privatization Liquidation rates of banks of banks of banks Benin Cameroon Cameroon Benin Burundi C6te d'Ivoire C6te d'Ivoire C6te d'Ivoire Congo Ghana Guinea-Bissau Guinea C6te d'Ivoire Guinea Madagascar Niger The Gambia Kenya Mauritania Rwanda Ghana Madagascar Senegal Senegal Kenya Mali Madagascar Mauritania Malawi Rwanda Mauritania Senegal Mozambique Tanzania Rwanda Uganda Tanzania Note: This table is not intended to be a comprehensive list of all the financial sector reforms undertaken. Source: World Bank staff. IIZ REFORMING THE PUBLIC SECTOR nancial repression-by aligning interest rates toward market equilib- rium levels and reducing directed credit programs. The second is to re- store solvency and improve the incentives under which banks operate. This involves restructuring and recapitalizing distressed banks, liquidat- ing some development credit institutions and some other specialized entities, and privatizing banks. The third objective is to improve the financial infrastructure, strengthening supervision, auditing, and ac- counting practices and providing more training and development op- portunities for staff. Improving the legal framework, though initiated with other adjustment reforms, is part of the longer-term development agenda for financial sector reform. The reforms have had only limited success. Financial systems con- tinue to finance the deficits of the central government and the overex- tended public enterprise sector. Continuing heavy regulation in the real sector discourages the emergence of profit-making private sector bor- rowers. A weak legal framework and an inability or unwillingness to en- force financial discipline further compromise efforts to improve bank portfolios. Although bad loans have been stripped from the balance sheets of many banks, apparently improving bank performance, there are new signs of balance-sheet problems. More encouraging, however, are the progress in reducing financial repression and the move to priva- tize more financial institutions. African financial systems continue to be an extension of the fiscal sys- tem, in part because of heavy government ownership. Additionally, both central banks and commercial banks are often major lenders to the gov- ernment. The public sector has a large share of total domestic credit in many African countries, crowding out the private sector. The monetiza- tion of budget deficits also undermines or prevents the development of interbank and imoney markets, and high taxation reduces the financial sector's profitability.6 In addition, there is significant public interference in the manage- ment of both public and private banks, eroding the quality of their port- folios. Credit allocation procedures are poor, as the public banks gener- ally allocate credit following directives from the government (Tenconi 1992). Loans to parastatals are seldom evaluated, because they have the implicit or explicit backing of the government, and loans to the private sector are not systematically scrutinized. Government inspection and supervision of banks is grossly inadequate. Indeed, supervision is essen- tially nonexistent in countries where the banking system has been na- 113 AJUSTMENT 0IN AFRICA tionalized. Even in countries with nominally private banks, there may be a conflict of interest because members of the supervisory agencies often sit on the boards of the banking institutions. In most instances, prudential ratios have been ineffective in preventing the degradation of loan portfolios. The ratios sometimes were poorly designed, and banks sometimes lacked rules for classifying credit according to risk and for defining and handling delinquent loans. Not surprisingly, adjustment under such conditions has generally been slow and difficult. Some Progress in Reducing Financial Repression The rationalization and liberalization of interest rates, among the most common features of adjustment programs, have been somewhat successful in easing financial repression. Interest rates were fully liberal- ized in Burundi, The Gambia, Ghana, Kenya, Madagascar, Malawi, Mauritania, and Zambia (table 4.5). Rwanda followed a slightly differ- ent approach, with a program that set a floor for deposit rates and a ceil- ing for lending rates. The central banks in the CFA franc zone raised in- terest rates and eliminated preferential rates to maintain competitiveness with France. These reforms have not always succeeded in reducing financial re- pression in countries with highly negative real interest rates. Although several countries had moved from significantly negative to acceptable real interest rates by the end of the adjustment period, Rwanda, Sierra Leone, and Zimbabwe had average rates ranging from -10 to -31 per- cent during 1990-91 (appendix table A.4). At the other end of the spec- trum, in another problematic group, were eleven countries with highly positive real interest rates. Ten of these-all with convertible currencies and overvalued real exchange rates-had average real deposit rates above 5 percent during 1990-91. The limited success in achieving reasonable real interest rates, despite considerable attention to interest rate reforms, is not surprising given limited competition and continuing intervention in Africa's financial systems. Because most countries have few banks, and because most of those banks have significant public ownership, there is little scope for "true" market-determination of interest rates. Burundi, Kenya, Rwanda, and Zaire have moved toward the use of treasury bill auctions to establish a benchmark rate for a market- determined interest rate structure. But too much reliance on this mech- anism as a "true" market rate could be deceptive, particularly if the gov- "14 REFORMING THE PUBLIC SECTOR Table 4.5 Government lnterven1tlomi in tFe Financial Iv,arks8 Interest rate for deposits Interest rate for loans Before Late Before Late Country reforms 1992 reforms 1992 Benin * Burkina Faso 0 0 0 Burundi * 0 0 0 Cameroon 0 @ * 0 Central African Rep. * 0O * 0 Chad OX 0 0 Congo * 0 S 0 C6te d'Ivoire 0 6 o Gabon 0 0 * The Gambia *0 0 0 Ghana * 0 * 0 Guinea 0 - - - Guinea-Bissau 0 * * 0 Kenya *0 0 0 Madagascar 0 C S 0 Malawi * 0 0 0 Mali 0 0 Mauritania * 0 * O Mozambique 0 0 0 0 Niger 0 0 0 0 Nigeria 0 0 * 0 Rwanda 0 0 * (0 Senegal 0 0 0 0 Sierra Leone *0 0 Tanzania 0 0 0 0 Togo 0 03 * 0 Uganda * 0 * S Zambia * 0 0 0 Zimbabwe * 0 0 0 * Rate set. 3 Spread regulated. 0 Minimum deposit rate maximum lending rate set. O No government control. - No data av7ailable. Source: World Bank staff. ernment domirLates the market. Transparent, fair auctions and trading in government securities may be more useful for fostering a trading cul- ture in these countries, but they have not proved to be a reliable proxy for market-determined interest rates in the short and medium term. "15 ADJUSTMENT IN AFRICA Reducing Directed Credit Countries have apparently had some success reducing directed credit. Many countries are scaling back their efforts to target credit to specific sectors, especially to unsound public enterprises or other quasi-public agencies. Several development finance institutions heavily involved in providing directed credit have been closed. For example, the 1989 re- forms of the crop marketing system in the West African Monetary Union significantly reduced preferential credit programs. Such reforms can be difficult to sustain, however. In Tanzania about two-thirds of all bank lending at the end of 1987 was to cover the oper- ating deficits of the crop marketing parastatals. During 1988 the gov- ernment officially took over 40 percent of these liabilities, but the crop marketing parastatals were more indebted to the banking system at the end of the year than at the beginning of 1987. While many countries have nominally eliminated directed credit policies or institutions, infor- mal pressure to lend to the politically well connected may remain. No systematic analysis has been undertaken to determine if the elimination of formal directed credit programs has resulted in sounder portfolios. Restructuring Balance Sheets, Recapitalizing, and Privatizing Restructuring and recapitalizing insolvent or undercapitalized banks has been another important part of adjustment programs. The objective is to restore the viability of these banks and improve the quality of fi- nancial intermediation. In a typical restructuring operation, a newly cre- ated government agency absorbs the nonperforming loans of the banks, taking them off the banks balance sheet, and the government and the shareholders infuse banks with new capital to enable them to restart normal operations under a revised incentive structure. For example, in Ghana in 1990, the government recapitalized ten state-owned commer- cial and development banks through a central bank bond issue and transferred the related bad loans and assets to a recovery trust for non- performing assets. The internal restructuring of banks-including downsizing, changing incentives for managers (and changing the man- agers themselves), and insulating management from political interfer- ence-is essential for avoiding past mistakes but difficult to achieve. Bank recapitalizations in Africa-without corresponding changes in the real sector, such as restructuring the parastatals-have generally ii6 REFORMING THIE PUBLIC SECTOR failed. In the Central African Republic, one bank has been restructured three times. In Mauritania, five major state-owned banks were recapital- ized at a cost of nearly 15 percent of GDP in 1988, but they are again suf- fering large losses of up to 50 or 60 percent of total loans. In Kenya in 1989, eight failed institutions were merged into a "turnaround" bank, Consolidated Bank Limited, now in difficulty. The recapitalization operations have also been expensive. The fiscal cost in some cases was between 1 and 2 percent of GDP-1.5 percent of GDP in Ghana, 2.0 percent in Guinea, and 1.5 percent in Madagascar; in other cases, it was much larger. In Senegal the cost came to about 15 percent of GDI', mainly because the restructuring program covered nearly all losses for the distressed banks, rather than stripping out a por- tion of the bad debts. In Tanzania the cost of a partial restructuring is es- timated at roughly 40 percent of GDP. Moreover, recovery of bad debts has generally been disappointing. Tenconi (1992) reports that the amounts Guinea recovered in the first five years after closing all govern- ment-owned banks amounted to about 3 percent of the portfolio. In Cameroon the total recovery rate for banks being liquidated was less than 5 percent by 1990, and liquidation costs, including severance pay, absorbed more than half the recovered amounts. In a few cases, the terms and conditions of the first restructuring operations involving for- eign shareholders were overly generous to these shareholders and set a costly precedent for restructuring other banks. Guinea-Bissau, Madagascar, and Senegal privatized all or at least a few of the restructured banks. Ghana and Tanzania kept the restructured banks in the public sector. In the very few cases in which restructuring was accompanied by privatization or by the creation of private banks- or both-there is no evidence that banks are generally performing well. Guinea-Bissau found that the banks rapidly accumulated bad debts without an ap,propriate interest rate structure. Privatization without at- tention to the LLnderlying policy framework and to restructuring the real sector is thus unlikely to improve the performance and efficiency of the banking sector. In many countries, the public sector dominates the banking system. For the region's adjusting countries, the government has equity in ap- proximately half of the banks and is the majority owner of more than a third. In 1992 the government was the majority holder of all banks in Congo and Tanzania. In four other countries-Burkina Faso, Burundi, Chad, and Rwanda-the public sector holds at least a minority stake in 117 AJT I AFRICA all banks (Tenconi 1992). Most countries have fewer than ten banks (most of them government-owned). And only C6te d'Ivoire, Kenya, Nigeria, and Zambia have six or more private banks, facilitating some competition and independence from public interference. In the few countries with substantial private capital participation, the domestic shareholders are usually closely linked to the political authorities. Despite these difficulties, more banks are being privatized, and new private financial institutions are entering the markets-a welcome trend that reverses the nationalization of financial institutions in the 1960s and 1970s. In the twenty-nine countries examined in this report, the number of banks in which the government holds a controlling interest was reduced from 106 to 76 as a result of liquidation and privatization between 1982 and 1992 (table 4.6). The government-owned banking systems of Benin and Guinea were dosed and replaced by privately owned banks (some with government participation in Guinea).7 While the total number of commercial banks increased by 15 percent between 1982 and 1992, the number with no government participation almost doubled. Nigeria had twenty-four fully private banks in 1992, com- pared with just nine in 1982. These are clear signs of a more open atti- tude toward private ownership, but the public sector still dominates the financial system, and further progress is needed in reducing its role. Improving Financial Infrastructure and Regulatory Capabilities Financial markets in Africa are particularly lacking in skilled profes- sionals-a result of relatively low education levels and relatively short periods of independence in these countries. Although things are now changing, foreign banks continued to dominate the financial sector in many countries during the first few years after independence, narrowing the possibilities for African managers to gain significant experience and handicapping bank officers even more. The same was true for staff in central banks and supervisory agencies. African governments have recently placed more emphasis on devel- oping regulatory capabilities and skills. Many of the early financial ad- justment programs in West Africa focused on restructuring balance sheets, with regulatory issues often relegated to studies. The 1988-89 programs in Ghana and Kenya were the first to emphasize the develop- ment of the legal and regulatory framework. More attention has also been given to improving the financial infrastructure, typically through a II8 REFORMING THE PUBLIC SECTOR TaSNe 4. Gwernmn-sn Fari cpz9n 3 inlte I1t'a3 oIfIB OVra tE C 1bP (a iwibe;/ OJfnk:) 1982 (213 banks) 1992 (245 banks) Majority Minority No Majority Minority No Country shareholder shareholder share shareholder shareholder share Benin 3 0 0 0 0 5 Burkina Faso 4 0 1 4 1 0 Burundi 3 0 3 3 4 0 Cameroon 5 6 0 1 7 2 Central African Republic 1 2 0 1 0 2 Chad 1 2 1 1 2 0 Congo 4 0 0 3 0 0 C6te d'Ivoire 2 5 5 2 4 9 Gabon 1 4 1 2 4 5 The Gambia 1 1 0 1 0 2 Ghana 7 2 2 5 3 4 Guinea 6 0 1 1 2 3 Guinea-Bissau - - - - - - Kenya 3 1 9 4 2 20 Madagascar 3 0 0 2 1 1 Malawi 3 1 1 1 3 1 Mali 3 1 2 2 2 2 Mauritania 4 1 0 1 0 3 Mozambique 1 1 0 0 0 3 Niger 4 0 5 2 1 4 Nigeria 24 4 9 22 6 24 Rwanda 1 1 0 1 2 0 Senegal 2 5 3 2 3 3 Sierra Leone 3 1 3 2 0 4 Tanzania 6 0 0 5 0 0 Togo 3 2 2 3 2 2 Uganda 2 5 2 2 4 3 Zambia 4 1 4 2 0 10 Zimbabwe 2 1 6 1 1 3 Total 106 47 60 76 54 115 - Not available. Sources: Tenconi (1992); World Bank staff. variety of technical assistance operations. The tasks range from simple training activities (teaching domestic bank officials how to shorten the time that a check takes to clear) to more complex operations (designing an adequate legal and regulatory framework). Ghana, Kenya, Madagas- car, and Mauritania have made efforts to strengthen supervision and fi- nancial expertise. In general, though, not enough has been done to de- velop accounting and audit standards or to form a core of accounting II9 ADJUSTMENT IN AFRICA professionals-tasks that go beyond the short time-frame of adjustment programs. Changing regulatory standards is one thing; getting the information to apply them, given the scarcity of accountants, is another. Eighteen countries in the region have fewer than fifty fully qualified accountants (UNCTC 1991). In some countries, such as Rwanda, the banks must go outside the country to hire an external auditor, because they have no qualified chartered accountants. Requiring a bank (much less an enter- prise) to be audited becomes time-consuming and expensive. Audit standards for banks and enterprises are often uneven and haphazardly enforced, and even if institutions conform to the letter of the regula- tions, their financial data may be of such questionable quality as to make the audit report meaningless. Progress in creating new financial instruments, such as interbank and money markets, has been mixed, and it is clear that efforts to develop money and capital markets cannot succeed where competition is lacking or where governments still intervene extensively to set interest rates. The most advanced money markets, in Ghana and Zimbabwe, consist of an interbank market and discount houses. The discount houses trade in short-term instruments that are usually rediscountable by the central bank, such as central bank or treasury bills, crop bills, export and import bills, and bankers' acceptances. Short-term markets deepen and diversify financial markets, and they can develop if the fiscal and budgetary situ- ations remain stable. Improving Public Sector Management: A Long-Term Challenge A CHIEVING BETTER MANAGEMENT IN THE PUBLIC SECTOR IS essentially a long-term development objective, and evaluating Africa's success in this regard is beyond the scope of this report. Upgrading the quality of the civil service, developing career streams, putting in place promotion and pay systems based on merit, strengthening budgetary and investment management, and building policy analysis and regulatory capacity-all these take considerable time and require considerable improvements in the institutional envi- ronment. They also require good governance. Adjustment programs 120 REFORMING THE PUBLIC SECTOR have laid some of the foundations, but the efforts will have to continue long after the basic reforms are in place. The main short-term objec- tives of adjustment reforms have been to trim the civil service to a manageable size, increase real public sector wages to reasonable levels where necessary, get control of the payroll system, eliminate the worst "white elephants" from investment programs, create a rolling invest- ment plan that includes all the investment projects being undertaken, and reform tax policy and administration. Progress even in these areas is difficult to quantify, partly because well-defined norms for measur- ing progress do not exist. Civil Service Reform Since independence, most African countries' policies toward the civil service (and public sector employment) have had three common features, each undermining institutional capacity. First, they expanded the size of the public sector faster than the economy grew. Second, they favored em- ployment growth over income growth in the public sector, driving down the real wages of public sector employees. Third, they favored pay in- creases in the lower ranks, reducing pay differences between skilled and unskilled employees. The result: civil services are larger than countries need, more costLy than they can afford, and less effective and productive than they should be. Reform programs have responded to these problems with short-term cost-containment measures and medium-term programs to build institutional capacity to increase productivity. A narrow, short-term approach to civil service reform is likely to yield limited benefits where governance problems are serious (Dia 1993). In such settings, recruitment is based on subjective criteria, public employ- ment is part of the social welfare system, pay levels are unrelated to pro- ductivity, and loyalty is to the political leader in power rather than to the state. Building a clear public consensus on the need for good governance may be a precondition for creating a more effective civil service. En- couraging professional and other interest groups, allowing freedom of association and expression, and reforming the judicial and legal systems will enable a more productive and responsive public sector and sharply curtail patronage and corruption. The short-term record of civil service reform is mixed. Most adjust- ing countries tried to trim excessive public sector wage bills, shed "sur- plus" civil service staff, reverse salary erosion, and decompress the wage 121 ADJU;S TMENT IN AFRICA structure. There has been some progress in reducing public sector wages and salaries as shares of GDP and current expenditure (net of interest). Between 1985-86 and 1990-9 1, fifteen countries cut their wage bill as a share of GDP, while eleven increased it (figure 4. 1). That eleven of fif- teen countries with flexible exchange rates managed to cut their wage bill, while increases were recorded for seven of eleven fixed exchange rate countries, suggests that real depreciations and the relatively greater wage flexibility in the flexible exchange rate countries were important in re- ducing the wage bill. Many countries also took steps to reduce the number of surplus civil servants. They used a variety of measures: ensuring attrition through Progress in controlling the wage bill hiring freezes, enforcing mandatory retirement ages, abolishing job has been mixed. guarantees for high school and university graduates, introducing volun- ;.9P... - pg --.- ,, ,Thvaga 131H Wages and Salaries Wages and Salaries as a Share of Current as a Share of GDP Expenditure Excluding Interest Payments Ugaanda -Uganda Nigeria Nigeria Sierra Leone -198S86 Zambia 198586 Malawi E 1990-91 Malawi t 1990-91 Ghana Tanzania Madagascar Mauritania Zambia The Gambia Tanzania Rwanda __ Mail Countries Mali Countries Kenya _ with Kenya with Mauritania decreases Ghana s decreases Central African Rep. = Senegal Senegal Madagascar = Benin Central African Rep. Mozambique Mozambique Niger A2 . Sierra Leone Rwanda se.s. Gabon Burundi Zimbabwe Za Gabon Burundi _ Togo No , CutisNiger - - - Countries Cameoonwith Congo wit CBerkina Faso ,increases C6te d'lvoire increases Burkina Faso Togo , C6te d'ivoire Cameroon Congo Benin ______ Zimbabwe Burkina Faso ___ 0 2 4 6 8 10 12 14 16 18 0 10 20 30 40 50 60 70 80 90 Percent Percent NVote: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. Sources: IMF staff estimates; Nashashibi and Bazzoni (1994). 122 REFORMING THE PUBLIC SECTOR tary departure schemes, and making outright dismissals. Cameroon, Ghana, Guinea, Tanzania, and Uganda significantly reduced employ- ment by exorcising payroll "ghosts"-paycheck recipients who did not actually work. The Central African Republic, Congo, Gabon, Ghana, The Gambia, K:enya, Mauritania, Mali, Nigeria, and Senegal used vari- ous forms of hiring freezes. Guinea and other countries achieved reduc- tions by strictly enforcing the retirement-age provisions. And several countries adopte d early retirement and voluntary departure schemes- the Central Afirican Republic, Ghana, Guinea, Guinea-Bissau, Mali, and Senegal. Dismissals were less widely used, primarily in the Central African Republic, Guinea, Ghana, and Senegal. Despite these efforts, the gross reductions shown in table 4.7 considerably overstate net re- ductions, because of new hiring. Only a handful of countries have cut the number of civil service em- ployees by more than 5 percent since they began structural adjustment Table 4.7 Reduc:tions of Civil Service Personnel in Selected Countries, 1981-90 Removal Enforced! Retrenchment of 'ghost" early Voluntary Regular Temporary Other Country employees retirement departure staff staff mechanisms Totala Cameroon 5,840b 5,000 _ 10,830 Central African Republic 2,950- _ 1,200 350-400 - - 4,500-4,550 Congo - - - - - 2,848c 2,848 The Gambia - - 919 2,871 3,790 Ghana llOOOd 4,235e _ 44,375e - _ 48,610 Guinea 1,091 10,236 1,744 - - 25,793g 38,864 Guinea-Bissau 8006 945 1,960 921 - 3,826 Mali - - 600 - - - 600 Senegal 497 747h 1.283 - 2,527 Uganda 20,000i - - - - 20,000 a. Gross figures, not adjusted for new recruitment and attrition. b. Includes elirmination of double payments as well as "ghosts" (fictitious employees). c. Attrition and hiring freeze. d. Includes "ghosts" identified but nor necessarily removed. e. Includes staff in district assemblies and in the education sector. f. Includes "ghosts" in the Conakry area only. A second census in 1989-90 identified many more "ghosts." g. Parastatal liquidations and the transfer of joint-venture mining workers to company rolls accounted for 14,983 of these. The remaining 10,810 officials were assigned to a personnel bank and placed on administrative leave; it is unclear whether all have left the civil service. h. An additional 2,133 officials applied for voluntary departure or early retirement. i. Estimate based on savings from 'ghost" removal divided by average civil service wages. Source: World Bank (199 ib). I23 ADJUSTMENT IN AFRICA (table 4.8).' Moreover, personnel cuts have usually had little fiscal im- pact, because the cuts disproportionately affected low-paid employees at the bottom of the civil service. Salary erosion has continued and has fre- quently been more important in reducing the wage bill than personnel reductions. Although a few countries, notably Guinea, reversed real wage erosion before the adjustment period, this has been the exception rather than the rule. Scattered evidence suggests that the erosion of real wages has been more prevalent in countries with higher inflation. Few adjusting countries have effective control over the payroll sys- tem; in fact, all but six have significant problems (table 4.9). To manage the payroll system, governments need to maintain information on who is employed in the public sector and verify the information against the payroll. Personnel censuses can become quickly outdated unless systems are in place to maintain them and link them to the payroll. Again, gov- ernance problems undermine reforms in these areas. The democratic transition in Africa has also led to reversals of wage containments and personnel reductions in some countries. In Congo the wage bill ballooned by about 60 percent in 1991 because of salary increases and a 15 percent increase in the number of public employees. In Ghana in 1992-a year of political transition-wage increases of about 80 percent compromised the fiscal stability maintained since 1986, casting doubt on the sustainability of the present level of public spending. Restoring fiscal balance has proven difficult. Increase of 5 percent No significant Decrease of Information or more change 5 percent or more not available Burundi Burkina Faso Central African Rep. Benin Cameroon Chad The Gambia Gabon Congo Guinea-Bissau Ghana Madagascar C6te d'Ivoire Togo Guinea Mozambique Kenya Zimbabwe Mali Nigeria Malawi Mauritania Rwanda Niger Senegal Tanzania Sierra Leone Uganda Zambia Source: World Bank staff. 124 REFORMIING THE PUBLIC SECTOR Public sector salary spreads have widened substantially in several K:e F uL[X': countries-a desirable development in that it may facilitate recruitment 1-he ,-c r. of talented individuals and provide public employees greater incentives LT E LkiK to improve performance. But the data are too limited to establish Country Effectiveness whether this is a regionwide trend. In Ghana the ratio of the highest- Benin C paid echelon to the lowest-paid widened from about 5:2 in 1984 to 10:1 Burkina Faso 0 in 1991, and in Mozambique from 2:1 in 1985 to 9:1. Salary spreads Burundi 0 appear also to have widened in Guinea and The Gambia, though they Cmeroon 0 remained unchaLnged in the Central African Republic. Central African Rep. 0 Several countries have also begun devising new administrative struc- Congo tures, redeploying personnel, redefining career streams, adopting formal C6te d'lvoire o hiring and promotion procedures, and instituting incentive-based com- Gabon C pensation systerns. These exercises, while useful, are very complex and Ghana c may yield few long-term benefits unless underlying governance prob- Guinea 0 lems are also addressed. Guinea-Bissau 0 Kenya C Madagascar c Public Expenditure and Investment Malawi C Mali 0 Adjustment programs (and related technical assistance) have aimed at Mauritania 0 Mozamnbique o improving the budgeting processes in the public sector, the monitoring Niger 0 and reconciling of expenditures, and the allocation of public funds. But Nigeria c such efforts are complex, and it takes time to develop the necessary local Rwanda 0 capacity. A recent assessment of efforts among countries benefiting from Senegal 0 Sierra Leone 0 the Special Program of Assistance for Africa (SPA) noted improvement in Tanzania 0 the government's capacity to plan and monitor public expenditure.9 The Togo c strength of that capacity remains a concern, though, as does the sustain- Uganda C ability and composition of government spending. Public spending is dif- Zambia c ficult to track because of numerous off-budget accounts (many estab- lished for donor--funded projects), the opacity of military budgets, and C Some problems. the financial operations of public enterprises and the banking systems. 0 No significant problems. There is still a long way to go in improving the monitoring of implicit Source: World Bank staff. and explicit financial flows between public enterprises and government. With respect to public investment, adjustment programs have tried to ensure that resources are used wisely and that low-priority projects are eliminated. One of the first steps necessary is to develop a rolling invest- ment plan that takes account of all projects. As of late 1992, eighteen of the twenty-nine adjusting countries had not put in place an effective system for tracking all projects in the investment program, and Cameroon, the Central African Republic, Chad, Guinea-Bissau, Nigeria, and Tanzania 125 AIHJUSTMENT IN AFRICA 7m[a ; were assessed as having substantial problems (table 4.10). In some cases, ProgFAn7 :zt @2 the sheer number of projects financed by external donors makes effective Recording control difficult. The public investment program in Burundi varies from Country of all projects 500 to 600 projects. In Tanzania the development budget lists about Benin c 2,000 projects, and even this list excludes many that are donor-funded. Burkina Faso c Maintaining a rolling investment program that includes all invest- Cameroon ment projects, even those financed by donors, is only the first step in en- Central African Rep. * suring adequate quality of investment. Nigeria improved the monitor- Chad 0 ing of investments, but it is planning large investments with dubious Congo C rates of return. Because there are no established benchmarks that would C6te d'Ivoire c Gabon Cenable us to compare the quality of public investment programs across The Gambia O countries, we were not able to systematically rank countries on the basis Ghana C of their efforts to reform their public investment programs. Guinea s Creating a core program of the most worthwhile investment projects Guinea-Bissau 0 Kenya O has been useful in improving the management of aid. Core programs Madagascar O push governments to set priorities for spending and to devise more ef- Malawi O fective mechanisms for channeling domestic counterpart resources to Mauritania important foreign aid-financed projects. Such a program helped Ghana Mozambique meet its most critical investment objectives (box 4.3). But in many Niger C countries, noncore projects have remained in the public investment pro- Nigeria gram and competed (often successfully) for funds. So, designating a core Rwanda C Senegal c program may not be as effective as properly sizing and ranking public Sierra Leone O investments in the first place. Tanzania * Many countries have made progress in consolidating expenditures Togo ° and revenues in the government budget, including better accounting of Uganda c Zambia c foreign aid and increased measurement of tax exemptions and quasi- Zimbabwe O fiscal expenditures. The composition of public expenditure is generally * Substantial problems. poor, however (see box 4.4 for a discussion of this as it relates to Kenya). C Some problems. Common problems include underspending in the sectors most vital for o No significant problems. development, funding of investment projects without allocation of suf- Source: World Bank staff. ficient resources to meet future recurrent charges, poor maintenance of existing capital stock, overspending on wages, and high levels of military spending. Only two of the twenty-one countries covered by a 1991 sur- vey for the SPA had a reasonable degree of efficiency in spending. Tax Reform Tax reform in Sub-Saharan Africa appears to have been driven by short-term revenue needs as much as other considerations, such as se- 126 REFORMING THE PUBLIC SECTOR t-Rzx -4.2 2_,~Pu,,bdcf. "Ff' igzance Re*):mfW in, Ghhi n AT THE START OF ITS ADJUSTMiENT PROGRAM IN increased dramatically-from 34.5 percent in 1984 1983, Ghana faced a number of critical public expen- to about 50 percent in 1989. Real per capita educa- diture problems. Economic decline and a shrinking tion spending increased by 150 percent between 1984 revenue base had severely compressed government and 1989, while real per capita health expenditure spending. Civil service wages and salaries had eroded, tripled. The share of primary education also grew, and wage differentials were narrow, and employment rolls there was greater cost recovery and a reduction of sub- were paddecL. Nonwage operations and maintenance sidies at higher levels. As a result, basic education en- were neglected, while public investment declined to rollments are up and a range of health indicators have less than 1 percent of GDP. The result was a near-col- improved. lapse of the country's economic infrastructure and se- The government also shed about 50,000 surplus vere cutbacks in social services. civil service staff, increased average pay levels, and In the course of adjustment, Ghana made substan- widened the salary differential between the highest- tial progress in addressing several of these problems. and lowest-paid workers. Another priority was re- Increased revenue and foreign financing (including shaping public investment. A task force prepared project and nonproject assistance) permitted total Ghana's first rolling public investment program in central government expenditure to rise from I I per- 1986 and established objective screening criteria for cent of GDP in 1984 to 19 percent in 1986, where it projects, scaling down to a core program of about 100 stayed for the rest of the decade. Capital expenditure projects consistent with the country's implementation for infrastructure rose to about 8 percent of GDP by capacity. That program emphasized rehabilitation the end of the decade. The composition of expendi- rather than new projects. In 1987 an inner "super- ture was rest.ructured to increase the share of nonwage core" program was identified, with twenty-nine proj- operations and maintenance expenditures in health, ects of special importance to be protected in budget education, and agriculture. The share of social services implementation. (including education and health) in total expenditure Source: World Bank (1992a). curing a given level of revenue while minimizing efficiency losses, mov- ing toward a more equitable distribution of income, and making the tax system simple and transparent both to ensure better voluntary compli- ance by taxpayers and to lower administrative costs. Despite the short- term focus on raising revenues, the ratios of tax revenue to GDP have been low in Sub-Saharan Africa, ranging from 6 to 20 percent at the start of the adjustment period in 1986 and ending with a median im- provement of less than 1 percent by 1992, with considerable variation among countries (appendix table A. 14). There were wide-ranging dis- cretionary rate aLnd base changes, while the expenditure-to-GDP ratio re- mained stubbornly higher, at 20-25 percent. These discretionary changes have t-aken several forms: eliminating exemptions for all nondiplomatic imports, imposing a minimum import levy, converting 127 ADJUSTMENT IN AFRICA Bx4.4 Chan:lges;2 h'i> g g Speojlh,n,9 ,, KENYA MANAGED TO REDUCE ITS FISCAL DEFICIT dent are more than seventy times those per primary in the 1980s and contain the total level of public student. At the university level, 30 percent of spend- spending. Like many other African countries, how- ing goes for boarding and allowances, and the gov- ever, it was less successful in improving the composi- ernment has taken some steps toward cost recovery. tion of that spending. In primary education, teacher salaries account for 85 Rapid growth in curtent expenditure, primarily for percent of spending, and milk distribution for an- wages, was fueled by growth in central government other 13 percent. The number of teachers has been employment (5.4 percent a year during 1981-87). growing much faster than the number of students, Employment of teachers by the Teachers' Service leading to falling student-teacher ratios and rising Commission grew particularly rapidly, at 8 percent a costs. Nonwage expenditure has dropped sharply, to year. Meanwhile, expenditures on nonwage opera- less than half the price of one textbook per student tions and maintenance declined both in real terms per year. More recently, some efforts have been made and as a percentage of total expenditure for much of to improve the composition of public spending in the 1980s, and recent increases have been insufficient the education sector. to reverse the decline. Cuts in public investment were The health sector also faces growing imbalances instrumental in reducing expenditures and the deficit between personnel costs and other operating expendi- until the early 1990s, when both shot back up. But tures. Personnel expenditures increased an average of the government has not established priorities for its 6.4 percent a year in real terms over fiscal 1985-88, public investment program. while nonwage operating expenditures fell by 4.4 per- Intrasectoral imbalances remain in both educa- cent, resulting in shortages of drugs and supplies. tion and health. Budgetary costs per university stu- Source: World Bank (1992a). import and excise duties from specific rates to higher ad valorem rates, raising tax rates for petroleum products, and moving to a current-year basis for assessing corporate income. The most prominent trend has been for taxes of goods and services to replace taxes on international trade as the main source of tax revenue, while the share of taxes on in- come and profits has stagnated. Efficiency considerations have been taken into account indirectly in several ways. High marginal rates of taxation on personal and corporate income have been significantly reduced. This, coupled with inflation- proofing of personal exemptions and tax brackets, has strengthened the incentive to work. In addition, taxing consumption rather than income has been stressed as a way to broaden the tax base while limiting adverse effects on the poor.10 Tax reform has also promoted tax neutrality be- tween the consumption of domestically produced goods and the con- sumption of imports by reducing the average level of effective protection for the former and reducing rate dispersion for the latter. Income redis- I28 REFORMING THE PUBLIC SECTOR tribution objectives, however, have been pursued primarily through ex- penditure programs rather than tax programs, using implicit subsidies and explicit income transfers. Some countries have also improved their tax administration proce- dures. Value-acdded taxes (VATs) or turnover tax thresholds reduce the number of taxpayers to a level compatible with administrative capabil- ity. A broad-based VAT and a single rate or a limited number of rates (as in Benin or Burkina Faso) also simplify administration. Simple business license fees ancl other systems of presumptive taxation aimed at small businesses are a step in the right direction. Benin and Senegal simplified their penalty systems for tax evasion, and Benin, Chad, and Niger en- forced a tempo:rary shutdown of taxpayers' premises to improve compli- ance. Some countries have also set up special units to monitor compli- ance of large taxpayers, in order to stabilize the major share of tax revenue. Improving compliance is difficult when tax administration is weak and perceived to be applied in a manner favoring the politically well connected. Notes The final section of this chapter, on tax reform, is based 6. Giovanini and de Melo (1993) and Chamley and on a contribution by the Fiscal Affairs Department of the Honohan (1992) show that Africa's formal financial sys- International Monetary Fund. tems are indeed subjected to higher implicit taxation than other sectors. High reserve requirements on deposits and 1. See World Bank (1991a) for a discussion of an ini- forced lending to the government at low real interest rates tiative to build institutional capacity. are forms of implicit taxation. 2. Burundi, Cameroon, Congo, Cote d'Ivoire, Ghana, 7. Guinea had one privately owned bank in the pre- Guinea, Kenya, Madagascar, Mozambique, Nigeria, Sene- reform period, which was the sole survivor of the massive gal, Tanzania, Togo, Uganda, and Zambia all had at least banking liquidation of 1985. seventy-five public enterprises. 8. A few countries, such as Rwanda, may not need large 3. Those countries include Benin, Burkina Faso, reductions in civil service personnel. Congo, C6te d'][voire, Ghana, Guinea, Malawi, Mozam- bique, Nigeria, Tanzania, Togo, and Zambia. 9. The SPA coordinates balance-of-payments assistance to low-income, debt-distressed African countries that are 4. This was not the case in Cameroon, Congo, Gabon, committed to adjustment programs. or Kenya. Data are from the International Monetary Fund. 10. Noncascading value-added taxes and final-stage re- 5. See McKinnon (1973), Fry (1988), King and Levine tail sales taxes that exempt staples are examples of con- (1992), and World Bank (1989b), for example. sumption taxes. 129 CHAPTER 5 The Payoffs to Reforms S ADJU',TMENT PAYING OFF IN SUB-SAHARAN AFRICA? THE answer is a qualified yes. Adjustment programs may not have raised all countries' GDP growth, exports, savings, and invest- ment rates to those of adjusting countries in other regions. But the stronger reformers in Africa have turned around the decline in economic performance and are growing for the first time in many years. There also are signs that new firms are being created, that exports are growing, that private investment is picking up, and that savings petformance is improving. For the six countries with large improvement in macroeconomic policies: *Median annual GDP per capita growth was almost 2 percentage points higher during 1981-86 than during 1987-91; it was 2.6 percentage points lower for countries with deterioration in macro- economic: policies. * Median industrial growth was up 6.1 percentage points, compared with an improvement of 1.7 percentage points for countries with deteriorating policies. * Median export growth increased 7.9 percentage points; it declined 0.7 perce.rrtage points for countries where policies worsened. * Median gross domestic savings climbed 3.3 percentage points; the median fell 3.3 percentage points for countries with policy deterioration. * Domestic investment also recovered, albeit slightly. Furthermore, for the seven countries that substantially reduced taxation of their export crop producers, median agricultural growth rose 2 per- centage points; it declined 1.6 percentage points for countries that '31 A$JUSTMENT IN AFRICA raised taxes on producers. Overall, countries with better macroeco- nomic policies and limited government intervention in markets are en- joying GDP per capita growth, while those with a poor policy stance and more market intervention are facing further declines. The strong acceleration in GDP per capita growth between 1981-86 and 1987-91 for African countries with improving macroeconomic poli- cies compared favorably with that in other regions. Mozambique and Nigeria managed bigger shifts in growth than did Bolivia, the Philip- pines, or Thailand. Ghana, Sierra Leone, and Tanzania accelerated growth more than Mexico or Colombia and about as much as Costa Rica. After controlling for policy changes and initial conditions, the in- crease in GDP per capita growth among African adjusters appears to be no less robust than the increase among adjusting countries elsewhere.' These outcomes, while encouraging, are not as positive as they might be. Current growth rates among the best African performers are still too low to reduce poverty much in the next two or three decades. So far, the rebounds have merely brought countries back to their historical trend of low growth, and it is not yet clear whether they are shifting onto a higher-growth path. Without further substantial increases in agricul- tural, investment, export, and GDP growth, Sub-Saharan Africa will con- tinue to lag behind other developing regions. How to Assess the Payoffs to Reforms I-S ADJUSTMENT WORKING? TO ANSWER THIS QUESTION, WE NEED to determine the most appropriate indicators of both adjustment effort and adjustment outcomes. In this report we use the indexes of policy change described in previous chapters to measure the extent of reform. It is worth highlighting that we are not using adjustment lending as a proxy for reform efforts, but rather assessing how much reform has actually occurred.2 We argue further that the most telling indicator of the success of ad- justment efforts is the change in GDP per capita growth. The rate of growth that countries attain after implementing adjustment programs, though important in the long term, is not the best measure of how well adjustment is working in the short and medium terms. For decades, GDP per capita growth has been slower in Sub-Saharan Africa than elsewhere, 132 THE PAYOFFS TO REFORMS and adjustment: policy reforms are not likely to remove deep-rooted im- pediments to growth. That is the broader task of development, which encompasses mLlch more than the implementation of an appropriate policy framework. It is not fair, then, to measure the short-term success of adjustment policies in Sub-Saharan Africa only by the yardstick of growth rates in other regions. Furthermore, most adjustment programs in Sub-Saharan Africa are young, with many reforms implemented only recently, and it is too soon to realize big payoffs. But to the extent that the programs are addressing major macroeconomic imbalances and improving the overall policy en- vironment, somae positive outcomes-however modest-should be ex- pected. In particular, programs that removed the import constraints of the early 1980s, and restored external competitiveness while controlling inflation should have resulted in some improvement in growth. In addition t-o growth, what other payoffs would be likely? Export performance is the leading indicator of whether adjustment policies are working, because adjustment programs seek to redress external imbal- ances. But there is no reason to anticipate an improvement in the trade balance or the current account early on. In fact, an initial deterioration in the trade balance is in some cases desirable: because most African economies were initially import-constrained, growth could not recover without an increase in imports. The improved incentives for farmers (proxied by higher producer prices or lower rates of explicit taxation) should also lead to higher output of some agricultural commodities. But not much can be expected during the first years of adjustment, because aggregate supply responds slowly. Likewise, there should be modest increases in investment, primarily in countries with stable inflation, a clear direction in policy reforms, and a larger domestic savings effort. Industrial performance should improve too, as greater availability of inputs relieves production capacity constraints. Better Macroeconomic Policies Boost Growth F OURTEE.N OF THE ADJUSTING COUNTRIES INCREASED THE average annual rate of growth between 1981-86 and 1987-91, while fourteen others decreased it (figure 5.1).3 Mozambique, Nigeria, Uganda, and Ghana had the largest gains, exceeding 3 per- 133 ADJUSTM[E0NT IN AFRICA Figure St 1 Change es, Average anuFL7u 0 .-;@ [r rhg.Ee,z^C<r^JWi, f_Z`i-035 o 'i987-9!1 Median change in average annual growth (percentage points) 2.0 i.5 'D 0 . . 0.5 0.0 -0.5 Increases in external income -o.0 contributed to higher growth. Countries with increases Countries with declines in external income in external income Soz,;Ye: World Bank data 137 o ale 5, .1 Difference Diffrrence H between between the two the two z Average annual periods Average annual periods H growth rate (percent) (percentage growth rate (percent) (percentage County 1981-86 1987-91 points) Country 1981-86 1987-91 points) Leroe improvement in macroeconomic hnh;cies" Dpepriortcin'a (cntin.n,ued) , Ghana -2.4 1.3 3.7 Togo -2.8 -1.4 1.4 - Tanzania -1.7 1.3 2.9 Zambia -3.2 -2.3 0.9 > The Gambia 1.2 0.3 -0.8 Mozambique -5.9 1.7 7.6 Burkina Faso 2.2 0.4 -1.7 Congo 4.1 -0.7 -4.9 Nigeria -4.6 2.4 7.0 C6te d'Ivoire -4.2 -6.8 -2.6 Zimbabwe 0.3 1.0 0.7 Cameroon 4.6 -7.9 -12.5 Mean -0.8 1.1 2.0 Gabon -2.7 1.9 0.9 Median -0.7 1.1 1.8 Mean -1.0 -2.6 -1.6 Median -2.1 -2.0 -2.6 Small improvement' Madagascar -3.7 -2.1 1.6 Unclassified Malawi 1.4 0.7 2.2 Chad 4.5 2.6 -1.9 Burundi 2.1 1.2 -0.9 Guinea - - - Kenya -0.5 0.9 1.5 Guinea-Bissau 2.9 1.5 -1.4 Mali 0.4 -1.2 -1.6 Medians Mauritania -0.9 -1.0 -0.1 coi1ris 0.1 0.3 Senegal 0.4 -0.2 -0.6 Low-income countries -1.2 0.6 1.1 Niger -4.9 -2.4 2.5 Middle-income cotntries 0.3 -1.3 -1.6 Uganda -1.5 2.8 4.3 Countries with fixed Mean -1.1 -0.1 1.0 exchange rates 0.4 -1.7 -1.8 Median -0.9 -0.2 1.5 Countries with flexible Deteriorationa exchange rates -1.5 0.9 1.5 Benin 1.1 -2.0 -3.1 Oil-exporting countries 0.7 -1.3 -2.0 Central African Republic -0. 1 -2.8 -2.6 - Nor available. Rwanda o.4 _5.0 -5.5 a. Classifications are based on the overall scores reported in appendix table Sierra Leone -2.1 0.8 2.9 B.1. Source: World Bank data. THE PAYOFFS TO REFORMS opposite directions. Some countries (Benin, Burundi, the Central African Republic, Chad, and Rwanda) had declines in growth, despite increased external income. Improving Policies Makes a Difference Do changes in policies affect growth? Yes. Countries that improved their exchange r ate policy and reduced inflation and budget deficits gen- erally increased their GDP per capita growth rates (table 5.1). Our methodology was to divide countries into three groups, depending on whether they hlad large positive changes in macroeconomic policies, small positive changes, or negative changes, and to determine whether the groups differed in their economic performance. The six countries that improved policies the most had the largest median improvement in GDP per capita growth (1.8 percentage points), and all countries in the group returned to positive (though very low) growth rates. In contrast, among the countries where policies worsened, median growth declined 2.6 percentage points. This group includes the six countries with the sharpest downturns in growth. Finally, the countries that made small improvements in macroeconomic policies had a median increase of 1.5 percentage poinls. This was progress, but not enough to restore positive GDP per capita growth (this group had a median decline in growth of 0.2 percent a year cduring 1987-91).5 We also usecl econometric analysis to investigate the relationship be- tween policies and growth, controlling for the initial rate of growth and external factors (box 5.1). The objective was to establish whether changes in policies affect growth even after taking into account the effects of changes in external income (discussed above) and a country's initial per- formance (because countries starting with negative rates of GDP growth are likely to show more improvement). The analysis supports the forego- ing conclusion: changes in macroeconomic policies have a positive and statistically significant effect on growth (figure 5.4). However, the econo- metric analysis (does not provide conclusive evidence on the impact of ex- ternal factors on growth. While transfers are positively associated with growth, the recent improvement (or deterioration) in growth perfor- mance has been largely unrelated to changes in the terms of trade. Improvements in macroeconomic policies are correlated not only with larger changes in the rate of GDP per capita growth but also with higher levels of growth. Countries that improved macroeconomic poli- 139 kADJUSTMENT IN AFRICA :; :n :: : :Hf 0 : :0:u :: ' ,:) ~~~~~~~~~i NNXD,rX 2t S S A 15'E" Z Ept Xi THE FIVE REGRESSIONS IN THE TABLE BELOW IL- short time period under study and should not be taken lustrate the impact of changes in policies and external to contradict the well-established positive relation in income on growth. The regressions control for the rate the long run between growth and the terms of trade. of growth before adjustment, because countries that The main difference across regressions is the were doing particularly poorly (or well) are more likely choice of policy indicators. Regression I shows the re- to experience an improvement (or decline). The five sults for the change in GDP per capita growth using regressions are robust for the importance of policies the index of change in overall macroeconomic policy and the initial growth rate, which are statistically sig- described in appendix B. Regressions II through V ex- nificant and have the right sign. External transfers are amine the link between the change in the growth rate generallypositive,though notstatisticallysignificantat and the changes in the individual policy indicators the 10 percent level. The coefficient ranges between that are components of the overall change index. The -0.06 and 0.58, with the average less than the coeffi- policy variables have the right sign, and with the ex- cient used to predict growth rates in figure 5.2. The ception of inflation they are statistically significant at terms-of-trade effect is not significant and generally the 5 percent level. Inflation is statistically significant has the wrong sign (an improved terms of trade slows (at the 10 percent level) in regression II, but not in re- growth). This result reflects the peculiarities of the gression IV, where it is the only policy variable. ass ,@ 5. t=$we:S;N,,M',5 `dj P241D7,'~,yl:ng7 Cg"en c ngs ~T 'L -2rnJnA +'y-4>2,a ^.)/'bal. ll5>/tn7Al wDP pe,^ Ca17{ )5Utbi, 1987-86 to X,)87--312 Change in Income Overall from GDP macro- Exchange Overall Net changes in per capita economic rate fiscal external the terms growth, Adjusted Regression Constant policies policy Infation deficit transfers of trade 1981-86 R2 I -1.84 2.11 0.45 -0.55 -1.04 0.75 (-3.5) (4.4) (1.6) (-1.3) (-6.3) II -1.85 0.59 -0.43 -1.04 0.20 -1.08 -0.98 0.85 (-4.4) (2.6) (-1.8) (-4.5) (0.8) (-2.9) (-7.2) III -1.13 1.09 -0.06 -0.05 -0.84 0.70 (-2.2) (3.8) (-0.2) (-0.1) (-5.1) IV -0.62 -0.47 0.22 0.04 -0.92 0.53 (-1.0) (-1.1) (0.6) (0.1) (-4.4) V -2.00 -1.28 0.58 -1.06 -1.16 0.79 (-4.0) (-5.0) (2.2) (-2.5) (-7.5) Note: Numbers in parentheses are t-statistics. 140 THE PAYOFFS TO REFORMS F;igure .4 ChAangs un hi hi]vam euroomu Ffe s andi G)V per UaRa urG1rVIh 2 21- 'IC, fv37-si Change in average annual growth (percentage points) 10 C Mozambique n Nigeria 5 -Uganda Tanzania 1G Ghana Sierra Leone F 2Niger Mala i 2Gabon Zambia T go Kenya s adagascar El 13abon Zambia Eg ~~~Zimbabwe 0Oa _IIMauritania Senegal Burundi cEThe Gambia Mali H EZ Burkina Faso C6te d'ivoire 7 Central African Republic Benin El 5 i Congo C R ianda -10 z Cameroon -15 Deterioration Small improvement Large improvement in policies in policies in policies Note: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. Sources: Table 5.1 and appendix table B. 1. Changes in macroeconomic policies are positively correlated with cies the most had positive rates of GDP per capita growth during increases in GDP per capita growth. 1987-91, while those whose policies deteriorated had negative rates. But even for the top improvers, the growth rates remained low (at a me- dian of 1.1 percent a year). Adjustment policies, though instrumental in generating the conditions for higher growth, are thus only part of the so- lution. Shifting economies to a new growth path requires a good overall policy environmtent and better use of more traditional development in- struments, such as investments in education and infrastructure. To the extent: that macroeconomic policies matter-and they do- getting the exchange rate right is one of the top priorities for short-term growth. Countries that significantly reduced the parallel market pre- mium (by devaluing) and adopted realistic macroeconomic policies en- I4I ADUTENT IN AFRICA joyed the biggest payoffs. Dividing countries into three groups based on the real depreciation during adjustment shows that countries that man- aged large real depreciations (40 percent or more) enjoyed the largest turnaround, with a median increase in GDP per capita growth of 2.3 per- centage points (appendix table A. 16). Of this group, Burundi was the only country whose GDP growth declined. In contrast, countries that ex- perienced real appreciations suffered a median decline of 1.7 percentage points. There is a difference of 4 percentage points in the median changes for the top and bottom groups. The gap between the fixed and flexible exchange rate countries is also sizable (3.4 percentage points), because of the different adjustment strategies followed (box 5.2). African countries had a clear need for real depreciations in the early 1980s because of deteriorating terms of trade and unsound exchange rate policies. These initial conditions largely explain the effectiveness of devaluations in increasing growth. For countries that already have real- istic real exchange rates, there is little scope to further boost growth by depreciating, though gains would be expected from maintaining exter- nal competitiveness. But for countries with unrealistic exchange rates, the expected payoffs to real depreciations are large. Having Good Policies Also Makes a Difference We have argued that the change in GDP growth rates is the most appro- priate indicator of whether policy reforms are working in the short and medium terms. Over the long term, however, it is the rate of growth itself that ultimately matters. And over the medium to long term, the policy stance is more important than the change in policies in explaining the rate of growth. This is because changes in policies are not necessarily correlated with a country's policy stance. Tanzania showed large improvements in macroeconomic policies, and yet it still had a very poor policy stance in 1990-91 because its policies were so poor to begin with. By the same token, a country starting with reasonable policies might maintain an ade- quate stance even if its policies deteriorate. In Togo, macroeconomic poli- cies worsened in the second half of the 1980s (as the real exchange rate ap- preciated and the budget deficit increased), and yet the country's policy stance remained fair thanks to a sound macroeconomic framework in the early 1980s. What do we find when we examine the link between policy stance and the rate of growth? Median GDP per capita growth in countries with 142 THE PAYOFFS TO REFORMS adequate or fair macroeconomic policies was 0.4 percent a year-low but at least positive, and better than the decline of 1.2 percent a year during 1981-86 (appendix table A.17). Meanwhile, countries with poor or very poor macroeconomic policies had staggering declines in GDP per capita, with the median at -2.1 percent a year during 1987-9 1. The median growth rates were also quite different for countries with limited, medium, and heavy government intervention in markets: 1.9 percent, -1.1 percent, and -1.4 percent, respectively. More than 4 per- centage points separated the three countries with the best macroeco- nomic policies and the least intervention from the two countries with the worst macroeconomic policies and heavy intervention-a powerful incentive for moving toward good policy. Agriculture Is Growing Faster T* ! HE MOST IMPORTANT WAYS THAT ADJUSTMENT AFFECTS agriculture are by reducing taxation through macroeconomic policy changes and by reforming the arrangements for market- ing. Chapter 3 showed that reforms shifted the terms of trade in favor of agriculture in many countries. The taxation (both explicit and implicit) of export crop producers fell in about two-thirds of the coun- tries, and in ten countries real prices for export crop producers were higher at the end of the decade than at the beginning. Few reliable data series are available on producer prices for food crops, but several coun- tries eliminated food crop marketing boards or substantially cut back on their activities, virtually eliminating the parallel market. These reforms, proxied by the change in real producer prices for ex- port crops and the change in overall taxation of export crop producers, appear to have buoyed agricultural growth.6 Countries that improved the real prices ior agricultural exporters increased agricultural growth, while countries Xhat let prices decline saw agricultural growth fall, for a difference in medians of more than 2 percentage points (top of figure 5.5 and appendix table A. 18). Even more striking is a comparison of the countries that slabstantially reduced the level of taxation and those that increased it: the difference in median growth rates was 3.6 percentage points (bottom of figure 5.5 and appendix table A. 19).7 But no clear pattern emerged distinguishing countries in which macroeconomic '43 A US ENT IN AFiRICA Box 5.2 7 Fanye _2di aii2t [EAdbe Ex hange, Rat, ltegimes WHEN REAL DEPRECUITION IS NEEDED, A FLEMBLE export growth stagnated. The flexible exchange rate exchange rate generally reduces the output costs of de- countries did better. Median GDP per capita growth preciating because it allows a quick initial real depre- climbed from -1.5 percent a year during 1981-86 to ciation to raise competitiveness. But it increases the 0.9 percent a year during 1987-91, with real export risk of a permanent increase in inflation. A fixed ex- growth showing even more improvement. change rate reduces that risk, generally at the expense The disparity in outcomes for the two groups of of output, because recession is needed to bring down countries extended to other economic indicators. domestic prices to achieve the required real deprecia- Gross domestic investment as a share of GDP fell about tion. The stronger the resistance of workers to wage 3 percentage points in the fixed exchange rate coun- cuts and the more opposed producers of nontradables tries, but increased slighdy in the flexible exchange are to price cuts, the longer it will take to effect the rate countries. And gross domestic savings in the fixed real depreciation-ancd the more costly (in forgone exchange rate countries, while increasing from low output) the fixed exchange rate strategy becomes. In levels, still fell short of the median level of savings in both, a well-designed adjustment package needs to in- the flexible exchange rate countries. Poverty worsened clude tight domestic credit and fiscal policies to re- in the fixed exchange rate countries because of the loss duce domestic absorption. of income-earning opportunities and the inability of The median real depreciation between 1981-86 fiscally constrained governments to provide basic so- and 1987-91 in the flexible exchange rate economnies cial services. And because the reforms did not lower was 50 percent, compared with an 8 percent real ap- nominal wages or the prices of nontradables relative preciation in the fixed exchange rate economies (ap- to the prices of tradables, the attempted internal ad- pendix table A.16). The devaluations in the flexible justment failed to restore competitiveness and exchange rate countries had an unavoidable short- growth. term impact on inflation, but by 1990-91, once the Relying solely on internal adjustment measures is initial effects had worked themselves out, inflation fell best for correcting small losses of competitiveness, par- back to the levels of the first half of the 1980s. The an- ticularly those from excess internal demand. Under an nual inflation rate declined in nine of sixteen coun- internal adjustment, competitiveness can be regained tries with flexible exchange rates, and overall, the me- only at the rate by which international inflation ex- dian remained virtually unchanged-19 percent ceeds domestic inflation. The fixed exchange rate during 1981-86 and just under 21 percent during countries have been able only to freeze-not reduce- 1990-91 (appendix table A.3). Meanwhile, the fixed domestic wages and prices. Meanwhile, because inter- exchange rate countries continued to have low infla- national inflation decelerated in the 1980s to 2 to 4 non-a median of 6.3 percent during 1981-86 and percent annually, the countries with fixed exchange 0.5 percent during 1990-91. rates can at best improve their competitiveness by 2 to Overall, however, the macroeconomic perfor- 4 percent a year. At that rate, it will take them a decade mance of the fixed exchange rate countries deterio- or more to become competitive-longer if their com- rated significantly in the second half of the 1980s and petitors further depreciate their exchange rates. With- was clearly worse than in other African adjusting out big changes in policies to restore competitiveness, countries (box figure 5.1). In the countries with fixed the outlook for these countries is grim: low invest- exchange rates, median real GDP per capita growth fell ment, financial crises, worsening poverty, faLling ex- on average by 1.7 percent a year during 1987-91, and ports, and continuing declines in per capita income. '44 THE PAYOFFS TO REFORMS Medlabis fog Selected hAlcatbre Real Effective Exchange Rates Index: 1980=100 160 1987-91 140 120 8 1987-91 100 1981-86 60 40 20 0 Countries with fixed Countries with flexible exchange rates exchange rates GDP per Capita Growth Real Export Growth Percent Percent 1.0 . . 0.5 1981-86 4 0.0 ~~1987-91 1981-86 3 -015 1-981-86 1987-91 2 -1.01918 -2.0 0 Countries with fixed Countries with flexible Countries with fixed Countries with flexible exchange rates exchange rates exchange rates exchange rates cross Domestic Investment as a Share of GdP Gross Domestic Savings as a Share of GDP Percent Percent 24 8 971 20 ~~~~~~~~ ~~~~~~~1987-91 1981-86 1981-86 ~1981-86 1987-91 6 16 1987-91 12 4 1981-86 8 4 0 0 Coujntries with fixed Countries with flexible Countries with fixed Countries With flexible exchange rates exchange rates exchange rates exchange rates a, An increase in tbe real exchange rate constirures a depreciation; a decrease constitutes an appreciation, Sources: World Bank data; table 5.1; appendix tables A.22, A.23, and A.24. 145 ADJUSTMENT IN AFRICA ?igaire 5.": Le0aTh 0strasr ugdwV rr Cvwc tIn, 7R t F dQgrJcLlFb?]g 7hF]inwf7 12'c=23(C Ž X3KY=W7 Producer Prices and Agricultural Growth Median change in average annual growth (percentage points) 3 2 0 -2 Countries with improvement Countries with deterioration in the real producer price in the real producer price of agricultural exports of agricultural exports Taxation and Agricultural Growth Median change in average annual growth (percentage points) 3 Better prices for farmers and less -1 taxation meant faster agricultural growth. -2 Countries with large Countries with small Countries with decrease in taxation decrease in taxation increase in taxation of export crop of export crop of export crop producers producers producers Sources: Appendix tables A. 18 and A. 19. policies improved from those in which they deteriorated (appendix table A.20). This is not too surprising, given that macroeconomic policy changes do not always translate into higher producer prices because of lags in marketing board reforms and other government interventions. Policy changes may be moving some countries back to their long-run agricultural growth trends, but the conditions are not yet in place every- 146 THE PAYOFFS TO REFORMS where for accelerated agricultural growth rates. Per capita agricultural growth rates are still negative for most countries. Many countries, such as Tanzania and Burundi, continue to operate below their agricultural potential because of macroeconomic and agricultural pricing distortions and inefficient rnarketing arrangements. Moreover, both policy and in- stitutional factors have strong effects on supply response, as the differ- ence in the agricultural performance of Ghana and Nigeria shows (box 5.3). The aggregate supply response takes time to materialize, and ex- trapolations of short-term supply responses may underestimate the long-term response to policy reforms (box 5.4). Bx 53 lhy Agicultur Is Growing Faster in igeria Tl h !-a $lan m ALTHOUGH- GHANA IS REGARDED AS HAVING a Fourth, Ghana's main crop, cocoa, has suf- made the most progress of any adjusting country in fered an extended period of depressed world Sub-Saharan Africa, its performance in the agricul- prices. Cocoa does not account for as large a tural sector has lagged behind that of others. Nigeria, share of Nigeria's agricultural exports, and the for example, has had substantially stronger growth in terms-of-trade effects on its other export crops agriculture. Whereas Ghana's agricultural sector have been less severe. grew about 2 percent a year during 1987-91, Nige- * Fifth, while Ghana implemented overall ad- ria's grew about 4 percent a year, and the net real re- justment policies earlier, reforms within agri- turn to farming increased fourfold. culture did not get under way until the late Several factors account for this difference: 1980s. Nigeria proceeded quickly on agricul- tural reform in 1986. * First, -Nigeria eliminated its export crop market- ing boards, which has allowed Nigerian produc- The lesson from this comparison is the impor- ers to obtain a larger share of export prices. In tance of removing bottlenecks in the agricultural sec- Ghana a number of marketing boards, including tor. Structural differences seem to outweigh different the Cocoa Board and the Cotton Mills, continue exchange rate policies in explaining outcomes. On to exert a major influence on producer prices. the policy side, Ghana appears to have an advantage * Second, Nigeria's investment in research and in allowing the exchange rate to adjust quickly. Nige- extension services has provided farmers with ria has relied upon periodic devaluations, a practice the technical know-how to improve small- that leads to occasional misalignments. Structural scale irrigation and soil fertility. conditions in Nigeria, however, are more favorable * Third, rural infrastructure in Nigeria is more to agricultural growth. In addition, experience in extensive than in Ghana, which can be charac- Nigeria suggests that small-scale irrigation schemes terized as largely a "footpath economy." The in which farmers participate in design and imple- cost of poor rural infrastructure is significantly mentation are more effective and less costly than reduced labor productivity and higher produc- large public irrigation systems. Furthermore, farmers tion costs, which translate into reduced profits are more likely to adopt new technologies if they see and fewer production incentives. the linkage to higher financial returns. 147 ADUSMNT IN AF RICA Box 5.4 Age gate SUPQipy Tends to Res §rdSlow y $ -n :ha SharFfm JAEGER'S (1992) ANALYSIS OF THE RESPONSE OF And real effective exchange rate depreciations may in- export crops and food crops to policy changes in crease the (official) price of imported staple foods- Africa provides furthel- evidence that aggregate sup- and thus the domestic price of foodstuffs that can ply responds to favorable policy changes. Specifically, substitute for imports. The effect on domestic prices Jaeger examines the response of export crops to may be minor, however, since prices in the parallel changes in real producer prices, the real effective ex- markets in large part reflect the foreign exchange pre- change rate, and variables representing weather and mium before the liberalization of import markets. disasters (war, civil strife, famine) for twenty-one Jaeger's analysis squares with other research find- Sub-Saharan countries having data for 1970-87 ings that document the responsiveness of African (countries that overlap with our sample of adjusting farmers to changes in the prices of individual crops countries). The estimated price elasticities of export (see Bond 1983). The price elasticities of supply for supply range from 0.1 io 0.3, and the elasticities with individual crops are high in the short run-and even respect to the real effecive exchange rate (where a fall D higher din the long run. But as Binswanger (1989) indicates an appreciation) range from -0.01 to points out, the aggregate supply response is more con- -0.25. Countries exporting annual crops had higher tentious. In the short run, the supply of the main fac- elasticities than countries exporting tree crops. The tors of agricultural production tends to be fixed, so results support the cornclusion that the aggregate re- that a relative price shift in favor of one crop shifts fac- sponse in the short run is positive but not large. tors of production into that crop at the expense of an- Jaeger also finds that total agricultural roduc- other. Consequently, aggregate agricultural supply tion, total food production, and staple food produc- elasticities tend to be small even if individual elastici- tion respond to increases in export crop prices and ties are large. But over the long run-and long can be depreciations of the real effective exchange rate. The ten to twenty years-the supply response can be large, results suggest no tradeoff betwveen food and export with less rural outmigration and greater investment in crop production-results consistent with other em- agriculture, roads, markets, education, and health. pirical findings of a positive correlation between the ATheshort-run price elasticitiesofaggregatesupply growth in export crops and that in food crops (see for various countries reported in Binswanger (1989) von Braun and Kennedy 1987). ranged from 0.05 to 0.25. Those for Sub-Saharan Jaeger offers severa l explanations for this. First, Africa were no lower than for other areas, suggesting food crops can benefit from applications of fertilizer that the lower levels of infrastructure and other factors to export crops when they are grown together or in do not pose significantly greater constraints to a supply succession. Second, highernincome from-export crop response in Sub-Saharan Africa. Very few method- production can lead t higher investment in foodt ologically sound studies have looked at the long-run crops. In addition, farrners may not have to supple- supply response-with none for Africa-but one for ment their incomes with off-farm employment out- Argentina estimated a supply elasticity of 0.4 after five side the growing season, wsothey may be able to devote years and 1.8 after twenty years. A study of Chilean more labor to food crops. Third, a better policy envi- agriculture cited by Valdes (1989) found an aggregate ronment, proxied by higher export pries and more supp elasticity of 0.3 after three years, 0.6 after four competitive exchange rates, helps both export and and five years, and more than 1.0 after ten years. This food crops by promotinlg growth and thus evince suggests that, if policies improve, we should aggregate demand. The dismantling of foodcrop expect substantial increases in aggregate supply not in marketing boards stimulates foodt crop production. the short run but in the medium and long run. I48 THE PAYOFFS TO REFORMS Industry Is Expanding RITICS OF STRUCTURAL ADJUSTMENT PROGRAMS IN AFRICA point to an anti-industrial bias in the policy package (see Lall 1992; Stein 1992). They fault policies that suddenly reduce trade protection, devalue the currency, cut public spending, remove government subsidies and price controls, increase real interest rates, and promote privatization. Such policies, they argue, not only cause industrial production to stagnate or decline in the short run, but also erode an important part of the industrial base for future growth. Put simply, the critics claim that structural adjustment programs are causing the deiridustrialization of Africa. One piece of evidence cited is that about half the countries in Sub-Saharan Africa had declines in in- dustrial output in the first part of the 1980s, while the rest of the region had output growth of less than 2 percent (Stewart, Lall, and Wangwe 1992). But manay of the countries with negative growth had not imple- mented adjustment programs, or had done so only recently, so their poor performance cannot be blamed on adjustment. Assessing whether deindustrialization has occurred is not easy, but there is no systematic evidence of it in our sample of adjusting countries (box 5.5). Countries that made large improvements in their macroeco- nomic policies had strong increases in the growth of industry and man- ufacturing-wi.th median increases close to 6 percentage points between 1981-86 and 1987-91 (figure 5.6 and appendix table A.21). The re- sponse was weaker in countries with small improvements in macroeco- nomic policies and weakest for countries with deterioration in policies. In all, only eiglht adjusting countries had contractions in their industrial sectors, and five of those were from the group with worsening policies. Surveys of manufacturing firms in Ghana, Kenya, and Tanzania fur- ther support the contention that strong reform efforts do not cause deindustrialization.8 The picture in Ghana, the country with the most extensive adjustment, is not one of stagnation and deindustrialization; instead, it shows much activity, particularly among smaller enterprises not included irn official statistics. Aggregate employment and output have been increasing about 2 percent a year for a sample of manufactur- ing firms in four sectors representing about 80 percent of manufactur- ing employment-food processing, metalworking, woodworking, and textile and garrrient manufacturing. Both old firms (those operating be- fore adjustment) and new firms (those started after adjustment) have 149 AJfST NT IN AFRICA Box 5.5 o tHow W e iel n deiz pon rrin gr PROPONENTS AND DT TORS OF STUC * Trends in industry's contribution to the do- adjustment agreethat Africa'sindustrialgrowthwas mestic economv, including changes in the :generally inefficientbefore adjustment (Stewart, Lall, growth of real manufacturing value added over and Wangwe 0 1992). ][n many cases, heavy-handed a period long enough for the economy to move government policies to promote import-substituting through the difficult transition phase immedi- industrialization-at::thle e-xpense of :agriculture and ately after policy reforms. Another indicator, othersectors-expanded the manufacturing sector though more flawed, is the relative share of beyond its efficient size and distorted its structure. manufacturing output (in constant prices) in Such policies can lead :o rapid growth of output and GDP over the same period. eployment in the short run, but that growth m Employment. Changes in emplovment indica- generally is not sustainable and soon peters out, leav- tors need to be interpreted carefully, because ing in its wake intlernal and external imbalances-and adustment programs can lead to a rationaliza- often economic crisis. The economy, distorted and,00 tion of the manufacturing sector and thus pos- piushed beyond its eflicient limits, requires down- sibly to a short-term reduction in jobs. As ad- sizing and restructuring. In such cases, any deindus- justment proceeds, however, employment trialization is a move toward more efficient outcomes, should increase as new activities come on not towad a structural maladjustment of the econ- stream and overall production capacity rises. oiny. But it is also possible that downsizing and re- a Indicators of trends in production capacity the structuring could be the unintended consequences of number of manufacturing company bankrupt- a policy package that inhibits manufacturing growth. cies, the number and size of establishments, To make the case that structural adjustment pro- and the average capacity in plants. Xgrams acausedeindustriifization in Afia, one woud Shifts in resources from activities with high do- have to show that (a) there are significant des in mestic resource costs to those with low costs, industrial output, output shares of GDP, a employ- providing evidence of efficiency improvements. ment because of policv reforms, (b) th decines in u Trends in productivity growth and changes in output and empl e am nthe growth rate of manufactured exports. vfoutcomes: of efficiet dustment, (c) t cag i the industrial sectr arento- Time series for most of these indicators are not ward greater effici nd(d)tpolicy reforms are avaable, making it impossible to resolve the deindus- iipeding long-r n growt alization debate conclusively. But the data we have tion by inhibiting thepace Sand pattern iof investment as; sembledi: from national accounts and from surveys of and thus the shiftto ahighergrowthtrajectory. 0 manufacturingV firms: do not support the hypothesis [ :What indicators mrnight be Sused tof examine: these t0 that structural adjustment programs have led to dein- issues? dustrialization in Africa. been growing, but new firms outpace older enterprises in every size class. Some old, larger firms are downsizing to become more efficient in re- sponse to adjustment policy changes, and some have shut down. Of forty-three medium-sized and large old firms, 63 percent had fewer em- ployees in 1991 than in 1983 (at the start of adjustment). For all new firms in these size classes, however, the number of employees rose be- 150 THE PAYOFFS TO REFORMS ligire 5£6 Ii anges in the Growth cf lnduistry and Manu ac uring, LS$,4$ t KO$T-$l Median Change in Industrial Growth Median Change in Manufacturing Growth Percentage points Percentage points 7 7 6 6 5 5 4 4 3 3 2 2 0 0 Countries with large Countries with small Countries with Countries with large Countries with small Countries with improvement in improvement in deterioration in improvement in improvement in deterioration in macroecononmic macroeconomic macroeconomic macroeconomic macroeconomic macroeconomic policies policies policies policies policies policies Source: Appendix table A.21. tween startup and 1991. Almost two-thirds of the increase in manufac- Macroeconomic policy reform turing employnment since the start of adjustment was attributable to new spurred industrial and manufacturing entrants. A significant proportion of new entry has been concentrated in growth. small enterprises in the woodworking and metalworking sectors, where domestic resource costs are relatively low. Adjustment programs radically change relative prices and the incen- tive environment that firms operate in-and thus the characteristics and behavior of firns that survive and prosper. For example, surviving en- trepreneurs would be expected to focus more on raising production skills and cutting costs than on seeking rents. The new firms in Ghana (those started after 1963) are run by entrepreneurs with better-educated employees.9 Armong small firms, three-fourths (77 percent) of the own- ers of new firms have gone through apprenticeship programs, compared with 53 percent of the owners of older firms. New firms also have a higher propensity to invest, and this has increased since 1987. And of the medium-sized and large firms, more new firms export than old. Contrary to the conventional view that policy conditions in Africa inhibit enterprise growth, there has been substantial movement of en- terprises from one size class to the next during the adjustment period. For example, 6,7 percent of the new medium-sized firms in Ghana began as small firms, compared with 50 percent of the old medium-sized firms. This greater upward mobility, combined with differences in en- I5' ADJUSTMENT IN AFRICA terprise attributes, suggests that new firms may be better able to take ad- vantage of new economic opportunities. The big question is whether the activity of small firms is a structural break with the past or simply a sign of distress. Are many of the smaller new entrants simply household efforts to survive at the margin, or are they dynamic new enterprises that can significantly increase employ- ment in the future? In every size class, the new entrepreneurs have more human capital than old firms-and new firms, on average, have higher export intensities, higher propensities to invest, similar or higher levels of total factor productivity, and faster movement through the size classes. All this supports the view that adjustment programs are leading to more efficient outcomes in the industrial sector. What is not yet clear is whether the conditions are in place to shift the adjusters to a much higher growth trajectory. Given Africa's lack of in- dustrial experience, an increase in technological capability would seem to be a prerequisite for further, fast-paced industrial growth. As Pack (1993, p. 1) argues, "prices are one-half of a scissor, the other being tech- nical skill." In Africa one might add upgraded infrastructure and busi- ness support services to that other blade of the scissor. Exports Are Growing A FRICA'S EXPORT PERFORMANCE BEFORE ADJUSTMENT WAS poor. Between 1965 and 1986, real exports from Sub-Saharan LAfrica merely doubled, while those of non-African adjusting countries quintupled. Between 1987 and 1991, the median rate of export growth for the African adjusting countries was 3.6 percent a year, compared with 9.2 percent in other developing countries. Although this is an improvement from 1.3 percent a year between 1970 and 1986, export growth needs to accelerate. Exports are expected to expand early in successful adjustment pro- grams, because successful programs increase external competitiveness and rely on export growth to offset the output costs of stabilization ef- forts. Although export growth remained positive in most adjusting countries-the exceptions being Benin, Cameroon, the Central African Republic, Mauritania, Sierra Leone, and Zambia-the performance was uneven, and for ten countries the rate of export growth slowed. Coun- 152 THE PAYOFFS TO REFORMS Figure 52 7Median :lange in ReaD Zspcst Gkoniha, dE2-i7 iC 0 L987-sl Percentage points 8 6 4 2 0 Macroeconomic policy reform paid -2 off in faster real export growth. Countries with large Countries with small Countries with deterioration in improvemient in improvement in macroeconomic policies macroeconomic policies macroeconomic policies Source: Appendix table A.22. tries with the largest improvements in macroeconomic policies enjoyed the highest median growth in exports and had by far the largest surge in export growth, wvith a median of almost 8 percentage points (figure 5.7 and appendix table A.22). In contrast, exports lost ground in countries with deteriorating macroeconomic policies, where the median growth rate fell 0.7 percentage points. Despite the higher exports, few African countries have managed to di- versify their export base much. Almost all African countries continue to depend heavily on primary commodities for their export earnings, with no more than three primary commodities constituting over 70 percent of the value of exports. There are, however, signs of export diversification in countries such as Burundi and Kenya (Husain and Faruqee forthcoming). No Quick Response in Investment or Savings A LTHOU GH INVESTMENT IN MANY COUNTRIES HAS NOT YET reachecl leve]s consistent with sustainable long-term growth, there are indications of improvement in some countries. The share of private investment is also increasing, albeit from a very small I53 ADUSMNT IN AFRICA base. Savings have traditionally been low throughout the region, and adjustment has had little impact on them so far. Investment Investment generally responds slowly to adjustment programs-in Africa and elsewhere (World Bank 1992a; Serven and Solimano 1992; Mosley, Harrigan, and Toye 1991). This slow response is understand- able. Governments cut capital spending as part of their fiscal stabiliza- tion, while the private sector adopts a wait-and-see attitude during the early phases of adjustment, mindful of the irreversibility of investment decisions and the reversibility of key policy changes (indeed, policies have frequently been reversed in the past). The problem is particularly serious where there is no consensus about the importance of private sector-led growth. Median gross domestic investment in African adjusting countries de- clined from more than 21 percent of GDP in the second half of the 1970s to 17 percent in the early 1980s and 16 percent in the second half of the decade. By comparison, investment in other adjusting countries re- mained stable at about 23 percent of GDP during the 1980s, after de- clining from 26 percent of GDP in the late 1970s. Africa needs to in- crease investment to about 25 percent of GDP to sustain GDP growth of about 6 percent a year. Investment fell in thirteen of twenty-eight African adjusting coun- tries in the second half of the 1980s, with performance related to changes in macroeconomic policies (appendix table A.23). It is impor- tant here to distinguish the oil exporters from the other countries, be- cause the oil exporters entered the adjustment period with much higher (and many times unsustainable) investment-to-GDP ratios (about 40 percent in Congo and Gabon). So, unlike most other adjusters in the re- gion, the large cuts in investment that these countries made were ex- pected and desirable. The ratio of investment to GDP declined signifi- cantly in the oil-exporting economies (a median fall of 7.7 percentage points). Despite the large reductions, the investment-to-GDP ratios in these countries were still among the highest in the region. Investment increased in the non-oil exporters that improved their macroeconomic policies and declined markedly in those where policies worsened (figure 5.8). The median change in investment as a share of GDP was 1.1 percentage points for countries with large policy improve- 154 THE PAYOFFS TO REFORMS Figure [&a.i&dhn [ghaLgt Pbs Utxinuzltii& [Iwartf Seignioragea Inflationb Seigniorage- Inflationb Country 1981-86 1990-91 1981-86 1990-91 Country 1981-86 1990-91 1981-86 1990-91 Extremely high seigniorage in 1990-91 Low seigniorage (continued) Tanzania 3.8 7.6c 30.6 21.0 Senegal 1.0 -0.8 11.0 -0.7 z Sierra Leone 6.2 6.2 57.1 106.8 Chad 1.8 0.8 5,Oe 0.6e Zambia 3.6 4.0 26.2 101.8 Mean 1.3 -0.3 12.5 3.5 Mean 4.5 5.9 38.0 76.5 Median 1.1 -0.3 7.7 0.6 Median 3.8 6.2 30.6 101.8 . 7 Unclass,fied C High seigniorage Burundi 1.0 - 7.7 8.0 > Nigeria 1.1 2.9 17.4 10.2 Guinea-Bissau - - 40.5e 35.0e Benin -2.4 2.7 403d OId Guinea - - 24.2c Zimbabwe 0.8 2.5 15.0 20.8 Mozambique - - 18.5e 35,9c The Gambia 1.8 1.8 20.7 W0.4 Uganda 87.6e 32.2e Kenya 1.2 1.8 11.8 13.3 Mean - - 38.6 27.1 Mean 0.5 2.3 13.9 11.0 Median - - 29.5 32.2 Median 1.1 2.5 15.0 10.4 Medians Moderate seigniorage All countries 1.1 0.9 10.6 8.0 Madagascar 1.6 1.5 19.4 6.5 ILow-income countries 1.6 1.3 12.6 10.4 Togo 1.9 1.4 6.5 1.0 Middle-incone countries 0.8 0.3 10.0 0.5 Mauritania 1.7 1.1 103e 6.60 Countries with fixed Malawi 1.0 1.0 13.3e 12.20 exchange rates 1.0 0.0 6.3 0.5 Rwanda 0.2 0.8 5.3 11.9 Countries with flexible Congo 0.1 0.8 9.8e -0.3e exchange rates 1.6 1.8 19.0 20.8 Mean 1.1 1.1 10.8 6.3 - Nnt available. Median 1.3 1.1 10.0 6.5 a. Scignioragc was calculatcd as (MI,-MM1 -)/GDP-g,(MI/GDP)t, where Low seigniorage Mlt is the stock of money at the end of period t, GDP, is gross domestic product Ghana 3.3 0.4 56.0 27.6 at time t, and g, is real GDP growth. C6te d'Ivoire 1.1 0.4 5.9 0.6 b. Inflation was calculated as the percentage change in the consumer price index. Gabon 0.6 0.2 92 52c. Based on M2. Gabon 0.6 0.2 9.2 5.2 d. 1se private consumption deflator was used when the consumer price index Cameroon 0.7 0.0 10.2 0.5 was not available. Burkina Faso 1.0 -0.1 6.2 1.9 e. The GDP deflator was tised when the consumer price index was not Mali 1.7 -0.6 6.0d 3,3' available. Central African Republic 1.1 -0.6 9.6e 0.3e SourceslMF and World Bank data. Niger 0.6 -0.7 6.0 -4.3 APPENDIX A Table A.4 Real nateress Rate for Deposits (percent) Country 1981-86 1990-91 Country 1981-86 1990-91 Highly positive interest rate in 1990-91 Highly negative Niger 6.3 16.0a Rwanda 1.5 -9.9 Senegal -1.7 8.9a Zimbabwe -3.3 -12.5 Congo 0.2 8.7a Sierra Leone -36.0 -30.7 Cameroon -1.5 8.7a Mean -12.6 -17.7 Central African Republic 1.9 6.8a Median -3.3 -12.5 Gabon 0.3 6.7a Uckusifid Benin 4.8 6.6a Burundi -1.8 i Togo 4.0 5.9b Guinea -1. C6te d'Ivoire 2.5 54a Ganea Chad 3.9 2b Guinea-Bissau - - The Gambia -9.8 3.2a Madagascar -4.8 Mean 1.0 7.5 Mauritania -4.0 - Median 1.9 6.7 Mozambique - - Tanzania -20.1 __d 'Acceptable " range Zambia -15.8 - Mali 2.4 2.8a Ghana -16.4 2.8a - Not available. Burkina Faso 2.7 2.6a Note: The real interest rate is calculated as [(1 +R) + (lI t+) Nigeria -5'9 1,7a -1] *100, where R is the nominal interest rate and iT is the Nigeria -3.9 10.7a inflation rate. Malawi -3 7 -0 3a a. The real interest rate for 1991 was calculated using the Kenya 0.9 -1.0 1991 inflation rate. Uganda -38.6 -2,9a' C b. The real interest rate for 1990 and 1991 was calculated Mean -8.4 0.8 using the 1990 inflation rate. Median -3.7 1.7 c. The inflation rate was based on the annual change in the consumer price index and on the implicit GDP deflator. d. Precise data were not available for Tanzania in 1990-91, but World Bank staff estimate the real interest rate to have been about 5 percent. Source: IMF data. 227 ADJSTENT IN AFRICA 7bl bt! Prle akREchrge lRate @rmup(ectt Country 1981-86 1990-91 Sub-Saharan adjusters Extremely high premium in 1990-91 Mauritania 121.5 166.6 Mozarmbique 2,110.8 62.6 Sierra Leone 49.4 104.4 Tanzania 248.8 74.5 Zambia 46.3 149.7 High premium Rwanda 43.7 47.5 Moderatepremium Burundi 24.1 20.9 The Gambia 13.8 21.3b Malawi 53.6 29.4 Nigeria 232.7 25.1 Uganda 190.0 24.6 Zimbabwe 81.3 23.5 Low premium Ghana 1,098.2 3.4 Guinea 655.2 7.6 Guinea-Bissau 59b7 Kenya 15.1 7.3 Madagascar 42.0 7jlb M 0 t00 04ean\l t 0 0 l0\0t 0000000 \000 t299.2 45.5 Excluding Mozambique 186.0 44.4 < 01 ~~~~~ ~ ~ ~~~~Median 59.7 24.6 Other adjusting countries Argentina 32.8 42.4b Bolivia 136.2 1.5b Costa Rica 204.7 15 9b Indonesia 4.2 2.6b Mexico 13.9 6.8b Morocco 6.0 13.1b Philippines 12.3 7.lb Thailand -2.2 2.0b Turkey 9.6 2.4b Venezuela 110.3 5.2b Mean 52.8 9.9 Median 1 3.1 6._0 Note: The parallel market exchange rate premium is calculated as the per- centage difference between the parallel market exchange rate and the official exchange rate (in domestic currenicy at the,end of the period). a. Data are only for counitries with flexible exchange rates. b. Data are for 1990. Sources: International Currency Analysis, Inc. (various years); IMF data. z28 APPENDIX A Table A.6 Change in the Real Effective Exchange Rate, 1980 to 1990-91 (pereent) Countries withfixed exchange rates Benin 10.8 Burkina Faso 10.3 Cameroon -18.4 Central African Republic 9.1 Chad 18.7 Congo -9.2 C6te d'Ivoire 2.8 Gabon 7.8 Mali 10.8 Niger 28.6 Senegal -4.0 Togo 9.7 Countries with flexihle exchange rates Burundi 37.3 The Gambia -26.0 Ghana 471.0 Kenya 64.2 Madagascar 109.9 Malawi 6.8 Mauritania 26.7 Mozambique 39.7 Nigeria 355.5 Rwanda -11.5 Sierra Leone -11.7 Tanzania 182.2 Uganda 945.3 Zambia 132.2 Zimbabwe 74.6 Note: An increase in the real effective exchange rate constitutes a depreciation; a decrease constitutes an appreciation. Guinea and Guinea-Bissau are excluded because of insufficient data. Soul ce: World Bank data. 229 ADJUSTMENT iN AFRICA Tabe A.7 Foreign Excha age Allocation and Import Controls for Selected Countries Controls on Implied import foreign exchange Parallel market restrictiveness allocation exchange rate (index scores)a Before Late premium (percent) Before After Country reforms 1992 1985-86 1990-91 reformzs reforms Severe overvalsation followed by lowering of trade barriers Ghana * 0 142 3 217 29 Severe overvaluation followed by substantial progress The Gambia 0 0 96 21 - - Guinea-Bissau * 0 66 -2 - - Madagascar * 0 7 7 37 11 Mozambique * 0 4,457 63 - - Nigeria 0 O 235 25 34 1 Tanzania 0 & 253 75 103 39 Uganda 0 0 337 24 - - Zambia 0 0 36 150 65 57 No oveivaluation, ongoing liberalization Burundi * 0 22 21 12 7 Kenya 0 0 2 7 45 60 Malawi 0 0 26 29 - - Rwanda 0 0 39 47 - Zimbabwe * 0 56 23 -9 -15 * All foreign exchange allocation is controlled. 0 25-50 percent is controlled. 0 Less than 25 percent is controlled. O Almost no controls. - Not available. a. The index measures the percentage of deviation of actual imports from their notional demanded level. Notional import demand is defined as tY + iGDP + yREER, where oe is a constant calculated to equalize actual and notional imports in the 1970s, f is an income elas- ticiry equal to 1.25, Sy is an exchange rate elasticity equal to 1, and REER is the real effective exchange rate. Sources: World Bank staff; International Curtrency Analysis, Inc. (various years); IMF data; Narasimhan and Pritchett (1993). 230 APPENDIX A Table A.8 Import Items Subject to Nontariff Barriers Before Late Country reforms 1992 Benin All 0 Burkina Faso Hundreds i9a Burundi Nearly all 0 Cameroon Hundreds Hundreds Central African Republic Chad Congo 33 categories 13 categories C6te d'Ivoire 37 percent Little change Gabon 30 4 The Gambiab Ghana All 2 Guinea Guinea-Bissau Kenya 24 percent 0 Madagascar All oc Malawi All Few Mali 58 0 Mauritania Hundreds 0 Mozambique 0 Niger Hundreds 9 Nigeria All 19 categories Rwanda Nearly all 0 Senegal Hundreds About 15 Sierra Leone Tanzania Nearly all About 100 Togo 20 2 Uganda - 5 Zambia Nearly all 0 Zimbabwe All Few - Not available. Note: Nontariff barriers include quantitative restrictions and special licensing require- ments other than for health and safety reasons. a. As of early 1993. b. The Gambia has undertaken a trade reform, but precise data are not available. c. Substantial intervention exists through the control of foreign exchange. Source: World Bank staff. 231 ADJUSTMENT IN AFRICA , ENt''1,, (M.: i.grb 1r6 or Z1j iYr5 ']8 J :re5S Domestic purchasing Export sales from producers Producer pricing Before Late Before Late Before Late Country Crop reforms 1992 reforms 1992 reforms 1992 Benin Cotton * * * * * I Burkina Faso Cotton * * * * * A Burundi Coffee * c * O * A Tea * * * * * U Cameroon Coffee * 0 * O * A Cocoa * 0 0 0 U * Central African Rep. Coffee * * 0 C * O Cotton * * 0 0 * U Chad Cotton * * 0 * * I Congo Coffee * * * * * O Cocoa * * * * * O C6te d'lvoire Cocoa 0 0 0 *XU Coffee O 0 0 0 * X Gabon Cocoa * * * * * U Coffee * * * * * The Gambia Groundnuts a C * C * =-V Ghana Cocoa * * * * * A Guinea Coffee * 0 * 0 0 Guinea-Bissau Cashews * 03 0 Kenya Coffee * * * * I1 1 Tea C C * Il Madagascar Vanilla 0 0 O 0 * P Coffee 0 C 0 0 * 0 Malawi Tobacco (smallholdings) * c * uI Tea 0 0 0 0 0 0 Mali Cotton * * * * * I 232 APPENDIX A Domestic ptrchasing Export sales from producers Producer pricing Before Late Before Late Before Late Country Crop reforms 1992 reforms 1992 reforms 1992 Mozambique Cashews * O 0 * * Niger Cowpeas * ( * O * Nigeria Cocoa * 0 0 0 U * Palm oil * 0 0 0 * O Rwanda Coffee * * Tea * * Senegal Groundnuts * * * * Cotton 0 * * * * Sierra Leone Cocoa 0 0 0 O * O Coffee * 0 0 * Tanzania Coffee * 4 0 * [* Cotton * * 0 * * Togo Cotton 0 *0 0 0 a Coffee * * * * * Uganda Coffee * O * O C Cotton 0 * * * * C Zambia Cotton * * * 0 * * Tobacco 0 (03 * ° Zimbabwe Tobacco 0 0 0 0 K K> Cotton 0 0 0 0 K O * Public sector monopoly (including cooperatives and de facto monopolies). * Price set at government's discretion. * Parastatals and private traders in competition. IU Price set but linked to world prices. 0 Exporters/private purchasing agents licensed by government or parastatals. 0 Indicative producer price recommended; export O Private sector competition. price linked to world market prices. Note: Mauritania is excluded because it has no major export crops. E Indicative producer price recommended. Source: World Bank staff. O No prices set. 2.33 ADJUSTMENT IN AFRICA z,b]B A.20 Mionopoly Activities before Reforms Petrokeu,m Fertilizer Exports Production activities 4,~~~~~~~~~~~~~~~~~~~~~~~~~~~Q 4,~~~~~~~~~~~~~~~~~~~~~4 Country < a s, Benin * * * * *1 n.a. O- El * *0 * * Burkina Faso * El 1i O1 * n.a. * n.a. *1 O F] [1 n Burundi F- :1 * * * * * n.a. * 0 0 n.a. O1 Cameroon * * * 0 * * j El * |- i - * n.a. Republic * * * * * * n.a. n.a. * na ..(C Chad C1 O: O j O 0- m na. n.a. n.a. n.a. 1ln.a. Congo * * C6te d'Ivoire 0 1: O EOl O * 0 A n-a * ~ * O o n.a. Gabon * FIG _ _ The Gaienbia * 1: O n.a. 0 n.a. 0 n.a. * n.a. Ghana * O7 * F- O n.a. O- O- O j O O S Guinea * i @ * * * n.a. * * * | * j * * a Guinea-Bissau * * * O n.a. _ _ j_ _ _ >< 0 Kenya * 1:1 * * * j 0 * * -a O .- j n.a. Madagascar * @ * * * * * O- * * * * El CnMalawi I * I O i O Mali 0 O 0 0 e n.a. _ Ol * 01 0 0 Mauritania s 0 El Eli E n 0 n.a. S n.a. n.a. El F0 na Mozambique | n.a. n.a. * 0n E n O I - Niger * ID ' 0 |* * na n.a. .n.a. - * * Nigeria * ' 0 0 * 0 * El - - * n Rwanda O i O 0[* 0 0 n n 0 n a CSenegal E El E 0 a A n.a. A E A na. Sierra L-eone * 0 0 0 0 * E Tanzania 0 El - - * 0 0 0 Togo a 0 [ 0 * * n * * n 0 n | na Uganda 0 [1E 0 ' 0 n E E E E E E - Zambia * 0 0 0 0 n 0 : 0 0 * 0 - Zimbabawe A O * | 0 O E El El A O E El a Publi monopoly. * Private monopoly. l No monopoly. n.a. Notapplicable. Datanotavailable. a. See table A.9 for specific crops for each country. Source: World Bank staff. 234 *il I 0101 IODODOSI *011 El * * | | D @ D n o E * o Wheat * II Sa I0 00 0 -0 u *l O I *ODDE OW u c * * Rice I I0 *I El | * * u * | * I | *W*I *I * * E Sugar DI I01 I I@*0*0* I I I * * I I *LO u n *I DD Otherstaplefoods I IDl * I I I | * * u I |Dl * u * I I DDlI * E Vegetable oil II IOI LIa I u I O |IDI ODDIOD uE * El El g |sI I * * Drugs uDI I DuI uI IO I E I ED Du oI I * u o Ieo u7'extiles * I I I I I *I 10 I IDOL *o * OW I * * u * IOI * * DDS ID * Cernent Elu o * S D n I I I * rI * OOO D ED D-D El D I o u o Banking *@0 0 I * E 0 0 0 00 * - - Telecommunications u] I WEIIu o * I l * El *oI D L *L * u D I o Hiring oflabor UO * 0 I * I | O | * * E |n UW * .0 . WO I * u ' @ Personil insurance (life, automobile, etc.) DO l0 : DO Dl * ODD D oOD 1O DO I OD Import insurance uIWu uII DW WI 0 0 D * D D D Du 0 DO * International shipping * I I 0 0 * I* I I .... * * LI D 0 D* DO D 000 D* Urban btus transportation ~~ a-ri -~~U a is 0-. NO a- "0C C )q > a~~~~~~2 ; A3fRln ;3 19~~~~~~~~~~~~~~~~~~~~ N a- a- w~~~~~~~~~~C Cii i ADJUSTMENT IN AFRICA 17a, e a. Monopoly lActivities, Late 1992 Petroleum Fertilizer Exports Production activities Country t ; }00 Benin * * n.a. * * Burkina Faso * O~ O F g* :n. 0: ::*0 ; n.a. 00 C 0 I:0030: El ri :0[ di:a W* Burundi C 1 1 0 C1 . C1 O * CH n.a. * ; * * n.a. E~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- ..na Cameroon * F F- O !C1 0 * [3I Cl _ 0* . n.a. Republic * 1, * * * * n.a. n.a. * ° - n.a. O Chad F- | [1 1 F[ 1 . ] n.a. n.a. n.a. * F- * n.a. n.a. Congo * * _ _ * * * * * * * * n.a. G6te d'Ivoire *0 El Cl 0 0 *A* n.a. * 0 * ]1 n.a. Gabon * [: _ _ * * * The Gambia O FC O na 1na n.a. ,2 Ghana * - F- F- 0 n.a. F- F[: O- 0: O- r- O W Guinea F-I o1 O _ n.a. _ Guinea-Bissau * * 7 F n.a. _ ,_ ' _ * - Kenya * |: O o] El O- |- O9 El | 0 * 1 * tf0 Madagascar * 0 * * 0 * C:I * Cl A 0 * * * Malawi * F- F- Cl El 1: r- O ] A El 7 A 0 Mali * F- Fi F * n.a. - - 0 F1 * 1 g E C Mauritania A |] F2 [ - .1.a. n.a. 2 | n.a. n.a. O- O 7 E Mozambique * F- El F'i F n.a. F-1 E] E FI C1 - 0 Niger * 1 * 2 n.a. * n.a. n.a. - * - Nigeria * F-I * * 11 F-I F-1 FI O-1 | O 0 a1 Rwanda 7 z -I a a [. *| Senegal A F- F [: * * A n.a. A * A A Sierra Leone O El O ] [2 * ] O F ° _ _ _ - - - O Tanzania * 1- * F] * * O- IF * El 0 F 6* Togo C1 12 Cl * * _ _ _|- - * Uganda El O O-I F-I a O O * O O O :a O O O Zambia 0 * *_ 0 na _ _ 0 _ Zimbabur e un* E El El * l n A | O nA El Gaeoo ~l El ElI0 E l -n rPublic monopolv. * PrivaL monopoly. El No monopolv. n.a. NoL applicabEl. - Data not available. a. See table A.9 for specific crolps for each country. Source: World Bank staff. 236 APPENDIX A I-Ports Services J It El El E _ : El D F- El iO ElIBenin EEl l El l El El El -E El10 El * * n.a. ElBurundi El El'A El1El - EI F El - Central African Republic El El 0 ElEl:El>- El El 0 0 El E El Chad * A 0 A S 0 El 0 El'S~~~~~~~~~~~~~~F1 0 lCongo A - I El *~~~~ ~~~ ~~~ ~~~ ~~~~~~~~El El El El El C&,e d'Ivoire El ElEl El.ElI El El> El A, F El l -1 El El The Gambia * El El El ElI El DEl El1 0 El - El S El Ghana - _ I - El 0 * - ---0 ~~~~~~~~~~~~~Guinea El - - - El El El~~~~~~~~~~~~~F- 0 0 - Guinea-Bissau * El El El A El> -1 El El *tEl El El El ElKey * E 0 El S0 El El El ElI 0 ElI S 0 El El Madagascar El E El n.a El El El A El1 El El El El El El Matawi El El El El~ ~~ El El El El E El ElI El ElI n.a El Mali El El 0 El El El ElEl ElI * E * @0 El O Mauritania El1 El El El El 0 El El El 0 El n.a n.a. El El Mozambique * 0 - - 0 0 E~~~~~~~ ~~~~~~~~~~~~~~LI El - - Niger * n.a El n.a. -- E E El El El El El Nigeria - 1- - - -~~~~~~~~~~~~~~Rwanda * S A El El El'El~~~~~~~F- El El El'El El El El Senegal - El ~~~~~- - - - - - El - Dr~~~~~~~~F- - lSierra Leone El El El El El El~~~~F- El n.a 0 0 El 0 ElF- El El Tanzania El El El El El '0IA - 0 El - --Togo El El El E El El D]El El 4 OrE ELI El El ELUganda - -- i- El 0K Ir@.4ti-X' !~~~~~~~~~~~~~~~~~~~~~~~~~t Tahle A.1i(- (-h.-qnpp in *he Rea! eis rffefan;,c Rach_ p,*lirrD,_ dnnJ ....................... nPar 'a, .....a "Gro-UVV,, ilJelw-eeii 1970JL-0 anlu 19o9i -................ Diffierence in Diffierence in> Change in the annual GDP C.hange in the annualgeD average REER per capita growth average REER per capita growtlh Country (percnt)a (percentage points) Country (percent)a (percentage points) N Large depreciation of the REER A ppreciation of(the RE19 Nigeria 404.4 7.0 Togo -4.7 i.4 Ghana 283.7 3.7 Buirkina Faso -5.7 -1.7 Tanzania 255.4 2.9 Congo i-6.7 -4.9 Madagascar 99.2 1.6 Mali -9.2 -1.6 Mozambique 86.9 7.6 Benin -12.2 -3.1 Zambia 65.0 0 9 Senegal f:-14.0 -0.6 BLuruwidi 56.7 -0.9 - Gabon -19.2 0.9 Zirnbabwe 49.9 0.7 Gote d'lvoire -20.8 -2.6 Uganda 46.5 4.3 Cameroon -27.1 -12.5 Kenya 42.7 1.5 Mean -13.3 -2.7 Mean 139.1 2.9 Median -12.2 -1.7 Median 76.0 2.3 Medtians Moderate depreciationi oj lhe REER All countries 8.4 0.7 Sierra Leone 34.6 2.9 Low-incotme countries 31.7 1.4 Mauiritania 31.7 -0.1 Middle-income countries -16.6 -1.6 The Gambia 16.7 -0.8 Countries with fixed Malawi 8.4 2.2 exchange rates -8.0 -1.8 Rwanda 8.1 -5.5 Countries with flexible Chad 5.6 -I .9 exchange rates 49.9 1.6 Niger 4.3 2.5 Oil-exporting countries -13.0 -2.0 Central African Republic 1.4 -2.6 ~~~~~Note; Guinea anid Guitnca-Bissati are excluded because of itnsufficient data. Mean 13.9 -o.4 a. An increase in the real effective exchange rate constitutes a depreciation; a Median 8.3 -0.5 decrease constitultes an appreciation. .Source: World Bank data. Table A.17 Average Annual GDP per Capita Growth for Countries Classified by Macroeconomic and Market Intervention Policies (Percenl) MARKET INTERVENTION, LATE 1992 Limited Medium Heavy All countries 1981-86 1987-91 1981-86 1987-91 1981-86 1987-91 1981-86 1987-91 Ganibia 1.2 0.3 Burundi 2.1 1.2 Burkina Faso 2.2 0.4 Median -1.2 0.4 Nigeria -4.6 2.4 Gabon -2.7 -1.9 Madagascar -3.7 -2.1 t Uganda -1.5 2.8 Ghana -2.4 1.3 Median -0.8 -0.8 o Median -1.5 2.4 Kenya -0.5 0.9 114 Malawi -1.4 0.7 Mali 0.4 -1.2 on C Mauritania 4.9 -1.0 Senegal 0.4 -0.2 L4, Togo -2.8 -1.4 Median -0.9 -0.2 Sierra l.eone -2.1 0.8 Benin 1.1 -2.0 Central African Median -0.9 -2.1 Median -2.1 0.8 Cameroon 4.6 -7.9 Republic -0.1 -2.8 C6te d'Ivoire -4.2 -6.8 Congo 4.1 -0.7 Mozambique -5.9 1.7 Median 2.0 -1.8 Niger -4.9 -2.4 Rwanda 0.4 -5.0 Tanzania -1.7 1.3 O 4 Zambia -3.2 -2.3 Zimbabwe 0.3 1.0 Median -1.7 -2.3 i X Chad 4.5 2.6 X Guinea-Bissau 2.9 1.5 > Guinea - - E Median 3.7 2.1 a Median -0.2 1.9 Median -1.2 -1.1 Median 1.0 -1.4 Median -0.7 0.1 c; ~~~~~~~~~~~~~~~~~~~~~~~~~~~> -Not available. Soter-es:Tables 5.1 and A.13. z > [77j ,>Si,--m.s-6M w ADJUSTMENT IN AFRICA Table A.18 Changes in Producer Prices of Export Crops and in Agricultural Growth Change in the Difference in average real producer price annual agricultural of export crops, growth rate between 1981-83 to 1989-91 1981-86 and 1987-91 Country (percent) (percentage points) Improvement in producer prices Ghana 96.5 2.2 Nigeria 46.5 2.7 Burkina Faso 30.6 -2.3 Benin 21.7 -0.1 Mozambique 16.2 5.2 Togo 15.8 -1.8 Tanzania 8.3 1.9 Mali 5.8 2.5 Madagascar 5.3 0.9 Niger 2.6 Mean 24.9 1.3 Median 16.0 1.9 Deterioration in producer prices Central African Republic -1.8 -2.9 Zimbabwe -4.8 -4.7 Malawi -8.8 2.3 Senegal -11.7 -2.2 Chad -12.9 6.9 Burundi -18.1 -1.6 Rwanda -23.3 -0.1 The Gambia -25.0 -10.0 Kenya -26.1 -0.3 Gabon -29.4 1.5 Congo -31.3 1.9 Uganda -36.8 6.4 Zambia -42.7 -1.6 Cameroon -44.3 -4.7 C6te d'Ivoire -49.6 8.2 Guinea-Bissau -52.1 -5.0 Sierra Leone -62.0 1.4 Mean -28.3 -0.3 Median -26.1 -0.3 - Not available. Note: Guinea is excluded because of insufficient data; Mauritania is excluded because it has no major export crops. Source: World Bank data. 244 APPENDIX A Table A.1S Changes in Agricultural Taxation and Agricultural Growth Difference in average Change in the real annual agricultural protection coefficient, growvth rate betwveen 1981-83 to 1989-91 1981-86 and 1987-91 Country (percent)a (peercentage points) Large decrease in overall taxation of export crop producers Ghana 341.0 2.2 Guinea 325.8 Madagascar 117.1 0.9 Malawi 78.3 2.3 Uganda 33.9 6.4 Central African Republic 31.5 -2.9 Tanzania 30.6 1.9 Mean 136.9 1.8 Median 78.3 2.0 Small decrease in overall taxation Burkina Faso 17.9 -2.3 Rwanda 15.2 -0.1 Burundi 15.0 -1.6 Togo 10.9 -1.8 Gabon 10.7 1.5 Mali 9.4 2.5 Kenya 8.9 -0.3 Congo 4.5 1.9 Nigeria 1.3 2.7 Niger 1.1 Mean 9.5 0.3 Median 10.0 -0.1 Increase in overall taxation Mozambique -2.0 5.2 Zimbabwe -3.1 -4.7 The Gambia -10.3 -10.0 C6te d'Ivoire -23.2 8.2 Chad -27.4 6.9 Benin -27.6 -0.1 Senegal -28.3 -2.2 Sierra Leone -33.4 1.4 Cameroon -34.7 -4.7 Guinea-Bissau -70.3 -5.0 Zambia -76.0 -1.6 Mean -30.6 -0.6 Median -27.6 -1.6 -Not available. NAote: Mauritania is excluded because it has no major export crops. a. An increase in the real protection coefficient constitutes a decrease in agricultural taxation. Source: World Bank data. 245 z Table A.20 Agricultural Growth H Average Difference Average Difference annual growth rate between 1981-86 annual grozwth rate beten 1981 86 (percent) and 1987-91 (percent) and 1987-91 Country 1981-86 1987-91 (percentagepoints) Country 1981-86 1987-91 (percentagepoints) Large improvement in macroeconomic policie? Deterioration!1 (continued) Ghana -0.2 2.0 2.2 Mozambique -0.9 4.3 5.2 Tanzaniia 3.4 5.3 1.9 Congo 2.3 4.2 1.9 The Gambia 7.4 -2.6b -10.0 C6te d'lvoire -4.0 4.2 8.2 Burkina Faso 5.0 2.7 -2.3 Cameroon: 3.6 -1.1 -4.7 Nigeria 1.2 3.9 2.7 Gabon 1.4 2.9 1.5 Zimbabwe 5.3 0.6 -4.7 Mean 1.9 2.6 0.6 Mean 3.7 2.0 -1.7 Median 2.3 3.3 -0.1 Median 4.2 2.4 -0.2 Ulse Small improvementa Clhad 2.0 8.9 6.9 Madagascar 1.6 2.5 0.9 Guinea - 2.5 : Malawi 1.6 3.9 2.3 Guinea-Bissau 7.7 2.7 -5.0 Burundi 4.1 2.5 -1.6 Medians Kenya 3.4 3.1 -0.3 All countries 3.1 2.8 -0.1 Mali 0.5 3.0 2.5 Mauritania 3.1 0.4b -2.7 Low-income countries 3.1 2.7 -0.1 Senegal 4.1 1.9 -2.2 Middle-income Senigerl 3.4 1.9 -2.2 countries 2.3 2.9 1.5 NUger -.4 4 6 Countries with fixed Mean 2.2 2.7c 0.7 exchange rates 3.2 3.0 -0.1 Median 3.1 2.8 0.3 Countries with flexible exchange rates 2.5 2.5 0.4 Deterioration' Oil-exporting cotntries 1.8 3.4 1.7 Benin 4.7 4.6 -0.1 - Not available. Central African Rep. 3.0 0.1 2.9 a. Classifications are based on the overall scores reported in table B. 1. Rwanda 0.2 0.2 -0.1 b. 1987-90 average annual growth late. Sietra Leone 1.9 3.3 1.4 c. 1984-86 average annual growtlh rate. Togo 5.2 3.4 -1.8 Sources:IMF and World Bank data. Zambia 3.8 2.1 -1.6 Table A.21 Growth in Industry and Manufacturing Industry Manufacturing Average annual growth Difference between Average annual growth Difference between rate (percent) 1981-86 and 1987-91 rate (percent) 1981-86 and 1987-91 Country 1981-86 1987-91 (percentage points) 1981-86 1987-91 (percentage points) Large improvement in macroeconomic policies' Ghana -1.8 6.4 8.2 -0.4 4.5 4.9 Tanzania -9.2 2.8 12.1 -4.5 3.3 7.8 The Gambia 5.2 2.8b --2.3 - - - Burkina Faso 2.5 5.0 2.5 -0.3 6.3 6.6 Nigeria -5.0 5.1 10.1 0.6 - - Zimbabwe -0.2 3.8 3.9 2.7 4.2 1.5 Mean -1.4 4.3 5.7 -0.4 4.6 5.2 Median -1.0 4.4 6.1 -0.3 4.4 5.8 Small improvement' Madagascar -2.8 1.1 3.9 - - - Malawi 1.0 5.9 5.0 3.0 5.4 2.4 Burundi 5.7 5.5 -0.2 6.1 6.6 0.5 Kenya 3.0 5.0 2.0 4.2 5.3 1.2 Mali 11.2 2.1 -9.1 - - - Mauritania 5.2 - Senegal 2.1 4.8 2.8 5.8 5.6 -0.3 Niger -3.0 - - - - - Uganda -3.7 13.7c 17.3 -6.3 17.2c 23.6 Mean 2.1 5.5 3.1 2.6 8.0 5.5 Median 2.1 5.0 2.8 4.2 5.6 1.2 (Table continues on the fllowing page.) z N NI~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Nr IaMe RLZ R. (Co'H,isuzud/) 00 Industry Manufacturing Average annualgrowth Difference between Averageannualgrowth Difference between H rate (percent) 1981-86 and 1987-91 rate (percent) 1981-86 and 1987-91 Country 1981-86 1987-91 (percentage points) 1981-86 1987-91 (percentage points) z Deterioration' Benin 8.3 4.0 -4.4 7.8 5.8 -2.0 Z Central African Republic 0.5 5.2 4.7 > Rwvar,da 2.G -1.5 -4.1 3.2 -0.2 -3.4 Sierra Leone -3.5 4.3 7.8 5.2 6.3 1.1 Togo -2.6 4.7 7.4 -0.4 8.3 8.7 Zambia 0.0 1.7 1.7 2.0 4.7 2.8 Mozambique -6.7 1.1 7.8 Congo 8.3 4.1 -4.1 11.2 3.3 -7.9 C6te d'Ivoire -1.2 -2.2 -1.1 Cameroon 12.5 -4.0 -16.5 17.2 - Gabon 1.8 3.6 1.8 4.8 20.0 15.3 Mean 1.8 1.9 0.1 6.4 6.9 2.1 Median 0.5 3.6 1.7 5.0 5.8 1.1 Unclassif ed Chad 16.0 -0.1 - 1.9 Guinea - 5.0 - 4.6 --- Guinea-Bissau 4.5 5.2 0.7 Medians All countries 1.4 4.1 2.5 3.0 5.5 1.9 Low-incomne countries 1.0 4.8 3.4 2.0 5.6 2.4 Middle-income countries 1.8 3.6 -1.1 8.0 4.2 1.5 Countries with fixed exchange rates 2.3 4.0 0.4 5.3 6.1 3.2 Countries with flexible exchange rates -0.1 4.6 3.9 2.7 5.0 1.9 Oil-exporting countries 5.0 3.9 -1.2 8.0 11.7 3.7 - Nor available. a. Classifications are based on the overall scores reported in table B.1. b. 1987-90 average annual growth rate. c. 1984-86 average annual growth rate. Sources: IMF and World Bank data. Tiame Ans22 Growrh in Vxerisd Average Diffierence Average Difference annualgrowth between 1981-86 annualgrowtb between 1981-86 rate (percent) and 1987-91 rate (percent) and 1987-91 Country 1981-86 1987-91 (percentage points) Country 1981-86 1987-91 (percentage points) Large imnprovement in macroeconomic policies' Deteriorationa (corntinued) Ghana 4.5 8.1 3.5 Togo -1.4 2.7 4. l Tanzania - - - Zambia -2.2 -2.8 -0.5 'T'he Gambia -O.6 11.3 11.9 Mozambique - - - BurkinaFaso 0.0 5.3 5.4 Congo 5.9 2.0 -3.9 Nigeria -5.5 4.9 10.3 C6te d'lvoire 3.9 4.9 1.0 Zimbabwe - - - Cameroon 13.7 -11.8 -25.4 Mean -0.4 7.4 7.8 Gabon 0.4 11.4 11.0 Median -0.3 6.7 7.9 Mean 1.3 -0.1 -1.4 Median 1.3 0.6 -0.7 Small imp rovemen#' Madagascar -7.6 6.1 13.7 Unclassified Malawi 1.6 3.6 2.0 Chad 14.8 5.0 -9.7 Burundi 11.8 4.9 -6.9 Guinea - - - Keniya 2.3 6.3 4.0 Guinea-Bissau - Mali 2.6 6.7 4.2 Medians Mauritania 8.9 -0.8 -9.7 All countries 2.2 3.6 2.0 Senegal 6.1 0.9 -5.2 Low-income countries 0.8 4.2 3.8 Niger -7.1 0.2 7.3 Middle-income countries 5.9 2.0 -3.9 Mganda 2 3 1 Countries with fixed Mean 2.3 3.5 1.2 exchange rates 2.4 2.4 0.1 5Countries with flexible Deterioration' exchange rates 1.6 4.9 3.5 E3enin 2.2 -4.1 -6.3 Oil-exporting countries 3.2 3.4 3.2 Central African - Not available. Republic --3.6 --4.4 -0.8 a. Classifications are based on the overall scores reported in table B.1. Rwvanda 4.5 2.1 -2.4 Sources: IMF and World Bank data. Sierra Leone -10.5 -0.8 9.7 ot < . N .. .. .~~~~~~~~~~ 19 Table A.23 Investment Gross domestic investment Public investment > Difference between Difference between Percentage of GDP 1981-86 and 1987-91 Percentage of GDP 1981-86 and 1987-91 Country 1981-86 1987-91 (percentage points) 1981-86 1987-91 (p ercentage points) Large improvement in macroeconomic policiesh Ghana 6.3 15.1 8.8 1.7 3.2 1.5 Tanzania 18.3 26.9 8.7 6.8 5.3 -1.5 The Gambia 19.0 18.4 -0.6 12.4 10.3 2.1 Burkina Faso 20.0 20.9 0.9 15.7 10.4 -5.3 Zimbabwe 19.6 20.7 1.1 6.0 8.0 2.1 Mean 16.6 20.4 3.8 8.5 7.5 -1.1 Median 19.0 20.7 1.1 6.8 8.0 -1.5 Nigeriac 16.5 15.4 -1.1 9.5 5.6 -3.9 Small improvement: Madagascar 9.1 12.4 3.3 6.6 7.5 1.0 Malawi 17.6 18.7 1.1 8.7 6.7 -2.1 Burtndi 16.4 18.0 1.6 13.7 13.7 0.0 Kenya 23.1 23.7 0.5 7.1 6.0 -1.1 Mali 17.2 21.6 4.4 10.5 11.1 0.6 Mauritania 29.8 16.6 -13.2 2.5 5.7 3.3 Senegal 11.2 13.0 1.7 3.8 2.8 -1.0 Niger 12.9 10.0 -2.9 8.7 8.2 -0.5 Uganda 7.8 11.1 3.3 2.2 5.0 2.8 Mean 16.1 16.1 0.0 7.1 7.4 0.3 Median 16.4 16.6 1.6 7.1 6.7 0.0 Deteriorationb Benin 16.0 12.4 -3.6 11.0 6.2 -4.9 Central African Republic 11.0 12.1 1.0 6.9 12.1 5.2 Rwanda 15.6 14.4 -1.2 8.2 8.1 -0.1 Sierra Leone 13.9 11.6 -2.3 4.3 3.2 -1.1 Togo 25.3 22.7 -2.6 10.7 7.4 -3.3 Zambia 17.2 12.8 -4.4 4.9 5.4 0.5 Mozambique 12.8 34.6 21.8 13.8 22.7 8.9 C6te d'Ivoire 17.9 11.5 -6.4 10.0 3.5 -6.5 Mean 16.2 16.5 0.3 8.7 8.6 -0.2 Median 15.8 12.6 -2.4 9.1 6.8 -0.6 Cameroon' 24.8 17.8 -7.0 10.6 8.6 -2.0 Congoc 39.4 16.1 -23.4 18.7 5.0 -13.7 Gabonc 37.2 29.0 -8.3 16.7 6.0 -10.7 Unclass4ifed Chad 5.9 8.9 3.1 - - - Guinea - - - Guinea-Bissau 27.2 31.1 4.0 - - Medians All countries 17.2 16.3 0.8 8.7 6.4 -1.0 Low-income countries 16.4 16.0 1.1 8.5 7.0 -0.3 Middle-income countries 22.2 16.9 -6.7 10.3 5.5 -4.2 Countries with fixed exchange rates 17.6 14.5 -1.3 10.6 7.4 -3.3 Countries with flexible exchange rates 16.9 17.3 1.1 6.8 6.0 0.0 Oil-exporting countries 31.0 16.9 -7.7 13.7 5.8 -7.3 - Not available. a. Capital expenditures and net lending. b. Classifications arc based on the overall scores reported in table B.l. c. Data for Camcroon, Congo, Gaboni. and Nigetia are llot counted in the group means and medians because of the strong intltuence of oil exports on investment lev- els in these coulntries. Sources: IMF and World Bank data. tl X -4 _~~~~~~~~~~~~~~~~~~~~~- > Table A.24 Savinigs , Gross domestic savinigs Public savings" 0ttt Difference between Differenee between Percentave of GDP 1981-86 and 1987-91 Percentage of GDP 1981-86 and 1987-91 Country 1981-86 1987-91 (percentage points) 1981-86 1987-91 (percentage-points) Large improvement in macroeconomic policies6 i Gliana 5.6 7.5 2.0 -1.7 2.6 4.3 ;i Tanzania 9.7 -2.6 -12.3 -4.2 -1.6 2.6 t)0:00 The Gambia 6.2 8.2 2.1 -1.4 1.0 2.4 Burkina Faso -4.9 1.5 6.4 -I.1 0.5 1.5 Nigeria 14.4 22.8 8.4 3.6 -1.2 -4.9 00 Zimbabwe 17.9 22.3 4.5 -3.7 -2.5 1.3 Mean 8.1 10.0 1.8 1.4 -0.2 -1.2 Median 7.9 7.9 3.3 -4.6 -o.4 2.0 Small improvement6 Madagascar 1.9 5.9 4.0 0.2 1.5 1.3 i Malawi 13.3 9.0 -4.3 -2.7 -0.3 2.4 Burtundi 3.1 1.8 -1.3 i.4 1.5 0.1 Ketiya 20.7 18.8 -1.9 -0.1I -1.1 -0.9 Mali -3.7 5.3 9.0 -2.3 0.8 3.1 Mauritania 4.3 8.1 3.8 --2.8 3.1 5.9 Scnegal -0.4 7 9 8.3 -3.3 -0. 1 3.2 Niger 5.1 4.6 -0.5 0.3 -2.7 -2.9 Uganda - - - -3.9 -0.7 3.1 Mean 5.5 7.7 2.1 -1.5 0.2 1.7 Median 3.7 6.9 1.6 -2.3 -0. 1 2.4 Deteriorationb Beniin 0.7 3.1 2.4 -2.6 -3.5 -0.9 Central Africani Repuiblic -2.0 -0.7 1.4 -0.7 -1.3 -0.6 Rwanda 6.0 4.8 -1.2 1.1 -2.1 -3.2 Sierra Leone 7.4 6.8 -0.6 -8.9 -8.0 0.9 Tlogo 18.8 13.4 -5.4 3.7 0.9 -2.7 Zambia 14.1 13.4 -0.7 -10.3 -7.4 2.8 Mozambique - - - -2.5 -2.7 -0.3 Congo 35.7 23.1 -12.7 11.1 -6.7 -17.8 Cote d'lvoire 21.3 15.4 -6.0 3.2 -9.8 -13.0 Cameroon 29.1 15.6 -13.4 10.6 0.3 -10.2 Gabon 50.1 36.3 -13.9 16.0 -1.0 -17.0 Mean 18.1 13.1 -5.0 1.9 -3.7 -5.6 Median 16.4 13.4 -3.3 1.1 -2.7 -2.7 Unclassified Chad -11.6 -15.7 -4.1 - _ Guinea - - Guinea-Bissau -4.7 -6.2 -1.5 - - _ Medians All countries 6.1 7.7 -0.7 -1.2 -1.0 0.5 Low-income countries 5.3 5.6 -0.6 -1.6 -0.9 1.1 Middle-inconie countries 25.2 19.0 -9.3 6.9 -1.7 -11.6 Countries with fixed excliange rates 2.9 6.6 -2.3 0.3 -1.0 -2.7 Countries with flexible exchange rates 6.8 7.8 -0.7 -2.5 -1.1 1.3 Oil-exporting cotntries 32.4 22.9 -13.1 10.8 -1.1 -13.6 - Not available. a. Current revenitue (exclutding grants) minus current expenidittre. b. Classifications are based on the overall scores reported in table B. 1. Sources: IMF and World Bank data. ti~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- z N~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~> ADJUSTMENT IN AFRICA Table A.25 Sources of Income of Poor, Rural Smaliholders in Selected Countries (percent) Agricultural income' Foods for Off-farm home Foods earned Nonearned Total Country consumption sold Total incomeb incomer income Burkina Faso, 1984-85 Sahelian zone - - 49 20 31 100 Sudanian zone - - 59 25 16 100 Guinean zone - - 56 38 6 100 C6te d'Ivoire, 1985-86 Forest 31 45 76 21 3 100 Savannah 40 41 81 17 2 100 The Gambia, 1991 Various regions 22 35 57 22 21 100 Ghana, 1988 Forest 37 20 57 40 3 100 Savannah 54 14 68 31 1 100 Kenya, 1985-87 South Nyanza Province 40 14 54 42 4 100 Madagascar, 1984 Coast 25 17 42 55 3 100 Plateau 31 8 39 58 3 100 South 37 11 48 49 3 100 Malawi, 1988 South 37 14 51 13 36 100 Rwanda, 1985-86 Northwest 33 12 45 38 17 100 Tanzania, 1976 50 23 73 25 2 100 - Not available. Note: Data may not add up to totals because of rounding. a. Includes income from livestock. b. Includes wages, salaries, and earnings from self-emplovment. c. Includes income from trnnsfers, remittances, and other nonearned sources. Source: Dorosh and Sahn ( 993), table 1. 254 APPENDIX A Table A.26 Sources of Agricultural Income of Poor, Rural Smaliholders in Selected Countries (pewenrt) Tradedfoodse Nontradedfoodsb Total For home For home Export agricultural Country consumption Sold Total consumption Sold Total crops' income C6te d'lvoire, 1985-86 Forest 8 6 14 32 9 41 45 100 Savannah 18 14 32 31 14 46 22 100 The Gambia, 1991 Various regions 19 44 63 20 17 37 0 100 Ghana, 1988 Forest 9 9 18 57 13 70 12 100 Savannah 16 10 26 63 10 73 1 100 Kenya, 1985-87 South Nyanza Province - - 35 - - 45 20 100 Madagascar, 1984 Coast 23 23 35 11 46 31 100 Plateau 28 2 30 51 18 69 1 100 South 33 3 36 44 15 58 6 100 Malawi, 1988 South 52 1 53 20 4 24 23 100 Tanzania, 1976 27 9 35 42 18 61 4 100 Neglible (less than 1 percent). - Not available. iVote: Data may not add up to totals because of rounding. a. Such as rice, maize, and groundnuts. b. Such as millet, cassava, sweet potatoes, and yams. c. Such as cocoa, tobacco, cotton, coffee, cola nuts, rubber, and sugar. Source: Dorosh and Sahn (1993), table 2. 255 z z? Tradedfoods Ground- Nontradedfoods Totalfood Nonfood Total Country Rice Maize nuts Other Total Millet Cassava Other Total expenditures expenditures expenditures C6te d'lvoire, 1985-86 Forest 6 5 1 2 15 0 4 46 50 65 35 100 Savannah 11 10 5 3 28 3 3 37 42 70 30 100 The Gambia, 1991 Various regions 15 2 16 34 - - - 33 67 33 100 Ghana, 1988 Forest 2 6 1 1 10 0 12 51 63 73 27 100 Savannah 8 16 1 1 26 16 5 33 54 80 20 100 Kenya, 1985-87 South Nyan7a Province 1 31 _ 1 33 2 45 2 49 82 18 100 Madagascar, 1984 Coast 13 .. .. 5 19 .. (-40-) 40 59 41 100 Plateau 16 1 0 16 .. (-49-) 49 65 35 100 Sourh 13 .. 3 16 .. (-46-) 46 62 38 100 Malawi, 1988 South 0 33 2 0 35 1 1 25 26 61 39 100 Tanzania, 1976 5 17 1 0 23 4 2 41 48 71 29 100 Negligible (between 0 and 0.5 percent). - Not available. Note: Data may niot add up to totals because of rounding. Source: Dorosh and Sahn (1993), table 4. APPENDIX A $l A RUAocalk103n (311 SIte lEdur<2a,i337 Bu!dgell 0-] S@u:Ms !C t - Percentage of budget Distribution of allocated budget (percent) allocated to specific Primary Secondary Higher Country and year Type of expenditure education levela education education education Benin, 1988 Both recurrent 79.0 54.4 27.8 17.7 and development Burkina Faso, 1990 Recurrent 94.0 45.7 24.5 29.8 Cameroon, 1992 Recurrent 100.0 42.1 25.0 32.9 Central African Both 100.0 54.0 25.0 21.0 Republic, 1990 Chad, 1988 Recurrent 75.2 3.2 68.1 28.7 Cote d'Ivoire, 1990 Recurrent 100.0 52.0 32.5 15.5 Development 100.0 27.4 52.8 19.8 Both 100.0 51.8 32.7 16.5 The Gambia, 1989 Both 84.0 48.8 39.3 11.9 Ghana, 1988 Both 100.0 85.9 ( 14.1 ) Guinea, 1990 Recurrent 92.8 36.0 32.3 31.7 Guinea-Bissau, 1989 Recurrent 71.2 76.3 21.3 2.4 Kenya, 1990 Recurrent 94.0 57.1 24.7 18.3 Development 83.5 0.0 27.4 72.6 Both 92.6 49.8 25.1 25.2 Madagascar, 1990 Both 100.0 ( 72.0 ) 28.0 Malawi, 1988 Recurrent 81.5 58.0 16.3 25.6 Development 87.4 4.1 21.5 74.4 Both 82.8 45.7 17.5 36.8 Mali, 1988 Recurrent 92.0 38.0 41.4 20.7 Mozambique, 1990 Recurrent 100.0 37.2 42.7 20.1 Development 100.0 8.2 59.9 31.9 Nigeria, 1984 Both 100.0 ( 81.2 ) 18.8 Senegal, 1991 Recurrent 94.0 48.9 25.5 25.5 Tanzania, 1986 Recurrent 91.2 62.1 24.2 13.7 Development 76.1 15.9 63.7 20.4 Both 87.9 52.7 32.7 14.7 Togo, 1990 Recurrent 97.0 38.9 33.6 27.5 Uganda, 1989 Recurrent 89.1 29.9 51.0 19.1 Development 93.0 41.6 14.6 43.8 Both 90.0 32.9 42.0 25.2 Zambia, 1986 Recurrent 90.1 45.0 30.9 24.1 Zimbabwe, 1990 Both 95.8 55.3 30.4 14.3 Median Recurrent 93.4 47.3 29.6 20.4 Development 90.2 12.1 40.1 37.9 Both 94.2 52.3 30.4 18.8 a. Excludes nonallocable items such as general administrative expenditures. Source: Sahn and Bernier (1993), table 1. 257 ADJUSTMENT AIF RI CA Table A.29 Change in Net External Transfers as a Share of GDP, 1981-86 to 1987-91 (percentage points) Countries with improvement Countries with deterioration in macroeconomic policies in macroeconomic policies Countries with increases Tanzania 19.7 Mozarmbique 42.0 in net external transfers Malawi 8.4 Cameroon 2.9 Madagascar 4.7 Central African Republic 2.8 Kenya 4.3 Gabon 2.7 Uganda 4.2 Sierra Leone 2.6 Burundi 3.9 Benin 2.4 Ghana 2.7 Rwanda 1.7 Mean 6.9 Zambia 0.4 Median 4.3 Mean 7.2 Median 2.6 Countries with decreases Burkina Faso -1.9 Togo -0.8 in net external trasnfer Zimbabwe -2.5 Core d'lvoire -1.3 Nigeria -3.4 Congo -9.2 The Gambia -5.5 Mean -3.8 Niger -0.7 Median -1.3 Senegal -4.2 Mali -6.2 Mauritania -12.1 Mean -4.6 Median -3.8 Note: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. Source: World Bank data. 258 Appendix B The Indexes of Macroeconomic Policy Change and Stance HIS APPENDIX HAS TWO OBJECTIVES. ONE IS TO explain how the index of change in macroeconomic policies and the index of macroeconomic policy stance were constructed. The second objective is to explore the robustness of our findings about the rela- tion between policies and GDP per capita growth. Change in Macroeconomic Policies and Change in Growth T HE INDEX OF CHANGE IN MACROECONOMIC POLICIES MEA- sures the changes in three policy indicators: fiscal, monetary, and exchange rate policies between 1981-86 and 1987-91. Numerical scores from -3 to +3 were assigned to each country based on the size of the change in each indicator, with a higher score indicat- ing more improvement in policy (table B.1). * For fiscal policy, the index scores were based on the change in the budget deficit excluding grants, which provides a more accurate measure of the domestic fiscal effort than does the deficit includ- ing grants. Changes in domestic tax revenue were also taken into account, because they are a proxy for the quality of the fiscal ad- justment. the index score was increased (or decreased) by one point if revenue as a share of GDP rose (or fell) by more than 3 per- centage points. 259 ADJUSTMENT IN AFRICA Fiscalpolicy Monetary policy Change in overall fiscal balance Che Ch aCnge Change dChange excluidinggrants in total revenue h in seignioraged in inflation _____ _____ ____ ____ _____ ____ in fiscal Percentage Percentage policy Percentage Percentage Country points Scorea points Scoreb (score)' points Score' points Scoref Benin 4.0 2 -3.6 0 2.0 3.6 -2 -2.7 1 Burkina Faso 6.8 3 0.4 0 3.0 -0.7 1 -5.2 1 Burundi 0.1 0 2.0 0 0.0 - - 0.2 0 Cameroon -8.3 -2 -6.2 -1 -3.0 -0.5 0 -7.1 1 Central African Republic -5.9 -2 -1.9 0 -2.0 -1. 1 1 -11.6 2 Congo -4.2 -1 -12.8 -1 -2.0 0.2 0 -7.9 1 C6te d'Ivoire -6.5 -2 -6.0 -1 -3.0 -1.7 1 -3.8 1 Gabon -6.3 -2 -10.7 -1 -3.0 -0.2 0 -7.9 1 The Gambia 4.5 2 2.6 0 2.0 -0.4 0 -7.9 1 Ghana 2.8 1 5.7 1 2.0 -1.0 1 -25.7 2 Kenya 0.2 0 -0.5 0 0.0 -0.3 0 -1.9 0 Madagascar 0.3 0 -0.7 0 0.0 0.7 0 -6.7 1 Malawi 4.4 2 0.0 0 2.0 0.7 0 5.9 -1 Mali 2.5 1 2.3 0 1.0 -2.5 2 -3.8 1 Mauritania 2.6 1 2.0 0 1.0 -0.2 0 -3.0 1 Mozambique -9.2 -2 -0.5 0 -2.0 - - 47.1 -3 Niger -2.4 -1 -1.0 0 -1.0 -0.9 1 -10.0 1 Nigeria -1.0 0 3.7 1 1.0 1.4 -1 9.9 -1 Rwanda -3.1 -1 0.7 0 -1.0 0.0 0 1.1 0 Senegal 4.1 2 -0.6 0 2.0 -1.6 1 -12.4 2 Sierra Leone 2.0 1 -0.5 0 1.0 0.2 0 40.7 -3 Tanzania 4.1 2 1.0 0 2.0 2.4 -2 -4.8 1 Togo 0.6 0 -6.2 -1 -1.0 -2.9 2 -6.5 1 Uganda 0.3 0 -2.6 0 0.0 - - 27.6 -2 Zambia 2.4 1 -3.8 0 1.0 1.1 -1 60.3 -3 Zimbabwe -0.8 0 5.8 1 1.0 1.3 -1 -0.1 0 - Not available. n.a. Not applicable. Xote: Chad, Guinea, and Guiinea-Bissau are excluded because of insufficient data. a. A score of-2 reflects a change in the fiscal balance of-9.9 to -5.0 percentage points; -1, a change of-4.9 to -2.0 percentage points; 0, a change of-1.9 to 0.9 percentage points; 1, a change of 1.0 to 2.9 percentage points; 2, a change of 3.0 to 4.9 percentage points; and 3, a change of 5.0 percentage points or more. b. A score of-I reflects a change in total revenue of-4.0 percentage points or more; 0, a change of-3.9 to 3.0 percentage points; 1, a change of 3.1 percentage points or more. c. Calculated by adding the scores for change in overall fiscal balance and change in revenue. d. Seigniorage is based on MI; however, for Tanzania in 1985-86 and 1990, it is based on M2. e. A score of 2 reflects a change in seigniorage of-3.0 to -2.1 percentage points;i, a cfhange6of-2.0 to-0.6 percentage points; 0, a change of-0.5 to 0.9 percentage points; -1, a change of 1.0 to 1.9 percentage points; and 2, a change of2.0 to 3.9 percentage points. f. A score of 2 reflects a change in inflation of -49.0 to -10.0 percentage points; 1, a change of-9.9 to -2.5 percentage points; 0, a change of-2.4 to 4.9 percentage points; -l, a change of 5.0 to 9.9 percentage points; -2, a change of I O.0 to 30.9 perceotage points; and -3, a change of 31.0 percentage points or more. 260 APPENDIX B Exchange rate policy Change in the Change in the parallel Change in Overall Chanige in real effective market exchange exchange change in monetary exchange rate1 rate premiumm rate macroeconomic policy Percentage Percentage policy policies (score)9 points S&orei points Score'I (score)k score)I Country -0.5 --12.2 -2 n.a. n.a. -2.0 -0.2 Benin 1.0 -5.7 -1 n.a. n.a. -1.0 1.0 Burkina Faso 0.0 56.7 3 -4.9 0 1.5 0.5 Burundi 0.5 --27.1 -2 n.a. n.a. -2.0 -1.5 Cameroon Central African 1.5 1.4 0 n.a. n.a. 0.0 -0.2 Republic 0.5 -6.7 -1 n.a. n.a. -1.0 -0.8 Congo 1.0 -20.8 -2 n.a. n.a. -2.0 -1.3 C6te d'lvoire 0.5 -19.2 -2 n.a. n.a. -2.0 -1.5 Gabon 0.5 16.7 2 3.1 0 1.0 1.2 The Gambia 1.5 283.7 3 -1,080.6 3 3.0 2.2 Ghana 0.0 42.7 3 -4.6 0 1.5 0.5 Kenya 0.5 99.2 3 -28.0 1 2.0 0.8 Madagascar -0.5 8.4 1 -27.1 1 1.0 0.8 Malawi 1.5 -9.2 -1 n.a. n.a. -1.0 0.5 Mali 0.5 31.7 3 52.3 -3 0.0 0.5 Mauritania -3.0 86.9 3 -2,039.5 3 3.0 -0.7 Mozambique 1.0 4.3 1 n.a. n.a. 1.0 0.3 Niger -1.0 404.4 3 -197.4 3 3.0 1.0 Nigeria 0.0 8.1 1 -7.0 0 0.5 -0.2 Rwanda 1.5 -14.0 -2 n.a. n.a. -2.0 0.5 Senegal -1.5 34.6 3 398.1 -3 0.0 -0.2 Sierra Leone -0.5 255.4 3 -164.1 3 3.0 1.5 Tanzania 1.5 -4.7 -1 n.a. n.a. -1.0 -0.2 Togo -2.0 46.5 3 -41.3 2 2.5 0.2 Uganda -2.0 65.0 3 239.8 -3 0.0 -0.3 Zambia -0.5 49.9 3 -37.2 2 2.5 1.0 Zimbabwe g. Calculated by averaging the scores for change in seigniorage and change in inflation. h. An increase in the real effective exchange rate constitutes a depreciation; a decrease constitutes an appreciation. i. A score of-2 reflects a change in the real effective exchange rate of-10.0 percentage points or more; -1, a change of -9.9 to -5.0 per- centage points; 0, a change of-4.9 to 2.0 percentage points; 1, a change of 2.1 to 14.9 percentage points; 2, a change of 15.0 to 30.9 per- centage points; and 3, a change of 31 .0 percentage points or more. j. A score of 3 reflects a change in the premium of -1 00 percentage points or more; 2, a change of -99 to -30 percentage points; l, a change of-29 to -10 percentage points; 0, a change of-9 to 4 percentage points; -1, a change of 5 to 15 percentage points; -2. a change of 16 to 50 percentage points; and -3, a change of 51 percentage points or more. k. Calculated by averaging the scores for change in the real effective exchange rate and change in the parallel market exchange rate premium. 1. Calculated by averaging the scores for change in fiscal policy, change in monetars policy, and change in exchange rate policy. A score of 1.0 or more reflects large improvement; 0 to 0.9, small improvement; and below 0, deterioration. Soerces: World Bank data; IMF data; staff estimates. 26I AiFJUSTMENT IN AFRICA * For monetary policy, the index scores were based on the average of changes in seigniorage and inflation.' Changes in real interest rates were not factored in because they are very similar to changes in inflation. * For exchange rate policy, the index scores for the fixed exchange rate countries were based on the change in the real effective ex- change rate (PEER); for the flexible exchange rate countries, the scores were based on a simple average of the change in the REER and the change in the parallel market exchange rate premium. The individual scores for each of the three indicators were then ag- gregated by simple averaging to arrive at a composite score for overall change in macroeconomic policies. Based on their composite scores, the adjusting countries of Sub-Saharan Africa were divided into three groups: countries that had large improvement in macroeconomic poli- cies (scores above or equal to 1), small improvement (scores below 1 but above 0), or deterioration (scores below 0). Before deciding upon this methodology, we experimented with other ways of scoring countries. For example, we used weighted averages that gave more importance to the REER and inflation, and we widened the range of scores from -9 to +9 so as to give more weight to large distor- tions. While country rankings were somewhat altered as a result, the overall picture did not change significantly. We are thus confident that the scoring procedures themselves were robust. To check the robustness of the results obtained using this index, we explored two other ways of constructing an index of overall change in macroeconomic policies. Substituting other policy indicators. The first approach was to use alter- native indicators of change in macroeconomic policies. For fiscal policy, changes in three different indicators were used: the fiscal deficit including grants, the primary deficit including grants, and the primary deficit excluding grants. For monetary policy, an indica- tor was developed based only on the rate of inflation. For exchange rate policy, we continued to use the change in the REER for the fixed exchange rate countries, and, for the flexible exchange rate coun- tries, the change in the REER together with the change in the parallel market premium. These indicators were then combined in various ways to create eight variants on the measure of overall policy change (table B.2). z6z APPENDIX B Table 6.2 Alternative Approaches for Calculating Overall Change in Malacroeconomnic Policies Indicators used Approach Cbange infiscal policy Chanige in monetary policy Change in exchange rate policy Approach used in Change in the overall fiscal Average of the changes in Change in the real effective the study balance excluding grants, plus seigniorage and infla[ion exchange rate (for countries change in total revenue with fixed exchange rates) or the average of the changes in the parallel market premium and the real effective exchange rate (for countries with flexible exchange rates) Alternative I Change in the overall fiscal See above See above balance including grants Alternative 2 Change in the overall fiscal See above See above balance excluding grants Alternative 3 Change in the primary fiscal See above See above balance including grants Alternative 4 Change in the primary fiscal See above See above balance excluding grants Alternative 5 Change in the overall fiscal Change in inflation only See above balance including grants Alternative 6 Change in the overall fiscal See above See above balance excluding grants Alternative 7 Change in the primary fiscal See above See above balance including grants Alternative 8 Change in the primary fiscal See above See above balance excluding grants Comparing different time periods. The second approach was to retain the original indicators of policy change, but to calculate the change over alternative time periods: first, between 1981-86 and 1990-91, and second, between 1985-86 and 1990-91. The choice of time period depends on whether one is interested in period averages or endpoints in time. We had opted to compare the policies during 1981-86 with 263 AbDJUSTMENT IN AERICA those during 1987-91 because we were interested in the average change from the pre-reform period to the adjustment period. But for countries that have only recently begun to make reforms, the change in policies from 1981-86 to 1990-91 may be more meaningful. And if one wants to track progress starting from a time when many countries had very poor policies, it might be more appropriate to look at the change between 1985-86 and 1990-91. Robust outcomes. Although choosing different policy indicators or time periods does shift some countries from one group to the next, the correlation between policy reform and growth remains robust. The evi- dence indicates that the group of countries with the largest improve- ment in macroeconomic policies consistently had the largest increase in economic performance, while the countries whose macroeconomic policies deteriorated generally fared the worst (tables B.3 and B.4). The major exception to this pattern is in agricultural growth, which exhibits no clear-cut link to policy change in any of the three time periods used. Also, when we assess the change in policies between 1981-86 and 1990-91, the group with deterioration in policies did better in GDP, export, and manufacturing growth than the group whose policies improved slightly. The country groupings themselves exhibit a certain amount of ro- bustness. Regardless of the policy indicators used, six of the countries in the study (Burkina Faso, The Gambia, Ghana, Nigeria, Tanzania, and Zimbabwe) consistently showed a large improvement in overall macro- economic policies, while four countries (Mali, Niger, Senegal, and Uganda) were always among the group with small improvement. Five countries (Cameroon, Congo, C6te d'Ivoire, Gabon, and Togo) consis- tently appeared among the countries with deterioration in policies. Of the remaining ten countries with data available, four (Kenya, Madagas- car, Malawi, and Mauritania) switched between the large- and small-im- provement groups. Five countries (Benin, the Central African Republic, Rwanda, Sierra Leone, and Zambia) shifted between the group with small improvement and the group with deterioration. Only Mozam- bique jumped between the top and bottom groups, showing a large im- provement in policies if grants were included in the fiscal indicator and a deterioration in policies if they were not. Country groupings exhibited a similar robustness across time periods, with the same core groups of countries consistently registering large improvement or deterioration in overall policies. 264 APPENDIX B 'fat0 a. MA W0 iE my n,!z S3_ 22,Pa .z u '.sl 'l~ zi,7JDLz~ C - :~ ; I 'I '' ""I ,L _ :.1 -I,,-,Cj.LJ (. Median change between 1981-86 and 1987-91 (percentage points) GDP Agri- Manu- Gross Gross Approach used to calculate overall per capita cultural Industrial facturing Export domestic domestic change in macroeconomicpolicie? growth growth growth growth growth investmentb savings Approach used in the study Countries with large improvement 1.8 -0.2 6.1 5.8 7.9 1.0 3.3 Countries with small improvement 1.5 0.3 2.8 1.2 3.0 1.6 1.6 Countries with deterioration -2.6 -0.1 1.7 1.1 -0.7 -3.6 -3.3 Alternative 1 Countries wvith large improvement 1.9 1.4 4.5 4.9 7.9 1.1 3.1 Countries with small improvement 0.4 -1.6 2.4 1.2 -0.7 1.3 0.4 Countries with deterioration -2.8 0.7 -2.6 -0.4 -0.7 -5.0 -5.7 Alternative 2 Countries with large improvement 1.8 -0.2 6.1 5.8 7.9 1.0 3.3 Countries with small improvement 1.5 0.3 2.8 1.2 3.0 1.6 1.6 Countries with deterioration -2.6 -0.1 1.7 1.1 -0.7 -3.6 -3.3 Alternative 3 Countries with large improvement 1.5 0.3 3.9 3.2 5.4 1.4 2.1 Countries with small improvement 0.4 -0.1 3.8 1.7 -0.5 -0.1 -0.5 Countries with deterioration -2.8 0.7 -2.6 3.3 -1.5 -6.7 -9.3 Alternative 4 Countries with large improvement 1.5 0.3 3.9 4.9 7.9 1.0 3.1 Countries with small improvement 1.0 1.4 3.9 1.1 2.0 1.4 -0.5 Countries with deterioration -2.6 -0.1 0.3 0.4 -0.8 -4.0 -5.4 Alternative 5 Countries with large improvement 1.6 0.9 4.5 4.9 5.4 1.1 3.8 Countries with small improvement -0.6 -0.2 2.4 0.5 -0.8 1.6 1.4 Countries with deterioration -0.8 0.7 0.3 1.9 0.2 -5.4 -5.7 Alternative 6 Countries with large improvement 1.6 0.9 3.9 5.8 10.3 1.1 4.0 Countries with small improvement -0.4 -0.3 2.4 0.8 -0.8 1.1 1.4 Countries with deterioration 0.9 1.4 1.7 1.9 0.2 -4.4 -5.7 Alternative 7 Countries with large improvement 1.5 -0.3 3.9 3.2 4.7 1.1 2.9 Countries with small improvement -0.6 0.7 3.8 0.4 0.6 1.0 0.4 Countries with deterioration -0.8 -0.1 0.3 5.7 0.2 -6.7 -9.3 Alternative 8 Countries with large improvement 1.5 -0.3 3.9 4.9 5.4 0.9 3.8 Countries with small improvement -0.6 0.7 3.8 0.8 0.6 1.1 0.4 Countries with deterioration -0.8 0.7 0.3 2.8 -0.5 -5.4 -6.0 a. See table B.2 for a description of the policv indicators used under each alternative approach. b. Calculations include the oil-exporting countries (Cameroon, Congo. Gabon, and Nigeria). in contrast to table A.23. Source: World Bank data. 265 ADJUSTMENT IN AFRICA %abla 3.4 Economic On¢tcDmes Using Alternative Time Periods for Calculating Zierall Change in rylacroeconornic Policies Median change (percentage points) Time period used to GDP Agri- Manu- Gross Gross calculate overall change per capita cultural Industrial facturing Export domestic domestic in macroeco nomic policies growth growth growth growth growth investmenta savings Period used in the study (1.981-86 to 1987-91) Countries with large improvement 1.8 -0.2 6.1 5.8 7.9 1.0 3.3 Countries with small improvement 1.5 0.3 2.8 1.2 3.0 1.6 1.6 Countries with deterioration -2.6 -0.1 1.7 1.1 -0.7 -3.6 -3.3 Alteriative 1 (1981-86 to 1I990-91) Countries with large improvement 2.9 2.2 8.2 6.4 10.3 3.3 2.0 Countrieswithsmallimprovement -0.8 -1.9 2.0 0.8 -3.9 1.0 4.1 Countries with deterioration 0.9 -0.1 1.8 1.9 0.2 -2.9 -0.9 Alternative 2 (1985-86 to 199091) Countries with large improvement 2.9 2.2 8.2 6.4 3.5 1.1 1.7 Countries with small improvement 0.9 0.3 2.6 1.5 5.4 0.9 1.8 Countries with deterioration -2.8 -0.9 -4.1 -2.0 -5.1 -5.0 -3.3 a. Calculations include the oil-exporting countries (Carneroon, Congo, Gabon, and Nigeria), in contrast to table A.23. Source: World Bank data. Policy Stance and GDP per Capita Growth T tHIS REPORT ALSO USES AN INDEX OF POLICY STANCE TO assess the relation between macroeconomic policy stance and growth. The index, which reflects the state of macroeconomic policies in adjusting countries as of 1990-91, uses the following policy isindicators: * Fiscal policy stance is based on the budget deficit including grants, because this provides a good measure of the current fiscal imbal- ance. In using this indicator, we implicitly assume that grants will continue in the short term. * Monetary policy stance is based on seigniorage, inflation, and the real interest rate.2 * The exchange rate policy stance is based on the change in the REER between 1980 and 1990-91 for the countries with fixed exchange rates; for the countries with flexible exchange rates, the stance is based on the parallel market exchange rate premium. For each policy, each country was classified as having a good/ade- quate, fair, poor, or very poor stance and assigned a numerical score Z66 APPENDIX= Bf from 1 to 4, with a smaller number indicating a better policy stance (table B.5). The individual scores were then aggregated by simple aver- aging to arrive at a composite score for overall macroeconomic policy stance. The cutoff points were unavoidably arbitrary, since there is no solid analytical basis for differentiating sharply between fair and poor policy stances, or between adequate and fair. For example, there is prob- ably little difference in the macroeconomic policy environment in Bu- rundi and Ghana: Burundi had a larger parallel market exchange rate premium and a slightly bigger budget deficit in 1990-91, but lower in- flation. And yet Burundi fell into the fair category, while Ghana was classified as adequate. The labels-adequate, fair, poor, and very poor- provide a useful basis for ranking countries according to generally ac- cepted standards, but not too much weight should be attached to the precise rankings. To test the robustness of the index of macroeconomic policy stance, we recomputed the index using the primary deficit, rather than the over- all deficit, as an indicator of fiscal policy stance. Again, there was some movement of countries from one group to the next. The main change was that Congo and Zimbabwe shifted from the groups rated poor or very poor to those with fair or adequate stance. With respect to the mean and median growth rates for each group, however, there was little noticeable change, indicating that the choice of indicator for fiscal pol- icy stance does not make much difference for our results. Notes 1. Although inflation is an outcome, rather than an in- 2. Interest rates are included in the stance index be- dicator, of monetary policy, it is useful to include it as an cause they generally do not follow the same pattern as the indicator in the index, because it reflects the effectiveness rates of inflation. (Interest rates were not taken into ac- of monetary policy. The effectiveness of monetary policies count in the index of overall change in macroeconomic in controlling inflation is as important as the monetary policies because changes in the interest rate are closely re- policies themselves. lated to changes in the rate of inflation.) 267 ADJUSTMENT IN AFRICA Tab3e B.5 Components af Macroeconomic Policy Stance, 1990-91 Fiscal policy Overall fiscal balance including Monetarypolicy grants Fiscal (percentage policy Seigniorage Inflation Country of GDP) (score)a Percent Score Percent Score Benin -6.3 3 2.7 3 0.1 1 Burkina Faso -3.4 2 -0.1 1 1.9 1 Burundi -3.3 2 - - 8.0 1 Cameroon -8.6 4 0.0 1 0.5 1 Central African Republic -6.5 3 -0.6 1 0.3 1 Congo -7.7 4 0.8 2 -0.3 1 C6te d'Ivoire -13.0 4 0.4 1 0.6 1 Gabon -1.7 2 0.2 1 5.2 1 The Gambia 2.7 1 1.8 3 10.4 1 Ghana 0.8 1 0.4 1 27.6 3 Kenya -5.6 3 1.8 3 13.3 2 Madagascar -5.1 3 1.5 2 6.5 1 Malawi -2.5 2 1.0 2 12.2 2 Mali -5.3 3 -0.6 1 3.3 1 Mauritania -0.9 1 1.1 2 6.6 1 Mozambique -8.9 4 - - 35.9 3 Niger -7.2 4 -0.7 1 -4.3 1 Nigeria -4.5 3 2.9 3 10.2 1 Rwanda -7.0 3 0.8 2 11.9 2 Senegal -1.1 1 -0.8 1 -0.7 1 Sierra Leone -7.7 4 6.2 4 106.8 4 Tanzania -0.9 1 7.6 4 21.0 2 Togo -3.3 2 1.4 2 1.0 1 Uganda -4.1 3 - - 32.2 3 Zambia -8.5 4 4.0 4 101.8 4 Zimbabwe -8.3 4 2.5 3 20.8 2 - Not available. n.a. Not applicable. Note: Chad, Guinea, and Guinea-Bissau are excluded because of insufficient data. a. A score of 1 is considered good or adequate; 2, fair; 3, poor; and 4, very poor. b. A score of 1.0 to 1.3 is considered adequate; 1.4 to 2.3, fair; 2.4 to 3.0, poor; and 3.1 and above, very poor. c. An increase in the real effective exchange rate constitutes a depreciation; a decrease constitutes an appreciation. Sources: IMF data; World Bank data; staff estimates. z68 APPENDIX B Exchange rate policy Change Parallel in the real Overall market effective macro- Monetary exchange rate exchange rate Exchange economic Real interest rate policy premium since 1980 ratepolicy policies Percent Score (score)b (percent) (percent): (score)a (score)' Country 6.6 2 2.0 n.a. 10.8 3 2.7 Benin 2.6 1 1.0 n.a. 10.3 3 2.0 Burkina Faso - - 1.0 20.9 n.a. 2 1.7 Burundi 8.7 3 1.7 n.a. -18.0 4 3.2 Cameroon Central African 6.8 2 1.3 n.a. 9.1 3 2.4 Republic 8.7 3 2.0 n.a. -9.2 4 3.3 Congo 5.4 2 1.3 n.a. 2.8 4 3.1 C6te d'Ivoire 6.7 2 1.3 n.a. 7.8 3 2.1 Gabon 3.2 2 2.0 21.3 n.a. 2 1.7 The Gambia 2.8 1 1.7 3.4 n.a. 1 1.2 Ghana -1.0 1 2.0 7.3 n.a. 1 2.0 Kenya - - 1.5 7.1 n.a. 1 1.8 Madagascar -0.3 1 1.7 29.4 n.a. 2 1.9 Malawi 2.8 1 1.0 n.a. 10.8 3 2.3 Mali - - 1.5 166.6 n.a. 4 2.2 Mauritania - - 3.0 62.6 n.a. 4 3.7 Mozambique 16.0 3 1.7 n.a. 28.0 2 2.6 Niger 1.7 1 1.7 25.1 n.a. 2 2.2 Nigeria -9.9 2 2.0 47.5 n.a. 3 2.7 Rwvanda 8.9 3 1.7 n.a. -4.0 4 2.2 Senegal -30.7 3 3.7 104.4 n.a. 4 3.9 Sierra Leone - - 3.0 74.5 n.a. 4 2.7 Tanzania 5.9 2 1.7 n.a. 9.7 3 2.2 Togo -2.9 1 2.0 24.6 n.a. 2 2.3 Uganda - - 4.0 149.7 n.a. 4 4.0 Zambia -12.5 2 2.3 23.5 n.a. 2 2.8 Zimbabwe 269 Appendix C Agricultural Policy Indicators D ECOMPOSING THE REAL PRODUCER PRICE INTO its policy and exogenous components shows how policy and world price trends interact in determin- ing real producer prices for agricultural exports. The equation is as follows: RPP- e~p PI -NPC .RER .PB, CPI PB e CPI WPI where RPP is the real producer price for export crops, PF is the farmgate producer price, PB is the border price in dollars, e is the nominal ex- change rate, cPi is the consumer price index for the country in question, and wPi is the U.S. wholesale price index. Let us look more closely at the three terms of the decomposition. The first term, P/P,,e, is the nominal protection coefficient (NPC). An NPC of less than 1 means that producers are being taxed rather than pro- tected.1 An increase in the NPC means that the producer's share of the border price is increasing, and thus that explicit taxation is decreasing. All other things being equal, this will raise the RPP. The second term in the decomposition, e(WPI/CPI), is the bilateral real exchange rate (RER). If the real exchange rate depreciates and the depreciation is passed back to the producer, the RPP will increase. The third term in the decomposi- tion, PB/WPI, is the price of the country's exports (at the border) de- flated by the wPI, which we call the real border price, p. A fall in the real border price will result in a decline in the RPP, unless the fall is offset by a sufficiently large reduction in explicit taxation or by a depreciation of the real exchange rate. 271 ADJUSTMENT IN AFRICA Because real prices for export producers are influenced by world price trends as well as domestic variables, shifts in the real price for export pro- ducers do not tell us whether overall taxation of agricultural producers has increased or decreased. For this we need a measure combining ex- plicit taxation (from policies that affect the producer price) and implicit taxation (through overvaluation of the real exchange rate).2 The real protection coefficient (RPc) measures overall taxation. It is identical to the nominal protection coefficient (NPC), except that it calculates the border price at the equilibrium exchange rate instead of at the official ex- change rate. It is defined as follows: c F= _ F .e e RP = - - = NPC , - PB E PB e E E where E is the equilibrium exchange rate. The first term, PF /PBe, is the nominal protection coefficient, which shows how much producers are being taxed explicitly by governments. The second term, elE, is the ratio of the nominal exchange rate to the equilibrium exchange rate, which shows how much producers are being taxed implicitly by overvalued ex- change rates. The change in the Rmc is a better proxy for assessing the change in taxation than the change in the product of the NPC and the RER (the first two terms in the decomposition of the real producer price). This is il- lustrated by the following example. Suppose the real exchange rate de- preciates by a small amount at the same time that the equilibrium ex- change rate moves increasingly out of line (because of a large terms-of-trade shock, for example). Producers would be implicitly taxed more by the more overvalued exchange rate (indicated by a decrease in the RPc), while at the same time the real producer price would improve because of the depreciation in the real exchange rate. Hence, using the change in the product of the NPC and the RER does not show how much the level of taxation has changed, but rather how much of the change in the real exchange rate is transferred back to producers, taking into ac- count changes in direct taxation. In practice, the Rpc is not easy to compute because the equilibrium exchange rate is not an observable variable. It can only be estimated. We used the parallel market premium as a rough proxy for the real exchange rate misalignment in the flexible exchange rate countries. As a very rough proxy for the misalignment in the countries with fixed exchange 272 APPENDIX C rates, we compared the change in each country's real effective exchange rate between 1980 and 1990-91 with the extent of real exchange rate depreciation in the reference group of countries over the same period (see chapter 2). 'This gives a crude indication of how much additional depreciation was needed in 1990-91 to restore competitiveness. Notes 1. Generally, the producer price is calculated at the and other interventions in these services. It would be im- border price equivalent-that is, transportation and mar- portant to take account of differences in these costs in keting costs are added to the producer price. Our esti- comparing NPCs across countries. mates of the NPC do not include transportation and mar- keting costs because of a lack of data. Also, we did not 2. Ideally, a method similar to that of Kreuger, Schiff, want to add in marketing and distribution costs, because and Valdes (1988) would be used to calculate total taxa- they may reflect substantial inefficiencies due to parastatal tion-something beyond the scope of this study. 273 References Akiyama, Takamasa, and Donald F. Larson. 1989. "Recent Trends and Prospects for Agricultural Commodity Exports in Sub-Saharan Africa." PPR Working Paper 348. World Bank, International Economics Department, Washington, D.C. . 1993. "Adding Up Problem-Strategies for Primary Commodity Ex- ports in Sub-Saharan Africa." PPR Working Paper 1245. World Bank, In- ternational Economics Department, Washington, D.C. Alderman, Harold, Sudharshan Canagarajah, and Stephen D. Younger. 1993. Consequences of Permanent Layofffrom the Civil Service: Results from a Survey of Retrenched Workers in Ghana. Cornell Food and Nutrition Policy Pro- gram Working Paper 35. Ithaca, N.Y.: Cornell University. Barrett, Scott. 1990. "Macroeconomic Policy Reforms and Third World Soil Conservation." Draft. London Business School, Department of Economics. Barro, R. J. 1991. "Economic Growth in a Cross-Section of Countries." Quar- terlyJournal of Economics 106(2):407-43. Bates, Robert H., and Paul Collier. 1993. "The Politics and Economics of Pol- icy Reform in Zambia." In Robert H. Bates and Anne 0. Krueger, eds., Po- litical and Economic Interactions in Economic Policy Reform. Oxford, U.K., and Cambridge, Mass.: Basil Blackwell. Berg, Elliot. 1993. "Privatization in Sub-Saharan Africa: Results, Prospects, and New Approaches." Consultant's report prepared for World Bank, Pol- icy Research Department, Transition and Macro-Adjustment Division, and for Financial Sector Development Department, African Economies in Transition Project, Washington, D.C. Binswanger, Hans. 1989. "The Policy Response of Agriculture." In World Bank, Proceedings of the World Bank Annual Conference on Development Eco- nomics 1989, pp. 231-58. Washington, D.C. Birdsall, Nancy, and Richard Sabot. 1993. "Virtuous Circles: Human Capital, Growth and Equity in East Asia." Background paper for The East Asian Miracle. World Bank, Policy Research Department, Washington. D.C. 275 ADJUSTMENT IN AF RICA Blejer, Mario I., and Adrienne Cheasty, eds. 1993. How to Measure the Fiscal Deficit: Analytical and Methodological Issues. Washington, D.C.: Interna- tional Monetary Fund. Bond, Marian E. 1983. "Agriculture Responses to Prices in Sub-Saharan African Countries." International Monetary Fund Staff Papers 20(4): 703-36. Chamley, Christophe, and Patrick Honohan. 1992. "Taxation of Financial Intermediation: Measurement Principles and Application to Five African Countries." PRE Working Paper 421. World Bank, Policy Research De- partment, Washington, D.C. Cleaver, Kevin M. 1993. A Strategy to Develop Aggriculture in Sub-Saharan Africa and a Focus for the World Bank. Technical Paper 203. Washington, D.C.: World Bank. Cleaver, Kevin, and Gotz Schreiber. 1993. "The Population, Agriculture and Environment Nexus in Sub-Saharan Africa." Africa Technical Department working paper. World Bank, Washington, D.C. Collins, Susan M., and Won-Am Park. 1989. Book II, "External Debt and Macroeconomic Performance in South Korea." In Jeffrey D. Sachs and Susan M. Collins, eds., Developing Country Debt and Economic Perfor- mance. Vol. 3, Country Studies-Indonesia, Korea, Philippines, Turkey. Chicago: University of Chicago Press. Conservation International. 1989. The Debt for Nature Exchange. Washing- ton, D.C. Corbo, Vittorio, and Klaus Schmidt-Hebbel. 1991. "Public Policies and Sav- ings in Developing Countries." Journal of Development Economics 36(1):89-115. Cromwell, C., and J. T. Winpenny. 1991. Has Economic Reform Harmed the Environment?A Review of StructuralAdjustment in Malawi. London: Over- seas Development Institute. Delgado, Christopher L. 1992. "Why Domestic Food Prices Matter to Growth Strategy in Semi-Open West African Agriculture." Journal of African Economies 1(3):446-71. Demery, Lionel. Forthcoming. "C6te d'Ivoire: Fettered Adjustment." In Ishrat Husain and Rashid Faruqee, eds., Adjustment in Africa: Lessons from Country Case Studies. A Regional and Sectoral Study. Washington, D.C.: World Bank. Demery, Lionel, and Ishrat Husain. 1993. "Assessment of the Interrelation between Adjustment Financing and Economic Performance in SPA Coun- tries." Draft. World Bank, Africa Region, Washington, D.C. 276 REFERENCES Denizer, Cevdet, and Alan Gelb. 1992. "Mongolia: Privatization and System Transformation in an Isolated Economy." Policy Research Working Paper 1063. World Bank, Policy Research Department, Washington, D.C. Devarajan, Shantayanan, and Jaime de Melo. 199 1. "Membership in the CFA Zone: Odyssean Journey or Trojan Horse?" In Ajay Chhibber and Stanley Fischer, eds., Economic Reform in Sub-Saharan Africa. A World Bank Sym- posium. Washington, D.C.: World Bank. Devarajan, Shantayanan, Jeffrey D. Lewis, and Sherman Robinson. 1993. "Ex- ternal Shocks, Purchasing Power Parity, and the Equilibrium Real Ex- change Rate." WorldBank Economic Review 7(1):45-63. Dia, Mamadou. 1993. A Governance Approach to Civil Service Reform in Sub- Saharan Africa. Technical Paper 225. Washington, D.C.: World Bank. Dollar, David. 1992. "Outward-Oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976-1985." Economic De- velopment and Cultural Change 40(April):523-44. Dornbusch, Rudiger, and Stanley Fischer. 1993. "Moderate Inflation." World Bank Economic Review 7(1):1-44. Dorosh, Paul A., and David E. Sahn. 1993. A General Equilibrium Analysis of the Efect ofMacroeconomic Adjustment on Poverty in Africa. Cornell Food and Nu- trition Policv Program Working Paper 39. Ithaca, N.Y.: Cornell University. Dutz, Mark, and Claudio Frischtak. 1993. "Industrial Adjustment in Tanza- nia: An Enterprise-Level Analysis." Working paper. World Bank, Private Sector Development Department, Washington, D.C. Easterly, William. 1992. "Projection of Growth Rates." Outreach 5. World Bank, Policy Research Department, Washington, D.C. Easterly, William, and Ross Levine. 1993. "Is Africa Different? Evidence from Growth Regressions." Draft. World Bank, Policy Research Department, Washington, D.C. Easterly, William, Michael Kremer, Lant Pritchett, and Lawrence H. Sum- mers. 1993. "Good Policy or Good Luck: Country Growth Performance and Temporary Shocks." Journal of Monetary Economics 32(3). Easterly, William, and Klaus Schmidt-Hebbel. 1991. "The Macroeconomics of Public Sector Deficits: A Synthesis." PPE Working Paper 775. World Bank, Policy Research Department, Washington, D.C. Elbadawi, Ibrahim A. 1992. "Have World Bank-Supported Adjustment Pro- grams Improved Economic Performance in Sub-Saharan Africa?" Policy Re- search Working Paper 1001. World Bank, Policy Research Department, Washington, D.C. 277 ATEN AFRICA Elbadawi, Ibrahim, and Nader Majd. 1992. "Fixed Parity of the Exchange Rate and Economic Performance in the CFA Zone: A Comparative Study." Pol- icy Research Working Paper 830. World Bank, Policy Research Depart- ment, Washington, D.C. Fischer, Stanley. 1982. "Seigniorage and the Case for a National Money." Jour- nal of Political Economy 90(April) :295-313. Fosu, Augustin Kwasi. 1992. "Political Instability and Economic Growth: Evi- dence from Sub-Saharan Africa." Economic Development and Cultural Change 40(July):829-41. Fry, Maxwell J. 1988. Money, Interest, and Banking in Economic Development. Baltimore, Md.: Johns Hopkins University Press. Gelb, Alan. 1988. Oil Windfalls: Blessing or Curse? New York: Oxford Univer- sity Press. Giovannini, Alberto. 1983. "The Interest Elasticity of Savings in Developing Countries: The Existing Evidence." World Development 1 1(7):601-07. 1985. "Saving and the Real Interest Rate in LDCs." Journal of Develop- ment Economics 18(2-3):197-217. Giovannini, Alberto, and Martha de Melo. 1993. "Government Revenue from Financial Repression." American Economic Review 83(4):953-63. Glickman, Joan, and Darius Teter. 1991. "Debt, Structural Adjustment, and Deforestation: Examining the Links-A Policy Exercise for the Sierra Club." Consultants' report. Sierra Club, Washington, D.C. Godfrey, Martin. 1985. "Trade and Exchange Rate Policy: A Further Contri- bution to the Debate." In Tore Rose, ed., Crisis and Recovery in Sub-Saharan Africa. Paris: Development Centre of the Organization for Economic Co- operation and Development. Grootaert, Christiaan. 1993. "The Evolution of Welfare and Poverty under Structural Change and Economic Recession in C6te d'Ivoire, 1985-88." Policy Research Working Paper 1078. World Bank, Policy Research De- partment, Washington, D.C. Grut, Mikael, John A. Gray, and Nicolas Egli. 1991. Forest Pricing and Conces- sion Policies: Managing the High Forests of West and Central Africa. Techni- cal Paper 143. Washington, D.C.: World Bank. Gupta, Kanhava L. 1987. "Aggregate Savings, Financial Intermediation, and Interest Rate." Review of Economics and Statistics 69(2):303-1 1. Haggard, Stephan, and Steven B. Webb. 1993. "What Do We Know about the Political Economy of Economic Policy Reform?" World Bank Research Ob- server 8(2): 143-68. 278 REFERENCES Hardin, Garrett. 1968. "The Tragedy of the Commons." Science 162: 1243-48. Helleiner, G. K. 1992. "The IMF, the World Bank and Africa's Adjustment and External Debt Problems: An Unofficial View." World Development 20 (June):779-92. Hodd, Michael. 1991. The Economies of Africa: Geography, Population, His- tory, Stability, Structure, Pe?formance Forecasts. Aldershot, U.K.: Dart- mouth Publishing. Husain, Ishrat. 1993. "Trade, Aid and Investment in Sub-Saharan Africa." Paper presented at Royal African Society Conference on Africa '93: Gover- nance, Business Aid, in Oxford, U.K., March 21-23. Husain, Ishrat, and Rashid Faruqee, eds. Forthcoming. Adjustment in Africa: Lessonsfrom Country Case Studies. A Regional and Sectoral Study. Wash- ington, D.C.: World Bank. IMF (International Monetary Fund). Various issues. International Financial Statistics. Washington, D.C. International Currency Analysis, Inc. Various years. World Currency Yearbook. Brooklyn, N.Y. Jaeger, William K. 1992. The Effects of Economic Policies on African Agricul- ture. Discussion Paper 147. Washington, D.C.: World Bank. Kessides, Christine. 1993. Institutional Optionsfor the Provision ofInfrastruc- ture. Discussion Paper 212. Washington, D.C.: World Bank. Kiguel, Miguel, and Stephen A. O'Connell. 1993. "Parallel Exchange Rates in Developing Countries: Lessons from Eight Case Studies." Draft. World Bank, Policy Research Department, Transition and Macro-Adjustment Division, Washington, D.C. Kikeri, Sunita, John Nellis, and Mary Shirley. 1992. Privatization: The Lessons ofExperience. Washington, D.C.: World Bank. Killick, Tony. 1992. ExplainingAfrica's Post-Independence Development Expe- riences. Working Paper 60. London: Overseas Development Institute. King, Robert G., and Ross Levine. 1992. "Financial Indicators and Growth in a Cross Section of Countries." PRE Working Paper 819. World Bank, Pol- icy Research l)epartment, Washington, D.C. Krueger, Anne C. 1978. Liberalization Attempts and Consequences. Published for the National Bureau of Economic Research. Cambridge, Mass.: Ballinger. Krueger, Anne O., Maurice Schiff, and Alberto Valdes. 1988. "Agricultural Incentives in Developing Countries: Measuring the Effect of Sectoral and Economywide Policies." World Bank Economic Review 2(3) :255-71. 279 ADJUS,TMENT IN AFRICA Lall, Sanjaya. 1992. "Structural Problems of African Industry." In Frances Stewart, Sanjaya Lall, and Samuel Wangwe, eds., Alternative Development Strategies in Sub-Saharan Afiica. London: Macmillan. Lambert, Sylvie, Hartmut Schneider, and Akiko Suwa. 1991. "Adjustment and Equity in Cote d'Ivoire: 1980-86." World Development 19(11): 1563-76. Lopez, Ramon. 1992. "The Environment as a Factor of Production: The Eco- nomic Growth and Trade Policy Linkages." In Patrick Low, ed., Interna- tional Trade and the Environment. Washington, D.C.: World Bank. Maddison, Angus. 1989. The World Economy in the 20th Century. Paris: De- velopment Centre of the Organization for Economic Cooperation and Development. Mayfield, Malcolm R. 1992. "The Effects of Policy Reforms on the Perfor- mance of Agriculture in Tanzania." Consultant's report. World Bank, Eastern Africa Department, Agriculture and Environment Division, Washington, D.C. McKinnon, Ronald. 1973. Money and Capital in Economic Development. Washington, D.C.: Brookings Institution. Meade, J. E. 1967. "Population Explosion, the Standard of Living and Social Conflict." EconomicJournal77(June):233-55. Mills, Bradford, David Sahn, E. E. Walden, and Stephen Younger. Forth- coming. Public Finance and Public Employment: An Analysis of the Public Sector Retrenchment Programs in Guinea and Ghana. Cornell Food and Nu- trition Policy Program Working Paper. Ithaca, N.Y.: Cornell University. Mosley, Paul, Jane Harrigan, and J. F. J. Toye. 1991. Aid and Power: The World Bank and Policy-Based Lending. London and New York: Routledge. Narasimhan, Bhanu, and Lant Pritchett. 1993. "The Evolution of Import Re- strictions in Sub-Saharan Africa in the 1980's: An Empirical Analysis." Draft. World Bank, Washington, D.C. Nashashibi, Karim, and Stefania Bazzoni. 1994. "Alternative Exchange Rate Strategies and Fiscal Performance in Sub-Saharan Africa." International Monetary Fund Staff Papers 41 (1). Nellis, John R. 1989. Contract Plans and Public Enterprise Performance. Dis- cussion Paper 48. Washington, D.C.: World Bank. OED (Operations Evaluation Department). 1993. "Adjustment in Sub-Saha- ran Africa: Selected Findings from OED Evaluations." Report 12155. World Bank, Washington, D.C. Pack, Howard. 1993. "Productivity and Industrial Development in Sub- Saharan Africa." WorldDevelopment 21(1): 1-16. 280 REFE RE NCES Perrings, Charles, Hans Opschoor, Jaap Arntzen, A. Gilbert, and David W. Pearce. 1988. Economics for Sustainable Development-Botswana: A Case Study. Gaborone: Ministry of Finance and Development Planning. Pimentel, David, Beth Floyd, Wayne Teel, and Julie Bourns. 1991. "Defor- estation, Biomass Depletion, and Land Degradation: Linkages to Policy Reform in Sub-Saharan Africa." Paper presented at a seminar. Cornell University, Department of Natural Resources, Ithaca, N.Y. Pritchett, Lant. 1991. "Measuring Outward Orientation in Developing Countries: Can It Be Done?" PRE Working Paper 566. World Bank, Policy Research Department, Trade Policy Division, Washington, D.C. Reed, David, ed. 1992. Structural Adjuistment and the Environment. Boulder, Colo.: Wesrview. Repetto, Robert. 1989. "Economic Incentives for Sustainable Production." In Gunter Schramm and Jeremy J. Warford, eds., Environmental Management and Economic Development. Baltimore, Md.: Johns Hopkins University Press. Rutihinda, G. 1992. "Tanzania's Experience." In I. G. Patel, ed., Policies for African Development: From the 1980s to the 1990s. Washington, D.C.: In- ternational Monetary Fund. Sahn, David E. 1992. "Public Expenditures in Sub-Saharan Africa during a Period of Economic Reforms." World Development 20(5):673-93. Sahn, David E., and Rene Bernier. 1993. Evidencefrom Africa on the Intrasec- toral Allocation of Social Sector Expenditures. Cornell Food and Nutrition Policy Program Working Paper 45. Ithaca, N.Y.: Cornell University. Sarris, Alexander H., and Rogier van den Brink. 1993. Economic Policy and HIousehold Welfare during Crisis and Adjustment in Tanzania. New York: New York University Press. Schadler, Susan, Franek Rozwadowski, Siddarth Tiwari, and David Robin- son. 1993. EconomicAdjustment in Low-Income Countries: Experience under the Enhanced Structuiral Adjustment Facility. IMF Occasional Paper 106. Washington, D.C.: International Monetary Fund. Schiff, Maurice, and Alberto Valdes. 1992. The Plundering of Agriculture in Developing Countries. Washington, D.C.: World Bank. Schloss, Miguel. 1993. "Does Petroleum Procurement and Trade Matter? The Case of S ub-Saharan Africa. " Finance and Development 30 (1):44-46. Schwartz, Gerd, and Paulo Silva Lopes. 1993. "Privatization: Expectations, Trade-Offs, and Results." Finance aizd Development 30(2):14-17. Seckler, David, ed. 1993. Agricultural Transformation in Africa. Arlington, Va.: Winrock International Institute for Agricultural Development. 281 Selowski, Marcelo, and Herman G. van der Tak. 1986. "The Debt Problem and Growth." World Development 14(9): 1107-1124. Serven, Luis, and Andres Solimano. 1992. "Private Investment and Macro- economic Adjustment: A Survey." World Bank Research Observer 7(1): 95-114. Shaikh, Hafeez, Sunita Kikeri, and Daniel Swanson. 1993. "Privatization and Public Enterprise Reform in Africa." Background paper prepared for Ad- justment in Africa: Reforms, Results, and the Road Ahead. World Bank, Washington, D.C. Sherif, Khaled Fouad. 1993. "Public Enterprise Reform and Privatization in Africa." Working paper. World Bank, Africa Technical Department, Washington, D.C. Stein, Howard. 1992. "Deindustrialization, Adjustment, the World Bank and the ILIF in Africa." World Development 20(Jan.):83-95. Stewart, Frances, Sanjaya Lall, and Sarnuel Wangwe, eds. 1992. Alternative Development Strategies in Sub-Saharan Africa. London: Macmillan. Stryker, J. Dirck, Robert L. West, Jeffery C. Merzel, B. Lvnn Salinger, Peter M. Haymond, Alison T. Slack, and Bruce Aylward. 1989. "Linkages be- tween Policy Reform and Natural Resource Management in Sub-Saharan Africa." Draft. Fletcher School, Tufts University, Somerville, Mass., and Associates for International Resources and Development, Cambridge, Mass. Summers, Lawrence H., and Lant H. Pritchett. 1993. "The Structural Ad- justment Debate." American Economic Review, Papers and Proceedings 83(May):383-89. Tan, Jee-Peng, and Alain Mingat. 1992. Education in Asia: A Comparative Study of Cost and Financing. Washington, D.C.: World Bank. Tanzi, Vito. 1978. "Inflation, Real Tax Revenue, and the Case for Inflation- ary Finance: Theory with an Application to Argentina." International Mon- etary Fund StaffPapers 25(3):417-5 1. - 1991. "Fiscal Policy, Growth, and Design of Stabilization Programs." In Ana Maria Martirena-Mantel, ed., External Debt, Savings, and Growth in Latin America, pp. 121-41. Washington, D.C.: International Monetary Fund. Tenconi, Roland. 1992. "The African Banking Systems in the 1980s and Early 1990s." Draft. World Bank, Occidental and Central Africa Depart- ment, Industry and Energy Operations Division, Washington, D.C. Torfs, Marijke. 1991. "Effects of the IMF Structural Adjustment Programs on Social Sectors of Third World Countries." Discussion paper. Friends of the Earth, Washington, D.C. Z82 REFERENCES UNCTC (United Nations Centre on Transnational Corporations). 1991. Ac- countancy Development in Africa: Challenge of the 1990s. New York: United Nations. Valdes, Alberto. 1989. "Comment on 'The Policy Response of Agriculture' by Binswanger." In World Bank, Proceedings ofthe World Bank Annual Confer- ence on Development Economics 1989, pp. 263-68. Washington, D.C. van den Brink, Rogier. 1993. "A Review of Agricultural Statistics of Mainland Tanzania." Revised version of background paper for Agricultural Sector Memorandum for Tanzania. World Bank, Eastern Africa Department, Agriculture and Environment Division, Washington, D.C. von Braun, Joachim, and Eileen Kennedy. 1987. "Cash Crops versus Subsis- tence Crops: Income and Nutritional Effects in Developing Countries." In J. Price Gittinger, Joanne Leslie, and Caroline Hoisington, eds., Food Policy: Integrating Supply, Distribution, and Consumption. Baltimore, Md.: Johns Hopkins University Press. WCED (World Commission on Environment and Development). 1987. Our Common Future. Oxford: Oxford University Press. World Bank. 1989a. Sub-Saharan Africa: From Crisis to Sustainable Growth. Washington, D.C. 1 1989b. World Development Report 1989: Financial Systems and Develop- ment. New Yotk: Oxford University Press. .1990. World Development Report 1990: Poverty. New York: Oxford Uni- versity Press. - 1991 a. The African Capacity Building Initiative: Toward Improved Policy Analysis and Development Management. Washington, D.C. 199 lb. Reform ofPublic Sector Management: Lessonsfrom Experience. Pol- icy and Research Series Paper 18. Washington, D.C. - 1991 c. World Development Report 1991: The Challenge of Development. New York: Oxford University Press. - 1992a. Adjustment Lending and Mobilization of Private and Public Re- sources for Growth. Policy and Research Series Paper 22. Washington, D.C. - 1992b. Governance and Development. Washington, D.C. - 1992c. World Debt Tables 1992-93: External Finance for Developing Countries. Washington, D.C. - 1992d. World Development Report 1992: Development and the Environ- ment. New York: Oxford University Press. - 1993a. Thc East Asian Miracle: Economic Growth and Public Policy. New York: Oxford University Press. 283 ADJUSTMENT IN Al-RICA . 1993b. "Ghana 2000 and Beyond-Setting the Stage for Accelerated Growth and Poverty Reduction." Economic Report 11486. Washington, D.C. - 1993c. Implementing the World Bank's Strategy to Reduce Poverty: Progress and Challenges. Washington, D.C. . 1993d. World Development Report 1993: Investing in Health. New York: Oxford University Press. Young, Crawford. 1982. Ideology and Development in Africa. New Haven, Conn.: Yale University Press. Younger, Stephen D. 1992. "Aid and the Dutch Disease: Macroeconomic Management When Everybody Loves You." World Development 20(11): 1587-97. 284 T HE W OR L D B AN K TO REVERSE THE ECONOMIC DECLINE THAT BEGAN IN THE 1970S, MANY SUB-SAHARAN AFRICAN countries have undertaken structural adjustment programs. These programs are designed to pave the way for long-term development and prosperity by fundamentally restructuring African economies. Continent-wide growth of gross domestic product (GDP) per capita remains low, how- ever, leading many to question the effectiveness of adjustment efforts. But is this poor performance the result of a failure to reform policies or a failure of those policies to restore growth? To address this question, Adjustment in Africa examines the extent of policy reforms and their impact on growth and poverty in twenty-nine Sub-Saharan countries that were undergoing adjustment in the second half of the 1980s. How much have policies actually changed? This book marshals a wealth of data to assess progress in improving the macroeconomic framework, liberalizing trade, deregulating markets and prices, privatizating public enterprises, and strengthening management of the financial and public sectors. The evidence shows that the pace of policy reform has been uneven, both across sectors and across countries, Many governments are moving closer to macroeconomic stability, increasing their competitiveness in world markets, and providing better incentives for their agricultural pro- ducers. They have been less successful, however, in reforming their public enterprises and financial sectors. And even those countries that have undertaken substantial reforms have yet to achieve a macroeconomic policy stance considered sound by international standards. One of the report's key findings is that improving policies paid off in higher GDP and sectoral growth rates, which are vital to reducing poverty. In countries where policies deteriorated, eco- nomic performance generally worsened. Part of the explanation, then, for Africa's disappointing aggregate growth is the lack of sustained reform, not a failure of the reforms themselves. The chal- lenge for the future is to pursue policy reforms with stronger commitment and with a rethinking of the adjustment strategy in the areas that have met with least success. Although adjustment can work in Africa, the report recognizes that it cannot work miracles. Achieving long-term, equitable growth also requires more investment in human capital and infra- structure, greater expansion of institutional capacity, and better governance. The focus on adjust- ment in this book complements the broader development perspective of an earlier World Bank report, Sub-Saharan Africa: From Crisis to Sustainable Growth. NONE This book is the second in a series of Policy Research Reports, intended to bring to a broad audience the results of World Bank research on development policy issues. These reports take stock of what is and is not known about these issues and contribute to the debate on appropriate public policies for developing economies. The first book in the series, The East Asian Miracle: Economic Growth and Public Policy} was published in 1993. COVER DESIGN BY BRI|AN NOYES / THE MAGAZINE GROUP