Shantayanan Devarajan David R. Dollar 22118 Torgny Holmgren Editc.rs S ~~~~~WO RL D A - ~ Ethiopia -Ghana Ke' M'; NiJen TanZania uLana - zzmb_l ReX a rmX 1?kFRIC Aid and Reform in Africa Lessons from Ten Case Studies Edited by Shantayanan Devarajan The World Bank David R. Dollar The World Bank Torgny Holmgren Ministry of Foreign Affairs, Sweden The World Bank Washington, D.C. Copyright (© 2001 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, USA All rights reserved Manufactured in the United States of America First printing April 2001 The findings, interpretations, and conclusions expressed in this book are en- tirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Execu- tive Directors or the countries they represent. The World Bank does not guaran- tee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denomi- nations, and other information shown on any map in this volume 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. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission to reproduce por- tions of the work promptly Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use is granted by the World Bank, provided that the appropriate fee is paid directly to the Copy- right Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clear- ance Center before photocopying items. For permission to reprint individual articles or chapters, please fax a request with complete information to the Republication Department, Copyright Clear- ance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the Office of the Publisher, World Bank, at the address above or faxed to 202-522-2422. Cover design by Drew Fasick and Ultra Designs. Library of Congress Cataloging-in-Publication Data Aid and reform in Africa: a report from ten countries / edited by Shantayanan Devarajan, David Dollar, Torgny Holmgren. p. cm. Includes bibliographical references and index. ISBN 0-8213-4669-5 1. Economic assistance-Africa. 2. Africa-Economic policy. I. Devarajan, Shantayanan, 1954- II. Dollar, David. III. Holmgren, Torgny, 1954- HC800 .A628 2000 338.96-dc2l 00-050359 Contents Foreword xi Acknowledgments xiii 1 Overview 1 PART I SUCCESSFUL REFORMERS 2 Ghana 45 3 Uganda 101 PART II POST-SOCIALIST REFORMERS 4 Ethiopia 167 5 Mali 227 6 Tanzania 287 PART III MIXED REFORMERS 7 C6te d'Ivoire 363 8 Kenya 469 9 Zambia 533 PART IV NONREFORMERS 10 Democratic Republic of Congo 627 11 Nigeria 645 Index 679 iv Contents Box 1.1 Measuring Economic Policy 3 TABLES 1.1 Total Effective Development Assistance in Relation to GDP, 1978-97 13 1.2 Total Effective Development Assistance Per Capita, 1980-96 14 1.3 Relative Shares of Types of Aid (Official Development Finance), 1980-96 37 2.1 Ghana: Policy-Based Loans from Multilateral Institutions, 1983-96 50 2.2 Phasing and Sequencing of Reform Policies in Ghana, 1983-96 52 2.3 Ghana: Effective Development Assistance, 1980-96 93 2.4 Ghana: EDA Loans and Donor Grants by Sector, 1980-96 94 3.1 Uganda: Total Disbursements on Loans and Grants (ODF), 1980-96 107 3.2 World Bank and IMF Policy Reform Loans to Uganda, 1980-98 108 3.3 Uganda: Import Support, 1991/92-1996/97 110 3.4 Uganda: Growth Patterns Since 1989 125 3.5 Uganda: Sector Distribution of Budgetary Allocation, 1965-70 and 1994-99 134 3.6 Aid Commitments (ODA) to Uganda, 1962-97 147 3.7 Uganda: Official Development Finance, 1980-98 148 3.8 Uganda: Effective Development Assistance, 1980-96 149 3.9 Uganda: Total Official Development Finance by Creditor, 1980-96 150 3.10 Uganda: Total Effective Development Assistance by Creditor, 1980-96 152 3.11 Uganda: Total Disbursements on Loans by Creditor, 1980-96 154 3.12 Uganda: Total Grant Equivalents of Loan Disbursements by Creditor, 1980-96 154 3.13 Uganda: Total Non-Technical Assistance Grants by Creditor, 1980-96 155 3.14 Uganda: Total Technical Assistance Grants by Creditor, 1980-96 155 3.15 Uganda: Disbursements on Official Loans by Sector, 1980-96 156 3.16 Uganda: Debt Tables, 1980-95 157 3.17 Uganda: Multilateral Debt Agreements with Official Creditors, 1980-97 158 3.18 Uganda: Debt Relief, 1991-97 159 Contents v 4.1 Ethiopia: Basic Economic Indicators 172 4.2 Ethiopia: Fiscal Trends, 1980-99 175 4.3 Ethiopia: Official Development Assistance, 1980-97 176 4.4 Ethiopia: Effective Development Assistance, 1980-96 179 4.5 Ethiopia: Net ODA Loans and Grants by Creditor or Donor, 1970-97 182 4.6 Ethiopia: Bilateral ODA Commitments by Purpose, 1973-97 183 4.7 Ethiopia: Total Debt Stocks, 1980-97 184 4.8 Ethiopia: Debt Stocks and Debt Service Indicators, 1980-97 185 5.1 Mali: Net ODA Disbursements, 1964-97 247 5.2 Mali: ODA Commitments by Sector, 1973-97 249 5.3 Mali: Technical Assistance, 1962-97 252 5.4 Mali: Official Development Finance and Its Components, by Source, 1975-97 276 5.5 Mali: World Bank Adjustment Credits, 1980-98 277 5.6 Mali: IMF Adjustment Loans, 1969-96 278 5.7 Mali: Multilateral Debt Restructuring Agreements with Official Creditors, 1980-98 279 5.8 Mali: Debt Ratios 280 5.9 Mali: Structure of Long-Term, Public, and Government- Secured Debt Outstanding, 1970-97 280 5.10 Mali: Structure of Total Service of Long-Term, Public, and Government-Secured Debt, 1970-97 281 5.11 Mali: Budgetary Convergence Criteria 282 5.12 Mali: Tariff Structure by Rate Categories in 1998 282 5.13 Resources for Mali's Cereals Market Restructuring Program, by Phase 283 6.1 Total Aid to Tanzania, 1970-96 291 6.2 Tanzania: Effective Development Assistance in the Form of Loans, 1970-96 293 6.3 World Bank Loans to Tanzania, 1986-99 303 6.4 Tanzania: Data for the Pre-Arusha Period, 1961-67 313 6.5 Tanzania: Data for the Pre-Crisis Period, 1968-78 315 6.6 Tanzania: Data for the Crisis Period, 1979-85 327 6.7 Tanzania: Data for the Reform Period since 1986 329 6.8 Tanzania: Shares of Gross Fixed Capital Formation by Type of Economic Activity at Current Prices, by Sector 331 6.9 Total Grants to Tanzania by Creditor 343 6.10 Tanzania: Total Disbursements on All EDA Loans 345 6.11 Tanzania: Total Disbursements on EDA Loans by Sector 346 6.12 Tanzania: Total Disbursements by Sector 350 vi Contents 6.13 Tanzania: Disbursements on EDA Loans by Creditor 352 6.14 Tanzania: Total Grant Element by Creditor 354 6.15 Technical Assistance Grants to Tanzania as a Share of Total TA Grants 356 7.1 ODA Commitments to C6te d'Ivoire, 1975-98 368 7.2 C6te d'lvoire: Total ODA Commitments by Sector, 1975-97 368 7.3 C6te d'Ivoire: Total Disbursements on Official Loans by Sector, 1975-96 370 7.4 C6te d'Ivoire: Official Development Finance and Total Disbursements on Loans by Creditors, 1975-96 371 7.5 C6te d'Ivoire: Technical Assistance Grants by Creditor, 1980-96 375 7.6 C6te d'Ivoire: Aggregate Net Transfers by Donor, 1995-97 377 7.7 C6te d'Ivoire: Selected Economic and Financial Indicators, 1987-96 384 7.8 C6te d'Ivoire: Total Net ODAby Donor Type, 1960-97 452 7.9 C6te d'Ivoire: EDA Flows by Type, 1970-96 453 7.10 C6te d'Ivoire: Total Disbursements by Donor Type, 1970-96 454 7.11 C6te d'Ivoire: World Bank Policy-Based Loans, 1980-98 455 7.12 C6te d'Ivoire: IMF Policy-Based Loans 456 7.13 C6te d'lvoire: Multilateral Debt Agreements with Official Creditors, 1980-98 457 8.1 Kenya's Share of Total ODA to Sub-Saharan Africa, 1970-96 472 8.2 IMF and World Bank/IDA Policy Reform Loans to Kenya, 1974-96 476 8.3 Stages in Kenya's Structural Adjustment Program 495 8.4 Kenya: Foreign Direct Investment, 1980-96 511 8.5 Total Aid Flows to Kenya, 1970-96 520 8.6 Kenya: Loan and Grant Components of ODA Flows, 1970-96 521 8.7 Kenya: Gross and Net ODA Flows, 1970-96 522 8.8 ODA to Kenya by Source: Multilateral and Bilateral 523 8.9 Kenya: Shares of Major Aid Donors/Lenders in ODA 524 8.10 Grant Aid to Kenya for Technical Assistance, 1970-96 525 8.11 Balance of Payments Aid to Kenya: Multilateral, Bilateral, and IMF Loans, 1970-96 526 9.1 Policy-Based Loans to Zambia from Multilateral Institutions 542 9.2 Net Aid Transfers to Zambia, 1991-99 547 9.3 Consultative Group Meetings on Zambia, 1991-99 550 9.4 Zambia: Macroeconomic Indicators, 1973-90 552 Contents vii 9.5 Zambia: Macroeconomic Indicators, 1991-98 557 9.6 Zambia: Total Aid in Loans by Sector, 1980-96 600 9.7 Zambia: Total Disbursements by Sector, 1980-96 601 9.8 Zambia: Total Grants by Creditor, 1980-96 602 9.9 Zambia: Total Disbursements by Creditor, 1980-96 603 9.10 Zambia: Total Disbursements, Bilateral and Multilateral, 1980-96 604 9.11 Zambia: External Debt, 1990-98 605 9.12 Zambia: Government Expenditures, 1990-98 606 9.13 Zambia: Employment by Economic Sector, 1990-98 607 9.14 Zambia: Marketed Production of Selected Crops, 1990-98 608 9.15 Zambia: Social Indicators, 1986-95 609 11.1 Nigeria: Contributions by Donor, 1970-97 655 11.2 Sectoral Distribution of Aid to Nigeria, 1970-96 658 11.3 Nigeria: External Debt Service Profile 660 11.4 Nigeria: Structural Adjustment and Implementation Record 668 FIGURES 1.1 Aid and Policy in the 1990s 2 1.2 Policy and Growth in the 1990s 3 1.3 Actual Aid, Poverty-Efficient Aid, and Policy 5 1.4 Aid and Policy: Successful Reformers 15 1.5 Aid and Policy: Post-Socialist Reformers 16 1.6 Aid and Policy: Mixed Reformers 18 1.7 Aid and Policy: Nonreformers 19 1.8 Aid and Policy in Reforming Countries 22 1.9 Zambia: Aid and Policy 25 1.10 Total Official Development Finance Flows to Successful Reformers, 1980-96 38 1.11 Total Official Development Finance Flows to Post-Socialist Reformers, 1980-96 38 1.12 Total Official Development Finance Flows to Mixed Reformers, 1980-96 39 1.13 Total Official Development Finance Flows to Nonreformers, 1980-96 39 2.1 Trends in Aid to Ghana, 1970-96 49 2.2 Ghana: Percentage of EDA Loans by Creditor, 1970-96 51 2.3 Composition of Total Aid to Ghana, 1970-96 53 3.1 Official Development Finance Flows to Uganda, 1980-96 105 3.2 Uganda: Debt Indicators, 1980-97 111 viii Contents 3.3 Uganda: ODF and EDA from Bilateral Donors, 1980-96 113 3.4 Uganda: ODF and EDA from Multilateral Institutions, 1980-96 114 3.5 Inflation and Growth in Uganda, 1980-96 124 3.6 Uganda: Aid and Policy, 1970-96 138 4.1 Ethiopia: Average Grace Period and Maturity of ODA Loans 180 4.2 Ethiopia: Average Grant Element and Share of Multilateral Debt 181 4.3 Ethiopia: Growth Rates of GDP, Exports, and Net ODA 187 4.4 Ethiopia: Sources and Uses of Foreign Exchange 188 5.1 The Different Aid Concepts Used (in Terms of Disbursements) 232 5.2 Mali: Total Net Official Development Assistance 269 5.3 Mali: Aid Intensity, 1975-97 269 5.4 Mali: Official Development Finance (Disbursements), 1970-96 270 5.5 Mali: Composition of Flows of ODF (Disbursements), 1970-96 271 5.6 Mali: ODF from Bilateral Sources (Disbursements), 1975-96 272 5.7 Mali: ODF from Multilateral Sources (Disbursements), 1975-96 273 5.8 Mali: Comparison of ODF and TOF, 1970-96 274 5.9 Mali: Net Transfers from the Public Sector, 1975-97 274 5.10 Mali: Debt Ratios, 1975-96 275 5.11 Mali: Structure of Long-Term, Public, Government-Secured Debt Outstanding, 1970-97 275 7.1 C6te d'Ivoire: Total Net ODA, 1970-96 367 7.2 C6te d'Ivoire: Official Development Finance Flows (Disbursements), 1970-96 369 7.3 C6te d'Ivoire: EDA Flows by Type, 1975-96 370 7.4 C6te d'Ivoire: Comparison of ODF and TOF, 1970-97 372 7.5 C6te d'Ivoire: Aid Intensity, 1975-97 373 7.6 Cote d'Ivoire: Net Transfers on Aid, 1975-97 375 7.7 C6te d'Ivoire: Bilateral Official Development Finance Flows, 1975-96 458 7.8 C6te d'Ivoire: Multilateral Development Finance Flows, 1975-96 459 7.9 Cote d'Ivoire: Composition of Net Private Flows, 1960-97 460 7.10 C6te d'Ivoire: Debt Ratios, 1975-96 460 Contents ix 9.1 Zambia: Total Grants (EDA), 1980-96 544 9.2 Zambia: Total Aid in EDA Loans, 1980-96 548 11.1 Nigeria: Per Capita Income, 1970-97 648 11.2 Nigeria: Nonconcessional External Debt to Exports, 1980-96 651 11.3 Nigeria: Net ODA Per Capita, All Donors 652 11.4 Nigeria: Oil Exports and Net ODA 653 11.5 Nigeria: Net ODA Per Capita, Multilateral Donors 654 11.6 Nigeria: Grants as a Percentage of Total Disbursements 655 Foreword Three years ago, in 1998, we published a study, Assessing Aid, from which two themes emerged. First, effective aid requires the right timing, and second, it requires the right mix of money and ideas. This volume builds upon the lessons learned from that study. In 1999 1 outlined a proposal for a comprehensive development frame- work, a key element of which-and the one I feel most strongly about- is ownership. I have always argued, and continue to argue, that a development program must be country-owned, not owned by donors or the World Bank. The ten case studies that make up this volume also show that country ownership is the way to make assistance effective. These studies of aid and reform in Africa confirm that when aid sup- ports a country-owned development strategy, it can lead to sustained growth and poverty alleviation. The case studies also show that when reform is imposed from abroad, even as a quid pro quo for aid, it is not sustainable. In other words, this book provides empirical foundations for the new development partnership, with donors supporting country- owned programs with their knowledge, ideas, and money. Each recipient country of aid in Africa has demonstrated different results in the reforms it has undertaken. Two have grown rapidly and reduced poverty. Two others, as this volume outlines, have shown sig- nificant reform in recent years, but it may be too early to judge sustainability. In yet other countries, matters have changed little, or they have even become worse. The divergent outcomes have resulted from several factors, all of which are elaborated upon in the case studies. It is significant that the authors of the case studies are all nationals of the countries themselves, writing in collaboration with members of the donor community. Together, and with outside support, Africans can and will address their challenges-in their own ways-for lasting results. This message was underscored by all the 22 African presidents I met on my last trip to Africa. And this is why the subject of aid and reform is so crucial to Africa. xi xii Foreword Even a cursory gleaning of development statistics will make clear the scale of the problems that Africa faces, and the challenges that must be overcome. Let me present just a few. Between 1990 and 1998, the number of people living in poverty actu- ally increased in Sub-Saharan Africa, from 242 to 291 million people. And while the number of people living on less than a dollar a day-in absolute poverty-declined in this period by five percentage points the world over, in Africa the change was barely discernible. Even with faster economic growth, the number of people living on less than a dollar a day will increase from nearly 291 million in 1998 to nearly 330 million in 2008. Under conditions of slower growth and ris- ing inequality, that number could be as high as 406 million. In 17 of the 48 countries in Sub-Saharan Africa, life expectancy declined between 1990 and 1998, while it went up from age 63 to age 65 in all developing countries. Africa is home to 70 percent of the adults, and 80 percent of the children, living with HIV in the world. Of the nearly 22 million people who have died from AIDS and AIDS-related diseases, 17 million have been Africans. It is estimated that by the year 2014, the population of the world will be 7 billion. Of the increase, nearly a quarter-240 million-will be in Sub-Saharan Africa. To deal with the challenges that these problems pose-and those of coping with a rapidly globalizing digital economy-aid must be effec- tive for reform to take hold. The lessons drawn from the studies of the relationship between aid and reform in Africa become critical to its people, and to their future. This book also reflects the World Bank's new style of research. Just as Voices of the Poor taught us about poverty, the authors of the case studies in this book teach us about how aid really affects reform in Africa. The studies capture the richness, the texture, and nuances of aid and reform. And the conclusions show that we are not afraid to be critical of ourselves. The authors of the studies in Aid and Reform in Africa deserve to be commended. This volume represents the World Bank doing what it does best: asking hard questions; building on 55 years of experience; reaching out to partners; pushing the frontiers of knowledge; delivering results; and doing all this while keeping its eyes on the main goal-a world free of poverty. James D. Wolfensohn President The World Bank Acknowledgments This book has enjoyed an unusual degree of collaboration. We would like to express our gratitude to Paul Collier, who encouraged us to un- dertake the project, helped us frame the study, and provided invaluable advice and support throughout the process. We are also grateful to Alan Gelb and Kwesi Botchwey for comments and suggestions at various stages of the manuscript. We benefited from the research assistance of Charles Chang and Jacqueline Vanderpuye-Orgle, and the logistical sup- port of Raquel Luz, Emily Khine, Shanaz Khan and Rina Bonfield. Fi- nally, our thanks go to the World Bank publications team and to Catherine Sunshine, who did an excellent job of editing the manuscript. For helpful comments, we are grateful to participants at the follow- ing workshops and seminars, where the papers were presented: Frank- furt, Germany, June 7-8, 1999; Washington, D.C., September 27, 1999; Dar es Salaam, Tanzania, November 18-19,1999; Nairobi, Kenya, Sep- tember 19, 2000; and Clermont-Ferrand, France, September 20-21, 2000. Finally, for financial assistance, we are grateful to the World Bank's Research Support Budget, and the governments of France, Germany, the Netherlands, Norway, Sweden, and Switzerland. xiii Overview Since the early 1980s, virtually every African country has received large amounts of aid aimed at stimulating policy reform. The results have varied enormously. Ghana and Uganda were successful reformers that grew rapidly and reduced poverty. Cote d'Ivoire and Ethiopia have shown significant reform in recent years, but it remains to be seen if this is sustained. In other countries, policies changed little or even got worse. This book synthesizes the findings from 10 case studies that investigate whether, when, and how foreign aid affected economic policy in Africa. The 10 countries covered in the study reveal the range of African policy experience. The countries are shown in figure 1.1, which relates foreign aid in the 1990s to a measure of overall economic policy (the World Bank's Country Policy and Institutional Assessment). This is a broad measure of economic policy that covers macroeconomic management, as well as effectiveness of the public sector in providing the services that are es- sential for growth and poverty reduction (see box 1.1). All of these coun- tries received large amounts of aid; all had structural adjustment programs with the international financial institutions. Yet the policy outcomes are quite diverse. Ghana and Uganda achieved sustained good policy (and good economic outcomes). The other eight countries did not. In examining their cases, it is useful to divide them into three groups: the post-socialist reformers (Ethiopia, Mali, and Tanzania), the mixed reformers (Cote d'Ivoire, Kenya, and Zambia), and the nonreformers (the Democratic Republic of Congo [formerly Zaire], and Nigeria). These categorizations do not reflect policy changes in the past one or two years; nor are they meant to be an assessment of current policy. The case stud- ies look at the reform process during the 1980s and 1990s. The average policy ratings for the 1990s give an indication of the pace of reform dur- ing the period covered by the studies. What constitutes "good policy" and how to measure it are somewhat controversial. Development experience and research have established a 1 2 Overview FIGURE 1.1 AID AND POLICY IN THE 1990S Aid as a percentage of GDP (1991-97) 8 Zambia Mali 6 Tanzania Ethiopia Uganda 40 Cote d'Jvoire Congo, Ghana Dem. Rep. of *Kenya 2.- Nigeria 0 Poor Mediocre Good Policy rating (average 1993-99) fair amount of knowledge about policies that promote growth and re- duce poverty: absence of high inflation, functioning foreign exchange and financial markets, openness to foreign trade, effective rule of law, and delivery of key services such as education. There remain, however, policy areas about which we know less and about which there is ongo- ing debate. Openness of the capital account and school vouchers for poor children are two examples. Nevertheless, we know enough about devel- opment policies to make a fair assessment of the quality of policies across countries and over time. We emphasize, however, that learning about development policy is an ongoing process. In fact, one of the main roles of effective aid is to support countries' learning about good policy, both drawing on existing knowledge and creating new knowledge. Finally, the notion that we are doing a reasonable job of measuring policy across countries is supported by the fact that our broad measure of policy pre- dicts fairly well the gross domestic product (GDP) growth rates of the four categories of countries in our study. The successful reformers have grown well in the 1990s; the nonreformers have gone backwards rap- idly; and the performance of the other groups is in between (figure 1.2). That the 10 countries in our study all received large amounts of aid, including conditional loans, yet ended up with vastly different policies suggests that aid is not a primary determinant of policy. That is the con- clusion of a number of recent cross-country statistical studies summarized Overview 3 Box i.1 MEASuiNG ECONOMIC POLICY The meAsureof poliy used in the figuresthoughout tispaperis abroad one. Corneptualy, it meaue the Xex to whih goverment policy creates a good envonment for broadbased growth and poverty reduc- tion. in practice, it has for components:, a Microe o poiiewhete fisc4 moetary, and exhae rate polcies provide a stable environent for economic activity * Stn etarpoldew ffie extent to winch trade, tax, an sectorad poh ces create good nentives for acumulion by househl and * Pubic saecwomanagemet th etet to which public sedor insti- tuons effectively provide sers to pae ii- tiatve, such as the re of law (functioning of the juiciary, pole), ifrastrucue, and sol seravies * Soda? domte exent to whichpolcy eres hfpartici- patio othe society thosocial sevesthat reach e poor ad disadvantagd, inclding wom and eth m,mities., FIGURE 1.2 POLICY AND GROWTH IN THE 1990S Real GDP per capita growth (1990-96) 5 Successful reformers 3 1 Post-socialists Mixed reformers -3 0 Nonreformers -5 Poor Mediocre Good Policy rating (average 1993-99) 4 Overview in the World Bank report Assessing Aid (1998). Specifically, Burnside and Dollar (2000) found that the amount of aid countries received had no effect on the quality of their macroeconomic policies. Dollar and Svensson (2000) investigated the determinants of the success or failure of reform programs supported by adjustment assistance. They found that reform outcome could be predicted by underlying political-economy variables such as length of tenure of the government and whether the leader was democratically elected. Variables under donors' control had no influ- ence on the success or failure of reform. Other studies have found no systematic relationship between aid and policy (Killick 1991; Ranis 1995; Mosley, Harrigan, and Toye 1995). More generally, recent literature indicates that policy formulation depends primarily on domestic political-economy factors (Rodrik 1996; van de Walle and Johnston 1996). There is no consistent relationship be- tween aid and reform. It is important to emphasize, though, that cross- country studies reveal what is true on average. The average relationship is likely to disguise the fact that aid supported policy reform in some cases and sustained poor policy in others (Sachs 1994). The purpose of country studies is to delve more deeply into the relationship between aid and reform, capturing the nuances of both successful and stalled reforms. More specifically, our study addresses the following questions: * Are there common characteristics of successful and failed reform- ers that enable us to understand better the political economy of reform? * Do donors tailor their assistance to different types of countries and to different phases in the reform process, or just provide undiffer- entiated assistance? * Have large amounts of financial aid to countries with poor policy sustained the bad policies? * Has aid played a useful role in the successful reformers? If so, how? These are important questions for the development community. Im- provements in economic policy are the key to more rapid growth and poverty reduction in Africa and elsewhere in the developing world. How can foreign aid best support these improvements? If policy is truly inde- pendent of aid, then-to have the maximum impact on poverty reduc- tion-aid should be allocated on the basis of how poor countries are and the observed quality of their economic policy. Collier and Dollar (1999) estimate the poverty-efficient allocation of aid, assuming that aid has no effect on policy. But the effect of aid on growth and poverty reduction increases with the quality of policy. For a given level of poverty, there- fore, aid should rise with the quality of policy (figure 1.3). Figure 1.3 also shows the actual allocation of aid in 1996, after controlling for the level of poverty and for population. Donors tend to concentrate their assis- Overview 5 FIGURE 1.3 ACTUAL AID, POVERTY-EFFICIENT AID, AND POLICY Aid as percentage of GDP 7 6 Poverty-efficient aid 5 4- 3 Actual aid 2 / ~ 0 Poor Mediocre Good Policy tance in countries with mediocre policies and actually discriminate against poor countries that have reformed. A number of factors account for this concentration of assistance in countries with weak policies, but one of them is the expectation that aid can spur policy reform and thus have a big payoff in these cases. Our study asks whether this pattern of aid allocation makes sense: that is, has the large amount of aid to coun- tries with weak policies actually led to reform? The main findings from the 10 case studies are that: * Policy formation is primarily driven by domestic political economy. It is hard to find simple generalizations about successful and failed reformers, but one striking fact is that countries that have truly severe economic and political crises (Ghana, Uganda, Nigeria, and the Democratic Republic of Congo) tend to move to the extremes, either developing coherent reform movements or declining rap- idly. Of these crisis countries, it is the ones rich in natural resources that fell into civil war (Nigeria, the Democratic Republic of Congo). The countries that avoided crises did not fall apart, but have found it hard to develop coherent reform movements that can overcome vested interests who benefit from the perpetuation of poor policies. 6 Overview * Successful reformers have consultative processes that build con- sensus for change. However, in our sample there is no relationship between formal democratic institutions and good economic policy. * Large amounts of aid to countries with bad policy sustained those poor policies. The funding allowed the delay of reform. Attaching conditions to the aid in these cases has not successfully led to policy change, nor has it delayed the disbursement of funds. * Aid played a significant and positive role in the two sustained re- formers (Ghana, Uganda). It helped with ideas, though in both cases there was also strong local talent formulating policy. In these cases, finance grew as policy improved and increased the benefits of re- form, helping to sustain political support. * In general, donors have not discriminated effectively among dif- ferent countries and different phases of the reform process. Do- nors tend to provide the same package of assistance everywhere and at all times. The lessons from the Ghana and Uganda cases are that donors should concentrate on technical assistance and other soft support without large-scale budget or balance of payments support in the phase before governments are serious about reform. If a reform movement develops, finance can be increased as poli- cies actually improve. In the early stage of serious reform, political leaders and technocrats actually welcome conditionality in order to bind themselves to policy change. Once the reform movement is well entrenched, conditionality becomes less useful because it lim- its participation and disguises ownership. In summary, if donors thought that they had no effect on policy at all, the rational way to allocate aid would be based on how poor countries are and the observed quality of their policies. Our study shows that this approach is also the best way to ensure that aid has a positive effect on policy. It avoids disbursing large-scale finance into bad policy environ- ments, where the funds can actually sustain corrupt and incompetent governments. It puts resources into the good policy environments in which financial assistance has a positive effect, increasing the benefits of good policy and sustaining political support for reform. In addition to quantity, the composition of aid is important. In the pre-reform period, technical assistance and policy dialogue are most supportive of policy reform. During periods of rapid reform, policy dialogue is important, as is finance. This is the phase in which conditional loans tend to be useful and effective. At a later stage of reform, conditionality is less useful, while finance remains important. By altering both the quantity and com- position of aid, donors could more systematically support genuine policy reform in the developing world. Overview 7 The remainder of this overview chapter is organized as follows. The next section examines the political economy of reform in the 10 coun- tries covered by the case studies. The third section then looks at the pat- tern of aid to the countries. In the fourth section we link the two previous sections and investigate the effect of aid on reform. Section 5 concludes. POLITICAL ECONOMY OF REFORM IN 10 AFRICAN STATES At the beginning of the 1980s, the typical African country was following a "statist" model of development, with government intervention in al- most every sector of the economy. Most notable were government re- strictions on foreign trade (import tariffs, quotas, and export restrictions), maintenance of an overvalued exchange rate, and production in the hands of public enterprises. Not only did these policies fail to improve living standards in Africa-despite substantial amounts of foreign aid-but they seemed to exacerbate the effects of the external shocks of the 1970s. A consensus emerged in the donor community that improving economic policy in Africa was the highest priority (see, for example, World Bank 1982). Many donors, including the World Bank, began channeling their aid toward "structural adjustment programs"-loans and grants aimed at inducing governments to reform their policies. Twenty years later, there does not seem to have been a systematic relationship between the volume of aid and the extent to which African countries reformed their economic policies. Some countries, such as Ghana and Uganda, undertook sustained reforms; others, such as Kenya, Tanzania, and C6te d'Ivoire, reformed in fits and starts; while others, notably Nigeria and the Democratic Republic of Congo, deteriorated. Yet, all of these countries received large amounts of aid. Clearly, there- fore, the process of policy reform was significantly affected by domestic factors. In this section, we investigate what those factors were, and how they spurred or hampered policy reform. In particular, we look at the roles of crises, technocrats, ideology, leadership, and institutions. It Takes a Crisis It is difficult to find a case where reform occurred without a crisis. In January 1983, inflation in Ghana was over 100 percent. Meanwhile, Nigeria announced it was expelling 1.2 million Ghanaians-10 percent of Ghana's population-back to their homeland. Brush fires caused by a recent drought destroyed much of the cocoa crop. An aid mission to "friendly" socialist countries in 1982 generated no financial support. The country had no choice but to undertake a stabilization and structural adjustment program supported by the International Monetary Fund 8 Overview (IMF) and the World Bank. In short, Ghana embraced serious reform only after it had exhausted all other options. Similarly, in Uganda in 1986, the banking system was insolvent; the country had two weeks of imports in reserves; and inflation was 300 percent. The government had little choice but to restart the IMF program. When a country is in a crisis, even those who benefited from the dis- torted policy regime begin to suffer.' The crisis in Ghana galvanized political support for reforms from those who were most severely hurt by the previous regime. An unlikely coalition of neo-Marxist students, ur- ban workers, and lower-ranking army officers formed the nucleus of a pro-reform group. While the army as a whole may have supported ear- lier regimes, the 1981-83 crisis caused a split between the higher and lower ranks. Jerry Rawlings, the flight-lieutenant who led the December 31, 1981 coup, came from the lower ranks. In Mali, likewise, most of the significant reforms occurred when the government's "back was against the wall"-when civil servants had not been paid for eight months (1988), when cotton prices fell sharply (1992), or when a drought led to unprecedented power cuts (1999). The other mixed reformers in our 10 countries either did not have a crisis or had other options when faced with a crisis-and chose to exer- cise them. For instance, Kenya was rescued from the first oil crisis by the coffee boom of the mid-1970s. Zambia met the first copper crisis of 1974 with commercial borrowing, and thereby avoided having to negotiate policy reforms with the Bretton Woods institutions. That Zambia's re- forms in the early 1990s came out of a "honeymoon" period, not a crisis, may have something to do with the weak implementation and stop-go nature of those reforms. Moreover, C6te d'Ivoire continued to receive aid from France even when its policies were extremely poor. Only when the spigot was turned off in 1993 did C6te d'Ivoire face its first real crisis and undertake its first real reform. Second, leadership in a crisis matters. How the chief executive re- sponds to pressure from vested interests can make a difference. Fearing that structural adjustment programs would cut their food subsidies, ur- ban workers in Zambia opposed these reforms until 1991. President Kaunda, who faced opposition to reform from his own political party and bureaucracy, felt his political base would be undermined by reforms.2 But urban workers supported the Ghanaian and Ugandan reform pro- grams, not least because the programs were championed by the head of state in both countries. Ugandan president Museveni's support for the reform program did not waver even when the International Coffee Agree- ment collapsed and the 1990-91 Gulf War sent oil prices skyrocketing. Ayittey (1989) observes that when the stature of the head of state is inex- tricably linked to the state of the economy, reform requires immense cour- Overview 9 age, since it implies failure as a leader (Houphouiet-Boigny of C6te d'Ivoire and Nyerere of Tanzania are examples). Third, the complexity of many of these reforms makes it difficult to judge before the fact who will win and who will lose. The private sector in Kenya supported price decontrol, even if the ensuing trade liberaliza- tion may have hurt some firms. The business community in Ghana was too diffuse to either support or oppose the Economic Recovery Program: they did not benefit from the earlier regime, and probably suffered un- der trade liberalization. In short, rather than instilling a status quo bias (Fernandez and Rodrik 1991), uncertainty about reform outcomes may- in a crisis atmosphere-break down vested interests. A final point about crisis is that the combination of economic crisis and the existence of significant natural resources provides a fertile ground for civil war and political disintegration (Collier and Hoeffler 2000). Thus, in Nigeria and the Democratic Republic of Congo, economic crisis resulted in civil wars in which different groups vied for control of the resources. Backsliding Just as every reform episode in Africa is preceded by a crisis, it is fol- lowed by reform slippage. In Ghana, an 80 percent wage increase to civil servants in 1992 undermined the macroeconomic program. A value added tax (VAT) introduced in 1995 had to be rescinded because of po- litical opposition. The VAT was reintroduced in 1998 with significantly lower rates. Uganda's stabilization program began slipping in 1991-92, with the fiscal deficit approaching 10 percent of GDP and inflation ris- ing. The donors cut back assistance that year, and the IMF/World Bank program was delayed. The reasons for this backsliding become more apparent when we ex- amine the experience of mixed reformers. In Tanzania, there was retro- gression in the early 1990s because the second-generation reforms-civil service reform, budget management, privatization-were much more intrusive. They required fundamental changes to public sector institu- tions, something that the first-generation reforms, such as trade liberal- ization or exchange rate devaluation, did not. Even in a star performer like Uganda, public sector reform has yet to happen. Zambia's reform program has been undermined by falling copper prices and a severe drought in 1994-95. But here too the intrusiveness of the second-genera- tion reforms is a factor. Zambia has privatized 224 of the 275 parastatals, but the government's inability to privatize the copper mine, ZCCM, brought negotiations with the international financial institutions to a standstill. In Kenya's case, the first-generation reforms-trade liberal- ization, price decontrol, and exchange rate flexibility-were undertaken 10 Overview by a small circle of reformers in the Central Bank and Ministry of Fi- nance. They were ill-equipped to undertake the next round of reforms, which were of a structural and institutional nature, and required much broader consultation, at the very least with line ministries. After adopt- ing an ambitious reform program in 1992, Ethiopia may be going through its backsliding phase now (the war with Eritrea did not help). Privatization, corruption, and transparency are all issues the govern- ment has been reluctant to address, and may lead to a stalemate with the donor community. Despite major strides in macroeconomic and trade policy reform, Mali has been unable to privatize many inefficient and loss-making public enterprises. Technocrats Any reform program has winners and losers in the short run, and the losers will attempt to block the policy change. It takes disinterested eco- nomic analysis to gauge whether on balance the economy will benefit from the reforms. Countries with a strong cadre of economists would therefore have an edge. For instance, the appointments of Joe Abbey and Kwesi Botchwey to the National Economic Review Committee in Ghana contributed to the success of the economic program. Sensitive to popular sentiment against devaluation, Botchwey was able to sell the program as one of bonuses and surcharges that accomplished the same thing. But it is important to note that in Ghana the technocrats began to lose power after the 1992 parliamentary elections, when elected legisla- tors and not a group of highly skilled but unelected technocrats began designing policies. In Uganda as well, technocrats such as Emmanuel Tumusiime-Mutebile have played a critical role in developing the re- form program. Technocrats alone, however, cannot ensure reform. Kenya had some of the best, but its reform record is very mixed. Technocrats anticipated all the public sector reform issues in Kenya in the 1990s through their sessional paper series. But without political support from the leader- ship, they have been unable to accomplish much. The same was true in Zambia in 1986, when President Kaunda replaced his economic team, who had just negotiated an adjustment program with the Bank and Fund, with a group of anti-reformers. Ideology It is sometimes argued that the resistance to, and impetus for, reforms are driven not by economic analysis but by ideology. This interpretation is especially tempting in Africa, given the legacy of socialist ideology in the continent. The case studies reveal that reality is more complex. One Overview 11 of the most left-leaning governments in Africa, Ghana, is a star reformer. The left-wing factions in Ghana did oppose aid from the West (and, by extension, devaluations and associated reforms). But when the leftist intellectual Samir Amin was brought in to advise the government in 1982, he told them Ghana had no choice but to deal with the interna- tional financial institutions. Kenya may have accepted the label of "Afri- can socialism" at independence, but its economic policies followed a more market-oriented and foreign investment-friendly path. Neverthe- less, its reform record has been mixed. In 1981, Tanzania was unable to present a credible reform package to the World Bank and IMF (despite encouragement from both institutions) because the Nyerere government believed the problems causing the crisis were external and not internal policy weaknesses. Later, however, when the anti-reform minister of plan- ning, Malima, was put in charge of drafting the government's reform package, he became pro-reform. Institutions Perhaps the most important lesson that emerges from the 10 case studies is that the way in which societies organize vested interests, technocrats, and the leadership drives the success of reform. It is a fair generalization to say that successful reformers have consultative processes that result in a broad consensus for reform. On the other hand, it is interesting that in our sample there is no relationship between formal democratic institu- tions and reform. The two strongest reformers-Rawlings and Museveni- came to power through military coups. Authoritarian governments in Nigeria and the Democratic Republic of Congo, on the other hand, never seriously embraced economic reform. The democratic government that replaced Kaunda in Zambia did not make much progress with reform. While not formally democratic, the Ugandan reform was broadly con- sultative at every step of the way. President Museveni established the Presidential National Forum to debate reform issues in 1987. The Ugan- dan Manufacturers Association sponsored seminars and discussion pa- pers in the 1987-89 period. The Presidential Economic Council had open debates on reform and sponsored a December 1989 conference on trade liberalization that has been described as a turning point in public opin- ion. In Tanzania, when economists at the University of Dar es Salaam began holding public meetings on economic liberalization, party lead- ers and policymakers had to start listening to the reformist elements in government. That same year (1984), the government allowed own-funded imports, devalued the currency, and increased agricultural prices by about 30 percent. By contrast, policymaking in Kenya appears to be restricted to a small circle. While prescient, the sessional papers are not debated in Parliament. 12 Overview As a result, the reforms are not always "owned" by even the line minis- tries, and other stakeholders are not consulted. The establishment of the Export Promotion Council in 1992 was a step toward involving the busi- ness community in policy decisions. Zambia's reforms were pushed through during the "honeymoon period" following multiparty elections. As Finance Minister Kasonde put it, "Necessary but unpopular deci- sions had to be quick. I was very interested in using the political status of the MMD government to make economic advancements." By moving quickly before opposition to the reforms could mobilize, the Zambian government may have contributed to the backsliding. Interest groups, none of which were consulted before the reforms, slowed down the pace of implementation. Overall, support for the Zambian reforms outside of a few cabinet ministers has been shallow. FOREIGN AID FOR AFRICAN REFORMERS Trends and Directions of Foreign Aid A striking feature in the case studies is the persistence of aid to indi- vidual countries. Aid was disbursed at an even level irrespective of policy reforms undertaken (see the appendix for a definition of aid). Two ex- amples run counter to this average trend. The first is the smooth increase in aid to the successful reform countries (Ghana and Uganda), and the second is the trend in the late 1990s toward greater selectivity in aid allocations. In this section we first review those broad trends and then characterize the aid relationship for the four country groups. We con- clude by commenting upon some common findings regarding the roles of nonfinancial aid and donor coordination. One reason why aid is so persistent is that in many cases it was a foreign policy tool rather than a tool for economic development. This has led to a high degree of inertia in bilateral aid flows over the years. Every single reduction of aid levels, however small, was deemed a sign of distrust of the recipient government rather than a judgment of its eco- nomic reform program. Alesina and Dollar (forthcoming) find that aid allocations for many bilateral donors have been driven by strategic vari- ables such as colonial relationships and voting patterns in the United Nations. Exceptions to this pattern are the Nordic countries, Canada, and the Netherlands. In our sample, the post-socialist reformers received the highest GDP share of aid and the mixed reform countries the largest amounts of aid per capita (tables 1.1 and 1.2). C6te d'Ivoire and Zambia top the list with US$74 and US$112 per capita, respectively, in the mid-1990s. The most successful reformers started off with very low aid levels. Aid to Ghana and Uganda increased as they reformed, but by the late 1990s they were getting half as much aid per capita as the mixed reformers. The Overview 13 TABLE 1.1 TOTAL EFFECTIVE DEVELOPMENT ASSISTANCE IN RELATION TO GDP, 1978-97 (percent) 1978-81 1982-85 1986-89 1990-93 1994-97 Successful reformers 1.12 1.54 3.62 3.89 3.56 Ghana 1.30 1.56 3.66 2.95 2.43 Uganda 0.94 1.52 3.58 4.83 4.68 Post-socialist reformers 4.89 4.94 6.00 6.05 5.20 Ethiopia 1.85 3.17 4.32 5.32 3.78 Mali 6.00 5.97 7.65 7.06 7.33 Tanzania 6.81 5.69 6.02 5.79 4.49 Mixed reformers 2.36 2.42 3.47 4.95 8.46 C6te d'Ivoire 0.85 0.65 1.22 2.52 5.61 Kenya 2.22 2.38 2.93 2.98 2.18 Zambia 4.00 4.24 6.26 9.36 17.59 Nonreformers 1.34 1.02 1.50 1.14 1.00 Congo, Dem. Rep. of 2.63 1.98 2.90 2.13 1.88 Nigeria 0.05 0.05 0.09 0.15 0.11 Source: Chang, FernAndez-Arias, and Serv6n 1999. nonreformers, finally, score very low on these measures with Nigeria receiving on average just US$1-$2 per capita in aid, among the lowest in Africa. (Donors give less aid per capita to populous countries, and that is an important factor explaining the low aid to Nigeria.) The successful market-based reformers received more multilateral aid while the post-socialist countries have mainly received bilateral aid (see table 1.3 in appendix 1.2). The reform programs in Ghana and Uganda started off with an unusually high share of multilateral aid, around 70 percent. Bilateral aid has increased over the years and now accounts for some 50 percent of the aid. An opposite trend can be noted among the post-socialist countries where multilateral aid is on the rise but on a much smaller scale. The mixed reformers received aid from almost all bilateral and multilateral donors, in equal amounts. There is a close association between the stability of aid flows and the share of multilateral versus bilateral aid. Aid levels are more stable in countries where bilateral do- nors dominate. Where the multilateral donors have had a larger pres- ence, the flow of aid depends more on the actual policies pursued. This holds as long as the international financial institutions have not commit- ted too many funds?as they then tend to lend in order to get paid back. The interesting exception to the pattern of aid persistence is the aid to Ghana and Uganda, which has risen in lockstep with improvements in policies-at least until the mid-1990s when aid started to taper off. Start- ing from very low aid levels and an initial period of technical assistance and policy dialogue, financial support took off as policies developed. 14 Overview TABLE 1.2 TOTAL EFFECTIVE DEVELOPMENT ASSISTANCE PER CAPITA, 1980-96 (U.S. dollars) 1978-81 1982-85 1986-89 1990-93 1994-97 Successful reformers 11.6 13.0 27.9 35.3 28.4 Ghana 13.0 12.2 34.3 35.2 31.4 Uganda 10.4 13.6 22.0 35.4 26.3 Post-socialist reformers 19.5 17.6 26.2 28.4 23.1 Ethiopia 7.9 10.3 16.0 19.9 15.1 Mali 33.1 32.1 46.7 46.0 47.4 Tanzania 37.8 26.9 39.8 39.3 29.6 Mixed reformers 26.7 25.6 32.1 43.7 54.2 C6te d'Ivoire 17.9 11.7 21.6 36.8 74.2 Kenya 23.0 28.4 30.2 34.5 24.5 Zambia 50.1 37.4 54.1 82.0 111.7 Nonreformers 4.7 3.7 4.4 2.8 2.0 Congo, Dem. Rep. of 13.3 9.4 15.0 8.9 4.9 Nigeria 0.9 0.7 1.4 2.1 1.4 Aid per capita in Sub- Saharan Africa 23.7 22.5 30.7 37.3 32.3 All ten countries' share of total aid to Sub- Saharan Africa (%) 30.7 29.5 32.9 31.3 34.4 Source: Chang, Fernandez-Arias, and Serven 1999. Also, the general persistence of aid may be changing in the 1990s. With increasing aid fatigue, budget cuts in countries of the Organisation for Economic Co-operation and Development (OECD), the end of the Cold War, and the introduction of "governance" conditionality, aid funds have become more movable. The emerging global consensus in the early 1990s on what forms good development policy has, together with re- search evidence, also helped induce a greater willingness to reallocate aid in favor of good policies. Successful Reformers Ghana and Uganda are among the most successful reformers in Sub- Saharan Africa during the past 20 years. There are many noticeable simi- larities in their reform programs as well as in the donors' support of them. Starting in 1986-87, with a time lag of some four years, Uganda followed a path more or less identical to the one Ghana had taken. In fact, a number of Ugandan delegations visited Accra to learn from the Ghanaian experience. Overview 15 FIGURE 1.4 AID AND POLICY: SUCCESSFUL REFORMERS Aid as a percentage of GDP Policy index 9 4.0 8 Policy (right axis) / \ - 3.5 6 3.0 5 4 2.5 3 AID/GDP (left axis) 2.0 0 1.5 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 1994-97 The two countries' new governments-which came to power through military coups-started off with initial programs built on price controls, barter trade, and seeking aid from the Eastern bloc. Neither government was convinced that market-friendly policies were desirable. The policy failures that followed helped trigger the introduction of market-based reform programs. Those programs in turn received good support from the donors, both multilateral and bilateral, after an initial "wait and see" period. Aid volumes picked up in the mid-1980s when the reform pro- grams took off and the level of aid has since then closely matched the policy improvements inlockstep fashion (figure 1.4). However, aid started to taper off in the late 1990s when policies (in Uganda) were still improv- ing. The aid level reached at that stage was far below what other coun- tries in the sample have received (excluding the nonreforming countries). Lending played the predominant role in the aid program in the 1980s, while grant-based aid has taken over in the 1990s (see figure 1.10 in Appendix 1.2). The shift is closely related to the gradual change from multilateral aid (loans) to bilateral aid (grants). Policy dialogue with the IMF and the World Bank played a critical role in the early years of reform. Both countries relied in the early stages of reform on small groups of dedicated technocrats and politicians. 16 Overview Conditional financial aid was later most helpful in pushing the reform agendas and in implementing the reforms. The two countries have a high degree of success in adjustment lending provided by the World Bank, with outcome ratings considerably higher than the average for Africa. Loans to Ghana during the first years of reform were judged as performing better than loans in the 1990s, when the macroeconomic situ- ation deteriorated. The success rate of adjustment lending to Uganda followed an inverse pattern with better outcomes in the 1990s than in the 1980s, when reforms were either rejected or implemented reluctantly. Post-Socialist Reformers A common pattern emerges in the reform path among the three post- socialist countries. Tanzania experienced two waves of reform with a period of backsliding in between, and it appears that Ethiopia and Mali are going through a similar cycle. How has aid responded to the differ- ent periods of reforms? Aid volumes did not follow this cyclical pattern but were quite steady over time (figure 1.5). The three countries, particularly Mali and Tanza- nia, received large amounts of aid irrespective of policy performance. FIGURE 1.5 AID AND POLICY: POST-SOCIALIST REFORMERS Aid as a percentage of GDP Policy index 9 4.0 8 7 3.5 6 3.0 5 4 Policy (right axis) 2.5 3 AID/GDP (left axis) 2 2.0 0 . _ 1.5 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 1994-97 Overview 17 When policies deteriorated in the early 1980s, aid in relation to GDP actually increased. The worse the policies got, the more aid they received. From the late 1980s there is a closer correlation between the (improving) policy level and the level of aid in relation to GDP. All three countries have mainly received grant aid (figure 1.11 in Appendix 1.2). Technical assistance (TA) has played an important role, counting for as much as 18 percent of overall aid, the highest share among any of the group of coun- tries studied. Lending accounts for approximately half of the assistance in the 1980s and one-third in the 1990s. Bilateral aid is two-thirds of total aid but a shift to multilateral aid is gradually taking place. Tanzania has been critically dependent on bilat- eral donors from the OECD's Development Assistance Committee, and it was mainly bilateral organizations that provided loans to Ethiopia during its socialist period. As a noncolony, Ethiopia did not until re- cently attract the level of aid it needed and could effectively absorb. The country now receives aid from both multilateral and bilateral donors but many policymakers express a strong preference for greater multi- lateralization of aid and disbursement in block grants. Mixed Reformers Among mixed reformers, aid has moved in inverse relation to policy reforms (figure 1.6). When policies were rather good (1970s and early 1980s) the countries received less assistance than they did when policies started to deteriorate. In the early 1990s, however, both policies and aid levels moved in a synchronized manner. For instance, Cote d'Ivoire re- ceived a large inflow of aid after the 1994 devaluation of the CFA franc. Zambia poses an interesting example of donor behavior. A reformist government came to power in Zambia in 1990. Donors rewarded this locally owned policy process with large increases in aid. Major reforms were initiated, but over time the commitment to reform has waned. Dif- ferent views among donors have led to counter-cyclical aid flows from bilateral and multilateral donors. The bilateral donors have increasingly tied their aid to political conditionality in the 1990s and have frozen disbursements on occasion. However, the multilateral institutions, stress- ing economic performance, have not cancelled any loans in the 1990s, while the 1980s was characterized by a stop-go pattern in their relation- ship with the Zambian government. In the 1980s, in the absence of aid coordination, bilaterals even increased their aid portfolios when Zambia cancelled the agreements with the multilaterals. Lending has played the key role among aid instruments even though the share has decreased a bit over the years from 70 percent in 1980-81 to 60 percent in 1994-96 (figure 1.12 in Appendix 1.2). The three coun- tries have all been large recipients of adjustment loans. Kenya was in 18 Overview FIGURE 1.6 AID AND POLICY: MIXED REFORMERS Aid as a percentage of GDP Policy index 9 4.0 8 Policy (right axis) 7 6 3.0 5 4 2.5 3 2 2.0 1 AID/GDP (left axis) 0 l I 1.5 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 1994-97 fact the first country in Africa to get structural adjustment funding from the World Bank. Technical assistance has been very stable, around 16 percent of over- all aid. Kenya, with the highest relative share of TA among all countries in the sample, has received over US$3.5 billion of grant TA since 1970. Nonreformers Aid flows relative to GDP have been much lower for the Democratic Re- public of Congo and Nigeria than for the other countries in the study (figures 1.7 and 1.13 in Appendix 1.2). This is partly due to the two coun- tries' large populations and partly due to their extraordinarily poor eco- nomic management over the past 30 years. Aid flows to Nigeria have also been influenced by that country's oil income, with less aid provided during the oil boom years. Aid, though, has not been nearly as important a component of Nigeria's political economy as it has for other African countries. Debt rescheduling, includ- ing the rescheduling of private debt, has been more important. The Democratic Republic of Congo has made two genuine attempts Overview 19 FIGURE 1.7 AID AND POLICY: NONREFORMERS Aid as a percentage of GDP Policy index 9 4.0 8 7 3.5 6 3.0 5 4 Policy (right axis) - 2.5 3 2 AID/GDP (left axs)\ 2.0 0 1.5 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 1994-97 at economic reform since independence, in 1967 and 1983-85. The re- form program in the 1980s was supported by an IMF standby arrange- ment, but little additional support came from the donor community. Improved fiscal management allowed the Democratic Republic of Congo to pay its external debt, resulting in a net transfer of resources to its creditors. This gave strong ammunition to the opponents of fiscal rigor and the reform program was abandoned in late 1985. The donor commu- nity, reacting with delay to the net transfer problem, stepped up its as- sistance and net transfers became positive in 1986 when it was too late. Both countries have mainly received loans and multilateral aid, with Nigeria having the highest multilateral share of all countries in the sample. Nigeria progressively became more dependent on World Bank loans throughout the period as the country rejected IMF loans and their conditionalities. Technical Assistance and Nonfinancial Aid The case studies show that technical assistance works when it re- sponds to genuine demand and is well managed. Unfortunately, a lot 20 Overview of TA has not followed that concept. This is a great loss since TA has accounted for a large share of the aid to the countries studied, some 13 to 18 percent. TA in general has long had a low rate of success in annual World Bank (Operations Evaluation Department) evaluations. Even though the post-socialist countries have received as much as one-fifth of overall aid as TA, the recipients are surprisingly critical of technical assistance: A fundamental problem is that there has been no clear government policy on the role or limits of technical assistance. The government has finally [in 1997] decided to address the problem in a system- atic fashion. It has drafted a policy with wide-ranging proposals to restore better use of external resources allocated for technical as- sistance recruitment and to enhance ownership by the government in the management of these resources. (Tanzania: 300) But bilateral donors are reluctant to shift responsibilities to the recipi- ent government. They find that Tanzania is doing the right things in terms of policy reforms but that administrative capacity and competence is very low. Ethiopian policymakers express mixed feelings about some conven- tional aspects of technical assistance. They support technical assistance where knowledge transfer is programmed, the advice is grounded in Ethiopian conditions, and the country is given some freedom to buy ex- pertise in the free market as needed. However, they tend to look askance especially at those consultants who "parachute in." (Ethiopia: 209) Demand and management aspects have been the fundamental causes of successful TA. Uganda has made unusually good use of TA. One main reason for this has been that Uganda made greater use of TA when its commitment to reform grew stronger. The success of its reform program can to a large extent be attributed to nonfinancial assistance. Even if the overall record of TA to Africa is mediocre, policy dialogue and TA have in several cases had a substantial impact on economic policy formulation. The donor community has played a central role in support- ing key ministries and agencies and, increasingly, local research institu- tions. In Kenya, which has received the most technical assistance of all the countries studied, TA has undoubtedly influenced the analytical capabilities of Kenyan technocrats and institutions and their approach to analyzing eco- nomic issues.... Although outsiders may have influenced the pri- ority given to various economic policy issues, it is nevertheless clear that most of the reform agenda implemented over the past 20 years has been developed internally. (Kenya: 475, 506) Overview 21 Similarly, TA to the Central Bank and the Ministry of Finance in Zam- bia has provided ammunition to reformist technocrats. But the owner- ship of the reform process has been weak, and Zambia fits a more general portrait of a country in which policy choices are driven by donor fund- ing rather than domestically formulated policies (Zambia: 587). Nigeria exposes an interesting aspect of TA. The country has had the basic technocratic capacity to design reforms; TA has never been important in policy design. Any TA would have belied the government's insistence that the reforms were designed by Nigerians for Nigerians. The other nonreforming country, the Democratic Republic of Congo, has, on the other hand, received a fairly large share of its aid as TA (22 percent). In the early 1980s the donor community provided assistance to help run public institutions and the large public enterprise sector. A number of local technicians were trained in various sectors and be- came the backbone of the economy. Aid Coordination In many of the countries studied-all huge recipients of bilateral and multilateral aid-inefficiency in, or lack of, aid coordination over the years is remarkable. The main coordination mechanisms in most coun- tries have been the Consultative Group (CG) meetings and local coordi- nation meetings led by a lead ministry or the World Bank. In addition, the Special Program of Assistance for Africa (SPA) has definitely had a great impact on aid coordination, as have the Paris Club debt resched- ulings for countries with heavy debt. Aid coordination has improved considerably during the 1990s in many countries. The governments are increasingly playing a more active role, in Ghana and Uganda for instance, and the local donor coordination has been intensified. Consultative Group meetings are being held in the field. In Uganda the 1998 CG meeting took place in Kampala for the first time, with broad local participation of the president, cabinet ministers, non- governmental organizations, civil society, and the media. Tanzania is at present in the forefront of discussions on a change in the partnership between donors and recipient. The donors and the gov- ernment agreed on a list of actions in 1997 and there have been signifi- cant advances in a number of fields such as establishing government leadership in macroeconomic management. Nevertheless, the donors will probably still have a high degree of control of their own slices of the "joint" programs to make sure that the activities are in accordance with their list of priorities and that the money can be accounted for (Tanzania: 306). Donor coordination has been an ongoing challenge in Kenya given the multiplicity of multilateral and bilateral donors. Almost all Western 22 Overview donors have been active, putting heavy demands on the time of senior government officials. HAS AID SUPPORTED REFORM? Countries that have successfully reformed have had clear political move- ments leading to these changes, what is often referred to as local "own- ership" of reforms. Countries that have made less progress typically have had powerful vested interests blocking change. Either way, economic policies are primarily domestically grown. That was the main message of section 2. While countries have different political economies and go through different phases of reform, donors tend to do the same things every- where and at all times-with a few interesting exceptions that we high- lighted. But the overall similarity of donor behavior everywhere was one of the strong messages of section 3. In this section we bring aid and reform together and examine the ex- tent to which aid has influenced policy. The general finding is that dif- ferent instruments work in different phases of the reform process. If donors use the wrong instrument at the wrong time, it will be either a waste or downright counterproductive. Since Ghana and Uganda are the countries in the sample that achieved sustained good policy, we can use their experiences to trace out three phases of reform. Figure 1.8 shows the time path of economic policy for FIGURE 1.8 AID AND POLICY IN REFORMING COUNTRIES Policy index Aid as a percentage of GDP 3.5 5.0 4.5 3.0 4.0 2.5 Aid/GDP (right axis) 3.5 20 53.0 2.5 1.5 2.0 1.0 1.5 1.0 0.5 Policy (left axis) 0.5 0.0 0.0 t-3 t-2 t-I t-0 t-1 t-2 t-3 t-4 t-5 t-6 t-7 Overview 23 three successful low-income reformers. (Vietnam's experience has inter- esting parallels with those of Ghana and Uganda, so we have included it in this figure.) Each country has a "pre-reform" phase of very poor eco- nomic policy. Then there is a period of rapid reform, in each case associ- ated with a new government. The rapid reform begins in 1983 for Ghana, in 1987 for Uganda, and in 1991 for Vietnam. Within a period of three or four years there is very significant improvement in policy. After that the pace of reform slackens, but in each case the country has put in place policies that are well above average for low-income countries. All the other countries in our study have had poor policy before re- form too. But none has achieved sustained good policy. Where exactly to place countries in this framework can be a matter of some debate. For example, Zambia appeared to be embarking on serious reform with the advent of a new government in 1991, but implementation has fallen short of expectations. C6te d'Ivoire and Ethiopia are arguably in a phase of rapid reform now (that is, since 1994), but it remains to be seen if it will be sustained. Nevertheless, most of the categorizations are quite clear. None of the countries other than Ghana or Uganda progressed beyond a "pre-reform" phase through the end of the 1980s. It is also useful to ask, what signals the transition from one phase to another? How do we know if a country has started the rapid reform phase? In the case of the successful reformers, the beginning of rapid reform was marked by a new government coming to power. In addition, there are usually some dramatic, "stroke of the pen" reforms that are needed and that signal a real shift to reform. Stabilization, price liberal- ization, devaluation-these are the classic examples of dramatic changes that signal a new economic regime. As for the shift from the "rapid re- form" to the "good policy" phase, this occurs when all of the important macroeconomic reforms have been completed (the above plus fiscal re- form and trade liberalization), and then the country moves on to the second-generation reforms. Most of the second-generation reforms re- quire institutional change and reform of the public sector (privatization, civil service reform, judicial reform, and budget reforms to make public expenditures more effective and to increase accountability). In general, these reforms take more time and require more participation than the macroeconomic reforms. Aid in the Pre-Reform Phase What is distinctive about the pre-reform phase is that a country has poor policies and no coherent political movement to change the situation. Ghana and Uganda (and Vietnam too) are interesting in that during their pre-reform period they received small amounts of aid from Western donors (shown in figure 1.8), and most of this was in the form of technical assistance. This is not typical of donor behavior. All of the other countries 24 Overview received large amounts of aid in their poor policy periods, including quick-disbursing policy-based loans. (The different treatment for Ghana, Uganda, and Vietnam probably reflects the fact that their governments were politically estranged from the West during their poor policy peri- ods, and for them reform came hand-in-hand with political openings to the West.) What are some of the lessons about aid in the pre-reform pe- riod, gleaned from the different case studies? First, there is general agreement that technical assistance and policy dialogue are often-though not always-useful in this phase. When Ghana was dealing with macroeconomic crisis in the early 1980s, for example, it had well-trained economists to develop policy proposals, and these technocrats found the policy dialogue with the international financial institutions to be helpful in working out plans. A few years later, when Uganda's leaders were looking for new policies, one thing that helped was donor-financed study tours to Ghana. In the successful cases, political leaders learn from other countries and from their own mistakes. Low-key assistance can help with this policy learning, which generally has to take place at a country's own pace. Even in countries that remain in the pre-reform phase for a long time, technical assistance can lay the foundation for policy learning. In Kenya, for example, do- nors are supporting the Kenyan Institute for Policy Research and Analy- sis to help develop local capacity for policy analysis and formulation. This kind of capacity building is not going to have a large payoff as long as vested interests block serious reform, but it is an essential foundation if a political movement for change develops. A second common theme in the case studies is that conditionality does not work in the pre-reform phase in which there is no serious movement for change. Conditional adjustment loans were used in the pre-reform phase in all of the countries. The Zambia study argues forcefully that conditionality was not effective in bringing about change, and that the failure to reform could be tied to powerful vested interests: Western donor institutions had insisted since the early 1980s that market mechanisms should replace state controls. The demands and conditions became ever more pronounced and insistent. Zambia's ability to ignore these demands was limited as a result of its financial need and lack of alternative financial sources for se- curing foreign exchange.... In theory, therefore, the leverage of the Western financial institutions and donor governments was high. However, when one looks at the efforts to implement reforms in the 1985-87 period, the lack of influence of the external institu- tions in the face of unraveling agreements is striking. The imple- mentation of the 1985 structural adjustment program indicated that the international donors were virtually powerless and unable to Overview 25 interfere after the president changed his economic team in April 1986 from the group who had negotiated the program to some of the strongest critics of a market-based economy Experiences in Zambia in the 1980s highlight the importance of local ownership of the reform process: The conditionality mechanisms were unable to bring about policy change as these measures were at odds with the economic ideology prevailing at the time, and therefore were not supported by the political leadership. (Zambia: 572) In Zambia, policy failed to improve-and actually got worse-de- spite a long series of adjustment loans and a mounting volume of aid (figure 1.9). The Kenya study comes to similar conclusions about the effectiveness of conditionality during a period in which there was no strong commitment to reform: Did [structural adjustment lending with conditionality] induce the Kenyan government to adopt reforms that it might not otherwise have undertaken? Kenya received huge amounts of aid in return for policy reform agreements-almost US$3 billion over the entire 1970-96 period. How effective was this aid in "buying" reforms? We would argue that at times of severe economic crisis, as in 1980- 82 and 1993, the government's need for financial support was FIGURE 1.9 ZAMBIA: AID AND POLICY Policy index Aid as a percentage of GDP 1.2 12 1.0 - Policy (left axis) AID / GDP (right axis) 10 0.8 - -8 0.6 - -6 0.4 - -4 0.2 - 0 2 -0.2 l I l l 0 1966-69 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 26 Overview sufficiently desperate that the promise of such support did induce the government to come to agreement relatively quickly on far- reaching reform programs. However, as noted above, these agree- ments were not always implemented. Sometimes the probability of successful implementation was low from the outset. Other times the lenders or donors may have aligned themselves with well-in- tentioned technocrats who wished to achieve the results contracted for but lacked the political support to do so. It is our view, there- fore, that donor aid can have an influence on the form of agree- ment reached and on the agreed timetable for implementation, but whether implementation is carried out depends in the end much more on domestic political and economic factors than on donor money. (Kenya: 514) The impotence of conditionality in the face of powerful vested inter- ests against reforms comes through as well in the case studies of the Democratic Repubic of Congo, Ethiopia, and Nigeria: The dialogue with technicians and high officials was open and pro- ductive. There was a broad consensus on the reform agenda at this level, but, as became increasingly clear, President Mobutu had no intention of making any change in his ways of managing the coun- try. In his view, the reform program was fine as long as it did not impose any limits on his prerogatives. (the Democratic Republic of Congo: 640) Consistent with the experiences of many countries, in Ethiopia there appear to be few cases where conditionalities overcame stiff government resistance . . . Aid does not seem to buy good reform when the conditionalities are perceived, rightly or wrongly, to undermine the core of the regime's ideology or its power base. (Ethiopia: 205) The Babangida government attempted to address the complex is- sues that contributed to the impoverishment of Africa's largest coun- try, but failed to implement a coherent program. These failures came about because the logic of austerity and economic reform is anath- ema to the clientelistic system of Nigerian politics. (Nigeria: 674) The use of conditional loans in these situations is not merely waste- ful; it may actually be harmful. As noted above, Ghana, Uganda, and Vietnam were able to engage in policy learning without a massive aid relationship. Aid and conditionality may actually obstruct this policy learning, as in Zambia: In aid-dependent countries such as Zambia, donor conditionality undermines genuine policy learning. Once they understand that Overview 27 donors mean to set policy, ministries become passive. Individual officials have negative incentives to disagree with the donors since this will only serve to delay the arrival of the much-needed re- sources. (Zambia: 587) A third general point about aid in the poor policy period is that it may have perverse incentive effects. That is, finance may deter reform, and the absence of finance may encourage reform since it removes one easy way out of macroeconomic problems for the government. This argument is particularly applicable to large-scale budget or balance of payments support, which in a bad policy environment may reduce the urgency of reform. The C6te d'Ivoire, Kenya, the Democratic Republic of Congo, Nige- ria, and Tanzania studies all conclude that aid financing in the 1980s actually led to worse economic policies: The former colonizer [France] paid its ex-colony's debts to save it from default. It can be argued that by so doing, it relieved the pres- sure for structural adjustments, thereby retarding reform. (C6te d'Ivoire: 379) Is it possible that a large volume of aid could make it easier for a government to ride out a crisis without undertaking needed re- forms? It is probable that the heavy infusion of budget support Kenya received during the 1980s helped the government to finance the budgetary cost of an overstaffed civil service and inefficient public enterprises, thus enabling the government to defer reforms in these areas until the 1990s. (Kenya: 515) External support ... was indirectly used to enlarge the capacity of the regime to spend for nondevelopment purposes. The reform- ers who counted on the help of the international community had great difficulties fighting against the populist sentiments that were stirred up by the regime as a means to better maintain its staying power. (the Democratic Republic of Congo: 644) Of course, the money provided by donors gave the Babangida government some breathing room.... When these inflows com- bined with the additional oil revenue following the invasion of Kuwait in 1990, it allowed the Babangida government to try to revert to politics as usual. The government did not have enough money to return to the status quo ante but the immediate impe- tus for adopting the reform program was no longer there. (Nige- ria: 674) Initially aid probably delayed reforms by helping to finance schemes that would have been wholly nonviable without aid back- ing ... (Tanzania: 338) 28 Overview Conversely, several studies note occasions in which the absence of aid finance encouraged reform. In Ghana, Five factors help to explain why the PNDC chose to reform ... First, there was undoubtedly a liquidity crisis. Foreign exchange reserves, trade credit, and medium-term financing had all dried up . . . What gave the government political room to carry out the reform? Not many alternative approaches existed that would gen- erate the financial resources necessary for the economy's recovery. Few Ghanaians needed to be convinced of this fact; for the average citizen who had already suffered more than a decade of economic hardship, it was clear that the old order had failed. (Ghana: 79) The important point here is that in the absence of large-scale aid, the impact of poor economic policies was clear. The political leaders felt that they had little choice but to undertake serious reform. In Kenya too, re- forms such as liberalization of foreign exchange and foreign trade in the early 1990s were introduced when relations between the government and donors were strained and financing was in decline (Kenya: 478). To summarize conclusions about the pre-reform, poor policy phase: large-scale finance has, if anything, a negative effect, reducing the need to reform; and conditionality has typically failed in the absence of a seri- ous domestic movement for change. Technical assistance and policy dia- logue, on the other hand, have helped governments and their civil societies learn about policy from neighbors and from their own experi- mentation. It is noteworthy that the successful reformers all had modest aid programs in the pre-reform period. It should be stressed that the recom- mendation here is not to withhold all aid from countries with bad poli- cies. Rather, the point is that, in an environment of poor policies, large-scale budget support has sustained those bad policies and has not produced good outcomes. Aid in the Period of Rapid Reform The successful reformers have each had a period of three to four years in which very significant policy changes were effected. It is not easy to know in advance when a large reform movement is about to take off, but in most cases the change is brought about by a relatively new govern- ment. In the cases of Ghana and Uganda the new governments did not come to power intending to put market-oriented policies into place. But they did come to power intending to improve people's lives and they were open to learning and experimentation. Zambia is an interesting case as well, as a new government came to power in the early 1990s intending to put market-friendly policies into place, yet achieved only limited success. Overview 29 What can we say about the role of aid during the phase of rapid policy improvement? First, both the Ghana and Uganda studies are quite em- phatic that aid finance played an important supporting role. In the case of Ghana, balance of payments support "provided the government with the breathing space it required to contain domestic opposition to mar- ket-based reforms.... [It] allowed imports that helped fill the shelves of supermarkets and other traders. The filled shelves provided a psycho- logically-induced breather for the government because . . . people saw this as a sign of better things to come" (Aryeetey and Cox 1997). Referring back to figure 1.8, it can be seen that in the successful re- formers aid flows rose in lockstep with policy improvements. When coun- tries actually reform, finance increases the benefits of those reforms. That is, the growth impact of a particular improvement in policy is enhanced by the flow of aid. There are two reasons for this. Aid increases confi- dence in the reform program and calls forth greater private investment. Also, it enables the government to provide public services that are com- plimentary to private investment. By increasing the benefits of reform, aid enhances the likelihood that it will be sustained. As the Ghana study notes, Coming back to politics, economic reform proved politically sus- tainable in the end only because some results emerged quickly ... (Ghana: 67) While Ethiopia has not made as much progress with reform, its case nevertheless reveals the importance of aid in consolidating reform once it has started: Aid had a minimal impact on growth or poverty in the 1980s since Ethiopia was embroiled in a protracted civil war along with a bold program of building a socialist economy. Market-oriented reforms were initiated in 1990 in response to profound economic and po- litical crises. The reform program initially attracted a sizeable increase in external assistance in support of liberalization, stabili- zation, and rehabilitation. Subsequent aid helped to deepen reform commitments and supported high growth rates. (Ethiopia: 170) A second point about the rapid reform phase is that technical assis- tance and policy dialogue retain their usefulness. "Ownership" of re- form means that there is political commitment to change and able technocrats who can work out the details of reform. This layer of techno- crats is often rather thin, however, so that the training/policy advice part of aid is quite important. It helps deepen the layer of capable offi- cials and provides useful sounding boards for the technocrats who are working out the details of reform. 30 Overview When government officials are committed to reform, technical sup- port is often important to actually put reforms into place. A good ex- ample of this is the return of nationalized property to private citizens in Uganda, a highly politicized reform but one that was important in es- tablishing the government's commitment to private property rights. While success required support at the political level, it also required for- eign technical assistance to bring about a good outcome (Uganda: 128). Aid is particularly critical in the "rapid reform" phase because of the tendency for backsliding noted in section 2. When Ghana's macroeco- nomic program deteriorated in the early 1990s, donor support contin- ued because the government had some credibility as a reformer-and Ghana proceeded to undertake major structural reforms. Meanwhile the lack of donor support for debt reduction in the Democratic Republic of Congo's brief reform episode of 1985 may have contributed to that country's rapid reversal. The third important finding about the rapid reform phase concerns conditionality. If political leaders are committed to reform and techno- crats are implementing the changes-with technical advice as needed- one might think that there is no role for conditionality. But both the Ghana and Uganda studies agree that the technocrats actually welcome condi- tional assistance in this period. This is the one time at which conditional loans tend to be successful and useful. What technocrats are doing with conditional loans in this period is getting clear decisions from political leaders and signaling publicly the seriousness of reform. Thus, in the case of Uganda, The conditions helped reform-minded elements in the government push the reform agenda forward with the support of the president. The conditionality was therefore a helpful tool for generating, implementing, and cementing reforms. When the government was in this transition process many reforms were driven by the require- ments of conditionality. (Uganda: 136) In order for conditionality to play this useful role, the key measures that form the conditions for loan disbursement have to be ones that the recipient government truly supports. The Ghana study notes several rea- sons why conditional loans worked during that country's rapid reform phase: The first was the caliber, coherence, and continuity of the Ghana- ian team. Its relatively small size, highly qualified staff, and skill at implementing measures allowed some flexible interpretation of conditions. In addition, it appears that the Ghanaian team was usu- ally extremely well prepared. The second reason was the nature of the World Bank and (to a lesser extent) IMF teams. They proved Overview 31 over time adept at understanding the context in which certain de- cisions had to be undertaken and gave the Ghanaians some room. (Ghana: 86) Thus, to be useful, conditionality must reflect measures that the gov- ernment wants to carry out, not ones imposed by outside agencies. A fourth point about the rapid reform phase is more speculative. There are several pieces of evidence suggesting that thefailure of instruments in the pre-reform period limits their utility if a country ever gets serious about reform. Recall again that rapid reform in Ghana, Uganda, and Vietnam was not preceded by a period of large-scale Western aid and failed adjustment loans. Contrast this with Zambia, where a reform gov- ernment came to power in the early 1990s in an environment of aid de- pendency and with a long legacy of failed adjustment loans. After years of failure, it is hard to see how Zambian officials could have used condi- tional loans in the same manner that Ghana and Uganda did. Aid in Good Policy Regimes There are two points to make about the role of aid once good policy has been achieved-one about volume, the other about composition. Cross-sectionally, donors tend to discriminate against the poor coun- tries that have put good policy into place. There is also some element of this in the time-series relationship between aid and policy in Ghana and Uganda. Aid rose as policy improved, and as noted in the preceding section, this played a positive role in sustaining reform. However, once good policy was achieved, aid declined (figure 1.8). There are several factors at work here. For one thing, policy leveled off, which is almost inevitable after a period of rapid improvement. Donors like to think that their money is attached to policy change, and in Ghana and Uganda when policy change slowed down, aid declined. This behavior is not particu- larly rational given that the level of policies is much higher in these coun- tries than elsewhere in Africa. A second factor may be that donors think that these countries can now attract private capital flows. However, the evidence is that poor countries with good policies do not find it easy to attract private foreign investment. Perceptions of policies and of risk change slowly. Aid has a high return in Ghana and Uganda as it stimu- lates growth and poverty reduction. These are exactly the environments in which high levels of aid flows should be sustained. A second point about aid in the good policy environments is that do- nors continue to use conditional assistance, when in fact it has outlived its usefulness by this point. As noted, conditionality is useful during a period of rapid reform in which the government is trying to establish its credibility as a reformer and get macroeconomic policy measures well 32 Overview entrenched. But there is a cost to using conditionality in this way. To some extent, conditionality disguises the ownership of reforms. And it tends to limit participation in policymaking. The Uganda study argues that there was strong ownership of reforms by the government, but because of the continued use of conditionality by the donors, many stakeholders in civil society are confused about where reform measures come from. Activists in nongovernmental orga- nizations and private sector representatives interviewed for the case study thought that much of the policy package was imposed from out- side (Uganda: 145). If this view is widespread, it reduces the credibility of the reform program. People have a natural and understandable ten- dency to resent and resist outside pressure. Hence, if the reforms are believed to be primarily externally driven, it is harder for them to take root. A careful reading of the case studies suggests that there is a brief period during which the benefits of conditionality may outweigh the costs; however, once reform measures are well established it is impor- tant for the donor community to move away from conditionality so that the local management of reform is clear. In both Ghana and Uganda, donors have continued to use condition- ality well after the end of the rapid reform period. In fact, in Ghana con- ditionality became more stringent over time, as the overall quality of policies improved: * Conditions became more numerous. IMF conditions rose from 20 in 1983 to 40-50 in 1988-89. This reflected the Fund's increased involvement in structural issues, some of which the Bank regarded as its turf. * Conditionality-based lending became more important. The share of policy-based loans rose sharply in Ghana's overall loans from the Bank. * Conditions became tighter and deeper, and also increasingly spe- cific: for example, personnel cuts in the civil service were explicitly specified (Ghana: 86). Not only does conditionality disguise the ownership of reform, it also tends to limit the participation in decisionmaking. In Ghana, "the nature of the dialogue and Bank/Fund conditionality favored the use of a small group of technocrats for making decisions" (Ghana: 85). If each major dose of assistance requires detailed agreement on 40 to 50 policy mea- sures, then it is difficult to have wide debate and participation in eco- nomic decisionmaking. Finally, conditionality cuts against the grain of the second-generation reforms that even successful reformers find elu- sive. Public sector reform, for instance, requires a long-term commitment to institution building, something that cannot be achieved by condition- ality-based lending. The institution building of the second-generation Overview 33 reforms is also more likely to require wide participation, which as noted is limited by conditionality. Finally, as we stressed in the introduction, with some of the second-generation reforms we simply do not know what is good policy, and the issue for donors is to support genuine policy learning rather than to impose models through conditionality. CONCLUSIONS Donors have three basic instruments that they can use to encourage the adoption of good economic policies in developing countries: money, conditionality, and technical assistance/policy dialogue. The 10 case stud- ies in the Aid and Reform in Africa project reveal examples in which each of these has made positive contributions to countries' efforts to improve their policies. However, the striking impression that comes through is that donors have used these instruments fairly indiscriminately. That is, donors tend to provide the same types of aid in all cases, when in fact the instruments have different payoffs at different points in the reform process. Using the wrong instrument at the wrong time is at best wasteful, and in some cases may have actually retarded reform. Several of the case studies argue eloquently that aid could have been used more effectively to support reform: While the responsibility for the failure of the Democratic Republic of Congo as a state lies squarely with the political elite under Presi- dent Mobutu, which ruled the country until 1997, it is clear also that the interaction with the international donor community has not been as helpful as it could have been. The major donors were also the major creditors. By financing nonviable projects under com- mercial credit conditions in the early 1970s, and not providing ap- propriate debt relief ten years later when the country was in the process of reforming, they contributed to making adjustment diffi- cult to sustain. (the Democratic Republic of Congo: 629) Successful reform is undergirded by a strong commitment to good governance. This is essential for rebuilding business confidence, including among the growing Ethiopian diaspora. In the final analy- sis, political reform is a domestic affair. If the govemment man- ages to undertake the remaining reforms while building up its capacity for implementation, Ethiopia stands to benefit doubly from an increasingly selective international aid regime and a reinvigo- rated private sector. Donors can aid this process with a better cali- bration of aid and reform. (Ethiopia: 212) In summary, how should donors use their different instruments in order to have the maximum impact on policy change? 34 Overview Money If money had no impact on policy at all, the poverty-efficient allocation of aid would be conditioned on the level of poverty and the quality of policies: countries with a better level of policy should get more money This is also the allocation of money that is most likely to have a positive impact on policy. Giving large amounts of money to countries with poor policy has not stimulated reform. In fact, several case studies argue per- suasively that money in this situation allows a government to avoid re- form. The evidence is that in the poor policy situation, money does not provide broad benefits reflected in growth or poverty reduction. But it has to go somewhere, so presumably it allows the government to finance favored sectors and groups, maintaining support for the inefficient sta- tus quo. As the level of policy improves, the evidence is that money be- gins to have a broad positive effect on the whole economy. That is, it increases the benefits from the better policy environment. In Ghana and Uganda, this effect clearly helped sustain difficult reforms. Thus, money can help improve policies, but the key is to disburse it as actual policy improvement is achieved. In our two successful reform cases, finance increased in lockstep with policy gains. Conditionality All of the case studies agree that economic policy is primarily driven by domestic politics, not by outside agents. The key to successful reform is a political movement for change, and donors cannot do very much to generate this: What is needed is to develop new ideas about how Nigerian poli- tics should operate, something that the World Bank, the IMF, and the bilateral donors cannot provide. Only the Nigerians can do that. (Nigeria: 675) However, once reform movements developed in Ghana and Uganda, there was a brief period of rapid reform in which conditionality was welcomed by the government and was useful. It enabled a government that was serious about change to signal and commit to policy measures. To be effective in this way, conditionality must focus on measures that the government believes in and that are truly important. In Ghana and Uganda, the government itself proposed policy conditions to which it should be held. Using conditionality in other situations has not been productive. In the pre-reform phase in which the government is not committed to re- form, conditional loans have generally been a farce in which the govern- ment agrees to measures it does not believe in as a way to get funding, fails to carry them out, and then receives the funding from donors any- Overview 35 way. All of the case studies are very negative about this traditional use of conditionality to try to cajole governments to do things that they are really not interested in. These failed adjustment loans have discredited the instrument and made it less useful in the cases in which conditional aid actually could have been helpful. (To the extent that true reformers would like to signal that they are serious about reform, conditional loans are less useful as a signal if they are given indiscriminately to reformers and nonreformers alike.) Conditionality is also less useful once a country has achieved sustained good policy, as in the case of Ghana and Uganda. The ongoing use of conditionality disguises the true ownership of the re- form program, takes up valuable government time, and limits participa- tion in the debate and decisionmaking about economic policy. In these cases donors need to find simpler ways of transferring the financial re- sources that in fact have a high payoff in the good policy environment. Technical Assistance/Policy Dialogue Technical assistance or policy dialogue is an instrument that can be use- ful at any phase of the reform process. If it is well managed, it assists the country in its own policy learning process, that is, learning from other countries and from its own policy experiments. The case studies reveal that TA often is not well managed: it tends to be ineffective if it is sup- ply-driven from the donor side. But in each phase of the reform process there are groups inside and outside of government that are analyzing policy experiences and working out the details of reform, and these groups generally benefit from a dialogue with foreign experts. Where technical assistance has worked well, there has been genuine mutual learning on the part of the local counterpart and the foreign experts. In sum, foreign aid has often contributed to policy reform, but it could do so more consistently if donors were more selective in the recipients and instruments. With governments that have poor economic policies and no serious interest in reform, donors need to operate on a small scale, providing technical support where there is real demand (perhaps to groups outside of government). In this situation, large-scale finance and conditional loans have not been useful. The development of a re- form movement is primarily in domestic hands. If one develops, foreign aid can help consolidate it through financial support and conditionality. Conditionality outlives its usefulness fairly quickly, however, and has the problem that it may disguise domestic ownership of policies and limit participation in decisionmaking. In the low-income countries that have made good progress with policy reform, donors should find simple ways of providing finance without loads of conditionality. Most impor- tantly, rather than reducing support, as they have tended to do in the few countries that have achieved sustained good policy, donors should maintain a high level of finance in these productive environments. 36 Overview APPENDIX 1.1 DEFINITION OF AID The term "aid" in this paper is based on a measure called effective de- velopment assistance (EDA), recently developed by the World Bank. EDA is defined as the sum of the grants of all development flows disbursed in a given period. It covers grants and the grant equivalent of official loans. In this study we have also included grants tied to technical assistance in this measure. A second concept used is official development finance (ODF), which consists of EDA plus the nongrant component of official loans. The EDA measure provides a more accurate measure of true aid flows than the standard aid measure known as official development assistance (ODA). ODA consists of financial aid and technical cooperation, where financial aid includes grants and concessional loans having a grant ele- ment of at least 25 percent. The measure tends to overestimate aid flows since it includes the full face value of concessional loans. For further information on the definitions of EDA and ODF see Chang, Fernandez-Arias, and Serven (1999). Overview 37 APPENDIX 1.2 ADDITIONAL DEVELOPMENT FINANCE DATA TABLE 1.3 RELATIVE SHARES OF TYPES OF AID (OFFICIAL DEVELOPMENT FINANCE), 1980-96 (percent of total) Successfiul Reformers Technical Period Bilateral Multilateral Loans Grants assistance 1980-81 48 52 65 21 14 1982-85 28 72 69 20 11 1986-89 37 63 64 26 10 1990-93 46 54 53 34 13 1994-96 48 52 46 38 16 Average 42 58 57 30 13 Post-Socialist Reformers 1980-81 74 26 53 32 15 1982-85 77 23 57 29 14 1986-89 68 32 37 45 18 1990-93 57 43 27 53 20 1994-96 56 44 29 49 22 Average 65 35 39 43 18 Mixed Reformers 1980-81 49 51 70 13 17 1982-85 48 52 69 16 15 1986-89 55 45 52 30 18 1990-93 55 45 49 33 18 1994-96 45 55 60 28 12 Average 51 49 57 27 16 Non reformers 1980-81 55 45 64 12 24 1982-85 48 52 75 9 16 1986-89 39 61 73 12 15 1990-93 34 66 62 22 16 1994-96 22 78 58 25 17 Average 39 61 68 16 16 Source: OECD. 38 Overview FIGURE 1.10 TOTAL OFFICIAL DEVELOPMENT FINANCE FLOWS TO SUCCESSFUL REFORMERS, I980-96 Millions of 1996 constant U.S. dollars 1,800 Nongrant component of official loans 1,600 Grant equivalent of official loans - Technical assistance grants 1,400 - Non-technical assistance grants 1,200- 1,000 800 - 600 - <000 0:l7 400 200 0 9oo°~ ~~0 1P 9S IR PR 9p' io P@9t90 9 9 9o 9 9N9N9 FIGURE 1.11 TOTAL OFFICIAL DEVELOPMENT FINANCE FLOWS TO POST-SOCIALIST REFORMERS, 1980-96 Millions of 1996 constant U.S. dollars 4,500 3 Nongrant component of official loans 4,000 = Grant equivalent of official loans - Technical assistance grants 3,500 - - Non-technical assistance grants 2,500070 0 0 00 ;. l-00 __ 2,000 1,500 1,000 500 0 O? N N? 9oOt 1k 11S 9ap -11 -l -'q Nq _Iqt 9009 992 -lq -lq3 9N Overview 39 FIGURE 1.12 TOTAL OFFICIAL DEVELOPMENT FINANCE FLOWS TO MIXED REFORMERS, 1980-96 Millions of 1996 constant U.S. dollars 6,000 r Nongrant component of official loans = Grant equivalent of official loans 5,000 _ Technical assistance grants _Non-technical assistance grants| 4,000 3,000 -'/ \ 2,000 / 1,000 FIGURE 1.13 TOTAL OFFICIAL DEVELOPMENT FINANCE FLOWS TO NONREFORMERS, 1980-96 Millions of 1996 constant U.S. dollars 2,500 = Nongrant component of official loans Grant equivalent of official loans _ Technical assistance grants / 2,000 - Non-technical assistance grants 1,500 1,000 500 0\,9oOO , op ; 9 pp P9N p p I k 9 40 Overview NOTES 1. Grosh (1994) points out that clientelist networks unravel during a general- ized political and economic collapse, whereas during moderate economic dete- rioration, state elites are reluctant to encroach on rent-seeking coalitions. 2. Van de Walle (1994) and Lewis (1996) have described the typical African leader as seeking to placate key constituencies while marginally observing policy conditions to maintain the flow of external resources. REFERENCES Alesina, Alberto, and David Dollar. Forthcoming. "Who Gives Aid to Whom and Why?" Journal of Economic Growth. Aryeetey, E., and A. Cox. 1997. "Aid Effectiveness in Ghana." In Jerker Carlsson, Gloria Somolekae, and Nicolas van de Walle, eds., Foreign Aid in Africa: Learn- ing from Country Experience. Uppsala, Sweden: Nordic Africa Institute. Ayittey, George. 1989. "Political Economy of Reform in Africa." Journal of Eco- nomic Growth 99 (3): 4-17. Bumside, Craig, and David Dollar. Forthcoming. "Aid, Policies, and Growth." American Economic Review. Chang, Charles C., Eduardo Fernandez-Arias, and Luis Serven. 1999. "Measur- ing Aid Flows: ANew Approach." Policy Research Working Paper 2050. World Bank, Development Research Group, Washington, D.C. Collier, Paul, and David Dollar. 1999. "Aid Allocation and Poverty Reduction." World Bank, Washington, D.C. Processed. Collier, Paul, and Anke Hoeffler. 2000. "Greed and Grievance in Civil War." Policy Research Working Paper 2355. World Bank, Development Research Group, Washington, D.C. Dollar, David, and Jakob Svensson. Forthcoming. "What Explains the Success or Failure of Structural Adjustment Programs?" Economic Journal. Fernandez, Raquel, and Dani Rodrik. 1991. Resistance to Reform: The Status Quo Bias in the Presence of Individual-Specific Uncertainty. American Eco- nomic Review 81:1146-55, December 1991. Grosh, Barbara. 1994. "Through the Structural Adjustment Minefield." In Jenni- fer A. Widner, ed., Economic Change and Political Liberalization in Sub-Saharan Africa. Baltimore: Johns Hopkins University Press. Killick, Tony. 1991. "The Development Effectiveness of Aid to Africa." Policy Research Working Paper 646. World Bank, Intemational Economics Depart- ment, Washington, D.C. Lewis, Peter M. 1996. "Economic Reform and Political Transition in Africa: The Quest for a Politics of Development." World Politics 96(49): 99-129. Mosley, P., J. Harrigan, and J. Toye. 1995. Aid and Power. 2d ed. Vol. 1. London: Routledge. Overview 41 Ranis, Gustav. 1995. "On Fast Disbursing Policy-Based Loans." Department of Economics, Yale University. Processed. Rodrik, Dani. 1996. "Understanding Economic Policy Reform." Journal of Eco- nomic Literature 34 (1), March 1996. Sachs, Jeffrey. 1994. "Life in the Economic Emergency Room." In J. Williamson, ed., The Political Economy of Policy Reform. Washington, D.C: Institute for In- temational Economics. Tumusiime-Mutebile, E. 1995. "Management of the Reform Programme." In P. Langseth, J. Katobor, E. Brett, and J. Munene, eds., Uganda Landmarks in Re- building a Nation. Kampala: Fountain Publishers. van de Walle, Nicolas. 1994. "Political Liberation and Economic Policy Reform in Africa." World Development 94 (22): 483-500. van de Walle, Nicolas, and T. Johnston. 1996. Improving Aid to Africa. Washing- ton, D.C.: Overseas Development Council. World Bank. 1982. Accelerated Development in Sub-Saharan Africa: An Agendafor Action. Washington, D.C. . 1998. Assessing Aid: What Works, What Doesn't, and Why. New York: Ox- ford University Press. Successful Reformers Ghana Yvonne M. Tsikata Economic and Social Research Foundation Dar es Salaam, Tanzania ACRONYMS AND ABBREVIATIONS AFRC Armed Forces Revolutionary Council AGC Ashanti Goldfields Corporation/Ashanti Goldfields Company Ltd. BOP Balance of payments EDA Effective development assistance ERP Economic Recovery Program ESAF Enhanced Structural Adjustment Facility (of the IMF) GDP Gross domestic product GQ Grant equivalent of a loan IMF International Monetary Fund INCC Interim National Coordinating Committee JFM June Fourth Movement NATCAP National Technical Cooperation and Assessment Program NDC National Democratic Congress NDM New Democratic Movement NERC National Economic Review Committee NPP New Patriotic Party NRC National Redemption Council NUGS National Union of Ghana Students PAMSCAD Program of Actions to Mitigate the Social Costs of Adjustment PDC People's Defence Committee PNDC Provisional National Defence Council PNP People's National Party SAP Structural Adjustment Program SMC Supreme Military Council SOE State-owned enterprise SPA Special Program of Assistance (for Sub-Saharan Africa) TA Technical assistance TUC Trades Union Congress VAT Value added tax WDC Workers' Defence Committee A significant body of work in recent years has focused on the links between foreign aid and the economic performance of countries.1 Researchers have explored the roles that different types of aid may play in reform-financial versus nonfinancial aid, policy-based versus un- conditional aid. They have also examined ways to maximize aid effec- tiveness and the role of donors in achieving that goal. One troubling conclusion is that "recent cross-country evidence has shown that for- eign aid has a strong, positive effect on a country's economic perfor- mance if the country has undertaken certain policy and structural reforms. However, the evidence shows that aid in general has not been going to countries that have undertaken these reforms. Donors give less assistance to countries with good policies than to ones with poor or me- diocre policies" (Holmgren 1998). Ghana's experience with foreign aid supports this conclusion. Dur- ing the 1980s, when it undertook far-reaching economic reforms, Ghana was one of the most successful adjusters. Yet, it received relatively small amounts of aid compared to other African countries. In 1996 it received 2 percent of purchasing power parity GDP in aid, compared to 7 to 8 percent for Mali or Zambia and 4 to 5 percent for C6te d'Ivoire and Tan- zania. An important point is that during the 1980s, the volume of aid largely tracked the policy level-it was small when policies were bad, and increased when policies improved. In the 1990s, though, aid to Ghana has declined, while policy slippages-in the larger context-have been modest. Ghana is thus a particularly interesting case for understanding why countries choose to reform, and under what circumstances aid is associated with better economic performance. From a state of economic collapse, Ghana's economy rebounded with sustained economic growth during the first decade of reform. This stel- lar performance was accompanied by an exponential increase in aid in- flows from both bilateral and multilateral sources. The reform program was characterized by a high degree of ownership by top Ghanaian policymakers and the leadership, and was implemented by a high-cali- ber group of technocrats. The impetus for reform appears to have been both the economic crisis in which the country found itself and an evolu- tion in the thinking of policymakers as they strived for political survival. Aid, especially of the financial kind, was important in sustaining reform. Derailment of economic reform in the 1990s coincided with increased democratization and greater demands on institutional capacity. While economic reforms have resumed, the implementation of structural re- forms appears to have become more difficult. This paper examines the underlying reasons that Ghana chose to re- form and the role external assistance may have played in that decision. It analyzes economic policy, institutional and political developments, the main constituencies for reform, and other internal processes that 47 48 Successful Reformers shaped the reforms. It also evaluates the role of aid and donors in initiat- ing, catalyzing, and sustaining reforms. The analysis is based on inter- views carried out in Accra in March 1999, as well as on the available literature. Section 2 describes the different types of aid that Ghana receives and traces the evolution of aid flows. Section 3 places economic policy in Ghana in historical perspective and provides a brief overview of the reforms that took place during the 1980s. Section 4 analyzes the politi- cal-institutional framework within which Ghana pursued economic re- forms, and the reforms are analyzed in detail in section 5. Section 6 assesses the links between aid and reform and presents conclusions. AID: COMPOSITION AND TRENDS This study uses the concept of effective development assistance (EDA) presented in Chang, Fernandez-Arias, and Serven (1999). EDA is an ag- gregate measure of aid flows that combines total grants and the grant equivalents of all official loans. It is computed on a loan-by-loan basis to reflect the financial cost the creditor incurs in making loans on concessional terms. More precisely, EDA is defined as the sum of grant equivalents and grants, excluding technical assistance and any bilateral debt forgiveness. This adjusted measure uses the same conventional grant data but aggregates grant equivalents of loans (GQ) rather than the full face value of all loans deemed concessional. The grant equivalent of a loan is defined as the difference between the present values of the loan's disbursements and stream of expected debt service payments, or the resources that borrowers receive in excess of their interest and repay- ment obligations. Conceptually, for each loan, this amount is equal to the net loss to the lender-equivalent to a grant. The computation of grant equivalents involves discounting two cash flow streams for each loan's disbursements and interest/amortization payments. It uses dis- count rates that reflect both the term structure of each loan and the mar- ket conditions specific to the loan's currency of issue at the time of valuation. Finally, the paper defines a broader concept of aid, total aid (TOTAID), as EDA plus technical assistance. Patterns and Trends in Aid The overall trend in aid to Ghana, as measured by effective develop- ment assistance, reflects the country's economic and political history. In that sense, aid flows to Ghana may be said to have been endogenous. Aid flows remained at a low level in the 1970s, a period characterized by chronic domestic economic mismanagement. A default on foreign loans Ghana 49 FIGURE 2.1 TRENDS IN AID TO GHANA, 1970-96 Millions of U.S. dollars 900 800 Total aid (TOTAID) 700- 600 - Effective 500 - development assistance 400 ( 300 r 200 - Technical assistance (TA) 100- 0 -,--.r r---¶ --- Source: World Bank. by the military government in 1972 further discouraged foreign assistance.2 With the emergence of a democratically elected government in Sep- tember 1979, aid flows rose for two consecutive years. This trend re- versed after the 1981 coup d'etat by the armed forces. Starting in 1985, however, a clear and sustained increase in aid flows occurred as donors perceived greater commitment by government to better economic man- agement and economic reform. Indeed, between 1985 and 1996 total aid flows to Ghana tripled, from US$150.7 million to US$450.8 million in 1995 (see figure 2.1 and table 2.1). However, during the 1980s Ghana was still receiving less aid than most other African countries. The espe- cially rapid increase between 1990 and 1991 was linked to upcoming multiparty democratic elections and was driven primarily by increased grants to support various institution-building activities. Following the elections, total aid fell, but still remained higher than pre-election levels. The drop was due to fiscal "slippage" in the reform program. This was linked to an 80 percent increase in wages to civil servants, among other factors.3 The immediate consequence was a sus- pension of World Bank disbursements between November 1992 and the middle of 1993. This episode was short-lived, and by the end of 1993 both the World Bank and the International Monetary Fund (IMF) were disbursing funds, programs were back on track, and flows resumed to pre-interruption levels. Additional fiscal slippage in 1996, however, led 50 Successful Reformers TABLE 2.1 GHANA: POLICY-BASED LOANS FROM MULTILATERAL INSTITUTIONS, 1983-96 IMF Amount (millions of special Loan type Year drawing rights) Standby 1983-84 238.50 Standby 1984-86 180.00 Standby 1986-87 81.80 Extended Fund Facility 1987-90 245.40 Structural Adjustment Facility 1987-90 129.86 Enhanced Structural Adjustment Facility 1988-91 368.10 Enhanced Consultation 1992-95 0.00 Enhanced Structural Adjustment Facility 1995-99 164.40 Total 1,408.10 World Bank Amount (millions of Loan type Year US. dollars) Reconstruction Import I 1983 40.00 Export Rehabilitation I 1984 17.10 Export Rehabilitation II 1984 40.12 Export Rehabilitation III (TA) 1984 17.10 Reconstruction Import II 1985 60.00 Health and Education 1986 15.00 Industrial Sector I 1986 24.95 Industrial Sector II 1986 28.50 Structural Adjustment 1987 80.9 Education Sector 1987 34.5 Structural Adjustment 1987 14.66 Agricultural Services 1987 17.02 Structural Adjustment 1987 34.00 Structural Adjustment Institutional Support 1988 10.80 PAMSCAD 1988 10.60 Financial Sector 1988 100.00 Public Enterprise Assistance 1988 10.50 Cocoa Rehabilitation 1988 40.00 Education Sector Adjustment II 1990 50.00 Structural Adjustment III 1991 13.0 Financial Sector Adjustment II 1991 100.00 Agriculture Sector Adjustment Credit 1992 80.00 Private Sector Development 1994 13.00 Sources: Armstrong 1996; World Bank 1998b. Ghana 51 FIGURE 2.2 GHANA: PERCENTAGE OF EDA LOANS BY CREDITOR, 197o-96 Percent 100--- - - 60 40 20- 9eXo~ ~ \' ,k, 0&w 10 y 0, + 9 A 909 9 - Multilateral Bilateral Source: World Bank. to a temporary derailment of the IMF-supported program under the Enhanced Structural Adjustment Facility. Multilateral aid has risen dramatically as a share of effective devel- opment assistance to Ghana since the late 1970s, when it accounted for less than 10 percent (see figure 2.2). The most dramatic increase occurred in the mid-1980s as the World Bank and IMF supported Ghana's eco- nomic reform efforts with a series of adjustment loans and facilities. Be- tween 1983 and 1984 the multilateral share of total aid doubled, to 79.1 percent. While this share has fluctuated since then, it has never fallen below 71 percent; in 1996 it was 85 percent. The increase in multilateral aid relative to bilateral reflects smaller aid budgets for the bilateral do- nors as well as competing demands from Eastern Europe. Aid given to different sectors has evolved over time, reflecting both the donors' interests and changes in government priorities. In the first 52 Successful Reformers TABLE 2.2 PHASING AND SEQUENCING OF REFORM POLICIES IN GHANA, 1983-96 Stabilization Adjustment Oscillating reform Area of reform (1983-86) (1986-91) (1992-96) Pricing reforms * Currency devalu- * Market-based * Licensing of ation foreign exchange foreign * Removal of price auction exchange controls * Interest rate bureaus * Wage restraint liberalization * Rationalization or deregulation of energy and utility (infrastructure) prices Fiscal policy * Tax reform * Medium-term * Medium-term * Consumer expenditure expenditure subsidies planning framework * Producer subsidies Structural * Tariff regime * Financial sector u Cocoa price to policies simplified and reforms farmers raised made more u Investment * All cocoa uniform promotion subsidies for * Reduction of * Divestiture of production negative list SOEs removed * Removal of u AGC floated import controls * Additional SOEs sold/ liquidated Institutional * Management * Creation of reforms of state enter- multisector prises regulatory * Civil service agency reform * Coordination * Capacity build- of donor ing in core support for ministries public sector * Planning management process * Statistical and information sys- tem improvement Sources: Armstrong 1996; various World Bank and government of Ghana documents. Ghana 53 two years of the Economic Recovery Program, aid served mainly to fi- nance imports (see table 2.2). For the rest of the decade, however, aid was targeted at balance of payments (BOP) support and the transport sector, the latter reflecting the tremendous rehabilitation needs. Between 1986 and 1989, balance of payments support averaged just over 50 per- cent of total aid. During the 1990s, BOP support continued to be impor- tant but aid increasingly went to community and social services; between 1993 and 1996 this sector averaged 37.1 percent of total aid. Both agri- culture and the transport sector (again) were important recipients of aid in the late 1990s as well. Technical assistance (TA) to Ghana declined during the first years of the reform program (see figure 2.3). This probably reflected the fact that donor support initially focused on essential BOP support for imports and reconstruction. Starting in 1986, TA rose steadily, though it remained a stable or declining share of overall aid until 1990. In the 1990s TA rose considerably, both in dollar terms and as a share of overall aid, account- ing for about one-fifth of the total. This trend can be linked to a growing realization that the efficiency with which aid is used and the viability of aid-funded projects depend on strengthened capacity in key institutions. FIGURE 2.3 COMPOSITION OF TOTAL AID TO GHANA, 1970-96 Percent of total aid 100 90 Effective development 80 assistance 70 (EDA) 60\ 50- 40- 30 Technical 20 10- 0 Source: World Bank. 54 Successful Reformers Multilateral Institutions The World Bank and the IMF have been leaders in providing external support for Ghana's reform efforts. Due to Ghana's perceived commit- ment to reform and its good results early on, financial support from the Bank and Fund steadily increased. The two institutions collabo- rated significantly, especially in the early 1980s. The World Bank re- sumed Consultative Group (CG) donor meetings for Ghana in 1983, and has convened them annually since then to help mobilize large vol- umes of external assistance for the country. In addition to the CG meet- ings, the Bank organized a social sector donors' meeting in Vienna in 1986. A meeting of donors to PAMSCAD-the Program of Actions to Mitigate the Social Costs of Adjustment-was held in Geneva in 1988. The biannual meetings of the Special Program of Assistance for Africa (SPA) are used to mobilize aid and coordinate activities. The Bank also took the lead until recently in fostering communication and coordina- tion among the donors. Increasingly, however, the government is play- ing this role. Earlier assessment of the World Bank's role in the 1996 Country As- sistance Review, as well as recent structured interviews with donor agen- cies in Accra, indicate that overall the Bank has been successful in its coordination efforts. Nonetheless, the plethora of donor projects and in- terests means that even stronger government aid coordination efforts are required. World Bank lending initially focused on rehabilitating Ghana's dete- riorated infrastructure, as well as on BOP support. Adjustment lending emerged in response to the need for deeper structural reforms and spanned a number of sectors. Between 1983 and 1994 the Bank commit- ted US$2.4 billion, of which US$1 billion, or about 40 percent, consisted of adjustment lending (Armstrong 1996: 36). If sectoral adjustment lend- ing is included, the adjustment-lending share increases to 42 percent. Table 2.1 gives the list of policy-based loans by the World Bank and the IMF to Ghana since 1983. These adjustment loans have had a mixed record. In general, how- ever, and certainly during the first years of reform, internal World Bank reviews found the adjustment loans to be performing better than the rest of the portfolio. Agreed policy actions were almost all taken, and with little delay. Agreed studies were carried out, albeit with more delay. The reasons for the success of the early adjustment cred- its seem to be high commitment on all sides and the administratively simple steps required. The Ghana country assistance review showed that up to 1994, approximately 80 percent of policy-based loans re- ceived a satisfactory rating. This was higher than the Bank-wide Ghana 55 average of 73 percent, and considerably higher than the 59 percent reported for the Africa region. Bilateral Aid Bilateral aid is given for political, strategic, humanitarian, or commer- cial reasons. It may be intended to meet foreign policy objectives, main- tain a historical (usually previously colonial) relationship, or create commercial opportunities. The motivation tends to differ across donor countries, and it is often reflected in the sectors where donors are active and the form in which they choose to give aid. During the 1970s the most important donors in Ghana were the United Kingdom, the United States, and Canada. Some changes have occurred since then in the ranking of bilateral donors by importance. Since the early 1980s, the United States has fallen to second place, and Japan has been the largest donor for about a decade. Canada, Den- mark, France, Germany, the Netherlands, Spain, and the United King- dom and also rank high by volume of aid. The rationale for Japanese aid is based on Ghana's importance in the region, its good economic record, and its progress with democratization. In terms of sectors, the Japanese have been most active in infrastructure (fishing harbors, highways, bridges, rural water supply, and electrification) and food aid. One im- plication of this rise in Japan's importance is that since Japan is not in- volved with policy-based lending, the Ghanaian government has access to a large volume of nonconditional lending. The United States had a low-level program between 1982 and 1992, reflecting political tensions between Washington and Accra.' With in- creased democratization in Ghana, its good economic record, and the end of the Cold War, the United States has again become an important bilateral donor. The new U.S. aid paradigm places emphasis on contain- ing and reducing subregional conflicts. U.S. aid goes increasingly to coun- tries in conflict, and tends to taper off as countries do better. The United States is also more interested in the comprehensive framework for re- form than previously Washington has always supported the private sec- tor, but now also examines other indicators of competitiveness such as the exchange rate and inflation. Finally, U.S. agencies bring actors to- gether-Ghanaians with Ghanaians, and Ghanaians with international investors. For example, in October 1997 the U.S. Agency for Interna- tional Development organized a well-attended conference in North Caro- lina on investing in Ghana. Several bilateral donors have provided co-financing for adjustment programs. While bilateral donors often tie their support to either World 56 Successful Reformers Bank or IMF conditions, several of them also negotiate their own bench- marks with the government (for example, Canada). Donor Coordination As indicated above, there has historically been a close relationship be- tween the World Bank and other donors in Ghana, with the Bank play- ing a leadership role. After 1986, the Bank's adjustment operations typically attracted substantial co-financing from other donors. In addi- tion to the CG and SPA meetings, the World Bank currently leads the Head of Agencies Meeting in Accra. There are donor sectoral groups and donor lunches, to which ministers and other government officials are invited. Finally, the Bank helps to coordinate the sector investment programs in roads, education, and health. The quality of donor coordi- nation varies by sector. In some, individual interests mean there is still much confusion. Particularly close collaboration is now evident within the United Nations system in both program planning and operations implementa- tion. The World Bank and U.N. agencies have completed a joint "com- mon country assessment" that provides the overarching framework for the collaboration. Increasingly, however, the government is taking the lead in setting the agenda for aid/donor coordination meetings on a regular basis. The International Economic Relations Division in the Ministry of Finance is responsible for managing the country's external aid. The Bank of Ghana and the controller and accountant general's office are also important in monitoring aid flows. Problems with disbursement information have ameliorated over the years, but they remain an issue. Several donors forgave debt or converted debts to grants in the late 1980s and early 1990s. The impact of this debt relief is estimated to have averaged US$30 million from 1990 to 1994. ECONOMIC REFORM IN PERSPECTIVE The uninterrupted rule of Jerry Rawlings from the end of 1981 to the present would seem to provide a neat period for analysis. However, one must delve into the late 1960s and the 1970s to understand the economic quandary in which Ghana found itself by 1981 and the nature of the alliances that emerged following the 1981 coup d'etat. This provides background for analysis of the economic policies the government fol- lowed during 1982-96. Ghana 57 Antecedents to the 1981 Macroeconomic Crisis State-Led Development in the First Republic, 1957-66 At independence in 1957, Ghana was one of the brightest stars in Africa. It enjoyed one of the highest per capita incomes on the continent and was known for its relatively large and well-educated middle class. Eco- nomic growth was respectable, averaging 4 percent during the 1960s. This growth was heavily driven by Ghana's most important export, co- coa, which accounted for about 70 percent of foreign exchange earnings. The country was the world's largest producer of cocoa, averaging about a third of world supply in the late 1950s (Killick 1978). Heavy government intervention in the economy after independence, however, adversely affected economic health. The role of the public sec- tor was greatly enlarged on both the productive and distributive sides. Increasingly Prime Minister Nkrumah carried out more dirigiste poli- cies, and in 1961 he officially introduced socialist state-led planning. Deteriorating terms of trade compounded the impact of these policies. Between 1959 and 1965, the world market price of cocoa fell substan- tially; in fact the 1965 price was only 40 percent of the 1958 level. For all these reasons, large fiscal deficits emerged, the balance of payments de- teriorated significantly, and the economy took a nosedive. Political Turmoil and Short-Lived Reforms, 1966-72 Following a military coup d'etat in February 1966, a new economic team was put in place and attempts were made to reform the economy (see appendix 2.1 for a chronology of events). The National Liberation Coun- cil (NLC) reopened negotiations with the IMF, reaching agreement on a standby arrangement in March 1966. The currency was devalued by 43 percent in July 1967, public expenditure was reduced, and the activities of public enterprises were streamlined or eliminated. To address bal- ance of payments difficulties, import controls were tightened and for- eign debt payments were rescheduled (Dordunoo and Nyanteng 1997). Dr. Kofi Busia, a former academic who headed the new civilian gov- ernment elected in 1969 (the Second Republic), continued these economic reforms. Improving terms of trade enabled the government to reduce significantly the number of goods on the negative list for imports. With a fall in the world price of cocoa in 1971, however, rapid import liberal- ization was no longer affordable. Both the fiscal and capital accounts deteriorated. In late 1971, the cedi was devalued by 44 percent in an attempt to improve the balance of payments position. Whether this would have achieved that objective will never be known, because less than a 58 Successful Reformers month after the devaluation, on January 13, 1972, a military coup d'etat led by Colonel Ignatius Kutu Acheampong overthrew the Busia regime. Economic Chaos, 1972-79 Under Acheampong and his National Redemption Council (NRC)-later the Supreme Military Council (SMC)-the misguided policies of the First Republic were reintroduced and, indeed, intensified. The NRC revalued the currency by 26 percent, completely at odds with the indications from the parallel market. They imposed stringent import controls, introduced additional price controls, and unilaterally suspended foreign debt servicing. While favorable terms of trade initially resulted in good economic performance, by 1974 economic decline and decay had begun to set in. Consumer price inflation skyrocketed from 9.6 percent in 1972 to 77.2 percent by 1977. As the exchange rate remained fixed, the cedi was in- creasingly overvalued, discouraging exports. Government expenditures rose as the state became more heavily involved in production and regu- lation. These expenditures were financed by complex and high rates of taxation on imports, goods and services, and exports. Ghana experienced continued stagflation in the second half of the 1970s. More pernicious, perhaps, was the widespread corruption and moral decline that emerged during the rest of the decade. The roots of the cor- ruption lay in the system of import licenses and the high rates of taxa- tion on goods and services. High taxes encouraged rent-seeking behavior (Jebuni, Oduro, and Tutu 1994), while the issuance of chits (associated with the release of goods from state-owned factories) and import licenses became a widespread form of state patronage. During the 1970s, the phrase kalabule entered the Ghanaian lexicon.5 It referred to widespread corruption or a licensed beat-the-system approach to survival (Chazan 1983: 194-97). The regime had an especially damaging impact on the country's in- stitutions. To begin with, corruption had weakened most state institu- tions. Civil service employees had to engage in pervasive moonlighting because of inadequate pay. Educated Ghanaians sought better-paying jobs elsewhere in West Africa, as well as in North America and Europe. This brain drain further weakened the universities and other institu- tions. The diaspora of Ghanaian academics and professionals and the weakening of institutions were to have serious consequences when eco- nomic reforms were finally initiated. As pressure rose from various groups (Association of Professional Bodies, People's Movement for Freedom and Justice, Front for the Pre- vention of Dictatorship), Acheampong attempted unsuccessfully to form a coalition government of civilians, the military, and the police known Ghana 59 as UNIGOV. Under increasing pressures from junior army officers well as civil society, Acheampong's colleagues in the SMC forced him to re- sign on July 5, 1978. Lieutenant-General Fred Akuffo took over as head of the revamped Supreme Military Council (SMC 1I). While some senior officers were dis- missed, SMC II prosecuted no one and the UNIGOV concept was still presented as the only model for a transition to "civilian" government. On the economic front, SMC II held preliminary discussions with the IMF but was unable to conclude negotiations. Following widespread strikes in almost every sector of the economy, however, Akuffo finally announced the "unbanning" of political parties on January 1, 1979. Old political groupings and alliances were resuscitated. The People's Na- tional Party took over the Convention People's Party/Nkrumahist mantle while the Popular Front Party was the new version of Busia's Progress Party. Assorted smaller parties also entered the fray. General and presi- dential elections were set for June 18, and Ghana seemed on a smooth path to democratic rule again. The lower ranks of the armed forces were apparently unhappy with the failure to prosecute SMC officers, even after evidence of corruption. On May 15 Flight Lieutenant Jerry Rawlings and a small group of sol- diers conducted what would later be termed a "mutiny or uprising" against the officers. They were unsuccessful and were jailed. Rawlings and the others were subjected to a public trial during which it appears they gained more support from the army rank and file. On June 4, a group of soldiers sprung Rawlings from prison and the "coup" was launched. The new Armed Forces Revolutionary Council (AFRC) indi- cated that the elections would still go on and stated their intention to hand over power to the elected civilian government in September 1979 as planned. In the meantime, the next three months would be used to "clean house." On the economic front, the AFRC reimposed price controls on consumables, raised the cocoa price paid to farmers, evacuated some of the rotting cocoa from rural areas with the assistance of students, and collected large amounts of outstanding tax payments during this pe- riod. While some of these measures were successful in the short run, many would prove unsustainable (for example, the price controls). The Limann Era (Third Republic), 1979-81 The new civilian administration of the People's National Party (PNP) led by Dr. Hilla Limann inherited a difficult economic situation. The massive economic decline required tough decisions. It was perhaps not surprising that the new civilian government, having been out of power for several years, seemed to take a long time to make these decisions. 60 Successful Reformers Shillington (1992) reports that some ministers at the time also indicated that they found it difficult to work with a "civil service not used to work- ing within the confines of a strict constitution." In general, the percep- tion was that the administration did not have an economic vision to pull the country out of the malaise. This was illustrated vividly when its bud- get to Parliament in July 1981 was defeated-a first in independent Ghana's short history. The situation was also tricky on the political side. Despite the new president's professional credentials, the fact remained that he was a rela- tively new face in Ghanaian politics.6 According to Shillington, although the fact that Dr. Limann was not affiliated with any particular political base was a political advantage in the run-up to the election, it probably became a liability once he was in power. In particular, Limann was be- holden to the old party hands who had worked hard to ensure his elec- tion.7 By the end of the year, the impression was that the government had come to a "virtual standstill because of the in-fighting" within the PNP.8 Increasing public dissatisfaction aggravated the economic and politi- cal difficulties.9 Union strikes increased in all sectors and the incidence of armed robbery (until then uncommon) rose. Left-wing groups such as the June Fourth Movement and the New Democratic Movement be- came focal points for opposition to the regime. Under these circumstances, it was perhaps not surprising that the coup d'etat of December 31, 1981 took place. Rumors had been circulating for several weeks in military circles (Shillington 1992: 79). Nonetheless, the immediate rationale for the coup was not purely economic but encom- passed moral and populist reasons as well. This was particularly evi- dent in the statements of Flight Lieutenant Jerry Rawlings and members of the new Provisional National Defence Council (PNDC) at the time. Economic Policy Dr. Limann and his government, then, came into power under some- what inauspicious circumstances. In addition, as outlined above, the general economic situation was desperate and financial liquidity pre- carious. Lax fiscal and monetary control over almost a decade had re- sulted in hyperinflation. Inflation worsened over the next two years, reaching 77 percent by 1981. Production had contracted in all sectors of the economy. Cocoa exports, which were Ghana's leading foreign ex- change earner, had declined, in part because of the absence of basic trans- port infrastructure to move the harvest to the ports. The shortage of foreign exchange meant that critical spare parts and inputs were in short supply, adversely affecting the performance of the industrial sector. Not surprisingly, both social and physical infrastructure had deteriorated severely. Ghana 61 Following the coup, the PNDC was quick to recognize the seriousness of the situation. It was less quick to act, however. This reflected the inter- nal tug-of-war at the time between various supporters of the regime over the direction of economic policy (discussed farther on). When the struggle was over, a far-reaching program of economic stabilization and reform was introduced. For the rest of the 1980s, Ghana continued to reform its economy and indeed became a leading example of a successfully adjust- ing country. During the 1990s this momentum slowed somewhat. Stabilization and Reform, 1983-91 In 1983 the PNDC introduced an Economic Recovery Program (ERP) to stabilize and liberalize the economy. The ERP and its follow-up, the Struc- tural Adjustment Program (SAP), received substantial assistance from the international financial institutions and donors. Rehabilitation of the country's deteriorated ports, roads, and railway was given priority early in the program. Input and produce marketing of most crops was liberal- ized over the reform period. In general, price controls on goods and ceil- ings on interest rates were removed. In the decade following the introduction of the ERP, real GDP growth was impressive, averaging 5 percent annually. Real income grew by 2 percent per capita during the decade. Inflation remained high and variable, however, discouraging private investment. Oscillating Reform, 1992-96 The mostly successful adjustment was partially derailed in 1992, in part because of an election-related wage increase. With loss of fiscal control came macroeconomic instability, reflected, for example, in higher infla- tion. Much of the 1990s were spent trying to regain sustained fiscal bal- ance. Not surprisingly, the decade was marked by increased difficulties between the government and the Bretton Woods institutions as policy slippage occurred. On the positive side, accelerated government efforts to increase investment paid off in the form of higher investment rates and greater foreign investment. INSTITUTIONAL FRAMEWORK The institutional features of Ghana, as well as political developments during the period under review, influenced the economic reform pro- cess. To understand how, it is necessary to examine aspects of the under- lying political culture, the political economy of reform, and the role of democratization and various interest groups. 62 Successful Reformers Political Culture From the time of Ghana's independence until the 1981 coup d'etat, the state (and by extension the politicians who ruled the state) was viewed as a provider. This concept, which existed under Nkrumah, was linked to the idea of the head of state as a chief who must take care of his subjects- in essence, a "provider-in-chief." There was a general expectation that the state would provide jobs directly, provide subsidies for enterprises, or even provide free social services."' Moreover, at a personal level there was an expectation that members of a family who were well off would provide for others less fortunate. Politicians were no exception to this rule. During the Acheampong era, this concept evolved into a form of state patronage in which import licenses, chits, and even cars were the give- aways."1 The beneficiaries were executives of state-owned enterprises (SOEs) and their families and friends, but the largesse was not restricted to them. Even employees of SOEs benefited because in lieu of adequate pay, the government gave them goods that they were expected to sell on the black market. During the 1980s this system was overturned. The December 31 coup was clearly targeted at those who had been benefiting from kalabule. Rawlings felt very strongly that this was a moral issue, and from his speeches, it is clear that he felt these kalabule "entrepreneurs" were against the people. The populism of the PNDC made it clear where they stood: with the people. How did these developments affect economic reform and decisionmaking? The evidence suggests that they helped in the begin- ning. Clearly, all the rent-seekers and others who had benefited from the system were a ready-made anti-PNDC group. However, the PNDC had effectively, and decisively, broken links with those groups. Hence, they were not a factor initially. Chazan (1991) points out that between 1983 and 1991, the state was insulated from strong interest group demands. She is referring in particular to what she terms the "postcolonial elites" with whom the PNDC had broken. The relationship previous govern- ments had maintained with the urban elite (consulting them, making them part of the decisionmaking process, and ensuring that they were kept happy) was shattered. In its place stood the "people" as represented by the People's Defence Committees (PDCs) and Workers' Defence Com- mittees (WDCs). While this should have complicated decisionmaking in principle, as discussed earlier, by early 1983 decisionmaking was quite centralized, carried out by a small group of technocrats. The Political Economy of Reform, 1983-91 A key question is when and why economic policy reform followed the path it did in Ghana, and why it faltered in the early 1990s. This requires Ghana 63 an understanding of the evolution of economic thought in Ghanaian eco- nomic policymaking. The discussion draws on the extensive literature on Ghanaian politics in the past two decades as well as on interviews carried out in Accra in March 1999.12 Support for the regime was especially strong among those who felt they had not benefited under the previous regime. The establishment of People's Defence Committees and Workers' Defence Committees that excluded management also reflected the gap between the "haves" and "have-nots." The resulting array of coalitions in support of and against the govern- ment was predictable. Urban workers, students, and lower ranks of the armed forces were especially strong supporters during the first years of the regime. The workers and students had been highly vulnerable to the dramatic decline in living standards during the 1970s. In addition, the unions had clashed with the Limann government during the Third Re- public over better pay and working conditions."3 With respect to labor, Herbst (1991) notes that the PNDC was against the trade union leader- ship. It perceived that leadership to have sold out its constituents to pre- vious leaders. According to Herbst, the WDCs were an effective mechanism for reaching union members outside the normal channels. University students were also an important supporting constituency at first. Two major groups emerged initially from the student body: the June Fourth Movement (JFM) and the New Democratic Movement (NDM). As discussed in Shillington (1992), the NDM was formed in 1980 and consisted primarily of students and academics. The JFM was also composed entirely of students to begin with, but broadened to include militant union leaders as well as community groups. Ideologically, both of these groups were viewed as neo-Marxist. Finally, the lower ranks of the army were an important supporting group. In large part because of the June 4 uprising, when he took full responsibility and absolved his army colleagues, Rawlings was perceived as a "ranks" man. The most vocal opponents of the new regime were what could be viewed as the "establishment"-those in the middle and upper mana- gerial classes that the regime was criticizing vociferously. They included persons who had benefited through their connections from the exten- sive system of chits, import licenses, and kalabule more generally. Among them were former politicians, traders, and executives in state-owned enterprises, professionals (including the Ghana Bar Association), and entrepreneurs. While there was a sense that little economic progress had occurred under Limann, redressing this was initially less important in the coupmakers' minds. This was because real concerns existed about the consolidation of power militarily in the face of both domestic and exter- nal threats. Domestically, the PNDC could not discount the possibility of a counter-coup by the officer corps of the army. Perceived external 64 Successful Reformers threats centered on Nigeria and the West, in particular the Central Intel- ligence Agency (CIA).14 Once attention turned to economic issues, it was unclear what was to be done. Disagreement quickly emerged among the supporters of the regime as to what path to follow. On the one hand, the more ideological leftists felt that it would be inappropriate for Ghana to be beholden to the same "imperialist" powers they had been denouncing. In their view, Ghana should go it alone, or failing that, draw on resources from the Eastern bloc. On the other hand, there were those who felt that given the dire economic situation, the country would inevitably need some assis- tance from the West.'5 Because of these differing points of view and the multitude of parallel new institutions (such as the PDCs and WDCs) with ill-defined roles, the entire decisionmaking process lent itself to endless discussions. In an attempt to coordinate the PDCs and WDCs better, the regime established an Interim National Coordinating Committee (INCC) in Feb- ruary 1982. It comprised JFM activists and members of the People's Revo- lutionary League of Ghana. While the PNDC retained ultimate policymaking authority, by virtue of its links with people at the grass roots, the INCC became extremely important. This created some ten- sions. Resolution of these was expected (but did not occur) when Rawlings replaced the INCC with the National Defence Council (NDC), which he chaired, in July 1982. The stated aim was to strengthen the defense committees and bring them under greater control. As we shall see, the underlying tensions between the left-wing elements in the NDC and the PNDC ultimately led to a serious breach. Deciding on an economic path was complicated by the fact that Rawlings himself was not very ideological, but instead more populist.'6 Regardless of the points of view, what was not in dispute was the nature of the government's financial difficulties. In December 1981, foreign ex- change reserves stood at US$125 million, less than two months of im- port cover. Short-term export credits and other forms of financing had dried up as external creditors downgraded the country's creditworthi- ness. This situation had undoubtedly worsened given the PNDC's anti- imperialist sentiments. Most foreign donors had halted flows as they adopted a wait-and-see attitude. The solutions proposed for these difficulties differed according to po- litical background. Initially the leftists appeared to hold sway in the eco- nomic debate as orthodox economic reform proposals were shot down. Shortly after assuming office, the PNDC had appointed a National Economic Review Committee (NERC) to review the economic situation and make recommendations for short- and long-term measures to reha- bilitate the economy. Two important members of the committee were Dr. Joe Abbey, previously commissioner of finance under General Akuffo, Ghana 65 and Dr. Kwesi Botchwey, the PNDC secretary for finance and economic planning. Dr. Botchwey was previously a lecturer at the University of Ghana and was at the time one of the most influential leftist intellectuals on the university campuses. The NERC technocrats had concluded after a comprehensive review that the government would need to seek exter- nal financial assistance and that a stabilization program involving de- valuation of the cedi was necessary. Technocrats in the Ministry of Finance had been working on an economic reform program. The committee viewed the upcoming World Bank/IMF annual meetings as an opportu- nity to present the government's views. Cognizant of the need for PNDC support, Dr. Abbey outlined the program to a full meeting of the PNDC, the Committee of Secretaries, and the National Defence Council in Au- gust 1982. The meeting's reaction to the August 1982 presentation was initially favorable as participants agreed with Dr. Abbey's exposition on the ex- isting (poor) incentives for exporters. They also seemed to agree with the principle of giving exporters bonuses and surcharging importers. However, a seemingly innocuous question about the distinction between the surcharges and devaluation shifted the momentum. Dr. Abbey's (cor- rect) reply that "it was a matter of semantics" highlighted to the leftists that devaluation was an integral part of the program.'7 In the uproar that followed, no decision was made on the economic path, and several weeks of debate followed. After the August 1982 meeting a new committee was established to reconcile differences, headed by Chris Atim of the JFM and Yao Graham of the NDM. They came up with what has been characterized as a more leftist document, the Alternative Economic Programme. However, this characterization is too simplistic. As reported in Nugent (1996) and con- firmed in interviews, the program was remarkably clear-eyed on certain harsh realities: the need for a realistic exchange rate, for example, as well as the need to contain wages and to link wage increases to produc- tivity. The main thrust of the program, however, was linking economic to political development. In the view of the authors, Ghana's problems were deep-seated and demanded a fundamentally different economic path free of imperialist interference. From their viewpoint, it was incon- sistent to depend on the West to rescue Ghana. Implicitly then, the pro- gram would rely on future assistance from the socialist countries or generation of the country's own resources. However, it was already obvious that the Ghana's socialist friends could not provide the level of support needed. Aid-seeking missions to "friendly" countries such as Libya, Cuba, Eastern European countries, and the Soviet Union in the first half of 1982 had generated little finan- cial support.'8 Respected leftist intellectuals such as Samir Amin of the Dakar-based Institute for Planning and Development (IDEP) were 66 Successful Reformers brought in. However, their conclusion was the same: the government would have to deal with the international financial institutions to finance any economic recovery. Although the Alternative Economic Programme did not have a financing plan and disappeared from the debate, political difficulties remained in implementing any stabilization program because of ideological differences among supporters of the PNDC. By October 1982, elements of the left wing associated with the Na- tional Defence Council were convinced Rawlings had abandoned the revolution and were increasingly frustrated. They set an October 28 date for a planned coup d'etat to remove him. The plot leaked, however, and the coup plotters were met at the barracks by troops loyal to Rawlings. The coup leaders backtracked and sought to make amends. That some WDCs and PDCs had already assembled in anticipation of a successful coup highlighted the potential dangers to Rawlings. On November 23, another coup was attempted, this time by soldiers who had previously supported the PNDC. Rawlings easily crushed the effort. Riding high on this success, he sought to tie the two conspiracies together.19 The failed coup attempts provided the PNDC with an opportunity to remove those left-wing elements of the JFM opposed to the economic package, as well as perceived political enemies. Several were arrested and others went into exile. The National Defence Committee was dissolved and restruc- tured to ensure tighter control. While the debate was going on, behind the scenes the economic team continued to design a comprehensive economic policy for discussion with the IMF. With the elimination of those left-wingers opposed to a deal with the IMF and more market-oriented reform, the team now had a chance to begin to put forward an economic reform plan. In December 1982, Dr. Botchwey, the finance secretary, presented the government's Program for Reconstruction and Development. This pro- gram is discussed in more detail below. Pending a formal arrangement with the Fund, financing was arranged through a bridge loan from Stan- dard Chartered Bank in London. Dr. Frimpong-Ansah, a former gover- nor of the Bank of Ghana during the Busia years, and then a senior adviser at Standard Chartered, was crucial in arranging this loan (Frimpong- Ansah 1991; interviews in Accra). A number of exogenous crises and increasing financial difficulties in early 1983 further narrowed the options for reform and concentrated the minds of government officials. In January 1983, Nigeria announced that it was deporting all foreign nationals who lacked valid immigration papers. An estimated 1.2 million Ghanaians who had taken advantage of Nigeria's oil boom to escape Ghana's economic crisis were forced to pack up hurriedly and return home. Over a two-week period, Ghana's population essentially swelled by 10 percent. At the same time the coun- try was experiencing a serious drought, reducing opportunities for rural Ghana 67 employment for the returnees. Bush fires had destroyed much of the cocoa crop. In April 1983 the Economic Recovery Program was presented in the budget. By this time, political momentum had shifted in favor of the technocrats. The ERP's economic thrust was essentially the same as the government's program announced on December 30,1982. The main dif- ference was the language: all references to exploitation and anti-imperi- alism had been dropped. The program was supported by a standby arrangement with the IMF and a multisector rehabilitation credit from the Bank. By 1984, technocrats held the day. A small group of highly trained professionals made economic decisions. Many in the group already knew each other, and there was a high level of trust among them. The empha- sis was on finding meaningful and practical solutions for Ghana's prob- lems, and on "packaging" the reforms for presentation to the PNDC. For the rest of the decade there was remarkable continuity as there was little turnover in the group. The group was also sheltered from the political battles to some extent. This concentrated, centralized decisionmaking worked well in the early stages where there was more of a crisis mode, and quick and flexible decisionmaking was essential. However, it be- came increasingly difficult to implement reform this way as the decade went on. This was the result in part of organizational changes in the decision- making apparatus that made coordination more complex. In 1985 a Na- tional Development Planning Commission was established. The Ministry of Finance reorganized itself in 1986: the Internal Revenue Service and the Customs, Excise and Preventive Service became autonomous agen- cies, and new departments were added (Toye 1991). In the late 1980s Dr. Abbey, an essential member of the economic team, became Ghana's ambassador to the United Kingdom and subsequently the United States. While the Washington appointment facilitated dialogue with the multi- lateral institutions, it probably weakened the team. Coming back to politics, economic reform proved politically sustain- able in the end only because some results emerged quickly, and because the government was able to overcome political opposition skillfully through the rest of the decade. Electoral and Constitutional Changes In 1991, a confluence of internal and external pressures led the PNDC to call for presidential and parliamentary elections to be held in 1992.20 Domestic groups agitating for reform included professionals, students, trade unions, and former politicians. The wave of political liberalization sweeping through Africa was a source of external pressure. Finally, 68 Successful Reformers greater donor emphasis on governance and use of conditionality to bring about improved governance hastened democratic reform on the conti- nent. Ghana's elections were staggered, with the presidential elections held in November 1992 and the parliamentary elections set to follow in December. Rawlings won with 58.3 percent of the vote, and the runner-up gained 30.4 percent. Rawlings won more than 50 percent in every region except Ashanti and gained great support in the rural areas. While the opposi- tion parties disputed the results and subsequently boycotted the parlia- mentary elections, by August 1993 they had officially recognized the government's legitimacy. The next elections, in 1996, by all accounts were fairer and the results were accepted by all. Rawlings and his party (NDC) won again, but with a reduced majority. The form of government specified in the 1992 constitution avoids con- centration of power in the hands of the president and cabinet or indeed any branch of government. It is designed to foster tolerance and the con- cept of power sharing. A president, a parliament, a cabinet, a council of state, and an independent judiciary share powers. Legislative functions are vested in the National Parliament, which consists of a unicameral 200-member body plus the president. The struc- ture and the power of the judiciary are independent of all other branches of government. The Supreme Court has broad powers of judicial review; it rules on the constitutionality of any legislative or executive action at the request of any aggrieved citizen. The legal system is based on the constitution, Ghanaian common law, statutory acts of Parliament, and assimilated rules of customary (traditional) law. The 1992 constitution contains the most explicit and comprehensive provisions in Ghana's post-colonial constitutional history regarding the system of local government as a decentralized form of national adminis- tration. These provisions were inspired largely by current law and by the practice of local government under the PNDC. The right to form political parties is guaranteed-an especially im- portant provision in light of the checkered history of political parties in post-colonial Ghana. The constitution explicitly requires political par- ties to have a national character and membership, and says they should not be based on ethnic, religious, regional, or other sectoral divisions. While these changes were well intentioned and indeed necessary, the net result was to make decisionmaking consultative and more difficult. Influence of Democratization on the Pace and Pattern of Reform The macroeconomic situation in Ghana deteriorated significantly in the early 1990s. Two of the main causes were political: the effect of the po- litical cycle and the role of unions. In the run-up to the multiparty elec- Ghana 69 tions, an 80 percent wage hike for all civil servants resulted in large fis- cal imbalances and monetary growth. The other major factor was the significant increase in the number of actors involved in economic decisionmaking. Parliament has played an important role in this regard. For example, in 1993 Parliament failed to approve a petroleum tax in- crease to raise revenues, resulting in worsened fiscal balance. Overall, the time horizon for policy decisions is now shorter. Since elections now require strategic planning, political types rather than technocrats are now ascendant. The centralized decisionmaking process that had served the country well in the earlier period of reform was possible precisely because decisionmaking and consultation had not been institutionalized. So there was no broad and popular support for macroeconomic reforms. This was not much of an issue under military rule, but became a problem with multiparty democracy There were also difficult issues related to second-generation reforms, including privatization, financial sector reform, civil service reform, and taxation. The government faced public (and in some cases violent) pro- tests against reforms. For example, a demonstration against structural adjustment in July 1995 resulted in the deaths of seven people. The new value added tax introduced in March 1995 was withdrawn after riots in two cities. After an extensive public education campaign, it was reintro- duced in 1998 (at a lower level of 10 percent rather than the 15 percent recommended by the government). Ghana's experience thus stands in contrast to the empirical findings of Dollar and Svensson (1998) that democratically elected governments are more likely to reform successfully. Ghana's periods of most rapid reform occurred under a non-elected government. Democracy has com- plicated the process of reform. But Ghana's experience is also a function of the nature and maturity of democracy in the country. For example, after the opposition boycotted the 1992 parliamentary elections, which they viewed as unfair, there was no real opposition in Parliament. In- stead, the opposition parties used lawsuits to capture government at- tention and as the basis for their dialogue. Consequently, the government spent much time fighting these lawsuits. Political Parties The PNDC banned all political parties following the December 31,1981 coup. With the announcement that elections would be held in 1992, po- litical parties were "unbanned" on May 18, 1992. In practice, however, the parties existed in all but name throughout most of 1990-91. There was a flurry of activity as new parties registered and established them- selves. When the dust settled, it could be seen that the major new parties 70 Successful Reformers all were essentially alliances of previous political groupings. The most important were the New Patriotic Party (NPP), the People's National Convention, the National Convention Party, the People's Heritage Party, the National Independence Party, and the unregistered People's Party for Democracy and Development (PPDD), which did not contest the elec- tion.21 A group of pro-government clubs and organizations formed the National Democratic Congress (NDC) and nominated Rawlings as their candidate. More importantly from the viewpoint of economic reform, the eco- nomic platforms of most parties were very similar to the ruling PNDC and to each other. For example, the liberal economic position of support for an open market-based economy, for which Busia's Progress Party/ NPP was known, had been appropriated by the PNDC. The presidential candidates of the other three parties also had little disagreement with the broad direction of the Structural Adjustment Program. One was a wealthy entrepreneur who supported market-based reforms. A second had an economic adviser who was Central Bank governor during the Busia era when market-oriented reforms were attempted. The third ac- knowledged that Nkrumahism and the welfare state were no longer sus- tainable (Nugent 1996: 243). The push for political decentralization and rural development was already being implemented by the PNDC. Ac- cordingly, the structural adjustment program was never an issue in the campaign. Reflecting the disappearance of economic issues from the agenda, the pre-election debates centered on the human rights record of the PNDC and attacks on Rawlings. Support for NDC came from a nationwide grassroots network as well as from aspirant businessmen, the rural elite, and those members of the middle class who have benefited from the rapid expansion of the service sectors. This includes, one might add, the cottage industry of aid-fueled projects and consultant assignments. As the incumbent, the PNDC was able to use a form of state patronage to bolster its support. Aid enabled the PNDC to expand access to infrastructure and social services (Nugent 1996: 206). New roads were constructed and existing ones improved all over the country, and the PNDC expanded the electricity grid to several regions. The New Patriotic Party emerged from the Danquah-Busia tradition.22 According to Nugent, it represented a "significant section of the Ghana- ian elite, chiefly professionals and businessmen, drawn from the Akan heartland of the defunct Progress Party." The Nkrumahist camp was represented by a plethora of divided parties. In the end, Rawlings won in 1992 (albeit controversially) and in 1996 (more universally accepted) in part because of the weakness of the op- position forces. Especially in 1992, they were often divided. One could also make the case that the NDC won because the opposition parties did Ghana 71 not present a convincing alternative economic plan. Furthermore, Rawlings, as head of state for more than a decade, was able to exploit the advantages of incumbency. His name had become a household word, and he had been able to successfully mobilize grassroots support in the tradition of the Convention People's Party,23 winning favor with a wide range of interest groups, influential chiefs, and local leaders. Helping Rawlings disseminate his message were a well-established nationwide network of Committees for the Defence of the Revolution, the 31st De- cember Women's Movement, other so-called revolutionary organs, and dedicated district secretaries and chiefs. All these had been active long before the fractious political parties-whose rival leaders were hardly known beyond the major cities-had struggled into existence. Clearly, the benefits of incumbency are especially marked in Africa and will con- tinue to be until a sustainable way to finance political parties is found. Interest Groups Politically active and influential interest groups in Ghana include the Trades Union Congress (TUC), the National Union of Ghana Students (NUGS), the Ghana Bar Association, the Association of Recognized Pro- fessional Bodies, and various Christian organizations such as the Catho- lic Bishops' Conference and the Christian Council of Ghana. During the early years of the PNDC, the TUC and NUGS were important support- ers of the regime, but did not influence economic policy. In later years they became important pressure points for political reform; this in turn influenced economic reform. In contrast to many other countries, in Ghana the business commu- nity was not an important interest group for a number of reasons. First, there was no monolithic "business community." On the one hand, old- time businessmen were resentful of the government and sensitive to its criticism, especially the charges of kalabule, even though many of these people had not necessarily benefited greatly under Acheampong or Limann. Many of them suffered under the rapid trade liberalization imposed by the Economic Recovery Program. On the other hand, aspir- ing entrepreneurs who were benefiting from the liberalization of the economy were grateful for the new opportunities. The second reason was that with increased PNDC scrutiny of tax payment records and the Citizens' Vetting Committee, few business owners were interested in drawing attention to themselves. The Trades Union Congress The TUC had a complex relationship with the PNDC regime that varied over time and depended on the specific affiliated union. Of the 17 unions 72 Successful Reformers in the TUC, the most important are the Ghana Private Road Transport Union, the Ghana Mine Workers' Union, and the Railway Workers' Union. In the early 1990s total membership of the TUC was more than 650,000. Between 1982 and 1984, the regime enjoyed widespread support from the unions. With their leadership discredited due to past close associa- tion with Acheampong, the workers sympathized with the stated ideals of the PNDC. In April 1982, the Association of Local Unions (ALU)-a group of more militant trade unionists-unseated the TUC leadership and heads of the 17 national unions. An interim management committee made up of radical supporters of the regime replaced it.24 How did the government sell and sustain reforms under the ERP to labor? After all, the program implicitly required wage restraint and po- tential mass retrenchments, which were bound to hurt organized labor. Consequently, it could be viewed as a betrayal of the regime's populist and left-wing credentials. However, the government was able to sur- vive the budget for a number of reasons. To begin with, the labor leadership was already compromised by its close links with the PNDC. Other important points were Rawlings's personal credibility and the fact that some of the measures were me- dium-term. Cukierman and Tommasi (1994) contend that policies are more likely to be implemented by "unlikely" leaders (such as popu- lists or left-wingers) than by those viewed as ideologues. If the popu- lace is not fully informed about the short-term costs, then the populist reformer is better placed to implement reforms. People are more likely to be convinced that the reforms are essential to enhance efficiency rather than ideologically motivated (Dollar and Svensson 1998). This was certainly the case in Ghana. Rawlings's public speeches increas- ingly emphasized the need to be pragmatic, to enhance efficiency and productivity in order to reduce Ghana's dependency. Other reasons that the PNDC survived the budget speech include its concerted at- tempt to explain the ERP through public seminars, its direct negotia- tions with the TUC-resulting in a modest increase in the minimum wage-and, to some extent, fear of the government. Ultimately there was still a bit of a political honeymoon, giving Rawlings the benefit of the doubt. By the end of the year, the fortunes of the ALU workers had changed. TUC elections voted most of them out and put many of the former ex- ecutives back in power. Some informants for this study believe that work- ers viewed ALU as too close to the regime. Whatever the reason, the TUC elections marked the beginning of a more confrontational relation- ship between organized labor and the PNDC. Workers tried to further their economic interests and focused on the issues of the minimum wage, civil service reform, and privatization. They were constrained, however, Ghana 73 by the environment. Heightened opportunities for removal of staff weak- ened their bargaining position. They were also hampered by the fact that they did not have any constructive alternative vision25 and the po- litical skills of government. Moreover, different unions benefited differ- ently, reducing union unity. In mining, Ashanti Goldfields workers negotiated a wage of US$2 a day compared with much lower wages at state-owned mines. In the late 1980s the PNDC tried to co-opt moderate elements of the union movement. By the end of the decade, the gap between the union members and their leaders had widened again. National Union of Ghana Students The student organization, NUGS, which represented more than 8,000 students at the country's three universities, was also important in the political dynamics. Initially students had been strong supporters of the regime, but over time they became vociferous critics. The students' sup- port dated back to the original June 4 uprising, when they had been more radical and to the left of the AFRC on many issues. In mid-1982, a crucial time, the students had demonstrated on behalf of the government. But by the end of that year the more conservative students had taken control of NUGS, and as a result support for the PNDC diminished. The change in NUGS support has been linked to student fears over the proposal to institute a two-year National Service before graduation and plans to revive the Student Task Force. In addition, the increased Christian revivalism on campus (and its associated anti-communism) was at odds with the perceived stance of the PNDC ("Ghana: Roots" 1983). By early 1983, NUGS was calling for the PNDC to hand over power to a government of national unity pending new elections, and following the April budget the students engaged in mass demonstrations against the government. At both the University of Ghana and the University of Science and Technology the students clashed with workers.26 Ultimately the students were unable to muster broad support for their views. They were widely perceived as being elitist and clashes with campus PDCs had not endeared them to the workers. Apart from their political viewpoint, the students were hostile to edu- cation reforms introduced after 1987, which required them to make greater financial contributions. With increasing student protests in 1987 the government closed the university campuses and arrested various student leaders. During much of the following academic year, the uni- versities were in turmoil as student boycotts occurred and the universi- ties were shut down repeatedly. The closures served effectively to neutralize the protests. 74 Successul Reformers POLICY REFORMS As far as economic policy was concerned, 1979-82 was essentially a pe- riod of muddling through. With elimination of the opposition to reform in late 1982, policy changes began in earnest. Economic Reforms Economic reform in Ghana can be broken down into three main periods: (a) successful stabilization, from 1983 to 1986; (b) structural adjustment, from 1986 to 1991; and (c) oscillating policy reform, from 1992 to 1996. Successful Stabilization, 1983-86 The Economic Recovery Program, supported by World Bank loans and three IMF standby arrangements, was the first comprehensive economic policy package in more than a decade. The objectives of the ERP were as follows: * Reversing the long decline in production of goods and services by realigning relative prices in favor of production and away from trading and rent-seeking activities; * Reducing and stabilizing the inflation rate; * Reducing the large budget deficits and improving the fiscal posi- tion of government more generally; * Rehabilitating social and economic infrastructure; * Eliminating smuggling and black market activities with respect to the currency; and * Realigning the currency (the cedi) with major currencies (Ghana 1987). Structural Adjustment, 1986-91 The Structural Adjustment Program, also known as Economic Recovery Program II, was a follow-up intended to focus on more deep-seated struc- tural issues in the economic recovery. Its specific objectives were to: * Sustain economic growth at between 5 and 5.5 percent a year over the medium term; * Increase the level of public investment from about 10 percent of national income to about 25 percent by the end of 1989; * Increase domestic savings from about 7 percent at the end of ERP I to about 15 percent at the end of the decade; * Further improve the management resources in the public sector; Ghana 75 * Effectively mobilize the resources thus generated to improve the overall well-being of the people of Ghana, particularly the under- privileged, deprived, and vulnerable sectors of the population (Ghana 1987). The overall assessment of several observers of the Ghanaian economy is that the ERP and SAP were successful in reversing the economic de- cline that had taken place. Growth moved from negative to positive and the economy was successfully stabilized (Nyanteng 1997; Kapur 1991). Throughout the 1980s Ghana earned high marks for its reform efforts. We now turn to the individual record of the main policy indicators dur- ing the reform periods. MACROECONOMIC MANAGEMENT. An immediate economic concern of the PNDC was to improve the external payments position and con- trol rampant inflation. From this perspective, the first stage of stabili- zation was well managed. Indeed, its centerpiece was the correction of the chronically overvalued exchange rate and successful reduction of inflation by 1991. EXCHANGE RATE. This was an early and important indicator of re- form as it was a very visible symbol of the large distortions in the Gha- naian economy. The government introduced innovative reforms in exchange rate management. Over the reform period, the exchange rate became more realistic and flexible. Exchange rate management can be characterized as one of the ultimate policy successes despite the high political risk that accompanied adjustment of the rate. As mentioned earlier, the PNDC inherited a grossly distorted exchange rate. In early 1982, the official cedi-U.S. dollar rate was 2.75:1, while the black market rate was over 19:1 (Chibber and Shafik 1991). A straightforward devalu- ation was perceived by some to be politically damaging. To get around this problem, the Ghanaian economic team devised a system of "bonuses" and "surcharges." This was of course a de facto devaluation, but it avoided the use of the dreaded "D" word.27 Over time, the system of bonuses and surcharges was discarded and step devaluation took place. Between 1983 and 1986, the currency was progressively devalued in nominal terms from 2.75 to 90 cedis per U.S. dollar. In 1986 the government moved closer to a fully flexible exchange rate with the introduction of a retail foreign-exchange auction market. The government allowed the establishment of foreign exchange bureaus in February 1988 as part of the continuing liberalization of the trade and exchange system. This liberalization continued and culminated in the establishment of an inter-bank market for foreign exchange in 1991. 76 Successful Reformers FISCAL POLICY. Fiscal discipline was introduced and, despite periodic slippage, ultimately sustained during 1983-91. The government turned a fiscal deficit of 4 percent into a surplus of 2 percent. Early success in reducing the deficit rested on the ability to increase revenues rather than on expenditure reduction. The Citizens' Vetting Committee succeeded in expanding the tax base, resulting in greater revenue collection, and revenues rose as a share of GDP from 6 percent in 1983 to 14 percent by 1991. The successful fiscal adjustment was instrumental in lowering inflation. INFLATION. By reducing deficit financing the government removed one source of inflation. In 1981, CPI inflation was 77 percent. It fell to 22 percent by 1986 and dropped further to 9.5 percent in 1992, its lowest level in two decades. STRUCTURAL POLICIES. Under the ERP, the prevailing extensive web of administered prices and price controls was largely dismantled. Inter- est rates were liberalized in the financial sector. By 1986 virtually all price controls had been removed. The external trade regime was liberal- ized rapidly, but progress was slow on the divestiture of state-owned enterprises. TRADE POLICY. Significant trade liberalization began under the ERP in 1983. Tariffs were simplified and lowered to create a uniform pattern of protection. Initially, the number of rate bands was reduced to three (0 percent, 25 percent, and 30 percent). Import tariffs fell further in 1986 when rates were lowered to 20-25 percent. In addition, the authorities shortened the negative list for imports. While the exchange rate had moved to a more realistic level, this was still far from equilibrium level or the parallel market rate. Accordingly, import controls remained in place until 1986. Overall, by 1986 Ghana was classified as "open" ac- cording to the Sachs-Warner index of openness, in contrast to the previ- ous years when it was classified as "closed." DIVESTITURE OF STATE-OWNED ENTERPRISES. Under the ERP and SAP, progress in reducing the state's role in public enterprises was slow due to vested interests, political sensitivities, and, in some cases, lack of tech- nical capacity. By 1991 the government had divested only 42 of approxi- mately 300 SOEs. PUBLIC SECTOR MANAGEMENT. Public sector management was not at the forefront of the reform agenda under the ERP. In part, this was be- cause the first stage of stabilization consisted of reforms that were auto- matic and involved announcements by government rather than detailed Ghana 77 implementation. These included, for instance, the removals of price con- trols and adjustments in the exchange rate. Under the SAP, the govern- ment and donors became increasingly concerned about implementation capacity within the public sector, as well as economic policy coordina- tion. Consequently, measures to strengthen key policy functions in the Ministry of Finance and to reform the civil service were part of the SAP. The government launched civil service reform in 1987 with support from the World Bank but its commitment was weak, and targets for retrench- ments were not met. The judiciary and the rule of law were even further removed from the agenda during both the ERP and SAP. Oscillating Policy Reform, 1992-96 This period marked the first major interruption of lending by the multi- lateral institutions in two decades. Between November 1992 and mid- 1993, Ghana's adjustment program veered off track and the World Bank suspended disbursements. Fiscal dislocation associated with the large (80 percent) wage increases awarded to civil servants before the 1992 multi- party elections was the main cause of the policy derailment. The imme- diate effect was the emergence of a fiscal deficit of about 5 percent of GDP in 1992 from a surplus of 1.5 percent in 1991 as the money supply increased by more than 50 percent. In the second half of 1996, reacting to significant fiscal slippage, the IMF delayed negotiation of the second annual Enhanced Structural Adjustment Facility and the Policy Frame- work Paper for a year. More generally, it appeared that the government faced increasing difficulties in implementing additional structural reforms. Ghana survived these temporary suspensions because of strong sup- port from the bilateral donors. By the early 1990s, Ghana had become the ultimate test of structural adjustment and donors were unwilling to see it fail. Canada (which had always been supportive), Japan (which did not generally get into policy discussions), and the United Kingdom in particular all exerted pressure on the multilateral institutions. During this period the government struggled to reduce inflation, which continued to be high and variable. Contributory factors were the creep- ing fiscal deficits and associated growth in the money supply, which remained high despite a decline from 57 percent in 1992 to 31 percent in 1996. With respect to trade policy, the government introduced a number of incentives targeted at exporters to redress the worsening external pay- ments position. The pace of divestiture of SOEs accelerated. Fiscal con- siderations led the government to sell shares in Ashanti Goldfields (AGC) on the Ghana and London stock exchanges. The government followed the sale of AGC with many other offerings. The privatization program 78 Successful Reformers picked up, and by 1997, more than a third of the 300 public enterprises had been sold or liquidated. With respect to public sector management, by 1994 all parties acknowl- edged that donor efforts to enhance capacity had been largely ineffec- tive. The government launched the National Institutional Renewal Program (NIRP) in late 1994 to coordinate various public sector reforms (financial management, perforrnance improvement, and pay). In the mid-1990s the government made progress on defining the regu- latory framework and including the enforcement of contracts and prop- erty rights. A liberal Investment Act (1994), a Free Zones Act (1995), the Statutory Corporations (Conversion to Companies) Act (1995), and the National Communications Act (1996) facilitated this. In addition, a multisector regulatory agency has been created to facilitate further pri- vate sector activity. The World Bank has supported many of these initia- tives. Table 2.2 summarizes the sequencing of individual reforms in the 1983-96 period. High turnover on the economic teams of the World Bank and the gov- ernment did not facilitate policy reform during the 1990s. A number of World Bank internal reorganizations and the proliferation of projects (and hence task managers) led to a lack of continuity that Ghanaian officials found frustrating. In addition, some officials felt tensions between Bank and IMF staff prevented clear guidance on a number of policy issues, such as civil service reform. In sum, Ghana's progress and evolution becomes evident when its record of macroeconomic management and structural reforms be- tween the 1970s and the mid-1990s is compared to that of other Afri- can countries. On the fiscal side, Ghana's fiscal deficit was significantly higher than the average for the rest of Africa in the 1970s; it surpassed 11 percent in 1976, more than double the average of the continent. Beginning in 1982, Ghana's fiscal position improved tremendously, and from 1983 to 1991 it registered smaller deficits (and eventually larger surpluses) than the rest of the continent. While Ghana's fiscal deterioration in 1992 resulted in a larger deficit than average, by the mid-1990s it was registering surpluses and was again performing above the Africa-wide average. The story on inflation reflects the struggle to contain inflationary pres- sures that Ghana has faced throughout its reform efforts. Ghana started the reform program with inflation that had been significantly higher than the rest of Africa for almost a decade. While inflation was eventu- ally reduced below the continent average in 1991-92, the fiscal dislo- cation and other pressures after that caused inflation to exceed the average again. On the trade side, Ghana made early and rapid progress compared to other African countries. While it was classified as "closed" according to Ghana 79 the Sachs-Wamer definition in 1980, by 1985 it was one of only four coun- tries in Africa (out of 36) that was classified as "open." By 1996 several other countries were classified as "open" (14 of 36). The World Bank's Country Policy and Institutional Assessment ratings for 1998 bear out Ghana's superior macroeconomic management and record of accomplish- ment in structural reform. In both cases, Ghana scored above the aver- age for countries in Africa. Triggers for Reform When and why countries choose to reform is an issue that interests many researchers. Five factors help to explain why the PNDC chose to reform when it did in 1983. The first two provided the immediate impetus for reform, while the last three have to do with the political dynamics that gave the government needed political space to carry out the reforms. First, there was undoubtedly a liquidity crisis. Foreign exchange re- serves, trade credit, and medium-term financing had all dried up by early 1983. At the same time, the government faced unavoidable obliga- tions. Several former officials interviewed for this paper alluded to im- pending oil payments and the need to pay for a new DC-10 aircraft. Second, drought and bush fires, as well as the return of more than 1 million Ghanaian citizens from Nigeria in early 1983, compounded this liquidity crisis. These two factors (the extreme financial crisis, and the exogenous shocks) are central in explaining how the government came to consider reform. In essence, the situation was so desperate that the costs of doing nothing greatly outweighed the costs of reform. What gave the government political room to carry out the reform? First, not many alternative approaches existed that would generate the financial resources necessary for the economy's recovery. Few Ghana- ians needed to be convinced of this fact; for the average citizen who had already suffered more than a decade of economic hardship, it was clear that the old order had failed. Second, the PNDC had credibility with interest groups that could have created difficulties for the government- notably trade unionists and students. These groups were sympathetic and in essence gave the PNDC a policy honeymoon. Other political op- position, both within and outside, had either been eliminated or was demoralized. The PNDC made a huge effort to educate military groups and the Workers' Defence Committees on the rationale for the reforms. Third, the crisis of the returnees created some national unity. Indeed, the situation gave the regime some political cover to accept the World Bank and IMF loans, which were presented as bolstering the government's own efforts to deal with the crisis. Overall, though, the economic debate that had taken place during 1982 was necessary for the evolution of economic thought from populism 80 Successful Reformers to more market-based reforms. It is inconceivable that these reforms could have taken place in 1982; neither Rawlings nor essential members of the PNDC were yet convinced that market-based reforms were necessary.28 In that sense, the reforms of 1983 were a result both of a gradual evolu- tion in the government's thinking and of crises that served as a trigger in the early part of that year. The evolution of economic thought was to some extent influenced by staff at the World Bank and IME. Both institutions had highly committed individuals on the macroeconomic country team who were willing to discuss issues with the government in a nonideological fashion. Indeed their commitment to the process led them to recommend the ERP, al- though officials back in Washington thought the program was too gradual and questioned the government's commitment to reform. Several people interviewed also referred to the efforts of a senior Ghanaian economist at the Bank who was instrumental in analytical discussions.29 As Ghana was at the forefront of African countries in terms of structural adjust- ment, there was little learning from the country's neighbors. Homegrown Policies As alluded to earlier, different elements in the government attempted different economic reform packages. The contents of the so-called "more leftist" programs have been discussed already. In December 1982, Secretary for Finance Dr. Botchwey presented the government's Programme for Reconstruction and Development (Ghana 1983). This program "could be seen as a balancing act between the nuts- and-bolts reformism of the National Economic Review Committee and the long-range concerns of the Alternative Economic Programme. In di- agnosing the causes of the crisis, Botchwey struck a middle position, highlighting both the vagaries of the international economy and the pursuit of unwise monetary, fiscal, foreign exchange and pricing poli- cies since the 1970s." A broadened tax base, fewer state enterprises, and a more realistic exchange rate were presented as solutions to the crisis. The program also included some statements associated with the lan- guage of the left. For example, there were references to ending exploita- tion and to imposing state control over certain aspects of import-export trade (Nugent 1996: 110). In short, the reconstruction and development program was broadly similar to the final Economic Recovery Program. The important differ- ence was that the reference to mass mobilization of the people was dropped. In practice, with the expulsion of the Ghanaians from Nigeria and the need to resettle them, food aid from donors was used to mobi- lize the returnees in the rural areas for assistance with harvesting the cocoa crop. Ghana 81 Ownership The Economic Recovery Program and the Structural Adjustment Pro- gram (to a lesser extent) enjoyed high Ghanaian ownership-if owner- ship is narrowly defined to mean the government's conviction that it had a stake in the program. Ghanaian officials could rightly point out that they had drafted the original program on which the ERP was based.30 Public statements by the authorities at the time indicated that they took responsibility for the program. In addition, the authorities did a signifi- cant amount of background work on one of the important issues: de- valuation. The idea of PAMSCAD emerged not only from concerns raised by outside agencies (principally UNICEF, the United Nations Children's Fund), about the social costs of adjustment, but also from government concerns. There was a great deal of back-and-forth discussion between the Bank/Fund teams on the one hand, and the government under the ERP. Within the Bank, the Ghanaians came to be known as "tough nego- tiators, able to define the limits of which issues they were or were not prepared to take on, and determine how far they could go" (Armstrong 1996:54). In an external assessment of the World Bank program in Ghana, the reviewer comments that this ownership led to a more gradual pro- gram than either the World Bank or the IMF would have preferred. He notes that this was essentially the correct decision, however, given the political economy of the time (Armstrong 1996:54). That is not to say of course that the gradualism was optimal. In the 1990s, ownership seems to have declined. This may be a re- sult of several factors, including changes in personalities leading to the need to build new relationships of trust, and adjustment fatigue on the part of both Ghanaians and donors. Perversely, Ghana's successful adjustment in the 1980s led to increased donor flows and with them, greater dependency. This has resulted in more donor-driven agendas and less ownership. Ownership more broadly defined-to include the private sector, civil society, academics, and trade unions-was not even on the agenda in the beginning.3" There was less participation by civil society in the pro- grams of the 1980s. The political situation affected this stance to some extent. As discussed above, the PNDC was under continuing political pressure throughout most of its tenure, but especially in the first few years. Opening a dialogue on the reform program was viewed as politi- cally dangerous. While some attempt was made to provide a forum to discuss the ERP and SAP during the 1980s, very little real discussion was ultimately achieved. Some observers interviewed in Accra suggest that the World Bank itself was concerned about having "endless discussions" of the reform process. So, in general, there was little pressure from any source to have 82 Successful Reformers wide participation. In the short run this enabled the authorities to move decisively and firmly on a number of potentially controversial economic issues. However, it has been argued elsewhere that the failure to institu- tionalize the reform underpinned the difficulties the government faced in consolidating reform in the 1990s (Nugent 1996). With the exception of only one aspect of the reforms, the various in- terest groups did not influence the design of the two reform packages in the 1980s. The exception was the social costs of adjustment, particularly as they affected the poor. Given the PNDC's heritage, however, it seems that one reason criticisms of the SAP in this regard were taken seriously was that there was genuine worry about how the poorest people were being affected. To put it another way, it is hard to imagine the PNDC at the time agreeing to such a program targeted at the middle class. In ad- dition, external agencies, notably UNICEF, were instrumental in bring- ing the issue to the forefront. In 1988 the donor-funded PAMSCAD initiative was introduced. PAMSCAD targeted vulnerable communities, most of them rural, and self-help projects within them. To the extent possible, local materials, technology, and labor were to be used. In total US$87 million was initially committed for PAMSCAD. THE LINK BETWEEN AID AND REFORM The overall objective of this study is to assess whether and how aid influ- enced reforms in Ghana. Exposures to ideas and policy conditionality, discussed below, are two of the channels through which aid can influ- ence the decision to reform. We conclude by assessing what the Ghana case study suggests about the larger project's three hypotheses: * Governments choose to reform, or regress, independent of the aid relationship; * Nonfinancial aid has a better impact than financial aid on genera- tion of policy reforms; * Financial aid works when policy reforms and institution building are underway Generation and Implementation of Reforms Generation of Ideas Ideas for the ERP and SAP were generated through a partnership be- tween a small group of technocrats in Ghana and a similar group in the World Bank and the IMF. The Bank in particular carried out high-quality macroeconomic analysis that informed the government's decisions and Ghana 83 was viewed as useful in the policy dialogue in the mid-1980s (Armstrong 1996). This was especially important given that the intellectual environ- ment for economists in Ghana had become increasingly barren during the 1970s with the departure of many economists and the closed nature of the economy. On the other hand, since the work of the World Bank was often devoid of any sense of the political economy of reform, the Ghanaian team was critical in providing the context and parameters of reform. This partnership is viewed as having worked well. Both exter- nal reviewers and internal Bank audits judged that the first and second Bank structural adjustment credits had met their objectives quite well. Once we move from the macroeconomic to the sector level, it appears that World Bank economic and sector work was uneven in quality or entirely absent. In these cases, reform suggestions were unconvincing to government officials. The Country Assistance Review on Ghana cites the areas of privatization and public enterprise reform as examples of this phenomenon. From this viewpoint, the poorer project performance in these areas was not surprising.32 Notwithstanding the importance of World Bank economic analysis in the mid-1980s, given the financial constraints at the time, financial aid was clearly crucial and more important in sustaining the program. While some of the technical work may have helped to persuade officials on particular issues, this would have been a hollow victory in the absence of financial resources that they could use to revive the economy In addition to formal economic analysis, particular individuals out- side the system were important.33 As alluded to earlier, key Ghanaians working overseas (at the World Bank and at IDEP in Dakar) were also viewed by some policymakers as critical to formulating ideas. The aca- demics by contrast did not influence the reform process very much. They seemed to have had a healthy dose of skepticism about how long the regime would last. The clashes with students, which led to the shutting down of the universities, did not create favorable conditions for dia- logue between the government and the academics. What is striking is that sources of ideas on reform tended to be avail- able to a rather small group of people within Ghana initially. One con- clusion to be drawn is that if these ideas are to be disseminated and discussed beyond government partners, then the supporting studies must be made widely available. The Bank's recent move toward making its documents more accessible is an important step in this direction. Technical Assistance Technical assistance by donors has had a mixed record and the overall impact appears weak. The Ghanaian civil servants interviewed perceived technical assistance rather negatively. There was a feeling that it arrived 84 Successful Reformers late in the reform process and was not properly designed. One Ghana- ian policymaker who has dealt with several technical assistance projects characterized TA as "excessively paid consultants from the countries funding that same technical assistance with little or no appreciation for the political and institutional dynamics." Similar comments were com- mon among working-level staff interviewed in Accra. Recent comments by the former minister for finance and economic planning, Dr. Botchwey, echo this characterization, suggesting that junior and senior staff shared these views.' The National Capacity Assessment of 1996 found that "Ghana's capac- ity to handle various technical and managerial activities in most sectors of the economy has worsened in the last two decades." Efforts to rebuild capacity, it said, had not been successful. The report identified "lack of investment in the institutions/processes that form capacity" and the "slow destruction of the enabling environment for maintaining existing capaci- ties" as the core underlying reasons for the decline (Ghana 1996). Technical assistance in Ghana has suffered from a number of difficul- ties. One of the most serious has been the lack of coordination by do- nors, resulting in duplication of efforts. The United Nations Development Program's National Technical Co-operation and Assessment Program (NATCAP) was supposed to improve the effectiveness of TA. However, as of 1996, the NATCAP program in Ghana was having difficulties achiev- ing this objective and was apparently receiving little support from other donors. As of 1996, there was no formal aid coordination mechanism for TA. Attempts to have Ghanaian professionals living overseas return to work in the civil service has had mixed results. These "returnees" have engendered substantial resentment among Ghanaian civil servants be- cause of their better working conditions. Sustainability Aid to Ghana was very important in sustaining reforms. As Ghana imple- mented wide-ranging reforms in the mid-1980s, it gained the attention and admiration of bilateral and multilateral donors. Aid flows to Ghana increased accordingly, as many donors were eager to support a "win- ner" that was making tough decisions-especially in a continent where success stories were rare. As outlined above, the amount of aid to Ghana grew exponentially during the 1980s. Initially a great deal of this sup- port was for balance of payments and rehabilitation. Over time, donors moved into sectors. This massive aid flow was instrumental in offsetting the terms-of-trade decline that Ghana faced in the late 1980s. Aid, in essence, allowed Ghana to escape the potentially contractionary effects of a reduction in absorption. Ghana 85 Aid also helped to sustain reforms during the ERP and SAP by reduc- ing their political costs. Two examples discussed in Aryeetey and Cox illustrate this point. First, balance of payments support "provided the government with the breathing space it required to contain domestic opposition to market-based reforms ... [It] allowed imports that helped fill the shelves of supermarkets and other traders. The filled shelves pro- vided a psychologically-induced breather for the government because . . . people saw this as a sign of better things to come" (Aryeetey and Cox 1997). The second specific example is PAMSCAD. While the implemen- tation of the program was disappointing for a number of reasons, it still indicated that the government was concerned about the rural poor and "silenced opponents." Finally, to the extent that aid (and the programs it supported) resulted in faster economic growth, rehabilitation of the country's roads, and re- vitalized exports, it helped the government stay the course. On the other hand, aid may have inadvertently reduced the long- term sustainability of reform in some instances. The nature of the dia- logue and Bank/Fund conditionality favored the use of a small group of technocrats for making decisions. In the context of a non-elected gov- ernment, the net effect was to discourage consultation. The donors were also directly culpable. Rothchild (1991) reports that the very success of the PNDC in implementing reform led many donors and the IMF to ignore political opposition for the first few years of reform. One hypothesis in the literature on aid is that strong institutions im- ply stronger sustainability of reforms. In the case of Ghana, the aid-insti- tutions-sustainability nexus does not meet conventional expectations. Reforms thrived initially despite the absence of strong institutionalized decisionmaking. Aid did not have a positive effect on institutions dur- ing the first years of reform because donors focused on (short-term) out- comes, not processes. Given the political environment in Ghana, achieving these favorable outcomes was dependent on the "inner circle" approach that implied keeping other institutions out of the decisionmaking process. Conditionality Conditionality and the manner in which it was implemented were im- portant in sustaining policy reform during the ERP and SAP. Aryeetey and Cox (1997) discuss how the government often used the existence of conditionalities to gain leverage when negotiating with lo- cal groups. The example they give is education sector reform, which was pushed through rapidly with little discussion. Domestic complaints were greeted with the response that World Bank requirements on dis- bursements and conditionality necessitated that approach. 86 Successful Reformers The nature of conditionality changed in a number of specific ways between 1983 and 1996: * Conditions became more numerous. IMF conditions rose from 20 in 1983 to 40-50 in 1988-89 (Martin 1991: 245). This reflected the Fund's increased involvement in structural issues, some of which the Bank regarded as its turf. * Conditionality-based lending became more important. The share of policy-based loans rose sharply in Ghana's overall loans from the Bank. * Conditions became tighter and deeper, and also increasingly specific: for example, personnel cuts in the civil service were explicitly speci- fied (Martin 1991). These conditionalities were not as onerous as they could have been, for two important reasons. The first was the caliber, coherence, and con- tinuity of the Ghanaian team. Its relatively small size, highly qualified staff, and skill at implementing measures allowed some flexible inter- pretation of conditions. In addition, it appears that the Ghanaian team was usually extremely well prepared.35 The second reason was the na- ture of the World Bank and (to a lesser extent) IMF teams. They proved over time adept at understanding the context in which certain decisions had to be undertaken and gave the Ghanaians some room. An example of this two-way dialogue was the introduction of the exchange rate auc- tion. With the support of the Bank, the PNDC convinced the IMF to al- low it to "manage" the auction in order to conserve foreign exchange. Instead of pure speculative bidding, it introduced a number of measures that ultimately reduced the number of bids made or accepted. These included "Dutch" bidding in which the successful bidder paid the rate it bid, as well as strict eligibility and documentation requirements. In addition to this "flexible implementation," the World Bank also redesigned conditions or delayed them if they seemed not to be work- ing. For example, cost recovery in the health sector fell in this category (Martin 1991). The willingness to engage the Ghanaians in discussions and to allow this kind of slippage was due to a confluence of factors. There was un- doubtedly a sense that the government was committed to reform and had sustained it longer than any other country in Sub-Saharan Africa. Bilateral donors who were pleased with Ghana also exerted significant pressure on the international financial institutions. It seems that this flex- ibility also enabled the government to carry out reforms that were fea- sible, generated greater sustainability, and strengthened policy reform. By contrast, since the mid-1990s the international financial institutions have seemed less accommodating, at precisely the time when the gov- Ghana 87 ernment has to carry out more consultation with Parliament. From in- terviews with donors, this change appears to be a consequence of slip- page in the early 1990s and hence disillusionment. This "tougher conditionality" has resulted paradoxically in the likelihood of more slip- page, not necessarily because the government is not committed, but be- cause of the need to bring others along. The World Bank and the IMF had a close working relationship during the Economic Recovery Program and the Structural Adjustment Program. This being the pre-Policy Framework Paper era, the closeness was all the more remarkable. Ghana was one of six countries chosen for an en- hanced Bank-Fund collaboration pilot. Therefore, closer and more fre- quent communication took place between staff in the two institutions working on Ghana relative to other member countries. Despite this close relationship, the IMF and the World Bank did not always coordinate their conditionality well. For example, small farmers and the private sector were expected to lead sustained growth and the Bank programs were designed accordingly. However, the devaluation and tight credit policies recommended by the IMF made it difficult for small farmers and entrepreneurs to obtain the necessary credit to pur- chase foreign exchange on the auction for imported capital goods (Mar- tin 1991). The Bank staff members felt that the Fund's monetary targets were too tight and began questioning them openly in discussions with the government.36 Another example of poor conditionality coordination was cocoa producer pricing, where over time higher cocoa prices re- sulted in a trade-off between fiscal and incentive objectives. Selectivity In the case of Ghana, donors initially took a wait-and-see attitude re- garding the PNDC and its commitment to reform. As already shown, it was not until 1985-two years after the reforms began-that bilateral donors and the international financial institutions committed apprecia- bly greater resources to the Ghana programs. By then, it was clear that the government was willing and able to undertake difficult reforms and there was a scramble to reward Ghana accordingly. This led to Ghana becoming the success story that everyone wanted to shower with aid based on its good policy environment. When the policy environment worsened in the early 1990s, many donors were unwilling or found it difficult to withhold aid. This may have reflected not only donor reluctance to label Ghana a "bad per- former," but also the belief that the government, which up to then had been a reliable partner, was facing genuine political difficulties. Information on the evolution of the composition of aid to meet coun- try circumstances is not easy to document and varies across donors. 88 Successful Reformers Nonetheless, some observations can be made. During the early to mid- 1980s the World Bank had an appropriate mix of analytical work and financial assistance. A great deal of lending was underpinned by analy- sis that laid out the case for particular policy recommendations (Armstrong 1996: 52). During the late 1980s and early 1990s, there was more need for technical assistance and fresh ideas as the institutional requirements of adjustment increased, and as program design became more complicated. However, the Bank, among others, was late in pro- viding this and the technical assistance that it offered was often ineffec- tive (Armstrong 1996: 103). With the advent of multiparty democracy, several donors began pro- viding a new kind of support. Financial and technical assistance to sup- port elections became part of aid. In the run-up to the 1996 elections, all the major donors participated in meetings of the Inter-Party Advisory Committee, which was an important forum for discussion and consen- sus building. Altogether, the donor community gave US$12.3 million to the National Electoral Commission to "facilitate free and fair elections" by enhancing the capacity of the NEC. A wide range of donors contrib- uted: the United States, Canada, the European Union, Germany, the Netherlands, and China (Ninsin 1998: 185-202). A recent innovation is the move toward sector programming of aid. Pilots of this approach are already in place for the health and education sectors. The sector approach marks recognition by donors that the mul- tiplicity of actors in particular sectors with different procurement and disbursement procedures as well as different strategies has not served the country or the donors well. Finally, with multiparty democracy and the increased importance of civil society, donors seem to be emphasizing aid that can facilitate inter- nal dialogue and debate and help the adjustment to democracy. The Bank is shifting its mix of instruments toward non-lending policy advisory services in Ghana. These include, for example, World Bank Institute sup- port for consensus building in civil society and short "just-in-time" tech- nical advice. According to the Country Assistance Strategy, this shift is partly in recognition of the need for vigorous communication and dis- cussion of policies given an active media and parliamentary opposition. Canada and the United States also provide nonfinancial aid aimed at strengthening civil society. U.S. aid particularly targets research and business organizations. Conclusion At the beginning of the reform program, even one of the program's stron- gest supporters perceived the link between aid and reform as something of a chicken and egg question. World Bank lending to Ghana had been Ghana 89 suspended (no new loans approved between June 1981 and January 1983) pending macroeconomic reform. Nevertheless, an important internal Bank document also noted that "the Ghanaian authorities' receptivity and interest in pursuing an economic dialogue with the Bank was very much dependent upon an early resumption of Bank lending" (Armstrong 1996). On that basis it was felt that supporting the (however gradual) ERP was an important way of gaining credibility for the Bank that could be parlayed into future advice. To conclude, let us revisit our three hypotheses. The first asserts that countries choose to reform or regress independent of aid. Ghana's expe- rience shows that aid was an important part of the decision to reform as the government anticipated that aid would enable it to meet its economic and political objectives. Aid was also important in sustaining reform as it offset economic shocks, generated demonstration effects, and reduced political pressure. The earlier discussion of the political economy con- siderations underlying reform and the episodes of policy regression in the early 1990s suggest that aid cannot, however, buy reform. It raises an interesting question as to whether potential aid cutoffs by donors hold any credibility for countries that donors have "invested" in heavily, such as Ghana. The second hypothesis posits that nonfinancial aid (technical coop- eration, advisory services, and analytical work) has a better impact than financial aid on the generation of policy reforms in bad policy environ- ments. The first round of reforms in Ghana included rehabilitation of basic infrastructure and balance of payments support. Both of these re- quired more cash than intellect. Indications are that the lack of donor finance during the poor policy period was an important consideration in the decision to reform. But this paper has also shown that the macro- economic policy dialogue was useful to the reformers. In addition, their implementation requirements were not as onerous as some later reforms. While the mid-1980s economic and sector work that the World Bank carried out was important in the economic dialogue, it is unlikely that it would have generated reforms in the absence of finance. After all, the authorities had already convinced themselves of the basic framework for stabilization. The third and final hypothesis is that financial aid works when policy reforms and institution building are underway. Ghana's experience sug- gests that already existing policy reforms are probably necessary (though not sufficient) for financial aid to work effectively. The existence of the reforms or a program for reforms implies prior commitment by the gov- ernment and indicates that debate regarding reform's merits or demer- its has taken place already. Institution building appears to be neither necessary nor sufficient in the short run for financial aid to work. The Economic Recovery Program was implemented in a situation where ex- 90 Successful Reformers tremely weak institutions inherited from the previous regime were the norm. Yet, the ERP gets high marks for its ability to introduce effectively a wide-ranging set of economic policies. The reason, as discussed above, was the high-caliber (though small) group of policymakers who remained in place for a long period. This was, however, a unique situation. With the inevitable transition in the group and the need to institutionalize decisionmaking, it has proven more difficult to use reform-based aid effectively. Our study suggests that the relationship between democracy and re- form in Ghana has been complex. Donors have expended much energy and resources pushing democratization, in the belief that political and economic liberalization go hand in hand. Ghana's experience illustrates that democratization will not necessarily lead to good economic policy. Ghana 91 APPENDIX 2.1 GHANA: CHRONOLOGY OF POLITICAL AND ECONOMIC DEVELOPMENTS 6 March 1957 Ghana wins independence from Britain; Nkrumah is prime minister. 1 July 1960 Inauguration of the First Republic. 24 February 1966 Coup d'etat by National Liberation Council overthrows Nkrumah. 22 August 1969 National Liberation Council hands over power to civilians. Second Republic inaugurated. Busia and Progress Party rule. 13 January 1972 Coup d'etat led by Acheampong and the National Redemption Council. October 1975 National Redemption Council superseded by Supreme Military Council. 5 July 1978 Palace coup; Akuffo overthrows Acheampong, and Supreme Military Council II emerges. 4 June 1979 Uprising by junior officers; Armed Forces Revolutionary Council comes to power. 18 June 1979 Elections. 24 September 1979 Inauguration of Third Republic. Limann is president. June 1981-January 1983 IDA lending suspended to Ghana. 31 December 1981 Coup d'etat by Rawlings's Provisional National Defence Council. 28-29 October 1982 Attempted coup. 23 November 1982 Attempted coup. 30 December 1982 Secretary for finance and economic planning announces Program for Reconstruction and Development. January 1983 1.2 million Ghanaians expelled from Nigeria. April 1983 Budget outlines Economic Recovery Program to be supported by financial assistance from World Bank and IME 19 June 1983 Coup attempt. November 1983 First Consultative Group meeting for Ghana in 13 years. November 1984 People's Defence Committees and Workers' Defence Committees replaced by Committees for the Defence of the Revolution. 18 May 1992 Ban on political parties lifted. 3 November 1992 Presidential elections. Rawlings wins. 29 December 1992 Parliamentary elections, boycotted by opposition. Late 1992-early 1993 Adjustment credit tranche releases withheld. 7 January 1993 Fourth Republic inaugurated. 92 Successful Reformers 1994 Tranche releases withheld on World Bank Financial Sector Adjustment Credit and Agricul- tural Sector Adjustment Credit due to privatization delays. 1996 Presidential and parliamentary elections. Rawlings wins 57 percent of popular vote. National Democratic Congress wins 66 percent of parliamentary seats. Opposition accepts results. 1996 ESAF negotiations suspended due to slippage. Ghana 93 APPENDIX 2.2 ADDITIONAL DEVELOPMENT FINANCE DATA TABLE 2.3 GHANA: EFFECTIVE DEVELOPMENT ASSISTANCE, 198o-96 (millions of U.S. dollars) GQ GQ (multi- Year TOTAID EDA GQ Grants TA (bilateral) lateral) 1970 22.3 7.2 1.7 5.5 15.2 0.0 1.7 1971 37.1 21.1 12.7 8.4 15.9 1.2 11.6 1972 40.3 25.4 17.6 7.8 15.0 3.7 13.9 1973 37.9 20.3 14.7 5.5 17.7 1.4 13.3 1974 39.7 19.9 7.9 12.0 19.8 5.0 2.9 1975 53.5 27.7 19.8 7.8 25.9 7.6 12.2 1976 67.8 40.3 20.4 20.0 27.5 12.5 7.9 1977 88.9 59.7 34.1 25.7 29.2 20.6 13.4 1978 98.3 60.9 30.5 30.4 37.5 13.5 17.0 1979 118.9 83.5 51.8 31.7 35.4 4.4 47.4 1980 140.2 97.8 75.3 22.5 42.4 11.5 63.8 1981 131.4 89.0 55.4 33.7 42.3 13.9 41.5 1982 110.7 76.1 46.2 29.8 34.6 18.6 27.6 1983 130.3 101.2 55.0 46.2 29.2 21.8 33.2 1984 189.4 164.1 58.3 105.8 25.3 46.1 12.2 1985 181.3 150.7 76.1 74.6 30.6 55.9 20.2 1986 345.0 306.1 172.6 133.5 39.0 142.8 29.8 1987 336.9 290.8 190.3 100.5 46.1 162.7 27.6 1988 436.6 376.6 206.0 170.7 60.0 165.2 40.8 1989 454.1 394.7 186.3 208.4 59.4 132.9 53.3 1990 707.4 645.3 196.5 448.7 62.1 161.8 34.8 1991 779.6 697.6 227.6 470.0 82.0 162.9 64.8 1992 489.2 386.7 168.9 217.8 102.5 130.7 38.2 1993 507.3 409.3 186.9 222.4 98.0 153.9 33.1 1994 483.3 395.1 176.8 218.3 88.2 130.7 46.1 1995 559.1 450.8 212.6 238.2 108.3 172.4 40.2 1996 523.0 426.5 207.7 218.8 96.5 177.3 30.4 Notes: TOTAID = Effective development assistance + TA (GQ + grants + TA). EDA = Effective development assistance (GQ + grants). GQ = Grant equivalent of ODA loan. TA = Technical assistance. Source: Chang, Fernandez-Arias, and Serven 1999. TABLE 2.4 GHANA: EDA LOANS AND DONOR GRANTS BY SECTOR, 1980-96 (millions of U.S. dollars) Electricity, Financial, General Trade, Balance of gas and insurance, purpose Other restau- Transport payments Communi- Social Construc- Current water real estate, contri- Manufac- Mining, contri- rants, and Year support cation services tion imports production business butions turing quarrying butions lodging storage 1970 0.00 0.00 0.40 0.00 1.30 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1971 0.00 0.00 5.47 0.00 6.11 1.11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1972 0.00 0.00 4.40 0.00 8.27 4.69 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1973 0.00 0.00 4.90 0.00 7.65 1.61 0.00 0.00 0.18 0.00 0.00 0.00 0.00 1974 0.00 0.00 1.19 0.00 0.90 0.86 0.00 0.00 4.10 0.00 0.00 0.00 0.05 1975 1.46 0.00 1.55 0.00 1.18 4.69 0.00 0.00 2.36 0.00 0.00 0.00 7.12 1976 3.01 -0.02 1.84 0.00 2.79 1.09 0.03 0.00 2.29 0.00 0.00 0.00 7.65 1977 6.88 -0.01 0.21 0.00 2.95 9.44 0.00 0.00 2.96 0.00 0.00 0.00 9.22 1978 0.63 -0.03 2.17 0.00 1.24 19.35 -0.02 0.00 3.09 0.00 0.00 0.00 2.39 1979 0.28 1.63 3.78 0.00 17.51 15.68 0.07 0.00 6.04 0.00 0.00 0.00 0.12 1980 2.20 -0.15 2.80 0.00 26.26 24.62 0.04 0.88 13.27 0.00 0.00 0.00 0.14 1981 0.00 -0.06 5.75 0.00 20.36 12.33 1.12 0.00 4.58 0.00 0.00 0.00 10.04 1982 0.00 0.03 2.76 0.00 1.85 24.64 3.94 0.00 2.80 0.00 0.00 0.00 7.61 1983 0.00 0.05 15.90 0.00 15.56 10.05 3.72 0.00 1.51 0.00 0.00 1.23 4.73 1984 1.41 1.85 4.04 1.03 20.37 6.59 4.70 0.00 1.93 0.00 4.39 0.00 9.88 1985 19.27 1.67 3.30 0.51 8.08 8.13 2.54 0.00 2.11 -1.02 6.60 0.00 19.31 1986 73.31 0.12 6.85 16.10 9.93 8.12 -0.01 0.00 6.50 -1.27 23.35 0.00 24.22 1987 89.39 0.32 7.75 10.83 4.31 16.56 -1.88 0.00 7.59 -2.28 14.61 0.93 34.36 1988 94.22 16.03 13.20 3.86 5.33 19.49 -0.50 0.00 12.67 -1.56 4.33 1.12 26.73 1989 110.21 7.05 11.62 4.97 2.91 8.20 -0.03 0.00 6.32 5.60 1.32 0.00 23.10 1990 121.62 1.70 14.44 6.45 4.11 15.26 -0.04 0.00 7.05 2.55 0.00 0.00 11.77 1991 94.91 2.97 64.15 6.83 0.00 15.21 0.00 0.00 12.62 5.39 0.00 0.00 16.13 1992 22.60 2.34 19.06 6.95 0.00 12.87 29.55 0.00 8.13 4.23 0.00 0.00 24.18 1993 21.78 2.12 70.74 6.14 0.00 10.20 33.92 0.00 3.87 2.36 0.00 0.00 19.01 1994 28.07 1.68 57.41 10.81 0.00 9.99 1.93 0.00 4.09 3.42 0.00 0.00 23.93 1995 9.61 0.71 79.04 14.17 0.00 15.19 3.97 10.06 1.11 3.21 0.00 0.00 32.52 1996 3.45 1.26 84.81 12.99 0.00 35.24 1.65 10.63 0.12 1.16 0.00 0.00 36.17 Source: Chang, FemnTdez-Arias, and Serven 1999. Ghana 95 NOTES 1. See, for example, van de Walle and Johnston (1996); Dollar and Burnside (1997); World Bank (1998); Dollar and Svensson (1998). 2. In January 1972 Colonel I. K. Acheampong took power in a coup d'etat. One of the first actions of his National Redemption Council was to repudiate several of the country's foreign debts. A popular cry at the time was "yentua, yentua" (we won't pay, we won't pay). 3. The policymakers interviewed expressed differing opinions as to whether the fiscal targets were missed solely or even primarily because of the wage increase. 4. The Ghanaian government accused the United States of trying to over- throw it, and subsequent events lent some credence to these accusations. In 1985, a sting operation involving an American CIA agent resulted in the ex- change of a Ghanaian intelligence agent arrested in the United States for several Ghanaian CIA operatives. In March 1986 a ship carrying American mercenaries headed for Ghana was arrested off the coast of Brazil. 5. Frimpong-Ansah (1991) notes that the word is assumed to derive from the Hausa kara bude meaning "keep it quiet" or "hide it." 6. Dr. Limann's academic qualifications included a B.Sc. from the London School of Economics, a master's degree in constitutional law, and a doctorate in political science from the University of Paris. He had also served as Ghana's representative to the International Labour Organisation and the World Health Organization. 7. The initial choice, Imoru Egala, who founded the PNP, was under a 12- year ban from public office dating to 1969. He was appealing this ban at the time of the election and was hence ineligible to run for president. 8. Limann confirmed this in a 1982 interview (West Africa, 10 May). In re- sponse to a question regarding the virtual paralysis of the government, he re- sponded: "Yes. For more than two weeks we did nothing but these continual meetings, the whole day. I was really angry ... I warned them that this was not going to do them or anybody any good." 9. In the second half of 1981, PNP high-level functionaries begun to fight among themselves over the issue of future party policy in the run-up to the next election. This resulted in a member of the PNP national executive committee issuing a writ in the Accra High Court in November 1981 to restrain the party hierarchy (chairman, general secretary, and publicity) "from railroading through party policy in advance of the national party congress." While the case was ultimately settled out of court, the three officers were ordered to refund within 21 days any monies they had received on behalf of the party. This was followed by the revelation that a senior member of the PNP had received a commission worth 2.7 million pounds sterling on a currency printing contract that was go- ing to cost 22 million pounds total. 10. This role of the state, which has been common in Africa and elsewhere, does not necessarily depend on ideology. Two countries with very different economnic 96 Successful Reformers philosophies-South Africa and Tanzania-each used public sector employment to fulfill their social contract with the people. In the case of South Africa under apartheid the state essentially provided guaranteed lifetime jobs for Afrikaners, while in Tanzania the underlying rationale was to fulfill a socialist vision. 11. See Frimpong-Ansah (1991: 110-11) and Ocquaye (1980) for accounts of the system of state patronage during this era. 12. The literature includes, for example, Hansen (1991), Herbst (1991), Ninsin (1998), Nugent (1996), Rothchild (1991), Shillington (1992), and Yeebo (1991). 13. Striking workers from the Ghana Industrial Holdings Corporation were dismissed following their occupation of the Parliament buildings in June 1990. The organization set up to support the sacked workers, the Workers' Solidarity Front, later transformed itself into the Association of Local Unions. This in turn became a militant alternative to the Trades Union Congress. 14. There was some basis for these fears. The government of Nigeria had cut off oil supplies to Ghana during the AFRC era (1979). Having gone through its own political transition, the new Nigerian democratic government was wor- ried about the "contagiousness" of Ghana's coup. The anti-imperialist rhetoric of the PNDC and its close ties with Libya and Cuba worried the Americans. Subsequently, evidence emerged around the time of the "spy swap" showing that the CIA had been actively involved in trying to topple Rawlings during the first few years of PNDC existence. 15. Some policymakers interviewed interpreted the fact that the secretary of finance was the last position to be filled as evidence both of PNDC awareness of the position's importance and of the struggle among PNDC supporters. 16. Chazan (1991) points out that "[while] the ideologues ... stressed the significance of the capture of the state as the first step toward revolutionary change, emphasis [was] placed by Rawlings on moral rectitude, accountability, reciprocity, and ethical reform. While other leaders of the PNDC underlined the revolutionary nature of the regime, Rawlings expressed himself in missionary and prophetic terms. The former employed the tools of class analysis, whereas Rawlings consistently focused on the people in relation to officialdom. But since the influence of the neo-Marxist intelligentsia on the political education of Jerry Rawlings was still strong at this time, the class-based view prevailed." 17. The description of this meeting is based on interviews with key persons present at the meeting and on Shillington (1992). 18. While financial aid was not forthcoming from most of these countries, Libya played a crucial role by providing 28,000 metric tons of crude oil in July 1982. Over the next six months 12 such consignments were provided, each on free credit for a year and interest-free repayment after that. 19. Nugent (1996) describes some links between the two coups and suggests that some alliances had emerged between unlikely bedfellows. 20. The PNDC had held local govemment elections in 1988, but they also intensified pressure for national elections. 21. The National Convention Party entered into an alliance with the NDC, earning it a vice-presidential slot. Ghana 97 22. Dr. J. B Danquah was an important leader in the United Gold Coast Con- vention political party before independence. Nugent (1996) uses the "Busia- Danquah tradition" to refer to a political tendency that veers to the right, relatively speaking. 23. The Convention People's Party was political successful in the late 1950s and early 1960s and was known for its ability to mobilize effectively at the rural grassroots level. 24. This section draws on Jeffrey Herbst, "Labor in Ghana Under Structural Adjustment: The Politics of Acquiescence," in Rothchild (1991). 25. Rothchild (1991: 184) suggests that they lacked analytical skills to take on the government. 26. Nugent (1996) reports that some of these anti-student protests may have been stage-managed. Nonetheless, it appears undeniable that the students were unable to rally mass support for their cause. 27. In a 1989 interview (Africa News, 23 January), Dr. Botchwey indicated that approaching the exchange rate adjustment this way was a hard sell to the IME 28. Shillington (1992) describes the evolution of Rawlings's economic think- ing over this period. 29. This person, a respected Harvard-trained economist, had been an ad- viser to Limann in 1981 while on leave from the World Bank. He was a friend of key advisers in the PNDC and was invited, on several occasions, to advise the government on economic policy In addition, he was valued for his perspective on how certain policies would play at the Bank and Fund given his familiarity with the two institutions. 30. This view contrasts with Martin (1991) who claims that the government officials were too busy providing the IMF with data to carry out economic analysis. 31. Aryeetey (1997: 79) notes that even the World Bank's definition of "'own- ership' laid particular emphasis on the commitment of the leadership . . . this probably explains why the Bretton Woods institutions did little to encourage the government to consult other institutions." 32. These results are in line with a study by Deninger, Squire, and Basu (1998) that found that prior analytic economic and sector work improved project out- comes and had a high payoff. 33. One World Bank resident representative (a respected academic before joining the Bank) taught courses at the university. He was viewed very favor- ably for his ability and willingness to engage both policymakers and academics at the University of Ghana in economic issues that had no easy answers. 34. These findings regarding technical assistance are similar to those of Aryeetey and Cox (1997). 35. According to one member of the negotiating team, position papers were always prepared in advance and an internal consensus was reached before negotiations. The Ghanaian team also sometimes arrived early in Washington to consult with Ghanaian economists at the World Bank and the IMF. 36. Personal communication from World Bank staff member. 98 Successful Reformers REFERENCES Abbey, J. L. S. 1989. "On Promoting Successful Adjustment: Some Lessons from Ghana." 1989 Per Lecture. Per Jacobssen Foundation, Washington, D.C. Armstrong, R. P. 1996. "Ghana Country Assistance Review: A Study in Devel- opment Effectiveness." World Bank, Operations Evaluation Department, Washington, D.C. Aryeetey, E., and A. Cox. 1997. "Aid Effectiveness in Ghana." In J. Carlsson, G. Somolekae, and N. van de Walle, eds., Foreign Aid in Africa: Learningfrom Country Experiences. Uppsala, Sweden: Nordic Africa Institute. Chang, Charles C., Eduardo Femrndez-Arias, and Luis Serven. 1999. "Measur- ing Aid Flows: ANew Approach." Policy Research Working Paper 2050. World Bank, Development Research Group, Washington, D.C. Chazan, N. 1983. An Anatomy of Ghanaian Politics: Managing Political Recession, 1969-82. Boulder: Westview. -____ 1991. "The Political Transformation of Ghana under the PNDC." In D. Rothchild, ed., The Political Economy of Recovery. Boulder: Lynne Rienner. Chibber, Ajay, and Nemat Shafik. 1991. "The Inflationary Consequences of De- valuation with Parallel Markets: The Case of Ghana." In Ajay Chibber and S. Fischer, eds., Economic Reform in Sub-Saharan Africa. Washington, D.C.: World Bank. Cukierman, Alex, and Mariano Tommasi. 1994. "Why Does It Take a Nixon to Go to China?" Harvard University Processed. Deininger, K., L. Squire, and S. Basu. 1998. "Does Economic Analysis Improve the Quality of Foreign Assistance?" World Bank Economic Review 12 (Septem- ber): 385-418. Dollar, D., and C. Burnside. 1997. "Aid, Policies, and Growth." Policy Research Working Paper 1777. World Bank, Development Research Group, Washing- ton, D.C. Dollar, D., and J. Svensson. 1998. "What Explains the Success or Failure of Struc- tural Adjustment Programs?" Policy Research Working Paper 1938. World Bank, Development Research Group, Washington, D.C. Dordunoo, C., and V. K. Nyanteng. 1997. "Overview of Ghanaian Economic Development." In V. K. Nyanteng, ed., Policies and Optionsfor Ghanaian Eco- nomic Development. Legon, Ghana: Institute of Statistical, Social and Economic Research, University of Ghana. Frimpong-Ansah, J. H. 1991. The Vampire State in Africa: The Political Economy of Decline in Ghana. London: James Currey. "Ghana: Roots of Student Protest." 1983. West Africa (6 June): 1343-45. Ghana. 1983. "The PNDC's Programme for Reconstruction and Development." Statement by the PNDC secretary for finance and economic planning on ra- dio and television on December 30, 1982. Information Services Department, Accra. - 1987. "National Programme for Economic Development" (revised). Accra. Ghana 99 . 1996. "National Capacity Assessment." Accra. Haggard, S., and S. B. Webb, eds. 1994. Voting for Reform: Democracy, Political Liberalization, and Economic Adjustment. New York: Oxford University Press. Hansen, E. 1991. Ghana under Rawlings: Early Years. Lagos and Oxford: Malthouse Press. Herbst, J. 1991. "Labor in Ghana under Structural Adjustment: The Politics of Acquiescence." In D. Rothchild, ed., The Political Economy of Recovery. Boul- der: Lynne Rienner. . 1993. The Politics of Reform in Ghana, 1982-1991. Berkeley: University of California Press. Holmgren, T. 1998. "Aid and Reform in Africa: Terms of Reference." World Bank, Development Research Group, Washington, D.C. Processed. Husain, I., and R. Faruqee, eds. 1994. Adjustment in Africa: Lessonsfrom Country Case Studies. World Bank Regional and Sectoral Studies. Washington, D.C.: World Bank. Jebuni, C., A. Oduro, and K. A. Tutu. 1991. "Trade and Payments Liberalization and Economic Performance in Ghana." Paper presented at the African Eco- nomic Research Consortium workshop, Nairobi, Kenya. December. . 1994. "Trade and Payments Regime and the Balance of Payments in Ghana." World Development 22 (8). Kapur, I., M. T. Hadjimichael, P. Hilbers, J. Schiff, and P. Szymczak. 1991. "Ghana: Adjustment and Growth, 1983-91." Occasional Paper 86. International Mon- etary Fund, Washington, D.C. Killick, T. 1978. Development Economics in Action: A Study of Economic Policies in Ghana. New York: St. Martin's Press. Martin, M. 1991. "Negotiating Adjustment and Extemal Finance: Ghana and the Intemational Community, 1982-1989." In D. Rothchild, ed., The Political Economy of Recovery. Boulder: Lynne Rienner. Ninsin, K., ed. 1998. Ghana: Transition to Democracy. Legon, Ghana: Freedom Publications. Nowak, M., R. Basanti, B. Horvath, K. Lochar, and R. Prem. 1996. "Ghana, 1983- 1991." In M. T. Hadjimichael, M. Nowak, R. Sharer, A. Tahari, and a staff team from the Africa Department, eds., Adjustment for Growth: The African Experi- ence. Occasional Paper 143. Intemational Monetary Fund, Washington, D.C. Nugent, P. 1996. Big Men, Small Boys and Politics in Ghana. Accra: Asempa Pub- lishers, Christian Council of Ghana. Nyanteng, V. K. 1997. Policies and Options for Ghanaian Economic Development. Legon, Ghana: Institute of Statistical, Social and Economic Research, Univer- sity of Ghana. Ocquaye, M. 1980. Politics in Ghana, 1972-79. Accra: Tomado Publications. Rothchild, D., ed. 1991. The Political Economy of Recovery. SAIS African Studies Library. Boulder: Lynne Rienner. Sackey, H. A. 1998. "External Aid and Real Exchange Rate in a Developing Economy: The Ghanaian Situation." Paper presented at the African Economic Research Consortium workshop, Nairobi, Kenya. 100 Successful Reformers Shillington, K. 1992. Ghana and the Rawlings Factor. London: Macmillan. Toye, J. 1990. "Ghana's Economic Reforms, 1983-87: Origins, Achievements and Limitations." In James Pickett and H. Singer, eds., Towards Economic Recovery in Sub-Saharan Africa: Essays in Honour of Robert Gardiner. London and New York: Routledge. - 1991. "Ghana." In Paul Mosley, Jane Harrigan, and John Toye, eds., Aid and Power: The World Bank and Policy-Based Lending. Vol. 2. London and New York: Routledge. van de Walle, N., and T. A. Johnston, eds. 1996. Improving Aid to Africa. Policy Essay 21. Washington, D.C.: Overseas Development Council. World Bank. 1998a. Assessing Aid: What Works, What Doesn't, and Why. New York: Oxford University Press. _____. 1998b. "Ghana: Country Assistance Strategy." Washington, D.C. Yeebo, Z. 1991. Ghana: The Strugglefor Popular Power. Rawlings: Saviour or Dema- gogue. London: New Beacon Press. Younger, S. 1992. "Aid and the Dutch Disease: Macroeconomic Management When Everybody Loves You." World Development 20 (11): 1587-97. Uganda Torgny Holmgren Ministry of Foreign Affairs, Sweden Louis Kasekende Michael Atingi-Ego Daniel Ddamulira Bank of Uganda Kampala, Uganda ACRONYMS AND ABBREVIATIONS AfDB African Development Bank ADF African Development Fund CG Consultative Group CPIA Country Policy and Institutional Assessment EDA Effective development assistance EPRC Economic Policy Research Centre ERC Economic and Recovery Credit ERP Economic Recovery Program ESAF Enhanced Structural Adjustment Facility (of the IMF) GDP Gross domestic product HIPC Heavily indebted poor countries ICOR Incremental capital-output ratio IDA Intemational Development Association (of the World Bank Group) IMF Intemational Monetary Fund MFPED Ministry of Finance, Planning and Economic Development NGO Nongovernmental organization NRM National Resistance Movement ODA Official development assistance ODF Official development finance PEC Presidential Economic Council SAC Structural Adjustment Credit SAD Sector Adjustment Credit SAF Structural Adjustment Facility (of the IMF) SAL Structural Adjustment Loan sh Uganda shilling SIL Sector Investment Loan TA Technical assistance TAL Technical Assistance Loan UDN Uganda Debt Network UMA Uganda Manufacturers Association UNDP United Nations Development Programme UNICEF United Nations Children's Fund URA Uganda Revenue Authority Uganda's development over the last two decades makes a fascinat- Ling case study of the causes of policy reforms and the links between aid and reforms. Starting from an unstable, uncertain, and insecure eco- nomic situation, the country has managed to develop a successful re- form program. Aid has certainly been part of this development. This study seeks to determine what role it has played and what lessons can be leamed from the Ugandan experience. Uganda received an increasing amount of aid from 1980 to 1995, after which aid started to level off. Until recently aid was provided mainly as loans from multilateral institutions, in contrast to the typical pattern for African countries. Bilateral assistance and grants only came to play the predominant role in the 1990s. Roughly half of total assistance since the mid-1980s has been associated with structural adjustment programs. Technical assistance makes up a large fraction of the aid and has amounted to about one-third of the grants. In addition, Uganda was the first country to receive support through the Heavily Indebted Poor Coun- tries (HIPC) Debt Initiative. Aid coordination improved considerably during the 1990s, with the government playing a more active role and with intensified local donor coordination. Uganda's reform program has been one of the most successful in Sub- Saharan Africa. The gross domestic product (GDP) grew by an average 6.4 percent during 1987-96, and inflation has been reduced from more than 100 percent in 1987 to single-digit figures. The government has achieved both stabilization and substantial decontrol of the economy. While its early stance was marked by rejection of market-based reforms (1986), followed by reluctant implementation (1987-92), the government has since then moved to full ownership of such reforms. The basic fac- tors behind the success of the reforms include strong political leader- ship, a process of learning from mistakes, evolving government commitment, and in recent years, broadening domestic ownership of reforms. A more active policy dialogue is now taking place involving the Parliament, the private sector, trade unions, nongovernmental orga- nizations, the media, and other actors. The challenge ahead is to further develop and sustain the reform program as well as to broaden and deepen the domestic ownership. Aid in various forms helped to support the generation and imple- mentation of the policy reforms. When the government reluctantly in- troduced market-based reforms, the policy dialogue, advisory services, training, and technical assistance provided by donors were of critical importance, both for the decisions to reform and for the direction of the reforms. The Ugandan government has made unusually good use of tech- nical assistance. Study visits to countries within and outside the region, notably Ghana, were also very valuable during this period. 103 104 Successful Reformers When the government decided to reform in the late 1980s, financial aid and conditionality became a powerful tool. Conditionality, which earlier had been regarded as externally imposed, was now used by re- formers within the government to help push the reforms forward. It was pivotal in generating, implementing, and cementing the reforms. With a very weak revenue base, the financial support was, and still is, neces- sary for implementation of most reforms. Since 1992, however, govern- ment ownership of the reforms has meant that conditionality has become less relevant for inducing reforms. From the donors' perspective, policy dialogue and advisory services can once again become the most impor- tant instruments for supporting the generation of reforms. This study is based on a review of current literature, analysis of aid and policy data, and interviews conducted in Kampala in March 1999 (see appendix 3.2). It consists of five parts. Section 2, which follows, de- scribes aid flows to Uganda, highlighting aid linked to the support of policy reforms. Section 3 provides an overview and analysis of the country's economic policy and institutional development. Section 4 ad- dresses the overall question of the links between aid and reform in Uganda, and section 5 concludes by presenting some of the lessons drawn. AID TO UGANDA In this paper the term aid will refer to effective development assistance (EDA).1 This measure captures the true grant element of aid flows more accurately than does the commonly used official development assistance (ODA). We will also use the measure official development finance (ODF), which in addition includes nongrant components of official loans, in order to capture all official flows. In this way we include all grants and loans to Uganda that are undertaken by the official sector in order to promote economic development and welfare. Loans from the International Mon- etary Fund (IMF) for structural adjustment programs, in the form of the Structural Adjustment Facility (SAF) and Enhanced Structural Adjust- ment Facility (ESAF), are included in our analysis, as are technical assistance grants. But we exclude assistance mobilized directly by non- governmental organizations and any form of military assistance given to Uganda.2 Official Aid Flows Uganda's post-independence experience with aid flows can be divided into three periods: 1962-78, 1979-85, and 1986-97 (see table 3.6 in Ap- pendix 3.1). Commitments were tiny during the first period, amounting only to US$357 million (4 percent of total aid since independence). A Uganda 105 FIGURE 3.1 OFFICIAL DEVELOPMENT FINANCE FLOWS TO UGANDA, 198o-96 Millions of 1995 U.S. dollars 1,000 = Nongrant component of official loans 900 - Grant equivalent of official loans 80- Technical assistance grants 800 _ Non-technical assistance grants/\ 700 600 500 / /_ 400 /l 300 ~ 100_ 0- ,9P, ,%1i E%9 & ,r9^v %1 qS) b c ' co\ lb 2 99N kqZ '0 ¢b q11 P E Ri0 Note: The figure shows official development finance (ODF) as well as effective development assistance (EDA) for 1980-96. The three lower graphs add up to EDA, and all graphs together show ODF. Source: World Bank. striking feature of these years is the low level of support from multilat- eral donors, especially the World Bank, and the general absence of West- ern countries. Very few bilateral donors committed resources to Uganda. Factors that discouraged the interest of Western donors in this era in- cluded the nature of the political regimes in Uganda, the tendency of the regimes to lean toward socialism, and poor macroeconomic manage- ment. Only after the fall of Idi Amin in 1979 did the World Bank and a number of European bilateral donors step in. This marked the second phase of aid relationships. The third phase and the real take-off in terms of aid volume occurred in the mid-1980s, when the current reform program started. Yoweri Museveni and the National Resistance Movement (NRM) assumed power in January 1986. The economy had at that time shrunk by more than 20 percent from its peak value in 1970, and inflation had an annual rate in excess of 240 percent. As shown in figure 3.1, Uganda has re- ceived an increasing amount of aid since 1986. The first half of the 1980s 106 Successful Reformers was characterized by a rather steady flow of EDA, around US$150 mil- lion to US$200 million a year (1995 U.S. dollars). A sharp upturn began in 1987, with continuous growth in annual aid flows until 1995, when aid started to level out. Official loans made up the major part of ODF in the 1980s, while grants, including technical assistance, have taken over that role in the 1990s. Support from multilateral sources has dominated official development finance over the years, but it has decreased from 3:1 to 1:1 in relation to the bilateral flows over the 16-year period studied. Bilateral ODF flows be- came larger than the multilateral flows in 1996. This had already been the situation with regard to EDA flows since 1991 (except for 1992). Multilateral institutions have thus controlled the aid scene for a num- ber of years in Uganda. They have set the tone for the types of aid to be provided and sectors to be supported. With bilateral flows now larger than multilateral flows it remains to be seen whether the bilateral orga- nizations will determine the direction of aid in the future. Financial Aid The distribution of financial aid between loans and grants has been fairly even over the period as a whole, with 48 percent loans and 52 percent grants including technical assistance.3 As shown in table 3.1, loans domi- nated the aid flows during the 1980s, when they accounted for some 50- 60 percent of the annual ODF flows. Grants have, however, increased steadily over the years and outnumber loan disbursements since 1990. The overall grant program took a giant leap in 1987-88 when it almost doubled in size compared to the 1986 figure. Since 1991 the average level of the grant program is US$455 million a year. Bilateral donors channel their aid mainly as grants, which on average have been 2.5 times the size of the grants by the multilateral institutions. The increase in grants as a share of total aid during this period can be attributed to two factors: (a) the increase in the number of bilateral do- nors (mainly the Nordic countries), and (b) the government's deliberate decision in 1992 to control its debt by limiting financing of its activities first to grants and then to concessional loans with a grant element of at least 75 percent. The multilateral institutions have directed a large part of their assis- tance as policy-based lending oriented to reform. Fast-disbursing sup- port for policy reforms has played a significant role in aid to Uganda, especially in 1987-92 when the basis of the reform program was taking shape. During this period, as discussed in more detail below, financial aid and conditionality were critical in driving the reform program. Bi- lateral donors in turn have provided substantial co-financing to the ad- justment operations. For example, World Bank Structural Adjustment TABLE 3.1 UGANDA: TOTAL DISBURSEMENTS ON LOANS AND GRANTS (ODF), 198o-96 (mnillions of current U.S. dollars) Type of aid 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Loans 95.8 207.3 173.1 199.4 106.4 153.5 135.5 317.4 207.3 321.1 343.5 258.9 305.7 391.4 323.7 288.2 288.5 Grants 82.7 134.3 85.0 110.5 98.9 84.3 131.6 153.8 262.8 264.2 351.3 421.4 458.0 383.9 443.7 547.3 478.6 Total 178.5 341.6 258.1 309.9 205.3 237.8 267.1 471.2 470.1 585.3 694.8 680.3 763.7 775.3 767.4 835.5 767.1 Source: Chang, Fernandez-Arias, and Serven 1999. 108 Successful Reformers TABLE 3.2 WORLD BANK AND IMF POLICY REFORM LOANS TO UGANDA, 1980-98 World Bank (millions of U.S. dollars) Amount Approval Closing Loan Type approved year year Reconstruction Credit ERC 72.5 1980 1982 Technical Assistance TAL 8.0 1980 1985 Reconstruction Credit II SAL 70.0 1982 1985 Technical Assistance II TAL 51.3 1983 1992 Agricultural Rehabilitation Sector Adjustment Credit SAD 66.1 1983 1992 Reconstruction III SAL 50.0 1984 1987 Economic Recovery Credit SAL 65.0 1987 1991 Technical Assistance III TAL 18.0 1988 1995 Public Enterprises TAL 15.0 1988 1995 Economic Recovery I/II Supplement SAL 26.7 1989 Economic Recovery Credit II SAL 125.0 1990 1993 Poverty and the Social Cost of Adjustment SIL 28.0 1990 1995 Agricultural Sector Adjustment SAD 115.0 1990 1996 Structural Adjustment Credit I SAL 125.0 1991 1994 Economic and Financial Management Technical Assistance TAL 29.0 1992 1999 Institutional Capacity Building TAL 36.4 1992 2000 Financial Sector Adjustment Credit SAD 100.0 1993 1997 Structural Adjustment Credit II SAL 80.0 1994 1996 Structural Adjustment Credit III SAL 125.0 1997 1999 Education Sector Adjustment Credit SAD 80.0 1998 (continued on next page) Credits (SAC I and III) through the International Development Associa- tion were co-financed with US$69.4 million and US$50.0 million respec- tively. Table 3.2 lists all IMF and World Bank loans for policy reforms during the period 1980-98. Balance of payments support and IMF disbursements together have accounted for approximately 50 percent of the official development lend- ing during the study period. Table 3.3 shows the amount of import sup- port provided in the 1990s as part of the overall balance of payments support. The support corresponds to approximately one-quarter of over- all ODF during those years. The social sectors accounted for 15 percent of the aid flows, while the primary sectors (agriculture, forestry, fish- ing), power and electricity, manufacturing, and transport accounted for Uganda 109 TABLE 3.2-continued International Monetary Fund (1980-84 in millions of SDR, 1987-97 in millions of U.S. dollars) Amount Amount Loan Time Year approved drawn Standby Operation 1 year 1980 12.5 12.5 Trust Fund Loan 1 year 1980 22.3 22.3 Compensatory Financing Facility 1 year 1980 25.0 25.0 Compensatory Financing Facility 1 year 1981 45.0 45.0 Standby I 1 year 1981/82 112.5 112.5 Standby II 1 year 1982/83 112.5 112.5 Standby III 1 year 1983/84 95.0 65.0 Structural Adjustment Facility 2 years 1987 69.7 49.8 ESAF 3 years 1989 179.3 179.3 ESAF, additional arrangement 1 year 1992 39.8 39.8 ESAF 3 years 1994 180.0 180.0 ESAF 3 years 1997 138.0 78.Oa - Not available. Note: ERC = Economic Recovery Credit. SAL = Structural Adjustment Loan/Credit. SAD = Sector Adjustment Credit. TAL = Technical Assistance Loan. SIL = Sector Investment Loan. ESAF = Enhanced Structural Adjustment Facility. a. Disbursements until end of 1998. Sources: World Bank; IMF. 6-7 percent each over the period (see table 3.15 in Appendix 3.1 for a detailed sector distribution). Uganda has also received commodity support from the United States, Denmark, Japan, Austria, and the World Food Programme. Since 1987, a total of US$99 million has been committed. The management of com- modity aid was very weak before 1992. It was only then, following the merger of the Finance Ministry and the Planning Ministry, that the new Ministry of Finance and Economic Planning4 was able to incorporate all the funds generated from the various commodity aid programs into the budgeting framework. Nonfinancial Aid Technical assistance (TA) grants are the only category of nonfinancial aid that is registered in the official development aid statistics.5 TA grants have expanded from about 10 percent of all ODF flows to Uganda in the 110 Successful Reformers TABLE 3.3 UGANDA: IMPORT SUPPORT, 1991/92-1996/97 (millions of U.S. dollars) Source of aid 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 Multilateral 77.62 62.94 101.04 115.90 54.37 69.09 Bilateral 108.10 120.56 53.79 63.11 31.76 41.71 Total 198.60 196.43 194.65 220.17 138.32 164.10 Source: Bank of Uganda. early 1980s to 16-17 percent on annual basis since 1992. However, the relative share of TA reached a peak in 1986 shortly after the new govern- ment came to power (see tables 3.7 and 3.8 in Appendix 3.1). TA made up almost one-third of total grants during the period 1980- 96. Bilateral TA grants have increased from 40 percent of all TA grants in the mid-1980s to more than 70 percent since 1994 (see table 3.14 in Appendix 3.1). The United States, the United Kingdom, and Germany are the main bilateral providers of TA, accounting for nearly two-thirds of the total. Since 1991, Denmark, the Netherlands, and Sweden have also been substantial providers of grant-based TA. The United Nations Development Programme (UNDP) and the United Nations Children's Fund (UNICEF) have been responsible for 76 percent of the total mul- tilateral grant-based TA. Both institutions increased their TA substan- tially in 1988-89. Technical assistance can take the form of freestanding TA loans or can be part of project loans. The World Bank provided six freestanding TA loans to Uganda between 1980 and 1996 for a total amount of US$158 million, corresponding to 12.3 percent of total funds committed by the Bank (see table 3.2). Five of these operations had as their central objec- tive to help core economic ministries and agencies design and imple- ment economic reforms, and to strengthen systems and capacities within those agencies. Evaluations show that Uganda has made unusually good use of tech- nical assistance, especially TA directed to the finance and planning min- istries and the Bank of Uganda. We will discuss this further in the section on financial versus nonfinancial aid, and examine the impact of nonfi- nancial aid on the policy formulation process. Debt Relief Uganda has been burdened by heavy debt for the last 15 years. The debt ratios were extremely high in the early 1990s, with total debt-to- export ratios over 1,400 percent and debt-service ratios over 60-70 per- Uganda 111 cent, well above what are commonly regarded as sustainable debt levels (see figure 3.2). The external debt increased in lockstep with the economic reform pro- gram during the 1980s. The total debt stock increased by some 16-17 percent annually toward the end of the decade. At the end of June 1996, public and publicly guaranteed external debt amounted to US$3,147 million, or 62.7 percent of GDP. Uganda has borrowed mainly from multilateral institutions. They accounted for most of the external debt at that time (76 percent), followed by non-Paris Club bilateral creditors (11 percent), Paris Club creditors (10 percent), and commercial creditors (3 percent). Among the multilateral institutions, the International Devel- opment Association (IDA) of the World Bank Group was the single larg- est creditor (53 percent), followed by the IMF (12 percent), and the African Development Bank (IDA 1997). Because of its severe debt and its credible reform program, Uganda became the first country to be declared eligible for assistance under the Heavily Indebted Poor Countries (HIPC) Debt Initiative.6 Uganda reached the "decision point" in April 1997 and the "completion point" one year later, receiving relief of close to US$650 million (US$350 mil- lion in net present value) from its external creditors. Assistance by IDA FIGURE 3.2 UGANDA: DEBT INDICATORS, 1980-97 Percent Percent 120 1,600 External debt as a External debt percentage of 1,400 100 as a percentage exports of goods of GNP and services 1,200 80 (left axis) (right axis) 1,000 60 800 600 40 Debt service as a percentage of 400 20 exports of goods 20 ~ and services 200 (left axis) 0 0 Source: World Bank. 112 Successful Reformers under the HIPC Initiative was conditional on structural and social devel- opment performance criteria agreed with the government and monitored under IDA-supported programs. The bilateral creditor governments in the Paris Club have agreed to debt rescheduling for Uganda on six occasions since 1980 (see table 3.17 in Appendix 3.1). Uganda was the first country to receive debt resched- uling on Naples terms in February 1995, corresponding to a reduction in eligible debt of 67 percent in net present value terms. The government is now committed to borrowing only on highly concessional terms and grant financing as part of its strategy to achieve debt reduction and sustainability. The effort is also reflected in table 3.18 in Appendix 3.1, which shows an increase in donors' debt relief disburse- ments from US$12.9 million in 1991/92 to US$53.31 million in 1996/97. Before the introduction of the HIPC Initiative, repayment of multilateral debt was eased by a number of bilateral donors through the Multilateral Debt Fund. Disbursements from this fund equaled the magnitude of debt relief eventually provided under the HIPC Initiative annually. On commercial debt, in February 1993 Uganda concluded an agree- ment sponsored by the IDA Debt Reduction Facility to purchase US$153 million of eligible principal debt owed to commercial banks. The debt was bought back at 12 cents per dollar, at a total cost of US$22.6 million. Bilateral and Multilateral Institutions Bilateral Aid The bilateral donors have accounted for 50 to 60 percent of total aid (EDA) to Uganda during the 1990s. Bilateral assistance took a dramatic turn upward in 1988, when EDA more than doubled from its 1987 level of US$83 million. The shift was closely related to the improvements in political arrangements with the new government from 1986 on, and the change in economic policy that occurred in mid-1987. The government's economic reforms clearly paid off in terms of more aid flowing in, and the donors adapted quickly to the improved circumstances for utilizing aid. This change occurred just as the Cold War was coming to an end, prompting some donors to re-evaluate their geostrategic reasons for aid allocations. Bilateral donors have extended their assistance principally in the form of grants (see figure 3.3). Grants including technical assistance accounted for 85 percent of ODF between 1980 and 1996. The share has grown even stronger in the 1990s, with grants amounting to 89 percent of total bilat- eral flows. Technical assistance makes up about one-quarter of this. Since aid from the bilateral donors has increased in size, in both absolute and relative numbers, Uganda now receives aid on softer terms than before. Uganda 113 FIGURE 3.3 UGANDA: ODF AND EDA FROM BILATERAL DONORS, 198o-96 Millions of 1995 U.S. dollars 500 I Nongrant component of official loans 450 - Grant equivalent of official loans - Technical assistance grants 400 - Non-technical assistance grants 350- 300 - 250 - 200 - 150 - 100 - 50 0 - Source: World Bank. Some 21 countries have been active donors since 1986, with a core group of seven countries that each provided more than US$10 million annually during the 1990s (tables 3.9 and 3.10 in Appendix 3.1 list all donors to Uganda). Before 1988 only four donors had annual programs of more than US$10 million (the United Kingdom, the United States, Germany, and Italy). The United Kingdom remains the largest bilateral donor, mainly because of the historical ties, and the United States stands second in grant allocation to Uganda over this period (1989-96). Other significant donors are Germany, Denmark, Sweden, the Netherlands, and Japan. Multilateral Institutions The International Development Association of the World Bank Group is the most important creditor to Uganda. IDA dominates not only multi- lateral aid to Uganda but also the total aid flow. IDA accounted for ap- proximately 20 percent of total aid to Uganda during the period. IDA, the IMF, and the European Union are responsible for 65-75 percent of 114 Successful Reformers the multilateral financing. Other multilateral organizations with material programs in Uganda are the African Development Fund, UNDP, UNICEF, the United Nations High Commission for Refugees, and the World Food Programme (see tables 3.9 and 3.10 in Appendix 3.1). Figure 3.4 indicates that financing by multilateral institutions, like that of bilateral donors, took a giant step in the mid-1980s. ODF financ- ing almost doubled between 1986 and 1987, from just under US$162 million to US$300 million, but toward the end of the period it flattened out at a level of about US$400 million a year in disbursements. A decline in support in real terms can be noted in 1996, when disbursements by the European Union and IDA diminished sharply. The multilateral institutions essentially provide assistance in the form of loans. Lending has on average accounted for 72 percent of the multi- lateral ODF flows. However, the European Union has provided most grants among all donors (including bilateral donors) to Uganda, total- ing an average of US$39 million per year in 1980-96. In 1992 alone Uganda received an equivalent of US$156 million from the European Union, largely because of the collapse of world coffee prices. FIGURE 3.4 UGANDA: ODF AND EDA FROM MULTILATERAL INSTITUTIONS, 198o-96 Millions of 1995 U.S. dollars 600 Nongrant component of official loans Grant equivalent of of ficial loans 500 - Technical assistance grants _ Non-technical assistance grantsX 400 300 200 100 0 Source: World Bank. Uganda 115 Aid Coordination A considerable improvement in the government's coordination of aid has taken place in the 1990s. Today, all coordination is handled by the Aid Liaison Department in the Ministry of Finance, Planning and Eco- nomic Development (MFPED). Previously, there existed at least five co- ordinating units in different ministries, making coordination very scattered. Furthermore, in 1995 the Parliament passed an External Loans Act that stated that only the Ministry of Finance could raise foreign bor- rowings on behalf of the government. This has also helped reinforce dis- cipline in the government's aid coordination efforts. Improved coordination has led to aid being more accurately reflected in the budget than before. At the same time donors tend to channel more of their funding as budget support, for instance through the Poverty Action Fund. All government projects are subject to screening and approval by the Development Committee. The Aid Liaison Department then identifies potential donors for projects approved by the committee. However, do- nor-funded projects are still to a large extent initiated through direct contacts between line ministries and donors. Sector coordination is of- ten managed by line ministries even if the Ministry of Finance some- times calls for tripartite discussions with line ministries and donors to help shape sector programs and ascertain that they are consistent with the overriding reform objectives. The Finance Ministry is also involved in all final negotiations with donors. More recently the ministry has been insisting that donor support should be discussed in the context of sec- tor-wide dialogue on policy reforms and investment programming and the rolling Medium Term Expenditure Frameworks. It has also instituted quarterly meetings with donors, nongovernmental organizations, and the media to report on the Poverty Action Fund. Consultative Group (CG) meetings chaired by the Wold Bank are routinely held every 12 to 15 months. The meetings have focused on the reform program and financing needs as well as on topics such as poverty, social sectors, debt, and corruption. In December 1998 the CG meeting took place in Kampala for the first time. The meeting had broad participation of donors (16 bilateral and 18 multilateral organizations), cabinet ministers, other government officials, nongovernmental orga- nizations, stakeholders, and the media. The president took part in al- most the entire meeting. The participation of a wide range of stakeholders in the local CG meetings is in line with the World Bank's Comprehen- sive Development Framework. Any continuation of locally held broad- based meetings will primarily depend on the government's interest and leadership. The government and the donors meet quarterly in local coordination meetings chaired by the minister of finance or the Treasury secretary in 116 Successful Reformers the periods between CG meetings. The group sometimes meets to discuss thematic issues. Donors meet among themselves monthly under the chair- manship of the World Bank resident representative. The heads of the United Nations agencies also meet monthly to coordinate U.N. activities in Uganda. In addition, a number of ad-hoc sector working groups are in place, each led by a donor with particular interest in or knowledge of the sector concerned. Those groups also include government representatives, and their meetings provide a forum for the exchange of information and in some cases for formulation of joint strategies. Some donors said in interviews that the gap between the World Bank-IMF and the bilateral donors in terms of their assessment of the overall situation in Uganda had narrowed lately thanks to the intensified coordination. This also resulted in the preparation of joint donor statements at the CG meeting held in Kampala in 1998. POLICY REFORMS AND INSTITUTIONAL DEVELOPMENT Economic Policy 1980-98 From 1980 to 1985 Development finance literature links economic performance measured in terms of GDP growth to a healthy and vibrant financial sector (see, for example, McKinnon 1973; Shaw 1973; Fry 1995). However, the link be- tween the financial and real sectors in Uganda has been weak. This is largely attributed to the weakness in the financial sector itself, poor credit culture among borrowers, and the financial repression that existed dur- ing the 1970s and 1980s. Furthermore, after years of decay in the 1970s a number of business enterprises lacked a credit history and suffered from low rates of capacity utilization and lack of working capital despite sound prospects over the medium term. There was need for a residual direct support from government to bolster such enterprises as a way of kick- starting the productive sector of the economy. Between 1970 and 1980, the financial markets in Uganda were distin- guished by regulation and lack of robust financial policies. Price controls that were applied to the goods and financial markets without regard to market conditions resulted in, first, the misallocation of resources to sec- tors that appeared productive only because they were highly protected, and second, the emergence of informal markets (Atingi-Ego and Matthews 1994; Kasekende and Atingi-Ego 1996; Kasekende and Malik 1993). Consequently, the basic indicators of financial development added up to a rather discouraging picture. Real interest rates continued be Uganda 117 negative, and the returns on formal financial assets declined. The result was disintermediation in the formal financial system with the savings and investment ratios declining. Investment became biased toward short- term projects, largely because of the increased variability in the rate of returns caused by high inflation levels. To restart growth the government had to place high priority on re- establishing macroeconomic stability. In addition, both working and long- term finances had to be secured from external sources. This culminated in the first standby agreement with the IMF in 1981 that was aimed at restoring macroeconomic stability. The arrangement emphasized polices to encourage the mobilization of domestic resources. To this end, infla- tion had to be contained through an agreed reduction in overall budget deficit, domestic credit to government, stabilizing the value of the shil- ling, and a recovery in the productive capacity of the economy. Selective credit policies of the authorities resulted in the crowding out of some private sector agents and the government tried to bridge this gap by mobilizing some donor resources for on-lending though certain com- mercial banks. In addition, government sought to use the interest rate policy to increase domestic savings and hence loanable funds, so as to meet the credit needs of the private sector in the economy. There was, however, little progress in improving the tax structure and tax collection. This combined with weak expenditure controls and grow- ing military opposition brought about a deterioration in budgetary discipline in 1984-85, resulting in expansionary fiscal policies. The gov- ernment accumulated huge expenditure arrears equivalent to 21 per- cent of total expenditure commitments, and then stepped up the issue of long-term bonds to finance the arrears. The available monetary policy instruments could not effectively mop up the overhang in domestic li- quidity and this built up inflationary pressure on domestic prices and on the floating exchange rate.7 In response to the rapid depreciation, the government tightened restrictions on the exchange system. These devel- opments led to violation of the benchmarks of the monetary programs agreed by the government and the IMF, and as a result the program with the IMF collapsed in 1984. Afterward, economic performance deteriorated. Real GDP declined by 10 percent between 1984 and 1985, inflation bounced back from double digits in 1984 to triple digits by 1985, and an overvalued exchange rate reduced the export base to a single export-coffee. Import volumes also declined, reflecting the reduced capacity of the economy to finance im- ports even to obtain essential intermediate goods and spare parts. The principal lesson of this period was that the stabilization program was not accompanied by commitment on the part of the government to institute fiscal and monetary policy reforms to curb inflation,8 improve public service delivery, and allocate resources to productive enterprises. 118 Successful Reformers These were the main causes of economic instability in the first place and ought to have been addressed. From 1986 to 1998 The economy of Uganda suffered from serious policy reversals in 1985- 86. The foreign exchange constraint tightened, budgetary discipline worsened, and the institutional framework suffered from further dislo- cation. Inflation skyrocketed to the triple-digit range, reaching 296 per- cent in 1986, while deposit rates remained at about 35 percent. A number of problems observed in the early 1980s re-emerged with greater sever- ity. Unregulated agencies in foreign exchange (the kibanda market) strengthened in their management of foreign exchange (Kasekende and Ssemogerere 1991), and a number of residents preferred foreign ex- change-denominated assets to the shillings-denominated assets (Kasekende and Malik 1994). 1986: TURNING POINT. The government that took office in January 1986 faced daunting challenges in almost all facets of the economy. There was the problem of liquidity overhang and runaway inflation at a time of huge financial requirements to rehabilitate the economy. The banking system was insolvent despite the urgent need to encourage productive investment. Foreign exchange reserves could hardly cover two weeks of import requirements. Many firms lacked working capital and were not creditworthy. The system for procurement and processing of cash crops was highly inefficient, and the mismanagement of credit resulted in the nonpayment of farmers. The new government invited a team of international economists and a number of influential local academics, entrepreneurs, and others to carry out an in-depth study of the economy and advise on the best way to restart economic growth. The team was asked to provide a compre- hensive macroeconomic framework for the country's rehabilitation and development and to identify a mechanism and projects for sound man- agement and development of the economy over the long term. There was lack of consensus among team members on certain key areas of macroeconomic management, in particular the choice between a closed and open economy. A minority report of the study recommended revaluing the shilling from shs 50 to shs 14 per U.S. dollar. The argu- ment, to use the exchange rate as a nominal anchor to inflation, was in line with the thinking of the new government. Indeed, the government chose the closed model and opted for a revaluation as recommended in the minority report. Implementing these recommendations turned out to be a serious mistake, as they fueled macroeconomic instability and worsened external viability. In addition, the government also began a Uganda 119 program built on price controls and ideas about barter trade with Libya, North Korea, and similar countries. The "Washington consensus" was regarded as an imperialistic policy package.9 There was, however, dis- agreement within the government about which type of reform program to follow. The government chose a "control" model despite opposition from critics who argued strongly that such a strategy would fail. The opponents of the reform program were nonetheless kept in the govern- ment; this later proved to be a very wise decision that benefited the economy (Tumusiime-Mutebile 1995). Macroeconomic imbalances meanwhile intensified and all indicators pointed to a further deterioration in the economy. It became evident that the "control" model was doomed to fail. This failure, however, became a landmark in the policymaking process in that officials were reshuffled and the direction of policy choices changed. It also softened the official attitude toward the once-unpopular market policies and paved the way for liberalization. In other words, a congruence of events and the failure of the 1986 recommendations to deliver economic stability led to a reas- sessment of policies. 1987-91 Economic Reforms As mentioned above, the initial economic policies of 1986 were incon- sistent and by 1987 it was evident that a new program was needed. The exchange rate and inflation levels were out of control and the whole situation was steadily deteriorating. Faced with these problems, the gov- ernment resumed dialogue with IMF and the World Bank, seeking their assistance in designing and implementing an economic recovery pro- gram. The main objectives of the reform program were to promote long- term growth through substantial investment in order to rehabilitate physical assets and the infrastructure. Policy reforms were needed to promote structural changes, whose implementation would depend on mobilization of necessary external resources. The president provided the highest level of backing for the new eco- nomic reforms, although there was still some reluctance within the gov- ernment. Government officials were sent to training programs and to Ghana on study tours to prepare for the new reform program. After pro- tracted negotiations with the IMF, the government decided to start an Economic Recovery Program (ERP) in May 1987, supported by the IMF and the World Bank. The ERP had the following broad objectives: * Economic growth of at least 5 percent per year in order to permit growth of real per capita GDP of at least 3 percent per year. This was to be achieved through improved producer incentives, agricultural produce marketing, and increased industrial capacity utilization. * Restoration of price stability. 120 Successful Reformers * Halting and reversing the deterioration in the balance of payments through restructuring of debt, export promotion, and increased capacity utilization in the import substitution industries. * Strengthening the institutional framework. Key to the reform package was a national currency reform. The gov- ernment introduced a new currency at a rate of exchange of one new shilling to 100 old shillings. It also imposed a 30 percent conversion tax on all shilling holdings (currency in circulation and deposits). The ex- change rate was simultaneously devalued by 76.6 percent and the gov- ernment said that in the future the exchange rate would be adjusted whenever necessary. Not all policymakers were fully convinced that the stabilization and structural adjustment programs were the right policies. Advocates of the reform process engaged in education and persuasion during 1987- 89. To influence public opinion, they sponsored seminars and prepared public discussion papers (Tumusiime-Mutebile 1995). On the policy front, the Presidential Economic Council (PEC) played a crucial role during the initial period of the country's reform efforts. The council, chaired by the president, is a decisionmaking rather than advisory body It greatly assisted reform advocates by holding open debate on the policies, and it also provided an important forum for discussion of major economic policy proposals requiring immediate attention. Proponents and oppo- nents of reforms were invited to articulate and defend their points of view in the PEC. Interviews with government officials indicated that technocrats also used the PEC as a forum to convince the president of the need for reform. To this end the council secretary had a certain de- gree of independence in shaping the agenda and inviting people to meet- ings, which helped in pushing through reform efforts (Lamont 1995). The leadership of the government remained committed to the reforms despite some lingering reluctance among officials and exogenous shocks to the economy (collapse of the International Coffee Agreement and ris- ing oil prices due to the Gulf War). As part of the reform process, the government convened a consultative conference in December 1989 to critically discuss the state of Uganda's economy. All stakeholders in the economy were invited and the conference turned out to be another land- mark in the reform process. It laid the foundation for a radical liberaliza- tion of the foreign exchange market system by legalizing the parallel (kibanda) market, which internal observers viewed as a key reform. The government made this decision entirely on its own, with no conditional- ity from the IMF or World Bank (Tumusiime-Mutebile 1995). The most important outcome of the meeting may have been the widespread sup- port, including that of the president, for the economic reforms that emerged (IMF 1998). Uganda 121 The rehabilitation of the economy relied to a large extent on donor support to finance the required increase in import volume. This in turn contributed to a worsened debt situation, as mentioned. The govern- ment's Economic Recovery Program was supported by about 25 policy- based credits from the World Bank, amounting to more than US$1 billion in the five-year period 1987-92. The lending started with two Economic Recovery Credits focusing first on stabilization and later on demand management, including interest rate reform and measures to contain the fiscal deficit, trade liberalization and revitalization of the private sector, and public sector management. A Structural Adjustment Credit (SAC I) in 1991 supported coffee sector reforms, improvements in the trade regime, increased revenue mobilization, civil service re- form, and actions to speed up operations of the Departed Asians Prop- erty Custodian Board. The principal results of the economic reform program implemented between 1987 and 1991 included: (a) a lowering of the inflation rate from an average of 190 percent per year to about 28 percent per year by 1991; (b) a revival of the GDP growth rate to an average of 6.3 percent per year over 1988-91; and (c) a slow recovery of exports, and of private and public sector investment. The program for 1990-91 specifically provided for the strengthening and development of the banking industry and the generation of up-to- date and accurate information on the balance sheets of financial institu- tions. Following the legalization of the parallel market foreign exchange bureaus were licensed in July 1990 to trade in foreign exchange. Interest rates were further reviewed to provide adequate incentives for domes- tic resource mobilization and term lending. AFTER 1992: FULL OWNERSHIP. In 1991 and 1992, the government implemented wide-ranging policies intended to eliminate the structural and financial bottlenecks that had constrained progress in economic sta- bilization. These included the excessive growth in domestic credit dur- ing the 1991/92 financial year, which resulted from weaknesses in monitoring the program and failure to take appropriate corrective mea- sures as the problems emerged. Prominent among the problems was the monetization of the budget deficit, which averaged 10.2 percent of GDP in 1991/92 compared to 3.5 percent the previous year. Consequently, government expenditure measured as a percentage of GDP rose from 16 percent in 1990/91 to 23 percent in 1991 /92; this caused inflation to surge to 58 percent on annualized rate by March 1992. This in turn led to vio- lation of the benchmarks in the monetary program and the delay of the IMF/World Bankprogram. Donors delayed disbursements in 1991 pend- ing implementation of some reforms (foreign exchange auction), and this caused the fiscal crisis. 122 Successful Reformers The experience made clear the increasing dependence of the budget on foreign savings and the need for a tool to make the budget flexible. It also showed the need to widen the range of instruments in economic manage- ment, particularly monetary policy instruments, and to move faster into market-oriented policies. These lessons provided an opportunity to reassess the policy envi- ronment, and led the government to reform the management of its fiscal and monetary policies. The following reforms were implemented to strengthen the conduct of fiscal and monetary policy: * The president announced that borrowing from the Bank of Uganda constituted indiscipline, and broad reform policies were introduced in the financial sector. * A cash budget rule was adopted beginning April 1992 and a cash- flow management committee was instituted, including officials from the Bank of Uganda and the Ministry of Finance and Eco- nomic Planning. The flexible management of the budget greatly helped to reduce financial imbalances and provide adequate re- sources to the banking sector to support private sector credit. * The Bank of Uganda was mandated to auction Treasury bills as necessary to achieve the monetary targets. A decision was also made to review interest rates regularly to ensure real interest rates of at least 4 percent on 12-month deposits. * More market-oriented methods were introduced in the manage- ment and allocation of foreign exchange and determination of the exchange rate. * The government linked the one-year deposit rate, the lending rates on credit to the agricultural sector, and the interest rates on term credits to the auction-determined discount rate in an effort to strengthen resource mobilization and efficient allocation of re- sources. This took effect in October 1992. All other interest rates were completely decontrolled. Interest rates were fully liberalized in 1993/94. * Beginning in 1992/93, the Reserve Money Program was developed as a key instrument in liquidity management and implementing monetary policy. Between 1993 and 1998, then, management of the economy was largely dictated by the lessons drawn from the slippage that occurred in 1991. Another important factor was institutional reform, namely the president's decision to strengthen the economic team by combining the ministries of Finance and Economic Planning into one. In fact, the team from the Ministry of Planning that had contributed immensely to the reform ef- forts took over the leadership of the combined ministry. Thanks to this Uganda 123 merger, the government strengthened its commitment to macroeconomic stability, and ownership of the reform program deepened. The continu- ity and long tenure of the economic team has turned out to be critical in the process leading to full ownership of the reform program and Uganda's entry into a period of sustained reforms. Following the restoration of macroeconomic stability, a new three- year ESAF program was agreed with the IMF in 1994. The program fo- cused on structural reforms, fiscal objectives, and reserves accumulation. The government introduced tax rate changes, established an indepen- dent revenue authority, abolished the export tax on coffee, and removed the monopoly of the Coffee Marketing Board. Under this program, the IMF and other donors sought to finance the deepening of reforms, espe- cially the structural and institutional reforms (discussed below). The World Bank followed up with a second Structural Adjustment Credit, also in 1994. The results of implementing the above policies between 1992 and 1998 have been quite satisfactory: * Private sector investments in Uganda increased from 9 percent of GDP in 1991/92 to 13 percent in 1998/99. * Growth rate of real GDP increased to an average of 7.6 percent per year. * Inflation declined to 8 percent, down from 190 percent over the period 1987-91. * The current account deficit was narrowed from 6 percent in 1992 to 4 percent in 1996, measured as current account deficit/GDP (in- cluding grants). * Monetary deepening (M2/GDP) was experienced, from 8.2 per- cent in 1992 to 11.7 percent in 1996. * Foreign exchange cover, measured in terms of months of imports cover, increased from two to about five months. All these measures resulted in lower inflation levels and satisfactory growth rates, as shown in figure 3.5. The IMF and Uganda agreed on a new ESAF in 1997. It focused on maintaining macroeconomic stability, private sector promotion, export- oriented growth, and structural and institutional reforms to further re- duce impediments to growth and job creation. Uganda accepted the obligations of Article VIII and removed current account restrictions, and went to full capital account convertibility in 1997. In the same year the World Bank approved a third Structural Adjustment Credit. The credit aimed at stabilization and the acceleration of economic growth through fostering long-term fiscal sustainability, involving a more efficient tax system as well as better management of public expenditures. 124 Successful Reformers FIGURE 3.5 INFLATION AND GROWTH IN UGANDA, 1980-96 Percent Percent 10 180 9 160 8 < Pr l 2140 120 6 100 5 80 4 3 60 2 GDP growth (left axis) Inflation (right axis) 40 1 20 0 0 1980-84 1985-89 1990-94 1995-97 Source: World Bank. Growth and Poverty Trends GROWTH. As stated above, Uganda has recorded reasonable economic growth rates between 1986 and 1998, averaging 6.5 percent per year. Before this period, Uganda had suffered from massive economic mis- management that led to total collapse of the productive capacity of the economy. A very important requirement for the economic revival of the country was the adoption of proper macroeconomic management poli- cies. Indeed, economic growth in Uganda in recent years owes much to the adoption of prudent macroeconomic policies by the government as discussed above. Growth in Uganda has been driven by recovery in ag- ricultural production and an increase in capacity utilization in the in- dustrial sector. However, sustained growth can only be expected through investment by both the private and public sectors. The recovery of the agricultural sector since 1987 followed govern- ment recognition of the sector's potential to lead the country's recovery and the economic reforms undertaken to realize this potential. Monetary and nonmonetary agriculture together accounted for about 54.8 percent of GDP in 1987. Efforts were made to remove the main constraints to agricultural production and diversification (Kasekende and Atingi-Ego Uganda 125 1999). These included the lack of foreign exchange to import agricul- tural inputs, as well as monopoly and other institutional inadequacies in the marketing of the output. The removal of some of these constraints, together with the peace prevailing in most parts of the country, led to increased land use and high agricultural production. Output in agricul- ture grew by 62 percent between 1986 and 1998. The turnaround has mainly benefited rural people, who produce mostly on small land holdings. In the industrial sector, the initial effort was geared toward increas- ing capacity utilization of the existing factories. Low capacity utiliza- tion had resulted from lack of foreign exchange for the importation of spare parts and raw materials. Measures taken to revive this sector included the introduction of the Open General License and Special Import Program in 1988, both of which were intended to allocate scarce foreign exchange to priority areas. Success has been achieved in this area, with the manufacturing sector recording average growth of 12.74 percent per year between 1989/90 and 1998/99 and increasing its share of GDP by more than 4 percentage points. Table 3.4 highlights the mon- etary GDP performance in selected sectors of the Ugandan economy in1989/90 and 1998/99. A significant change associated with the reform period has been the shift of production from the public to the private sector. This is expected to lead to sustained growth and poverty reduction. The share of private capital formation increased from 7 percent of GDP in 1989/90 to 13 per- cent in 1998/99. At the same time public capital formation declined from 7 percent of GDP to 5 percent during the same period. POvERTY. Poverty reduction has been a major concern of government during the reform period. Both human and financial resources have been TABLE 3.4 UGANDA: GROWTH PATTERNS SINCE 1989 Share of GDP (percent) Average growth 1989/90 - 1998/99 Sector (percent) 1989/90 1998/99 Agriculture 5.98 24.3 23.1 Manufacturing 12.74 5.4 9.6 Construction 10.46 5.1 7.3 Trade 8.41 11.0 13.1 Hotels and restaurants 12.10 1.1 1.8 Transport and communication 9.03 4.1 5.2 Total GDP 6.40 51.0 60.1 Source: Uganda Bureau of Statistics. 126 Successful Reformers deployed to tackle poverty at all levels, in both the urban and rural areas. Government recognizes that poverty reduction can be addressed in two ways in addition to maintaining macroeconomic stability. The first is by putting in place an environment that promotes investment and fast eco- nomic growth, in order to provide employment opportunities to the popu- lation and increase their incomes. Second, the government believes that there are additional intervention measures that can also work directly to alleviate poverty. The two approaches are seen as complementing each other in the fight against poverty. Given the high level of macroeconomic instability during the initial period of the reform effort, the government concentrated on programs that would promote stability and efficient use of national resources. The structural adjustment programs supported by the IMF and World Bank were used as the framework for economic management. During the lat- ter part of 1980s, there were no direct interventions aimed at reducing poverty. A special Program to Alleviate Poverty and the Social Costs of Adjustment was designed at the turn of the decade when it was recog- nized that some vulnerable groups would not benefit in the short run from the Economic Recovery Program. The World Bank supported the program with a credit of US$28 million. It was basically intended to mitigate the social costs that arose from demand management measures of government (particularly fiscal tightening), associated with the ad- justment programs. With economic stability restored for the past few years, the govern- ment is now focusing its efforts on poverty reduction through direct inter- vention. This is contained in a Poverty Eradication Action Plan that was developed by government in 1997. Key features include programs aimed at improving the road network, agricultural modernization, universal primary education, primary health care, and promotion of the private sector. Government plans to increase spending on those sectors that are believed to have a direct impact on the quality of life of poor Ugandans. To further boost the Poverty Action Fund, government has directed all savings on debt repayment arising from the HIPC Initiative toward the priority sectors-an indication of its determination to fight poverty. The savings from HIPC are in the region of US$40 million per year, and do- nors are expected to contribute an additional US$40 million to the fund. The government strongly believes that provision of security is key to ensuring that the poorest members of society enjoy the benefits of growth. In addition, the Poverty Eradication Action Plan emphasizes improved governance through enhanced transparency and accountability for re- sources directed to the Poverty Action Fund. As an example, resources released for the primary education program are advertised in the daily newspapers, and the regular donor meetings on the Poverty Action Fund receive financial reports on the fund's receipts and disbursements. Uganda 127 Although there are several gaps in poverty data, there is some evi- dence of a reduction in the level of poverty in Uganda following years of steady economic growth. The household budget survey of 1997 indi- cates improvements in the living conditions of the poor. The report indi- cates that household expenditure increased by 128 percent in the rural areas, while in the urban areas average household expenditure increased even faster, by 133 percent between 1992 and 1996. In addition, the re- port shows that the share of iron-roofed houses in total residential houses increased from 38 percent in 1991 to 48 percent by 1996. Reflecting in- creases in income, the share of food and beverages in total household expenditure has fallen by 5 percentage points, from 60 to 55 percent during the same period. Institutional Development and the Governance System The restoration of economic growth was largely supported by the adop- tion of prudent macroeconomic policies that initially increased the utili- zation of existing capacity; after 1992-93, new investments also played a role. However, according to Kasekende and Atingi-Ego (1999), this kind of growth will taper off as spare capacity diminishes, and therefore new capacity will have to be generated if current growth rates are to be sus- tained. This calls for increased investment-GDP ratios and incremental capital-output ratios (ICORs). It is very likely that with poor service de- livery in either the public or the private sector or both, the associated ICORs will be lower than would be the case with efficient service deliv- ery. Consequently, increased investment-GDP ratios need to be accom- panied by efficient public service delivery in order to tap the maximum ICORs. In line with this, some institutional reforms have been under- taken in an attempt to increase the effectiveness of service delivery through building relevant institutions. Institutional Development PRIVATE SECTOR DEVELOPMENT. In line with the recommendation that the role of government in production and the allocation of productive resources should be reduced, some institutional reforms were under- taken to increase the effectiveness of the private sector. These included, among others, civil service reforms, public sector reform, and a frame- work for market determination of prices of productive resources (for example, the exchange rate). Investment laws and procedures were streamlined in order to attract both domestic and foreign investors. To this end, a new Investment Code Act (1991) replaced the Foreign Invest- ment Act (1977), and the Industrial Licensing Act (1969) was abolished. The Investment Code Act sought to attract, promote, and facilitate 128 Successful Reformers investment, provide fiscal incentives, protect foreign and local investors, and introduce a mechanism for the repatriation of dividends and fees. In 1991 the Uganda Investment Authority was created to administer the new investment code and also to act as a one-stop investment center for investors. Contentious issues such as the Departed Asians Property Custodian Board had to be resolved as a step toward reaffirming government re- spect for private property rights. By 1994, a total of 640 out of 690 claims by noncitizens and 1,860 out of 2,000 claims by citizens had been veri- fied and property returned to the original owners. This was a good per- formance given that the non-returned properties were either not claimed or the verification needed more time. By 1996 this process was cleared. It is worth noting that this is an area in which technical assistance played a large role, and it also demonstrated the role of support for a reform at the highest political level. In addition to drawing on technical assistance, reform-minded tech- nocrats have used conditionality in order to overcome pockets of resis- tance to reforms. In some cases they have asked for specific conditions to be embedded in the support programs. It was therefore not uncom- mon to see structural benchmarks in the performance criteria of the ESAF programs being used as a means of speeding up reforms. This may be seen as a step in the right direction insofar as deepening the reforms is concerned. There have, however, been reports of less transparent reform undertakings in the privatization of some state-owned enterprises. Those few cases notwithstanding, most Ugandans appear to have embraced the idea of privatization. With respect to the exchange rate and trade policy reform, the gov- ernment in 1991 was more than committed to conform to Article VIII of the IMF agreement, which stipulated that no restrictions be placed on international transactions. By July 1990, the authorities had already le- galized the buying and selling of foreign currencies in the foreign ex- change bureaus at a market-determined rate. Further reforms in the financing of international trade were implemented in January 1992 fol- lowing the government's replacement of the Open General License and Special Import Program with the weekly foreign exchange auction of donor-provided import support funds. What is important here is that the donor aid flows provided the needed foreign exchange for the weekly private sector foreign exchange auction, thus boosting private sector confidence. Finally, in 1993, the auction and the surrender requirements were abolished and replaced with a unified foreign exchange system with the commercial banks and foreign exchange bureaus as the key institutions in the market. In addition, in July 1997 the capital account of the balance of payments was opened up in an attempt to increase pri- vate capital inflows. Uganda 129 Other measures taken to revamp export earnings included the liberal- ization of export commodity marketing. A number of private companies and cooperative unions were licensed to compete with the once monopo- listic marketing boards. The main result has been the on-spot delivery of cash for value offered by the producers, as opposed to the earlier practice of issuing promissory chits. On the macroeconomic front, the central bank financing of the marketing boards was often associated with the injection of high-powered money into the system and this was a principal cause of the inflationary episodes in Uganda. Since the liberalization of marketing arrangements, the competition in marketing has completely ended this type of financing. As part of the government's commitment to economic stabilization, a coffee stabilization tax was imposed when world market prices for coffee surged during the 1994-95 coffee boom. This was to en- sure that the exchange rate did not appreciate markedly and adversely affect competitiveness of the export sector. Government also took steps to create institutions that could help to promote the growth of the private sector. These actions included: * Provision of an appropriate legal framework that strengthens com- mercial jurisdiction, reinforces property rights, and facilitates settle- ment of business disputes * Support for the development of well-organized groups represent- ing various interests in the private sector * Establishment of a properly functioning financial system and safety net in order to meet liquidity needs of the private sector and mini- mize liquidity crises, for example through deposit insurance schemes. The financial sector reforms were also necessary because the high intermediation costs were retarding private sector development. PoLicy DIALOGUE. Since 1987, a few institutions and organizations have become active in the policy dialogue. These include the Parliament, the independent Economic Policy Research Centre at Makerere Univer- sity, the Uganda Manufacturers Association, the Private Sector Founda- tion, unions such as the National Organization of Trade Unions, nongovernmental organizations (mainly through Uganda Debt Net- work), and, to a lesser extent, the Uganda Farmers Association. The Parliament acts as a useful instrument and watchdog. Major eco- nomic policy changes are now being reviewed and adopted by the Par- liament. It has been especially active in the field of privatization and resisted, for instance, the privatization of Uganda Commercial Bank. It has also exposed cases of corruption, leading to impeachment, removal, or reshuffling of cabinet ministers and resulting in greater transparency and accountability in public administration. 130 Successful Reformers Research institutions can potentially play a strong role in the identifica- tion of problem areas and in deepening the domestic policy dialogue. Institutions such as the Economic Policy Research Centre (EPRC) could become facilitators for the intemal dialogue on policy issues as well as local watchdogs, roles often undertaken by the World Bank today in driv- ing the dialogue on policy reforms. Rather than having a voice of its own in the policy debate, EPRC would plan a catalytic role, providing analytical input for the process and bringing together the different stake- holders to debate the various issues. The private sector enjoys the respect of the government, which sup- ports private sector participation in policy formulation. A transition has taken place since 1987, when the private sector was regarded mainly as a subject for government regulation. Current thinking on private sector development is discussed with the private sector before new policies are developed and proposed by the government. The Uganda Manufactur- ers Association (UMA) has evolved into a key institution representing the interests of the private sector. Over the years, it has developed a strong partnership with government in policy formulation. In 1991 it spearheaded the setting up of the Presidential Forum, which includes all key stakeholders in policymaking and the private sector. However, the role of the forum has recently waned. The private sector through UMA participates in the budget process, which has been well received by the government and the president. UMA has participated on a num- ber of occasions in determining the speed and sequencing of the trade reform program. Academic researchers and local consultants have also taken part in the policymaking process through UMA. Private sector representatives emphasized in interviews that the gov- ernment should not make any rules for the sector without consulting them. Some also complained that even though the sector rightly is en- gaged in policy discussion, its real influence is rather marginal. The private sector and the government agree on many points, but mem- bers of the UMA said that the government is unwilling to listen seri- ously to the private sector when the two sides disagree (interview, UMA, 1999). Officials at the Ministry of Finance and Economic Plan- ning responded that someone has to make the final decision, weighing all factors, and that not all actors could be equally satisfied with the outcome. The policies implemented reflect inclusion of concerned par- ties in the process, if not always a consensus. Private sector representa- tives also said in interviews that there needs to be an element of competition in the political system so that private sector participants can voice their concerns to different parties if their ideas are not re- flected in actual policymaking. "Once you lack competition you risk being patronized." Other private sector actors are the Private Sector Foundation and the Uganda Exporters' Association. The Uganda National Chamber of Com- Uganda 131 merce and Industry is not as active in lobbying, although it has some political influence through members of Parliament (IMF 1998). Thefarmers are barely active in the policy dialogue. The Uganda Farm- ers Association only participates to a very limited extent in the discus- sion. However, the key institution charged with spearheading agricultural sector reforms is the Agricultural Secretariat in the Bank of Uganda. It was set up with financing from IDA, which provided resident advisers while consultants came from the U.N. Food and Agriculture Organiza- tion and the World Bank, and other local researchers were recruited as needed. The secretariat is credited with restructuring of both coffee and cotton subsectors as well as the land reform. Nongovernmental organizations (NGOs) have also become more in- volved in policy discussions. A large number of local and foreign NGOs are active in the country, an estimated 1,500 to 2,000 organizations, mak- ing coordination very difficult. Nevertheless, NGOs have made their voices heard on the debt issue in particular through the Uganda Debt Network (UDN). The government has called UDN for consultation on policy issues, but the network's representatives found the meetings of- ten to be information dissemination by the government rather than real consultations. As with the private sector, many NGOs welcome the open- ing of dialogue but find that their influence on policy formulation con- tinues to be minimal. There apparently exists a political will to involve NGOs and other stakeholders in the policy dialogue, but more investments seem to be needed on both sides to form this partnership. A group consisting of NGOs and the government has been formed to monitor the Poverty Action Fund. NGOs attended the CG meeting in December 1998 but were not engaged in its preparation; this made their participation in the meeting less effective, according to NGO sources. Some donors expressed doubts in interviews as to whether NGOs and the private sector can contribute effectively to the policy dialogue. Ac- cording to this view the NGOs are in many cases too consensus-oriented and the private sector is still too dependent on the government to be ac- tive critical voices in the debate. NGO representatives believe on the other hand that a major constraint to active participation is lack of consultation by donors. They say that instead of holding discussions with and relying only on the government, donors should meet with NGOs, the private sector, research institutions, and others as they formulate their support strategies. Aid organizations should focus on the civil society, private sector, and autonomous institutions in addition to the government. Public Sector Management The government realized in 1992 that the successful achievement of macroeconomic stability would rest largely on the control of inflation. 132 Successful Reformers To this end, monetization of the budget deficit would have to be reduced, and fiscal discipline based on tax administration and control of govern- ment expenditures would have to be implemented. A thoroughgoing reform of the civil service would also be necessary. In light of these ob- jectives, the government undertook major efforts on tax reform and this led to creation of the Uganda Revenue Authority (URA). The decision to establish the URA outside the existing civil service structure was intended to allow the entity much better terms of employment and tools to effec- tively carry out its duties of increasing tax collection. As of 1999, the tax- to-GDP ratio is nearly 12 percent, up from 6 percent at the time the URA was created. Financial assistance toward the creation of the URA came from the United Kingdom, UNDP, and the World Bank through IDA. A civil service reform was badly needed in order to establish a well- paid, motivated, and efficient civil service while at the same time strength- ening the government's ability to deliver services to the population. As a first step toward achieving this objective, in 1992 the number of minis- tries was reduced from 38 to 21. Under the Structural Adjustment Credit, the downsizing of the civil service was made possible through, first, the displacement of irregularly employed staff and the elimination of ghost workers, and second, retrenchment, natural attrition, and continued con- trol over recruitment. As a result of these measures, by the end of 1993 the number of civil service employees on the payroll was reduced to 170,000, down from the 270,000 recorded in July 1991. In addition, ap- proximately 50,000 soldiers were demobilized. To solidify this reform, donor-funded technical assistance to govern- ment helped to establish a computerized payroll and begin a system of personnel management. Once again, the commitment of government to civil service reform was very important because there was intense social pressure against this government policy. Equally important was the fi- nancial assistance of US$13 million, administered by IDA, that the do- nors provided to pay retrenchment packages. As a result of these policy measures, civil service pay more than tripled. In addition, a number of donors helped to provide salary top-ups, though this policy has recently been discontinued. Equity of Public Expenditures Public expenditure reforms were designed to address the inadequate spending on the social sector and economic infrastructure. Beginning in fiscal 1991, the budget increased allocation toward so-called Priority Program Areas. Since then, these areas have been protected from expen- diture cuts. Overall, improvements and transparency in public expendi- ture programs have been achieved largely through (a) keeping the expenditures within sustainable limits; (b) strengthening annual recur- Uganda 133 rent expenditure programming; and (c) monitoring actual expenditure through releases. The development budgets were rationalized through the screening of new projects, improving quality at entry points, and identifying the nec- essary funding. Beginning in 1995, projects under public investment pro- grams also received their full allocation of counterpart funds. Public spending on basic services such as health and education gen- erally increased between 1990 and 1995, largely because the budgetary allocations to these services were protected. In their study, Ablo and Reinikka (1998) attempted to determine whether the increase in budget- ary allocations translated into enhanced delivery of health and educa- tion services. They found that school enrollment had increased by 60 percent during 1990-95, contradicting earlier reports by the Education Ministry that school enrollment had not gone up despite the budgetary increase. This in turn cast doubt on the reliability of the officially re- ported data. This finding may be supported by the view that improved social, economic, and political conditions allowed more parents to send their children to school at the same time as the central government in- creased the contributions to the primary schools. Another finding of the study was that increased budget allocations might not matter in a situation where existing institutions are weak. Governance and accountability are key in ensuring that budgetary allo- cations result in expenditure outlays for the projects they are intended to benefit. But the study confirmed hypotheses of poor governance and accountability. Competing needs within the sector, or at various levels of government itself, lead to the misuse of public funds. The survey showed that not all the budgetary allocation to increased teacher sala- ries filtered through to the intended beneficiaries. The institutional con- text and incentives also played a role across the sectors. Schools were found to keep relatively good records on enrollment and financial flows, but this was not the case with the health centers. This can be attributed to the role of the Parent Teacher Associations (PTAs) that were the main financiers of the primary schools and, generally speaking, insisted on accountability from the schools. It is, however, doubtful whether under the current system of Universal Primary Education, which abolished the financing from the PTAs, the schools will continue to keep relatively good records. Table 3.5 compares budgetary allocations to different sectors of the economy in the 1960s and the 1990s. The data reveal the increasing em- phasis on education and security at the expense of agriculture and infra- structure. The health sector's share has hardly changed, and given the reduced share of agriculture there are worries that the fight against pov- erty may be weakened. The government, however, recognizes the prob- lem of poverty in the country and the urgent need to eradicate it as part TABLE 3.5 UGANDA: SECTOR DISTRIBUTION OF BUDGETARY ALLOCATION, I965-70 AND 1994-99 (percent of total) Sector 1965/66 1966/67 1967/68 1968/69 1969/70 1994/95 1995/96 1996/97 1997/98 1998/99 Agriculture 8.20 10.00 10.50 8.50 9.60 2.60 1.50 1.40 1.30 1.50 Education 12.30 12.20 6.30 8.11 9.40 19.80 18.80 21.80 23.40 22.70 Health 6.90 7.80 8.00 4.30 9.60 6.50 6.40 6.50 5.50 6.40 Road/works 5.10 4.60 6.30 8.30 9.50 4.30 4.30 6.80 4.90 6.90 Law/order 11.00 10.70 10.60 8.90 9.70 8.90 9.70 8.60 7.50 6.70 Defense 10.00 9.20 11.60 9.50 10.70 17.20 16.50 16.50 14.00 13.50 Other 46.50 45.50 46.70 52.39 41.50 40.70 42.80 38.40 43.40 42.30 Total 100 100 100 100 100 100 100 100 100 100 Source: Government of Uganda, "Background to the Budget,"1970/71, 1995/96, and 1998/99. Uganda 135 of sustainable development. This is seen in the increased budgetary allo- cation for education, mainly through the Universal Primary Education scheme. This represents the greatest social service investment toward poverty eradication to date. To ease constraints on the marketing of ru- ral produce, the government eliminated monopoly marketing arrange- ments, decontrolled producer prices, and increased budgetary allocations toward road works. It has drawn up a 10-year road sector development program that includes road rehabilitation and expansion. Government's focus in the health sector is toward preventive health care. The national immunization program has been vigorously pursued and this has led to a reduction in child mortality. Planning is underway to set up a commu- nity health department in every hospital, and government is currently developing a Health Policy White Paper and a National Health Plan Framework. Governance and the Political System Dictatorship, corruption, centralization, and personalization of power have characterized Uganda's political history. In the 1960s, the regime then in power abolished the federal system enshrined in the 1962 consti- tution and put in place a contested constitution that created an over- centralized system of governance. The decade of the 1970s was marked by military dictatorship, while the first half of the 1980s witnessed civil wars. Beginning in 1986, political stability was restored throughout the country, although beginning in the late 1980s insurgency in the northern part of the country, and more recently in the western part, eroded the stability. The insurgency notwithstanding, progress has been made in adhering to tenets of good governance. A homegrown constitution was recently promulgated following some amendments in the draft consti- tution by the constituent assembly. The draft constitution that the as- sembly debated was a summary of ideas that a constitutional commission had collected from Ugandans from all walks of life. This constitution is now the pillar of political system and governance in the country. The government has attempted since 1986 to politically empower its people. The direct election of local leaders in 1997 and 1998, for instance, provided an opening for Ugandans to demand accountabil- ity from their elected leaders. However, the process is still far from complete, and is constrained by persistent high levels of illiteracy and poverty as well as lack of civic competition. The presidential and par- liamentary elections held in 1996 and the local government elections of 1997 and 1998 reflected these problems. The system of secret ballot and adult suffrage still needs to be extended to elections at all levels. According to the Uganda Human Development Report (UNDP 1998), the presidential elections were marred by widespread vote-buying that 136 Successful Reformers was fueled by the high levels of poverty among the majority of the rural electorate. As a result, the electorate did not fully exercise their voting rights under the law. AID AND REFORM The Impact of Aid on Reforms We have seen that poor economic policies and social disturbance led to a huge decline in the economy in the latter part of the 1970s. By the early 1980s a reform mentality was beginning to take hold, and the country ran a relatively strong economic program based on fiscal consolidation in 1981-83. The IMF and the World Bank released standby and recon- struction credits during this period and a Consultative Group meeting was convened in 1982. Although delayed, the credits gave rise to fruit- ful dialogue on economic policy and management issues (World Bank 1998a). However, during 1983 the fiscal program started to loosen and the program went off track in 1984-85. Since the present government came to power in 1986, Uganda has gone through three distinct phases of reform: 1986, when the govern- ment rejected market-based reforms; 1987-92, when it implemented re- forms reluctantly; and from 1992 on, characterized by full government ownership of the reform program. Aid in various forms has supported reforms undertaken during these three phases. Technical assistance, train- ing, and policy dialogue were the key ingredients in shaping the reform program. An IMF representative interviewed in Kampala remarked, "In the early days of the NRM government, IMF and the World Bank more or less ran a macroeconomic classroom with the president" (interview, IMF, Kampala, 1999). During the period of reluctant reforms (1987-92), nonfinancial assis- tance was important but the conditionality linked to financial support probably had an even greater impact on the reform efforts. The condi- tions helped reform-minded elements in the government push the reform agenda forward with the support of the president. The conditionality was therefore a helpful tool for generating, implementing, and cement- ing reforms. When the government was in this transition process many reforms were driven by the requirements of conditionality (Tumusiime- Mutebile 1995). Many of the reforms were of a defensive nature, repair- ing aspects of the economy that had been destroyed over the preceding 15 years. Donor support was also critical in helping to build capacity. A World Bank representative interviewed in Kampala said, "The World Bank stayed on in 1986-88 even during periods when it did not fully subscribe to the policies undertaken by the government. This strength- Uganda 137 ened the credibility and trustworthiness of the relationship" (interview, World Bank, Kampala, 1999). Since 1992, with full ownership of the reform program, conditionality no longer plays a decisive role in inducing reform undertakings. The government has taken over the design of the reform agenda, and the success of the reform program during this period can largely be attrib- uted to this strengthened government ownership. The financial support is still needed, along with debt rescheduling through HIPC, but from the donors' perspective policy dialogue can once again be the single most important instrument in supporting generation of reforms. Uganda has had good relations with its donors due to the agreement on the need for reforms. There has been "broad acceptance of the mea- sures required to bring them about"; moreover, "conditionality was never 'imposed,' it was an economic necessity" (Tumusiime-Mutebile 1995). After some six or seven years of reform efforts the government was able to design reforms proactively, sometimes running ahead of the donors in this respect. Examples include the exchange rate reforms, capital ac- count liberalization, and decentralization. All in all, during the period 1989-98 the government implemented far-reaching measures that sub- stantially reduced the role of government in the economy. The financial sector and the trade and foreign exchange regimes were substantially liberalized. These measures did involve wide consultations and discus- sions with donor agencies as well as substantial preliminary work by government and the stakeholders. This reveals not only government at- tempts to own the reform process but also to involve key players in the design of economic policy. Uganda has experienced a smooth relationship between aid and poli- cies since the mid-1980s. Figure 3.6 traces the relationship between aid/ GDP and the World Bank's Country Policy and Institutional Assessment (CPIA). The latter is a broad measure of economic policy that covers macroeconomic management, as well as the effectiveness of the public sector in improving services that are essential for growth and poverty reduction." A distinct change in the policy index can be noted in 1987- 88, when the new government radically changed its course. This trig- gered large amounts of financial aid, and more bilateral donors became active in Uganda. This was also the period, according to government officials, when aid played its most decisive role in shaping policy. Tech- nical assistance and especially financial aid and the conditionality linked to it helped push the balance in favor of the reformers in government during this period of reluctance (1987-92). However, when the govern- ment acquired full ownership of its reform program from 1992 onwards, the aid/GDP ratio started to decline. To have maximum effect on pov- erty reduction, aid should instead have increased at least in tandem with the policy improvements. Other studies show that Uganda and countries 138 Successful Reformers FIGURE 3.6 UGANDA: AID AND POLICY, 1970-96 Aid/GDP (percent) Policy index 6 4.0 - 3.5 5 -3.0 4 Policy 7 -2.5 3 2.0 -1.5 2 J .1.0 1 - ,"Aid/GDP -0.5 0 l l 0.0 1970-73 1974-77 1978-81 1982-85 1986-89 1990-93 1994-97 Source: World Bank. with similar characteristics get less aid than they ought to if effective poverty alleviation is indeed the overarching goal of foreign aid." Surprisingly, IMF, IDA, and the European Commission lowered their disbursement in nominal terms drastically in 1996. That does not neces- sarily indicate a permanent shift, and it will require a few years more to assess the situation. Audits and follow-up of World Bank adjustment lending point in the direction of rather successful operations during the 1990s. The overall rating for Structural Adjustment Credits I and 11 (1991-94 and 1994-96) are tentatively set at satisfactory to highly satisfactory, while the institu- tional development impact of the operations would be partial to sub- stantial (World Bank 1997,1999). The degree of government commitment to reform is closely correlated to the outcome ratings. The success rate of adjustment lending follows closely the improvements in government ownership, with better outcome in the 1990s than in the 1980s when reforms were either rejected or implemented reluctantly. Four of the six adjustment operations supported by the World Bank during the period Uganda 139 have been deemed successful, which is a high rating compared to similar operations in most other African countries. Financial Aid versus Nonfinancial Aid Financial Aid Meaningful reforms start, in general, with the government acting of its own accord. Ideally, aid donors should then come in and refine and re- direct the reforms. If donors dictate reforms there is a high risk of unsus- tainable results. When the government of Uganda prepares reforms it starts by discussing them internally and with the IMF and the World Bank. Officials of the Ministry of Finance, Planning and Economic De- velopment interviewed in Kampala in 1999 claimed that they focus on how to get the reforms right rather than on finding potential financial support for the reforms. Reforms should not be triggered by the pros- pects of acquiring funding, or at least this should happen very seldom. This is probably more true for the period after 1992 than it was in the late 1980s, when the government adopted the IMF program because it needed to obtain access to donor funds (IMF 1998). Funding is nevertheless needed in almost every case where reforms are undertaken, even if it is not the main cause of the reform. Most re- forms implemented in Uganda could not have been implemented with- out donor support. At the same time, there were a few reforms undertaken that did not depend on financial assistance, such as the liberalization of coffee marketing, the primary education program, and decentralization to the district level (interviews, MFPED, 1999). In the end, the commit- ment to reform can be measured by the political will to sustain the re- forms. One example is the national pride in keeping inflation down, which would have been unthinkable ten years ago (interview, IMF, Kampala, 1999). Nonfinancial Aid A key principle of the aid and reform relationship is that ideas "stick" but projects do not. Ideas-about good governance, trade liberalization, or lowering inflation, for example-may take time to become accepted, but once established they can have sustainable effects. Individual projects on the other hand can have an immediate impact, but this effect may be short-lived or taper off over time. The key challenge for donors is there- fore to support ideas in order to help reforms take off. Experience suggests that nonfinancial aid should come first, while financial aid has a more important role to play later in the process by 140 Successful Reformers supporting the chosen path. This principle is probably more applicable in cases where there is greater government ownership of the reform program. Technical assistance has had a great impact on capacity building in Uganda, especially in relation to the Ministry of Finance and Economic Planning and the Bank of Uganda. Local observers also said that the success of the Ugandan reforms to a large extent can be attributed to nonfinancial aid (interview, EPRC, 1999). The commitment to technical assistance is critical for its success, as is the government's vision and its capacity to articulate its needs. There needs to be a well-articulated demand for the advice. A committed gov- ernment will benefit from the assistance and bring the right people in. Successful TA also depends on the culture of the ministry, the availabil- ity of the local staff, and the behavior of the expatriate adviser. Furthermore, policy dialogue and analytical work, in particular by the World Bank and IMF together with the government, have played a critical role in terms of setting the agenda and formulating priorities in the reform program. The consultation process has helped give broad exposure to the reform ideas and justify the priorities. For example, World Bank economic and sector work has been critical in the recent focus on poverty eradication (interview, World Bank, Kampala, 1999). Uganda has a reputation for making good use of TA. TA has been effective and internalized. In the case of the Finance Ministry, its offi- cials said, technical assistance has had a true advisory role and has been integrated in the policymaking process. The ministry has also benefited from donor-provided allowances for training and salaries for local pro- fessionals. Young Ugandan economists have learned a lot from the ex- patriates. The long-term effects of TA can be noted in the increased levels of skills, knowledge, and policy reforms compared to ten years ago (in- terview, MFPED, 1999). According to IMF representatives, in general the Fund's nonfinancial support can help trigger reforms while its funding seldom does so. "It is critical that the outcome of TA can be measured, otherwise there is an imminent risk for failure. Performance criteria should therefore be built into the programs" (interview, IMF, Kampala, 1999). In annual ratings by the Operations Evaluation Department of the World Bank, technical assistance projects often score very low. In gen- eral it is hard to find successful TA projects, especially in Africa. How- ever, the department's audit of the second and third technical assistance projects in Uganda found that they had been rather successful. The two projects had the twin objectives of facilitating the short-term design and implementation of economic reform measures and fostering long-term institutional development and capacity building related to improving economic management. The principal lesson was that man- Uganda 141 agement matters most. TA 11 (1983-92) had a marginally satisfactory out- come, while TAIII (1988-94) had a highly satisfactory outcome. The main shortcomings of TA II were that the overall political and institutional environments were not conducive to either short-term policy reforms or long-term institutional development. The strengths of TA III were that all of its components were highly relevant, that the borrower was strongly committed to achieving project objectives, and that project management on both the government and World Bank sides was good. The correlation between the success rate and the timing of the two operations is consis- tent with general findings of this study: that the commitment to and own- ership of the reform program was being established and strengthened during this period. It seems plausible that Uganda made better use of TA when its own commitment to reform grew stronger. Conditionality Government officials claim that it is solely a domestic decision whether or not to reform. Conditionality can tip the balance in favor of a reform pro- gram but cannot by itself induce the reforms. When the government be- lieves that conditions facilitate a program's objectives then conditionality can come to reflect a negotiated understanding between the government and the donor. Conditionality becomes part of a consensus-building pro- cess in which the government and donors together design the reforms. Such consensus-oriented conditionality has increasingly come to charac- terize the Uganda reform program over the years. Conditionality was used by the reformers within the government to help convince others of the advantages of the planned reforms during the period of reluctant reform (1987-92). The proponents of conditional- ity also found that it introduced a necessary rule-based system with checks, balances, and benchmarks to measure progress. In some cases the reformers even asked the Bretton Woods institutions (primarily the World Bank) to introduce certain conditions that they wished to have included. In this way coffee liberalization (abolishing the floor price of coffee), which was resisted by the then-minister of agriculture, came to be included as a condition in Structural Adjustment Credit II by the World Bank. In certain cases the government has also taken reforms a step far- ther than required by the conditionality in the loan contracts. Trade lib- eralization, for instance, went farther than the conditions in the World Bank's adjustment loans (interview, MFPED, 1999). As discussed earlier, in a situation with established government ownership conditionality tends to become less relevant, especially given the huge increase lately in the number of conditions linked to each loan. Tumusiime-Mutebile (1995) believes that "there is no reason why sound economic management should be driven, or even constrained, by 142 Successful Reformers conditionalities attached to donor assistance." On the contrary, condi- tionality can have negative effects if loans and credits are overloaded with conditions. IMF and World Bank representatives in Kampala said that the role of the two institutions is to strengthen the hands of the reformers (inter- views, IMF and World Bank, Kampala, 1999). The Washington institu- tions and the reformers often have the same objectives and they seldom disagree on issues in the context of conditionality. Analytical work and conditionality have in many cases been a fruitful combination for set- ting the agenda, and conditionality works more ex post and functions as benchmarking. One school of thought is that conditionality needs to be precise and defined in numeric terms to be effective, making IMF pro- grams more useful than World Bank conditional lending. Several NGOs and private sector representatives, on the other hand, say that the government of Uganda has had to agree to the policies proposed and accept conditionality in order to receive financial aid. "The government owns the reform program but the program is based on IMF-World Bank conditionality rather than local participation. The government adapts to the donors' priorities and there are still signifi- cant policy conditions" (interviews, UDN and UMA, 1999). Some NGOs were grateful for certain donor conditions such as the social sector spending targets linked to the HIPC Initiative, but they also criticized donors for not sticking to their conditions and for being too weak in penalizing nonfulfillment. Triggers for Reforn Leadership Political leadership, government commitment, and domestic ownership are key to any reform program. Aid and external support can only be complementary in the process. In the end the success of policy reforms comes down to political leadership. This has to a large extent proven to be the case for Uganda, where a strong leadership has gradually given rise to a deepened government commitment as well as broader domes- tic ownership. The government's capacity to learn from its own mistakes has also been decisive for the formulation of the reform program. In the early days of his presidency Museveni rejected market-based solutions and did not want to deal with the IMF. A congruence of events and the poor economic performance in 1986 led to a reassessment of policies in 1987. The country was in a post-conflict situation and had terms-of-trade prob- lems arising from a slump in coffee prices and other circumstances. All these factors contributed to a rethinking of policies, and they were cer- Uganda 143 tainly more critical than external "pressure" in triggering the new reform program in 1987. Policy formulation benefited from a process of "learning by doing." Since the period of reluctance, and especially since the 1989 national conference on the economy and the 1992 cabinet restructuring, reforms have been driven by strong leadership with a core team at the Ministry of Finance, Planning and Economic Development. Most important is that the economic team has continued intact over the years, and that the presi- dent himself has taken an active interest in the reform program. As stated above it has also been claimed that the president and the government have decided to reform, irrespective of potential aid. This is probably more true for the period since 1992 than it was during the pe- riod of reluctant reforms, when aid and conditionality apparently played a critical role in the reform program. Learningfrom Neighbors Education, training, seminars, and study tours have been important in the generation of reform ideas, according to the government. This has included visits to other countries with successful reform experiences. Ghana has played a notable role in Uganda's generation of ideas and reforms; several delegations from different ministries have been sent to Accra to learn from the Ghanaian program. Today other African coun- tries are sending delegations to Kampala for the same type of advice and learning. Government Ownership Endogenous forces have to a large extent driven the reform agenda. There have been very few policy reversals. During the early reform program, beginning in 1987, it was primarily the Ministry of Planning and Eco- nomic Development and reform-minded members of the cabinet who drove the reform program. The base for the reform program has wid- ened considerably since then. What used to be debate within an inner circle in the late 1980s has now in many instances become a broad-based dialogue involving many stakeholders. The same goes for the government budget, which is now made with the participation of many actors. The budget process takes off with a two-day seminar in which donors, NGOs, and other stakeholders take part, resulting in broader understanding and ownership of the budget. Decentralization and district focus have contributed to a closer involve- ment of stakeholders in the budget process. Further evidence that government ownership is essential to a suc- cessful program is the fact that program implementation is still slowest 144 Successful Reformers in areas where disagreements exist between the government and the in- ternational financial institutions. This was the case, for instance, with regard to civil service reform. In the early years of reform there was a considerable goodwill toward the reform undertakings, which had few losers. The public sector was small, government employees could hardly survive on their salaries, and there was no large protected import substitution sector. The cost of ad- justment has been relatively low in Uganda (IMF 1998; Tumusiime- Mutebile 1995), as a result of the country's particular circumstances when it entered into the reform program. In a post-conflict situation with a rundown economy, Uganda had hardly any price controls or subsidies on essential commodities in place, so there was no need to liberalize prices or remove subsidies to the detriment of poorer groups. Food consumption had remained robust over the years, thanks to fertile soils. However, the gains of the reform program could be questioned if the income gap between different groups is widening. IMF (1998) finds that policy change has brought only limited benefits to many poor Ugandans because of their initial dependence on subsistence agricul- ture. The government's Poverty Eradication Action Plan of 1997 stated that "regardless of the stable and conducive macroeconomic environ- ment the poor households have not fully benefited from the macroeco- nomic stability." Participatory Process Public dialogue is very active in Uganda and the government is now more open to discussion with a greater degree of openness. Reform poli- cies have been more internalized by both civil servants and those out- side government. The discussion on reforms in the society has become fairly open in this new era of "government by workshop" (interview, World Bank, Kampala, 1999). The Ministry of Finance for its part "advo- cates inclusion of stakeholders in the reform dialogue. It strengthens the understanding and outcome of the reforms and makes them less revers- ible" (interview, Tumusiime-Mutebile, 1999). If the next step is to move from government ownership to strength- ened domestic ownership, this process of inclusion needs to be deep- ened and broadened. In the end it relies not only on the political will of the government to open up, but also on capacity building among the stakeholders to prepare them for this role. A broadening of ownership is particularly necessary now that the reform program is shifting its main focus from macroeconomic reforms to public service delivery. Nonetheless, tensions still exist between the government and outside stakeholders. Persons outside government voiced criticisms in interviews to the effect that the government holds seminars, meets with task forces, Uganda 145 and so forth, but that those activities do not influence real policymaking. "There is a political will to involve stakeholders but investments are also needed in order to form the partnership" (interview, Coopibo-Uganda, 1999). Representatives of stakeholders also stressed that it is important for outside actors to keep their independence and retain their role for policy lobbying. There are few well-organized groups, and those that exist often lack the necessary capacity, skills, and equipment to express their thoughts convincingly. Capacity Building Donor support for capacity building has to a large extent focused on the key players in policy formulation-the Ministry of Finance and the Bank of Uganda. In order to keep up with an evolving reform agenda and deepen the ownership of reforms, however, there is a need to expand capacity in other ministries and agencies as well. "If line ministries are constrained by capacity, they are not only inferior in generating ideas, but also in the dialogue with donors as well as within the government in the budget process" (interview, PRESTO, 1999). Limited institutional capacity is still regarded as a potential risk by the World Bank in imple- mentation of the structural adjustment credits. Capacity building is also needed among NGOs and different stake- holders to prepare them to take an active and well-informed role in the policy discussion. Furthermore, the knowledge necessary to identify, review, and analyze government policies would strengthen the domes- tic ownership of the reforms. Many of those interviewed said that the policy dialogue should be based on facts rather than beliefs. Such capac- ity building would also help to improve aid effectiveness, since a great deal of aid is channeled through NGOs. LESSONS OF UGANDA'S REFORM PROCESS A primary lesson to be drawn from the period since 1987 is that owner- ship of a reform program is critical to its success. The time element also suggests that reforms are a long process that requires political stability and macroeconomic continuity in addition to patience and transparency. It is important to note that over the last 12 years, many stakeholders have become involved in the broad-based dialogue with government regarding the reforms. Organized groups such as the Parliament, the private sector (notably the Uganda Manufacturers Association), NGOs, the press, and academics are now greatly involved in the policy debates about future reforms. However, there is still a problem in that only a few well-organized groups exist and these typically lack capacity to 146 Successful Reformers articulate their positions forcefully and convincingly using empirical analysis. Homegrown programs continued to be implemented in this period. The government initiated measures for which donors mainly provided financing and technical assistance. The Ugandan authorities have em- phasized pro-developmental and poverty-reducing strategies in addi- tion to good macroeconomic policies. These have included the introduction of universal primary education, for which substantial resources from both multilateral and bilateral sources have been mobilized, as well as other initiatives such as the drafting and promulgation of a homegrown consti- tution. Second, on the basis of the Uganda experience we contend that macro- economic and political stability trigger increased donor interest in the re- form program. And the development of and commitment to a good reform program open an avenue for providing financing with less conditionality. With respect to key lessons learned from Uganda's experience of re- form, our findings tend to agree with some earlier ones, such as those contained in the World Bank Implementation Completion Report on Uganda's SAC I and II. This report concludes, first, that the programs' focus and internal coherence improved the performance of government and the private sector environment. It says that as the first-generation problems are tackled, new issues emerge that may call for a simulta- neous macroeconomic stabilization supplemented by microeconomic and sector reforms. Consequently, the unfinished reform agenda calls for continued balance of payments support and adjustment. Secondly, institutional capacity is a very important determinant of success in implementation. Political support notwithstanding, sustain- able deployment of available managerial talent through donor-financed technical assistance and related programs has been useful in ensuring successful implementation of structural adjustment programs. Accord- ingly, the future reform program needs to incorporate appropriate insti- tutional building of local capacity. The design of future programs should seek to increase understanding and ownership of the reforms by all key stakeholders in the private sector, the NGO community, and the country's political leadership. Uganda 147 APPENDIX 3.1 UGANDA: DATA AND STATISTICS The External Aid Co-ordination Department in the Ministry of Finance, Planning and Economic Development in Kampala maintains a database on external aid to Uganda, including commitments, aid terms, disburse- ment, and the sectors receiving assistance. However, the department has not published aid statistics regularly. UNDP in Uganda produces an annual "Development Co-operation Report" that provides comprehen- sive data on external development assistance to Uganda. This report has been produced in collaboration with the government since 1995. In or- der to increase government ownership of the report, UNDP has sup- ported the participation of the Ministry of Finance in the report's production. TABLE 3.6 AID COMMITMENTS (ODA) TO UGANDA, 1962-97 (thousands of U.S. dollars) Creditor/donor 1962-71 1972-78 1979-85 1986-97 European Union - - 100,564 692,547 EU countries - - 100,824 1,942,793 North America - - 23,296 458,537 Russia 19,746 - - 255 Eastern Europe - - 9,034 34,930 Japan - - - 197,241 China - - - 175,618 Asia (excluding Japan) - - 8,000 93,512 Arab countries - 99,430 20,415 167,343 Other bilaterals - - - 49,109 IMF - 182,368 307,365 507,594 World Bank - - 206,323 1,694,324 African Development Bank 4,284 23,640 133,867 476,381 Other multilaterals - 27,090 216,864 700,245 Total 24,030a 332,527 1,126,552 7,190,428 - Not available. a. The disbursement for this period differs significantly from the US$94.08 million re- ported in the government of Uganda's "External Debt Management Report" of 1991. The totals for other periods are comparable. Source: Aid Data Unit, Ministry of Finance, Planning and Economic Development. TABLE 3.7 UGANDA: OFFICIAL DEVELOPMENT FINANCE, 1980-98 (millions of U.S. dollars) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 All sources Gross disbursement on official loans 95.8 207.3 173.1 199.4 106.4 153.5 135.5 317.4 207.3 321.1 343.5 258.9 305.7 391.4 323.7 288.2 288.5 Non-technical assistance grants 61.6 100.7 55.5 76.6 66.5 47.4 81.6 96.4 187.1 176.5 257.5 314.2 336.2 259.7 319.0 399.1 338.5 Technical assistance grants 21.1 33.6 29.5 33.9 32.4 36.9 50.0 57.4 75.7 87.7 93.8 107.2 121.8 124.2 124.7 148.2 140.1 Total official development finance 178.5 341.6 258.1 309.9 205.3 237.8 267.1 471.2 470.1 585.3 694.8 680.3 763.7 775.3 767.4 835.5 767.1 Bilateral sources Gross disbursement on official loans 11.2 3.0 7.3 5.8 6.8 8.3 24.9 89.4 30.7 38.7 41.1 8.9 45.3 89.5 18.0 24.4 46.0 Non-technical assistance grants 31.6 77.4 40.7 54.3 34.4 24.6 56.7 53.8 134.9 124.3 204.5 268.1 169.7 213.2 231.2 249.5 257.3 Technical assistance grants 11.1 13.6 13.5 15.0 12.8 15.4 23.7 28.0 38.4 43.5 53.6 63.8 76.0 78.9 88.5 116.7 100.7 Total official development finance 53.9 94.0 61.5 75.1 54.0 48.3 105.3 171.2 204.0 206.5 299.2 340.8 291.0 381.6 337.7 390.6 404.0 Multilateral sources Gross disbursement on official loans 84.5 204.3 165.7 193.7 99.6 145.2 110.6 228.0 176.6 282.4 302.4 250.0 260.4 301.9 305.6 263.8 242.5 Non-technical assistance grants 30.0 23.3 14.8 22.3 32.1 22.8 24.9 42.6 52.2 52.2 53.0 46.1 166.5 46.5 87.8 149.6 81.2 Technical assistance grants 10.0 20.0 16.0 18.9 19.6 21.5 26.3 29.4 37.3 44.2 40.2 43.4 45.8 45.3 36.2 31.5 39.4 Total official development finance 124.5 247.6 196.5 234.9 151.3 189.5 161.8 300.0 266.1 378.8 395.6 339.5 472.7 393.7 429.6 444.9 363.1 Note: The grant equivalent of a $100 million loan disbursement is $20 million when the loan is 20 percent concessional. Source: Chang, Femrndez-Arias, and Serv6n 1999. TABLE 3.8 UGANDA: EFFECTIVE DEVELOPMENT ASSISTANCE, 1980-96 (millions of U. S. dollars) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 All sources Grant equivalents of loan disbursements 35.0 19.4 47.5 43.0 58.4 87.2 87.4 124..0 123.1 195.8 219.9 202.8 209.1 158.5 231.3 194.2 132.2 Non-Technical assistance grants 81.8 100.7 55.5 76.6 68.5 47.4 81.6 96.4 187.1 176.5 257.5 314.2 336.2 259.7 319.0 399.1 338.5 Technical assistance grants 22.1 33.8 28.5 33.9 32.4 36.9 50.0 57.4 75.7 07.7 93.8 107.2 121.8 124.2 124.7 148.2 140.1 Total effective development assistance 117.7 153.7 132.5 154.4 157.3 171.5 199.0 277.8 385.0 460.0 571.2 624.2 867.1 542.4 875.0 741.5 610.8 Bilateral sources Grant equivalents of loan disbursements 8.3 2.1 1.5 1.5 2.2 0.7 1.3 1.2 3.5 43.3 9.0 7.1 8.5 28.2 7.6 8.0 21.1 Non-technical assistance grants 31.6 77.4 40.7 54.3 54.4 24.0 85.7 53.8 134.9 124.3 204.5 200.1 100.7 213.2 231.2 240.6 257.3 Technical assistance grants 11.1 13.6 13.5 16.0 12.8 15.4 23.7 26.0 38.4 43.5 53.6 83.8 76.0 78.9 86.6 116.7 100.7 Total effective development assistance 49.0 93.1 55.7 70.8 49.4 40.7 81.7 83.0 176.8 211.1 267.1 339.0 254.2 318.3 327.3 374.2 379.1 Multilateral sources Grant equivalents of loan disbursements 28.7 17.3 46.0 42.3 58.2 86.5 66.2 122.7 119.7 152.6 210.9 195.7 200.7 132.3 223.7 186.2 111.1 Non-technical assistance grants 30.0 23.3 14.8 22.3 32.1 22.8 24.9 42.8 52.2 52.2 53.0 46.1 166.5 46.5 87.8 149.6 81.2 Technical assistance grants 10.0 20.0 16.0 18.8 19.6 21,5 28.2 29.4 37.3 44.3 40.2 43.4 45.8 45.3 38.2 31.5 39.4 Total effective development assistance 68.7 60.6 76.8 83.5 107.9 130.8 117.4 194.7 208.2 248.0 304.1 285.2 413.0 224.1 347.7 367.3 231.7 Note: The grant equivalent of a $100 million loan disbursement is $20 million when the loan is 20 percent concessional Source: Chang, Femrndez-Arias, and Serven 1999. TABLE 3.9 UGANDA: TOTAL OFFICIAL DEVELOPMENT FINANCE BY CREDITOR, 1980-96 (millions of U. S. dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 53.9 84.0 61.5 75.1 54.0 48.3 105.3 171.2 204.0 206.5 299.2 340.8 291.0 381.6 337.7 390.6 404.0 Multilateral 124.5 247.6 198.5 234.9 151.3 189.5 161.8 300.0 266.1 378.8 395.6 339.5 472.7 393.7 429.6 444.9 383.1 Total 178.5 341.6 258.1 309.9 205.3 237.8 267.1 471.2 470.1 586.3 694.8 680.3 763.7 775.3 767.4 835.5 787.1 Bilateral Australia 1.8 3.2 4.6 4.2 3.5 0.3 0.2 0.1 0.6 0.6 0.5 0.5 0.9 1.0 1.3 1.2 1.6 Austria 0.1 0.1 0.0 0.0 0.1 0.0 0.1 0.1 0.1 11.4 2.1 2.8 4.3 17.2 13.1 15.4 15.3 Belgium 0.2 0.5 0.2 0.4 0.3 1.7 0.2 0.5 0.5 0.4 10.6 0.8 0.9 1.5 1.9 5.3 1.8 Canada 0.9 3.3 4.9 2.0 6.5 2.6 1.7 2.1 21.8 6.6 7.1 5.3 9.1 3.7 4.7 1.6 1.7 Denmark 1.7 1.2 1.3 2.3 2.6 0.8 2.8 4.9 11.6 15.6 25.4 32.4 37.1 45.4 53.0 60.1 68.0 Finland 0.2 1.2 0.8 0.9 0.1 0.4 1.6 0.7 0.4 7.8 1.5 1.5 1.5 0.5 0.3 1.3 1.1 France 12.0 2.1 3.0 0.9 3.1 0.8 0.9 1.1 1.1 7.0 8.9 8.8 9.8 4.8 3.5 10.3 13.7 Germany, Fed. Rep. of 8.4 11.4 7.7 29.4 3.5 4.4 12.9 18.4 19.9 19.9 27.0 34.2 23.0 45.3 29.5 47.2 40.1 Ireland 0.0 0.0 0.0 0.0 0.1 0.0 0.1 0.1 0.2 0.1 0.1 0.1 0.3 0.6 2.0 5.0 4.9 Italy 0.1 1.0 1.0 1.9 5.2 7.1 30.0 18.4 23.4 13.0 15.8 14.1 13.5 14.4 5.6 3.6 7.0 Japan 0.1 1.0 1.5 4.1 5.6 3.2 3.1 0.6 9.5 1.1 9.5 14.9 14.6 39.7 41.0 23.9 49.8 Luxembourg 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 Netherlands 4.3 21.1 2.1 7.3 4.6 1.9 6.3 2.2 3.6 3.1 3.8 22.9 15.1 23.1 25.5 31.8 32.6 New Zealand 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.0 Norway 0.4 1.2 7.0 1.2 1.1 1.2 2.2 2.7 3.5 4.5 7.4 5.2 8.2 13.3 18.8 20.5 20.5 Portugal 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Spain 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.1 2.7 5.1 0.0 4.4 13.8 15.8 1.8 0.2 13.8 Sweden 2.4 3.8 0.4 0.7 0.2 0.0 1.9 4.0 11.2 16.3 14.5 34.3 29.0 16.6 25.4 25.3 32.7 Switzerland 0.7 0.8 0.5 0.5 0.3 0.3 1.8 1.8 9.0 3.1 12.5 7.2 8.3 7.3 0.8 8.9 0.7 United Kingdom 8.7 30.8 18.6 9.7 10.6 15.1 24.4 23.8 52.8 40.7 38.0 51.5 40.1 54.8 52.0 85.4 82.2 United States 13.0 8.0 2.0 4.0 3.0 5.0 4.0 14.0 18.0 21.1 59.4 51.7 22.0 57.0 51.0 48.0 28.0 Arab countries 1.3 0.1 0.1 0.2 0.2 0.1 0.0 0.0 0.0 28.5 40.0 50.0 15.5 0.2 0.2 0.1 0.1 Unspecified 1.6 3.0 6.1 5.4 3.3 3.8 11.3 75.6 14.8 2.2 17.0 0.4 23.0 19.5 6.7 14.4 7.6 Total 53.9 94.0 81.5 75.1 54.0 48.3 105.3 171.2 204.0 206.5 299.2 340.8 291.0 381.6 337.7 390.8 404.0 Multilateral AfDB 4,1 6.6 8.0 8.7 7.4 32.0 5.2 8.8 2.7 5.3 0.0 2.4 3.0 2.5 0.2 6.7 9.7 ADF 0.0 0.0 8.1 0.0 0.0 0.0 0.8 2.0 21.2 38.1 11.5 20.1 32.1 25.7 17.9 32.0 32.3 EEC 25.1 13.1 11.4 17.8 22.0 17.2 17.8 32.3 38.8 38.4 38.0 27.9 158.1 29.8 53.5 108.3 46.1 EIB 0.0 0.0 7.4 2.2 0.0 0.0 0.0 0.0 1.8 0.0 0.2 1.5 0.7 7.9 2.0 0.0 2.9 IDA 1.1 15.4 31.1 32.6 58.2 91.8 62.1 112.0 73.9 97.7 208.1 136.5 155.7 141.1 223.2 159.8 123.5 IFAD 0.0 0.0 0.5 6.4 0.0 6.8 9.8 3.5 0.0 0.0 0.0 3.7 3.5 8.5 3.8 2.8 4.0 IFC 0.0 0.0 0.0 9.8 0.0 2.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 IMF 77.9 144.8 93.8 113.8 21.5 0.0 0.0 58.8 73.6 51.1 81.2 78.4 58.1 0.0 52.5 55.0 63.2 Other U N 0.2 0.1 1.5 0.4 0.5 0.8 0.7 1.2 5.8 5.8 5.0 2.3 1.4 0.9 1.2 2.1 2.5 UNDP 3.3 13.9 8.4 6.8 7.0 8.3 7.0 8.8 12.7 15.6 16.2 17.1 18.0 14.8 11.8 7.1 19.0 UNFPA 0.2 0.1 0.0 0.1 0.2 0.2 0.2 0.4 0.8 2.5 3.0 3.2 2.6 2.1 4.5 4.8 2.9 UNHCR 2.9 1.7 1.3 4.4 5.4 3.8 10.5 7.5 5.8 6.2 2.5 4.0 5.5 8.0 16.3 19.8 21.1 UNICEF 3.1 3.9 4.4 6.1 5.4 7.6 8.2 7.0 9.4 12.5 11.8 14.9 18.5 17.2 16.3 14.2 13.0 UNTA 0.3 0.3 0.4 1.3 1.1 1.2 1.1 1.5 1.2 1.8 1.3 1.9 1.3 2.1 1.8 3.4 1.7 WFP 4.9 12.7 9.7 9.9 13.5 7.7 9.2 14.3 15.4 15.0 16.5 18.1 10.8 17.0 18.4 23.6 14.3 Arab agencies 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.2 0.8 0.9 0.1 0.1 0.3 0.2 0.0 0.0 Unspecified 1.4 35.2 10.5 14.9 9.1 9.3 31.3 43.9 3.0 89.8 3.4 7.5 9.2 118.2 5.7 6.7 8.9 Total 124.5 247.8 198.5 234.9 151.3 189.5 161.8 300.0 288.1 378.8 305.8 330.6 472.7 303.7 420.8 444.0 363.1 Note: Official development finance = gross disbursements on all official loans + official grants. Source: Chang, Fernandez-Arias, and Serven 1999. TABLE 3.10 UGANDA: TOTAL EFFECTIVE DEVELOPMENT ASSISTANCE BY CREDITOR, 1980-96 (millions of U. S. dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 49.0 93.1 55.7 70.8 49.4 40.7 81.7 83.0 178.8 211.1 287.1 339.0 264.2 318.3 327.3 374.2 379.1 Multilateral 68.7 80.8 76.8 83.5 107.9 130.8 117.4 194.7 209.2 249.0 304.1 285.2 413.0 224.1 347.7 367.3 231.7 Total 117.7 153.7 132.5 154.4 157.3 171.5 199.0 277.8 385.9 480.0 571.2 824.2 887.1 542.4 675.0 741.5 810.8 Bilateral Australia 1.8 3.2 4.6 4.2 3.5 0.3 0.2 0.1 0.6 0.6 0.5 0.5 0.9 1.0 1.3 1.2 1.8 Austria 0.1 0.1 0.0 0.0 0.1 0.0 0.1 0.1 0.1 10.1 1.8 2.5 4.3 12.0 11.2 10.8 14.3 Belgium 0.2 0.5 0.2 0.4 0.3 1.7 0.2 0.5 0.6 0.4 10.8 0.6 0.9 1.5 1.9 5.3 1.8 Canada 0.9 3.3 4.9 2.0 8.5 2.8 1.7 2.1 21.8 8.6 7.1 5.3 9.1 3.7 4.7 1.6 1.7 Denmark 1.7 1.2 1.3 2.3 2.6 0.8 2.8 4.9 11.8 15.8 25.4 32.4 37.1 45.4 53.0 60.1 68.0 Finland 0.2 1.2 0.6 0.9 0.1 0.4 1.8 0.7 0.4 5.9 1.5 1.5 1.5 0.5 0.3 1.3 1.1 France 8.3 2.1 2.8 0.9 2.5 0.8 0.9 1.1 1.1 2.1 7.4 8.7 4.9 4.8 3.5 10.3 13.7 Germany, Fed. Rep. of 8.4 11.4 7.7 29.4 3.5 4.4 12.9 15.8 18.7 19.9 27.0 34.2 23.0 37.9 29.5 47.2 40.1 Ireland 0.0 0.0 0.0 0.0 0.1 0.0 0.1 0.1 0.2 0.1 0.1 0.1 0.3 0.6 2.0 5.0 4.9 Italy 0.1 1.0 1.0 1.9 4.9 7.1 30.0 18.4 18.2 10.8 15.8 13.1 11.7 10.2 5.5 3.6 7.0 Japan 0.1 1.6 1.5 4.1 5.6 3.2 3.1 0.6. 9.5 1.1 9.5 14.9 14.6 28.0 38.7 23.9 40.4 Luxembourg 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 Netherlands 4.3 21.1 2.1 7.3 4.8 1.9 8.3 2.2 3.5 3.1 3.5 22.9 15.1 23.1 25.5 31.8 32.6 New Zealand 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.0 Norway 0.4 1.2 7.0 1.2 1.1 1.2 2.2 2.7 3.5 4.5 7.4 5.2 8.2 13.3 18.8 20.5 20.5 Portugal 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Spain 0.0 0.0 0.0 0.0 0.0 0.0 0.0 -0.1 -0.1 0.0 0.0 1.6 5.0 6.2 0.6 0.2 6.0 Sweden 2.4 3.8 0.4 0.7 -.2 0.0 1.9 4.0 11.2 16.3 14.5 34.3 29.0 16.6 25.4 26.3 32.7 Switzerland 0.7 0.8 0.5 0.5 0.3 0.3 1.8 1.8 9.0 3.1 12.5 7.2 9.3 7.3 0.6 8.9 0.7 United Kingdom 6.7 30.8 18.8 8.3 8.6 10.8 12.0 13.4 50.0 40.7 36.0 51.5 40.1 64.0 52.0 85.4 62.2 United States 13.0 8.0 2.0 4.0 3.0 5.0 4.0 14.0 16.5 18.9 46.0 51.4 22.0 57.0 51.0 49.0 29.0 Arab countries 1.3 0.1 0.1 0.2 0.2 0.1 0.0 0.0 0.0 28.5 40.0 50.0 15.5 0.2 0.2 0.1 0.1 Unspecified 0.4 2.1 0.8 1.5 0.7 0.3 0.1 0.8 0.7 26.8 0.3 3.1 1.8 -5.4 1.6 2.8 1.7 Total 49.0 93.1 55.7 70.8 49.4 40.7 81.7 83.0 176.8 211.1 287.1 339.0 264.2 318.3 327.3 374.2 379.1 Multilateral AIDB 0.2 0.8 1.4 2.2 1.5 0.7 0.3 0.9 0.1 0.1 0.0 0.1 0.2 0.0 0.0 0.1 0.3 ADF 0.0 0.0 6.9 0.0 0.0 0.0 0.8 1.0 15.9 28.0 0.1 14.5 23.4 18.2 12.2 21.6 21.8 EEC 25.1 12.7 10.4 16.9 21.4 18.8 17.5 32.1 38.5 36.4 36.0 27.9 156.1 29.6 53.5 106.3 46.1 EIB 0.0 0.0 1.2 1.7 0.0 0.0 0.0 0.0 0.8 0.0 0.1 0.7 0.3 4.6 1.5 1.0 1.7 IDA 0.5 13.4 27.3 28.4 50.6 77.4 52.8 91.2 58.7 75.8 180.0 104.8 115.5 102.8 157.7 110.7 84.5 IFAD 0.0 0.0 0.4 5.2 0.0 5.3 7.5 2.6 0.0 0.0 0.0 2.5 2.5 4.5 2.4 1.7 2.5 IFC 0.0 0.0 0.0 -1.4 0.0 -0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 IMF 27.9 0.2 0.0 0.0 0.0 0.0 0.0 24.3 37.4 48.1 41.4 73.1 51.8 0.0 48.2 51.0 0.0 Other UN 0.2 0.1 1.5 0.4 0.5 0.8 0.7 1.2 5.8 5.6 5.0 2.3 1.4 0.0 1.2 2.1 2.5 UNDP 3.3 13.9 8.4 6.6 7.0 8.3 7.0 8.8 12.7 15.6 18.2 17.1 18.0 14.8 11.8 7.1 19.0 UNFPA 0.2 0.1 0.0 0.1 0.2 0.2 0.2 0.4 0.8 2.5 3.0 3.2 2.8 2.1 4.5 4.6 2.9 UNHCR 2.9 1.7 1.3 4.4 5.4 3.8 10.5 7.5 5.8 6.2 2.5 4.0 5.5 8.0 16.3 19.8 21.1 UNICEF 3.1 3.9 4.4 6.1 5.4 7.6 6.2 7.0 9.4 12.5 11.8 14.9 16.5 17.2 16.3 14.2 13.0 UNTA 0.3 0.3 0.4 1.3 1.1 1.2 1.1 1.5 1.2 1.8 1.3 1.9 1.3 2.1 1.8 3.4 1.7 WFP 4.9 12.7 9.7 9.9 13.5 7.7 9.2 14.3 15.4 15.0 16.5 18.1 10.8 17.0 18.4 23.6 14.3 Arab agencies 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.2 0.8 0.9 0.1 0.1 0.3 0.2 0.0 0.0 Unspecified 0.1 0.5 3.5 1.7 1.2 1.8 3.9 1.4 6.3 0.5 1.3 0.0 6.9 2.1 1.8 1.1 0.5 Total 68.7 60.6 78.8 83.5 107.9 130.8 117.4 194.7 209.2 249.0 285.2 304.1 285.2 413.0 224.1 347.7 231.7 Note: Effective Development Assistance = grant equivalents of official loan disbursements + official grants. Source: Chang, Femandez-Arias, and Serv6n, 1999. TABLE 3.11 UGANDA: TOTAL DISBURSEMENTS ON LOANS BY CREDITOR, 1980-96 (millions of dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 11.2 3.0 7.3 5.8 6.8 8.3 24.9 89.4 30.7 38.7 41.1 8.9 45.3 89.5 18.0 24.4 46.0 Multilateral 84.5 204.3 165.7 193.7 99.6 145.2 110.6 228.0 176.6 282.4 302.4 250.0 260.4 301.9 305.6 263.8 242.5 Total 95.8 207.3 173.1 199.4 106.4 153.5 135.5 317.4 207.3 321.1 343.5 258.9 305.7 391.4 323.7 288.2 288.5 Source: World Bank. TABLE 3.12 TOTAL GRANT EQUIVALENTS OF LOAN DISBURSEMENTS BY CREDITOR, 1980-96 (millions of U.S. dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 6.3 2.1 1.5 1.5 2.2 0.7 1.3 1.2 3.5 43.3 9.0 7.1 8.5 26.2 7.6 8.0 21.1 Multilateral 28.7 17.3 46.0 42.3 56.2 86.5 66.2 122.7 119.7 152.6 210.9 195.7 200.7 132.3 223.7 186.2 111.1 Total 35.0 19.4 47.5 43.9 58.4 87.2 67.4 124.0 123.1 195.8 219.9 202.8 209.1 158.5 231.3 194.2 132.2 Sources: World Bank; Chang, Femandez-Arias, and Serv6n 1999. TABLE 3.13 UGANDA: TOTAL NON-TECHNICAL ASSISTANCE GRANTS BY CREDITOR, 1980-96 (millions of dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 31.6 77.4 40.7 54.3 34.4 24.6 56.7 53.8 134.9 124.3 204.5 268.1 169.7 213.2 231.2 249.5 257.3 Multilateral 30.0 23.3 14.8 22.3 32.1 22.8 24.0 42.6 52.2 52.2 53.0 46.1 166.5 46.5 87.8 149.6 81.2 Total 61.6 100.7 56.5 76.6 66.5 47.4 81.6 96.4 187.1 176.5 257.5 314.2 336.2 259.7 319.0 399.1 338.5 Source:OECD. TABLE 3.14 UGANDA: TOTAL TECHNICAL ASSISTANCE GRANTS BY CREDITOR, 1980-96 (millions of dollars) Source 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Bilateral 11.1 13.6 13.5 15.0 12.8 15.4 23.7 28.0 38.4 43.5 53.6 63.8 76.0 78.9 88.5 116.7 100.7 Multilateral 10.0 20.0 16.0 18.9 19.6 21.5 26.3 29.4 37.3 44.2 40.2 43.4 45.8 45.3 36.2 31.5 39.4 Total 21.1 33.6 29.5 33.9 32.4 36.9 50.0 57.4 75.7 87.7 93.8 107.2 121.8 124.2 124.7 148.2 140.1 Source: OECD. TABLE 3.15 UGANDA: DISBURSEMENTS ON OFFICIAL LOANS BY SECTOR, 1980-96 (millions of U.S. dollars) Sector 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Agriculture, forestry, fishing 0.6 0.2 1.5 14.5 7.5 19.2 24.2 35.9 19.2 25.4 15.9 20.0 14.5 17.8 20.0 17.1 28.5 Balance of payments support 0.0 48.4 36.8 26.4 35.4 49.1 16.8 34.7 41.3 157.8 125.3 75.2 79.8 60.1 57.0 2.7 2.9 IMF disbursements 77.9 144.6 93.8 113.8 21.5 0.0 0.0 58.8 73.5 51.4 81.2 78.4 56.1 0.0 52.5 55.9 63.2 Contributions to finance current imports 0.0 0.0 1.3 0.0 2.0 0.0 0.0 0.0 0.0 0.0 1.1 4.1 12.7 40.0 5.8 2.1 22.9 Community, social, personal, and environ- mental services 0.8 3.6 11.8 8.5 9.4 15.6 44.8 11.3 7.5 14.5 38.3 20.2 69.6 132.3 51.2 117.7 48.1 Manufacturing 2.8 3.4 13.0 15.9 11.5 27.6 21.6 21.7 15.1 20.3 15.6 16.0 10.5 16.2 12.1 12.6 11.3 Electricity, gas, and water production 0.9 0.0 1.6 6.6 1.5 8.9 18.0 24.9 11.5 10.8 12.1 15.7 18.4 32.6 33.8 30.6 32.8 Transport and storage 12.7 3.8 2.4 0.0 1.7 5.7 3.1 16.5 24.0 32.5 29.5 21.4 22.5 13.5 12.5 9.2 32.0 Other 0.1 1.5 10.9 13.7 15.9 27.4 7.0 113.6 15.2 8.4 24.5 7.9 21.6 78.9 66.3 40.3 46.8 Total 95.8 207.3 173.1 199.4 106.4 153.5 135.5 317.4 207.3 321.1 343.5 258.9 305.7 391.4 323.7 288.2 288.5 Source: World Bank. TABLE 3.16 UGANDA: DEBT TABLES, 19895 (millions of dollars) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 Debt Restructurings Total amount of debt rescheduled (current US$) 0.0 .. .. .. .. 0.0 0.0 60.7 22.8 60.2 .17.8 8.8 99.6 37.8 0.0 172.3 Debt stock rescheduled (current US$) 0.0 .. .. .. .. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 143.1 Principal rescheduled (current US$) 0.0 .. .. .. .. 0.0 0.0 47.7 13.3 18.2 3.5 3.8 54.9 24.2 0.0 0.0 Official 0.0 .. .. .. .. 0.0 0.0 11.4 7.3 10.1 3.5 3.6 30.6 10.8 0.0 0.0 Private 0.0 .. .. .. .. 0.0 0.0 36.3 6.0 8.1 0.1 0.2 24.3 13.4 0.0 0.0 Interest rescheduled (current US$) 0.0 .. .. .. .. 0.0 0.0 11.1 9.1 10.5 1.0 2.7 37.9 10.1 0.0 30.4 Official 0 .. .. .. .. 0 0 5.5 8.3 8 1 2.7 24.2 9.4 0 30.4 Private 0 .. .. .. .. 0 0 5.5 0.8 2.6 0 0 13.7 0.7 0 0 Principal forgiven (current US$) .. .. .. .. .. 5.7 0.0 16.6 40.8 0.2 51.2 0.5 13.7 16.0 5.9 0.0 Memo: Interest forgiven (current US$) .. .. .. .. .. 0.8 0 0.2 0 0.1 0.6 0.5 1.5 18.8 0 0.7 Debt stock reduction (current US$) 0.0 .. .. .. .. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 138.9 0.0 41.5 of which debt buyback (current US$) 0.0 .. .. .. .. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 17.3 0.0 0.0 Debt Stock-Flow Reconciliation Total change in debt stocks (DOD, current US$) .. .. .. .. .. .. .. .. .. 253.8 406.9 194.3 150.3 101.3 342.8 201.5 Net flows on debt total (NFL, current US$) 115.8 176.5 177.6 151.4 63.5 74.7 81.3 337.8 91.0 231.7 277.2 167.2 236.6 243.5 224.1 185.6 Net change in interest arrears (current US$) .. .. .. .. .. .. .. .. .. 17.0 26.3 18.5 -20.3 -9.6 9.5 -22.1 Interest rescheduled (capitalized) (current US$) .. .. .. .. .. .. .. .. .. 10.6 1.1 2.7 37.9 10.1 0.0 30.4 Debt forgiveness or reduction (current US$) .. .. .. .. .. .. .. .. .. -0.2 -51.2 -0.5 -13.7-137.6 -6.9 -41.5 Cross-currency valuation (current US$) .. .. .. .. .. .. .. .. .. -20.2 54.2 -12.6 -43.4 -10.1 34.3 6.3 Residual .. .. .. .. .. .. .. .. .. 14.9 86.3 19.0 -48.5 5.0 81.8 42.8 Debt Ratios External debt (% of exports of goods and services) 209.4 257.3 228.3 265.6 266.3 301.9 383.5 473.1 580.6 716.2 1,050.0 1,395.6 1,501.3 1,172.6 1,0125 535.7 External debt (% of GNP) 55.6 71.4 53.9 60.4 30.3 35.5 36.4 30.8 29.8 42.0 61.1 85.1 105.6 96.5 85.6 62.8 Total debt services (% of exports of goods and services) 17.4 25.0 15.6 22.7 32.0 32.0 44.1 39.4 62.2 61.3 59.5 74.4 58.5 65.8 45.2 20.5 Source: World Bank, Global Development Finance 1999. TABLE 3.17 UGANDA: MULTILATERAL DEBT AGREEMENTS WITH OFFICIAL CREDITORS, 1980-97 Consolidation period Consolidation includes Share Amount Repayment termsa for current maturities Previously of debt consolidated Maturity Contract Length rescheduled consolidated (millions of (years! Grace Date of agreement cutoffdate Start date (months) Arrears debt (percent) U.S. dollars) months) (years/months) 18 Nov. 81 1 July 81 1 July 81 12 Y 90 63 9/0 4/6 1 Dec. 82 1 July 81 1 July 82 12 90 16 9/0 4/6 19 June 87 1 July 81 1 July 87 12 Y Y 100 102 14/6 6/0 26 Jan. 89 1 July 81 1 Jan. 89 18 Y Y 100 86 Menu Menu 17 June 92 1 July 81 1 July 92 17 Y Y 100 172 Menu Menu 20 Feb. 95 1 July 81 1 Feb. 95 Stock Y Y 100 110 Menu Menu a. Maturity is measured here from the end of the consolidation period to the date of the final amortization payment; the grace period is the time between the end of the consolidation period and the date of the first amortization payment. The secretariat of the Paris Club measures grace and maturity from the midpoint of the consolidation period. "Menu" terms refer to the options agreed to at the 1988 Toronto economic summit meeting. Source: World Bank, Global Development Finance 1997 (table A3.2). Data compiled from World Bank Debtor Reporting System and International Monetary Fund data. Uganda 159 TABLE 3.18 UGANDA: DEBT RELIEF, 1991-97 (millions of U.S. dollars) Type of creditor 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 Multilateral (EEC) 3.00 2.00 Bilateral 12.90 12.93 36.35 41.20 50.20 53.31 Multilateral Debt Fund 37.84 48.33 Total 12.90 12.93 39.35 41.20 90.04 101.64 Source: Bank of Uganda. 160 Successful Reformers APPENDIX 3.2 INTERVIEWS IN UGANDA, MARCH 1999 Government of Uganda (Ministry of Finance, Planning and Economic Development) Emmanuel Tumusiime-Mutebile, Permanent Secretary, Secretary to the Treasury. Damoni Kitabire, Director of Economic Affairs. Keith J. Muhakanizi, Economic Adviser to the Minister. Oode Obella, Agricultural Commissioner, Aid Liaison Department. Ocailap Patrick, Assistant Commissioner, Aid Liaison Department. Makerere University Patrick Asea, Professor, Economic Policy Research Centre (EPRC). Fred Opio, Professor, Economic and Policy Research Centre (EPRC). Germina Ssemogerere, MA, Economics and Planning. Private Sector Dr. William Kalema, Chairman, Uganda Manufacturers Association (UMA). Emmanuel Buringuriza, Assistant Manager, Business Uganda Develop- ment Scheme, Private Sector Foundation Project. Yonasani B. Kanomozi, Director, Community Management Services. Nongovernmental Organizations Zie Gariyo, Coordinator, Uganda Debt Network (UDN). Kevin Akoyi Makokha, Country Representative, Coopibo-Uganda. Bruce Mazzie, Chief of Party, Private Enterprise Support, Training, and Organizational Development Project (PRESTO). Sarah Kitakule, Senior Adviser, Policy Regulation and Reform, PRESTO. Carolyn Elliott-Farino, Project Administrator, PRESTO. IMF, Kampala Zia Ebrahim-Zadeh, Resident Representative. World Bank, Kampala Randolph Harris, Resident Representative. Robert Blake, Senior Resident Economist. Bilateral Donors Jens Rasmussen, Deputy Counselor, Royal Danish Embassy. Martin Koper, First Secretary, Embassy of the Netherlands. Uganda 161 NOTES 1. EDA is defined in this paper as the sum of grant equivalents and grants, including technical assistance grants. Debt forgiveness is excluded. ODF is the sum of EDA and the nongrant component of official loans. For further informa- tion on the definitions see Chang, Fernandez-Arias, and Serv6n (1999). 2. Appendix 3.1 contains information on locally produced aid data. 3. Financial aid normally includes grants and concessional loans having a grant element of at least 25 percent. In this discussion of loans and grants we use the ODF concept, which means that "loans" includes gross disbursement of official loans, and "grants" includes non-technical assistance grants as well as technical assistance grants. The aim of the study is not to analyze aid flows per se but rather to examine the relationship between donor behavior, as manifested in aid flows, and recipient country reforms. 4. The Finance and Planning ministries have been separated and merged during the period of the study as follows. 1980-92: Ministry of Finance, Minis- try of Planning and Economic Development; 1992-96: Ministry of Finance and Economic Planning; 1996-98: Ministry of Finance, Ministry of Planning and Economic Development; 1998 onwards: Ministry of Finance, Planning and Eco- nomic Development. The abbreviation MFPED is used throughout the text when discussing the current joint ministry. 5. Nonfinancial aid includes technical assistance grants to nationals of aid- recipient countries who are receiving education or training, and payments for consultants, advisers, administrators, and the like, serving in aid-recipient coun- tries. The statistics in the study are based on this definition of stand-alone tech- nical assistance grants. In the analysis of the aid-reform relationship we also take into account loan-based technical cooperation, technical assistance compo- nents in project loans, advisory services, policy dialogue, and similar nonmon- etary measures. 6. Eligibility under the HIPC Initiative is based on the net present value of debt/exports and the debt-service ratio at the "completion point" and beyond in the range of 200-250 percent and 20-25 percent respectively. 7. The government unified the dual exchange regime in 1984. See Kasekende and Ssemogerere (1991) for a full discussion of the exchange rate regimes in 1981-85. 8. Related to expenditure controls, revenue collection, and effective con- duct of monetary policy. 9. The Washington consensus refers to a set of internationally agreed mea- sures that are normally implemented during policy reform, including fiscal dis- cipline, financial and trade liberalization, deregulation, taxation and public expenditure adjustments, and privatization. 10. The CPIA, which measures the extent to which government policy cre- ates a good environment for broad-based growth and poverty reduction, has four components: macroeconomic policies, structural policies, public sector man- 162 Successful Reformers agement, and equity. The equity component measures the extent to which policy ensures the full participation of the society through social services that reach the poor and safety nets targeted to the disadvantaged. 11. Collier and Dollar (1998) indicate that Uganda should have received 3.68 percent aid/GDP in a poverty-efficient allocation in 1996, while it actually re- ceived 3.34 percent. REFERENCES Ablo, E., and R. Reinikka. 1998. "Do Budgets Really Matter? Evidence from Public Spending on Education and Health in Uganda." Policy Research Work- ing Paper 1926. World Bank, Development Research Group, Washington, D.C. Atingi-Ego, M., and K. Matthews. 1996. "Demand for Narrow and Broad Money in Uganda." African Review of Money, Banking and Finance. Supplementary Is- sue on Savings and development 1 (2). Bigsten, A. and S. Kayizzi-Mugerwa. 1992. "On Structural Adjustment in Uganda." Canadian Journal of Development Studies 13 (1): 57-75. Brett, E.A. 1995. "Adjustment policy and institutional reform: rebuilding orga- nizational capacity in Uganda." In P. Langseth, J. Katobor, E. Brett, and S. Munene, eds., Uganda Landmarks in Rebuilding a Nation. Kampala: Fountain Publishers. Chang, C., E. Femandez-Arias, and L. Serven. 1999. "Measuring Aid Flows: A New Approach." Policy Research Working Paper 2050. World Bank, Devel- opment Research Group, Washington, D.C. Collier, P., and D. Dollar. 1999. "Aid Allocation and Poverty Reduction." Policy Research Working Paper 2041.World Bank, Development Research Group, Washington, D.C. Fry, M. 1995. "Savings, Growth and Financial Repression." Paper presented at the Money, Macroeconomic and Finance Research Group Meeting at the Lon- don Business School. May. Harvey, C., and M. Robinson. 1994. "The Design of Economic Reforms in the Context of Political Liberalization: Uganda Case Study." Institute of Devel- opment Studies, University of Sussex, U.K. Processed. IDA (Intemational Development Association). 1997. "Uganda: Preliminary Docu- ment on the Initiative for Heavily Indebted Poor Countries." Washington, D.C. IMF (Intemational Monetary Fund). 1995. "Uganda: Adjustment with Growth, 1987-94." Washington, D.C. - 1998. "Extemal Evaluation of the ESAF." Washington, D.C. Kasekende, L., and M. Atingi-Ego. 1996. "Financial Liberalisation and Its Impli- cations for the Domestic Financial System: The Case of Uganda." Final re- search paper submitted to the African Economic Research Consortium workshop, Nairobi. May 24-30. Uganda 163 .1999. "Enhancing and Sustaining Growth in Africa: Uganda's Macroeco- nomic Management." EAGER/PSGE study (Equity and Growth through Eco- nomic Research/Public Strategies for Growth with Equity). Kasekende, L., and M. Malik. 1994. "The Financial System, Saving and Invest- ment in Uganda." Paper presented to the U.N. Economic Commission for Africa, Addis Ababa. Kasekende, L., and G. Ssemogerere. 1994. "Exchange Rate Unification and Eco- nomic Development: The Case of Uganda 1987-1992." World Development (22) 1183-98. Lamont, T. 1995. "Economic Planning and Policy Formulation in Uganda." In P. Langseth, J. Katobor, E. Brett, and J. Munene, eds., Uganda Landmarks in Re- building a Nation. Kampala: Fountain Publishers. McKinnon, R. 1973. "Money and Capital in Economic Development." Brookings Institution, Washington, D.C. Shaw, Edward. 1973. Financial Deepening in Economic Development. New York: Oxford University Press. Tumusiime-Mutebile, E. 1995. "Management of the Reform Programme." In P. Langseth, J. Katobor, E. Brett, and J. Munene, eds., Uganda Landmarks in Re- building a Nation. Kampala: Fountain Publishers. Uganda. 1991. "External Debt Management Report." Prepared by S.G. Warburg Co. Ltd. . 1997. "Poverty Eradication Action Plan." Ministry of Planning and Eco- nomic Development, Kampala. . 1998. "Policy Framework Paper 1998/99 to 2000/01." Kampala. UNDP (United Nations Development Programme). 1998. Uganda Human Devel- opment Report. World Bank. 1997. "Implementation Completion Report. Uganda: Structural Adjustment Credits I and II (SAC1 and SAC2)." Africa Regional Office. . 1998a. "The World Bank's Experience with Post-Conflict Reconstruc- tion. Volume IV: Uganda Case Study." Report No. 17769. Washington, D.C. . 1998b. Assessing Aid: What Works, What Doesn't, and Why. New York: Oxford University Press. .1999. "Preliminary Performance Audit Report. Uganda: First Structural Adjustment Credit and Second Structural Adjustment Credit." Operations Evaluation Department, Washington, D.C. Post-Socialist Reformers Ethiopia Berhanu Abegaz The College of William and Mary Williamsburg, Virginia ACRONYMS AND ABBREVIATIONS AfDB African Development Bank CBE Commercial Bank of Ethiopia CPIA Country Policy and Institutional Assessment CRS Creditor Reporting System (of the World Bank) DAC Development Assistance Committee (of the OECD) DRS Debtor Reporting System (of the World Bank) EDA Effective development assistance EDT Total external debt, including short-term and use of IMF credit EPRDF Ethiopian People's Revolutionary Democratic Front ESAF Enhanced Structural Adjustment Facility (of the IMF) ETB Ethiopian birr EU European Union FIG Financial and industrial group FDI Foreign direct investment GDP Gross domestic product GTZ German Agency for Technical Cooperation HIPC Heavily indebted poor countries IDA International Development Association (of the World Bank Group) IMF International Monetary Fund LDOD Long-term debt outstanding and disbursed MEDAC Ministry of Economic Development and Cooperation NBE National Bank of Ethiopia NFR Non-factor receipts NGO Nongovernmental organization ODA Official development assistance OECD Organisation for Economic Co-operation and Development PADEP Peasant Agricultural Development and Extension Program PFP Policy Framework Paper SAC Structural Adjustment Credit SAF Structural Adjustment Facility (of the IMF) SDOD Short-term debt outstanding and disbursed SDP Sector development program SIP Sector investment program SOE State-owned enterprise SPA Special Program of Assistance (for Sub-Saharan Africa) TGE Transitional Government of Ethiopia TPLF Tigray People's Liberation Front U.N. United Nations UNESCO United Nations Educational, Scientific, and Cultural Organization UNICEF United Nations Children's Fund UNDP United Nations Development Programme UNFPA United Nations Population Fund WFP World Food Programme WHO World Health Organization WPE Workers' Party of Ethiopia The primary concern of this paper is the relationship between devel- Topment aid and the process of economic reform in Ethiopia in 1980- 98. However, a few words on issues pertaining to aid effectiveness and selectivity provide a useful context. Development aid involves the transfer of resources from official or private institutions to low-income economies in the form of loans on concessional terms, technical assistance, and outright grants. The effec- tiveness of official development assistance has recently been the focus of comprehensive cross-national studies, most notably by the Task Force on Concessional Flows (Cassen 1994) and the World Bank (1998b). The Cassen report, relying primarily on country case studies, concluded that, first, the majority of official aid intended for long-term development has been successful; and second, aid has not necessarily gone where it is used most efficiently. For example, South Asia has lost its share of aid to Africa, where aid works relatively less well. The recent World Bank study, focusing on domestic factors, notes that (a) financial assistance has a big, positive impact where recipients have already undertaken sensible eco- nomic reforms; (b) financial aid flows do not discriminate between re- cipients with good policy and recipients with mediocre policy; (c) financial aid is generally fungible, thereby making the assessment of aid effectiveness difficult; (d) program finance enhances the prospects of success for sectoral and project aid; (e) nonfinancial aid is more effective than money in inducing nonreforming governments to seriously con- template removing distortionary policies; and (f) effective aid often comes as a package of finance and ideas that fosters the emergence of good institutions and sound policies. These findings, together with the decline of aid by one-third in the 1990s, have triggered a serious rethinking of the optimal composition and recipient-selectivity of aid programs. This trend is particularly of concern for Sub-Saharan Africa, where aid intensity is high and aid ef- fectiveness is low (van de Walle and Johnston, 1996; World Bank 1994, 1998b; O'Connell and Soludo 1998). The mixed record on the effective- ness of external economic assistance to Ethiopia is reviewed in Maxwell (1998) and IDS/IDR (1996) for European Union aid; in World Bank (1997a) for the International Development Association (IDA) loan port- folio; and in Clay, Molla, and Habtewold (1998) for food aid. External economic assistance has played an important role in the eco- nomic development programs of successive Ethiopian governments since 1960. The primary objectives of donors in Ethiopia have been the pro- motion of economic growth through support for investment and reform, alleviation of the unacceptably high rate of absolute poverty, and reduc- tion of the vulnerability of the economy to adverse natural and terms- of-trade shocks. Other donor motives, especially geostrategic ones, have 169 170 Post-Socialist Reformers also generated military assistance, which at times has dwarfed develop- ment aid and threatened to nullify the latter's positive impact. Ethiopia managed to attract more than US$17 billion in official devel- opment assistance in 1980-97. The annual inflow of some 16 percent of gross domestic product (GDP) was divided equally between multilat- eral donors, on the one hand, and bilateral donors as well as foreign nongovernmental organizations, on the other. The country today is heavily indebted to the tune of US$10 billion, almost equally split be- tween Russia and members of the Development Assistance Committee (DAC) of the Organisation for Economic Co-operation and Development (OECD). The main conclusions of this paper can be summarized as follows. Aid had a minimal impact on growth or poverty in the 1980s since Ethio- pia was embroiled in a protracted civil war along with a bold program of building a socialist economy. Market-oriented reforms were initiated in 1990 in response to profound economic and political crises. The re- form program initially attracted a sizeable increase in external assistance in support of liberalization, stabilization, and rehabilitation. Subsequent aid helped to deepen reform commitments and supported high growth rates. The residual dirigisme of the ruling party, however, appears to frus- trate the enactment of important second-generation reforms. The rest of the paper is organized as follows. Section 2 presents a brief country background, focusing on the economic profile of Ethiopia, while section 3 contains data on the level, source, and composition of aid. The political economy of institutional and policy reform under the Derg and Ethiopian People's Revolutionary Front govemments is discussed in section 4. Section 5 sets forth hypotheses concerning the relationship between aid and reform, and section 6 offers a summary and conclu- sions. A chronology of reforms undertaken during the period under re- view is included as an appendix. COUNTRY BACKGROUND Ethiopia has a number of peculiar characteristics that must be kept in mind when evaluating the record on the role of aid in the country's de- velopment strategy. They include marked demographic and ecological diversity, an extremely high rate of absolute poverty and economic inse- curity, a predominantly subsistence and coffee-based monocrop economy, and an old but resilient state plagued by autocratic rule and endemic social unrest. The country's 60 million people are highly diverse in ethnic, religious, and geographic terms. According to census figures (Ethiopia 1998b), there are more than 80 distinct ethnic groups. However, two groups of equal Ethiopia 171 size, the Oromo and the Amhara, together account for more than two- thirds of the population. About 60 percent of Ethiopians are Christian, a third are Muslims, and the remainder adhere to various indigenous reli- gions. In the urban areas there is distinct occupational specialization among the various ethnic groups. Ecological diversity is unusually high, with 18 agro-ecological zones, and there is marked contrast between the densely settled mixed farming regions of the temperate central high- lands and the transhumant lowlands. Ethiopia is arguably the poorest country of its size in the world by almost all measures of poverty. Despite its good agricultural endow- ments, it has yet to achieve sustained food security. A 1981-82 survey found that nearly 90 percent of the population was below the poverty line of US$2 per day. In 1995-96, 43 percent of farm households were food insecure. Famines, regional and nationwide, occur with distressing frequency, thereby converting much of the transitory poverty into chronic poverty (Ramachandran 1997; Clay, Molla, and Habtewold 1998). The worst rural poverty is found in the famine-prone northern highlands, especially in parts of Tigray and Wollo, and in the lowlands along the country's international borders, most notably, the Afar, Somali, Borana, Gedeo, and Omo areas. Interestingly, the rate of urban poverty may very well have approached the level of rural poverty in the 1990s. Middlebrook and Corzato (1999) document poverty in Addis Ababa. An assessment of the relationship between growth and poverty reduction is frustrated by the lack of reliable data on the historical trends of poverty rates. The country's social indicators reflect its low per capita income. Half the population does not have access to health facilities within 10 kilome- ters. Life expectancy at less than 50 years is below the average for Sub- Saharan Africa, as are school enrollment ratios. Only 40 percent of primary school-age children are enrolled in primary schools and just 15 percent of eligible students attend secondary schools. Ethiopian economic data are generally incomplete or inconsistently reported, especially for periods longer than five years. I have constructed a number of economic series for the period under review, piecing to- gether the more reliable birr-based series from disparate sources.! The basic macroeconomic indicators are reported in table 4.1. They are aver- aged over distinct subperiods paralleling notable reforms or shocks to the economy (see also appendix 4.2). Seven time segments are identified in order to capture major institutional changes or policy episodes: 1980- 83 (zemetcha years), 1984-85 (nationwide famine), 1986-89 (central plan- ning), 1990-91 (liberalization and political change), 1992-94 (recovery and stabilization), 1995-97 (structural reforms), and 1998-2001 (sector development programs). The growth rate of real GDP during the study period was modest but respectable. The predominantly agricultural economy (agriculture ac- TABLE 4.1 ETHIOPIA: BASIC ECONOMIC INDICATORS, PART A Indicator 1980-83 1984-85 1986-89 1990-91 1992-94 1995-97 1998-00a Annual average growth (%) (in 1990 birr) Real GDP 5.5 -5.3 5.3 3.0 7.6 7.8 2.5 Population 2.8 2.8 3.0 3.1 2.9 2.9 3.0 GDP per person 2.7 -8.1 2.3 -0.1 4.7 4.9 -0.5 Exports 8.8 28.0 3.1 -18.0 89.4 12.1 3.2 Imports 3.8 3.3 -1.2 -32.6 -16.9 12.6 5.4 Gross domestic investment 30.0 -40.2 0.0 -5.0 32.3 27.5 Money supply, Ml 12.2 17.0 10.7 19.6 13.0 4.4 10.0 GDP deflator 6.5 26.0 0.8 21.4 11.7 3.4 2.4 Consumer price indexb 3.8 29.1 4.2 35.7 5.7 -4.3 3.8 As percentage of GDP Gross domestic investment 12.5 13.8 16.7 11.5 12.9 18.0 20.3 Private consumption 78.8 78.6 75.4 76.5 84.8 81.7 80.0 Government consumption 15.5 16.3 16.4 17.9 11.6 12.0 11.0 Net exports and NFR -6.8 -8.8 -8.0 -6.5 -9.6 -11.8 -11.7 Gross domestic savings 5.7 5.0 8.7 5.4 3.3 6.2 8.5 TABLE 4.1 ETHIOPIA: BASIC ECONOMIC INDICATORS, PART B Levels 1980 1997 Real GDP (millions of 1990 birr) 13,485 28,361 Population (millions) 38 60 GDP per person (millions of 1990 birr) 358 473 Real exports (millions of 1990 birr) 1,912 4,238 Real imports (millions of 1990 birr) 10,555 20,033 Exchange rate (ETB/US$) 2.1 6.5 Quality of life indicators Poverty rate Below US$1/day (% of population) 46.0 44.0 (1995) Poverty gapc (% of GDP) 12.4 Below US$2/day (% of population) 89.0 Poverty gap (% of GDP) 42.7 Under-5 mortality rate (per thousand) 213.0 177.0 Gross enrollment ratio (ages 6-23, %) 16.0 18.0 - Not available. a. Projected based on constant factor cost in the 1998 Policy Framework Paper (Ethiopia 1998a). b. The consumer price index (CPI) is for Addis Ababa only, and excludes housing cost. A national CPI was introduced in 1995-96. For 1990-97, the growth rates of the three price indicators are as follows: GDP deflator, 8.9 percent; CPI, 6.9 percent; and food price index, 12.4 percent. The complete data for the GDP-based inflation rate are as follows: 1970-98, 7.8 percent; 1970-75, 4.8 percent; 1976-80, 15.0 percent; 1980-90, 5.2 percent; 1991-92, 21.0 percent; 1993-98, 3.6 percent. c. Size of income transfer needed to lift the poor above the poverty line, expressed as a fraction of GDP. Sources: IMF 1998b, 1999a; World Bank 1995, 1999b; UNDP 1998. 174 Post-Socialist Reformers counts for 55 percent of GDP) faced a number of shocks, including two waves of systemic change as well as fluctuations in commodity prices, civil strife, and droughts. During "normal" years, real per capita income grew at an annual rate of between 2.6 percent and 3.3 percent. The corre- sponding growth rate for GDP was 5.5 percent in the 1980s and 6.2 per- cent in the 1990s. These improvements, though hard to disentangle, are attributable to good weather, higher coffee prices, and allocational effi- ciency gains from the zemetcha (mass mobilization) campaigns, economic liberalization, and substantial inflows of development assistance. When the crisis years of 1984-85 and 1990-91 are included, however, real GDP grew at an annual rate of only 2.6 percent between 1980 and 1997-replicating the performance of the 1970s. Agricultural growth, which carries half the sectoral weight, averaged 0.6 percent per year in the 1980s and 3.0 percent in the 1990s. According to the Intemational Monetary Fund (IMF), the annual average growth rate of real GDP for 1970-98 was 2.86 percent. In light of the high rate of population growth, variously estimated at 2.5 to 3.0 percent per year, the average Ethiopian is at best as well off today as she was in 1970. On the other hand, the record on macroeconomic stability has been enviable. The inflation rate rarely exceeded single digits (the annual average was 7.8 percent for 1970-98), and the fiscal deficit was kept within manageable levels with some exceptions (see tables 4.1 and 4.2). AID: LEVEL, TYPES, AND SOURCES This case study uses two broad measures of official aid. The first, net official development assistance (ODA), is the sum of official grants (in- cluding technical assistance and food aid) and concessional loans with a grant element of at least 25 percent. It is calculated net of amortization payments. The second, effective development assistance (EDA), is the sum of discounted official grants and the grant equivalent of concessional loans, that is, the dollar value of the grant element. Dissatisfied with the shortcomings of net ODA, Chang, Femandez-Arias, and Serven (1999) constructed EDA as a superior measure of the real cost of aid to donors (though not necessarily to the recipient). Size of Aid Flows The data on the time pattern of resource flows to Ethiopia show some inconsistency even between the DRS (which appears to include portions of aid from the Council for Mutual Economic Assistance) and CRS data- bases.2 I rely primarily on CRS data. Alternative measures of aid will also be used as shown in table 4.3. Ethiopia 175 TABLE 4.2 ETHIOPIA: FISCAL TRENDS, 1980-99 Baseline: Revenue/expenditure 1980-85 1986-90 1991-95 1996-97 1998-99 Revenue and expenditure (% of GDP) Total revenue 15.4 24.6 17.8 21.5 20.2 Domestic 15.2 21.0 15.0 17.9 16.7 Extemal grants 0.2 3.6 2.8 3.6 3.6 Total expenditure 20.5 31.7 23.1 23.8 24.9 Recurrent - 22.1 14.8 13.5 13.9 Capital - 9.6 8.2 10.4 11.0 Fiscal deficit (cash basis) Before grants -5.3 -10.7 -8.1 -5.9 -7.8 After grants -5.1 -7.1 -5.3 -2.2 -4.3 Average GDP (millions of current birr) 10,841 15,790 29,523 43,335 45,204 Allocation of expenditure (% of total) General administration - 7.2 10.7 10.8 Defense - 30.8 10.6 7.9 18.0 Economic infrastructure - 5.6 10.7 16.5 of which: Roads - 1.9 5.0 10.3 - Economic development - 24.4 23.1 22.0 - of which: Agriculture - 7.9 7.4 6.6 - Social Services - 15.9 23.2 25.1 - of which: Education - 9.5 14.1 15 - of which: Health - 3.3 5.1 6.4 - Other - 16.1 21.6 17.9 - of which: Interest and charges - 4.8 10.5 8.6 - of which: External assistance - 6.1 2.8 0.0 - Total expenditure (millions of birr) - 4,937 6,808 10,471 11,917 - Not available. Sources: 1998-99 baseline data from World Bank 1998a, vol. 1, tables 1.2 and 3.1. All other data from IMF 1998b; World Bank 1999d. Ethiopia's ODA receipts exceeded US$17 billion in nominal terms (US$23 billion in real terms) during 1980-97. This comes out to US$1.0 billion (nominal) or US$1.3 billion (real) of annual inflow. In 1996 prices, the annual inflow of net ODA3 (including loans, grants, technical assis- tance, and food aid) to Ethiopia averaged US$1.2 billion per year in the 1980s. It rose to US$1.4 billion per year in 1991-96 before slowing subse- quently. In constant dollars, the pattern is M-shaped with peaks occur- ring in 1985 (famine effect) and 1992 (honeymoon effect). This pattern suggests that the ODA premium for the ambitious reform program of the 1990s has been a modest increase of 17 percent. TABLE 4.3 ETHIOPIA: OFFICIAL DEVELOPMENT ASSISTANCE, 198097 (millions of U.S. dollars) Net ODAl: Net Net loans + ODA ODA IDA IMF Technical Food Net Net Debt ODA grants' loansb grants credits credits FDIF assistance aid ODA2d ODA3' service deflator Year (CRS) (CRS) (CRS) (DRS) (DRS) (DRS) (CRS) (CRS) (DRS) (CRS) (DRS) (CRS) 1980 212 43 169 28 9 0 44 30 274 286 45 0.53 1981 245 46 199 28 73 0 64 54 1,251 363 58 0.51 1982 200 40 160 28 26 2 53 38 1,625 291 76 0.49 1983 339 126 213 42 0 -3 64 41 1,016 444 107 0.49 1984 361 75 286 41 0 5 81 91 929 533 130 0.47 1985 710 92 619 50 0 0 104 167 1,296 981 159 0.48 1986 629 67 563 38 41 -1 115 175 1,108 919 224 0.60 1987 627 134 493 86 0 -3 147 91 1,000 865 248 0.69 1988 963 188 775 75 0 2 197 157 1,263 1,317 309 0.75 1989 749 132 617 70 0 0 234 97 902 1,080 304 0.74 1990 1,020 161 858 74 0 12 254 195 1,201 1,469 236 0.83 1991 1,097 110 988 59 0 6 206 265 1,565 1,568 138 0.86 1992 1,182 106 1,076 112 20 6 190 222 1,484 1,594 109 0.91 1993 1,094 361 733 230 30 7 171 118 1,136 1,383 95 0.89 1994 1,076 311 764 150 20 3 126 48 958 1,250 112 0.93 1995 888 246 642 84 0 8 167 125 837 1,180 154 1.02 1996 850 227 623 142 21 5 202 70 657 1,122 347 1.00 1997 637 72 565 65 0 5 168 75 723 880 99 0.92 Note: CRS = Creditor Reporting System (OECD database). DRS = Debtor Reporting System (World Bank database). a. Net ODA loans + grants (= financial aid) b. Loan disbursements, net of amortization. c. Foreign direct investment (private). The FDI figures are gross underestimates. During 1992-98, the Ethiopian Investment Authority approved 163 projects worth US$1.4 billion (mainly Saudi and European). Of these, only 21 projects are operational and another 49 projects are under implementation (see UNCTAD/ ICC 1999, tables 5 and 6). d. Net flows on debt + grants + technical assistance + food aid. ODA2 and ODA3 are the broadest measures of aid given the sometimes dubious assumption that food aid releases dollars for investment that would otherwise have financed essential food imports. Food aid (93 percent of which comes in the form of grains) has recently been equivalent to a quarter of national grain sales (Clay, Molla, and Habtewold 1998). e. Net ODA loans + grants + technical assistance + food aid. Sources: DRS data from World Bank 1999c. CRS data from OECD 1999. 178 Post-Socialist Reformers In terms of aid intensity, the relevant figures are as follows. The ODA3- to-GNP ratio rose from 12 percent in the 1980s to 23 percent in the 1990s, yielding an average of 16 percent for 1980-97. In per capita terms, re- ceipts rose from US$15/$23 (nominal/real) in the 1980s to US$27/$30 in the 1990s, for a period average of US$20/$26 during 1980-97. The table shows other interesting patterns of aid flow. First, the share of grants in net cash disbursements rose substantially to exceed 50 per- cent after 1985. Second, debt service levels reached their peak, in abso- lute terms and relative to exports, in the second half of the 1980s. Third, Ethiopia relies heavily on ODA from four multilateral sources (the Inter- national Development Association, the European Union, the African Development Bank, and the United Nations) and a handful of European bilateral donors (mainly the Nordic countries, Germany, and Italy). Al- though it has managed to obtain a steady average annual flow of US$275 million in trade credit, Ethiopia remains distinctly unattractive to pri- vate investment from abroad. Foreign direct investment (FDI) has remained low, falling far below East African levels even in absolute terms (UNCTAD/ICC 1999). Data from the Ethiopian Investment Authority show significant FDI inflows in 1997 and 1998. The cumulative investment capital of the 200 approved projects for the period 1993-99 was Birr 8.5 billion (US$1.1 billion). How- ever, only 21 projects with an investment capital of Birr 3.2 billion (US$400 million) are currently in operation (IMF 1999b). Data on EDA (table 4.4) confirm much of the trend captured by the conventional measures. The "effective" grant equivalent (GEQ) of concessional ODA loans averages 30 percent. This proportion rose from about 21 percent in the 1980s to 46 percent in the 1990s, which compares unfavorably with an average grant element of 70 percent for 1991-97 (see figures 4.1 and 4.2). In real terms, the size of EDA loans was on average twice as high in the post-socialist period as in the socialist pe- riod. In terms of grant equivalency, bilateral loans declined to modest levels in the 1990s. Finally, the share of technical assistance in total aid was 20 percent-below the Sub-Saharan Africa average of 25 percent (see O'Connell and Soludo 1998, table 1). Sources and Uses of Aid Being a non-colony, Ethiopia did not until recently attract the level of development assistance that it needed and could absorb effectively. It was, nonetheless, the largest recipient of the modest U.S. economic as- sistance to Sub-Saharan Africa during 1961-76. The Commission of the European Communities under Lome was especially generous, provid- ing grants of US$2.5 billion (in terms of 1990 dollars) during 1976-94 (Maxwell 1998: 105). Other multilateral agencies (notably IDA and the TABLE 4.4 ETHIOPIA: EFFECTIVE DEVELOPMENT ASSISTANCE, 1980-96 (millions of U.S. dollars) GEQ: GEQ: Year TEDA EDAl EDA2 GEQ Grants TA multilateral bilateral TA/EDA2 1980 239 165 209 40 125 44 26 14 21 1981 444 326 390 191 135 64 33 157 16 1982 269 178 231 71 107 53 32 39 23 1983 406 302 365 152 150 64 50 102 18 1984 553 281 362 76 205 81 42 34 22 1985 923 651 756 137 514 104 65 72 13 1986 855 564 680 117 447 115 49 69 17 1987 742 504 651 158 346 147 95 63 23 1988 1,097 737 934 159 578 197 93 66 21 1989 848 517 751 134 383 234 83 51 31 1990 1,175 726 980 122 604 254 97 25 26 1991 1,347 876 1,082 95 781 206 81 8 19 1992 1,393 981 1,171 95 886 190 119 8 16 1993 1,098 809 980 247 562 171 246 0 17 1994 964 790 916 152 638 126 152 0 14 1995 898 606 773 131 475 167 125 0 22 1996 850 578 780 158 420 202 158 0 26 Notes: TEDA = Total effective development assistance (= GEQ + grants + TA+ food aid). EDAI = Effective development assistance (= GEQ + grants). EDA2 = Effective development assistance ( = EDAl + TA). GEQ = Grant equivalent of an ODA loan. TA = Technical assistance (= technical cooperation). Source: Chang, Fernandez-Arias, and Serven 1999. 180 Post-Socialist Reformers FIGURE 4.1 ETHIOPIA: AVERAGE GRACE PERIOD AND MATURITY OF ODA LOANS Number of years 45 40- 35- 30 3 Average maturity (years) I 25- /k 20 -/ \ 15- 10 Average grace period (years) 0 Source: OECD 1999. African Development Bank) and bilateral donors (especially Italy and Sweden) provided the bulk of the external assistance. The United Na- tions agencies, most notably the United Nations Development Programme (UNDP) and the World Food Programme (WFP), were also active. WFP alone provided US$600 in food assistance during 1990-93. Between 1980 and 1997, 60 percent of grants in cash or in kind came from Europe and another 20 percent from the U.N. agencies. Since the mid-1990s, there has been a gradual decline in commitments by the Eu- ropean Union (EU) and Italy. This trend has been countered by increased support from IDA, Germany, and the United States. Japan, the Nether- lands, and Canada also increased their aid in the 1990s. The Nordic coun- tries as a group present an interesting model of stable and predictable aid to Ethiopia (table 4.5). Nearly all the concessional loans come from two multilateral institu- tions: the International Development Association of the World Bank Group, and the African Development Fund of the African Development Bank (AfDB). The former made its first African loan to Ethiopia in 1950. The latter began its lending program to Ethiopia in 1975 (as did the Eu- ropean Community), and it has beefed up its commitments in the past five years (AfDB 1996). Ethiopia 181 FIGURE 4.2 ETHIOPIA: AVERAGE GRANT ELEMENT AND SHARE OF MULTILATERAL DEBT Percent 80 70 X 60 - Average grant element (%) 50 . 40- 30 \ 20 - 10 Share of multilateral debt in total external debt (%/O) 0 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 Source: OECD 1999. The CRS data reveal other interesting patterns concerning assistance to Ethiopia from the Development Assistance Committee of the OECD. The share of DAC bilateral aid in net ODAI fell from 60 percent in 1991- 93 to 45 percent in 1994-97. ODA accounts for 98 percent of official de- velopment finance from all sources. One-third of technical cooperation grants come from non-DAC sources. And there has been a noticeable decline of both net disbursements and commitments since 1995. The distribution of aid disbursement by sector provides useful in- formation on the priorities of donors, creditors, and the two Ethiopian governments. Table 4.6 presents data on bilateral commitments by purpose. For net loans and grants from all sources, the gap between commitments and actual disbursements is generally small, except dur- ing the second half of the 1980s when the surge in external support began on the heels of the 1984 famine. The table shows that bilateral ODA commitments are largely accounted for by expenditures on so- cial infrastructure, program aid, and production sectors. Program as- sistance was especially strong, accounting for nearly half of bilateral resources, during the unstable years of 1985-93. Production sectors such as industry and agriculture received much less, as is often the case in Sub-Saharan Africa. TABLE 4.5 ETHIOPIA: NET ODA LOANS AND GRANTS BY CREDITOR OR DONOR, 1970-97 (millions of U.S. dollars) GrantSb Loans All European United United Nordic Year All loans ADFa IDA grants Community Nations Germany Italy States countriesc Japan Netherlands 1970 14 0 3 26 0 4 4 3 7 6 0 0 1975 63 0 19 71 8 16 6 3 9 12 1 3 1980 43 3 28 169 32 48 10 8 20 33 1 4 1981 46 11 28 199 42 67 11 15 4 23 3 4 1982 40 7 26 160 27 58 11 4 3 21 1 3 1983 126 12 41 213 47 15 10 14 8 25 3 3 1984 75 5 39 286 58 64 29 12 21 30 4 9 1985 92 17 48 619 103 108 27 63 146 49 8 8 1986 67 15 35 563 85 90 24 134 94 54 7 8 1987 134 23 82 493 93 103 27 98 8 59 16 7 1988 188 17 72 775 134 156 38 158 69 86 14 23 1989 132 28 66 617 45 192 41 119 26 74 13 11 1990 161 43 69 858 102 203 46 131 53 97 11 25 1991 110 41 54 988 159 367 48 77 88 88 18 26 1992 106 50 106 1,076 292 239 204 41 68 87 11 29 1993 361 117 225 733 123 183 50 12 131 80 48 34 1994 311 61 156 764 182 76 89 41 122 62 44 33 1995 246 82 77 642 82 107 76 27 70 76 63 40 1996 227 79 128 623 53 117 81 40 56 72 50 60 1997 72d 24 50 565 42 142 58 32 60 75 37 35 a. African Development Fund (of the African Development Bank, AfDB) b. Two other less generous but still important DAC donors not listed here are Canada and the United Kingdom. c. Sweden, Norway, Denmark, and Finland. d. This figure is too low (taken as it appeared in the source). Souirce: OECD 1999. Ethiopia 183 TABLE 4.6 ETHIOPIA: BILATERAL ODA COMMITMENTS BY PURPOSE, 1973-97 (millions of U.S. dollars) Total Infra- Prod- commit- structure Infrastructure Program uction Year menta + program Total Social Economic Total Food sectors 1973 32 21 20 13 7 2 0 2 1974 42 10 10 4 5 0 0 21 1975 67 51 38 29 9 14 8 16 1976 17 12 5 5 0 7 7 5 1977 31 17 7 6 1 9 9 13 1978 53 26 18 3 15 8 8 10 1979 42 23 5 5 0 18 18 7 1980 48 29 6 6 0 23 23 12 1981 80 49 8 8 0 41 36 25 1982 69 56 31 31 0 25 25 8 1983 76 46 10 6 4 36 36 16 1984 202 114 41 8 32 73 73 52 1985 308 206 61 24 37 145 145 45 1986 231 162 22 17 5 140 140 26 1987 361 127 43 31 12 84 84 62 1988 486 336 164 64 101 172 135 78 1989 278 144 48 25 24 96 90 82 1990 353 197 38 29 9 159 157 37 1991 210 170 16 7 8 154 150 20 1992 430 266 106 95 11 160 155 33 1993 368 233 79 67 12 154 111 64 1994 306 169 119 74 45 50 50 48 1995 566 273 147 104 43 126 85 58 1996 393 221 156 142 13 65 46 98 1997 295 180 103 82 21 77 65 46 a. Reported total commitment does not always equal the sum of its components (e.g. 1988). b. Program aid includes food aid and balance of payments and budget support. Remain- der categories include action relating to debt, emergency assistance, multisectoral aid, and unspecified. Source: OECD 1999. External Debt: Stock and Service At the end of 1997, Ethiopia's external debt stock (EDT)-nearly all of which was public debt-was estimated at US$10.0 billion (see table 4.7). The distribution of debt by ownership is as follows. Some 62 percent is owed to bilateral lenders, mainly to Russia for military purchases dur- ing the Soviet era. Multilateral creditors accounted for over one-third of 184 Post-Socialist Reformers TABLE 4.7 ETHIOPIA: TOTAL DEBT STOCKS, 1980-97 (millions of U.S. dollars) Total IDA: Export Year EDT LDOD LDOD SDOD IMF Arrears' credit 1980 824 688 249 57 79 2 - 1981 1,842 1,638 276 59 145 2 - 1982 3,280 3,051 302 68 161 2 120 1983 3,845 3,649 342 63 134 2 177 1984 4,220 4,054 379 68 98 2 279 1985 5,206 5,057 437 77 71 1 297 1986 6,134 5,967 486 83 84 1 323 1987 7,364 7,191 601 98 76 55 389 1988 7,704 7,515 658 134 55 16 372 1989 7,842 7,700 718 112 30 69 323 1990 8,634 8,483 824 145 6 279 311 1991 9,119 8,843 883 276 0 1,088 292 1992 9,341 9,003 964 319 19 1,780 239 1993 9,703 9,287 1,187 368 49 2,105 235 1994 10,067 9,570 1,373 424 72 3,204 234 1995 10,309 9,775 1,470 461 74 4,067 239 1996 10,078 9,484 1,555 502 92 4,784 300 1997 10,079 9,427 1,532 565 87 5,298 237 - Not available. Note: The above include ruble-denominated defense loans from the former Soviet Union. The Ministry of Finance put the external debt as of June 1998 much lower at US$7.8 billion, of which US$4.1 billion was owed to Russia. EDT = Total debt stocks, including short-term debt and use of IMF credit. LDOD = Long-term debt outstanding and disbursed. SDOD = Short-term debt outstanding and disbursed. a. Arrears include principal and interest. Source: World Bank 1999c. the total. IDA, with a portfolio of $1.5 billion, owns half of the multilat- eral debt. Private creditors claim the remaining 5 percent of the total debt stock. The data also show that short-term debt accounts for less than 5 percent of the total debt stock. Until recently, Ethiopia has had a record of remarkable diligence in servicing its debt. Arrears of both principal and capitalized interest in- creased substantially after 1990, reaching a whopping 53 percent of EDT by 1998. The real debt stock per capita in the 1990s was about six times as high as aid flow per capita (US$180 as against US$30). In net present value terms, the servicing of Ethiopia's external debt burden is quite unsustainable (Ghani and Zang 1995). In current value terms, debt service payments averaged 27 percent of GNP in the 1980s Ethiopia 185 before falling to 23 percent in the 1990s. International reserves were, in fact, down to less than one month of imports in 1988-91. At the begin- ning of 1998, the debt inclusive of arrears to Russia represented 963 per- cent of exports (see table 4.8) or 159 percent of GNP. The forecasts in Policy Framework Paper 5 (PFP5) show that the debt-to-export ratio will fall to 265 percent in 2001 (Ethiopia 1998a). Ethiopia was granted debt relief on enhanced Toronto terms by the Paris Club in December 1992. The relief involved debt cancellation of US$101 million and debt rescheduling of US$271 million. Ethiopia also bought back some of its commercial debt (US$226 million) in 1995 at a rate of 8 cents per dollar, utilizing the World Bank's debt buyback facil- ity. The preliminary review of eligibility for the Heavily Indebted Poor Countries (HIPC) Debt Initiative has been completed for Ethiopia. TABLE 4.8 ETHIOPIA: DEBT STOCKS AND DEBT SERVICE INDICATORS, 1980-97 (percent) RES/MGS Year CL/EDT (months) RESITDS EDT/XGS TDS/XGS EDT/GNP 1980 68 4 585 140 8 16 1981 78 5 643 326 10 36 1982 84 3 367 610 14 60 1983 86 3 192 666 19 64 1984 85 1 84 674 21 74 1985 86 2 136 932 28 78 1986 86 3 148 891 33 88 1987 87 2 90 1,150 39 99 1988 86 1 48 1,189 48 101 1989 87 1 40 1,035 40 99 1990 87 1 23 1,276 35 127 1991 84 1 77 1,667 25 172 1992 84 3 247 2,037 24 169 1993 84 4 527 1,889 19 158 1994 84 6 526 1,788 20 209 1995 85 7 529 1,277 19 180 1996 86 5 211 1,224 42 169 1997 86 3 507 963 10 159 Note: Includes ruble-denominated defense loans from the former Soviet Union. CL/EDT = Concessional loans to total external debt. RES/MGS = International reserves to imports of goods and services. RES/TDS = International reserves to total debt service. EDT/XGS = Total debt service to exports of goods and services. EDT/GNP = Total external debt to gross national product. Source: World Bank 1999c. 186 Post-Socialist Reformers Finalization of the debt relief package is pending settlement of the Ethio- pia-Eritrea conflict (World Bank 1999d; IMF 1999b). Ethiopia also stands to benefit from the more generous Cologne sum- mit declaration on HIPC. The annual saving from debt relief, however, appears to be modest given the recent trends of resorting to arrears. As- suming a 70 percent reduction in debt service payments from debt re- lief, the money that would be available for additional social spending is about US$50 million to US$100 million (that is, given the wide range of estimated annual debt service payment of US$150 million to US$350 million in recent years). Ethiopia also continues to press for a major write- off of Russian debt.3 Timing and Stability of Aid Flows In a world of imperfect markets, the impact of aid on growth depends in part on its timing and degree of instability, that is, whether it is predict- able or not. In the case of Ethiopia, the critical policy focus is on the quantity (and quality) of investment. Foreign aid, which accounts for more than one-fifth of GNP, finances nearly half the capital expenditure in the federal budget. Furthermore, in the spirit of two-gap models, an important determinant of changes in Ethiopian capital-good imports is the time path of foreign exchange earnings from exports (Etherington and Yainshet 1988; Lensink and Morrissey 1999; World Bank 1998a). A glance at figures 4.3 and 4.4 shows some interesting patterns in this regard. First, the variation in the growth rate of net ODAI is much higher than that of exports (or GDP). Generally, multilateral aid is more stable than bilateral aid, and project aid is steadier than program aid. Second, the correlation between the growth rates of export revenues and ODA (grants and loans) is low and negative (Pearson r = -0.25). This means that the changes in the flow of aid do not have a systemati- cally countervailing relationship with fluctuations in export earnings. This can be attributed to a number of factors, including the changing share of nonproject aid and the bunching of aid receipts relative to com- mitments, which are driven largely by bureaucratic dynamics. It also means that ODA flows are not entirely geared to stabilizing the overall supply of foreign exchange. Third, the level of total receipts (inclusive of ODA, other official flows, and private flows) does compensate for the volatility of export earnings (r = -0.57) so that total foreign exchange flows are reasonably stable. As shown in figure 4.4, this produces the expected synchronization between total foreign exchange receipts and the import bill. Two inferences can be drawn from this cursory overview. Increases in the share of program aid, by stabilizing total foreign exchange inflows, would help to reduce growth-retarding uncertainty for investors (Lensink Ethiopia 187 FIGURE 4.3 ETHIOPIA: GROWTH RATES OF GDP, EXPORTS, AND NET ODA Annual growth rate (%) 120 100 Net ODAl growth 80 60 40 20 0 -20 / Export growth GDP growth - 4 0 . . . . . . . . . . . . . . . . . . . . . . . 9t° 9 9 9& E6 °o 9p' 9°' 0 , 9Ct 99 9 9b Note: Calculated in 1996 U.S. dollars. a. Net ODAI refers only to financial aid (grants and net loans) to make it comparable to export earnings. It averaged US$500 million (or US$700 million in 1996 prices) during 1970-97. Sources: IMF 1998 (for GDP); World Bank 1999c (for imports and exports); OECD 1999 (for ODA, receipts, and deflator). and Morrissey 1999; Collier and Gunning 1999). Furthermore, the rising share of nonproject aid may bring more flexibility, but at the expense of aid predictability since the tap may be more easily turned on and off by disgruntled donors. INSTITUTIONAL DEVELOPMENT AND POLICY REFORM During the period under consideration, Ethiopia underwent two waves of radical economic reform: one revolutionary and anti-market in orien- tation, and the other more modest and mixed-economy in orientation. The former reflects the pervasive influence of the left-wing student move- ment, and the latter a reluctant retreat from socialist economic ideas by the products of the same movement. The chronology of institutional and policy reforms introduced by the two regimes is shown in appendix 4.2. 188 Post-Socialist Reformers FIGURE 4.4 ETHIOPIA: SOURCES AND USES OF FOREIGN EXCHANGE Millions of U.S. dollars 3,000 2,500 - RMGS*/ FOREX 2,000v , 1,500 - 1,000 500 , ,, '. ' .' '*. '"RRECEIP RRECEIP = All foreign exchange receipts from ODA and non-ODA sources, excluding exports. FOREX = Total foreign exchange inflows, including export earnings. RXGS = Exports of goods and services. RMGS Imports of goods and services. Sources: IMF 1998 (for GDP); World Bank 1999c (for imports and exports); OECD 1999 (for ODA, receipts, and deflator). Institutional Transformation under the Derg The period between 1974 and 1990 was one of revolutionary turmoil following the coup d'etat against the monarchy, civil wars, a war with Somalia, and an experiment with agrarian socialism. By the end of the 1970s, the broad outline of a quasi-socialist economy was in place: all land, extra houses, and large as well as medium private enterprises were nationalized without compensation. Owner-cultivators and sharecrop- pers were given use rights (usufruct) to land in lieu of their customary freehold or communal ownership rights. The rural population was then organized into peasant associations. These state-sponsored associations were soon to serve as stepping stones for the establishment of producer cooperatives or quasi-collective farms. Much of inter-regional whole- sale and retail trade in key staples and important industrial goods was reserved for parastatals. Ethiopia 189 The military regime's vision of state-led industrialization included an over-zealous "villagization" campaign, to facilitate political control as well as the delivery of social services; an ambitious but poorly planned program of "resettlement" of the drought-vulnerable population; and a program of import substitution for key industrial goods. By mid-1988 villagization had grouped 12 million people-one-third of the rural popu- lation-in more than 12,000 large villages (Clapham 1988: 251). Though briefly distracted by the devastating 1984-85 national fam- ine, when GDP contracted by 13.2 percent, the Derg consolidated its power in the second half of the decade. It formulated a Ten-Year Per- spective Economic Plan (1985-94) with an unrealistic expectation of re- ceiving foreign aid to finance more than half of the planned investment (World Bank 1987; Griffin 1992). The plan nevertheless provided a more consistent framework for medium-term and annual plans than the pre- vious disparate-projects approach. The economy rebounded briefly but spectacularly at the rate of 9.3 percent and 13.8 percent in 1986 and 1987 respectively. To its credit, the government also expanded elementary and secondary education and rural health clinics at unprecedented rates. The ambitious literacy campaign was awarded a UNESCO prize. However, two-thirds of the adult population remains illiterate today The military regime also belatedly established a socialist party, the Workers' Party of Ethiopia (WPE), with military officers and the upper echelon of the civil service at its core. Ethiopia finally joined the "people's democracies" of the Soviet bloc in 1987 with the declaration of the People's Democratic Republic of Ethiopia under a constitution ratified by national referendum. The Derg intensified the administrative centralization program of the monarchy under a unitary state, and implemented an unprecedented degree of state ownership and control of the modern sector of the economy. The country's 15 regions were reorganized into 30 administra- tive units. Imitating the nomenklatura systems of the socialist states, the regime established parallel networks of government and party organs down to the level of the kebelle (neighborhood), all controlled by WPE functionaries. The Derg also introduced a still enduring cadre system of party control of the economic, military, and administrative sectors of the state bureaucracy. A number of developments accentuated the "normal" level of crisis in the waning days of war socialism, which eventually led to policy reforms. First, the burden of financing a huge army was too high for the feeble domestic economic base to support, especially since the Soviet Union was itself beset by a profound crisis after 1988. Second, the idea of so- cialism was discredited internationally and domestically. Third, the eco- nomic consequences of the campaigns of villagization, resettlement, and producer cooperatives proved disastrous for an already impoverished 190 Post-Socialist Reformers population. The economy cumulatively shrank by 7.0 percent during and immediately after the height of the civil war, in 1988-92. While most captive civil servants could do little to arrest the double squeeze on sala- ries of pay freezes and double-digit inflation, small producers increas- ingly resorted to the age-old but inefficient methods of survival-informal exchange and subsistence production. The IDA maintained a modest portfolio of loans (US$35 million an- nually) during 1976-86. It confined its funding to economic and sector work (technical assistance and country economic memoranda), and fi- nancing of selected social services, transportation and communication, and the important coffee sector. Loan disbursement rose to US$52 mil- lion per year during 1989-90 with an agreement to support an agricul- tural extension and credit program called the Peasant Agricultural Development and Extension Program (PADEP). As shown in table 4.6, the government managed to obtain a respectable level of grant money in the 1980s (on a per year basis):US$70 million from the EC-EU, US$68 million from Italy, US$135 million from the Nordic countries, US$25 million from Germany, and US$41 million from the United States, mostly food aid. Data on economic aid from the Council for Mutual Economic Assistance are sketchy but it was rather modest (US$30 million in 1983) and came primarily in the form of project aid (Clapham 1988).4 The Derg seriously considered liberalization after 1988 although the implementing legislation was not in place until the WPE Congress of March 1990, a year before the regime's demise (Haile Mariam 1990). The impetus appears to have come from the pressure created by the steady gains of the northern rebel movements as well as the disappointing eco- nomic results of agricultural cooperatives and parastatals in the distribu- tive sector. The reforms were in the nature of "perfection of control" rather than significant concessions to the market. Their significance lies in the dynamic they created for further reform. The most notable measures were the following. Inheritable use rights were granted on nationalized land, membership in producer coopera- tives became voluntary (which led to their spontaneous dismantling by jubilant cooperators), many prices were decontrolled, and the dra- conian restrictions on private trade were grudgingly lifted. The liber- alization of the grain market, the most important market in Ethiopia, paved the way for a steady although still incomplete integration of inter-regional markets. Most foreign owners of nationalized enterprises did, under pressure from the international financial institutions and bilateral donors, even- tually receive compensation. Most citizens whose property was confis- cated are still awaiting restitution. As we will see below, these issues keep coming up in the policy dialogue with multilateral donors and in the lingering complaints of the business community. Ethiopia 191 Institutional Reform under the EPRDF The year 1991 was a watershed for reform: a coalition led by two libera- tion fronts ended the civil war victoriously. The province of Eritrea be- came an independent state in 1993.5 A transitional government of Ethiopia (TGE) was formed in Addis Ababa by a loosely affiliated assortment of liberation fronts known as the Ethiopian People's Revolutionary Demo- cratic Front (EPRDF). The Tigray People's Liberation Front (TPLF), which serves as the vanguard of the coalition, accounts for most of the senior government leadership. This was a rather unusual conclusion for a civil war. Groups at the periphery in effect gained exclusive control of the state machinery. The country is still contending with all the contradictions entailed in such a dispensation for both victor and loser (Bevan 1994; Chole 1993). The large national army of 500,000 was disbanded and replaced by EPRDF's own army (Dercon and Ayalew 1998). The civil service was trimmed by about a quarter; many were fired for political reasons. Despite a secular decline in real wages and compression of the salary structure, the 300,000- strong civil service accounts for one-third of formal-sector employment in Ethiopia (IMF 1999b). The TGE proceeded to decree a number of important institutional reforms in 1991-94, before the national election. The administrative struc- ture of the country was revamped once again, this time into nine federal units (killils) and special administrations in the two richest cities (Addis Ababa and Dire Dawa). The regional states were further divided into 66 zones comprising 556 woredas or units of local government with about 100,000 people each. The 1994 Ethiopian Constitution enshrined an ethnic-based federal- ism. After the 1995 "administratively fair, yet uncompetitive" elections (World Bank 1998c), a bicameral parliament dominated by the EPRDF assumed power under the newly created Federal Democratic Republic of Ethiopia. The federal constitution is Soviet-like in granting to any eth- nic group the right to self-determination "including and up to seces- sion." While the idea of a decentralized governmental structure finds broad support, the wisdom of gerrymandering districts to create ethnic fiefdoms remains a highly contentious issue in Ethiopian politics. Ethiopia today employs an unusual form of fiscal federalism. The large vertical imbalance between the revenue-raising power of the federal government (which collects some 80 percent of revenues) and the re- sponsibilities of the regions is bridged by unrestricted block grants. These subsidies are based on a formula that purports to reflect a region's popu- lation, development level, and local revenue generation. The advantages of responsiveness to local needs which would normally result from de- centralization are, therefore, diminished by the lack of performance-based 192 Post-Socialist Reformers reciprocity, inter-regional gaps in technocratic capacity, and the relative neglect of the special needs of municipalities (World Bank 1999a). Policy Reform under the EPRDF A few months after assuming state power, the EPRDF issued its first economic program (Ethiopia 1991). The program proposed to deepen the price decontrol and domestic trade liberalization introduced by the previous government. It advocated continuation of state ownership and control of industry, finance, modern services, and land. The leadership's vision of the post-socialist economy, though amended after 1993, remains one of a "state-led market economy" with mixed ownership and active regulation to stem market failures. The first Policy Framework Paper (PFP), covering 1992-93, marked the formal start of an extensive stabilization and structural adjustment program for poverty reduction and transition to a market economy. Key elements of the various rounds of the first-stage reforms included (see appendix 4.2): * Prudent fiscal policies (mainly through retrenchment of the army and civil servants, tax reform, and lower defense spending) and monetary policies (limiting monetization of the fiscal deficit) * A substantial correction of the overvalued nominal exchange rate * Decontrol of many prices * Liberalization of the foreign trade and foreign exchange regime * Autonomy of state-owned enterprises (SOEs)and privatization of small and medium enterprises * Financial market reform, including the licensing of six local pri- vate banks and seven insurance companies so far, and interest-rate (except for deposit rate) liberalization * A liberal investment code with an investment authority that is or- ganized as a "one-stop shop." The last two PFPs have paid greater attention to agriculture and rural development, social sectors, infrastructure, energy, export and private sector development, and the financial sector. A civil service reform pro- gram is also underway. It encompasses the areas of financial manage- ment, judicial reform, human resource management, top management systems, human resource management, and service delivery. The government adopted a three-year adjustment program in mid- 1996 under the IMF's Enhanced Structural Adjustment Facility (ESAF). In addition to the stabilization program, which was also supported by a Structural Adjustment Credit (SAC I) loan from IDA, it introduced struc- tural reform measures aimed at liberalizing the economy. All retail prices, Ethiopia 193 except petroleum, have been decontrolled. Export taxes were eliminated except on coffee, and in that case license fees for exporters were reduced. Measures to revive the private sector included allowing domestic pri- vate participation in freight forwarding and clearing, and in banking and insurance activities. The parastatal monopoly in coffee marketing was effectively abolished by allowing the private sector into coffee trad- ing, marketing, and exports. Maximum import duties were lowered from 230 percent to 40 per- cent, and to a weighted average of 21.5 percent by 1998. The negative list used to determine eligibility for imports through the foreign exchange auction has been reduced. Exporters of goods and services are now al- lowed to retain 10 percent of foreign exchange earnings. The govern- ment also issued a new labor code that incorporates much international practice in the areas of employment conditions and compensation; it eliminated the monopoly power of some official marketing and trade corporations, and completed the decontrol of agricultural prices. The investment code was revised four times in as many years and presently allows greater private sector participation, including in public utilities where the state enjoys a de facto monopoly. In addition, a number of modest measures have been taken to promote exports. Ethiopian public enterprises have historically been commercially (if not always economically) profitable, and their privatization does have important fiscal implications. For example, they accounted for 90 per- cent of federally collected revenues in 1996, with a full 40 percent com- ing from just 10 of the largest SOEs (World Bank 1998a:46-47). Nonetheless, the Ethiopian Privatization Agency has recently stated that, during 1993-98, it undertook the divestiture of 176 firms (three-quarters of them retail shops and hotels) at a total price of 2.5 billion birr, or US$330 million. Financing and the inefficient land-lease policy are the major con- straints on the planned case-by-case privatization of 120 large enterprises over the next three years. Some 16 of the enterprises were sold to foreign capital. The largest investor, the Saudi conglomerate of Midroc Ethiopia, which accounted for 64 percent of privatization receipts, paid US$172 million for the Lega Dembi Gold Mines and another US$90 million for ten agro-processing enterprises. It also owns the new Sheraton Addis and a network of con- struction, service, and financial establishments (UNCTAD/ICC 1999). The birr has been allowed to float, as a result of which the ETB/US$ exchange rate rose from about 2 in mid-1992 to about 8 in mid-1999. This flexibility has resulted in the convergence of the official, auction, and parallel market exchange rates. The real exchange rate depreciated significantly as a result of the nominal devaluation and disinflation. The growth rate of the GDP deflator declined from 21 percent to 3 percent between 1991 and 1997. With additional support for improvements on 194 Post-Socialist Reformers the supply side, these reforms will result in a marked improvement in the competitiveness of Ethiopian exports. A number of tax policy reforms and other government revenue en- hancement measures were also introduced. The number of income tax brackets was reduced and the top marginal tax rate was lowered from 85 percent to 40 percent. Revenue mobilization, at less than 15 percent of GDP, remains inadequate. Tax arrears and corruption (notably in Cus- toms, parts of Inland Revenue, and at the regional level) continue to be major problems, as is the 1960-vintage commercial code (World Bank 1998a, 1998c). Reform Dialogue: Government and Stakeholders A striking feature of the landscape of reform in Ethiopia is the absence of an institutionalized mechanism for subjecting economic policy propos- als to discussion or debate between the government and the various business and civic organizations. The dearth of public information and the limited public participation in policy deliberation makes Ethiopia an odd man out among such progressive reformers as Uganda and Ghana. The marginalization of important stakeholders (opposition parties, various chambers of commerce and sectoral associations, farmers' groups, the independent press, professional associations, trade unions, local non- governmental organizations, the academic community, and the like) re- flects the autocratic legacies of the country's political systems. Successive governments have prevented, co-opted, or marginalized nascent civil society institutions. This has had several unsavory consequences. The business community is impelled to resort to corruption; the "watchdog" role of an independent, informed press is weakened; the sense of owner- ship of the reforms by the public is foregone; effective oversight by Par- liament is undermined; and issues of distributional faimess across regions or sectors become overly politicized. Recent research on aid effectiveness and sustainability of reforms points to a number of conclusions. First, equitable growth is predicated on sound macroeconomic fundamentals, good budget management, and institutional reforms that foster the mobilization of underutilized re- sources. Second, the sustainability of good reforms, in turn, depends on domestic ownership as manifested by the commitments of the leader- ship and the citizenry to the reforms along with the capacity to imple- ment them. Third, good governance ultimately undergirds the above in the form of an accountable executive, a professional bureaucracy, the rule of law, a stable and transparent policymaking process, and an en- gaged civil society (World Bank 1998a; Leonardo and others 1999). The formal decisionmaking framework under the current EPRDF government is clear enough. The overall development strategy and policy Ethiopia 195 direction are apparently scrupulously debated within the central com- mittee of the EPRDF. The prime minister's office is the focal point of coordination on major policy decisions between the executive branch and the legislative branch (EPRDF congresses and the Parliament, which is controlled by the ruling party). The Ministry of Economic Develop- ment and Cooperation (MEDAC) is the main gateway for donors. Line ministries and regional planning bureaus also have varying levels of input. Any loan secured has to be signed off by the Council of Ministers. Donors focusing on program support and sector programs have found this arrangement slow but expedient since it lends itself to efficient tech- nocratic relations with unambiguous lines of authority. However, the Ethiopian governance system, by short-circuiting broad participation, undermines the sustainability of reforms in the long run. The easy phase of reform (involving rehabilitation, liberalization, and stabilization) seems to have enjoyed the tacit support of most sectors of the population, con- sulted or not. Second-generation reforms involving ownership of land, large state enterprises, party-owned business groups, and competition policy are, however, too contentious for the viability of a model based on diktat. The angst of the business community and other stakeholders is elo- quently expressed by the well informed, if sometimes overly critical, economics correspondent of the Addis Tribune, a respected weekly The assessment is worth quoting at length since it aptly captures the agenda for future policy debate: It is no wonder that in Ethiopia today several risible economic poli- cies and practices are in effect: Individuals cannot own land; houses and other properties confiscated by the predatory Derg regime con- tinue to be in the possession of the present government; the official policy of ethnic bias prompts leaders to allocate disproportionate amounts of public resources to favoured regions; the mixing of politics with business is flagrantly practised; privatization of na- tionalized enterprises is going on without any regard to the issue of compensation; back-breaking increases have been made in rents on government-owned (but previously privately-owned) commer- cial premises; unprecedently high taxes are levied on rental income and capital gains; bona fide franco valuta imports are prohibited; private foreign exchange bureaus are not allowed; foreign partici- pation in the banking and insurance sector is banned; directed credit is commonplace; top-notch university lecturers and experienced civil servants have been dismissed or forced into early retirement; peasants critical of official economic policy and political thinking have been disposed of their plots of land; commercial farming is eyed with suspicion; and peasant agriculture is being saturated 196 Post-Socialist Reformers with fertilizer to the neglect of irrigation. Similar deficiencies are easily observed in the political arena as well, but there is no need to go into them here as they are quite well-known. (Addis Tribune 1998e) Donors, including usually outspoken bilaterals, have chosen to bifur- cate economic conditionality and political conditionality. Their com- plaints appear to have been largely confined to the more programmatic issues of public sector efficiency, transparency, and a blanket appeal for stakeholder participation and respect for basic human rights. A number of donors, for example, suspended program aid in 1999 for fear that fungible resources might be diverted to finance the Ethio-Eritrean war, which claimed 20 percent of the 1999 budget (World Bank 1999d; IMF 1999b; Economist 1999). Given the current outpouring of pan-Ethiopian patriotism, many hope that the EPRDF administration will take advan- tage of this opportunity to establish a more inclusive governance system.6 Donor Assistance Strategy Given the importance of multidonor participation, in terms of both ideas and money, to the post-Derg reform programs, it is useful to take a brief look at the goals and strategies of the donor community. I draw heavily on the World Bank's Country Assistance Strategy since it is well documented. The highlights of reform measures pursued by the multilateral insti- tutions and multidonor conferences are presented in appendix 4.2. The history of reform in the past three decades and the ODA data reviewed earlier both suggest that, with the exception of the international finan- cial institutions, donors on the whole have been unable or unwilling to effect a close linkage between aid and reform, perhaps for fear of pun- ishing the poor. Ethiopia enjoys a good relationship with the community of multi- lateral and bilateral donors. IDA assistance was devoted in the early phase of the TGE reform program to economic recovery, structural ad- justment, and the preparation of policy framework papers as well as technical documents. It then switched emphasis in 1997 to sector in- vestment programs (SIPs), primarily externally funded and spanning five to ten years. Other multilateral donors active in these programs include the African Development Bank, the European Union, and the United Nations agencies. Several bilateral donors also provide active support for these programs. Overall, the donor assistance strategy for use of aid to Ethiopia has been directed at four policy clusters. The first cluster focuses on policy dialogue together with economic and sector work designed to empha- size the importance of macroeconomic stability, private sector develop- Ethiopia 197 ment, and improvements in capacity for implementing projects or poli- cies. A number of studies, including five public expenditure reviews and an equal number of PFPs, have provided a coherent framework for link- ing project funding with sector and program aid in a forward-looking manner (World Bank 1997a). A second cluster is devoted to infrastructure development with a spe- cial emphasis on expanding the road network and the supply of electric- ity. The third cluster concentrates on growth-enhancing support for agriculture (mainly improved seeds and fertilizer), exports (designed to diversify the export basket and reduce dependence on large flows of aid), and small enterprises. A fourth cluster of reforms is aimed at the alleviation of poverty with the help of ambitious sector investment pro- grams in education and health and modest projects in the areas of popu- lation, gender equality, food security, and water supply. SIPs have a number of defining characteristics. They are sector-wide in scope; they provide a coherent framework for sector policy; they put local stakeholders in the driver's seat; they let all donors sign on; they enjoy common implementation arrangements with a multiyear horizon and transparency; and they require minimal long-term foreign technical assistance (Rose 1985; Lister 1998). Presently, there are SIPs in three pri- ority sectors: transport (roads subsector), education, and health. They are integrated into the government's budget and implemented by gov- ernment agencies in collaboration with the private sector, especially in health and road building.7 Similar initiatives in the areas of energy and food security are in the making. The formal mechanisms for aid coordination in Ethiopia are strong, at both the external and country levels. At the external level, aid coordi- nation is carried out through Consultative Group meetings, four of which have been held since 1992. The multidonor meetings of the Special Pro- gram of Assistance for Africa, also held biennially since 1992, provide a useful forum for aid coordination and review of country performance. At the country level, MEDAC is the overall coordinator of all external assistance delivery to Ethiopia. Other key participants include the prime minister's office, the Ministry of Finance, the National Bank of Ethiopia, and the Council of Ministers. Frequent consultation between local rep- resentatives of donors and the government is now routine. More than 15 donor countries are active in Ethiopia. Their areas of emphasis include relief and rehabilitation, food aid, social sector devel- opment, infrastructural development, and adjustment support in col- laboration with the multilateral development institutions. Germany, Japan, Italy, the United States, the Nordic countries, the Netherlands, and Canada are the leading bilateral donors (see table 4.5). Those giving smaller amounts include the United Kingdom, France, Switzerland, Austria, Australia, China, and the Republic of Korea.8 198 Post-Socialist Reformers Many nongovernmental organizations (NGOs) are also active in Ethio- pia although their number is low by the standards of other African coun- tries. According to the Federal Disaster Prevention and Preparedness Commission, there are 310 registered NGOs (one-third of which are for- eign) operating in the country. They are active in the areas of relief and rehabilitation as well as in development programs geared toward edu- cation, health, water supply, agriculture (most notably the Sasakawa/ Global 2000 Program), and off-farm income-generating activities for poor households. The combined annual budget of NGOs in Ethiopia is vari- ously estimated in the range of US$150 million. A new "Code of Con- duct for NGOs in Ethiopia" was signed by 165 NGOs in March 1999. This followed a period of friction with the government, which is accused of, among other things, favoring party-affiliated NGOs. Donor Assessment of Reform Record A World Bank mission to socialist Ethiopia recalls an oft-repeated obser- vation regarding the perplexing gap between the potential and the ac- tual performance of the Ethiopian economy: Ethiopia has some comparative advantages stemming from its size, natural resources, location, and its competent civil service .. . Ethio- pia has generally avoided heavily bank-financed deficit spending and consequent inflation. The country's external debt burden has so far been kept within manageable bounds, and in (favorable) contrast to the experience of many African countries, its rates of capacity utilization in industry have been quite high. Its record of project implementation has been relatively good, and its techno- crats and civil servants have a reputation for honesty, dedication and competence. These are indeed important strengths and successes of economic management to build upon. (World Bank 1987: 3) The report goes on to explain this paradox of high potential and low achievement in terms of "initial conditions" of poverty and underdevel- opment, vulnerability to numerous internal and external shocks, and the misguided economic policies of the Derg government. Carefully not- ing that it had no quarrels with socialism per se, the mission made a strong plea for policy reform in the name of economic efficiency. The first full-scale single-country evaluation of European Commu- nity/European Union aid was conducted on Ethiopia in 1995, covering the previous 20 years (IDS/IDR 1996; Maxwell 1998). It reached the fol- lowing conclusions, among others: (a) the effectiveness of EU aid was undermined by the unfavorable policy environment and the numerous shocks to the economy; (b) there was a mixed pattern of success (infra- Ethiopia 199 structure, water, energy, Shoa province PADEP, sectoral import program for agriculture) and failure (rural water supply, soil conservation, food aid, Stabex); (c) there was a gradual "learning by doing" on the part of the EU, since the agency was inadequately prepared but also contractu- ally constrained to engage in policy dialogue, especially under Lome I and Lom6 II; and (d) the bias of the Ethiopian governments toward overcentralization and overstandardization was not vigilantly countered. The review recommends a unified Ethiopian program instead of a mul- tiplicity of instruments and a long chain of decisionmaking back to Brus- sels, and greater emphasis on policy dialogue with the government. Recent World Bank reports suggest a much more upbeat assessment regarding the first-generation or "Washington consensus" reforms un- der the new government (that is, reconstruction, stabilization, liberal- ization, and some privatization). First, the leadership is described as knowledgeable and committed to serious economic reform.9 Second, policymakers are said to exhibit a strong sense of "ownership" of the reform programs-a point proudly underscored by the prime minister himself (Zenawi 1999). Ownership here means irreversibility of com- mitment to the reforms, and the capacity to implement them. Third, the economic record, in terms of growth and stability, has so far been impressive. The reports also express some concerns. Loan portfolio performance, despite significant improvements through the 1980s, still receives an only "satisfactory" rating. Although the government has shown a steep learn- ing curve in implementing promised reforms, it continues to confront a number of problems. Tax collections (especially income and domestic sales) are narrowly based and backlogged even in relatively developed areas such as Addis Ababa. The overregulated government system breeds corruption, although not on a grand scale comparable to many other African countries. The civil service is underpaid and initiatives from below undervalued; and the judicial system remains woefully inadequate (World Bank 1998a, 1998c). The net effect is that absorptive capacity (de- fined narrowly as the current rate of disbursement) is increasingly be- coming a binding constraint on the full utilization of pledged aid. That is not to say that donors are always efficient. A recent assess- ment of the three SIPs laments the fall in external financing to less than half of what was promised (Ethiopia 1999). Concerns about the unreliability of ODA disbursements are also reflected in the strong pref- erence expressed by regional governments for predictable federal subsi- dies over earmarked foreign aid of similar amounts. The relatively low effectiveness of EU aid is also partly attributed to the EU's weakness in policy dialogue and the callous neglect by the EU bureaucracy of the importance of good local administrative resources to oversee the imple- mentation process (Maxwell 1998). Poor food aid targeting has meant 200 Post-Socialist Reformers that only half of the food-insecure households benefit from grain distri- bution (Clay, Molla, and Habtewold 1998). The leadership is seen to be especially attracted to the East Asian model of development, and the dual-track and gradualist Chinese approach to post-socialist reform. Ethiopia, however, lacks many of the ingredients of East Asian success: a government committed more to nurturing than controlling weak markets, a highly meritocratic and politically insulated technocracy, high rates of saving and investment, a reservoir of off-farm production capability comparable to China's dynamic township and vil- lage enterprises, a region with lots of positive economic spillovers, and a much less fractious polity (Wade 1990; World Bank 1993; Woo 1998). The government has now begun to confront, albeit less decisively, the second generation of reforms in the areas of financial deepening, com- petition, managerial capacity, transfer of technology, diversification of exports, and labor-intensive private small enterprises (Stiglitz 1998). The World Bank laments the reluctance to go all-out for a radical transfor- mation into a market economy in these terms: The careful consideration of policy options and the "ownership" that it reflects are welcome but, in general, the Government tends to be over-cautious in undertaking the second generation of re- forms that are needed now. The boldness and decisiveness of the initial phase of reforms has given way to an all-round cautious- ness on virtually all fronts in the next phase. Also weaknesses in implementation capacity need to be addressed if the absorption of concessional resources is to rise substantially and the economy is to move to the "frontier" of the balance between stabilization and growth. (World Bank 1997a: 10) The IMF expresses similar sentiments in more specific terms: After a number of ambitious "first-generation" reforms in the early years of the present regime, the structural reform effort slowed somewhat in subsequent years. An anti-export bias persisted in the trade regime, exchange and trade regulations remained bur- densome, foreign direct investment was limited to a narrow range of activities, and capacity building in the government was con- strained by slow implementation of the civil service reform. More- over, despite the establishment of the Ethiopian Privatization Agency in 1994, the sale of SOEs continued to be undertaken at a slow pace. The rudimentary state of the financial sector and the outdated reach of the legal and regulatory framework also hin- dered private sector development. (IMF 1999b: 20) Ethiopia 201 Some Ethiopian economists have suggested that financial stability has been attained largely through budgetary restraint and significant in- creases in aid flows (Chole 1993,1994; Degefe and Nega 1999/2000; Addis Tribune 1998c). Achieving sustained and broad-based growth requires greater emphasis on a positive supply response. It is also worth noting here that the relevant standard for judging the efficacy of the current policy stance is no longer the dismal record of the Derg regime. Domes- tically, the relevant frame of reference ought to be the goal of doubling real per capita income every 10 to 15 years. This presupposes a growth rate of 8 percent per year during the period. Externally, the benchmark should be the incentives provided to foreigners relative to those of coun- tries that compete most directly with Ethiopia for investment or export markets. One challenge is to raise the discouraging domestic saving rate (6 percent) and the modest investment rate from 18 percent to at least 25 percent of GDP. This, in turn, cannot be accomplished with increased tax effort alone. Accelerating reforms in such areas as land use, banking, and transparency would improve Ethiopia's competitiveness by provid- ing an unambiguous signal to the overcautious private sector that the government is committed to an open, rule-based market economy (UNCTAD/ICC 1999; World Economic Forum 1998). The upshot is that Ethiopia is at a crossroads between the easy phase of liberalization and the hard phase of deep institutional reform. The post-1990 record shows that the country has made a decisive break with its socialist past in many, but not all, respects. Price controls have been largely dismantled, tax and tariff rates lowered, and the inflation rate kept low. The EPRDF government's strong sense of ownership of the reforms has until recently resulted in well-coordinated donor assistance, which has focused on sector investment programs in roads, health, and education. The momentum of reform seems to have been adversely affected by the 1998-2000 intermittent war along the border with Eritrea. It has cer- tainly diverted the attention of senior policymakers from issues of long- term development. Aside from the loss of tens of thousands of lives, the conflict is estimated to have cost the central government directly as much as US$1.5 billion. The switching of port traffic from Assab to Djibouti, the destruction of farms, the disruption of regional domestic trade, the reductions in outlays for capital expenditure and economic services, the temporary surcharges on key imports, and the disruption of flights for Ethiopian Airlines are additional economic costs borne by a fragile economy. On top of this, donors suspended virtually all aid save emer- gency assistance and projects already underway. It is, therefore, by no means clear that the good record in macroeconomic liberalization can be 202 Post-Socialist Reformers sustained much less replicated at the microeconomic level given the in- evitable need to kick-start the reform process. THE RELATIONSHIP BETWEEN AID AND REFORM The relationship between aid and reform, especially when it involves institutional transformation, is profoundly political. It involves redistri- bution of power within society, and therefore calls for both positive and normative analyses. The comments below are based on the history of the government's preferred policy posture and the negotiated reform measures actually undertaken, and what I was able to glean from vari- ous donor reports or gather from interviews primarily with government officials in March 1999. A more nuanced treatment of the political economy of Ethiopian re- form calls for an intimate knowledge of policy debates within the ruling party and the nature of negotiations between the donors and the gov- ernment. Compared to the case of other African reformers, the identifi- cation of pro-reform and anti-reform policy elites is made more difficult (and perhaps moot) by the opaque nature of the decisionmaking process under the collective model of party governance that currently prevails in Ethiopia. The Aid and Reform in Africa project, of which this paper is a part, sets out to test three hypotheses: (a) governments choose to reform, or regress, independent of the aid relationship; (b) nonfinancial aid has a better impact than financial aid on the generation of policy reforms; and (c) financial aid works when policy reforms and institution building are already underway. The overall conclusion is that the aid-reform linkage in Ethiopia is rather complex: aid appears to have induced or cemented some reforms but not others; and the optimal sequencing of the forms of assistance (money or ideas) or the efficient timing of aid is often hard to establish in the absence of a clear counterfactual. Selected examples will be used to illustrate these issues. Where Aid Succeeds in Inducing Reform In countries like Ethiopia, aid is likely to induce and lock in new reform or broaden existing reform when the foreign exchange constraint is bind- ing and (a) the payoffs to both the government and citizens are demon- strably high, or (b) the payoffs to the rulers are high even though the "masses" get little, or (c) successful reforms are introduced progressively and delicately so that even a reluctant government can conveniently claim ownership. The payoff to the ruling elite may take the form of directing externally funded projects and off-budget expenditures to favored re- Ethiopia 203 gions or sectors, shifting fungible resources to politically favored but low-return activities, and misappropriation of public funds where cor- ruption is rampant. It also depends on the presence of a crisis situation (thereby generating intense pressure in favor of high-stake reforms) at the time reforms are seriously contemplated (Grindle and Thomas 1991). In the case of Ethiopia, the willingness to engage in meaningful re- form seems to have prevailed since 1988. The Derg approached West- ern donors with hat in hand once it became clear that Soviet bloc aid was about to end and that the socialist experiment had failed. The tread- mill of modest reforms did not impress potential donors, however.10 Major bilateral donors such as Italy displayed little interest in pushing for reform, and more reform-minded ones like Sweden limited their efforts to projects that skirted policy barriers to reach the poor directly. The IMF and the World Bank were active in the preparation of PFP1 under the Derg, but it was not implemented because the government balked, particularly at the requirement of a substantial devaluation of the birr (Chole 1994). In the post-socialist period, quick-disbursing and highly fungible aid initially followed the restoration of peace and the enactment of policies focused on stabilization and rehabilitation. Subsequent aid financed jointly designed sector programs, financial reform, and improvements in public sector management. The EPRDF government was quite keen to obtain funds for essential imports, emergency recovery, and rehabilitation (Bevan 1994; World Bank 1997b; Maxwell 1998). Despite an initially poor loan portfolio perfor- mance, there was mutual eagerness among donors and the government to push for significant increases in quick-disbursing loans and grants during 1991-93. While calling for a convincing strategy for marketization, a consortium of donors orchestrated by the World Bank provided US$680 million for the Economic Recovery and Reconstruction Program, followed by a US$250 million structural adjustment credit (World Bank 1997b). The significant evolution in official thinking (or calculation) becomes most evident when one contrasts the EPRDF government's initial ap- prehensiveness about a rapid marketization of the economy, as expressed in "Economic Policy of Ethiopia during the Transition Period" (Ethiopia 1991), with the latest reform measures agreed to in PFP5, "Ethiopia: Policy Framework Paper 1998/99-2000/01" (Ethiopia 1998a). The 1991 policy paper, acknowledging the failures of the Derg regime, committed the TGE to deepening the liberalization initiated earlier by the Derg in the areas of road transport, domestic retail (but not wholesale) trade, and rural labor markets. It also put great emphasis on reconstruction of war- torn regions and rehabilitation of infrastructure. Critical decisions concerning external trade and foreign exchange sys- tem reform, financial reform, privatization of public enterprises, and land 204 Post-Socialist Reformers and civil service reform were postponed until after the 1995 national election. Many of these issues were sidestepped even after the end of the transitional period, as we will see below. The potential leverage at the disposal of the major donors (notably the World Bank, AfDB, EU, Germany, Italy, Sweden, and the United States) is indicated in part by the level of aid intensity. In 1996 Ethiopia received 5 percent of the net ODA of approximately US$17 billion to Sub-Saharan Africa. This works out to US$16 per capita, one of the low- est per capita aid levels among countries of similar size. It is, however, 16 percent of Ethiopia's GNP, above the average for Sub-Saharan Af- rica."' In the 1990s the aid share in GNP averaged 23 percent. Since total government expenditure is a quarter of GNP, these figures imply that ODA flows were equivalent to 90 percent of budgetary expenditure. Aid currently finances as much as half of the capital budget and one-sixth of federal government revenue (see table 4.2). During 1992-94, the World Bank made it clear that the coordinated support for recovery and stabilization was being provided with the ex- pectation that it would be followed with an expanded program of assis- tance tightly leashed to a number of policy "triggers." It is fair to say that many of the structural reforms would not have been considered seriously, much less fully implemented, without the infusion of substan- tial policy loans and grants (Abegaz 1994; Chole 1994; World Bank 1997a). The contents of successive PFPs clearly show that a number of policy measures are conspicuous by their recurrence. At donor insistence, the investment code was revised four times in as many years to bring it closer to best practice among African reformers. The foreign exchange system was progressively pushed beyond episodic devaluations to in- clude auctions and an interbank market. The international financial in- stitutions are currently pushing for the phasing out of the auction system and full current account convertibility (more on this below). The gov- ernment has reluctantly committed itself to the privatization of large SOEs although PFP5 still calls for improved modality that includes trans- parency in utilization of the proceeds. The persistent call for rationaliz- ing the decentralized budgetary system, including the introduction of rolling public investment plans, is beginning to be heeded. Finally, the call for a comprehensive food security plan is being taken seriously and might result in a full-fledged sector program. While the government initiated most of these reforms, the prima facie evidence suggests that donor pressure has made a critical difference in transforming many of them into more than piecemeal measures. How far the aid leverage could have been deployed by a "united front" of donors in an effort to buy even more ambitious reforms is nonetheless hard to establish. Ethiopia 205 Where Aid Fails to Generate Reform Cases where aid, for various reasons, failed to produce more than mar- ginal reforms are more readily identifiable. Consistent with the experi- ences of many countries, in Ethiopia there appear to be few cases where conditionalities overcame stiff government resistance (Collier and oth- ers 1997; Leonardo and others 1999). The following examples pertain to disagreements involving national security issues, or the regime's power base, or its view of the optimal mix between the public sector and the private sector (as in the case of certain conditionalities of the IMF). The threat of a reduction or withdrawal of aid appears on the whole ineffective whenever Ethiopia is in the midst of a major national crisis. When the Derg declared Ethiopian socialism, some multilateral donors (notably the World Bank) and important bilateral donors (the United States) withheld much of their already meager assistance to the country, losing potential leverage in the process. The United States went further and refused delivery of weapons already paid for. Facing secessionist rebellions at home and an invasion by Soviet-armed Somalia, the re- gime sought and received massive (mostly military) assistance from the Soviet Union itself. In the spirit of the Cold War, the two superpowers ended up switching sides. Many informed Ethiopians blame shortsighted Carter Administration policy for unwittingly aiding in the transforma- tion of an otherwise reformist committee of noncommissioned officers into one controlled by a dictator (Mengistu Haile Mariam) with a gran- diose vision of garrison socialism. Similar threats in early 1999 by most donors to suspend program aid, which were soon carried out, have been bitterly resisted by the EPRDF government. Aid also does not seem to buy good reform when the conditionalities are perceived, rightly or wrongly, to undermine the core of the regime's ideology or its power base. Although the World Bank and other major donors understandably regret the lingering mindset of control over fa- cilitation in the bureaucracy, slippage in economic management, and in- efficiencies in the land lease system, they have also chosen to sidestep or soft-pedal some of the more fundamental institutional or policy reforms needed to solve them. That is probably because they know that these and other reforms are unpalatable to the government (World Bank 1997b). A notable policy issue that is conspicuous by its absence from the economic reform agenda is the question of private ownership of agricul- tural and urban lands. Ethiopia had a long history of private ownership of land in a multitude of tenure forms (freehold, kinship, village, parish, clan, crown). In all these institutional forms, wealth was based less on ownership title than on entitlements (customary or legal) to the produce from the land. European-type feudal estates or Asian-type widespread 206 Post-Socialist Reformers landlessness were rare. This explains in large part why the nationaliza- tion of land, which transferred use rights to producers, had such a neg- ligible impact on output or productivity. Officials put the case against the granting of freehold tenure in terms of the need for access to land for all in order to ensure economic security for the rural poor (Zenawi 1999). The government's longstanding prom- ise to hold a national referendum on land ownership has yet to material- ize (Ethiopia 1991). Its current strategy of "agricultural development-led industrialization" focuses instead on long-term leases and such short- cuts as supplying fertilizer and improved seeds. The experience of Ethiopia and other countries, however, demonstrates that the economic case against reprivatization is unconvincing (Deininger and Feder 1998). The linking of land access to residency, the inability to use land as collateral, and the insecurity that arises from loss of land as a result of periodic redistributions (due to demographic pressure or as political punishment) have stifled agricultural development. Labor mo- bility has been reduced or frozen. Optimal farm sizes fail to emerge as the efficient farmers who produce much of the marketable surplus can- not increase their farm size in a country where 20 percent of farmers account for 80 percent of grain sales. More alarmingly, investment in permanent improvements of the land is discouraged. Finally, off-farm activities that would absorb those who choose to mortgage or sell their subsistence plots (some 60 percent cultivate less than one hectare) are stymied. A secure private land title (individual and communal) is, of course, a necessary but not sufficient condition for a green revolution to occur. This is far from the case of a "dry prairie" awaiting the drenching rains of private incentives to bloom. Ownership reforms will have to be comple- mented by a more efficient marketing system and enhanced access to modem inputs. Support in the form of public infrastructural investment, credit, and extension services remains minimal. The inescapable fact is that the most effective anti-poverty program is one that reinvigorates rural development by harnessing the willingness of Ethiopian farmers to embrace demonstrably profitable innovations (Howard and others 1998; Degefe and Nega 1999/2000). In the urban areas, some half a million poorly maintained housing units remain in government hands. The urban land leasehold system, whose inefficiency and corruption are the subject of perennial complaints by the business community, has contributed to a severe shortage of com- mercial space and housing. Addis Ababa, whose municipal limit boasts plenty of land, continues to suffer an unacceptably high rate of open and disguised unemployment; many of the jobless have secondary or even tertiary education. The binding constraint here is one of finance or institutional infrastructure rather than a shortage of disciplined and easily Ethiopia 207 trainable labor (UNCTAD/ICC 1999; Middlebrook and Corzato 1999; Addis Tribune 1988a). A second area of policy contention that has received little attention in the dialogue between the government and the various donors is the con- cern regarding the growing number of enterprises that are surreptitiously owned by the ruling party. While the use of political power by "big men" to accumulate personal wealth and grease a patrimonial system is all too common throughout Africa, the phenomenon of an extensive net- work of party-controlled enterprises is a rather new development in Ethiopia."2 According to several unpublished reports and reports in the private press, the reliability of which cannot be fully ascertained, a large num- ber of enterprises wholly or partially owned by EPRDF are being orga- nized into interlocking financial and industrial groups (FIGs). Their assets are estimated in the range of 1 billion to 2 billion birr. They are said to be registered in the name of various individuals, NGOs, or regional gov- ernments. Most of the TPLF business concerns are reportedly under the supervisory umbrella of the Endowment Fund for the Rehabilitation of Tigray, whose major divisions are headed by senior party members (Aynekulu 1996a, 1996b; "Ethiopian Non-Governmental Business" 1999). Some of the largest EPRDF companies or their affiliates are as fol- lows: in agriculture, Hiwot Agricultural Mechanization, Zeleke Agri- cultural Mechanization, Tesfa Livestock, and Rahwa Goat and Sheep Export; in finance and trade, Wegagen Bank, Africa Insurance, Guna Trading House, Ambassel Trading House, Dinsho Plc, Wondo Trading, and Dedebit Credit and Savings Institution; in industry and mining, Meskerem Investment, Ezana Mining, Sheba Tannery, Almeda Textiles and Garment, Addis Pharmaceuticals, Mesfin Industrial Engineering, Beruh Chemical, and Dashen Brewery; in construction, Addis Construc- tion, Sur Construction, and Mesebo Building Materials; in transport, Trans-Ethiopia, Blue Nile Transport, and Express Transit; and in services, Mega-Net Corporation and Experience Ethiopia Travel.13 The concern about the role of these holdings is multifaceted. First, such firms in effect create two distinct private sectors-one that is politi- cally connected, and another that is not. The former can exploit political connections (through party cadres in the bureaucracy) to practice unfair competition for government business and to engage in rent-seeking in the form of tailored state subsidies or protection and insider privatization. These firms can also potentially be used to undermine the economic position of the party's political opponents, all without violating the let- ter of the law. Second, the government's near monopoly over the bank- ing system lends itself to directing credit to favored enterprises on less than commercial terms. Third, one can only wonder how the existence of these organizational economic assets has shaped the government's 208 Post-Socialist Reformers attitude toward future reforms, especially in banking, the commercial code, and the judiciary. If party- and bank-organized FIGs are here to stay, damage from rent-seeking and stripping of public assets may be minimized with an introduction of enforceable antitrust and fair com- petition legislation along the lines of Russia or Korea (Johnson 1997). A third illustration of the limits of aid as leverage comes from finan- cial-sector reform which, along with land policy and Party-affiliated businesses, constitutes the "third rail" of Ethiopian political economy. The three-year $127 million ESAF loan (1996-99) was not fully imple- mented because of a row between the government and the IMF (IMF 1998a). The midterm review of the ESAF was allowed to lapse at the end of 1997 over a number of disagreements between the government and the IMF (Addis Tribune 1998b). The Fund released the second annual disbursement in November 1998 on account of the good overall record on macroeconomic stability, and in exchange for limited concessions in the foreign exchange regime (see appendix 4.2 and Ethiopia 1998a). The IMF argues that the low level of financial intermediation and poor securities markets in Ethiopia, even by African standards, are due to largely policy-imposed "structural rigidities." The barriers to financial deepening include: the near-monopoly of the banking system held by the Commercial Bank of Ethiopia (CBE), including the requirement that SOEs do their banking exclusively with the CBE; poor collatoralization of assets, largely tied to market failures linked to the land-lease sys- tem; a misguided ban on foreign banks, insurance companies and stand- alone forex bureaus; and inadequate bank supervision and prudential oversight. Concern was also expressed about weakening external ac- counts and lackluster tax collection which threaten Ethiopia's advance to the decision point of the HIPC Initiative (IMF 1999B; Demeksa 1999; Harvey 1996). TIhe government's reluctance to undertake rapid financial reform ap- pears to reflect two concems. For one, the authorities rightly view radi- cal liberalization of the banking system and the capital account as a risky proposition for countries like Ethiopia with weak regulatory and judi- cial systems. The recent experience of transition economies and emerg- ing markets provides an important cautionary tale. Furthermore, policymakers seem frustrated by the unrealistic expectation of donors that a country with limited capacity can satisfactorily implement four ambitious sets of interconnected reforms simultaneously: fiscal devolu- tion, civil service reform, sector development programs, and banking reform. The other reason is that a less-transparent control of key finan- cial institutions constitutes an important leverage for the ruling party. Again, where economics meets politics, the resistance against sensible reform is likely to be strong. The Ethiopian experience suggests that aid followed rather than led the initial wave of reforms. Since then, a good mix of technical and financial Ethiopia 209 assistance has had the effect of coaxing the government to move closer to the limits allowed by its own ideological or narrow organizational interests. The sequencing of Ethiopian reform is the traditional one of progres- sion from the easy and immediately necessary to the hard. The reform program began with emergency recovery and stabilization and moved on to marketization and modest privatization. While other post-social- ist reformers (Tanzania and Mali, for example) have made some head- way in the transition between first-generation and second-generation reforms, Ethiopia continues to hesitate as it ponders the difficult jump into a full-fledged market economy. The Coupling of Money and Ideas Nonfinancial assistance, in the form of policy dialogue, analytical work (country memoranda, specialized sector studies, policy framework pa- pers, public expenditure surveys, public finance management, and sta- tistical work), and the provision of equipment and training can make a critical difference when it is focused on upgrading domestic capability. Ideas concerning the virtuous circle of openness, stability, and growth or the importance of a medium-term economic framework for rational- izing overall policy are certainly being internalized in Ethiopia. Some Ethiopian policymakers, however, express mixed feelings about some conventional aspects of technical assistance. They support techni- cal assistance where knowledge transfer is programmed, the advice is grounded in Ethiopian conditions, and the country is given some free- dom to buy expertise in the free market as needed. However, they tend to look askance especially at those consultants who "parachute in." These complaints are also familiar in other African countries (Berg 1993). Technical cooperation can certainly be useful even in nonreforming economies, since it can often help minimize damage from distortionary policies. The small nonfinancial assistance provided during the Derg years (about 15 percent of net ODA) was indeed helpful in upgrading technocratic skills. It was, however, too modest in size (and not tied to the prospect of a big aid inflow) to tip the balance in favor of what amounted to a regime-undermining systemic reform. The environment was such that open criticism by reform-minded civil servants would have been reckless (Griffin 1992; World Bank 1987). Anecdotal evidence sug- gests, however, that various segments of the bureaucracy strove to mini- mize economic damage by selective and slow implementation of misguided policies. Recent multilateral assistance to Ethiopia has had a high policy con- tent and a good analytical foundation. Technical assistance has recently accounted for some 20 percent of ODA, about two-thirds of the level for developing countries (O'Connell and Soludo 1998: table 1). Aside from 210 Post-Socialist Reformers the expected problems of inexperience and staff turnovers, the new ad- ministration is praised for fulfilling the numerous conditionalities of the structural adjustment program faithfully. Ethiopia is nonetheless far from the policy frontier compared to its competitors, and analytical work and policy dialogue are judged to be at the formative stages (World Bank 1997a, 1997b).14 Reforning the Donor-Beneficiary Relationship The fashionable rhetoric of full partnership notwithstanding, both the donor-recipient relationship (for grants) and the donor-client relation- ship (for soft loans) are fraught with inequality in Africa (van de Walle and Johnston 1996). Donors typically come armed with money, superior analytical capability (but inferior knowledge of local conditions), and the mentality of a generous but calculating patron. Aid bureaucracies are partly motivated by narrower incentives such as maximizing dis- bursement rates even at the cost of quality. Anecdotal evidence suggests that the behavior of some officials from bilateral aid agencies betrays an unwillingness to maximize the pass-through of funds while insisting on micromanaging the aid relationship.15 For policymakers in countries like Ethiopia, where a strong sense of confidence and ownership prevails, some policy conditions and exces- sive documentation requirements appear as wasteful diversions of scarce administrative resources. In interviews, many express a clear preference for greater multilateralization of aid, a guaranteed flow of aid under a negotiated framework, and disbursement in the form of block grants (earmarked or unrestricted budget support known as channel 1 funds) to the extent possible (also see Maxwell 1998). A negotiated framework, these officials argue, provides flexibility of implementation while permitting donors to concentrate on the "rules of the game" and on "evaluating outcomes." Evaluation is, however, frus- trated by the multiplicity of often conflicting output indicators akin to the "success indicator problem" of Soviet planning. Finally, while there are many donor conferences, there is no Africa- wide forum for channeling the collective views or reform proposals of recipient countries. In addition to delinking development aid from nar- row foreign policy or commercial objectives, there are a number of issues to be addressed, including recipient input concerning the amount and form of technical assistance and the uneasy roles of mushrooming NGOs. CONCLUDING THOUGHTS Post-socialist economic reform in Ethiopia has so far been undertaken in two stages: rehabilitation, stabilization, and basic liberalization during Ethiopia 211 1990-95; and infrastructure-oriented SIPs, financial liberalization, and improvements in public sector management during 1996-98. Aid, flow- ing at an annual average of US$1.3 billion (in constant dollars) followed the first phase of homegrown reforms as a reward for good behavior. Well-coordinated aid in the form of policy dialogue and analytical work, as well as money, also played an important role in jointly forging the sec- ond phase of reform, this time as much by persuading as by rewarding. Despite the significant accomplishments of the past six years, there is a substantial opportunity to further revamp the still regimented economy. Such measures would include restoring private ownership of most land; reforming the banking system by breaking up the CBE and granting autonomy to the National Bank of Ethiopia (NBE); a vigorous competi- tion policy with respect to large SOEs and party-owned businesses; and reform of the governance system, both the political and bureaucratic aspects. Many of these reforms have frontloaded costs and backloaded benefits that render them unappealing to myopic policymakers. They may also entail concentrated opportunity costs to the ruling party but diffused benefits to the electorate. Multinational agencies often pose the basic question of reform in Africa in terms of rescuing nascent markets from the bundle of contra- dictions that characterize the post-colonial state: "authoritarian and yet fragile; highly centralised yet dispersed; overdeveloped in size yet underdeveloped in function; trumpeting national unity yet controlled by narrow personal cliques; nationalistically resistant to outside politi- cal pressures but weakly subordinate to economic ones" (Africa Confi- dential 1999: 5). While the Ethiopian state is old and resilient, it does share the African pattern of a center that is too weak to centrally plan a socialist economy or to "govern" the market effectively. It is neverthe- less strong enough for capricious interventions which, among other things, raise transaction costs substantially and stimulate growth- retarding informality. The Achilles heel of Ethiopian development (and reform) is its inabil- ity to nurture an equitable political system that secures lasting peace at home and stability in the Horn of Africa. In March 1978, Ethiopia won a bloody border war with Somalia. March 1990 saw the country's return to a market economy in the midst of a raging civil war between an alli- ance of two liberation fronts and the Derg government. Almost a decade later, in March 1999, an even bloodier war was re-ignited between the two erstwhile allies. More than US$6 billion worth of Soviet bloc and Western arms sup- plies to the impoverished Horn of Africa region has left a legacy of chronic violence and mutual destabilization. The region has witnessed two of Africa's most mechanized wars in just two decades. The resumption of a war economy is once again accompanied by a major famine in the northern provinces affecting 5 million to 6 million people. The economic 212 Post-Socialist Reformers distress is compounded by an intensifying AIDS pandemic. External de- stabilization and incessant domestic conflicts produce myopia born of a heightened sense of insecurity, destruction of capital (physical and so- cial) and information, militarization of life, and mutual distrust. The two contradictory faces of external aid, one destructive and the other con- structive, have yet to be explicitly contrasted in the assessment of aid to Ethiopia. It is no wonder, then, that Ethiopia presents a classic case of an economy with great potential struggling to overcome a "low-level equi- librium trap." The hopeful signs-periodic spurts of growth, driven mainly by positive shocks such as favorable weather or higher coffee prices and aid flows-prove fleeting time and again. One consequence of this is the desire of both educated Ethiopians and those with capital to emigrate. Another is the strong preference by risk-averse capital for myopic investments that produce quick profits. Such an environment is invariably inimical to long-term investment and growth (Collier and Gunning 1999; Bevan 1994; Chole 1993). All is not lost, however, and talk of aid triage (a severe form of selec- tivity) by some donors is rather misplaced. The past six years have shown that, with peace and modest economic reforms, the economy is capable of expanding at more than twice the rate of population growth. This growth record has been facilitated by a remarkable degree of macroeco- nomic stability. One lesson from successful open economies is that, while sustained stability is necessary for investment, removing the remaining barriers to growth on the supply side is just as essential. The domestic capability that is being built up in Ethiopia with the help of external assistance will have to be aimed at exploiting the untapped opportunities at home and abroad (especially the underutilized Middle Eastern, Lom6, and GSP market outlets). The bridge between domestic capability and opportu- nity is supplied by sufficiently deep economic reform. Successful reform is, in turn, undergirded by a strong commitment to good governance. This is essential for rebuilding business confidence, including among the growing Ethiopian diaspora. In the final analysis, political reform is a domestic affair. As the disap- pointing response to the highly liberal investment code amply demon- strates, a country that treats its citizens fairly is also a country that is attractive to foreign capital. If the government manages to undertake the remaining reforms while building up its capacity for implementa- tion, Ethiopia stands to benefit doubly from an increasingly selective international aid regime and a reinvigorated private sector. Donors can aid this process with a better calibration of aid and reform. Ethiopia 213 APPENDIX 4.1 ETHIOPIA: CHRONOLOGY OF POLITICAL AND ECONOMIC DEVELOPMENTS 1974-75 A military junta led by NCOs (Derg) overthrows the monarch in a creeping coup d'etat. The Derg proclaims socialism and nationalizes all land, medium and large enterprises, and extra houses. 1975-84 The Derg uses zemetcha campaigns to implement land redistribution, establish peasant cooperatives and urban kebelles, and consolidate state control of the economy and polity. A major border war with Somalia results in the switching of superpower alliance toward the Soviet Union. Nationwide famine affects millions of people, resulting in relief efforts worldwide. The era of favorable terms of trade for primary exports ends. 1987-89 Ten-Year Perspective Economic Plan is introduced. Ethiopia is formally declared a socialist "people's democracy." Workers Party of Ethiopia is established. Villagization and resettlement campaigns intensify. Civil war in the northern part of the country escalates. 1990-91 WPE Congress announces the decision to liberalize prices and make membership in the cooperatives voluntary. Peasants dismantle producer cooperatives. EPRDF and EPLF guerrilla alliance overthrows WPE government. Transitional Government of Ethiopia formed. 1992-93 Stabilization and rehabilitation program under- taken. Province of Eritrea formally secedes from Ethiopia (1993). 1994-97 Major economic reforms introduced, including liberalization, stabilization, and structural adjust- ment; significant gains are achieved in terms of growth and stability. Ethnic-based federal system introduced under a new constitution. 214 Post-Socialist Reformers 1998-2000 Ethiopia and Eritrea fight three rounds of a costly border war. Most official loans and grants suspended pending end of the conflict. Prolonged drought and a major famine plagues the country once again. Looming HIV/AIDS crisis enters the national agenda. Ruling EPRDF wins nearly all seats in the 2000 parliamentary elections. Ethiopia 215 APPENDIX 4.2 ETHIOPIA: CHRONOLOGY OF MAJOR ECONOMIC REFORMS, 1974-98 1974-78 Derg nationalizes all land, modern enterprises, and extra houses. Peasant associations (rural kebelles) established. Urban neighborhood associations (urban kebelles) established. Rural cooperative proclamation issued. 1979-83 Lower and advanced producer cooperatives established. State commercial monopolies control trade and transportation sectors. State industrial enterprises rationalized and expanded. Economic zemetcha (mass mobilization) campaigns launched. 1984-89 Resettlement and villagization programs ex- panded. National famine relief efforts. Ten-year Perspective Economic Plan introduced. 1990-91 Land-use rights become inheritable. Limited price decontrols introduced. Producer cooperatives decollectivized. Private interregional trade permitted. Demobilization of 455,000 soldiers after EPRDF takeover in mid-1991. 1991-95 Stabilization of prices and reduction of budget deficit. Rehabilitation and reconstruction of civil war damage. Economic Recovery and Rehabilitation Program. Structural adjustment program under SAC and SAF: * Decontrol of most prices * Foreign exchange auctions * Reduction of import tariffs * Decentralized land-lease system * Entry of private banks * Treasury bill market * Establishment of Ethiopian Privatization Agency 216 Post-Socialist Reformers First generation of civil service reform: * Salary reform * Retrenchment of government employees Birr currency area established with Eritrea. Ethiopian Investment Authority established. 1996-98 Sector development programs in health, education, roads, energy. System of fiscal federalism introduced with nine regional states. Structural adjustment program under ESAF. With respect to thefinancial sector: * Decontrolled lending rate, but floor for deposit rate. * Abolished automatic surrender of forex, and 10 percent retention. * Established interbank market, though not yet active. With respect to theforeign exchange and trade system: * Raised the frequency of forex auctions, reduced most restrictions, and permitted bank-run forex bureaus. * Eliminated most restrictions on current account transactions. * Eliminated forex surrender requirements and permitted a 10 percent retention of export proceeds. * Lowered average tariff to 19.5 percent and reduced tariff bands to eight. * Ended parastatal monopoly on customs clearing and forwarding services. With respect to export promotion and development: * Ceased price verification on non-coffee exports * Allowed exporters to engage in all implicit forms of foreign credit except formal loan agreements With respect to privatization: * Completed privatization of 175 enterprises. * Brought 12 SOEs to point of sale. * First phase of privatization brought in Birr 2.8 billion for the extrabudgetary privatization fund. * Second phase envisages sale of 115 corporatized SOEs. * Government to retain over 50 of the largest, profitable SOEs. Ethiopia 217 With respect to private sector development: * Abolished subsidies for and ban on private trade in fertilizers. * New legal and regulatory framework and authority. * Foreclosure law introduced. * Further liberalization of the foreign investment code. With respect to the civil service reform program of 1996: * New directives on financial and debt manage- ment, and procurement by federal government. * Budget management and control manuals prepared. * Outstanding challenges: quality of staff, performance-based incentives, and corruption. Sources: Abegaz 1994; World Bank 1997a, 1999d; IMF 1999b; Ethiopia 1996, 1998a. APPENDIX 4.3 ETHIOPIA: HIGHLIGHTS OF DONOR ASSISTANCE STRATEGY, 1980-2001 Policy and capacity Adjustment Physical Poverty and Agriculture and development and growth infrastructure human services environment 1980-91 Technical assistance Coffee processing Energy Education V-VII Irrigation planning agency IMF standby (1981-82) Transport/telecom. Family health PADEP CEM Towns/urban 1992-94 CG1-2 (1992,1994) Economic recovery Road rehabilitation Food security Improved seeds SPA and reconstruction Calub Gas Fertilizer program Gilgel Gibe hydro PFP1-PFP3 SAC I and SAF PERI Private sector devel- opment 1995- CG3-4 (1996,1998) ESAF (1996-99) Power distribution Family planning Sale of state retail 2001 PFP4-PFP5 Export promotion Water supply Social rehabilitation Support services SIP (or SDP) SME development Road SIP Education SIP Research and training MTEF/PIP Rail rehabilitation Energy Health SIP Food security PER2-PER6 Women's rights Conservation CG Consultative Group (multidonor conferences) SIP Sector investment program SAC Structural Adjustment Credit ESAF Extended Structural Adjustment Facility SME Small and medium enterprises CEM Country economic memorandum PFP Policy Framework Paper PER Public expenditure review MTEF Medium-term Economic Framework PIP Public investment plan Sources: World Bank 1992,1995,1997a; USAID 1999; Maxwell 1998; IMF 1999b; Ethiopia 1998a. Ethiopia 219 NOTES 1. For this reason, the growth rates reported here do differ (mostly insignifi- cantly) from those reported in official publications for shorter periods. The Ethio- pian economy is among the least researched in Africa. 2. DRS is the Debtor Reporting System of the World Bank database; CRS is the Creditor Reporting System of the OECD database. 3. The 1997 accord between Russia and the Paris Club calls for an upfront discount of 80 percent on the stock of debt owed and the application of Naples terms to the remainder (which means another 67 percent reduction in NPV terms of the remainder 20 percent). If these assumptions hold-that is, a 90 percent write-off of Russian debt-the debt stock will fall to US$2.8 billion and the debt service to 184 percent of exports by 2001. Almost all of the Russian debt, now valued at US$4.8 billion, is military-related. 4. The projects included Komolcha Textile Mill, Mugher Cement Works, Nazret Tractor Assembly, and Melka Wakena hydroelectric project. 5. The terms under which the province of Eritrea (3.5 million people) for- mally seceded from Ethiopia in 1993 turned out to be of strategic importance for both countries in economic terms. Notable issues such as border demarcation, access to seaports, division of assets and the national debt, and equitable treat- ment of each other's citizens were either postponed or handled in a manner that was widely perceived as prejudicial to Ethiopia's interests. These inequi- ties are generally believed to be at the root of the 1998-99 border war between the two countries. According to the Ethio-Eritrean agreements on economic and political coop- eration, Assab and Massawa were to be free ports; goods, capital, and labor would move freely; and the Ethiopian birr was to serve as a common currency. However, the two countries had contrasting strategies of development, harmo- nization of macroeconomic policy was slow, and some 150,000 Eritrean nation- als resident in Ethiopia were accorded equal treatment with Ethiopians without full reciprocity by Eritrea. These loopholes produced an estimated 1 billion birr of debt incurred by the Eritrean monetary authorities to their Ethiopian coun- terparts, a substantial foreign exchange leakage due to re-exports by Eritrea of birr-purchased Ethiopian export products and participation of Eritrean agents in foreign exchange auctions in Addis, rising charges for port services, and the like (Addis Tribune 1998d). The arrangements were de facto terminated in 1998, with Ethiopia switch- ing its major port to Djibouti and requiring trade in hard currency upon the introduction of the Eritrean nakfa in 1997. The immediate consequences of the war for Ethiopia and Eritrea have included deportation of each other's citi- zens on grounds of national security, tens of thousands of casualties on both sides, displacement of half a million people, and diversion of development funds (from investment and debt service funds in the case of Ethiopia) for the acquisition of sophisticated weaponry-not to mention an estimated daily 220 Post-Socialist Reformers expenditure of US$1 million by each to prosecute the war (Economist 1999; Ethiopia 1999). 6. According to the 1999 reports of Amnesty International, Human Rights Watch, and the Ethiopian Human Rights Council, there are more than 10,000 political prisoners in Ethiopia, most being held without formal charge or trial. The second national parliamentary elections were held in 2000. 7. The donors for the transport (road sector) SIP include the European Union, the Department for Intemational Development of the U.K., the German Agency for Technical Cooperation and the German Development Bank, the Nordic De- velopment Fund, and the Japan International Cooperation Agency. The educa- tion sector SIP was a cooperative effort of AfDB, Finland, Germany, Ireland, Italy, Sweden, the United States, UNESCO, UNICEF, and UNDP. Finally, the health sector SIP was partially funded by AfDB, the EU, Finland, Ireland, Italy, the Netherlands, Norway, Sweden, Austrian Development Cooperation, UNDP, UNICEF, UNFPA, the United States, and WHO. 8. There are "national flag" patterns here as in other African countries. The United States, for example, limited its cumulative aid of more than US$1 billion in 1984-94 to emergency food assistance and rehabilitation. It has since expanded its menu to include development assistance in the areas of education, health, food security, and democratization (USAID 1999). The German aid program, channeled mainly through GTZ, emphasizes skills transfer, poverty alleviation, food security, resource management, and selected social development programs. 9. The most recent Country Assistance Strategy (World Bank 1997: 2) points out that, notwithstanding the unfortunate ethnic flavor to politics, the developmentalist nature of the govemment is reflected in "an exceptionally strong sense of 'ownership' of economic policies and to probity in their imple- mentation. Policymakers are given to very serious consideration of policy op- tions, do not make policy commitments lightly and have a solid track record of delivering on those they do. The Prime Minister [Meles Zenawi] himself is very keenly involved in economic management and has an impressive grasp of eco- nomic issues to an extent that the label of an outstanding economic technocrat could be applied to him without much exaggeration." 10. The conflict over the Peasant Agricultural Development and Extension Program (PADEP) provides a good illustration. Drawing on the experience of the minimum package extension programs of the previous government and the successful training and visit system in India, IDA offered as much as US$400 million in the early 1980s to provide modem inputs and advice in potentially surplus-producing districts. This was conditional on government action to raise farmgate prices and to remove draconian restrictions on private trade. The Derg refused on account of its different priorities and ideological commitments. These terms, however, were accepted in principle by the same government at the end of 1987 and limited funding became available once the crisis of agrarian social- ism became undeniable (Clapham 1988). In 1989, IDA disbursed US$77 for PADEP I. Ethiopia 221 11. The comparable DAC figures for Ethiopia's peer group in 1996 were as follows: Ghana, US$38 per capita or 11 percent of GNP; Kenya, US$22 or 7 per- cent; Uganda, US$25 or 13 percent; Zambia, US$40 or 11 percent. For other com- parative figures from early 1990-95, which make Ethiopia's aid intensity even lower, see O'Connell and Soludo (1998: table 4). 12. This phenomenon of cadre capitalism, familiar from the early years of the KMT in Taiwan, is also ubiquitous in the Sudan and Eritrea in the 1990s. Although the 1994 Ethiopian constitution prohibits political organizations from engaging in profit-making activities, the public is keenly aware of the promi- nence of these corporations. For example, a recent World Bank mission on cor- ruption observes: "Private sector and civil society representatives told the mission that politi- cal parties belonging the EPRDF, especially the Tigrayan People's Liberation Front (TPLF), have established companies which can include banks and insur- ance companies. There is a common perception among some in the business community that these companies have preferential treatment over private sec- tor companies with respect to government contracts, government-controlled credit facilities, import and export licenses, and customs clearances. The mis- sion was not able independently to confirm these allegations. "Government interlocutors explained that a number of organizations devel- oped during the seventeen years of struggle against the Dergue, acquiring as- sets and property. Once the war was won, the winning side had to decide what to do with them. Since holding companies are not provided for in Ethiopian law, the EPRDF created foundations under which the party-affiliated compa- nies (and NGOs) are located. The intention of such companies and NGOs is to provide benefits to the survivors of war-wounded, widows, orphans, handi- capped, etc. Several of the party-affiliated organizations are headed by party members. These organizations are not allowed to acquire institutions which have been or are being privatised (World Bank 1998c: 7)." Senior government officials insist that the firms do not receive any special favors from the government (Zenawi 1999). Another argument advanced in their favor is that, being large, these firms compete solely with foreign enter- prises. They therefore play little role in displacing the domestic private sector (author's interview with Ato Neway Gebreab, chief economic advisor to the prime minister). 13. Other TPLF-affiliated NGOs that are active investors in the front compa- nies include the Tigray Development Association Endowment Fund for Relief and Reconstruction of Tigray, and the Relief Society of Tigray. The rest are owned or co-owned by the other allied parties of the EPRDF coalition: Amhara Na- tional Democratic Movement, Oromo People's Democratic Organization, and South Ethiopia People's Democratic Movement. 14. Ethiopia's overall CPIA (Country Policy and InstitutionalAssessment) score of 3.7 for 1998 is on par with Ghana (3.6), but behind such reform leaders as Botswana (4.4) and Uganda (4.2). The subcategory scores for Ethiopia (compared 222 Post-Socialist Reformers with the Africa-wide average) are as follows: 4.3 (3.2) for macroeconomic man- agement; 3.1 (3.1) for sustainable and equitable growth; 3.8 (2.9) for reduction of inequality; and 3.9 (2.8) for public sector management. The 1998 Africa Competi- tiveness Report ranks Ethiopia low in overall competitiveness (17 out of 23), but high for the optimism of the business community (5 out of 20) and for its FDI protection regime (World Economic Forum 1999). 15. A senior policymaker in a line ministry recently told me of an encounter with a representative of a bilateral agency who suggested that funds be de- posited in the name of the minister. When told that such a practice is not per- missible in Ethiopia, the response was that it is indeed common elsewhere in Africa. A director of the Economic Commission for Africa calls this phenom- enon "a race to the bottom": such donors seek out the least demanding recipi- ents, and Africans are almost invariably blamed for the resulting aid ineffectiveness. From the donor perspective, the challenge of managing an aid program in the complex environment of countries like Ethiopia is aptly captured by the metaphor "hitting a moving target from a moving platform" (Maxwell 1998: 107). REFERENCES Abegaz, Berhanu. 1994. "Ethiopian Economic Reform." In Berhanu Abegaz, ed., Essays on Ethiopian Economic Development. Aldershot, U.K.: Avebury. Addis Tribune ("by our economic correspondent/commentator"). 1998a. "Yet An- other 5-Year Action Plan-This Time Around for Addis Ababa." 20 February. . 1998b. "Latest Talks Between Ethiopia and IMF Fail to Lift ESAF Sus- pension." 13 March. - 1998c. "Foreign Aid is Propping up the Ethiopian Economy." 10 April. - 1998d. "Unfair Economic Advantages Eritrea Enjoyed from Its Hege- monic Relationship with Ethiopia." 31 July. . 1998e. "Ethiopia's Alarming Brain Drain." 12 December. Africa Confidential. 1999. "Africa Scrambles for Africa." 40(1):2-5. AfDB (African Development Bank). 1996. "Ethiopia: 1996-98 Country Strategy Paper" Abidjan. Aynekulu, Awualom. 1996a. "The Emerging Monopolies of the TPLE" Ethio- pian Register (July): 20-32. . 1996b. "The Emerging Monopolies of the TPLF, Part 2." Ethiopian Regis- ter (August): 14-22. Berg, Elliot. 1993. Rethinking Technical Assistance and Capacity Building in Africa. New York: United Nations Development Programme. Bevan, David. 1994. "Economic Aspects of Ethiopian Transition to Peace." In Jean-Paul Azam, David Bevan, Paul Collier, Stefan Dercon, Jan Willem Gun- ning, and Sanjay Pradhan, "Some Economic Consequences of the Transition Ethiopia 223 from Civil War to Peace." Policy Research Working Paper 1392. World Bank, Development Research Group, Washington, D.C. Cassen, Robert. 1994. Does Aid Work? New York: Oxford University Press. Chang, Charles C., Eduardo Femrndez-Arias, and Luis Serv6n. 1999. "Measur- ing Aid Flows: A New Approach." Policy Research Working Paper 2050. World Bank, Development Research Group, Washington, D.C. Chole, Eshetu. 1993. "The Dismal Economy: Current Issues of Economic Reform and Development in Ethiopia." Ethiopian Journal of Economics 2(1): 37-72. . 1994. "A Preliminary Appraisal of Ethiopia's Economic Reforms, 1991- 93." In Harold Marcus, ed., New Trends in Ethiopian Studies, vol. 2. Trenton, N.J.: Red Sea Press. Clapham, Christopher. 1988. Transformation and Continuity in Revolutionary Ethio- pia. Cambridge: Cambridge University Press. Clay, Daniel, Daniel Molla, and Debebe Habtewold. 1998. "Food Aid Targeting in Ethiopia." Grain Market Research Program Working Paper 12. MEDAC, Addis Ababa. Collier, Paul, and Jan Willem Gunning. 1999. "Explaining African Economic Performance." Journal of Economic Literature 37 (March): 64-111. Collier, Paul, P. Guillaumont, S. Guillaumont, and J. W. Gunning. 1997. "Rede- signing Conditionality." World Development 25 (9): 1399-1407. Degafe, Befekadu, and Berhanu Nega, eds. 1999/2000. Annual Report on the Ethio- pian Economy. Addis Ababa: United Printers for the Ethiopian Economic As- sociation. Deininger, Klaus, and Gershon Feder. 1998. "Land Institutions and Land Mar- kets." Policy Research Working Paper 2014. World Bank, Development Re- search Group, Washington, D.C. Demeksa, Bulcha. 1999. "Revising the Banking Sector." Addis Tribune, July 30. Dercon, Stefan, and Daniel Ayalew. 1998. "Where Have All the Soldiers Gone: Demobilization and Reintegration in Ethiopia." World Development 26 (9): 1661-75. The Economist. 1999. "Africa's Forgotten War." 15 May. Etherington, Dan, and Alasebu Yainshet. 1988. 'The Impact of Income Terms of Trade for Coffee on Capital Goods Imports and Investment in Ethiopia." Eastern Africa Economic Review 4 (1): 48-52. Ethiopia. 1991. "Economic Policy of Ethiopia during the Transitional Period." Prime Minister's Office, Addis Ababa. .1992,1993,1994,1996,1998a. "Ethiopia: Policy Framework Paper." Min- istry of Finance and National Bank, Addis Ababa. .1998b. "The 1994 Population and Housing Census of Ethiopia." Central Statistical Office, Addis Ababa. "Ethiopian Non-Governmental Business: Companies Controlled by or Associ- ated with EPRDF-Member Organizations." 1999. Anonymous document. Addis Ababa. Processed. 224 Post-Socialist Reformers Ghani, Ejaz, and Hyoungsoo Zang. 1995. "Is Ethiopia's Debt Sustainable?" Policy Research Working Paper 1525. World Bank, Development Research Group, Washington, D.C. Griffin, Keith, ed. 1992. The Economy of Ethiopia. London: Macmillan. Grindle, Merilee, and John Thomas. 1991. Public Choices and Policy Change: The Political Economy of Reform in Developing Countries. Baltimore: Johns Hopkins University Press. Haile Mariam, Mengistu. 1990. "Report by Mengistu Haile Mariam: Resolu- tions Adopted by the 11th Plenum of WPE." Workers' Party of Ethiopia, Addis Ababa. Harvey, Charles. 1996. "Banking Reform in Ethiopia." Working Paper 37. Insti- tute of Development Studies, Sussex, U.K. Howard, Julie, Mulat Demeke, Valerie Kelly, Mywish Maredia, and Julie Stepanek. 1998. "Can the Momentum Be Sustained? An Economic Analysis of the MOA/SG2000's Experiment with Improved Cereals Technology in Ethiopia." MEDAC, Addis Ababa. IDS/IDR (Institute of Development Studies/ Institute for Development Re- search). 1996. An Evaluation of Development Cooperation Between the European Union and Ethiopia, 1976-1994. 7 vols. Addis Ababa. IMF (International Monetary Fund). 1988a. "IMF Approves Second Annual ESAF Loan for Ethiopia." Press release 98/51. October. 1998b. International Financial Statistics 1998. Washington, D.C. 1998c. Ethiopia: Statistical Appendix. IMF Staff Country Report 9816. Washington, D.C. . 1999a. International Financial Statistics. February supplement. Washing- ton, D.C. .1999b. Ethiopia: Recent Economic Developments. IMF Staff Country Report 99/98. Washington, D.C. Johnson, Juliet. 1997. "Russia's Emerging Financial-Industrial Groups." Post- Soviet Affairs 13 (4): 333-65. Lensink, Robert, and Oliver Morrissey. 1999. "Aid Instability as a Measure of Uncertainty and the Positive Impact of Aid on Growth." University of Nottingham. Leonardo, Jose, Hartwig Schafer, and Gaspar Frontini. 1999. "Towards a More Effective Conditionality: An Operational Framework." World Development 27 (2): 285-99. Lister, Stephen. 1998. "Implementing Sector Development Programs in Ethio- pia." Consultant's draft report. Addis Ababa. Maxwell, Simon. 1998. "Effectiveness of European Aid to Ethiopia." In Marjorie Lister, ed., European Union Development Policy. New York: St. Martin's Press. Middlebrook, P., and E Corzato. 1999. "Displaced, Disempowered and Dis- tressed: A Focus on Poverty in Addis Ababa." Draft. World Bank, Washing- ton, D.C. Ethiopia 225 O'Connell, Stephen, and Charles Soludo. 1998. "Aid Intensity in Africa." Work- ing Paper Series 3-99. Centre for African Economies, Oxford University. OECD (Organisation for Economic Co-operation and Development). 1999. Geo- graphic Distribution of Financial Flows to Aid Recipients. Paris. Ramachandran, Mahadevan. 1997. "Food Security: Economics of Famine, Food Aid and Market Integration in Ethiopia." University Microfilms International, Ann Arbor, Mich. Rose, Tore 1985. "Aid Modalities: Sector Aid as an Instrument in Sub-Saharan Africa." In T. Rose, ed., Crisis and Recovery in Sub-Saharan Africa. Paris: Organisation for Economic Co-operation and Development. Stiglitz, Joseph. 1998. "More Instruments and Broader Goals: Moving Toward the Post-Washington Consensus." 1998 WIDER Annual Lecture, Helsinki, Finland. UNCTAD/ICC (United Nations Conference on Trade and Development and International Chamber of Commerce). 1999. An Investment Guide to Ethiopia: Opportunities and Conditions. New York: United Nations. UNDP (United Nations Development Programme). 1998. Human Development Report 1998. New York: Oxford University Press. USAID (United States Agency for Intemational Development). 1999. "Ethiopia: Congressional Presentation." Washington, D.C. van de Walle, Nicolas, and Timothy A. Johnston. 1996. Improving Aid to Africa. Policy Essay 21. Washington, D.C.: Overseas Development Council. Wade, Robert. 1990. Governing the Market: Economic Theory and the Role of Govern- ment in East Asian Industrialization. Princeton: Princeton University Press. Woo, Wing Thye. 1998. "Why China Grew." In Peter Boone, Stanislaw Gomulka, and Richard Layard, eds., Emerging from Communism: Lessons from Russia, China, and Eastern Europe. Cambridge: MIT Press. World Bank. 1987. "Ethiopia: Recent Economic Developments and Prospects for Recovery and Growth." Washington, D.C. . 1992, 1995, 1997a. "Ethiopia: Country Assistance Strategy." Washing- ton, D.C. . 1993. The East Asian Miracle: Economic Growth and Public Policy. New York: Oxford University Press. . 1994. Adjustment in Africa: Reforms, Results, and the Road Ahead. New York: Oxford University Press. .1995. World Tables 1995. Washington, D.C. . 1997b. "Ethiopia: Implementation Report on Structural Adjustment Credit (SAC I)." Washington, D.C. 1998a. Ethiopia: Review of Public Finances. 2 vols. Washington, D.C. 1998b. Assessing Aid: What Works, What Doesn't, and Why. New York: Oxford University Press. 1998c. "Ethiopia: Anti-Corruption Report." Washington, D.C. 1999a. "Ethiopia: Regionalization Study." Washington, D.C. 226 Post-Socialist Reformers . 1999b. World Development Report 1999/2000. New York: Oxford Univer- sity Press. . 1999c. Global Development Finance 1999. Washington, D.C. . 1999d. "Ethiopia Public Expenditure Review: Aide Memoire." 1999 PER Mission, Addis Ababa. World Economic Forum. 1998. Africa Competitiveness Report 1998. Geneva. Zenawi, Meles. 1998. 'Interview." The Financial Times, 2 March. Mali Patrick Guillaumont Sylviane Guillaumont Jeanneney Jacky Amprou Auvergne University Clermont-Ferrand, France Oumar Cheick Sidibe Ministry of Economic Affairs, Mali ACRONYMS AND ABBREVIATIONS AFD Agence Francaise de D6veloppement BCEAO Banque Centrale des Etats de l'Afrique de l'Ouest BEAC Banque des Etats de l'Afrique Centrale CCCE Caisse Centrale de Coop6ration Economique CFA Communaut6 financi&e africaine (franc zone) CFAF CFA franc CFDT Compagnie Francaise des Textiles (French Textile Company) CMDT Compagnie Malienne pour le D6veloppement des Textiles (Malian Textile Development Company) CPS Contribution pour prestation de service (service charge) DAC Development Assistance Cormittee (of the OECD) DFI Droit fiscal d'importation (fiscal import tax) DNAE Direction Nationale des Affaires Economiques (National Director- ate of Economic Affairs) EDA Effective development assistance EDM Societ6 d'Energie du Mali (Mali Electricity Company) ESAF Enhanced Structural Adjustment Facility (of the IMF) GNP Gross national product HIPC Heavily indebted poor countries IAS Impot sur les affaires et services (tax on goods and services) IDA Intemational Development Association (of the World Bank Group) IMF Intemational Monetary Fund ISCP Impot special sur certains produits (special tax on certain products) ODA Official development assistance ODF Official development finance OECD Organisation for Economic Co-operation and Development OHADA Organization for the Harmonization of Business Law in Africa OPAM Office de Commercialisation des Produits Agricoles du Mali (Mali Agricultural Products Office) OSRP Office de Stabilisation et de R6gulation des Prix (Price Stabiliza- tion and Regulation Office) PASEP Public Enterprise Sector Adjustment Program (Programme d'Ajustement du Secteur des Entreprises Publiques) PRED Economic Reform and Development Program (Programme pour la R6forme Economique pour le D6veloppement) PRMC Programme de Restructuration des March6s Cerealiers (Cereals Market Restructuring Program) PRODEC Program for the Development of Education SAF Structural Adjustment Facility (of the IMF) SAP Structural adjustment program SDR Special drawing right (of the IMF) SGS Soci6t6 G6n6rale de Surveillance SOMIEX Societ6 Malienne d'Import-Export (Malian Import-Export Company) TOF Total gross official flows (from the public sector) USAID U.S. Agency for Intemational Development VAT Value added tax WAMU West African Monetary Union WAEMU West African Economic and Monetary Union In the case of Mali, analysis of the relationship between aid and eco- nomic policy reform is particularly interesting on several counts. First, Mali was, and still is in relative terms, one of the top 10 aid recipients in the world. In 1997, net flows of official development aid represented 18.4 percent of GNP, 77.2 percent of investments, and 47.5 percent of imports of goods and services (World Bank 1999c: 353). It should be noted that the "aid intensity" ratio as a function of the country's 1997 gross national product (GNP) corresponds roughly to the average over the past 10 years. Second, Mali has a long history of reforms that have dealt with the monetary regime, trade policy, and public sector management. After having embarked shortly after its independence upon a socialist experi- ment that included the creation of a separate currency, controls over for- eign exchange and external trade, and widespread state control over economic activities, Mali rejoined the franc zone in 1967 and the West African Monetary Union (WAMU) in 1984, and then helped to create the West African Economic and Monetary Union (WAEMU) in 1994. These radical institutional changes were, as we shall see, followed at irregular intervals by a great number of economic reforms both macroeconomic and structural in nature. Third, and perhaps because of the two preceding factors, Mali has been the object of particular interest on the part of donors interested in improving their procedures and the efficacy of their aid. Thus, at the initiative of the Development Assistance Committee (DAC) of the Organisation for Economic Co-operation and Development (OECD), a novel experiment in the local coordination of external aid was set up under the name Aid Review (revue de l'aide). It gave rise first to an evalu- ation report (OECD 1998), followed by regular and joint examinations of aid operations conducted by the main dispensers of aid at the local level. Mali was once again chosen to be one of seven countries included in the Structural Adjustment Participatory Research Initiative, in which the recipient government evaluates the adjustment policies in collabora- tion with the main donors and nongovernmental organizations. Given the intensity of aid, the number of aid sources, their degree of involvement in the reform process, and the country's longstanding and varied history of reform, Mali was thought to be able to provide some insight into the ways in which aid can or cannot influence economic reform. The relationship between aid and reform depends on the nature of the aid provided. In the case of adjustment aid, which includes explicit economic policy conditionalities, the linkage can be extremely close. Sec- tion 2 of this chapter, which describes the structure and evolution of the external aid that Mali has received, points up the importance, in the to- tal aid picture, of macroeconomic or budgetary aid and of structural or 229 230 Post-Socialist Reformers sectoral aid, as well as the general evolution of aid and the cycles it goes through. These last are, as we shall see, quite closely associated with economic policy cycles. It also emerges that, when the evolution and structure of Mali's aid are placed in their context (monetary coopera- tion, external debt, etc.), the links between aid and reform seem to have been mutually reinforcing. Section 3 deals with Mali's reform experience, its phases, and its ar- eas of application. The pace and effectiveness of reform appear to be greatly influenced by Mali's initial socialist experiment and by the country's place within regional institutions. The sectors where reform has been most effective are macroeconomic reform (that is, budgetary and monetary reform) and economic liberalization, where both external and internal trade are concerned. Section 4 attempts to derive some lessons from the Malian experience concerning the conditions under which aid can best foster reform. Four hypotheses are presented, dealing with the role of crises, internal con- sensus, aid coordination, and the financing of the costs of reform. TRENDS AND CYCLES IN AID TO MALI In order to understand the evolution of aid to Mali, it is useful to refer to different concepts of aid, each of which has a different significance for the study of the aid/reform relationship. Analysis of the evolution of aid then makes it possible to discern recurring "aid cycles." Finally, an analysis of the origin and nature of aid flows supports the hypothesis that the link between aid and reform tends to become stronger over time. Five Useful Aid Concepts Five concepts of external aid are used in this study. The first, as defined by OECD's Development Assistance Committee, is official development assistance (ODA). This represents aid flows to developing countries that are provided by public entities, that is, central and local governments, and by multilateral institutions, and that meet two criteria: (a) their main objective is to promote economic development and the well-being of the beneficiary countries (they therefore include technical assistance grants); and (b) the terms are concessional, with a grant element of at least 25 percent. ODA is usually calculated net of amortization. The main disad- vantage of this concept is that it fails to take into account nonconcessional flows, including those of the International Monetary Fund (for example, standby arrangements and enhanced facilities) and the World Bank (a consideration relevant to some countries, but not to Mali). The second concept, effective development assistance (EDA), devel- oped by Chang, Fernindez-Arias, and Serven (1999), is defined as the Mali 231 sum of total grants except technical assistance and the grant equivalents of official loans.' This type of aid is calculated on the basis of data from the OECD's DAC, but lumps together the grant equivalents of loans (rather than their face value), including loans with a concessional ele- ment of less than 25 percent. This concept, unlike ODA, does not take into account debt adjustments, whether in the form of debt forgiveness or rescheduling. Only concessional flows from the IMF (such as the Struc- tural Adjustment Facility and Enhanced Structural Adjustment Facility) are taken into account here. Two broader concepts can be introduced here in order to account for the whole of official development finance. The first, calculated on the basis of data gathered by Chang, Fernandez-Arias, and Serv6n (1999), is the aggregate official development finance (ODF). It includes not only effective development assistance (grants and the grant element included in loans), but also the nonconcessional element of loans, technical assis- tance grants, and the IMF's nonconcessional financing. The other is cal- culated on the basis of data from OECD's DAC and is referred to as total gross official flows from the public sector (TOF). It includes ODA (gross) as well as other flows from the public sector. The latter are public sector interventions whose main goal is not development-related and which, although they are intended to foster development, have a grant element below 25 percent and are therefore excluded from the ODA category. They are essentially public credits for exports, equity holdings and in- vestments in the public sector portfolio, and nonconcessional IMF and World Bank financing (enhanced facilities and standby agreements). These two aggregates-ODF and TOF-are conceptually equivalent, except for the debt adjustment issue (see figure 5.1). They differ mainly in the statistical sources used. For purposes of examining the relationship between aid and reform, the value of the ODF or TOF concepts is that they include not only tech- nical assistance (as does the ODA concept), but also the IMF's noncon- cessional financing (which ODA does not do). The latter carries macro- economic conditionalities and was used by Mali from 1971 to 1988. The value of the ODF or TOF concepts also resides in the fact that they deals with raw aid flows, that is, those that are subject to negotiation between the country and the donors and that are in this respect liable to influence a government's behavior. At the other end of the spectrum, net transfers point up the true contribution that aid flows are making to domestic expenditures, since they deduct amortization and interest from the raw amount. This concept is not only of macroeconomic significance, but also represents for governments that portion of external contributions above and beyond that which helps avoid the accumulation of external arrears. All flows initially expressed in current dollars were deflated by the index established for this purpose by OECD,2 based on the assumption 232 Post-Socialist Reformers FIGURE 5.1 THE DIFFERENT AID CONCEPTS USED (IN TERMS OF DISBURSEMENTS) OECD-DAC World Bank Net ODA + Grant element of EDA (i.e. gross) Sum of face value of nonconcessional Sum of grant ele- concessional flowsa loans ments of all offiicial - Corresponding - Technical assis- aid flows (apart amortizations tance from technical - Nongrant element assistance and of concessional nonconcessional loans IMF loans) + Other flows from - Impact of debt the public sector rescheduling of (net) gross ODA + Nongrant element of loans + Technical assis- tance Total net official 1 + Nonconcessional flows from the Inerstloans from the public sector [0] - Interest an agreements and enhanced facil- ities (gross) + Amortization of principal Net transfers Total gross official Official flows from the - Impact of debt development public sector restructuring finance (TOF) (ODF) a. Debt restructuring (whether cancellation or rescheduling) is taken into account as a fictive grant (which increases gross ODA) corresponding to normal amortization, which is itself accounted for as an imaginary amortization: gross and net ODA amounts are thus increased by the same amount, which is equal to the fictive, or avoided, amortization. Mali 233 that 1996 = 100. They represent aid flows in constant (1996) dollars. Fig- ures 5.2 through 5.11 and tables 5.4 through 5.11 in the appendix sum- marize data relative to the different aid concepts applied in Mali and provide a basis for the commentary that follows. Aid Intensity: Cycles of Aid and Reform As indicated in the introduction, aid intensity-that is, the relationship of aid to some other economic aggregate, which may be the GNP, im- ports, or investment-is particularly high in Mali (see figure 5.3 in the appendix). Over the past 10 years, ODA has represented about 20 per- cent of GNP, 53 percent of imported goods and services, and 85 percent of investment. These ratios put Mali among the top 10 developing coun- tries in terms of aid intensity. Although it has always been high, this intensity has experienced fluc- tuations. As shown in figure 5.3, net ODA as a percentage of GNP rose steadily, from only 10 percent of GNP in 1976 to 32 percent in 1985. It fell again to a low of 14 percent in 1993, and rose to nearly 25 percent in 1994 before declining again. In 1997, it was on the same order of magnitude as in the early 1990s or early 1980s. However, the wide variations in aid intensity should be interpreted in the light of exchange rate movements. The spike in aid intensity in 1994 is partly due, of course, to the devaluation of the CFA franc, which doubled the CFAF equivalent of external aid to Mali. The fact remains, however, that net ODA in constant dollars increased slightly in 1994 (see figure 5.2 in the appendix). Similarly, the rapid hike in ODA observed between 1981 and 1985 and its subsequent drop starting in 1986 resulted in part from changes in the dollar in relation to the French franc and, therefore, in relation to the CFA franc. Thus, between 1980 and 1983, ODA in dollar terms diminished even though aid intensity increased greatly. The First Three Cycles: 1962-82 Despite these divergences associated with the trading currency, it is pos- sible to identify several cycles in aid to Mali, which seem closely linked to the country's political situation. A first cycle was observed during the 1960s, and corresponded to aid flows from Eastern bloc countries during the socialist experiment from 1962-67, which then declined markedly in 1967-68. The reentry into the franc zone occurred at this point, and then the coup d'etat that brought Moussa Traore to power. A second aid cycle began in 1969 and peaked in 1975. It corresponded to an economic recovery program that benefited in 1969 and 1971 from 234 Post-Socialist Reformers two standby agreements with the IMF (which was the first to get in- volved in the franc zone), and was also undoubtedly related to the ex- ternal aid connected with the great drought of the early 1970s. Aid levels then fell back down in 1976. A third cycle peaked in 1980 and ended in 1982 following some eco- nomic difficulties. It was at this time that Mali resumed its dialogue with the Bretton Woods institutions. The Last Three Cycles: 1982-99 Since the early 1980s, three new aid cycles can be discerned in connec- tion with phases of economic policy. These phases will be the main focus of the rest of this study. The first cycle lasted from 1982 to 1987, with a peak year in 1986. Mali reestablished its relationship with the IMF, signed three standby agree- ments (in 1982, 1983, and 1985), obtained an agricultural sector adjust- ment credit (in 1983) from the World Bank, and rejoined WAMU in 1984. But in 1986-87 the agreements with the IMF were suspended (and no adjustment credit was obtained from the Bank). It should be noted that the growth of ODA and its subsequent decline in 1986-87 was not the direct result of changes in flows from the IMF, since those flows were at that time still nonconcessional. Rather, it seemed to be associated with the pace of reforms as reflected in relations with the IMF. A second cycle began in 1987, peaked in 1989, and ended in 1993. It corresponded initially to a wave of important reforms undertaken from 1987 to 1989 and was marked by a last standby agreement with the IMF in 1988. This was followed immediately by a structural adjustment facil- ity and a structural adjustment credit from the World Bank in 1990. Aid and reform then slowed due to political troubles in late 1991 and the overthrow of the regime in 1992-despite the provision in 1992 of an Enhanced Structural Adjustment Facility (ESAF) from the IMF-and then a suspension of adjustment lending in the franc zone in anticipation of the devaluation. The disbursement of the second tranche of the ESAF was put on hold pending this measure. A third cycle started to take shape in 1993, first with an increase in aid obtained in 1994-95 as an accompaniment to the devaluation and in sup- port of democratization. This increase was attributable as much to bilat- eral as to multilateral donors, the latter being a bit more generous. But it was followed by a new decline in 1997 and, as we shall see, by a corre- sponding decrease in the pace of reform. The apparent cycles of aid, as identified on the basis of ODA levels, appear again when one applies the concept of official development finance (ODF) (see figure 5.4). In the case of Mali, the proportion of nonconcessional loans is small, with the exception of the IMF's sup- Mali 235 port up to 1988. As became apparent, IMF agreements can play a sig- nal role for other donors, accelerating and slowing the pace of their disbursements. The Origin and Nature of Inflows: A Growing Aid-Reform Linkage The sources and the nature of aid flows to Mali have changed consider- ably over the past 30 years. Three phenomena characterize this evolu- tion: first, the change in monetary cooperation with France that occurred when Mali joined WAMU in 1984; second, the change in the relative importance of various donors; and, finally, the changing nature of aid in a context of growing indebtedness. These three factors profoundly modi- fied the relationship between aid and reform in Mali. Changes in Monetary Cooperation In 1967, Mali, which had withdrawn from the franc zone in 1962, signed a new monetary cooperation agreement with France and reentered the franc zone, without reentering the WAMU. This meant that the Malian franc was henceforth fixed relative to the French franc (one Malian franc = one French centime, a 50 percent devaluation of the Malian franc). The Malian currency became totally convertible into French francs, the guar- antee of convertibility being provided by the establishment of a compte d'operations (operations account) within the French treasury where the Central Bank of Mali was supposed to deposit its reserves and where it had unlimited drawing rights. Adherence to agreed-upon rules of mon- etary issuance was monitored by the Central Bank's board of directors, which included French representatives on equal footing with their Malian counterparts. The operations account, unlike what happened with the Banque Centrale des Etats de l'Afrique de l'Ouest (BCEAO) or the Banque des Etats de l'Afrique Centrale, was in debt from the outset. This was increasingly the case until 1984, when Mali reentered WAMU.3 Paradoxically, although France occupied a very important position within the Central Bank of Mali (which had a French general director until 1983 and Malian and French managers serving on an equal footing on the board of directors), it undoubtedly did not wish to impose strict management of monetary policy. Despite the statutory rule stating that the cumulative level of the Central Bank's contributions to the govern- ment could not exceed 10 percent of tax revenues, Mali's money supply grew very quickly, through advances to the public treasury as well as through contributions to public enterprises experiencing deficits. In a holdover from the 1960s, these enterprises continue to play an impor- tant role in the Malian economy. 236 Post-Socialist Reformers Mali's actual entry into WAMU, which had been planned under the 1967 agreement, did not occur until 1984. Mali's currency then became the CFA franc, which replaced the Malian franc without any change in parity between the two currencies. This marked a break in relations be- tween the Central Bank-henceforth the BCEAO-and Mali's public trea- sury The variation in the credit of the monetary system to the Malian public treasury had an impact on the money supply that was almost always negative between 1984 and 1992. This change resulted from the fact that the multilateral constraints exerted by other WAMU members proved stronger than the control wielded by France in an exclusively bilateral relationship, as well as from the fact that Mali's macroeconomic policy was the main focus of IMF conditionalities. Two factors obliged Mali to have recourse to the IMF. The first was that debits from the operations account, according to the franc zone's monetary cooperation agreements, required the debitor to turn to other sources of foreign currencies and thus to the IMF. The second was that Mali needed the IMF to lighten the constraints weighing upon its public finances. Indeed, the help provided in the form of the operations ac- count was in fact balance of payments support, but unlike other multi- lateral assistance with the same objective, it was not simultaneously supporting public finances, since the flip side of France's guarantee of the Malian franc's convertibility was the statutory limitation of the Cen- tral Bank of Mali's contribution to the government. This is why Mali, like most other franc zone countries, resorted to the IMF fairly quickly after the financial situation deteriorated, in 1969 as well as in 1982. Thus, from one phase to another of Mali's monetary regime, one sees donors playing a growing role in the definition of macroeconomic policy. As a WAMU member, Mali was obliged to devalue once again in 1994 (by 50 percent). For Mali, the devaluation of the CFA franc was almost an exogenous decision to which it rallied at the 1994 meeting of heads of state of the franc zone in Dakar. But for WAMU members, this meeting was also an opportunity to sign the treaty creating WAEMU (the West African Economic and Monetary Union), which not only implied the creation of a common market, but also provided for multilateral moni- toring of budgetary policies, as well as for joint sectoral interventions. This new institutional stage therefore had important economic policy implications for Mali: the influence of aid on policy could now be ex- erted in part through this regional conduit. Changes in the Sources of Aid Tables 5.3 and 5.7 in the appendix show changes in the proportions of various types of aid provided to Mali, starting in 1964 for ODA and in 1975 for ODF. For the 1960s, however, ODA statistics, which are ob- tained from the OECD, do not include contributions from countries Mali 237 that are not members of the DAC, many of which (such as communist bloc countries and their Arab allies) provided aid that cannot be in- ventoried under ODA.4 THE BILATERAL-MULTILATERAL SPLIT. If one looks at the extremes of table 5.4, that is, the 1964-67 and 1994-97 periods, it would appear that the relative proportions of bilateral and multilateral aid have remained unchanged, at 54 percent and 46 percent respectively However, during the 30 years in between these percentages fluctuated quite widely. The proportion of bilateral aid increased until 1982-85, when it reached 69 percent (as against 31 percent for multilateral aid), and then declined. The trend is a little less clear if, instead of considering ODA, one fo- cuses on ODF. This includes non-ODA public flows, bilateral as well as multilateral, and especially, it should be noted, the IMF's contributions through 1988 when the structural adjustment facilities came onstream. Over the last period, the proportion of bilateral financing remained pre- ponderant, at 59 percent, despite its steady 20-year decline. BILATERAL AID: FRENCH AID DECLINES, BUT REMAINS IMPORTANT. The relative proportion of French aid within ODA, which it naturally domi- nated in the decade following Mali's independence, diminished steadily to 16 percent of the total in 1994-97. This is still a high proportion, how- ever, and is equivalent to the World Bank's share. Among the bilateral aid flows, French aid has been closely linked to certain critical phases of economic policy because of the particular his- tory of Mali's monetary regime outlined earlier: Mali's return to the franc zone in 1967, its entry into WAMU in 1984 (over the 1982-85 period, French aid regained its 28 percent share, whereas it had stood at 21 per- cent over the 1972-81 period), and finally the devaluation of the CFA franc in 1994. French aid also contributes to the fact that the share of bilateral grants (including technical assistance) is much larger than the multilateral share. Bilateral aid from the United States also played an important role in terms of its volume and relationship to economic reform. Although there was a steady drop in the proportion of ODA coming from North America (from 13 to 8 percent, see appendix table 5.4), the 1988-90 period was characterized by an interesting American experiment in aid in support of economic reform (the PRED program, discussed farther on). Finally, one should note the significant growth in aid from Japan, which, with 7 percent of total ODA in 1994-97, was the second-ranking bilateral donor, even ahead of the United States (North America as a whole represented 8 percent). MULTILATERAL AID AND MULTILATERAL DEVELOPMENT FINANCING. Multilateral ODA, although at the same level in 1994-97 as 30 years 238 Post-Socialist Reformers previously (that is, 46 percent of the total), came from very different sources by the 1990s. During the 1964-67 period, European ODA dis- bursements represented almost all multilateral aid; in 1994-97 they ac- counted for only 13 percent of the aforementioned 46 percent (or 28 percent of all multilateral aid). Still, it should be noted that this propor- tion had fallen between 1982 and 1987 to 7 percent, on average, of total ODA (a drop that would be even more noticeable if the percentage were calculated for ODF, since this period corresponded to significant contri- butions from the IMF on conditions that were not yet those of ODA). The fluctuation noted in the relative proportion of European multi- lateral aid could, at first glance, be interpreted as a political manifesta- tion of a desire to modulate aide as a function of judgments about the country's general policy, since this type of aid rose in the 1990s relative to the 1980s. The fluctuation also reflects the wide swings in aid flows from other sources that brought total ODA to a particularly high level in the 1980s. Total European aid over the 1994-97 period was roughly at its 1972-81 level (table 5.4). It is reasonable to wonder why European aid did not undergo the same fluctuations as other forms of aid. The main compensatory change that occurred was, of course, the growth of the Bretton Woods institutions, which accounted for 3 percent of total ODA in 1964-67 and for 22 percent 30 years later: 16 percent for the World Bank (in the form of IDA credits) and 6 percent for the IMF. The weight of the IMF, like that of the Bank, is actually more important than these figures indicate, because IMF agreements are a prerequisite for many other types of aid, especially for the World Bank's adjustment operations and those of various other donors, and also because the World Bank acts as the lead bank in various aid areas, as will be seen. This change naturally helped reinforce the potential link between aid and economic reform, since the aid of the Bretton Woods institutions is, in principle, associated with reform to a greater extent than that of other donors. Changes in the Nature of Aid and Growing Indebtedness Like most Sub-Saharan African countries, Mali experienced a financial crisis in the early 1980s due to wide fluctuations in the terms of trade, which had improved in the late 1970s but deteriorated sharply thereaf- ter. The experience of the 1970s had led to an excessively expansive policy characterized by rapid growth of the money supply and in public ex- penditures. Thus, for the second time, Mali had to implement an adjust- ment policy with the support of the Bretton Woods institutions. This was a turning point in the evolution of the aid received by the country. The aid received by Mali since the early 1980s has had three main characteristics. The first is, of course, the increase in nonproject aid, es- Mali 239 pecially for adjustment. However, it is very difficult to isolate statisti- cally the aid flows that are akin to adjustment aid. (Under the heading "General Program Assistance," which is OECD's classification, appen- dix table 5.5, which deals only with ODA, indicates only a part of the nonproject aid that is subject to economic policy conditionalities. For example, it includes neither the IMF's contributions through 1998, which do not come under ODA, nor sectoral adjustment credits, which are in- cluded under other headings.) It should be noted that adjustment aid provided to Mali in the early 1980s came not only from the Bretton Woods institutions, although they predominated, but also from bilateral donors, and especially France through its aid agency or in the form of direct bud- get aid.5 It should also be noted that adjustment aid, which had been on hold just before the devaluation, was very high after that event (in 1994-95), and came as often from multilateral sources as from bilateral ones. But it seems that since then, multilateral aid (for example, from the IMF, World Bank, and European Commission) has predominated in this area. Second, the growing importance of technical assistance should be noted, as figure 5.4 shows. Based on trends in the overall aid volume, the proportion of technical assistance in total ODA fluctuated, finally ending up in the 1980s at a level above that of the two previous decades. During the 1994-97 period, it reached about 26 percent of total ODA (nearly 40 percent of bilateral ODA, but only 11 percent of multilateral ODA). The weight of technical assistance may in itself be a factor in the stronger link between aid and reform. Third, although grants are traditionally a large proportion of aid re- ceived by Mali (that is, usually more than 50 percent of official develop- ment finance, as shown in figure 5.5), this proportion dropped slightly in 1982 and 1983 due to the increase in other contributions. In general, the high proportion of grants has not prevented the growth of external indebtedness, which was particularly severe in the first half of the 1980s, as can be clearly seen in figures 5.10 and 5.11. Mali is currently counted among the heavily indebted poor countries that are eligible for HIPC initiatives. In late 1998, its total external debt totaled US$1,239 million in net present value, or 220 percent of export receipts for that same year (World Bank 1999a: 103, 106). Figure 5.10 shows how the debt ratio, expressed in face value, has evolved in rela- tion to GNP: the ratio had been around 100 percent between 1984 and 1993, and stayed at a little over 115 percent after the spike due to the devaluation. The growth of debt, whatever its origin, has increased Mali's de- pendency on the IMF, with which it had four standby agreements and three structural adjustment facilities between 1982 and 1998 (see ap- pendix table 5.9). The agreements were themselves a precondition for 240 Post-Socialist Reformers renegotiation of the debt with the Paris Club, which has restructured Mali's public debt four times since 1988 (see appendix table 5.10). Fi- nally, Mali's eligibility for the HIPC Initiative makes it especially im- portant for Mali to adhere to the agreements concluded with the IMF and the World Bank, and therefore increases the potential impact of aid on reform. It should be noted (see figure 5.9) that the evolution of net transfers hardly differs from that of raw aid flows, although the size of the differ- ence between these two flows naturally becomes more pronounced as indebtedness increases (it currently stands at around one-fifth). MALI'S COMMITMENT TO REFORM AND SUPPORT FROM EXTERNAL DONORS Mali's reform experience has been profoundly influenced by its point of departure. Let us recall that, almost immediately after independence, Mali embarked upon an ambitious socialist experiment characterized by state control of the economy and closure to the outside world, through, most notably, withdrawal from the franc zone (1962). The Malian expe- rience was also marked by a strong and traditional inclination toward regional integration. Without going as far back as the ancient empire of Mali, one must remember that, at the time of independence and during the brief period preceding the socialist experiment, Mali joined with Senegal in what was called the Federation du Mali. Once it reentered the franc zone in 1967, Mali then became involved in various attempts at regional integration in connection with trade (the West African Customs Union, UDAO, and then the West African Economic Community, CEAO) or resource management (the Organization for Development of the Senegal River, OMVS). As previously discussed, its entry into the West African Monetary Union in 1984 provided a durable framework for its policy of regional rapprochement, which was consolidated once again in 1994 with the creation of the West African Economic and Monetary Union. These two factors-the socialist experience of the 1960s and Mali's regional commitment-explain the long history and the pace of reforms undertaken in Mali, as well as the areas in which they have advanced the most (macroeconomic policy and economic liberalization) and those where they have been more uneven (public sector reform and institu- tional framework). The result is that external aid has exerted more influ- ence on reforms in some sectors than in others. Here we will examine in turn the pace of reform in Mali since 1980, according to the three "cycles" distinguished previously, looking spe- cifically at the reform of public finances, the reform of external trade, cereals market reform, the reorganization of the cotton sector, and, finally, Mali 241 other areas of public sector reform. As we shall see, aid sources have been closely involved. The Pace of Reform in Mali Since 1980 Although Mali experienced an initial reform phase upon its reentry into the franc zone in 1967, and with its first IMF agreement shortly thereafter, it was not until the early 1980s that the reforms were incor- porated into a general structural adjustment program onto which many sectoral programs were grafted. Mali's main donors participated in these programs. Since 1980, as was noted previously, three cycles of reforms can be distinguished, corresponding to their start-up (1980-87), acceleration (1987-93), and consolidation (1993-99). During each phase, reforms have dealt simultaneously with incentive structures and with public resource management. The 1980-87 Period IMPROVEMENTS IN THE STRUCTURE OF ECONOMIC INCENTIVES. The most important reform had to do with the cereals sector. The Cereals Market Restructuring Program (Programme de Restructuration des Marches Cer6aliers, PRMC) was set up, and the Malian government agreed to liberalize the cereals market (see discussion of the PRMC below). The distribution monopoly held by some public enterprises also be- gan to erode. SOMIEX (Societe Malienne d'Import-Export), which had held the monopoly on external trade since the 1960s, progressively lost it (it was abolished a first time in August 1986, and then reinstated in September). Its monopoly was restricted to 11 products. The monopoly of the cement plant was also eliminated. MANAGEMENT OF PUBLIC RESOURCES. Several measures were taken, although they were tentative in their content. They dealt either with the administration of public finances (including the reform of budgetary nomenclature), or with the control of expenditures (that is, a reduction in the operating expenditures of certain ministries, a freeze on promo- tions in the civil service, etc.). Finally, the first indications of government disengagement started to appear with the privatization of public enter- prises without any economic importance, namely a brickyard (SOBRIMA) and a tannery (TAMALI). The end of this period saw the beginning of the restructuring of the Banque de Developpement du Mali and of the airline Air Mali. In the latter case, the money-losing routes were eliminated and some aircraft sold off. 242 Post-Socialist Reformers The 1988-93 Period IMPROVEMENTS IN THE STRUCTURE OF ECONOMIC INCENTIVES. The main measures applied in this area during this period were the following: * Total liberalization of prices between 1988 and 1990, except for hydrocarbons; * Elimination of the monopoly on hiring held by the National Em- ployment and Manpower Office (Office National de la Main d'Oeuvre et de l'Emploi, ONMOE); * Revision of the Labor Code; * Updating of the Commercial Code; * Liquidation of SOMIEX (in 1988); * Liberalization of external trade, including most notably the elimi- nation of import/export permits (discussed below); * Simplification and streamlining of customs tariffs; * Replacement of the tax on goods and services by a value added tax (VAT). MANAGEMENT OF PUBLIC RESOURCES. In the area of budgetary policy, the wage bill did indeed dirninish due to the voluntary departure of some civil servants. Staff reductions continued through 1993. The differ- ent government funds (fonds de l'Etat), which constituted annexes to the government budget, were integrated and/or eliminated in accordance with the principle of the single cashier (unicite de caisse). In the area of financial policy, the restructuring of the Bank of Mali was completed and the Postal and Telecommunications Office was di- vided into three entities: the Postal Checking and Savings Funds (Societe des Cheques Postaux), the National Post Office, and the Telecommuni- cations Company of Mali (Soci6t6 des Tel6communications du Mali). The privatization of enterprises continued at a rapid pace. Enterprises had to be either liquidated, privatized, or rehabilitated. A time frame for this was developed under a special program: the Public Enterprise Sec- tor Adjustment Program (Programme d'Ajustement du Secteur des Entreprises Publiques, PASEP). A new program for the agricultural sector was initiated. It took into account the restructuring of the cotton production stream, of the Mali Agricultural Products Office (OPAM), and of the different development offices. The 1994-99 Period This was a period of consolidation and continuation of reforms already initiated. Mali 243 IMPROVEMENTS IN THE STRUCTURE OF ECONOMIC INCENTIVES * All prices were liberalized, including prices of petroleum products; * In terms of the regulatory framework, trade and administrative tribunals were set up and became operational in three regions of Mali (Kayes, Bamako, and Mopti); * External trade procedures were simplified through the establish- ment of the "one window" system (guichet unique) for handling all formalities; * In the context of support to the private sector, a center for handling business-related formalities was set up, and funds were made avail- able to the Private Sector Development Project (Projet de Developpement du Secteur Prive, PDSP) for the restructuring of Mali's Chamber of Commerce and Industry; * In the area of fiscal reform, a Large Enterprises Directorate (Direc- tion des Grandes Entreprises, DGE) was created to monitor enter- prises with turnover exceeding 200 million CFAF. In addition, the list of products exempt from the VAT was reduced and an overall reform of income tax was adopted. MANAGEMENT OF PUBLIC RESOURCES * The principle of the program budget was adopted. Monitoring of certain expenditures was improved. This mainly concerned the wage bill, for which the use of electronic databases became wide- spread, and also such budget items as student scholarships, water, electricity, and telephone. * Education, health, and road maintenance were recognized as pri- ority sectors as far as public expenditures are concerned. * In the monetary area, external adjustment resulted in devaluation of the CFA franc. * The privatization of enterprises continued and the government disengaged from all nonstrategic enterprises. * In the agricultural sector, a performance contract between the gov- ernment, the cotton company, and producers was signed in June 1994, and the various performance plans between OPAM and the government were renewed. Phases of Reform and Cycles of Aid It is interesting to compare the preceding description of the pace of re- forms with the aid cycles presented earlier. Unfortunately, we were not able to trace an economic policy indicator like the one established for 244 Post-Socialist Reformers various countries by Bumside and Dollar. There were two reasons for this: one is that the data are unavailable, which would explain why Burnside and Dollar did not establish such an indicator for Mali. The three indicators they selected and aggregated into an average indicator are not particularly relevant to Mali. Actually, little is known about Mali's inflation rate during the 1960s, and it also jumped in 1994 due to the devaluation after having been low for several years (its increase reflects the reform, or rather a reform, rather than the other way around). The openness indicator developed by Sachs and Warner, a unitary indicator equal to 0 or 1, is ill-suited to grasping the different phases of extemal trade reform as it will be described. The only thing that could be shown was the evolution of the third indicator-that is, the fiscal balance-but this is too partial to provide a synoptic view of economic policy. What is striking, on the other hand, is that each new phase of reform (1982-83,1987-88,1993-94) was accompanied or immediately followed by an increase in external aid. The slowing down of reform also corre- sponded more or less to the declining phases of the aid cycle. This paral- lelism is an indication that, in Mali, aid and reform have not evolved independently of each other. These apparent links at the overall level are examined in more detail in the sections dealing with the main sec- tors subject to reform. Reform of Public Finances The reform of public finances was an essential part of economic reform. It fulfilled both the objective of macroeconomic equilibrium (through a planned reduction in the budget deficit) and the objective of develop- ment, since fiscal reform was supposed to help improve the incentive system, whereas improvement in public resource management was aimed at restructuring public expenditures in favor of priority development- related expenditures. The reform of public finances mobilized most donors. The IMF and World Bank were the most heavily involved, especially through the struc- tural adjustment loan (ESAF) of 1992. This touched off numerous cofinancing packages from the German, Dutch, Swiss, and Norwegian governments, as well as from the European Union and the African De- velopment Bank. The United States and France also contributed to the program at the same time in the form of budgetary aid. The United States first provided budgetary aid in the context of the Economic Recovery Program (1987-88), and then under the Economic Reform and Develop- ment Program (Programme pour la Reforme Economique pour le Developpement, PRED) of 1991-97, which had its own conditionalities. The main measures taken in the area of public finance can be grouped into three categories of reform: budgetary administration, taxation, and the structure of public expenditures. Mali 245 Reforms to Improve Budgetary Administration During the first phase of reforms, the first task accomplished was the development of a clearer nomenclature for public expenditures. The most important efforts to improve financial administration were implemented from 1988 onwards, however. An important reform goal was to unify the government budget by integrating the supplemental budgets (bud- gets annexes) and special funds into it, and this was done in 1991 for 6 out of the 10 supplemental budgets or special funds. The intention was also to improve the monitoring and control of bud- getary procedures, especially by improving the efficiency of the office responsible for payments to civil servants (Bureau Central de la Solde). Unfortunately, the destruction in 1991 of the equipment and administra- tion facilities of the Ministry of Civil Service greatly slowed the comput- erization of personnel management. Since 1994, there has been better monitoring of the wage bill, as well as of other public expenditures such as student scholarships, water, electricity, and telephone service. Considerable progress has also been made in programming and moni- toring public investments. The national capacity for programming and monitoring still appears to be weak, however, especially at the level of the line ministries. Starting in 1994, emphasis was placed mainly on improving tax re- covery by strengthening the tax administration and by creating a system of unique taxpayer ID numbers. The Large Enterprises Directorate has been operational since 1994 and has proved effective. However, the com- puterization of fiscal services is still embryonic and much remains to be done to improve tax collection. Tax Reform Tax reform was undertaken mainly under the second reform phase. An important aspect of this reform consisted of replacing the tax on goods and services (imp6t sur les affaires et services, IAS) by a value added tax, thereby eliminating the IAS's spillover effects and export distortions, since the IAS did not provide for reimbursement of exporters for taxes paid at intermediate stages. Taxation of the informal sector was facili- tated by the introduction of the "advance on various taxes" (avance sur imp6ts et taxes, ADIT), which was set at 5 percent of imports and public procurements and deductible from indirect taxes. Anew investment code went into effect in June 1991. This reform of excise taxes obviously complemented the reform of import and export taxes, which was part of the simultaneous liberaliza- tion of external trade (see below). The reform of direct taxes had originally been planned for 1992. The idea was to institute a single private income tax and a single tax on 246 Post-Socialist Reformers business profits. This reform was postponed because the Malian au- thorities feared that it might result in a decrease in tax revenues due to their inability to manage a new tax system. The reform of direct taxa- tion was finally achieved in the budget of 1999 with the elimination of the income tax surcharge and the introduction of restructured schedu- lar taxes, including a tax on business revenues, the expansion of the global tax (impot synthetique) to include all businesses having an an- nual turnover of less than 30 million CFAF, and the introduction of a progressive tax on salaries. Restructuring of Public Expenditures The goal of the reform was to contain the wage bill so as to be able to increase spending on education, health, and the maintenance and de- velopment of infrastructures. A lasting cap on the wage bill could not be achieved with a freeze on civil service salaries and hiring alone, as was the case during the first phase of adjustment. A true civil service reform (supported by the SAC) was undertaken in 1989-90. It was aimed at reducing staff numbers through a program of voluntary departures and increasing staff efficiency by adopting new institutional guidelines (cadres organiques) and revising the compensation policy applicable to government employees. The voluntary departures were partially (60 percent) accomplished, but not really under the conditions that had been envisioned initially. Some civil servants who chose to leave came from the priority sectors of education and health. The results of the program for the reinsertion of those leaving (through the creation of businesses) were disappointing and contributed to an increase in the banking system's dubious loans. The civil service's institutional guidelines were called into question in 1991 by labor unions that had not been involved in their definition, and the new project did not become operational, since it would have involved an increase in the number of civil servants. In 1994 a new civil service statute was adopted, under which salaries are to be based more on merit and less on seniority. The devaluation of the CFA franc in 1994, which was followed by an increase of only 10 percent in salaries and a resumption in promotions (increasing the total wage bill by 18 percent), contributed greatly to a reduction in civil servants' remuneration in real terms. Since 1991, the proportion of expenditures devoted to education, health, and infrastructure maintenance has gradually increased. In edu- cation, however, a considerable share of funds was absorbed by second- ary and higher education (especially student scholarships), and this led to the cancellation in April 1993 of the last two tranches of the World Bank credit for education. The World Bank's new educational adjust- Mali 247 TABLE 5.1 MALI: BUDGETARY CONVERGENCE CRITERIA (percent of tax revenues) 1995 1996 1997 1998 1998 norms Salaries and compensation 36.5 28.9 28.9 27.0 40.0 max Internally financed capital expenditures 14.1 14.1 16.3 20.4 20.0 min Base cash balance 31.6 40.1 25.7 28.2 15.0 min Variation in internal arrears (billions of CFAF) -14.0 -19.0 -8.4 -4.0 0.0 Variation in external arrears (billions of CFAF) 0.0 0.0 0.0 0.0 0.0 ment (approved in 1995) includes a substantial program for the devel- opment of primary education and for the streamlining of secondary and higher education. Taken as a whole, all of these interventions, in terms of both receipts and expenditures, resulted in a widely fluctuating budget deficit. In terms of the percentage of GDP, the overall deficit (that is, the government's net financing requirement) experienced, from the beginning of the ad- justment period, three years of sharp increase (1985,1991,1994) followed by an effort at reduction. Although it reached 18.2 percent of GDP in 1983, it was brought back down in 1996-97 to less than 8 percent of GDP. Since 1996, Mali's budgetary policy has been subject to multilateral monitoring by WAEMU, whose council of ministers establishes its guide- lines. As can be seen from table 5.1, Mali has adhered to all of the bud- getary convergence criteria since 1998. Reform of Foreign Trade In 1988, Mali reestablished a relationship with the Bretton Woods insti- tutions and embarked upon a vast reform program, one of the most im- portant components of which was the reform of foreign trade. From independence to 1988, foreign trade had been essentially controlled by the government, with a rigid system of quotas and prohibitions. State enterprises had by that time revealed their shortcomings in production, imports, exports, and distribution. The reform was implemented in three stages. From 1988 to 1989 All export products received automatic authorization, which had not been the case previously. All import products received automatic autho- 248 Post-Socialist Reformers rization, except for sugar, flour, tea, oil, soap, rice, and pharmaceutical products. From the institutional standpoint, the "one window" approach to in- ternational trade was established to facilitate foreign trade procedures. From 1990 to 1991 Foreign trade documentation was modified. Instead of the import or export certificate, a declaration of "intent to import or export" was es- tablished, thus casting off the constraints of foreign trade operations and introducing a simple statistical registration that is more flexible. The requirement for the stamp of approval of the Foreign Exchange Office was abolished. The Societe Generale de Surveillance (SGS) was asked to become involved at the stage where merchandise is loaded. The SGS determines the quantity, quality, and customs value of all mer- chandise valued at over 3 million CFAF. At the same time, the following tariff measures went into effect at customs: * Elimination of the export tax, including the service charge (contri- bution pour prestation de service, CPS). * Elimination of all administrative prices and price schedules on imports and exports, with the exception of administrative prices for petroleum products. * Elimination of the tax imposed by the Price Stabilization and Regu- lation Office (Office de Stabilisation et de Regulation des Prix, OSRP) and its incorporation into the fiscal import tax (droit fiscal d'importation, DFI). * Elimination of the mining fund tax (taxe fonds minier) and of the petrostock tax (taxe petrostock). * Merging of the fuel import tax (taxe d'importation sur le carburant, TIC) and the special tax on certain products (impot sp&cial sur certains produits, ISCP). * Revision of DFI rates so as to reduce the number of rates to three (0 percent for exempted products, 10 percent for special products, and 25 percent for all other products). * Revision of the ISCP (or excise tax), which was thenceforth applied to cola nuts, alcoholic beverages, tobacco, cigarettes, vehicles, salt, photographic film, and petroleum products. * Introduction of a special tax on certain products (taxe de protection degressive, TDP), set at a minimum rate of 40 percent and intended to decrease by 10 percent annually, in such a way as to drop to 0 percent after four years. This was applied to the following prod- ucts: flour, palm oil, household soap, plastic bags, bottles, flasks, Mali 249 handicrafts, khaki fabric, printed fabrics, packaging materials, gal- vanized metal tubs and pails, household items, electronic batter- ies, pasta, cookies, plaster and matches. * Introduction of the flexible import tax (taxe conjoncturelle d'importation), the application of which was limited to rice and sugar. Since 1992 The final step in the reform of foreign trade was supposed to result in the abolition of the requirement for declarations of intent to import or export, and its replacement with simple customs declarations. This mea- sure could not be implemented due to reluctance on the part of the Eco- nomic Affairs Directorate to relinquish its prerogative to demand a declaration of intent. Customs tariffs were further simplified, to the extent that, by 1998, import tariff regulations distinguished only three categories of goods and three types of levy (customs, import tax, and the contribution pour prestation de service or CPS), as shown in table 5.2. The tariff had to be changed and simplified as of January 1, 2000 in the context of the WAEMU treaty. The Cereals Market Restructuring Program (PRMC) The Situation before the PRMC With regard to cereals, as in other economic sectors, marketing was still tightly controlled by the government in the early 1980s. Indeed, the Mali Agricultural Products Office (Office de Commercialisation des Produits Agricoles du Mali, OPAM), created in 1965, held the monopoly over cereals purchases from producer groups and over the sale of these cere- als in urban areas and in rural areas experiencing shortfalls. In addition, TABLE 5.2 MALI: TARIFF STRUCTURE BY RATE CATEGORIES IN 1998 (percent) Customs Import tax CPS Totalfees Category 1: critical products' 0 0 5 5 Category 2: inter-mediate products and equipment 5 10 5 25 Category 3: other products 5 25 5 35 a. Since August 1997, raw materials, intermediate products, and spare parts intended for industrial use have been taxed as critical products (category 1). Source: Data supplied by the Customs Directorate (table reprinted from Chambas and others 1999). 250 Post-Socialist Reformers the Societe de Credit Agricole et d'Equipement Rural (SCAER) had a monopoly on the marketing of agricultural inputs. Producer prices were at the same time controlled by the Price Stabilization and Regulation Office (OSRP), which used a fixed-rate estimate of the costs of transport- ing, warehousing, and marketing cereals to determine a cost price and then consumer prices, which could be subsidized. It was the same with the price of agricultural inputs. In reality, OPAM never succeeded in completely controlling the mar- ket (Lecaillon and Morrisson 1986). It seems to have handled half of it, at best. Since the towns and the deficit-suffering zones of northern Mali were imperfectly supplied, a private trade in cereals-initially prohib- ited and then tolerated-sprang up. Farmers participated all the more willingly in this trade as the official producer prices were set at an ab- normally low level. This policy led to an impasse. On the one hand, it had created a con- siderable deficit in the public and semipublic sectors, and on the other hand, production growth lagged behind that of demand. Thus, the policy contributed at the same time to a deterioration of public finances and in the balance of payments. This is precisely the diagnostic arrived at by the National Seminar of 1980, which was convened to consider a restructuring of OPAM and the cereals market. Its recommendations were therefore geared to the need to liberalize the cereals market, reorganize OPAM, and change pricing policy. In order to facilitate the implementation of these recommenda- tions, donors agreed to contribute 250,000 tons of cereals (maize equiva- lent) over five years. These cereals were to be sold by OPAM and the proceeds used to create a Joint Counterpart Fund (Fonds Commun de Contrepartie). Objectives and Phases of the PRMC The PRMC went through five phases, each of which was first validated by a national seminar. A joint platform was then developed and became the PRMC's frame of reference. PHASE I: 1981-87. In March 1982, a law "setting the rules for the cere- als trade" provided a basis for liberalizing this sphere of activity. OPAM's monopoly was thus abolished. Any merchant meeting the criteria of the Commercial Code was free to buy and sell cereals. From 1986 onward, such individuals were free to import cereals (with the exception of rice). OPAM's mandate from that time on was to regulate the market by creat- ing buffer stocks. It entered into purchase contracts with producer groups (at a price that constituted a floor for them) and resold cereals when their selling price reached a ceiling set by the authorities. OPAM's oper- ating expenses were to be maintained for five years at their 1981 level. Mali 251 The new price policy provided for a gradual increase in producer prices over five years. The purpose of this increase was to encourage production (with a view to food self-sufficiency) and to avoid clandes- tine exports by bringing prices closer to those prevailing in neighboring countries. The rise in producer prices implied a parallel increase in con- sumer prices. Because of the political risks, this rise was gradual and the merging of official and black market was also gradual. PHASES II TO IV: 1988-97. Phase II began in May 1988 with the liberal- ization of prices for millet, sorghum, maize, and rice, except for rice from zones where rice is grown on perimeters. Rice prices were freed up in the entire country in December 1989.6 These decisions had been preceded in 1987 by the elimination of all consumer subsidies. The next phases of the PRMC essentially dealt with reforming OPAM. In accordance with performance contracts signed in 1988,1990, and 1994, OPAM stopped subsidizing producer prices. Its role was limited to cre- ating buffer stocks (the volume of which was reduced), managing food aid, and providing cereals to areas experiencing shortfalls. OPAM un- derwent a considerable restructuring. It reduced its staff from 718 to 204, changed its salary policy so as to boost productivity, and strength- ened its financial management. Thanks to the PRMC and the favorable weather conditions of the pre- ceding years, Mali became nearly self-sufficient and food aid was sub- stantially cut back. Mali's cereals deficit zones were supplied mainly out of the country's own production. PHASE V: 1997-99. OPAM was by this time less and less dependent upon external aid. The government now finances a growing proportion of its operating costs out of its own budgetary resources, and in 2001 all external financing of OPAM should cease. Phase V had two main objectives. The first was to recognize and deal with emergency situations by establishing a system of enhanced food security (securite alimentaire renforcee, SAR). More specifically, the idea was to improve the efficiency of the existing system and to create new instruments for preventing major crises (e.g., a monitoring unit and emergency intervention plan). A second objective was to reduce food security risks over the me- dium term by improving the conditions in which cereals markets oper- ate. More specifically, the goals were to improve the information system, set up structures for directing production streams, develop new tools suited to a liberalized environment, develop a policy regarding prod- uct quality, and, finally, facilitate access to financing for professional operators. The restructuring of the market, now well underway, should open a modernization phase characterized by greater professionalism on the 252 Post-Socialist Reformers TABLE 5.3 RESOURCES FOR MALI'S CEREALS MARKET RESTRUCTURING PROGRAM, BY PHASE Phase I Phase II Phase III Phase IV Phase V (6 years) (3 years) (3 years) (3 years) (3 years) External resources 15,377 8,253 5,596 - 3,710 Malian government budget 0 0 0 - 2,810 Total 15,377 8,253 5,596 - 6,520 Annual average 2,563 2,571 1,899 - 2,173 - Not available. Sources: For phases I, II, and III, Coelos 1994; for phase V, Deme 1999: 36. part of operators and an improvement in the quality of products and services. External Supportfor the PRMC It is noteworthy that, during the implementation of these different phases, donors have provided the Malian government with substantial resources with which to carry out the reform. Table 5.3 shows the levels of these resources by phase. During the first three phases, the proportion of resources obtained from the sale of food aid dropped considerably, falling from 87 percent in the first phase to 45 percent in the third phase. Resources correspond- ing to monetary substitutions for food aid or to direct donor contribu- tions have evolved in the opposite direction: their relative weight went from 13 percent to 55 percent between the first and third phases (Coelos 1994). The PRMC and the Reform Process The PRMC has been a key tool in the liberalization of the cereals market. Its success in this area is essentially the result of: * The commitment of certain Malian authorities to this reform; * The determination of certain donors to see the reform succeed; and * The fact that the structural adjustment program (SAP) included among its conditionalities certain measures stipulated in the vari- ous joint donor/government platforms. The PRMC acted as a catalyst in the cereals sector, since it helped the Malian government to implement the SAP measures. This accelerated Mali 253 the liberalization of the cereals market within the larger context of the overall economic liberalization accomplished by means of the IMF's and World Bank's preconditions. The PRMC also mitigated social tensions over the 1985-88 period. It helped increase producer prices, stimulating production and maintain- ing consumer prices at reasonable levels. Reorganization of the Cotton Sector After having been in the hands of the Compagnie Francaise des Textiles (CFDT), the development of cotton growing has since 1975 been entrusted to a mixed enterprise (40 percent CFDT, 60 percent Malian government). Its mandate was very broadly defined as the processing of seed cotton into cotton fiber, as well as rural development (and not only cotton de- velopment) in its zone of intervention, Mali Sud. From the outset, the Malian Textile Development Company (CMDT) has performed impor- tant public service functions such as extension, credit, and transporta- tion, but its performance has been closely associated with that of other organizations involved in this production stream, including SOMIEX, which handles cotton fiber exports, and OSRP, which is responsible for stabilizing the government-determined price paid to producers. This latter entity collected cotton revenues and paid the CMDT and SOMIEX on a fixed-rate basis. Neither of these companies was concerned with the equilibrium of the cotton sector, nor did they have any incentive to reduce costs and increase productivity. Since cotton represents 50 percent of Mali's exports and is the liveli- hood of more than 20 percent of the population (Laporte and Monier 1994), the abrupt drop in international cotton prices in 1985 radically modified Mali's economic and financial prospects by causing a large operating deficit in the production stream, which was partially covered by tax exemptions that contributed to reduced budgetary inflows. The cumulative deficits of the 1985-86 and 1987-88 cropping seasons amounted to more than 2 billion CFAF. The seriousness of the crisis forced the Malian authorities to undertake an in-depth reorganization of the production stream to increase its productivity, a move that was all the more crucial because Mali's membership in WAMU made it impossible to use the devaluation to increase productivity by means of a relative price movement. In 1986, and not only because of the problems with cotton described above, SOMIEX's monopoly was abolished and the CMDT took over cot- ton fiber exports. Reorganization of the production stream was the sub- ject of the "Selingue seminar" that gathered together Malian authorities and donors whose proposals were accepted by the government in No- vember 1988. This made it possible, after various preparatory studies, 254 Post-Socialist Reformers for the government and CMDT to sign their first performance contract in 1989, and a second in 1994. In order to increase the production stream's productivity, the perfor- mance contract stipulated that the CMDT must become an autonomous enterprise assuming financial responsibility for its industrial and com- mercial activities. Therefore, the CMDT from that time on collected the proceeds from cotton sales and was no longer subject to any export tax. However, the mechanism for setting producer prices and sharing profits (see below) stipulated that CMDT's fees must be capped and that over- runs must be debited from profits, while reduced costs were to be a gain for CMDT. On the other hand, CMDT was only obligated to perform tasks related to rural development to the extent that the government provided the corresponding funds. These public service tasks were clearly spelled out in an appendix to the performance contract. From both the social and economic standpoints, it also seemed desir- able for cotton producers to benefit from a purchase price that was prof- itable (and in keeping with the international price), as well as stable. For this reason, a system of price stabilization was created according to the following principles: * A minimum producer purchase price was defined. * To this minimum price, CMDT's stabilized operating expenses were added, along with a portion of amortizations, to arrive at an equi- librium price of cotton fiber for export. * Profits resulting from an export price above the equilibrium price as defined above were then to be divided four ways, according to a rule established in advance, among the stabilization fund, the pro- ducers' rebate, CMDT, and the government. * If the export selling price was below the equilibrium price, the sta- bilization fund covered CMDT's minimum expenses. In the event that the fund experienced a lasting shortfall, the plan was to lower the floor price paid to producers (a situation that occurred in 2000). It should be noted that institutional reform in this production stream has certainly been facilitated by the important donor-financed develop- ment project (Mali Sud III) in the main cotton growing area. This project's economic objectives are ambitious and the financial stakes are high (24 billion CFAF, of which 16 billion in foreign aid from France, the World Bank, and the Netherlands). The aim is primarily to promote the diver- sification of agricultural production and natural resource management, while at the same time preserving land capital for the future, providing training and support to village associations, and, finally, developing new types of management advisory services and methods of disseminating agricultural techniques. Mali 255 The combination of the cotton sector reform and the development of the Mali Sud project was deemed satisfactory by a World Bank assess- ment in June 1998. Other Areas of Public Sector Reform: Varying Degrees of Success In addition to the three main sectors of reform that have been described, a brief account is given below of three other areas involving the public sector, where progress has been uneven: the reform of the numerous offices created in the past, privatizations, and the reform of judicial institutions. Reform of Public Offices Under the socialist policy pursued in the 1960s and 1970s, numerous public offices had been created to manage the various sectors of eco- nomic activity. The offices that were performing inefficiently were gen- erally unprofitable. One of the essential policy elements was then to reform them or abolish them. Much has been done in this regard in vari- ous ways. First, the reform of the Rural Development Agencies (Operations de Developpement Rural, ODR) has led to a substantial reduction in their share of the state budget (World Bank 1998b). OTER (Rural Equipment Works Operation) was privatized and became profitable. ODIPAC (Grain and Groundnut Integrated Development Agency) and ODIK (Kaarta Integrated Rural Development Agency) have been abolished. ORS (Segou Rice Agency) has been restructured (unprofitable activities discontin- ued, action limited to training activities financed from water fees), as has OPSS (Improved Seed Development Agency), so that seeds are no longer produced by the office but by farmers under contract, also with cost recovery. Finally, OHVN (Upper Niger Valley Rural Development Agency) has also been reorganized and has increased its productivity. It should be noted that the cost of these restructurings and in particular the compensation paid to staff who were dismissed was largely financed from various sources of aid. A more important, older, and in a way symbolic department is the Office du Niger, created in 1932. It also underwent considerable restruc- turing in the mid-1980s, and some of its activities were transferred to the private sector. Finally, in 1995, a performance contract between the state, the Office du Niger, and the producers confirmed the liberalization of the sector covered by the office. It is now responsible for conducting public service missions (maintenance of secondary roads and advice on rural matters) and for supplying one service-water, for which users pay the actual cost. There has been new private involvement upstream 256 Post-Socialist Reformers and downstream of rice production, which increased considerably dur- ing 1984-90, as soon as the reform was initiated (with a smaller increase during the first half of the 1990s because of the sharp decline in the price of rice). The reform of the Office du Niger was financed with the help of the principal donors in Mali. Privatization As a legacy of the socialist experiment, Mali had a number of public enterprises, including some in the productive sector. Here too, the inef- ficiency and shortcomings of these enterprises required disengagement by the state, which proved to be more difficult than for the Rural Devel- opment Agencies. It was perhaps more difficult to restructure these en- terprises, to compensate the staff who were dismissed, and to find buyers on terms acceptable to the Malian authorities. In any case, the privatization of the Mali Electricity Company (EDM) was not accepted until a major crisis affected electricity supply (see be- low). The privatization of SOTECMA (the telecommunications company) was postponed, and in the case of SONATEM (National Tobacco and Match Company), which had been seriously mismanaged in the past, foreign capital participation amounting to only 49 percent is currently envisaged. Difficulties remain for the H6tel de l'Amiti6 and-more im- portant-for BMCD (Mali Bank for Credits and Deposits). Restructuring of Public Spending and Promotion of Education One area that is essential for the long-term development of Mali-edu- cation-is still lagging behind. The main aim of the government's long- term (10-year) development strategy for education, called PRODEC, is to achieve a considerable improvement in primary enrollment rates, particularly by increasing the share of the budget allocated to education (26 percent in 2000, compared with 23 percent in 1997). However, dis- agreements within the government have delayed the implementation of this plan, to which various sources of external financing are ready to provide assistance. Reform of Judicial Institutions In this domain, where much remains to be done, external assistance seems to have had some effect: first in the case of PRED, the Economic Reform and Development Program, financed by the U.S. Agency for Interna- tional Development (USAID) in 1991; second, through French technical Mali 257 assistance within the framework of the Organization for the Harmoni- zation of Business Law in Africa (OHADA); and finally, as a recent con- ditionality for adjustment assistance (that is, the 1996 ESAF). As far as the judiciary is concerned, the aim of PRED was to make the investment environment healthier and more stable by creating commer- cial and administrative courts and studying the advisability of estab- lishing an arbitral tribunal to settle disputes in a manner appropriate to the culture of Mali. During 1994, therefore, PRED assisted the govern- ment in the following activities: * A decree concerning the election and eligibility of members of com- mercial courts. Following this decree, the overall intellectual level of the judges improved, and the function was opened to categories of persons who previously were excluded, such as haulage con- tractors, bankers, industrialists, insurers, and brokers. This made it possible to have more specialized commercial courts. * Adoption of requirements for proper functioning of the courts. To deter any behavior that might impede the correct functioning of the courts, the National Assembly adopted a disciplinary system similar to that applicable to presiding judges. Reference is made to the obligation of due diligence, in order to guarantee speedy con- sideration of cases. * Equipment of the commercial courts, administrative courts, and Supreme Court. Typewriters and photocopiers have already been delivered. Computers have been provided to the Supreme Court and the software will be provided for legal databases. * Adoption of a new Code of Civil, Commercial, and Social Proce- dure. The code in force dated from August 18, 1961. It was out of date and had many gaps. To facilitate access to the courts by liti- gants, the principles governing deposits were reviewed thoroughly. The plaintiff is now no longer required to deposit 7 percent of the amount of his claim. In order for his pleas to be receivable, he will simply have to pay a lump sum ranging from 5,000 to 25,000 CFA francs, depending on the amount of the claim. The amount of the deposit was reduced to 3 percent. * Creation of administrative courts at Kayes, Bamako, and Mopti. * Ratification of the New York and Washington Conventions so as to attract investment (the former relates to the recognition and en- forcement of foreign arbitral awards and the latter relates to the settlement of investment disputes). It is noteworthy that USAID gave substantial financial assistance to the government through this project. 258 Post-Socialist Reformers Five years later, despite the establishment of OHADA, private ob- servers have found that the judiciary is not yet functioning properly. This is why the ESAF is again including it in its program for 1998-99, with the aim of creating commercial courts, effectively applying the OHADA arrangements, expanding the training and inspection of judges, developing a code of ethics for judges to be applied as of 1999, and try- ing strictly to apply the sanctions against judges who violate the code. This text, which corroborates the information obtained in Mali, indi- cates how much remains to be done if the judiciary is to function prop- erly. Indeed, this is the aim of the ten-year program (PRODEJ) being implemented in this area, where it is still too soon to assess the effective- ness of the reform. WHEN DOES AID FACILITATE REFORM? FOUR LESSONS FROM THE EXPERIENCE OF MALI The preceding analysis of aid and reform trends shows that the relation- ship between them is not simple. Aid and reform have undergone fluc- tuations that are apparently simultaneous but that do not indicate a one-way causality: aid may either spur reform or respond to reform. Even when donors cannot spur reform, they want to be able to demon- strate the effectiveness of their aid in this area. In Cocteau's famous words: "The scope of these mysteries is beyond us; let us pretend that we orga- nized them." Donors may also quite simply want to reward reformers and perhaps thus encourage other countries to follow suit. However, the experience of Mali does seem to show a real influence of aid on reform, but with both successes and failures. It appears that aid plays an important but not always decisive role. Its decisiveness depends on whether the political authorities and the local administration 'own" the reforms. And the factors conducive to this ownership are precisely the factors that ensure the effectiveness of aid for reform. Four conclusions, based on Mali's experience, can be drawn from this general principle: * Aid is more likely to facilitate reform if a critical situation has re- vealed the need for it. * Aid is more likely to lead to reform if the stakeholders are aware of its usefulness or if there is scope for mediation between opposing interests. * Aid is more likely to facilitate reform if there is no disagreement between the various donors on the details of the reform. * Aid is more likely to facilitate reform if it considerably reduces the costs incurred. Mali 259 "Back to the Wall": Reform Facilitated by Aid in a Crisis The first lesson suggested by the Malian experience is that aid is more likely to facilitate reform if the economy as a whole or the sector con- cerned is in a critical situation. In a crisis, the donors' advice is more easily heard and the drawbacks of not reforming are more obvious. The crisis, whatever its costs, convinces all groups concerned of the need for reform and facilitates ownership of it. Several examples from Mali may illustrate this point. Unpopularity of Domestic Arrears It has been seen above that an important aspect of reform concerned the reform of public finances and particularly reduction of the budget defi- cit. This deficit reduction has been achieved gradually, largely through a payroll decrease resulting from downsizing and a decrease in real civil service wages. Paradoxically, the commitment of the Malian authorities to take the measures that resulted in this improvement was facilitated by the serious consequences of the state's financial difficulties and in particular the arrears in domestic payments, including a large backlog of wages. On three occasions, domestic public arrears were reduced under a program financed by assistance: from 1983 to 1986, from 1988 to 1991, and again after 1994. In 1988, the backlog of civil service wages ranged from three to eight months, depending on the sectors and geographical areas. The trade union grouping itself was then in favor of reform: better to be sure of receiving reduced wages than to face uncertainty in the payment of nominally unaltered wages! Effect of International Prices A second example is the impact of the decline in the international price of the principal export, cotton. In 1985-86 and again in 1991-92, the in- ternational price of cotton dropped sharply. Because of its repercussions on the whole sector, this decline created a need for widespread restructur- ing: in 1986, responsibility for cotton exports was taken away from SOMIEX and given to CMDT, and in 1989 a new performance contract defined the obligations of CMDT and created a new stabilization machinery. These reforms in turn made it possible to increase and stabilize the real price paid to the producer and achieve a sizable increase in production. Energy Shortage and Privatization The reform of the public enterprise responsible for the distribution of water and electricity has for the last 20 years been the subject of lengthy 260 Post-Socialist Reformers discussions between donors and the Malian authorities. The situation of the Societe d'Energie du Mali (EDM) is characterized by recurring diffi- culties in balancing electricity supply and demand, resulting in frequent power cuts during the dry season and serious financial difficulties. The Malian authorities were strongly opposed to privatization, and in 1995 they agreed to the idea of a "global management delegation" entrusted to an international consortium (composed of SAOM Interna- tional, Electricit6 de France, and Hydro Quebec International, and fi- nanced by the World Bank, the Canadian International Development Agency, and Agence Francaise de D6veloppement). However, since they did not really own the decision, the Malian authorities found themselves in ongoing disagreement with the management delegation. This under- mined the delegation's actions and then led Mali to denounce the agree- ment at the beginning of 1998. In April 1998, the Malian government and the donors agreed on the need to privatize the enterprise and on ways of preventing the financial situation from deteriorating too seriously, with reallocation of certain external funding. However, it seems that the steps toward privatization were taken reluctantly by Mali. It was only in February 1999 that a com- mercial bank was recruited for this purpose. In the spring of 1999, however, the electricity shortage reached un- precedented proportions, with extended power cuts during the driest season. This had dramatic consequences for many small businesses, as well as for people's living conditions. This acute shortage was attribut- able to poor maintenance of the power stations and to the questionable choice of a combustion turbine with delays that obliged EDM to empty the Selingue dam completely in May. In addition, on April 27 the control panel at one of the two power stations caught fire, probably because of poor maintenance. In the face of this critical situation, there was a change in the Malian government's willingness to undertake a radical reform and in the will- ingness of the population to accept one. In October 1999, at a meeting with all the donors (including IMF), Mali established a definite time- table for privatization. At the same time, it accepted technical assistance for the interim management and received aid for the purchase of a new power station, so that electricity supply could be restored and privatization would not be jeopardized by the lack of a buyer. But in this case it was definitely the crisis that enabled aid to spearhead reform. Decline of State Structures and Liberalization of the Cereals Market Other examples could be given to illustrate this point. In particular, the restructuring of the cereals market was easier to implement in 1981 be- cause farmers were no longer selling to OPAM, which was supposed to Mali 261 have a monopoly, while cereals production was increasing. Earlier the drought in the 1970s and its dramatic consequences had also illustrated the need for far-reaching reform in this sector. The people involved at the time admit that it is not clear whether it was the Malian leaders or their partners who had the idea of the reform which would gradually free prices. In any case, the time was ripe. Consensus or Mediation: Reform Facilitated by Aid When Interests Can Be Reconciled Whatever aid is provided, internal political factors are crucial if the re- form is to be effective. Economic policy advice is more likely to be well received if the proposed reforms appear not to disfavor a large segment of the population. Mali's success story, the Cereals Market Restructur- ing Program, also illustrates this concept: each stage of the reform was validated by a "national platform" at which producers and consumers expressed their views. However, it is not always so easy to reconcile interests and a different approach was adopted depending on whether the prevailing regime was authoritarian or democratic. Two Opposing Examples: Liberalization of Foreign Trade and Privatizations Two opposing examples may be used to illustrate this point: the reform of foreign trade and privatizations. It is noteworthy that in these two areas reform ran counter to the country's initial policy choices favoring a dominant role for the state in the economy. Despite this ideological reluctance, liberalization proved to be much easier than privatization. This is precisely because there were fewer interests opposed to liberal- ization than to privatization. During the period 1988-93, the trade liberalization measures were welcomed not only by the government but also by the technical depart- ments responsible for implementing them (in particular the National Directorate of Economic Affairs, DNAE), by the economic operators, enterprises, and traders, and by consumers, who immediately under- stood that they would benefit. In this way, for example, the monopoly of SOMIEX, which was losing one billion CFA francs a year, was finally brought to an end. All the stakeholders were convinced that the liberal- ization of distribution would benefit the state, traders, and consumers, even if the company's employees would suffer. Even in this area, however, some doubts were expressed and re- sulted in backtracking. This was so, for example, in the case of SOMIEX, as well as with the abolition of the export tax which was a conditional- ity for the USAID-funded PRED. This tax, which was abolished in 1990 262 Post-Socialist Reformers so that the scheduled aid could be disbursed, was restored and then abolished again in 1997 under pressure from USAID. The large-scale privatization program desired by donors generally encountered obstacles of two types (in addition to ideological resistance). One was connected with the vested interests of the employees of the enterprises to be privatized. The other had to do with the multiplicity of government departments involved. When the privatization or liquidation of a public enterprise is an- nounced and it is not possible to compensate the workers rapidly, there is a risk of collusion between managers and workers in the enterprise to aggravate the situation of the enterprise and make its privatization more difficult. The disparity in the compensation offered to workers in public enterprises (ranging from 12 to 36 months' wages) was a source of frus- tration and led certain dismissed workers to sue the government (in 1995, the government had been required by the courts to pay 24 billion CFA francs). While DNAE was the only technical body responsible for trade liber- alization, the office of public enterprises responsible for privatizations is obliged to consult the technical departments of the Ministry of Agricul- ture, the Ministry of Industry, or the Ministry of Tourism, depending on the sector corresponding to the enterprise to be privatized. In addition, whereas trade liberalization involves only a decision by the government, privatization requires the approval of the National Assembly. This natu- rally slows down the process, but makes it more difficult to reverse. There are, however, two cases in which the restructuring of public enterprises succeeded (the Office du Niger and the Office of Hydrocar- bons), thanks to the availability of effective technical assistance and the necessary funding. Reconciliation of Interests under an Authoritarian Regime and under a Democratic Regime The role of aid in overcoming opposition to reform, whether motivated by vested interests or by principle, differed depending on whether the regime was authoritarian or democratic-in other words, before or af- ter 1991. When power was essentially concentrated in the hands of the head of state, it was he who mediated between his proreform ministers (gener- ally the ministers of finance and planning), who were more frequently in contact with the Bretton Woods institutions, and the more conserva- tive ministers, who were often in charge of sectoral departments. In this case, the input from outside in some sense bolstered the proreform posi- tion, both because it provided arguments in favor and because of the financial benefits of reform. Mali 263 According to the people who were responsible for the initial reforms, in the 1980s, it seems that many reforms were suggested by the Bretton Woods institutions (for example, introduction of VAT, abolition of Air Mali, increased flexibility in labor law) and that the proreform ministers advocating these ideas were supported intellectually by the donors, although within the party these ministers were accused of being their "yes men." In this political context, the financial benefit of reform often tipped the scales in its favor at the level of the head of state. This was the case when meat prices were to be decontrolled, against the advice of the min- ister of animal husbandry (in 1988), and also when the monopoly of SOMIEX, after having been suspended following the president's meet- ing with the SOMIEX workers in August 1986, was reinstated by the same president the very next month. This was the context in which an IMF negotiator told the minister of finance: "Your country has intro- duced a number of reforms, but it is one of the countries least support- ive of reform." Starting in 1991, democratization somewhat modified the influence of vested interests on the role of aid in the promotion of reform. Democ- ratization must naturally promote the ownership and internalization of reforms. At the same time, however, the experience of Mali has shown that it slows down this process when certain interests involved are irrec- oncilable and there is not sufficient mediation capacity at the govern- ment level. Two examples can be given of this risk of aid blockage and ineffectiveness. The first example, which is especially important in view of the low educational level in Mali, is the Program for the Development of Educa- tion (PRODEC). After having been negotiated at length by all donors and Malian stakeholders, and although substantial financing had already been mobilized, this program was blocked because of a dispute between the two ministers responsible for education. Another example, less far-reaching but highly symbolic, was the "pink market"-the central market of Bamako. After having burned, this mar- ket was faithfully reconstructed using a donation from France (AFD). More than a year after completion of the reconstruction, it remains un- occupied because of the lack of a policy decision on the method of as- signing spaces to traders, who cannot all be accommodated, and lack of a policy of space management. In the meantime, traders and their cus- tomers are crowded in unhygienic conditions around the empty space reserved for the market. The "pink market syndrome" is emblematic of a more general diffi- culty, noted by several observers, resulting from the slow pace of decisionmaking, even in areas that are crucial for the development of Mali's economy. While the social benefits of consensus have been fully 264 Post-Socialist Reformers understood, the same cannot be said of the social cost of delays and failure to take decisions. In this context, aid seems to be effective in pro- moting reform only if it is combined with local capacity to mediate be- tween conflicting interests. Donor Attitudes: Reform Facilitated by Aid When the Sources of Aid Are Permanently Involved and Coordinated The third factor determining the contribution of aid to reform is the ac- tual attitude of the donors. Here there appear to be two essential ele- ments. One is the donors' ability to become involved themselves and to involve their Malian counterparts in the reform. The other is their ability to coordinate their advice. It is in the social sector that the need for projects to be prepared in close collaboration with the Malians is most acute. One example is a health, population, and rural water project in the Kayes region financed by the World Bank in 1985 (Kayes-Bafoulabe-Kenieba project). The goal was to improve access to drinking water by drilling wells and opening health centers, applying the principle of "cost recovery." The project failed because it was not supported by the people. The financed infrastruc- tures rapidly deteriorated and the centers closed. A second example may be taken from the customs sector. Because of the importance of this category of revenue for the Malian government, donors have tried constantly to improve the collection of customs du- ties. In 1994 the European Union financed technical assistance to the customs offices, but the results were not up to expectations. On the other hand, in 1995 the Canadian International Development Agency launched a Program for the Improvement of Customs Revenue in close coopera- tion with the Ministry of Finance. At the request of the customs admin- istration, arrangements were made to improve the working conditions of the relevant staff, particularly by constructing suitably equipped cus- toms offices across the country and by motivating and training customs officers. Partly as a result of this program, revenue increased from 5-6 billion CFA francs a month in 1995 to 11-12 billion a month in 1999. Another example of the weak effect of aid on reform is provided by the courts. Their reform is one component of the PRED initiative financed by USAID from 1992 to 1997. Apparently the project originated when a Malian lawyer contacted the American aid agency. It was a far-reaching project, since it comprised both reform of laws and regulations and in- vestment in the equipment of the courts (with funding amounting to US$7 million) and seemed to be sound. However, it was designed with- out the initial participation of the Malian side and implemented by a private U.S. firm that does not seem to have worked sufficiently directly Mali 265 for the Malian government. The functioning of the commercial and ad- ministrative courts apparently still leaves much to be desired, because of the reluctance of the private operators to criticize their peers. Once again, a different example is provided by PRMC, since its suc- cess in cereals market restructuring is also due to the continuing support of external financing sources and to the ongoing presence of a technical assistant who was totally involved in the program. Another efficiency factor to be considered is the coherence of the ad- vice given by the various aid sources. When there are differing points of view, these may be echoed within the administration, resulting in post- ponement of the reform. It is sometimes claimed that disagreement among donors gives recipient countries more freedom. In fact, however, it delays financing and reform. Two examples illustrate this risk. The first concerns the program for the improvement of the road infra- structure, an effort that is essential in order to make Mali more acces- sible and one to which several of the main sources of financing (European Union and World Bank) are prepared to give considerable support. The sticking point is the creation of a road fund to be financed from specific revenues and used to guarantee proper road maintenance. The World Bank and the European Union favor the creation of such a fund; but IMF, which for several years has been persuading the Ministry of Fi- nance of the benefits of a single cashier at the level of the state, is op- posed, fearing that the fund would set a precedent for other sectors (such as mines). Another example of disagreement relates to the status of CMDT, the conditions in which producers may participate in the capital of the com- pany, and, more generally, how to introduce more competition in the sector. The difference of opinion involves, on the one hand, the World Bank, and on the other, France and other European donors, especially the Netherlands. Here too, the stakes are high, since they involve the country's principal export activity, which was one of the main levers of Mali's growth, and since today CMDT seems to be in a worse financial position. Consequently, the negotiation of the new performance contract between CMDT and the government, which was to be signed in Sep- tember 1999, has fallen behind schedule. This relatively recent difficulty does not detract from the fact that in the past the successful implementa- tion and outcome of the southern Mali operation, in which CMDT played an essential role, benefited greatly from the coordination among donors (Agence Francaise de Developpement, World Bank, Banque Ouest- Africaine de Developpement, African Development Foundation, Banque Europ6enne d'Investissment, the Netherlands, Guinea, and Belgium). Together with the past experience of the southern Mali project and the cotton sector, PRMC once again provides an example of good coor- dination of aid. 266 Post-Socialist Reformers Lastly, it should be noted that, in budgetary matters, the leadership of IMF, whose support for reform reflects the goals of multilateral monitor- ing of public finances pursued by WAEMU, was instrumental in the re- duction of the budget deficit and the restructuring of public spending. Generally speaking, it seems useful if in a number of areas one exter- nal aid institution can play a lead role and be more particularly respon- sible for monitoring conditionalities. The very large number of aid sources and the sometimes excessive number of conditions laid down for dis- bursement considerably complicate the task of the Malian administra- tion and mean that it has to spend an inordinate amount of time responding to donors' conditions. In addition to the 30 or 40 conditions for the structural adjustment loan, there are the conditions for the sec- toral programs (as for agriculture or public enterprises) or specific pro- grams (PRMC, PRED). Ideally, all these conditions should be monitored by the Ministry of Finance, which is responsible for mobilizing financial resources. Certain conditionalities are not respected because of lack of monitoring: the Ministry of Finance concentrates essentially on respect- ing the commitments given to major donors (OECD 1998). One way of limiting the number of conditionalities is donor coordi- nation, which has developed considerably in Mali in recent years, par- ticularly in major sectors, for programs such as PRODEC and PRODESS (Integrated Health Sector Development Program), and through regular organized consultations. Coverage of Costs: Reform Facilitated by Aid When Aid Covers Some of the Costs Mali provides several examples of cases in which, by covering part of the costs of reforms, external funding has been a decisive factor in their adoption and success. For instance, food aid supported the launching of PRMC, thus avoiding a situation in which liberalization is accompanied by an excessive increase in food prices, which would have jeopardized continuation of the program. Another particularly instructive example is the aid provided by USAID under programs of economic reform. In order to promote the develop- ment of the private sector, USAID assumed for one year the temporary shortfall resulting from the reduction (from 15 to 7 percent) of the lump- sum contribution payable by enterprises. A similar example is the financing (by the World Bank and USAID) of the termination payments due when staff are to be reduced. On the other hand, the restructuring of public enterprises and, conse- quently, privatizations have been hindered by the lack of funds to pay prompt compensation to dismissed workers. A fifth lesson could be learned from the Malian experience. In a way, it intersects with the four others and echoes the initial principle stated at Mali 267 the outset: aid truly contributes to reform only if it is owned in one way or another by the leaders of recipient countries. Indeed, the Malian experience can be quoted in support of a condi- tionality of performance to replace the conditionality of means that is still the rule today (on the principle of conditionality of performance, see Guillaumont and Guillaumont 1994, 1995; Collier and others 1997; Guillaumont and Chauvet 1999). The Malian authorities have shown a keen interest in this new way of thinking. CONCLUSION: RETURN TO INITIAL ASSUMPTIONS These lessons of the Malian experience provide some kind of response to the three questions that the Aid and Reform in Africa study set out to address. First, did the Malian governments decide to reform their economic policies independently of the aid relationship? Clearly they did not, but the answer differs depending on whether the regime was the authoritar- ian one in power before 1991 or the current democratic government. For the reforms that took place before 1991 and which, as noted, were far- reaching, aid strengthened the arguments of the proreformers within the government, in the context of the financial constraints facing the government. During the democratic period, the role of external aid in the reform process changed: aid also contributed to reforms because it reduced their immediate cost to the population. In a way, democratiza- tion made the traditional conditionality less effective, since reform re- quires greater ownership by the political authorities. Second, does nonfinancial aid have a greater impact than conditional financial aid on the implementation of reforms? The answer to this ques- tion depends partly on the answer to the first question. If financial aid were indeed ineffective, then nonfinancial aid would very likely be more effective. However, since we have decided that financial aid is quite ef- fective in the preparation of reforms in Mali, the question concerns the effectiveness of nonfinancial aid per se. In Mali, such aid has taken two forms: traditional technical assistance and institutional arrangements. Technical assistance undoubtedly plays an important role in Mali, in view of the administrative shortcomings in that country. During the mission, it became apparent that its effectiveness depended essentially on the quality of the staff, the ongoing nature of their presence in Mali, and the nature of their functions. Two contrasting examples have been given above (PRMC and PRED). Institutional arrangements themselves have a considerable influence on the adoption of reforms and their irreversibility. In this connection, it is extremely important that Mali belongs to the franc zone and even more that it was a member of WAMU and is now a member of WAEMU. This 268 Post-Socialist Reformers was important not only for monetary policy but also for the manage- ment of public finances and the easing of foreign trade restrictions. Con- cerning the latter point, the creation of a regional common market offers some guarantee that reform will be lasting. Finally, did financial aid work better when reforms of economic policy and capacity building were already underway? The answer to this ques- tion is tricky and is not clearly affirmative in the case of Mali. The far- reaching reforms implemented over the past twenty years initially collided with a strong socialist and interventionist tradition inherited from the 1960s. This is why donors often seem to have taken the initia- tive in reforms and to have played an important role at the beginning of the process. Conceivably, financial aid should again become effective under the current democratic political regime and in an economy in which most restrictions have been lifted. However, as noted, if aid in this new con- text is to contribute effectively to reform and development, it must un- doubtedly provoke a more rapid and more far-reaching mobilization of political leaders and the administration. This could result from a new type of conditionality, geared more to performance than to instruments. Mali 269 APPENDIX 5.1 SUMMARY OF DATA FOR COMMENTS ON AID CONCEPTS IN MALI FIGURE 5.2 MALI: TOTAL NET OFFICIAL DEVELOPMENT ASSISTANCE Millions of 1996 constant U.S. dollars 800 600 Source: OECD-DAC. FIGURE 5.3 MALI: AID INTENSITY, 1975-97 Percent Percent 250 35 30Net ODA as a 30 200 - Net ODA as a ( pright axis 25GN percentage of gross \(rgtai) 2 domestic investment/ /> 1 00 ( axis) 20 100'~~~~~~~~~~~~~~~~~~1 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~- 1 50 - 35 - - Net ODA as a percentage of rcen5 imports of goods and services (left axis) 0 0 Source: World Bank, World Development Rep rt 1999. 270 Post-Socialist Reformers FIGURE 5.4 MALI: OFFICIAL DEVELOPMENT FINANCE (DISBURSEMENTS), 1970-96 Millions of 1996 constant U.S. dollars 800 r Technical assistance grants 700 - = Nongrant component of official loans - Grant equivalent of official loans 600 - Non-technical assistance grants 500 400 300 200 AI 100 0~~~~~~~~~~~~~~~. Note: Public lending includes concessional loans from the IMF (SAF and ESAF), as well as nonconcessional loans (enhanced facility and standby agreement). Source: Chang, Fernandez-Arias, and Serven 1999. Mali 271 FIGURE 5.5 MALI: COMPOSITION OF FLOWS OF ODF (DISBURSEMENTS), 1970-96 Millions of 1996 constant U.S. dollars 800 a Grants (including technical assistance) - Loans 700 600 500 400 300 200 100 Note: Public lending includes concessional loans from the IMF (SAF and ESAF), as well as nonconcessional loans (enhanced facihty and standby agreement). Souirce: Chang, Fernandez-Arias, and Serv6n 1999. 272 Post-Socialist Reformers FIGURE 5.6 MALI: ODF FROM BILATERAL SOURCES (DISBURSEMENTS), 1975-96 Millions of 1996 constant U.S. dollars 500 450 400 /000 ;00000 350 300 0 >xz X 250 _ 100 50 0 Technical assistance grants vI Nongrant component of official loans - Grant equivalent of official loans _ Non-technical assistance grants Source: Chang, Fernandez-Arias, and Serven 1999. Mali 273 FIGURE 5.7 MALI: ODF FROM MULTILATERAL SOURCES (DISBURSEMENTS), 1975-96 Millions of 1996 constant U.S. dollars 300 150 200 /4 \ 0 t 150 100 50 0 . . . . . . . ._ T echnical assistance grants = Nongrant component of official loans - Grant equivalent of official loans _ Non-technical assistance grants Note: Public lending includes concessional loans from the IMF ( SAF and ESAF), as well as nonconcessional loans (enhanced facility and standby agreement). Source: Chang, Fernandez-Arias, and Serven 1999. 274 Post-Socialist Reformers FIGURE 5.8 MALI: COMPARISON OF ODF AND TOF, 197o-96 Millions of 1996 constant U.S. dollars 900 800 - ~- Total gross official flows 800 ~ ~ ~ ~ ~ ~ A ~~~from the public sector, TOF 700 Total gross official flows (OECD-DAC) from the public sector, TOF 600 - (GDF 1999) \ 500 l 400 , \J Of ficial development finance, ODF (Chang, Ferndndez-Arias, 300 y * and Serven) 100 0 Sources: Chang, Fernandez-Arias, and Serven 1999; OECD-DAC; World Bank, Global Development Finance 1999. FIGURE 5.9 MALI: NET TRANSFERS FROM THE PUBLIC SECTOR, 1975-97 Millions of 1996 constant U.S. dollars 800 - Total gross official flows from the public sector 700 -" Net multilateral transfers _ Net bilateral transfers 600- 300- 200 100- 0 N9$ N9K 9 , ) Source: World Bank, Global Development Finance 1999. Mali 275 FIGURE 5.10 MALI: DEBT RATIOS, 1975-96 Percent Percent 180 600 Total debt as a percentage 160 - of exports of goods 500 140 and services / (right axis) 120 - 400 120 / Total ~~~~~debt as a40 100 - ecentage of GNP 80 t _ / (left axis) 300 60 - Debt service as a - 200 40 - - percentage of exports 40 >of goods and services 100 20 -(left axis) 10 0 o Source: World Bank, Global Development Finance 1998. FIGURE 5.11 MALI: STRUCTURE OF LONG-TERM, PUBLIC, GOVERNMENT-SECURED DEBT OUTSTANDING, 1970-97 Millions of 1996 constant U.S. dollars 3,500 Private creditors 3,000 m Bilateral official creditors on- Other of ficial multilateral creditors \ 2,500 _ IDA 2,000 1,500 1,000 -0 500 Source: World Bank, Global Development Finance 1999. 276 Post-Socialist Reformers TABLE 5.4 MALI: NET ODA DISBURSEMENTS, 1964-97 Annual average in millions of 1996 constant US dollars 1964- 1968- 1972- 1982- 1986- 1988- 1991- 1994- 67 71 81 85 87 90 93 97 Bilateral 61.5 79.7 212.7 399.1 365.7 384.5 278.5 268.9 EU members 46.6 58.5 133.6 248.6 222.2 263.5 181.8 186.2 France 42.7 46.6 74.7 158.6 101.6 134.4 98.1 81.8 North America 14.8 15.8 43.4 70.0 74.9 61.6 61.8 42.3 Japan 5.2 9.4 10.8 21.6 13.4 34.1 Arab 5.3 28.2 54.9 38.4 9.8 1.5 -9.1 Multilateral 53.4 61.4 147.2 176.5 195.3 205.3 192.4 231.4 IDA 2.9 11.4 32.2 41.7 66.5 63.1 63.3 80.3 IMF 0.0 0.0 0.0 0.0 -8.1 10.2 10.4 31.5 European Com- mission 53.6 33.2 62.5 43.1 40.8 48.1 64.9 62.9 ADB -3.3 8.9 18.9 32.0 39.6 19.0 21.4 Total 114.9 141.1 359.9 575.6 561.0 589.7 470.8 500.3 As a percentage of the total Bilateral 54 56 59 69 65 65 59 54 EU members 41 41 37 43 40 45 39 37 France 37 33 21 28 18 23 21 16 North America 13 11 12 12 13 10 13 8 Japan 0 0 1 2 2 4 3 7 Arab 0 4 8 10 7 2 0 -2 Multilateral 46 44 41 31 35 35 41 46 IDA 3 8 9 7 12 11 13 16 IMF 0 0 0 0 -1 2 2 6 European Com- mission 47 24 17 7 7 8 14 13 ADB -3 0 2 3 6 7 4 4 Source: OECD-DAC. Mali 277 TABLE 5.5 MALI: ODA COMMITMENTS BY SECTOR, 1973-97 Annual average in millions of 1996 constant US dollars 1973- 1982- 1986- 1988- 1991- 1994- 82 85 87 90 93 97 Infrastructures and social services 40.2 111.2 43.5 93.2 81.3 140.2 Infrastructures and economic services 66.1 103.8 28.8 51.5 34.7 68.9 Production 68.7 91.6 131.8 145.7 75.8 57.4 General assistance programa 44.2 51.5 47.4 119.0 98.1 58.6 including IMF and IDA adjustment assistance 0.0 0.0 2.5 79.3 79.2 48.0 Debt restructuring 22.4 44.4 0.0 2.1 0.6 7.8 Emergency aid 7.8 9.7 5.2 0.3 0.6 2.0 Other 17.3 31.9 37.0 36.6 52.0 42.0 Total 266.7 444.2 293.6 448.4 343.2 376.9 As a percentage of the total Infrastructures and social services 15.1 25.0 14.8 20.8 23.7 37.2 Infrastructures and economic services 24.8 23.4 9.8 11.5 10.1 18.3 Production 25.7 20.6 44.9 32.5 22.1 15.2 General assistance program 16.6 11.6 16.1 26.5 28.6 15.5 including IMF and IDA adjustment assistance 0.0 0.0 0.9 17.7 23.1 12.7 Debt restructuring 8.4 10.0 0.0 0.5 0.2 2.1 Eimer ency aid 2.9 2.2 1.8 0.1 0.2 0.5 O)ther 6.5 7.2 12.6 8.2 15.1 11.1 a. Includes adjustment programs, food aid programs, and other balance of payments supports. b. Includes multisectoral aid, support to NGOs, and nonallocated/nonspecific aid. Source: OECD-DAC Creditor Notification System. 278 Post-Socialist Reformers TABLE 5.6 MALI: TECHNICAL ASSISTANCE, 1962-97 Annual average in millions of 1996 constant US dollars 1962- 1968- 1972- 1982- 1986- 1988- 1991- 1994- 67 71 81 85 87 90 93 97 Bilateral 21.0 30.3 60.9 84.3 101.4 88.1 112.7 106.6 Multilateral 46.0 11.2 22.2 33.1 34.3 31.3 31.0 25.5 Total 67.1 41.5 83.1 117.4 135.8 119.4 143.7 132.1 As a percentage of the total Bilateral 31.4 72.9 73.3 71.8 74.7 73.8 78.4 80.7 Multilateral 68.6 27.1 26.7 28.2 25.3 26.2 21.6 19.3 Source: OECD-DAC. Mali 279 TABLE 5.7 MALI: OFFICIAL DEVELOPMENT FINANCE AND ITS COMPONENTS, BY SOURCE, 1975-97 Disbursements, in millions of 1996 constant US dollars 1975- 1982- 1986- 1988- 1991- 1994- 81 85 87 90 93 97 Total official developmentfinance (loans and grants) Bilateral 282.0 384.7 402.3 361.6 319.6 298.3 Multilateral 155.6 239.4 233.0 221.6 191.5 207.3 Total 437.6 624.0 635.3 583.3 511.1 505.6 Loans Bilateral 71.1 143.4 109.4 77.1 54.2 29.3 Multilateral 67.7 137.0 139.6 134.9 94.0 131.3 Total 138.8 280.4 249.0 212.0 148.2 160.6 Grants (including technical assistance) Bilateral 210.8 241.2 292.9 284.5 265.4 268.9 Multilateral 87.9 102.4 93.4 86.7 97.5 76.0 Total 298.7 343.6 386.3 371.2 362.9 345.0 Grants (excluding technical assistance) Bilateral 140.4 157.0 191.5 196.4 152.6 165.1 Multilateral 63.9 69.2 59.1 55.4 66.5 53.6 Total 204.3 226.2 250.5 251.8 219.2 218.7 As a percentage of total official development finance Total official developmentfinance (loans and grants) Bilateral 64.4 61.6 63.3 62.0 62.5 59.0 Multilateral 35.6 38.4 36.7 38.0 37.5 41.0 Total Loans Bilateral 16.3 23.0 17.2 13.2 10.6 5.8 Multilateral 15.5 22.0 22.0 23.1 18.4 26.0 Total 31.7 44.9 39.2 36.4 29.0 31.8 Grants (including technical assistance) Bilateral 48.2 38.7 46.1 48.8 51.9 53.2 Multilateral 20.1 16.4 14.7 14.9 19.1 15.0 Total 68.3 55.1 60.8 63.6 71.0 68.2 Grants (excluding technical assistance) Bilateral 32.1 25.2 30.1 33.7 29.9 32.7 Multilateral 14.6 11.1 9.3 9.5 13.0 10.6 Total 46.7 36.2 39.4 43.2 42.9 43.3 Source: Chang, Femrnndez-Arias, and Serven 1999; OECD-DAC. 280 Post-Socialist Reformers TABLE 5.8 MALI: WORLD BANK ADJUSTMENT CREDITS, 198o-98 (millions of U.S. dollars) Date Amount Amount Loans Type approved approved disbursed Agriculture SECAC 1983 - 56.6 Public enterprises SECAC 1988 40 40 Agriculture SECAC 1990 53 53 Structural adjustment SAC 1990 70 70 Economic recovery SECAC 1994 25 25 Education SECAC 1995 50 50 Economic management SAC 1996 60 46.1 - Not available. Note: SAC = Structural Adjustment Credit. SECAC = Sectoral Adjustment Credit. Source: World Bank. TABLE 5.9 MALI: IMF ADJUSMENT LOANS Millions of Millions of SDR U.S. dollars Amount Amount Amount Amount Loans Duration Year approved disbursed approved disbursed Standby Agreement I 1 year 1969 - 3.0 - 3.0 Standby Agreement II 1 year 1971 - 4.5 - 4.5 Standby Agreement IH 1 year 1982 - 30.3 - 33.5 Standby Agreement IV 2 years 1983 - 40.5 - 43.3 Standby Agreement V 2 years 1985 - 22.8 - 23.1 Standby Agreement VI 2 years 1988 12.7 12.7 17.1 17.1 SAF 3 years 1988 35.6 25.7 47.2 34.1 ESAF 4 years 1992 79.2 79.2 111.5 111.5 ESAF 3 years 1996 62.0 62.0 90.0 90.0 -Not available. Sources: IMF; Marches Tropicaux. TABLE 5.10 MALI: MULTILATERAL DEBT RESTRUCTURING AGREEMENTS WITH OFFICIAL CREDITORS, 1980-98 Portion Amount Consolidation included of debt consolidated Date of Consolidation period Debt already consoli- (U.S. $ Repayment terms agreement Deadline Start date Duraction Arrears rescheduled date (%) millions) Maturity Grace period 27 Oct 88 1 Jan 88 1 Jul 88 16 Yes 100 48 Menua Menu 22 Nov 89 1 Jan 88 1 Nov 99 26 Yes 100 33 Menu Menu 29 Oct 92 1 Jan 88 1 Oct 92 35 Yes Yes 100 107 Menu Menu 20 May 96 1 Jan 88 20 May 96 Stock Yes Yes 100 33 Menu Menu a. The term 'menu" refers to options that were the object of an agreement reached at the 1988 Toronto Summit; for agreements effective as of 1995, the term "menu" refers to the so-called "Naples" options after December 1994. Source: World Bank, Global Development Finance 1997. 282 Post-Socialist Reformers TABLE 5.11 MALI: DEBT RATIOS 1975- 1982- 1986- 1988- 1991- 1994- 81 85 87 90 93 97 Debt service (% of exports of goods and services) 5 11 17 15 11 15 Total debt (% of exports of goods and services) 282 432 491 461 492 472 Total debt (% of GDP) 47 94 105 103 107 127 Source: World Bank, Global Development Finance 1999. TABLE 5.12 MALI: STRUCTURE OF LONG-TERM, PUBLIC, AND GOVERNMENT-SECURED DEBT OUTSTANDING, 1970-97 Millions of 1996 constant U.S. dollars 1970- 1972- 1982- 1986- 1988- 1991- 1994- 71 81 85 87 90 93 97 Official creditors 1282.8 1104.5 2008.7 2569.0 2653.6 2986.0 2761.3 Official bilateral creditors 1241.9 898.4 1334.0 1718.5 1659.7 1806.2 1330.6 Official multilateral creditors 40.7 206.1 674.7 850.6 993.8 1179.8 1430.8 Including IDA 40.7 161.0 380.5 471.5 565.1 685.5 897.9 Private creditors 34.1 56.5 83.2 75.0 25.8 8.2 1.1 Total 1316.9 1161.0 2091.9 2643.9 2679.3 2994.2 2762.4 As a percentage of total debt outstanding Official creditors 97.4 95.1 96.0 97.2 99.0 99.7 100.0 Official bilateral creditors 94.3 77.4 63.8 65.0 61.9 60.3 48.2 Official multilateral creditors 3.1 17.7 32.3 32.2 37.1 39.4 51.8 Iincluding IDA 3.1 13.9 18.2 17.8 21.1 22.9 32.5 Private creditors 2.6 4.9 4.0 2.8 1.0 0.3 0.0 Source: World Bank, Global Development Finance 1999. Mali 283 TABLE 5.13 MALI: STRUCTURE OF TOTAL SERVICE OF LONG-TERM, PUBLIC, AND GOVERNMENT-SECURED DEBT, 1970-97 Millions of 1996 constant U.S. dollars 1970- 1972- 1982- 1986- 1988- 1991- 1994- 71 81 85 87 90 93 97 Official creditors 3.7 12.8 33.4 38.1 51.1 51.6 83.5 Official bilateral creditors 3.4 11.0 16.9 22.3 21.9 27.3 43.9 Official multilateral creditors 0.3 1.8 16.6 15.9 29.2 24.3 39.6 Including IDA 0.3 1.1 3.9 5.7 6.7 7.6 13.2 Private creditors 0.3 1.3 6.3 8.8 2.5 1.5 0.2 Total 4.0 14.2 39.8 46.9 53.6 53.2 83.7 As a percentage of total debt outstanding Official creditors 93.1 90.3 84.0 81.3 95.3 97.1 99.7 Official bilateral creditors 86.6 77.4 42.4 47.5 40.9 51.3 52.5 Official multilateral creditors 6.4 13.0 41.6 33.9 54.6 45.7 47.3 Including IDA 6.4 8.0 9.7 12.2 12.5 14.3 15.8 Private creditors 6.9 9.4 15.9 18.7 4.7 2.9 0.3 Source: World Bank, Global Development Finance 1999. NOTES 1. All loans of public origin, except for those having a purely military or defense-related purpose. 2. The deflator is the one calculated by DAC (OECD), and corresponds to the average of the deflators of donor countries that are members of the DAC once prices are expressed in U.S. dollars at the current exchange rate (assuming that 1996 = 100). This average is weighted according to the volume of ODA dis- pensed by each country in 1996. 3. The account went from 156 million French francs in 1968 to 1,272 million as of December 31, 1983. This debit was not counted in ODA since it bore debitor interest indexed according to the Bank of France's discount rate, nor was it counted under total gross official flows from the public sector (TOF), a more disputable decision that is probably justified because its direct objec- tive was not development, but instead the maintenance of the currency's convertibility. 4. Also, it was sometimes difficult to determine if such aid met ODA crite- ria due to the secret nature of the conditions and swap agreements associated with it. 5. Until July 1990, the Caisse Centrale de Cooperation Economique (CCCE) was providing Mali with concessional loans that complemented budgetary aid. Subsequently, the CCCE only granted "structural adjustment grants" to Mali, as it did with other low-income countries. The CCCE subsequently became the Caisse Francaise de Developpement (CFD), and then the Agence Francaise de DOveloppement (AFD). Bilateral aid to adjustment also came from U.S., Dutch, and Canadian sources. 6. Liberalization of rice marketing and of its purchase price was a condition of aid granted to the Office du Niger. REFERENCES Banque de France. 1998. "Rapport Zone Franc 1998." Paris. Bumside, C., and D. Dollar. 1997. "Aid, Policies, and Growth." Policy Research Working Paper 1777. World Bank, Development Research Group, Washing- ton, D.C. Chambas, G., J. L. Combes, P. Guillaumont, S. Guillaumont, and B. Laporte. 1999. "Mali: les Facteurs de Croissance a Long Terme." Draft report. Organisation for Economic Co-operation and Development, Paris. Chang, C., E. Fernandez-Arias, and L. Serven. 1999. "Measuring Aid Flows: A New Approach." Policy Research Working Paper 2050. World Bank, Devel- opment Research Group, Washington, D.C. Coelos. 1994. Capitalisation de l'experience du PRMC. Processed. Collier P., P. Guillaumont, S. Guillaumont, and J. W. Gunning. 1997. "Redesign- ing Conditionality." World Development 25 (9): 1399-1407. Deme, M. 1999. "Plan d'Action a Moyen Terme (1997-1999) pour la phase V du PRMC." Processed. Guillaumont, P., and S. Guillaumont. 1994. "La Zone Franc, les Institutions de Bretton-Woods et la Conditionnalite." In Bretton Woods, melanges pour un cinquantenaire. Paris: Revue d'Economie Financiere. .1995. "La Conditionnalite a l'Epreuve des Faits." In Michel Rainelli, ed., La Negociation Commerciale et Financiere Internationale. Paris: Economica. Guillaumont, P., and L. Chauvet.1999. "Aid and Performance: A Reassessment." Communication ABCDE Workshop on Aid, Selectivity, and Conditionality. World Bank, Washington, D.C. Laporte, B., and P. Monier. 1994. "Les m6canismes de stabilisation des prix aux producteurs peuvent-ils faire face a la devaluation du franc CFA? L'experience du coton au Mali et du caoutchouc en C6te d'Ivoire." Marches Tropicaux no. 2535 (10 June): 1194-5. Lecaillon, J., and C. Morrisson. 1986. Economic Policies and Agricultural Perfor- mance: The Case of Mali. Paris: Organisation for Economic Co-operation and Development. OECD (Organisation for Economic Co-operation and Development). 1998. "Re- vue du systeme intemational d'aide au Mali. Synthese et analyse." Provi- sional report. Paris. . 1999a. "R6partition Geographique des Ressources Financieres Allou6es aux Pays Ben6ficiaires de l'Aide." Paris. . 1999b. "Activites d'Aide en Afrique. Systeme de Notification des Pays Creanciers." Paris. Sachs, J. D., and A. M. Warner. 1995. "Economic Reforms and the Process of Global Integration." Brookings Papers on Economic Activity, vol.1. Brookings Institution, Washington, D.C. World Bank. 1994. "Country Assistance Strategy for Mali." Report no. 13746- MLI. Washington, D.C. . 1996a. "Implementation Completion Report for the Public Enterprise Institutional Development Project." Report no. 15953. Washington, D.C. . 1996b. "Implementation Completion Report for the Structural Adjust- ment Credit." Report no. 15644. Washington, D.C. . 1996c. "Project Completion Report for the Economic Recovery Credit." Report no. 15819-MLI. Washington, D.C. . 1998a. "Country Assistance Strategy for Mali." Report no. 17775-MLI. Washington, D.C. - . 1998b. "Implementation Completion Report for the Agricultural Sector Adjustment/Investment Credit." Report no. 18027-MLI. Washing- ton, D.C. 286 Post-Socialist Reformers - 1998c. "Memorandum of the President on Assistance to the Republic of Mali under the HIPC Debt Initiative." Report no. P-7529-MLI. Washing- ton, D.C. . 1999a. Global Development Finance 1999. Analysis and Summary Tables. Washington, D.C. - 1999b. World Development Indicators. Washington, D.C. - 1999c. World Development Report 1999. Washington, D.C. Tanzania Arne Bigsten Goteborg University Goteborg, Sweden Deogratias Mutalemwa Yvonne Tsikata Samuel Wangwe Economic and Social Research Foundation Dar es Salaam, Tanzania ACRONYMS AND ABBREVIATIONS AfDB African Development Bank BIS Basic Industries Strategy CAS Country Assistance Strategy CEM Country Economic Memorandum CG Consultative Group CIS Commodity Import Support CTI Confederation of Tanzanian Industries DAC Development Assistance Committee (of the OECD) EDA Effective development assistance ERP Economic Recovery Program ESAF Enhanced Structural Adjustment Facility (of the IMF) ESAMI East and Southern African Management Institute ESRF Economic and Social Research Foundation EU European Union GDP Gross domestic product IDA International Development Association (of the World Bank Group) IFI International financial institution IMF International Monetary Fund MDF Multilateral Debt Relief Fund NBC National Bank of Commerce NESP National Economic Survival Program NGO Nongovernmental organization ODA Official development assistance OECD Organisation for Economic Co-operation and Development OGL Open General License PC Paris Club PER Public Expenditure Review PFP Policy Framework Paper PSRC Parastatal Sector Reform Commission REPOA Research on Poverty Alleviation SAC Structural Adjustment Credit SPA Special Program of Assistance (for Sub-Saharan Africa) SPA/JEM Special Program of Assistance/Joint Evaluation Mission Tshs. Tanzanian shilling UNDP United Nations Development Programme USAID U.S. Agency for International Development Tanzania has been at the center of development debate since the 1960s. It chose to pursue a development strategy aimed at self-reliance and African socialism, and President Nyerere became a respected spokes- man for the whole Third World. This sparked a great deal of positive interest in the Tanzanian experiment, particularly from social democratic governments in Europe, which provided generous aid to the country. However, the Tanzanian experiment ran into trouble, drawing more and more critique from the outside, and donor support was reduced. The ensuing economic crisis and increasing domestic critique eventually led to a policy reversal and the introduction of adjustment measures. The purpose of this study is to get a better understanding of the causes of policy reform in Tanzania, particularly how aid affected the reform process relative to domestic forces. The study is an exercise in economic history or political economy and is not primarily concerned with the impact of aid on economic outcomes. The study covers mainly the pe- riod since the onset of the severe economic crisis in Tanzania in 1979, but to put this in perspective we will also briefly outline the policies pur- sued before that. The paper has three main objectives: it examines the nature of exter- nal assistance, the causes and path of policy reforms, and the relation- ship between the aid and reforms. Section 2 examines the different types of aid that Tanzania receives, tracing the size and character of the aid flows, the relationship with multilateral and bilateral donors, and at- tempts at donor coordination. Section 3 describes the economic policy and institutional framework and analyzes more precisely the causes and implementation of the identified policy reforms as well as the internal dialogue and involvement of different actors. Section 4 assesses the links between aid and reform and presents conclusions. AID FLOWS TO TANZANIA The standard concept of aid used in most studies is that of official devel- opment assistance (ODA), which consists of financial aid and technical cooperation (also called technical assistance). Financial aid includes grants and concessional loans having a grant element of at least 25 per- cent. Technical cooperation includes grants to nationals of aid-recipient countries receiving education or training, and payments to consultants, advisers, administrators, and similar persons working on assignments of interest to the recipient countries. However, the standard measure tends to overestimate the aid flow, since it does not provide an accurate measure of the grant element in the official loans. The overestimation is less for Africa than for other regions due to the high grant equivalent in loans to Sub-Saharan Africa, but it is still significant. 289 290 Post-Socialist Reformers Because of these problems the World Bank has produced a new aid measure called effective development assistance (EDA), which provides a more accurate measure of true aid flows than the traditional ODA measure (Chang, FernAndez-Arias, and Serven 1999). EDA is the sum of grants and the grant equivalent of official loans, and aims to measure the pure transfer of resources. EDA is defined as the grant equivalents of all development flows in a given period. Grants tied to technical assis- tance are excluded from EDA, but since we use the grant data from the Development Assistance Committee (DAC) of the Organisation for Eco- nomic Co-operation and Development (OECD), they are included in the estimates presented in the following section. Official Aid Flows Volume and Composition of Aid Since independence in 1961, Tanzania has been one of the largest recipi- ents of aid in Sub-Saharan Africa in absolute terms. The country still receives considerably more aid as a percentage of gross domestic prod- uct (GDP) than most other countries in the region. Tanzania's share of total aid from all the DAC countries was 8.3 percent during the 1970s by the definition used in this study, and the country was the largest recipi- ent of aid in Sub-Saharan Africa in that period (see table 6.1). This share started to decline in 1981, and by 1985 it was down to 5.8 percent. Nev- ertheless, Tanzania still ranked as the second-largest beneficiary of aid (after Sudan) during the 1980s. Even more remarkable is that even in the mid-1980s, when the policy environment was very poor in Tanzania, the country received more than twice as much as did Ghana, which was considered at that time to be the reformer. This suggests that there is a high degree of inertia in aid flows. Once they are committed, it is diffi- cult to cut a country loose. Tanzania continued to rank second in the 1990s, slightly behind Mozambique. Between 1970 and 1996, the country received a total of US$16,632 million in foreign aid (see table 6.1). Aid is defined here as the sum of total grants according to the DAC definition plus the grant equivalent of loans according to the EDA definition of the World Bank. Aid has ac- counted for more than 80 percent of total net inflows of external capital, indicating that private capital flows have been small. The annual vol- ume of aid rose from US$37.7 million in 1970 to a peak of US$1,158.0 million in 1992, but then declined to US$830.5 million in 1996. In per capita terms, Tanzania's aid receipts peaked in the late 1980s and early 1990s at more than US$40. As a share of GDP, aid peaked at 30 percent in 1990, but due to the decline in aid flows and the increase in GDP during Tanzania 291 TABLE 6.I TOTAL AID TO TANZANIA, i97o-96 (millions of U.S. dollars) Shzare of Share of Tanzania in bilateral total aid to Total aid aid in Sub-Saharan Total Total per capita total aid Africa Year bilateral multilateral Total aid (US$) (%) (%) 1970 37.70 0 37.70 2.7 100.0 4.6 1971 39.04 0.06 39.10 2.8 99.8 3.8 1972 90.52 1.18 91.70 6.3 98.7 6.3 1973 153.62 1.78 155.40 10.4 98.8 9.6 1974 187.56 7.14 194.70 12.6 96.3 8.0 1975 271.04 32.26 303.30 19.1 89.3 9.8 1976 213.12 44.78 257.90 15.7 82.6 8.9 1977 239.86 67.04 306.90 18.1 78.1 9.3 1978 481.66 51.44 533.10 30.6 90.3 12.0 1979 489.51 67.59 557.10 31.0 87.8 10.5 1980 666.82 94.58 761.40 41.0 87.5 11.2 1981 495.43 165.47 660.90 34.5 74.9 9.5 1982 458.07 176.03 634.10 32.0 72.2 9.2 1983 451.71 132.59 584.30 28.6 77.3 8.4 1984 426.48 107.82 534.30 25.3 79.8 7.3 1985 384.49 98.11 482.60 22.2 79.6 5.8 1986 595.58 194.02 789.60 35.1 75.4 7.9 1987 607.00 144.50 751.50 32.4 80.7 7.1 1988 621.97 421.63 1,043.60 44.5 59.5 8.4 1989 660.73 181.57 842.30 34.1 78.4 6.7 1990 767.78 287.81 1,055.60 41.4 72.7 6.4 1991 775.33 315.37 1,090.70 41.5 71.1 6.9 1992 789.73 368.27 1,158.00 42.7 68.1 7.2 1993 918.51 235.99 1,154.50 42.3 79.5 7.8 1994 642.12 285.08 927.20 32.2 69.2 5.8 1995 622.17 231.83 854.00 28.8 72.8 5.7 1996 569.65 260.85 830.50 27.2 68.5 6.0 Sources: World Bank EDA database and OECD-DAC database. The grant estimates are from OECD and include technical assistance, while estimates of the aid element in loans are from the EDA computations. the 1990s the share has fallen very significantly since then, to about 13 percent of GDP in 1997. These figures can be compared with those of some other countries in the region. Aid to Kenya peaked at 27 percent of GDP in 1993, but fell to 5 percent by 1997. In Uganda aid was 27 percent of GDP in 1992, but fell to 13 percent by 1997. Aid to Zambia declined from 36 percent in 1991 to 292 Post-Socialist Reformers 17 percent in 1997, while aid to Ghana fell from 14 percent of GDP in 1989 to only 7 percent in 1997 (World Bank 1999c). It is evident that Tan- zania is not the only country that is highly dependent on aid. Still, aid dependence has declined significantly during the 1990s. Tanzania has received aid from more than 50 bilateral sources. This is a high figure, but not unusual for the larger African economies. The Nordic countries (Sweden, Norway, Denmark, and Finland) have been the major donors, accounting for more than 30 percent of the country's total ODA receipts between 1970 and 1996 (see table 6.9 for a break- down of flows by country). Sweden has provided almost half of that amount for most of this period. The second group of major bilateral do- nors consists of Germany and the Netherlands, each contributing around 8 percent of total aid over the past 20 years. The third group consists of Canada, the United States, and the United Kingdom, each accounting for about 6 percent of Tanzania's total aid receipts. Italy and Japan became important new donors in the late 1980s. China and the socialist bloc dras- tically reduced their aid programs for Tanzania starting in the 1980s. During the 1970s, most of the aid flows to Tanzania came from bilat- eral sources. Multilateral donors contributed an average of about 10 percent annually. Since the early 1980s, however, the share of multilat- eral donors has risen considerably. In 1988 the multilateral share of total aid was about 40 percent (see table 6.1). In the case of aid in the form of loans, the contribution of multilateral donors has risen much faster, from an average of less than 5 percent in 1970-75 to more than 85 percent in 1987-96 (see table 6.2). Major multilateral donors include the Interna- tional Monetary Fund (IMF), the World Bank, the African Development Bank (AfDB), the European Union (EU), the United Nations Develop- ment Program (UNDP), and the United Nations High Commission for Refugees, with the World Bank Group's International Development As- sociation (IDA) providing about half of total aid from multilateral do- nors. Total disbursements on EDA loans are shown in table 6.10. During the reform period the amounts have been in the range of US$200 million to US$350 million per year. In terms of sectoral distribution, most of the aid in the 1960s and the early 1970s was directed to the agricultural and transport sectors. Dur- ing the second half of the 1970s the emphasis shifted toward industry and energy. Transport became an important aid recipient in the late 1980s and early 1990s. The general pattern of aid distribution in the 1990s has been (in order of importance): transport and communications, receiving about half of the aid flows, followed by agriculture, human resources development, health, integrated regional development, and energy (Mutalemwa, Noni, and Wangwe 1998). The sectoral allocation of EDA loans is shown in tables 6.11 and 6.12. Much of the loan amount during the reform period has gone to the rehabilitation of infrastructure as well Tanzania 293 TABLE 6.2 TANZANIA: EFFECTIVE DEVELOPMENT ASSISTANCE IN THE FORM OF LOANS, 1970-96 (millions of U.S. dollars) Share of bilateral aid in total aid Year Total bilateral Multilateral EDA as loans (%) 1970 10.15 0 10.15 100.00 1971 8.50 0.02 8.48 100.20 1972 46.50 0.59 47.10 98.70 1973 89.73 0.88 90.61 99.00 1974 91.55 3.77 95.32 96.00 1975 98.17 16.88 115.05 85.30 1976 38.68 23.28 61.96 62.40 1977 54.91 36.52 91.44 60.10 1978 49.16 26.87 76.03 64.70 1979 33.59 35.82 69.41 48.30 1980 53.53 49.51 103.03 51.90 1981 58.87 92.70 151.57 38.80 1982 50.62 93.25 143.87 35.20 1983 82.58 71.89 154.46 53.50 1984 48.82 55.98 104.81 46.60 1985 22.54 52.06 74.60 30.20 1986 46.77 95.99 142.76 32.80 1987 11.72 71.96 83.69 14.00 1988 22.70 210.85 233.54 9.70 1989 13.38 90.60 103.98 12.80 1990 23.79 145.08 168.87 14.10 1991 16.78 160.11 176.89 9.50 1992 36.83 187.57 224.39 16.40 1993 9.88 119.74 129.62 7.60 1994 6.28 144.57 150.77 4.34 1995 18.88 117.81 136.69 16.02 1996 3.35 132.33 135.68 2.53 Source: World Bank EDA database. as manufacturing. The World Bank has been the predominant creditor during this period (see table 6.13). Other donors have been reluctant to make loans to the poor and heavily indebted country, but have chosen to provide grants instead. By comparing tables 6.13 and 6.14 we may note that the grant element of World Bank loans has also increased. Since grants are more fungible than loans (Feyzioglu, Swaroop, and Zhu 1998), it is generally more difficult to reallocate aid from the international fi- nancial institutions (IFIs) according to the preferences of the recipient govemment. 294 Post-Socialist Reformers The composition of aid has also shifted over time. For most of the 1960s and 1970s, investment projects assistance made up more than two-thirds of total aid. The preference for project assistance over pro- gram aid during this period can be attributed to three factors. First, projects appeared to be easy to plan, design, control, and supervise, and hence promised visible results while allowing for direct account- ability. Second, unlike programs, projects are easy to tie to the procure- ment of goods and services from the donor. Third, projects provided donors with opportunities to bypass national institutions and pursue their own objectives. Since the mid-1980s, there has been a shift in emphasis from project aid to program aid. Principal reasons for the change include Tanzania's growing balance of payments problems and the declining utilization of capacity in industry. These developments were perceived as evidence that the performance of project assistance had been unsatisfactory, and that import support was critically needed to raise the level of output through greater utilization of enterprise capacities. Another important change has involved a shift toward financing more explicitly the project forms of the government's recurrent budget. This has been a response to the growing awareness that insufficient allocations to the recurrent ex- penditure in the government budget had become a more serious con- straint to the delivery of services than allocations to the development budget. Trends in Aid Flows Over the period 1970-96, three broad phases of aid flows to Tanzania can be identified: the expansion phase, from 1970 to 1982; the contrac- tion phase, from 1983 to 1985; and the adjustment phase, from 1986 to 1996. Aid to Tanzania increased more than 20 times (in current dollars) between 1970 and 1982. It decreased by about 25 percent between 1982 and 1985, and then rose again sharply from US$482.6 million in 1985 to a peak of US$1,158 million in 1992. EXPANSION PHASE: 1970-82. The willingness of donors to extend sub- stantial amounts of aid to Tanzania during the late 1960s and the 1970s can be explained by the development policies the country pursued dur- ing this period. The country's emphasis on self-reliance and growth with equity, the targeting of governnent programs to basic needs, and em- phasis on cooperatives and Ujamaa villages motivated many donors to provide aid to Tanzania. Such policies were in line with the thinking of social democrats in the North, notably the Nordic countries, Germany (under Willy Brandt), and Canada; these governments took a positive view of Tanzania's development strategy. The World Bank doubled its Tanzania 295 lending program to Tanzania between 1973 and 1977 because Robert McNamara, who was president of the World Bank at the time, had adopted a policy of growth with equity that was in line with Tanzania's development strategy under the leadership of President Julius Nyerere (Bagachwa, Mbelle, and van Arkadie 1992). CONTRACTION PHASE: 1983-85. Toward the end of the 1970s, there was a change in donor attitudes toward Tanzania. Donors became in- creasingly critical of the country's development strategy. There was grow- ing evidence that in spite of the Arusha Declaration neither socialism nor self-reliance had been achieved, and that the effectiveness of aid had been low.' Following the breakdown of negotiations with the IMF in September 1980, the general view among Western donors was that aid to Tanzania could not be effective unless the country agreed to redress inappropriate domestic policies. Despite these developments, the Nor- dic countries did not abandon Tanzania but continued to provide aid although at a reduced rate. By 1983, however, most donors, including Sweden, had begun to scale down their support and aid flows declined sharply. Ultimately, the sheer magnitude of the crisis, pressure from the IMF, the World Bank, and bilateral donors, and the failure of indepen- dent national efforts to revive the economy all combined to persuade the government to adopt a far-reaching Economic Recovery Program (ERP I) in 1986 with IMF and World Bank support. ADJUSTMENT PHASE: 1986-96. The agreements with the IFIs in 1986- a standby agreement with the IMF and a structural adjustment program with the World Bank-helped to restore donor confidence. Under ERP I, the government adopted a broad range of policies aimed at liberalizing internal and external trade, unifying the exchange rate, reviving exports, stimulating domestic saving, and restoring fiscal sustainability. The ERP enjoyed strong donor support. Bilateral aid resumed, leading to a sec- ond aid boom that peaked in 1992 (table 6.1). The agreement with the IMF was renewed in 1987, 1988, and 1990. In 1989, the reforms entered a second phase under ERP II (1989-91). The new reforms continued earlier efforts at trade and exchange rate liberal- ization and macroeconomic stabilization, but widened to include reforms in the banking system, agricultural marketing, the parastatal sector, gov- ernment administration, and the civil service, together with a targeting of the social sectors. However, during 1993-94 relations between the government and major donors deteriorated over the government's failure to collect counterpart funds under the Commodity Import Support scheme and to increase tax collection efforts generally. There were increasing suspicions of corrup- tion. Donors started to question the effectiveness of their aid, as reflected, 296 Post-Socialist Reformers for example, in evaluation studies carried out by the Nordic countries and the Netherlands in 1994. The government commitment to reform was also questioned, and its failure to conclude a new Enhanced Struc- tural Adjustment Facility (ESAF) with the IMF worsened donor fears. The govemment, for its part, felt that the donors' demands were often unrealistic and too intrusive in matters that were essentially domestic. An independent group of local and international advisers was commis- sioned by Denmark to see how the climate of confidence between the two sides could be restored. The group, headed by Professor Helleiner, produced a report in June 1995 that has become an important policy document in both the donor community and the government (Mutalemwa, Noni, and Wangwe 1998). Strained relations between the government and donors in 1993-94 had a significant adverse effect on aid flows to Tanzania between 1994 and 1996. Donor confidence was restored and normal aid operations resumed after the new government came to power and, especially, after an ESAF agreement was concluded with the IMF in 1996. Other Types of Financial Aid Substantial aid has been disbursed as balance of payments support through the Commodity Import Support (CIS) and the Open General License (OGL) schemes. More and more donors are currently inclined to channel their support through a new Balance of Payments Import Sup- port scheme and the newly introduced Multilateral Debt Relief Fund. A few donors, for example the Japanese, still channel their assistance through the CIS arrangement. Commodity Import Support The CIS facility established under ERP I used donor funds to help pro- vide foreign exchange to cushion the impact of devaluation and stimu- late agricultural exports. The program was to improve levels of capacity utilization of existing production facilities. The CIS was a tied facility, with donors insisting on procurement of goods from their own coun- tries. Restrictions were also placed on the type of goods that could be imported. Applications for utilization of CIS funds had to be submitted to the Treasury, and after approval importers were required to pay a minimum of 30 percent cash cover on the allocated amount. The remain- ing 70 percent would be paid on arrival of shipping documents. The scheme ran into difficulties and many donors decided to aban- don it because of the government's failure to account for payment of counterpart funds against the foreign exchange disbursed. It should also be noted that the allocation of CIS reflected donors' preferences in that Tanzania 297 most donors supported "their" projects, that is, projects they had been responsible for financing in the first place. This meant that allocation did not necessarily follow criteria of efficiency. The program was in some ways similar to the Marshall Plan that helped Europe recover after World War II. The outcome in Tanzania was not as good, however. There was not yet recognition that many of the industries were obsolete and that new investment would be required to make them competitive. Open General License The OGL system was established in the early 1980s to provide a nonad- ministrative system for allocating foreign exchange for the importation of goods. The system was suspended in 1982 because of a shortage of foreign exchange, but resumed in 1988 with joint financing by the World Bank and other donors. The objective was to support a market-driven program of imports. Importers were free to import the goods they wanted, provided they were eligible under the OGL system. The minimum value of goods to be financed under the OGL facility was US$5,000. The Bank of Tanzania issued import licenses automatically, provided that import- ers fulfilled all the conditions pertaining to the utilization of funds un- der the facility. Foreign exchange was allocated by way of an import license, and importers were required to make a down payment and settle the difference on arrival of shipping documents. However, payment of counterpart funds eventually became a prob- lem, as importers did not pay promptly, and the scheme had to be aban- doned. A new nonadministrative system of allocation had to be established in 1993. Attempts by government to collect the overdue coun- terpart funds in relation to OGL and CIS, estimated at Tshs.160 billion at the end of 1992, have not been successful. The Economic and Social Research Foundation (ESRF) in Dar es Sa- laam has evaluated the extent to which Tanzania has implemented rec- ommendations from the Special Program of Assistance/Joint Evaluation Mission (SPA/JEM) project with regard to the CIS and OGL schemes (ESRF 1998). The picture is mixed. There has been some progress, but in other areas, for example debt recovery, little has been achieved.2 Balance of Payments/Budgetary Support In 1993, the government introduced a new system of importation as one of the measures adopted under liberalization of foreign trade. The sys- tem provides for importers to declare imports by filling in an import declaration form. Donors, through the Special Program of Assistance (SPA) for Sub-Saharan Africa, agreed to disburse balance of payments 298 Post-Socialist Reformers import support through a reimbursement method. Under the system, expenditures incurred from the country's own foreign exchange are re- imbursed upon presentation of specified import documents. After reim- bursement, the Bank of Tanzania promptly pays the equivalent in Tanzanian shillings to the Treasury. The agreement between Tanzania and the United States requires the Bank of Tanzania to deposit the coun- terpart funds to an account that the U.S. Agency for International Devel- opment (USAID) maintains with a local commercial bank. The Bank of Tanzania then sells the foreign exchange arising from donor resources to commercial banks in the Inter-Bank Foreign Exchange Market. This sys- tem has resolved the problem of nonpayment of counterpart funds to the government. In 1998 the donors who were active in providing balance of payments import support through this system were the United Kingdom, the Neth- erlands, the EU, AfDB, and USAID. In 1993-94 a total of 99.4 percent of the amount pledged was disbursed. The period 1994-96 experienced low disbursement of about 30 percent as most donors withheld disburse- ment. Thereafter disbursement rates picked up to 58 percent and 50.4 percent in 1996-97 and 1997-98 respectively (Mutalemwa, Noni, and Wangwe 1998). Debt Rescheduling and Debt Alleviation Tanzania's external debt has increased from US$4,970 million in 1986 to US$7,931 million in 1997, an increase of 59.9 percent. Actual debt service decreased from US$303 million in 1989 to US$122 million in 1990, then increased to US$233 million in 1997 and dropped again to US$174 mil- lion in 1997. Between 1986 and 1997, actual debt service averaged 50 percent of scheduled debt service, and as a result, debt arrears accumu- lated from US$664 million to US$3,227 million during that period. Paris Club Between 1986 and 1997 Tanzania went to the Paris Club (PC) five times. The first meeting took place in 1986, followed by meetings in 1988,1990, 1992, and 1997. At PC I, Tanzania obtained terms under which 97.5 per- cent (US$722.9 million) of eligible debts in arrears were rescheduled. PC II and PC III gave Tanzania the opportunity to benefit from Toronto terms, which provided for 33 percent debt reduction. Under PC II, debts worth US$22.3 million were cancelled and US$373 million rescheduled. Debts worth US$18.7 million were cancelled and US$199.4 million was resched- uled within the framework of PC III. At PC IV, the enhanced Toronto terms were applied and debts worth US$182.8 million were cancelled and those worth US$458.3 million were rescheduled. Under PC V, the Tanzania 299 government received the Naples terms. This made it possible for debts worth US$1,000 million to be eligible for consideration for cancellation and those worth US$700 million to be rescheduled. HIPC The Heavily Indebted Poor Countries (HIPC) Debt Initiative was estab- lished in 1996 to address multilateral debt. It lays down a set of criteria for providing across-the-board debt relief. Tanzania is one of the coun- tries potentially eligible for HIPC support. In April 2000 IDA and the IMF agreed to support a comprehensive debt reduction package for Tanza- nia under the enhanced HIPC Initiative. Total relief from all of Tanzania's creditors is worth more than US$2 billion in net present value terms. The whole debt relief operation will translate into debt service relief over time of US$3 billion, or about half of Tanzania's debt service obligations during fiscal years 2001-03 and about a third of debt service obligations thereafter. For the program to be fully implemented Tanzania has to sat- isfy a number of conditions, including the adoption of a participatory poverty reduction strategy paper. Multilateral Debt Relief Fund In July 1998, the government took the initiative of setting up a Multilat- eral Debt Relief Fund (MDF) which is intended to assist in servicing debt obligations to multilateral financial institutions, notably the IMF. Disbursement of funds is based on actual debt obligations that mature during a specific period. Funds disbursed under the system are depos- ited into a special MDF account and payments are made when the debt obligations fall due. The budgetary savings arising from this arrange- ment are to be used by government to enhance expenditures in social sectors, especially health and education. Countries that have shown par- ticular interest in contributing to MDF are Denmark, Finland, Ireland, the Netherlands, Norway, Sweden, and the United Kingdom. Nonfinancial Aid Technical Cooperation Between 1976 and 1996, the volume of technical cooperation resources to Tanzania increased by more than 12 times, from US$21.4 million in 1970 to a peak of US$266.7 million in 1995 (see table 6.15). The share of technical cooperation in total foreign assistance, which was 56 percent in 1970, had dropped to 19 percent by 1975. It rose to 29 percent in 1985 and stabilized at around 30 percent in the 1990s. 300 Post-Socialist Reformers Technical assistance has mainly been provided as part of specific projects or programs. In that form it did not address the wider objective of institutional capacity building. Projects have usually been approved by donors on condition that they would be executed by special project implementation units often staffed by highly paid expatriate personnel and selected local experts. This is driven by the urge to speed up imple- mentation, though it frequently sets up parallel management units that duplicate functions that can be performed by existing local institutions. In fact, donors go as far as managing the funds themselves. The behavior of bilaterals differs little from that of multilaterals in this regard. Both tend to doubt local capacities; the difference is that multilaterals are bound by stricter rules regarding procurement that en- sure that the Tanzanian government manages the funds and selection of consultants and experts. A fundamental problem is that there has been no clear government policy on the role or limits of technical assistance. The government policy is so undefined that responsibility is scattered among five ministries, with uncoordinated procedures for determining priority needs, recruit- ing expatriates, issuing immigration visas and work permits, extending contracts, training local counterpart and replacement personnel, negoti- ating donor funding, and evaluating expatriate performance (Mutalemwa, Noni, and Wangwe 1998). Following long-standing criti- cism of both donors and the government of Tanzania over this matter- which culminated with the Helleiner report of 1995 and a commitment in 1997 to do something to rectify the situation-the government has finally decided to address the problem in a systematic fashion. It has drafted a policy with wide-ranging proposals to restore better use of external resources allocated for technical assistance recruitment and to enhance ownership by the government in the management of these re- sources. Helleiner asserts in his latest assessment report (March 1999) that donors have started to provide more information upon request con- cerning their direct expenditures on technical cooperation. The Role of NGOs and Similar Institutions in Transmission of Ideas The community of nongovernmental organizations (NGOs) in Tanzania has undergone spectacular growth. While there were about 200 NGOs in the mid-1980s, by 1997 about 8,400 had been registered. The unprec- edented growth of civic organizations has made it possible for them to assert themselves more forcefully. They work in coalitions, or individu- ally, to lobby for change, as in the recent case of the movement to influ- ence the land bill. Women's organizations, which make up a significant percentage of all NGOs, have been especially active in this regard. They have been involved in lobbying for the reform of laws pertaining to in- Tanzania 301 heritance, domestic violence, land reform, children's rights, and child abuse. In this struggle they are not influenced by external assistance as such, although some of these issues have become transnational. The growing strength and influence of NGOs and other civil society associations has not come without conflicts. The government has sought ways to contain the NGOs or control their actions, yet at the same time it has been obliged to temper its concern and to consult them more frequently. More recently, the government has been trying to help them create an umbrella body with which it can engage in dialogue. Inde- pendent policy analysis groups have begun to speak out directly or through the media, and the government and donors are starting to lis- ten to them. For example, OXFAM, a British NGO, is working with the Mwalimu Nyerere Foundation, a local NGO, to facilitate a national debate on Tanzania's external debt and possible options. The debate has contributed considerably to the establishment of the Multilateral Debt Relief Fund, with pledges and financial contributions from sev- eral bilateral donors. Relations with Multilateral Institutions Tanzania has had a mixed relationship with the major multilateral insti- tutions-the World Bank and the International Monetary Fund-over the years. During the 1970s, admiration for the integrity of President Nyerere and the independent development path he was trying to forge led the Bank, in particular, to provide significant financial assistance to Tanzania. By the end of the decade, however, it was clear that the coun- try faced increasing economic difficulties due in part to policy mistakes. Both the external and internal accounts were in imbalance. The crisis worsened as a result of several shocks (the breakup of the East African Community in 1977, the war with Idi Amin's Uganda in 1978-79, the end of the coffee boom in 1978, and the second oil crisis in 1979). It seemed inevitable that some of the Arusha Declaration principles of socialism would have to be abandoned if the economy were to recover. The period 1978-82 was marked by difficult relationships with multi- lateral institutions. Tanzania attempted a number of homegrown eco- nomic reforms, described below. These were not very successful as the proposed programs were too limited and failed to attract foreign financ- ing. Increasingly, bilateral donors who had always been supportive of Tanzania were offering their financial support conditional on Tanzania making a deal with the IMF. Finally, in 1986, Tanzania reached agreement on a standby arrange- ment with the IMF. This agreement paved the way for increased lending by the World Bank and several other donors. Nonetheless, the level of bilateral aid did not recover to pre-1986 levels in real terms. Consequently, 302 Post-Socialist Reformers as discussed above, compared to the 1970s the multilateral donors have become much more important, accounting for more than 90 percent of effective development loans by 1996 (see table 6.13). The World Bank has taken the lead in mobilizing aid through the Consultative Group meetings. In December 1997 the meeting was held in Dar es Salaam for the first time and was regarded as generally suc- cessful (Helleiner 1999). Holding the CG in Dar enabled the president and the entire cabinet to participate. It also allowed trade unions, NGOs, and the private sector to interact with the donor community. Unfortu- nately, this positive step toward ownership was set back in 1999. The donors voted narrowly to hold the CG in Paris, arguing that they would be better able to attract high-level representation. The need for donor coordination has been obvious to donors in Tan- zania for a long time. Still, efforts to actually coordinate aid from differ- ent donors have seldom progressed beyond the organization of meetings, where donors exchange information with each other and the govern- ment. The World Bank has now taken the lead in renewed efforts on donor coordination, co-chairing the monthly joint government-donor meetings with UNDP. In addition, the Bank actively consults with the other donors through the Special Program of Assistance for Africa and the World Bank/IMF annual meetings. Nonetheless, it appears that sig- nificant duplication still exists among donors and little progress has been made in standardizing reporting systems and so on. For example, sev- eral donors are supporting capacity building in the Ministry of Finance, with little attempt to coordinate such support. Individual donors have their own parallel project management systems that essentially bypass the official channels for aid management. Indeed, the 1999 Public Ex- penditure Review estimates that more than 70 percent of aid did not enter through the budget in the last fiscal year. Aid sometimes does not tie in to national priorities or development plans. Attempts are underway to resolve some of these problems through sector-wide approaches in education and health. The recent multidonor- government review missions in these sectors are an encouraging sign. The World Bank is encouraging government ownership of major com- ponents of the macroeconomic policy dialogue. In 1999, for the first time, the government drafted the Policy Framework Paper (PFP) and made significant contributions to the Public Expenditure Review (PER). While the PFP took longer to draft than in the past, because of the need to consult, there seems to be general agreement that this new approach resulted in greater ownership by the government. World Bank lending during the reform period initially focused on re- habilitation of the country's dilapidated infrastructure, with loans for roads, power, and telecommunications. Tanzania has had several ad- justment loans from the World Bank and IMF (see table 6.3). A multisector rehabilitation loan supported macroeconomic adjustment in 1986. Tanzania 303 TABLE 6.3 WORLD BANK LOANS TO TANZANIA, 1986-99 (millions of U.S. dollars) Purpose of loan Principal Disbursed Approval date Roads Rehabilitation 50.0 58.7 6 May 1986 Power Rehabilitation/Energy 40.0 44.3 6 May 1986 Multisector Rehabilitation Ia 50.0 54.5 20 November 1986 Multisector Rehabilitation I 46.2 48.0 20 November 1986 Telecommunications II 23.0 24.0 28 May 1987 Multisector Rehabilitation I 30.0 30.1 19 January 1988 Multisector Rehabilitation 26.0 25.4 19 January 1988 Agriculture Exports Rehabilitation I 30.0 17.7 29 March 1988 National Agriculture and Livestock 8.3 8.6 13 December 1988 Industry and Trade Adjustment Credit 135.0 129.2 13 December 1988 Industry and Trade Adjustment Credit 12.5 12.6 13 March 1989 Agriculture Extension 18.4 19.3 21 March 1989 Tree Crops 25.1 19.2 27 June 1989 Industry and Trade Adjustment Credit 10.3 12.0 12 October 1989 Ports Modernization 37.0 31.7 27 February 1990 Health and Nutrition 47.6 42.5 6 March 1990 Agriculture Adjustment Credit 200.0 207.6 29 March 1990 Education Planning and Rehabilitation 38.3 33.8 22 May 1990 Roads I 180.4 157.6 31 May 1990 Agricultural Adjustment Credit 16.1 16.4 13 December 1990 Petrol Rehabilitation 44.0 29.4 15 January 1991 Railways Restructuring 76.0 48.3 13 June 1991 Urban Sector Engineering 11.2 10.9 30 July 1991 Financial Sector 200.0 211.0 14 November 1991 Agricultural Adjustment Credit 11.3 11.8 26 November 1991 Engineering Credit 10.0 10.8 14 January 1992 Forest Resources Management 18.3 18.5 11 February 1992 Financial and Legal Management Project 20.0 15.5 9 July 1992 Financial Sector 11.3 11.4 23 December 1992 Telecommunications III 74.5 47.1 27 April 1993 Power VI 200.0 164.5 6 May 1993 Public Sector Management -;:34.9 32.0 10 June 1993 Agriculture Sector Management Project: - 24.5 15.2 20 July 1993 Roads II 1702 20.4 7 April 1994 Mineral Sector Development -12.5 9.0 28 July 1994 Financial Institutional Development 10.9 10.3 3 August 1995 Urban Sector Rehabilitation 105.0 15.9 23 May 1996 National Extension Project Phase II 31.1 9.0 11 July 1996 River Basin Management 26.3 6.7 11 July 1996 Lake Victoria Environment 10.1 1.7 30 July 1996 Structural Adjustment Credit I 125.0 49.4 20 June 1997 Structural Adjustment Credit I 3.9 3.8 20 June 1997 Human Resource Development I 20.9 5.0 7 October 1997 Structural Adjustment Credit I 2.6 2.5 15 December 1997 Agriculture Research 21.8 2.0 29 January 1998 Tax Administration 40.0 0.0 30 March 1999 Note: Boldface indicates policy-based loan. a. Projects appearing more than once reflect the release of multiple tranches. Source: World Bank project files. 304 Post-Socialist Reformers Sectoral adjustment was supported through loans for industrial reform (1988) and the agriculture sector (1990). The most recent adjustment lend- ing was the Structural Adjustment Credit in 1997. Loans for improved public sector management and for some reforms in the judiciary appeared relatively late. Apublic sector management project was approved in 1993, and a financial and legal management project to "improve the legal regu- latory framework and to improve the administration of justice" was approved in 1992. General conditionality in Bank lending has been marked by increased cross-sector conditionality and the use of multiple floating tranches. For example, the most recent adjustment credit consists of five tranches. The first tranche was released when the cross-sectoral conditions of credit effectiveness were met. The other four (floating) tranches were designed to be released when specific sectoral conditions were met for (a) devel- opment expenditures and the social sector; (b) divestiture of the Na- tional Bank of Commerce; (c) parastatal reform; and (d) petroleum sector liberalization. Conditionality in the Bank adjustment credits has focused on numerical targets. The current adjustment program targets a fiscal balance of 0.9 percent (after grants) of GDP before Board presentation. For example, on trade, earlier adjustment lending targeted the number of items on the negative list for imports, as well as the number of tariff rates and their level. In some cases, the conditions have focused on the preparatory work required for effective implementation. This includes the issuance of terms of reference for consultants, establishment of par- ticular units, drafting of action plans, revision of legislation, and so on. The IMF targets have tended to be more strictly numerical and have focused on macroeconomic indicators such as inflation and money sup- ply growth. Tanzania seems to be facing more difficulties meeting the condition- ality criteria for structural reforms. Privatization and reform of the fi- nancial sector have been sticking points in the dialogue with the IMF and World Bank. Targets have been missed on a number of occasions. For example, financial sector reform (involving the privatization of a number of state-owned banks) has dragged on since 1995. These diffi- culties appear to be partly linked to political pressures and partly to inadequate technical capacity to implement the agreed-upon reforms. Multilateral lending has worked effectively for those aspects of liber- alization that were less complex administratively. For example, agricul- ture, foreign investment, and external trade have all been progressively liberalized. Less progress, however, has been made in those areas that are more complex administratively and that require more frequent moni- toring. For example, privatization of the National Bank of Commerce and civil service reform have taken longer than anticipated. Other gen- eral issues that have contributed to slow and unsatisfactory progress in Tanzania 305 the implementation of projects include the lack of counterpart funds for IDA-financed projects, cross-sector conditionality, and difficulties with Bank procurement procedures. Overall, however, disbursement ratios have risen for the Tanzania portfolio and they are roughly at par with the World Bank average. The multilateral organizations have been pushing the basic structural adjustment agenda in Tanzania as in other African countries, and dur- ing the 1990s the policymakers have largely accepted this. They have also supported local research foundations (such as ESRF and REPOA) financially or through consultant contracts, and this also may be con- tributing to the generation of ideas. In Tanzania, it seems the university (notably the economics faculty) was very influential in generating new ideas for reform. Economic and sector work by the World Bank on Tan- zania has become increasingly infrequent, as projects have overwhelmed the country program. The amount spent on economic and sector work has declined in each of the past four years.3 There now seems to be a renewed effort to carry out more economic analysis but it remains to be seen whether this will translate into more resources for that work. One exception to this trend was the collaborative Country Economic Memo- randum/Poverty Assessment in 1996, which was well funded and in- volved UNICEF, the World Bank, and bilateral donors. Relations with Bilateral Donors During the reform period (1986-96), Tanzania was critically dependent on DAC bilateral donors. As noted above, the socialist bloc and China, which became important donors in the 1970s, drastically reduced their aid programs to Tanzania during the 1980s because of their internal po- litical and economic problems. Within the donor group, three categories of donors can be distinguished: • The Nordic countries, most of which have Tanzania as their main aid recipient, and which have a long tradition of consulting among themselves on aid issues and of coordinating their actions at head- office and local levels. X Major industrial countries in the DAC group, including Germany, the Netherlands, Canada, the United States, the United Kingdom, Italy, Japan, Belgium, and (particularly for Tanzania) Ireland. * A large number of small donors such as the former socialist bloc countries, China, and the bilateral Arab funds. Since the early 1970s Tanzania has had a very special relationship with the Nordic countries, in particular Sweden. This country has con- sistently provided substantial resources to Tanzania and has positioned itself at the head of the list of the country's aid donors. By 1983 Tanzania's 306 Post-Socialist Reformers relations with practically all the bilateral donors were under great strain. The fact that even the strongest supporters scaled down their assistance contributed to the change in the country's policies. The view among the bilateral donors today is that Tanzania is trying to do the right things in terms of policy reform, but that administrative capacity and competence are very low. This means that donors are re- luctant to shift responsibilities to the recipient government, although in principle they realize that this is necessary for aid to be effective and sustainable in the long term. Donors are also reluctant to shift too much of their efforts into building administrative capacity, since this might jeopardize the management and success of their conventional projects. The Swedish slogan is "Partnership in Development," but much remains to be done before this becomes a reality on the ground. There is in any case an improvement in terms of policymaking capacity, insofar as it is now more often the government that writes policy and position papers spelling out the priorities. Previously ministries often referred to World Bank documents when asked about government policy. There is work underway to design comprehensive sector programs in education, health, roads, water, and local government, but how well these will function will depend both on the capacity of the government (par- ticularly at the local level) and on the willingness of donors to change to another type of control. The latter will depend on the former, which means that it is hard to be optimistic about things changing very much in the short run. Donors will probably maintain a high degree of control of their own slices of the "joint" programs to make sure that the activities are in accordance with their list of priorities and that the money can be accounted for. It is an eternal dilemma in the aid business: on the one hand the donors want to give control to the recipient government, but on the other hand they have to report to their home constituencies, par- ticularly the parliaments, about how the money allocated to aid is used. This also puts a brake on coordination efforts. Some donors are more willing to go down this road than others are, but it is very hard to be- lieve that all will agree to be coordinated. And this means that aid de- pendence will persist. The recipient should take a more strict line with the donors to make sure that aid fits in with domestic priorities, but in practice there is still a very flexible attitude to donor demands. Tanzania does not say "no" to donors who are willing to put their money into a project, even though it may be out of line with the government's priori- ties and hard to finance in the long term. An efficient partnership or efficient local control requires a well-run bureaucracy, and that in turn requires a competent staff and organiza- tion. As long as the government pays very low salaries relative to the private sector, it will find it hard to recruit and retain the best people. On top of this there is a virtual stop on new hiring, which means little new Tanzania 307 blood coming into the system. Donors are involved in various topping- up exercises, either to key personnel or key units to bring about a more efficient organization, but this in turn complicates the general overhaul of the civil service. The government has set forth principles of civil ser- vice reform that include gradually increasing real wages, but they are already behind on the latter count. Given the constraints of the cash bud- get and the limited ability to collect taxes, it will be very hard to achieve the goals set out in the plan. One aspect of partnership or increased ownership is that the recipient should shoulder a large part of the recurrent costs of a certain project. Typically it is written in the initial agreement that a certain percentage will be paid by the government and that this will gradually increase. The government is willing to sign agreements like this although it should be clear that in reality they will not be able to pay up. The assumption is that they will deal with that issue when it occurs, and typically the do- nor does not withdraw from their project but tries to find some way to salvage the project anyway. Thus there is a tendency toward hypocrisy on both sides of the table. It is also hard to coordinate within the government. In the case of the education program the coordination exercise involves as many as nine ministries, which suggests that in the end it will be hard to get a simple and efficient management structure. A wide range of players and inter- ests are involved on both the donor and the government side, and in a weak system such as the Tanzanian one it will be hard to get the bureau- cracy to work well. Moreover, in a wide-spanning aid relationship such as Tanzania's, it should not be surprising to find that the interests of players on both the donor and recipient sides frequently overlap or in- terlock, thus evolving into relationships that are outside the officially established channels. One early step toward changing the relationship between donors and the government was taken in September 1996, when the Nordic coun- tries (Denmark, Finland, Norway, and Sweden) entered into a partner- ship agreement with Tanzania. The agreement required the government to produce a policy platform, respect democracy and human rights, fight against corruption and for good governance, emphasize macroeconomic stability and domestic resource mobilization, and focus government ac- tivities on core functions, while civil society and the private sector would complement these efforts. Five general principles for the collaboration were set forth: * The program must be fully owned by Tanzania and formulated according to Tanzania's visions and policies. * Tanzania is responsible for the programs and accountable for re- sources provided. 308 Post-Socialist Reformers * There should be shared financing, with the Tanzanian share in- creasing over time. * There should be a move from discrete projects to sector program support. * The Nordic countries should support only those activities to which the government of Tanzania gives priority (Embassy of Sweden 1998: 4-5). These principles obviously represent a radical break with past prac- tice; but the actual practice is, of course, still a long way from the above. Sweden, for example, has a long list of goals that must be met for aid to be considered at all, so it is somewhat hypocritical to say that the pro- grams must be fully owned and formulated according to Tanzanian pri- orities. When the government formulates plans, it certainly keeps in mind what the donors want to hear. It is only when transfers are in a general form, for example through transfers to the Multilateral Debt Relief Fund, that one can really talk about ownership. Still, other forms of collabora- tion can be characterized as partnerships, although the donor is in a much more powerful position than the recipient government. The do- nor can always threaten to withdraw its money, while the recipient gov- ernment has no such course of action open to it. In contrast to the Nordic countries, which have been working behind the scenes to provide encouragement and support to Tanzania, the next category of major donors, led by the United States, the United Kingdom, and Germany, have traditionally taken a hard line in their relations with Tanzania. They have tended to identify more strictly with the IFIs and to trigger their aid disbursements more in line with World Bank and IMF judgments. However, support for reform considerations is also balanced against a complex set of foreign policy interests. For instance, it has been argued that during the second half of the 1980s, Tanzania did not do well in aid terms with the British government of Margaret Thatcher because of its role as a Frontline state, supporting independence for countries in south- ernAfrica and vocally criticizing Western countries' continued economic ties with South Africa. With the change to a Labour government in the United Kingdom, the country's aid disbursements to Tanzania doubled. Domestic economic interests of the donors have also been a signifi- cant influence. For instance, it has been reported that in the late 1980s U.S. pressure on Japan to reduce its balance of payment surpluses led to a major increase in Japanese aid to developing countries, including Tanzania. Japan then hosted the first Tokyo International Conference on African Development (TICAD I) in 1993, followed by TICAD II in 1998, as an indication of its intention to shape an independent African political and economic cooperation policy. The other main objective of Tanzania 309 the TICAD conferences was to highlight Africa's development prob- lems and prospects, elevating their importance in the eyes of the inter- national community. Similarly, France in recent years was reported to be planning a significant expansion of its aid program to Tanzania and other countries in order to counterbalance efforts by the United King- dom and the United States to expand their operations in francophone Africa. As a former colonial power, Germany has maintained a special interest in Tanzania and has, among other things, tended to fund some projects that reflect the historical connection. The Arab funds have con- tinued to provide concessional lending without much policy linkage, except when they are involved in co-financing with the Bretton Woods institutions or the AfDB. In 1988-92, Italian aid to Tanzania rose sharply, mainly because of a large increase in support for projects operated by Italian contractors. Tanzania's relations with donors came under great strain in 1994-95, toward the end of the government of President Ali Hassan Mwinyi. But recent assessments indicate a substantial improvement in donor rela- tions. A factor that has contributed to this improvement has been cited before in reference to the Helleiner report of June 1995. Following an extensive discussion of the report, an agreement was reached in January 1997 between the donors and the government on a new mode of interac- tion based on a partnership approach, as discussed below. Many donors are cautious, slow, and reluctant or unable to make the shift, especially in terms of making contributions to "basket funds" in support of these sectoral programs. The tendency is for each donor to cling to its particular area of specialization (for example, capacity build- ing for private sector development in the case of the United States, inte- grated rural development in the case of the Irish, and structural adjustment in the case of the Swiss). However, an overall assessment is that today many bilateral donors see great potential in Tanzania. While they recognize that the country's structural problems persist-an inefficient civil service, a slow rate of project implementation, poor accountability in project management- they seem to be encouraged by an improved policy environment and perceived willingness and determination on the part of the government to implement reforms. Donor Coordination: Transition in Slow Motion Tanzania is presently in the forefront of the discussions on transforming the aid relationship into a partnership between donors and recipient. Sida, the Swedish development cooperation agency, carried out an in- formal evaluation of what has been achieved since the Helleiner work- shop in January 1997 (Embassy of Sweden 1999a). The Swedish study 310 Post-Socialist Reformers reviews the list of aims set out in the agreed minutes of that meeting, and reports on progress or lack of it: 1. Donors should provide the government of Tanzania with information about planned and actual aidflows. This has proved to be very difficult to achieve because of the quite different ways of accounting for flows among the various donors. Attempts are again made in this year's Public Expenditure Review. 2. Donors shouldfollow the Tanzanian government's prioritiesfor develop- ment projects and seek to exitfrom noncore activities. Tanzania claims that the number of projects has been reduced very significantly from about 1,500 to between 500 and 600, but the information is scanty It should also be noted that development expenditures in 1997-98 were 60 percent below target. This suggests that the sys- tem still absorbs resources inefficiently. The shortfall may be due to, for example, the lack of counterpart funds, failure to meet con- ditions for the project, or problems in implementation. 3. Aid coordination should be improved. The education and health sec- tors should become test cases for aid coordination modalities. There have been steps in this direction (see, for example, Tanzania 1999), but disagreement among donors as to how to proceed has slowed the process. Still, the education sector development plan has just undergone a very comprehensive appraisal by joint teams com- posed of, on the one hand, foreign consultants appointed by do- nors who are active in supporting the sector, and on the other hand, local consultants appointed by the Ministry of Education. The 60 consultants spent two weeks making the appraisal and submitted the report to the ministry The report will provide input into the negotiations between donors and Tanzania-an example of joint action in the policy reform process. 4. The Policy Framework Paper should be made public and the government should consult with civil society. The PFP has been made public, but so far the consultations with civil society have not taken off. 5. Donor discussions on debt relief and development assistance should be integrated. A Multilateral Debt Relief Fund has been established, although it is a bit unclear what has been achieved in terms of inte- grating the discussions. 6. Steps should be taken to develop the planning andfinancial management capacity in the Treasury and the Planning Commission. There are sev- eral programs underway that seek to build capacity. The further role of the Planning Commission is unclear, however. 7. Tanzania should take the lead in the preparation of Policy Framework Papers and Public Expenditure Reviews. This has happened. Tanza- nia has taken a lead role in the drafting and discussion of the PFP, Tanzania 311 and the relevant ministries are more involved than they were be- fore. The same can be said of the PERs. The work is now done on three levels. First, there is a steering committee that oversees the process at the policy level. This consists of the permanent secretar- ies in the relevant ministries and is chaired by the permanent sec- retary in the Ministry of Finance. Second, there are technical advisory groups that oversee the implementation of the technical studies and the PER process. This group consists of senior officials from the relevant ministries and is chaired by the deputy perma- nent secretary in the Ministry of Finance. Third, there is a working group to monitor and coordinate progress on the technical studies. This is also chaired by the deputy permanent secretary in the Min- istry of Finance and includes key personnel from the relevant min- istries and representatives of donors and research institutions. This group has met twice a month during the process. There is also a small secretariat, consisting of representatives from the ministries and the World Bank, that provides logistical support and other help. The approach is very participatory and is open to ministries, the research community, and donors. The decentralization of the Pub- lic Expenditure Review, the Country Assistance Strategy, the Struc- tural Adjustment Credit, and the Country Economic Memorandum to the local World Bank office seems to have helped to achieve a more consultative approach. The fact that the highly respected pro- fessor Benno Ndulu is located at the World Bank office in Dar es Salaam has also helped to improve local control and access. 8. Information, accounting, and reporting of assistance should be standard- ized. This has not been done. 9. Best practice projects in Africa should be collected and analyzed. This has not been done by the government, and maybe it should not be its job to do so. 10. The governmnent's capacity to manage should be enhanced and the donors should be willing to transfer more responsibility to the government of Tanzania. This is an ongoing process in which Sweden is taking the lead. In practice the process is very slow, however. There is deep concern about the problem of corruption, which makes it hard to transfer responsibility. Unless further steps are taken to control corruption and improve transparency, progress will be difficult to achieve. 11. With regard to the relationships with the civil service, parallel or top- ping-up arrangements should be minimized. Again, not much has hap- pened apart from a report identifying and discussing the problem. 12. Coordination meetings should be rationalized. The Ministry of Finance has started quarterly sectoral reviews in some sectors (health, educa- tion, and roads) to deal with the donors in a more coordinated way. 312 Post-Socialist Reformers 13. The use of technical assistance should be reviewed. This has not been done. 14. The recommendations should be followed but by a joint committee. A development forum has recently started to function after a long period of dormancy. 15. Further stocktaking should take place at the CG meetings. This did oc- cur at the CG meeting of December 1997, which was held in Dar. The venue of that meeting also allowed broader participation of Tanzania officials in the CG. The 1999 meeting was also supposed to take place in Dar, but the donors preferred to meet in Paris. Helleiner himself recently returned to Tanzania to review the progress in implementing the agreed actions. He found that there has been a sig- nificant advance (Helleiner 1999), especially in establishing government leadership in macroeconomic management. However, his review also notes that progress at the sectoral level has been slow. An important test of the new relationships has been the introduction of sector-wide pro- grams in education, health, and a refurbished road sector. This calls for a significant change from the traditional approach based on directly con- trolled projects. It also requires the timely provision of complete infor- mation, increased emphasis on sector-wide policies (and the capacity and willingness for frank dialogue on them), the turning over of some budgetary control and procurement functions to a local authority, and, above all, the acceptance of government leadership and ultimate respon- sibility for the sector program. In conclusion, much remains to be done for the process to move for- ward. The ideas from the Helleiner workshop make sense, but there are problems on both the donor and the government side. A general obser- vation is that Tanzania is rather good at launching new initiatives and setting up interesting plans, but has poor capacity for implementing them and living by the new principles once they have been established. This was largely true in the past, when such initiatives were not backed by institutional arrangements to keep the implementation going. With the current emphasis by the government on institutionalization of initiatives, there seems to be a chance for better performance than in the past. Tran- sition in Tanzania does not happened with a "big bang." It is, rather, transition in slow motion. This is of course a concern, because if there is no take-off the system may eventually implode. POLICY REFORMS AND INSTITUTIONAL DEVELOPMENT The following sections analyze the causes and processes of the reforms. However, before entering into this discussion we need to provide a brief background of the events leading up to the Tanzanian economic crisis. Tanzania 313 We will then provide a chronology of the process of policy reform since 1979, identifying the main phases in the donor-recipient relationship. We look specifically at inflation and fiscal policy (macroeconomic man- agement) and at trade policy (an example of structural policies). The macroeconomic and structural policies considered here have been shown in empirical studies to affect developing countries' growth (Sachs and Warner 1995). But successful development also depends on institutional development, so we discuss institutional reform such as the privatization process and public sector management. From the Arusha Declaration to Crisis Following independence in 1961 the government concentrated on the Africanization of the public sector, but otherwise there was continued reliance on the private sector. An import substitution policy served to promote industrial development, but the economy remained fairly open. It continued to be highly dependent on revenue from commodity ex- ports. During 1961-67 per capita income grew by 2 percent per year, the highest rate of any period in independent Tanzania (see table 6.4). This was a time of macroeconomic stability with low inflation and a satisfac- tory balance of payments, and capital formation increased steadily be- tween 1963 and 1967. However, the government was not satisfied. It was concerned about the apparent increase in inequality, and found industrial growth to be below TABLE 6.4 TANZANIA: DATA FOR THE PRE-ARUSHA PERIOD, 1961-67 Indicator 1961 1962 1963 1964 1965 1966 1967 Per capita income growth (%) -7.5 4.3 1.2 3.6 -0.2 9.9 1.8 Population growth (%) 2.8 2.8 2.9 2.9 2.9 3.0 3.0 Urbanization (%) 4.8 4.9 5.1 5.2 5.3 5.6 5.9 Labor force in agriculture (%) 92.4 92.1 91.9 91.7 91.4 91.2 90.9 Monetary growth (%) - 21.7 25.6 -15.2 31.8 476.7 13.1 Gross investment (% of GDP) 13.7 11.6 10.7 12.0 13.9 15.1 18.9 - Not available. Sources: Compiled by Bigsten and Danielsson 1999. Income and investment data from the Bureau of Statistics, Dar es Salaam. Other data from World Development Indicators 1998. 314 Post-Socialist Reformers expectations, as was the inflow of external resources. President Nyerere felt that a new strategy was necessary to speed up development. The Arusha Declaration of 1967 was a watershed in Tanzanian politi- cal and economic history. The government abandoned the cautious poli- cies pursued up to that point and launched a strategy of "African socialism." There was concern that external dependence was too high, and that steps should be taken to create an internally integrated economy The necessary structural transformation would be facilitated by increased state control. By the early 1970s, the government had consolidated its hold on most aspects of formal economic activity. Banking and large portions of the industrial sector had been nationalized, the bulk of international trade and private retail trade had been confined to state agencies, and admin- istered prices had largely replaced market prices. A foreign exchange allocation system was developed in response to the balance of payments crises of the early 1970s, and a National Price Commission was estab- lished to set prices for a vast range of commodities. By the mid-1970s, most of the rural population had been forced to resettle in "Ujamaa vil- lages" as part of a strategy to promote cooperative agriculture, and mo- nopoly government marketing boards had replaced peasant marketing cooperatives. In 1974 the government began to implement its Basic Industries Strat- egy (BIS) of import-substituting industrialization, but it soon had to post- pone it in the balance of payments crisis that accompanied the first oil shock. The external position improved dramatically during the coffee boom (1975-78), and the government revived the BIS in a decision that came to dominate the allocation of foreign exchange and the pattern of domestic investment. Table 6.5 shows that investment was high through- out the 1970s, although it declined somewhat in 1973-75. The invest- ment program was state-led, but many of the investments undertaken in the 1970s were supported by donors. The crisis was heralded by the breakup of the East African Community in 1977 and the rapid depletion of international reserves following the end of the coffee boom in 1978 and an abortive import liberalization that same year. Donors disagreed about Tanzania in the second half of the 1970s. Sev- eral bilateral donors were supportive, while the World Bank and the IMF became more and more critical of policies pertaining to agriculture, domestic pricing, the BIS, and the exchange rate (Collier 1991). Since President Nyerere's position was strong both domestically and with many bilateral donors, he was able to take a hard line with the IFIs. During 1968-78 growth in per capita income was still positive at 0.7 percent per year (see table 6.5). The growth was led by the public sector, while Tanzanian exports, which continued to be dominated by tradi- tional agricultural exports such as coffee and cotton, stagnated. The policy TABLE 6.5 TANZANIA: DATA FOR THE PRE-CRISIS PERIOD, 1968-78 Indicator 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 Per capita income growth (%) 2.1 -0.7 3.0 0.8 2.3 0.5 -0.5 2.9 2.3 -2.7 -1.9 Population growth (%) 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.1 3.1 3.1 Urbanization (%) 6.1 6.4 6.7 7.4 8.1 8.7 9.4 10.1 11.0 12.0 12.9 External debt (millions of U.S. dollars) - - 212 284 407 619 900 1,170 1,380 1,700 1,970 Labor force in agriculture (%) 90.6 90.3 90.1 89.7 89.3 88.9 88.5 88.1 87.7 87.2 86.7 Monetary growth (%) 17.8 9.2 12.0 18.2 17.7 18.2 22.1 24.4 25.1 20.2 12.6 Gross investment (% of GDP) 18.4 16.3 22.9 26.8 23.6 22.6 21.6 20.8 29.0 29.4 33.8 - Not available. Sources: Compiled by Bigsten and Danielsson 1999. Income and investment data from the Bureau of Statistics, Dar cs Salaam. Other data from World Develop- ment Indicators 1998. 316 Post-Socialist Reformers environment in Tanzania was hostile to both traditional and nontradi- tional exports. Government monopolization of marketing in the 1970s undercut the rewards to peasant production for export, while inefficien- cies drove down producer prices. Other factors included the relocation of rural producers in the "villagization" drive, the high effective protec- tion of the import-substituting industrial sector, which turned relative prices in favor of the urban areas, and the use of trade controls rather than exchange rate adjustment as a means of adjusting to external shocks. The coffee bonanza concealed the weaknesses of the system for a few years, but when it ended the inefficiencies of the economy became apparent. Chronology of Policy Reforms 1979-99 In looking at the overall pattern and phases of the reform process, it is possible to identify five policy phases in the last 20 years in Tanzania. 1979-82: Stalemate After the collapse of the commodity boom of 1975-78, the Uganda war in 1978-79, and the second oil crisis in 1979, the economy faced very severe problems. In response, Tanzania agreed to a standby credit with the IMF in 1979, but after the government failed to observe the agreed budget ceiling the agreement was cancelled. During discussion of the release of the second tranche, the government ordered the head of the visiting IMF mission to leave the country The minister of finance, Edwin Mtei, was perceived by President Nyerere as being too close to the IMF and was removed. Mtei was succeeded by Jamal, who was less eager to reform the economy. Nyerere made a speech in which he criticized the IMF for trying to set itself up as the "International Ministry of Finance" and rejected its demands. Tanzania then turned to the World Bank for assistance, but although the Bank had been supportive throughout the 1970s, it now refused to support the Tanzanian government. There was then little communication for about two years. In1981 the Bank indicated that it could provide quick-disbursing aid if the Tanzanian government undertook certain policy changes and reached an agreement with the IMF. The latter demanded a 50-60 per- cent devaluation, a significant reduction of the budget deficit, removal or reduction of consumer and producer subsidies, positive real interest rates, and higher agricultural prices and import liberalization. The gov- ernment proposed some new measures in the National Economic Sur- vival Program (NESP) of 1981, which was essentially part of an application to the IMF. The goals stated in the plan were to increase ex- port revenue and eliminate food shortages through various supply-side Tanzania 317 measures and to reduce public expenditures. This was to be achieved through tighter state control of public expenditures and increased pro- duction. The plan was wholly unrealistic, and the IMF found it to be completely unsatisfactory. Anti-reform elements in Tanzania argued that the causes of the crisis were external problems and not internal policy weaknesses. The government's position was that Tanzania was a struc- turally weak economy and that the remedies proposed, including the dismantling of the socialist structures and deep cuts in social spending on education, health, and water, would not help to revive the economy. The government was not ready to agree to carry out significant reforms. This partly explains why NESP I (1981) and NESP 11(1982) were more exhortations than policy change documents. Moreover, at this time there were very few senior officers in central ministries capable as a group of undertaking independent macroeco- nomic and sectoral analysis to generate new ideas and develop new pro- grams to address the crisis. Sufficient local intellectual capacity was not available to make it possible for the government to engage in a mean- ingful debate with the IFIs or even to counterbalance the ruling party's inclinations toward maintaining the status quo. There was excessive re- liance on a handful of expatriate advisers, most of whom preferred to toe the party line. Sometimes these advisers held divergent policy per- spectives and it was a task to maintain harmony among them. The faculty at the University of Dar es Salaam had generally been weakened, either by socialist ideology or by opportunists within the insti- tution who were eager to please the party leadership. To identify with the party could open opportunities for accelerated promotions or for transfer to more remunerative positions in parastatal organizations or political structures. Because of a severe shortage of qualified people, such trans- fers were frequent at the time and contributed to the lack of consolida- tion of intellectual policy groups within the government or around it. 1982-85: Hard Internal Debate The country was now in a desperate situation, with the exchange rate greatly overvalued and exports at very low levels. The bilaterals and the World Bank were trying to find some middle ground and sponsored a technical advisory group with well-known economists, who were asked to come up with a compromise package. The group proposed a more moderate devaluation, but the government did not accept even this. The government used the report to produce its own reform package, which was a dilution of the policy changes proposed by the advisory group. This revision was done by the Planning Ministry, then under Minister Malima, who in 1980-92 was basically anti-reform although he later (1992-94) became more pro-reform. The government turned to the Nordic 318 Post-Socialist Reformers bilaterals to fill the gap left by the IFIs, but by now they also were be- coming more critical and did not provide the extra funds. There was thus a steep decline in aid flows between 1982 and 1985, which was period of extreme economic crisis. Still, a sign of some movement on the government side was the ap- pointment in 1982 of the prime minister, Cleopa Msuya, as minister of finance. He was more aware of the need to mend fences with the IFIs. Together with Rutihinda, who was the permanent secretary in the Min- istry of Finance, Msuya tried within the government to push for reforms. There were similar attempts at the Bank of Tanzania. Still, the support- ers of reform in the Ministry of Finance had to work hard to build up support for it within the government. The cabinet was dominated at the time by hard-line left-wingers opposed to the reforms, with Finance Minister Msuya as one of the few advocates of reform. The minister of planning, Malima, was a leading opponent of the reforms, and he quickly brought in Ajit Singh from Cambridge to write a counterproposal to the reform plan. There was a head-on collision between Msuya and Malima, in which the latter even wrote a memo to the president accusing the reformers of being "traitors." The parastatal establishment was also a stumbling block in the process, although some could see that devalua- tion would make it possible for them to improve their results. It took time for the government to realize that the parastatals were a problem. First, there were attempts to peg the desired exchange rate at a level that would be sufficient to clear the losses of parastatals. This ap- proach was much in favor in 1983 and 1984. Second, the extreme oppo- nent of the reforms, Malima, submitted a paper to the ruling party suggesting that the problem of parastatals was rooted in the existence of the private sector. The private sector would have to be phased out in order to give room for the parastatals to recover! Apparently he believed that the private sector was doing harm to the parastatals through its dealings with them, for example by supplying them with overpriced goods and services. Public opposition to the policies of the early 1980s was limited, as most Tanzanians chose instead to withdraw into subsistence activities or find other ways to deal with shortages and regulations. Parallel mar- kets emerged, as did illegal cross-border trade. The rapid expansion of the underground economy as well as capital flight undermined the tax base. Semi-public discussions were started at the behest of the reformers within government. An informal network of economists from the Uni- versity of Dar es Salaam, together with private consultants and person- nel from key ministries such as Agriculture, Transport, Industry and Trade, and the Planning Commission, was set up to provide a forum for debate and to increase the understanding of the politicians and policymakers. Tanzania 319 The strong political leadership made it possible to push through im- portant decisions (or to reject others) without proper analysis and de- bate. Nevertheless, there were courageous individuals, notably Mtei and Msuya, who could put forward reform ideas to the party hierarchy and the cabinet. Such people faced stiff resistance from the party, and, as so often happens, vacillating positions developed within the cabinet de- pending on who could get the president's ear. But the increasingly pre- carious social situation arising from severe shortages of consumer goods, especially in the large cities, forced the leadership to search for solutions and ultimately to consider seriously the adoption of more comprehen- sive reform packages. Up to 1983 the government essentially preserved existing policies. Instead of introducing measures to deal with the imbalances, the gov- ernment pursued campaigns against corruption and profiteers. Stewart, Klugman, and Neyapti (1999:82) argue that a side effect of the political stability that characterized Tanzania was that it postponed pressures for economic reforms that were essential for economic development. Adam and others (1994) argue that the backing from a range of bilateral donors also made it possible to delay adjustment. Nonetheless, a second homegrown structural adjustment program was launched in 1983. It was more ambitious than the previous one, but still did not change any of the major parameters of policy. The pro- gram sought to identify a series of structural problems in the economy and then adjust the type of state intervention so that the economy would do better. It was not a matter of liberalization and deregulation. The main aim was to increase agricultural production and exports, and the measures included a modest devaluation. This did have some positive effect on exports and agriculture, but fell far short of closing the gap between the official and parallel rates. There was a further tightening of fiscal and monetary policies, but the measures were still insufficient to reverse the trend. The situation remained grave, and there was mounting public criti- cism of the shortages of goods. The increasingly tough security mea- sures against smugglers and profiteers only increased the shortages, and it was mainly the smaller culprits who were sentenced in the end. Cam- paigns against corruption and profiteers were directed at improving the distribution of goods and curbing the practice of stockpiling since it was believed that hoarding or unfair distribution of goods caused the short- ages. When the shortages became worse in spite of the campaigns, the champions of this approach began to realize that the problem was more deeply rooted. This realization influenced the decision to permit partial import liberalization in 1984. In fact, Sokoine, then prime minister, led the campaign against profiteers, and later he also led the initiative to- ward partial liberalization starting with transport. Imports of pick-ups 320 Post-Socialist Reformers were allowed and supply improved, and these positive results were used to argue for further import liberalization. The overall failure of the structural adjustment program was aggra- vated by two factors. First, the receding aid inflows reduced the already low levels of capacity utilization and thus worsened the shortage of goods. Second, the allocation of whatever foreign exchange was avail- able through CIS tended to favor projects, which earlier had been funded by the respective donors rather than being allocated according to effi- ciency criteria. The few resources available were spread too thinly to be effective. In 1984 economists from the University of Dar es Salaam started to hold public meetings at which they advocated liberalization measures. The initial discussions targeted policymakers and politicians and some private sector leaders. These meetings were, of course, against the offi- cial party line, but they were supported by the reformers within the gov- ernment, for example in the Ministry of Finance and the Bank of Tanzania. That same year, in the face of extreme goods shortages, the government finally made a more significant move. It chose to allow own-funded imports, it devalued the currency, and it increased agricultural prices by about 30 percent. There was an export retention scheme, and user charges were introduced in health. This was the first step on the road to the res- toration of relations with the IFIs. There was now some improvement in the economy and imports could increase somewhat. The government tried to mend relations with the IMF, which felt that the new reform program was a least a basis for further discussions. Talks were reopened between the government and the IMF at the end of the Nyerere era. 1986-92: Initial Reforms and Aid Boom The years between 1986 and 1992 saw the second aid boom (the first having occurred in the 1970s).When Mwinyi took over as president some initial steps had already been taken on the reform route with the Eco- nomic Recovery Program. The government reached an agreement with the IMF in 1986, which also meant that other donors came back in force. Joint government-donor meetings were initiated and have continued with more or less intensity since then.4 ERP I covered the period from 1986 to 1989. The targets were to: * Increase output of food and export crops through appropriate pric- ing strategies * Increase capacity utilization in industry from 20-30 percent to 60- 70 percent * Restore internal and external macroeconomic balance through "pru- dent" monetary, fiscal, and trade policies Tanzania 321 * Reduce inflation from 30 percent to 15-20 percent * Increase foreign exchange earnings from exports from US$400 mil- lion to US$630 million by 1989 * Rehabilitate and maintain basic social services. The policy instruments to be used included: * Raising agricultural prices by 5 percent per year or to 60-70 percent of the world market price, and reducing the use of price controls * Sufficient devaluation to eliminate exchange rate overvaluation by mid-1988, with a subsequent move to market-based allocation of foreign exchange * Deregulation of both imports and local trade * Control of public expenditure by continuing the austerity measures started in 1984-85, limiting government recurrent expenditure, re- ducing deficit financing from the banking system, and rationaliz- ing and reducing parastatal activities and subsidies to them * Tightening monetary policy by raising interest rates until positive in real terms, adopting tight credit ceilings, and limiting the growth of money supply to 15-20 percent per year * Improving revenue collection through close monitoring and look- ing for ways to increase efficiency and broaden the tax base * Non-price measures to increase agricultural production and steps to improve access to consumer goods, distribution of agricultural inputs, and transport infrastructure; general freeing of consumer goods markets. The Multisector Rehabilitation Credit agreed in 1986 was funded with US$300 million, half from IDA and half from other donors. It concerned prices, fiscal management, and exchange rate adjustment. The main aim of this program was to increase agricultural production through devalu- ation and improvements in marketing and transport. The allocation of the funds was left to the government, conditional on compliance with certain macroeconomic conditions that were imposed through the dis- bursement in tranches. At the end, in 1988, some money went through the OGL. Under the program there was a sharp devaluation, limits on the budget and balance of payments deficits, and limits on domestic credit. Still, the implementation was not as smooth as was hoped ini- tially. The second program entered into was the Industrial Rehabilita- tion and Structural Adjustment Credit of US$242 million, agreed in December 1988. Its main focus was trade liberalization, tariff and sales tax reforms, and industrial restructuring. Funds were now channeled through the OGL system. A new agreement covered 1989-92. It consisted of an IMF Enhanced Structural Adjustment Facility (ESAF) plus three different World Bank 322 Post-Socialist Reformers projects. The ESAF covered liberalization of the foreign exchange mar- ket. The other programs were the following: * The Tanzania Agricultural Adjustment Credit of US$338 million started in 1990, again with IDA as the major source. The main con- cern was now liberalization of the marketing of agricultural inputs and outputs. The money again went through the OGL system, now with a negative list, indicating that the credit was not restricted to agriculture only. * The Financial Sector Adjustment Program from 1991 covered US$275 million. This focused on privatization of the banking sec- tor and allowing foreign exchange bureaus and foreign banks to enter the market. * The Structural Adjustment Credit of 1991 focused on restructuring and privatization of parastatals, reform of the civil service, and further restructuring of the markets for agricultural exports. Nonetheless, although reforms were undertaken during the second half of the 1980s, the cabinet was still dominated by hard-liners who went along very reluctantly. Former president Nyerere, who had stayed as chairman of Chama Cha Mapinduzi (CCM), the ruling party, made his views known from time to time, but seldom intervened directly. It was just a handful of ministers that really supported the reforms, and the debates in the cabinet were very difficult until the Eastern European collapse. After that, from about 1990 onwards, the tone changed dra- matically and the hard-liners essentially conceded defeat. In fact a few of them-such as Gertrude Mongela, Alfred Tandau, and David Mwakawago, all strong in the ruling party-were made ambassadors to various countries, keeping them away from active domestic politics. Although resistance to reforms has been reduced over time and the purge of the hard-liners was a big step forward, elements of resistance remain and have taken different forms. Since the ruling party conceded ideological defeat in the Zanzibar Resolution in 1991, the opposition to reforms that has surfaced from time to time has been mainly issue-ori- ented rather than based on specific ideological lines. The opponents of reforms have also changed tactics and have been taking nationalistic lines (in the name of indigenous people or Tanzania in general) rather than ideological lines (in the name of socialism). These tactics have emerged most clearly in relation to the process of privatization. In the first few years of the 1990s reforms took off. Donor confidence in the government increased significantly, and the IFIs presented very positive reports on Tanzania's achievements. Since that time there has been less emphasis on comprehensive macroeconomic programs and targets; rather, the emphasis has shifted to issues such as privatization Tanzania 323 and civil service and parastatal reform. The five-year plans were aban- doned and were replaced by Policy Framework Papers, rolling plans, and forward budgets. Public debate in the early 1990s mainly concerned the poor quality of social services, but there was also criticism of the lack of democracy. The adjustment measures initially met with limited public opposition, but in 1992 riots took place at two universities to protest the introduction of cost-sharing measures in association with the World Bank Higher Edu- cation Credit. The government's suppression of civil society has gradually eased. It allowed independent newspapers in 1988 and limited public debate on human rights violations in 1991, and in that year it also reduced the supervisory power of the ruling party over the cooperatives. Labor and cooperative unions received some autonomy after 1990 (Stewart, Klugman, and Neyapti 1999). These reforms made it possible to debate issues that previously had not been allowed. The leadership code was abolished in 1992, and several officials have taken advantage of their newfound freedom in unethical ways. Some of this has been exposed in the press. Nyerere eventually threw his support behind the introduction of de- mocracy. By this time several opposition parties had been formed in Brit- ain and underground in Tanzania. In 1991 a presidential commission took a survey of attitudes toward the introduction of multiparty democ- racy and found that 77 percent supported the existing one-party system. However, CCM, the ruling party, succumbed to external pressure and legalized the registration of opposition parties in 1992. CCM continued to win by-elections, although the legitimacy of these was contested. The party still had privileged access to the media and there were accusations of irregularities. The opposition parties were either fragmented in ideo- logical identity or were distinguished by their geographical concentra- tion of membership. 1993-95: Off-Track In 1992 President Mwinyi appointed Malima as minister of finance. Al- though the move is hard to explain politically given Malima's back- ground, he seemed to have changed his views on economic reform. The cabinet team also changed. This was the start of a new period of crisis, however, and President Mwinyi seemed less emphatic as a reformer during his second term in office. Some argue that he lost control. During the years 1993-95 reforms veered off track, fiscal policy went out of con- trol, large tax exemptions were granted by the minister of finance, and there was a general feeling that corruption and tax evasion were ram- pant. Government revenue fell, and the government reverted to heavy 324 Post-Socialist Reformers borrowing in the Central Bank. Much of the fiscal control and discipline that had been developed since 1986 was lost. Why did the reform process stagnate? The main reason is probably that the donors shifted their demands to new areas and raised the stakes. Second-generation reforms are more intrusive than first-generation re- forms. They demand the closing up of various points of leakage in the system, for example through banking reforms. This confronts vested in- terests more directly, and the government then seeks to find new ways to allocate rents. For such reforms to be successful, therefore, the gov- ernment must be fully committed to radical reform, and in this case the government was not fully committed. Many of the donors now lost faith again. In November 1994 the Nor- dic donors except Denmark suspended aid payments because of the tax exemptions to politically influential people. The IMF failed to engage the government in meaningful discussions, and in 1994 the Fund can- celled the adjustment program and essentially pulled out. Malima was soon forced to resign as minister of finance, mainly on account of the tax exemption scandals, but also because of his inability to communicate with the IFIs. He was moved to the Ministry of Industries and Trade, but before the elections in 1995 he joined an opposition party and left the government. A new minister of finance, Kikwete, was appointed for a short period. Politically it was important to replace a Muslim with a Muslim to avoid turning it into a religious issue, especially because Malima had often indicated in private that he was the first Muslim min- ister of finance. The transfer of Malima from Finance was clearly engi- neered by donors on account of the way CIS funds and exemptions were handled. In fact, a few months earlier Parliament had tried to fight out these issues and strongly criticized Malima for the exemptions, but he sur- vived. President Mwinyi was not ready to let him go. A few months later, when the donors put their foot down, Mwinyi let Malima go but transferred him to Industries and Trade rather than sacking him. This demonstrates the relative influence of donors compared to that of do- mestic constituencies. In matters of policy reforms it would appear that donors exert greater influence than local groups. This imbalance has been changing gradually, and is likely to continue changing as ownership of development policy shifts to Tanzania. 1996-99: Back on Track A number of opposition parties took part in the first multiparty elec- tions, but CCM retained power in 1995 with 62 percent of the vote. The new government was elected in November 1995, and in early 1996 the new president, Mkapa, installed a team and from then on the reforms Tanzania 325 started again. Mkapa had always been one of the few reformers in the government. The government set off on a shadow program with the IMF for the period January-June 1996, which was successfully implemented. This program was watched carefully by all the donors and earned praise from the IMF. The IMF then came back with a new three-year ESAF and other donors followed. Since then the relations have improved. Still, aid inflows are smaller than they were before the break, due to the general decline in foreign aid. The agenda introduced by the new government at the end of 1995 largely reflects the agreement that had been reached between Tanzania and donors, as expressed very well in the Helleiner report. Mkapa put high priority on fighting corruption and mending relations with donors. He was also determined to keep the reforms on track. In the first few months of the new government a commission was formed to examine the problem of corruption. Efforts to form a Tanzania Revenue Author- ity were accelerated, and the TRA was established in July 1996. In order to get the new cabinet to come to grips with economic reforms a work- shop was organized for all ministers and principal secretaries in which various aspects of reforms were discussed. Resource persons were drawn from within and outside Tanzania, organized by the president's office (Civil Service Department), the Commonwealth Secretariat, and the East and Southern African Management Institute (ESAMI). The week-long workshop was held in Arusha in January 1996. The government signed a deal at the Paris Club in January 1997 and has already received Naples terms. In April 2000 Tanzania was included in the HIPC Initiative. Private debt is small and has been reduced mainly by donor-financed debt buybacks. The Multilateral Debt Relief Fund helps the government service its debt to the IFIs. At the beginning of the new century, the status of the adjustment pro- gram is relatively satisfactory. The country performs well on the macro- economic indicators, but struggles with some of the structural performance indicators. During 1996-99 there were a few cases in which local constituencies were able to express their interests and influence policy by seeking as- sistance from policy analysis institutions. First, in 1996 a group of busi- ness people felt the tax regime was unfavorable to them. The tax rates were too high and they wanted to convince the government that high taxes were inimical to revenue growth. They sought assistance from ESRF, which did a study that synthesized their views and helped them argue their case. The main argument was that if tax rates were lowered and simplified, compliance would increase and tax revenue would go up. This argument was contrary to what the IMF was advising. However, in the 1996-97 budget the government gave in to some of these demands and the following year more gains were made in that direction. In order 326 Post-Socialist Reformers to blunt the influence of the IMF on the government the business com- munity representatives even arranged to meet for discussions with the major donors and the IFI representatives. It is not clear how much this helped, but they won some concessions in their favor. In the second instance, in 1997 the Confederation of Tanzania Indus- tries (CTI) wanted to convince the Dar es Salaam city commission that it could lower tax rates and revenue would not suffer. The CTI approached ESRF to undertake a study of the issue; this yielded recommendations for lowering the tax rates but widening the tax base. The study helped in the negotiations and CTI reached an agreement with the city commis- sion. In the end CTI was happy with the lower rates and the city com- mission was happy because revenue that year exceeded the previous year's revenue by a considerable margin. In this case as well as the pre- ceding one, the local groups influenced policy by generating ideas and by getting a local policy research institute to help analyze the issues and craft the argument in a manner that could be understood better by policymakers. Civil society is also beginning to exert some influence on policy, but progress in this direction is still gradual. In the case of the restructuring of the National Bank of Commerce (NBC), the government basically agreed with the World Bank about a three-way split (into trade bank, regional bank, and microfinance bank). This idea was not discussed in- ternally, but was sold to Tanzanian policymakers as the option that would most likely receive prompt support from the World Bank; budget time was approaching and support for the budget was crucial. However, a group of concerned citizens undertook to discuss the matter. The group was organized by the Development Policy Centre and held a one-day meeting at ESRF. After analyzing the options, they concluded that the three-way split was not in the national interest. They wrote a letter to the president requesting him to delay the matter for three months to allow more careful discussion of the options for restructuring the NBC. In response, the president decided to reopen the discussion and ap- pointed additional members to the steering committee dealing with the issue. Some of the new members came from the citizens' group that had submitted its views to the president. The expanded steering committee developed a recommendation for a two-way split (between microfinance and the rest). This option was adopted and continues to operate. Three of the four new members of the steering committee, who had opposed the three-way split, were subsequently appointed to the boards of the new banks. In effect, the Tanzanian policy advisers had selected an option that they had been led to believe the World Bank preferred on the assump- tion that this choice would simplify subsequent negotiations on these issues. On paper it appeared to be a decision by Tanzania's policy ana- Tanzania 327 lysts and policymakers, but in reality little analysis had been done. The recommendations reflected what Tanzanians perceived to be the posi- tion of key people in the World Bank. The objective was to simplify ne- gotiations, and not necessarily to choose the option that was best for Tanzania. What was deemed best for Tanzania was the option that would make it easier to attract aid. When the debate was reopened, the choice changed. Under conditions of genuine ownership of policymaking, the restructuring options would have been discussed openly from the start, and it would not have required a protest note from citizens to start that process. The experience confirms that policy analysts sometimes read the minds of IFI officials through informal discussions and later come up with their "own" policy choices that are in reality a reflection of ex- ternal influence; but it also shows that civil society is beginning to ex- press its views on policy reforms. Economic Policy and Growth Outcomes 1979-99 Between 1979 and 1985 per capita income fell by 1.5 percent per year according to official estimates (see table 6.6).5 However, the estimates are particularly uncertain for this period because the price system was in disarray and many economic activities had moved to the parallel economy. Estimates by Bevan and others (1988) suggest more signifi- cant income declines. Manufacturing output collapsed on account of the lack of imports, and agricultural growth declined. Public administra- tion was still growing, although the growth rate had fallen to less than 4 percent. State control of the "commanding heights" of the economy was TABLE 6.6 TANZANIA: DATA FOR THE CRISIS PERIOD, 1979-85 Indicator 1979 1980 1981 1982 1983 1984 1985 Per capita income growth (%) 0.2 -0.2 -3.7 -2.6 -5.6 0.2 1.4 Population growth (%) 3.1 3.2 3.2 3.2 3.2 3.2 3.2 Urbanization (%) 13.9 14.8 15.4 15.9 16.5 17.0 17.6 External debt (millions of U.S. dollars) 2,070 2,450 2,880 3,130 3,390 3,620 4,030 Interest rate spread (%) 7.5 7.5 8.0 8.0 9.0 9.0 7.8 Labor force in agriculture (%) 86.3 85.6 85.6 85.5 85.4 85.2 85.1 Monetary growth (0/4 46.9 26.9 18.1 19.5 17.8 3.7 30.3 Gross investment (% of GDP) 33.6 33.1 28.6 26.0 19.3 20.2 18.7 Sources: Bigsten and Danielsson 1999; World Development Itidicators 1998. 328 Post-Socialist Reformers a central tenet of the Arusha Declaration, and by the late 1970s Tanzania had constructed a very comprehensive parastatal sector (Bagachwa, Mbelle, and van Arkadie 1992). From about 40 entities in 1966, the parastatal sector grew to encompass approximately 450 entities by the mid-1980s, covering the entire range of economic activities. In the manu- facturing sector the parastatal firms produced half the output and con- trolled two-thirds of the fixed assets. Export performance deteriorated as domestic inflation increased and the government compressed imports through direct foreign exchange rationing rather than accommodating inflation through depreciation. Export data for the crisis period must be interpreted with care, since the black market premium on foreign exchange was high and provided strong incentives for smuggling (Kaufmann and O'Connell 1991). The decline in official exports during the crisis period therefore reflects in part a diversion of exports from official to unofficial channels in response to an increasing parallel premium (it rose from 45 percent at the end of 1978 to more than 800 percent at the end of 1985). Some estimates even put the level of smuggled exports at about 50 percent of official exports (Adam and others 1994). The period of more significant reforms began in 1986 with a maxi- devaluation and the initiation of a standby agreement with the IMF and a structural adjustment program with the World Bank. The agreements with the IFIs helped restore donor confidence. They came back in force and there was a second aid boom that peaked in the early 1990s. This support was probably needed to make the transformation possible, but it is hard to know for sure what would have happened without this re- source inflow. Tanzania was obviously in a desperate situation and would have had to go down the reform route in any case. The reforms managed to increase growth again (table 6.7). Per capita incomes have so far grown by 0.6 percent per year during the reform pe- riod. Agricultural growth increased significantly, and the manufacturing sector turned from a decline of 4 percent per year to positive growth, in spite of the fact that many industries have collapsed in the face of in- creased import competition. There were strong recoveries in construc- tion, trade, and transport as well. The development of the transport sector may be considered a necessary ingredient in the re-creation of a viable market economy. The lagging sector has been public administration. Drought and floods undermined agricultural production in 1997. Harvests of coffee and cotton, Tanzania's main exports, were particu- larly badly hit, undermining the country's trade balance. Economic growth in 1998 is estimated to have been only about 3 percent, barely keeping pace with the population growth of 2.8 percent. From the Arusha Declaration in 1967 until the reforms in the mid- 1980s, policies encouraged public sector investment while they actively TABLE 6.7 TANZANIA: DATA FOR THE REFORM PERIOD SINCE 1986 Indicator 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Per capita income growth (%) 3.4 2.7 1.3 -0.6 3.1 -0.3 -1.3 -4.1 -0.1 0.6 1.4 0.5 1.4 Population growth (%) 3.2 3.2 3.1 3.1 3.1 3.1 3.1 3.0 3.0 2.9 2.8 2.7 2.6 Urbanization (%) 18.2 18.9 19.5 20.2 20.8 22.0 23.2 24.5 25.7 26.9 28.1 29.3 30.5 Current account (% of GDP) -11.6 -15.5 -18.1 -17.2 -21.8 -24.5 -25.5 -24.4 -15.8 - - Exports (% of GDP) 9.8 14.6 13.2 8.5 8.4 15.2 18.5 20.5 21.5 - - Imports (% of GDP) 27.3 30.7 35.5 35.3 36.8 44.0 45.4 42.4 36.2 External debt (millions of U.S. dollars) 4,610 5,490 6,010 5,850 6,410 6,540 6,620 6,800 7,260 7,430 7,410 Government consumption (% of GDP) 16.8 16.3 17.0 18.4 19.3 19.5 18.1 16.1 13.2 - - Gross investment (% of GDP) 23.5 17.4 19.3 28.2 28.5 28.9 26.7 26.4 21.2 23.8 Domestic saving (% of GDP) 1.3 1.3 0.3 -0.6 -1.6 -2.8 -2.0 0.0 3.4 Interest rate spread (%) 10.0 11.8 12.2 14.0 18.2 23.6 - - Labor force in agriculture (%) 85.0 84.8 84.7 84.5 84.4 Manufacturing (% of GDP) 8.1 8.1 8.9 9.1 8.5 7.8 7.4 7.3 7.3 Monetary growth (%) 27.9 32.1 32.6 32.1 41.9 30.1 40.6 39.2 35.3 33.0 8.4 - Not available. Sources: Bigsten and Danielsson 1999; World Development Indicators 2000. 330 Post-Socialist Reformers discouraged private sector investment by means of a complex system of regulations and licenses (Likwelile 1998). The public sector also received preferential treatment in credit allocation (Ndulu and Hyuha 1984). The public sector dominated investment activity and came to dominate pro- duction in major parts of the economy, and trade was confined to a lim- ited number of parastatals (Moshi and Kilindo 1995). There was a preference for large and capital-intensive investment projects, often in- frastructure projects, rather than directly productive investments. Gross fixed capital formation stood at 23.7 percent on average during the pre- crisis period of 1967-78, and then remained above 30 percent of GDP during the first years of the crisis before gradually declining to 18.7 per- cent in 1985. Investment levels in Tanzania have been high throughout, but there has been a strong shift toward private sector investment in recent years. The relative expansion of private sector investment is partly due to the shift in policy and partly due to the lack of resources to fi- nance public sector investments. Table 6.8 shows the distribution of investment across sectors during the reform period. Agricultural investments have remained fairly low, although this sector would have been expected to respond more strongly to marketing liberalization measures and improvements in the trans- port infrastructure. There has been an expansion from low levels in min- ing, where substantial investments have been made in, for example, gold mines. There has been very little investment in manufacturing for most of the period, which may reflect the lack of profitable investment oppor- tunities in the manufacturing sector as well as lack of faith in the reform measures. If investors are not certain that the new system is going to continue, they usually prefer to put their money in short-term assets such as Treasury bills. The risk rating has improved in Tanzania, but it is still much below that of other countries in the region (Ndulu and Wangwe 1997). Much of the new investment has gone instead into infrastructure, trade, and finance. This is probably a necessary first step to restore the economic and market infrastructure after a long period of decline. We cannot be sure, however, that there will be a substantial recovery of in- vestments in the directly productive sectors, although there are some recent signs of a recovery. Investment rates have continued to be fairly high in Tanzania, although they are much lower during the reform period than what was implied by the old, nonrevised national accounts. It is obvious that low levels of capital formation cannot explain Tanzania's growth problems. The cru- cial issue is the quality of investments rather than the quantity. Low qual- ity of investment during the pre-crisis and crisis periods is explained by the highly distorted macroeconomic environment and, in the crisis pe- riod, by the govemment's persistence with the Basic Industries Strategy in the face of collapsing foreign exchange availability. The capacity TABLE 6.8 TANZANIA: SHARES OF GROSS FIXED CAPITAL FORMATION BY TYPE OF ECONOMIC ACTIVITY AT CURRENT PRICES, BY SECTOR Sector 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Total Agriculture 0.052 0.039 0.040 0.042 0.049 0.063 0.063 0.063 0.064 0.063 0.097 0.066 Mining 0.030 0.026 0.025 0.032 0.048 0.046 0.041 0.043 0.043 0.043 0.106 0.053 Manufacturing 0.003 0.007 0.002 0.002 0.002 0.003 0.003 0.003 0.003 0.003 0.133 0.026 Electricity and water 0.277 0.213 0.229 0.256 0.237 0.235 0.240 0.244 0.240 0.241 0.111 0.218 Construction 0.182 0.248 0.256 0.178 0.161 0.148 0.152 0.136 0.145 0.144 0.110 0.147 Trade 0.111 0.095 0.098 0.120 0.111 0.116 0.119 0.122 0.119 0.120 0.109 0.116 Transport 0.015 0.051 0.022 0.010 0.010 0.008 0.009 0.010 0.009 0.010 0.111 0.029 Finance and real estate 0.314 0.283 0.288 0.340 0.314 0.324 0.331 0.335 0.330 0.332 0.117 0.290 Public administration 0.016 0.038 0.040 0.020 0.068 0.058 0.042 0.044 0.047 0.044 0.104 0.056 Total 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Note: These calculations are taken from national accounts in Tanzanian shillings. Source: Tanzania Bureau of Statistics, "National Accounts of Tanzania Mainland 1987-1997," 23. 332 Post-Socialist Reformers utilization rate in manufacturing was driven below 30 percent in the early 1980s. The first phase of reforms in Tanzania was clearly designed to address the more drastic distortions associated with quota protection and nonmarket pricing of foreign exchange. A second constraint on investment productivity in Tanzania, how- ever, has been the dominance of the public sector in the allocation of credit. Until the banking sector reforms that began in 1992, the parastatal sector had automatic access to credit from the monopoly commercial banking sector, which in turn had automatic access from the Bank of Tanzania. The vast bulk of NBC's lending was to the parastatal sector; in 1980 the private sector accounted for less than 8 percent of NBC's out- standing loans. The parastatal sector made enormous losses through the 1980s (see Eriksson 1998), representing a large drain not only on the budget but also on credit from the banking sector. The point is simple: without hard budget constraints, relative price changes are not likely to produce significant reallocations of resources or big productivity gains. To a large degree, the parastatals, and particularly the marketing boards and (quasi-public) cooperative unions during the reform period, were able to pass a major portion of the relative price changes they faced into greater losses that were financed by the budget or banking system. The reforms of Tanzania's financial system have just started, and as yet there is relatively little long-term financing available for private firms. Tanzania Housing Bank has folded and the National Bank of Commerce is being transformed and provides mainly short-term loans. Private in- vestments therefore often have to be financed by foreign money. Institutional Reforms and Governance Governance is a crucial constraint on development prospects in Tanza- nia. Immediately after being elected, President Mkapa set up a commis- sion that produced a 900-page report on corruption and mismanagement, known as the Warioba report. Since then, however, little concrete action has been taken, although some low-level officials were dismissed from the government. There are two main pressure groups, the press and the donors, that are capable of bringing about a change in government policy or forcing the government to take action against malpractice. So far, how- ever, people who misbehave tend not be fired or prosecuted, but are simply transferred to some other activity. The lack of action may be be- cause the elite is a rather tightly knit network. In addition, the Parlia- ment does not really play an independent watchdog role in Tanzania; throughout the country's history its role basically has been to support the government's development efforts.6 The civil service reform program was officially launched in July 1991, but it was not until 1993 that the government began designing a pro- Tanzania 333 gram to strengthen management of the public sector. Phase one of the Public Sector Reform Program sought to reduce the scope of govern- ment operations to affordable levels, to rationalize the government ma- chinery, to develop an open, objective, and competitive pay structure, to decentralize executive responsibilities to local government, and to im- prove the quality and performance of civil servants (Tanzania 1998a: 1). The main achievement of phase one was the redefinition of the role of the state. The government is in the process of withdrawing from pro- duction activities and is trying to reduce its role to affordable levels and enhance the participation of the private sector and NGOs in service de- livery. Between 1993 and 1997 there was a 30 percent reduction in the size of the work force, from 355,000 to 270,000. The new five-year strategy for public sector management and em- ployment builds on what has been achieved so far, seeking better levels of performance. Efforts will concentrate on strengthening public sector discipline and developing clear lines of responsibility and accountabil- ity based on modern management principles. The aim is to deliver qual- ity public services under severe budget constraint. The new program has an interesting budget-based incentive built in. When a ministry or agency has managed to operationalize a services improvement program on the basis of the performance improvement model, it will be graduated out of the "cash budgeting" exchequer re- lease mechanism. This means that the Ministry of Finance will from the outset commit to fund fully and predictably the approved budget of the organization. This is regarded as a key output of the public financial management reform program, and it seems to be a rational way out of the tight cash budget system. Those that have managed to achieve an effective system will be able to run their operations more smoothly in the future. A second incentive mechanism is that institutions will be al- lowed to retain a significant portion of the cost savings realized on the implementation of restructuring and private sector rationalization pro- grams. Monitoring and evaluation mechanisms will be strengthened. There is also going to be increased private participation in public sector service delivery. The crisis years in Tanzania led to a dramatic deterioration of real civil service salaries. By the late 1980s the real salary of a civil servant was only one-fifth of what it was in the early 1970s (Tanzania 1998b: 2). Severe distortions and inequities had developed in the compensation structure, and as a result the morale of the service had deteriorated sig- nificantly, Performance was poor, and absenteeism, moonlighting, and corruption were growing. The government has tried to rectify these distortions by rationalizing the wage scales and increasing the real income levels. Many donors of- fer remuneration incentives (local cost compensation) in various forms 334 Post-Socialist Reformers to civil servants working in their projects. According to a recent esti- mate, about 1,000 civil servants benefit from local cost compensation (Guidelines 1988:4) at a cost of 13 billion Tshs. This system of course may make the donor project function better, but over the long term it tends to undermine efforts to build capacity in the government sector. The strategy now is that there shall be a convergence between donor pays and government salaries over the next five years. In the future a civil servant who takes up a project job will have to resign from his or her civil service job. The government's aim is to achieve reasonable and competitive wages for its employees. For low-level workers the govern- ment salaries are often higher than those in the private sector, but for higher grades the salaries are not competitive. The government will do away with various kinds of benefits, but housing allowances will re- main for top personnel. There will also be medical benefits and post- employment security. In the drive toward socialism and political control, the government had set up a large number of parastatals. This was one of the most dra- matic examples of dysfunctional government.7 The performance prob- lems of the parastatals were at their peak during the crisis period, 1979-85, but the system remained intact and even expanded after that. Parastatal losses were growing, but the firms had by now learned how to operate the system and they were successful in acquiring various forms of fi- nance from the government, banks, and foreign donors. This made it possible for even chronic lossmakers to survive. The reforms have gradually changed this. Since reforms in the parastatal sector initially were very limited while at the same time rela- tive prices changed and competition increased, the losses of the parastatals grew rapidly. The soft budget constraint delayed structural adjustment and reduced the effectiveness of stabilization measures. In 1990 the ruling party finally decided to move toward complete private ownership. The Parastatal Sector Reform Commission (PSRC) was formed in March 1992, but was not fully operational until mid-1993. The privatization process has moved more slowly than anticipated. First the PSRC dealt with the smaller units, but now it has started to work on larger companies such as the utilities. These were originally considered to be strategic, and the original idea in 1992 was to bring in private management and let them work under performance contracts. This was not as effective as hoped, so from 1996 on private ownership was sought for the strategic industries including railways, telecommu- nications, harbors, and Air Tanzania. Of the 383 firms that the PSRC had on the books when it started, 230 had been dealt with by June 1998. When its task is completed, the PSRC will be phased out. In the wake of mounting criticism, especially in the Parliament, that public firms were being sold to foreigners with no meaningful participation by Tanzani- Tanzania 335 ans, the government in 1998 established the Privatization Trust to pro- mote wide-share ownership among citizens. In sum, there has been progress with regard to the privatization pro- gram, which the World Bank has characterized as one of the better in Africa. The dismantling of the parastatals will be concluded within a few years. The problem then becomes one of regulating the various utili- ties with a virtual monopoly position; the system for handling these is- sues is still very weak. Thus, efforts to effect a division of responsibilities between the functions of government and those of producers have moved forward, but much remains to be done before this is translated into an efficient industry. The judiciary is exceedingly weak, although a commercial court is to be instituted. At present it can take years for a firm to get a decision in a lawsuit. The labor laws are obsolete, and even firms that fire dis- honest or corrupt workers can be taken to court for expensive trials. This means that they end up having to buy off the dishonest workers that they dismiss. It seems clear that Tanzania had made considerable progress with regard to the economic policy regime. Each year the World Bank under- takes a policy and institutional assessment that ranks performance with regard to 20 specific indicators in four areas: (a) macroeconomic man- agement and sustainability of reforms; (b) policies of sustainable and equitable growth; (c) policies for reducing inequalities; and (d) public sector management. As expected, Sub-Saharan Africa is the region with the lowest average ranking, but Tanzania ranks above the average for Africa (World Bank 1999a). However, the impact of the reforms on eco- nomic performance is still limited. The impact of similar reforms in Uganda, for example, has definitely been better. One reason may be that Uganda stayed on course for a much longer period of time. Uganda under Museveni has had more genuine ownership and commitment than Tan- zania, where there have been setbacks and a lack of force in implemen- tation. Tanzania always seems to be on the verge of take-off, but so far it has not happened. A big problem is that the government lacks the competence to man- age an essentially market-dominated economy. The provision of public services remains poor, and corruption has probably increased as a con- sequence of the low salaries in government. From 1994 onward the rela- tionships with the donors deteriorated again. The donors had, for example, pledged substantial balance of payments support, but on ac- count of their view that the government has not been effective enough in tax collection efforts and other areas, they have withheld substantial amounts. Still, Tanzania has certainly come a long way since 1985. Few countries have undergone such a dramatic change of their entire social and economic set-up in such a short time. 336 Post-Socialist Reformers THE RELATIONSHIP BETWEEN AID AND REFORMS In the late 1970s negotiations with the IMF collapsed, since the demands of the IFIs implied a critique of the fundamentals of the Tanzanian ap- proach based on socialist economic management. We have noted that political resistance to reforms was strong, because the liberalization of the economy represented a U-turn relative to the development strategy that was outlined in the Arusha Declaration. The reforms were seen as counter to the ideology of the ruling party, which therefore resisted them as long as possible. Moreover, the behavior of donors during the crisis was neither uniform nor consistent. It took a long time for donors to take a critical look at Tanzania's post-Arusha economic development poli- cies. The Nordic countries and the Netherlands did not take a clear posi- tion in favor of liberal reform and the IMF until 1983. The economic crisis of the first half of the 1980s, plus the withholding of aid and open donor critique of the policy even by the country's "tra- ditional friends," spurred a debate within the government and also out- side it, for example at the university. This external pressure likely helped to tip the balance in favor of domestic groups supporting the reforms. There were small policy changes during the first half of the 1980s, but comprehensive reforms were still resisted. By the mid-1980s, however, the crisis was so acute and the external support so small that the govern- ment had to budge. The reforms, which were undertaken in 1984 and 1985, were done partly in anticipation of the negotiations with IFIs. At one level they were inter- nally generated, but at another level they were undertaken with a view to preparing for negotiations with the IFIs. The retirement of Nyerere as presi- dent in 1985 opened the way for the agreement on the Economic Recov- ery Program of 1986-89 and later on a new ERP for 1989-92. Although some people within the government believed it was an appropriate strat- egy for the future, many saw it as a temporary setback. Nonetheless it is true that aid made reforms possible, at least in the sense that the govern- ment saw reforms as a way to get aid. The agreement on reforms also attracted resources, but it was the initial deprivation of resources that speeded preparations for reform and stimulated the internal debate. After 1986 the tempo of internal policy debates slowed and reforms were basically directed by the IFIs and donors. The reform measures un- dertaken were in the standard structural adjustment mold, and the level of ownership decreased. The policy changes that were required in the case of Tanzania were unusually large, however, and all parties involved were not really supportive. There was a lot of foot-dragging, but the reformers gradually increased their influence. After the early 1990s the anti-reform- ers essentially lost control of the party, but even so it was hard to stay on course. Soon there was slippage in the program again and there seemed Tanzania 337 to be an increase in corrupt practices, which precipitated a new crisis in donor relations between 1993 and 1995. Donors were also dissatisfied with the government's efforts with regard to revenue collection. The decline in aid inflows following the deterioration in relations with donors again had a significant impact on Tanzanian behavior. It seems to be the case that aid can more easily induce governments to undertake first-stage reforms, which in Tanzania's case included macroeconomic stabilization and some price liberalization. Aid seems to be less effective in stimulating the more intrusive second-generation reforms such as the freeing up of key markets, institutionalization of the market economy, privatization, and opening to foreign participation. These reforms are more directly in conflict with vested interests than first-generation re- forms and are therefore more likely to meet resistance. Still, the new government, from November 1995, put improvement of donor-Tanzania relations high on the agenda. This time the withdrawal of donor support led to more substantive discussions of the need to en- hance ownership of the policy reforms. The post-1986 resumption of aid had eroded ownership of the reform process, while the resumption of aid relations in 1996 came with strong expressions of the need to en- hance ownership of the reform process. Donor support to institutions of reform such as the Tanzania Revenue Authority showed that donors were willing to see the problems of revenue collection, which had been cen- tral to the collapse of Tanzania-donor relations in 1993-94, resolved through institutional reform. The new government has increased reform efforts, although it has had continual problems in undertaking some of the institutional and structural reforms set out. Structural and institu- tional reforms are at present high on the agenda, and there have been serious problems in their implementation. Still, it is hard to foresee any reversals of the reforms at this stage, although the effectiveness of many aspects of the reform process is in doubt. During the reform phase Tanzania has managed to bring about an improvement in its policy environment, and it has seen positive per capita income growth. Still, in spite of an increase in investment, it has not been enough to accelerate growth very much. The returns on investments have been poor, which is an indication that the reforms have not yet funda- mentally changed the economic environment. Many reforms are still in the planning stage, and those that have been implemented are often ineffective. Evaluation of the Impact of Aid on Reforms Our discussion of the link between aid and reforms in Tanzania has of necessity been tentative. It is virtually impossible to identify a counter- factual pattern of change. The key question is whether aid bought reform, 338 Post-Socialist Reformers or whether the reforms would have come anyway since they were an economic necessity. Initially aid probably delayed reforms by helping to finance schemes that would have been wholly nonviable without aid backing, but once aid started to decline in the first half of the 1980s, the cutbacks probably speeded up reforms. Of course, a country that is not committed to reform may choose to initiate a reform in order to receive foreign assistance. Then, once aid is disbursed, the recipient may not exert much effort to make the reform succeed. This may well have been the case in Tanzania at that time. We will consider three aspects of the aid-reform relationship: the im- pact of aid on the initiation and pace of reform, on the content of poli- cies, and on the effectiveness of implementation. Generation of reforms has largely originated with the IFIs or donors, hence the concern about lack of ownership as pointed out in the Helleiner report. The donors have also had a significant impact on the pace of reform. Some argue that things would be much slower if no deadlines were set by the IMF. The relationship between the World Bank and the government is good at present, and the World Bank has taken the lead in coordinating donor activities. The content of policy has largely been the standard structural adjust- ment package, with limited original input by Tanzanians. The weakness of domestic agents in generating ideas has gone hand in hand with lack of coordination in the management of reforms: reforms have been man- aged piecemeal, each one carried out by the respective government de- partment in isolation. There has been no effective mechanism for examining reforms as a whole package. It is only recently, after the be- ginning of efforts to draft the PFP, that a broad perspective on reforms has started to take shape. Added to this is the new initiative to prepare Sector Development Programs, although technical advisers have played an important role in formulating some of these programs. It should also be realized that local groups have historically been ex- cluded from policy debates and policymaking, which has been regarded as the responsibility of government. As economic and political liberal- ization takes root, it should become increasingly possible for local groups to play a role in generating ideas for reforms. The various local groups have a better chance now to articulate their interests, and over time they will build the capacity to do so. Given the very strong leadership in the period before the reforms and the fundamentally different development policy framework that was adopted, the process of consensus building toward reform was bound to be difficult. Still, in recent years there has been a more open debate that may help to build national consensus and commitment. With regard to the content of the policies, there have been improve- ments during the current government. With the search for direction high Tanzania 339 on the agenda, the government appointed a team of local experts to craft a development vision for Tanzania. The team submitted its "Vision 2025" document to the government in 1998, which approved and adopted it in May 1999. Some of the issues addressed by the vision document are re- duction of aid dependence and promotion of self-reliance (on domestic resources), building capacity to own the development agenda, human development including poverty eradication, building a strong and com- petitive economy, and broadening and institutionalizing participation in development. Today it is largely the government that sets sector priorities. Donors push mainly for health and education, while the government is also con- cerned with roads, water, and other areas. The big problem now is not a lack of insight about what the appropriate policy package is, but the fact that the reforms still are not implemented effectively. It is not easy to determine how much of this ineffectiveness is due to structural charac- teristics and the generally low level of competence, and how much to the lack of government commitment and lack of support and pressure from the domestic constituency. There have obviously been several setbacks in the Tanzanian reform process, but it seems unlikely that it would have progressed to its cur- rent stage without the pressure from donors. It is not possible to say whether it was technical assistance and policy advice or the financial flows that brought about the change. Since the nonfinancial aid typi- cally is combined with financial aid, it is hard to separate their effects. The problem of the counterfactual also applies to the issue of whether the reforms can be sustained without aid in the future. The answer to this question also hinges on whether aid is buying reforms that the govern- ment otherwise would be unwiUing or even unable to undertake, or whether the reforms are done willingly since it is obvious that they are the best option in any case. In the case of Tanzania, aid initially bought reforms that otherwise at least would have been delayed, but in the 1990s it seems more likely that there is some genuine backing for the reform process. It appears that the ineffectiveness of the system to a considerable ex- tent depends on the aid relationship. There is certainly need both for a more efficient governance structure and for changes in the nature of the donor-recipient relationship. It is not clear exactly how and to what ex- tent changes in the aid relationships can contribute to improvements of the governance structure, but some thoughts on this question follow. Where to Go? Collier (1997) contends that the link between aid and reforms in Sub- Saharan Africa is weak and that governments generally have decided to reform due to pressure from domestic political forces, largely indepen- 340 Post-Socialist Reformers dent of the aid relationship and aid conditions. According to this view, it is only a combination of broad, informed, and empowered domestic con- stituencies and the presence of successful neighbors that will induce re- forms. Conditionality therefore should not be used to induce policy change but to ensure that aid goes to environments where it can be effec- tive. The flow of aid should reflect the level of the policy environment. In its review of the ESAF (IMF 1998), the IMF argued that high condi- tionality programs generally do not do well. When conditionality is tight the recipient country perceives a loss of control over the policy content and the pace of implementation of the reform program. It is also sug- gested that there is a correlation between the degree of ownership and the success of program implementation. Aid to Tanzania has bought a certain reform agenda. This does not necessarily mean that the government would follow a strategy of no reform without aid, but it might pursue a strategy more in line with its "real" priorities. It is not easy to discern in what way such a reform agenda would differ from the current one, and therefore also hard to determine its viability. Still, to make the reform process more effective the govern- ment should be allowed to take control over priorities, and if donors are not willing to support the agenda-because they believe it to be nonvi- able or detrimental to development as they see it-they will have to with- draw from supporting the country. There is a problem in implementing this approach, however, as reflected in the slowness of change in donor flows to Tanzania. It is easy to withdraw if the support is of a general nature such as balance of payments or budget support, but once the donor is closely involved in project aid it becomes very difficult to pull out and in so doing probably kill the project. This is actually another argument for more general forms of aid. If the notion of country choice is to be- come meaningful, there is a need for more general forms of transfers. It has been pointed out that conditionality is more effective when it focuses on a small range of quantifiable indicators. Still, the number of conditions has not declined much over the years, and they have become increasingly important since the bilaterals now require that the govern- ment of Tanzania meet the demands of the IFIs. Here the donor commu- nity should move toward conditions that reward good behavior or good policy environments based on expert evaluations, rather than reward promises. Such a strategy also requires improved donor coordination, since this would make it easier for the government to take control. There is some movement toward more general forms of support, as in the case of the tax reform program where the World Bank put up some money and the donors then provided the rest without their own specific requests. This is not the norm; instead, the donors fall over each other in some areas, such as capacity building in the Ministry of Finance. Tanzania 341 Given the magnitude of aid that Tanzania has received, the poor growth record poses an obvious question of aid effectiveness. It is clear on the face of it that aid has not had a strong payoff in terms of growth. Constructing a proper counterfactual is extremely difficult, however. What would have happened if less aid had been given? Aid has clearly exerted two kinds of influence in Tanzania: on the supply of external resources, and on policy formation. Disentangling these effects is crucial for assessing the impact of aid on growth in Tanzania. The answer to the question of whether aid works only in good policy environments is left hanging. The policy environment as we normally define it has certainly improved in Tanzania, but as yet there is no strong evidence that aid has a better growth impact now than in the early stages of the reform pe- riod. On the basis of this paper we are therefore not able to say much about the impact of aid on the real economy. Still, aid has unquestion- ably had a major impact on the reform process in Tanzania. 342 Post-Socialist Reformers APPENDIX 6.1 TANZANIA: CHRONOLOGY OF POLITICAL AND ECONOMIC DEVELOPMENTS 1961 Tanzania wins independence. 1967 Arusha declaration. 1975-78 Coffee boom. 1978-79 Tanzania-Uganda war. 1981 Government proposes new measures in the National Economic Survival Program (NESP I). 1982 NESP II initiated. 1983 Launching of second homegrown structural adjustment program. 1984 Currency devaluation and import liberalization. 1985 Nyerere retires as president. 1986 Maxi-devaluation and initiation of a standby agreement with the IMF and the SAP with the World Bank. 1986-89 ERP I implemented. 1989-92 ERP II. 1991 Financial sector adjustment program implemented. July 1991 Civil service reform program launched. March 1992 Parastatal Sector Reform Commission formed. 1992 Riots to protest cost-sharing measures associated with the World Bank higher education credit. Leadership code abolished. Registration of opposition parties legalized. November 1994 Nordic donors except Denmark suspend aid payments because of tax exemptions. 1994 IMF cancels the adjustment program and essentially pulls out. 1995 First multiparty election. January-June 1996 Government sets off shadow program with IMF. July 1996 Tanzania Revenue Authority established. 1996 ESAF agreement with IMF. September 1996 Nordic countries enter into partnership agree- ment with Tanzania. December 1997 Consultative Group meeting. July 1998 Multilateral Debt Relief Fund established. APPENDIX 6.2 TANZANIA: AID DATA, 1970-96 TABLE 6.9 TOTAL GRANTS TO TANZANIA BY CREDITOR (millions of U.S. dollars) Creditor 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Canada 2.3 3.5 3.9 3.5 6.1 14.8 7.4 7.3 92.1 28.6 19.4 27.2 35.2 Denmark 1.5 2.7 4.1 5.5 8.8 6.5 9.7 6.9 19.7 50.1 22.6 32.5 39.1 European Union 0 0 0 0 0 15.1 5.7 10.6 17.8 22.9 22.0 32.2 20.8 Finland 0 1.0 2.1 3.1 5.8 4.5 6.3 5.2 8.0 33.8 17.7 12.8 12.7 France 0 0 0 0 0 0 0 0 0 0 1.6 1.9 2.7 Germany 4.2 3.6 5.6 7.8 10.6 13.8 16.9 18.2 22.4 40.8 209.0 54.5 59.0 Italy 0.5 0 0.1 0.4 0.1 0.2 0.4 0.8 1.1 1.0 1.8 10.7 12.3 Japan 0.7 0.7 0.7 1.0 1.9 3.1 2.9 2.7 3.2 13.8 13.3 17.5 10.9 Netherlands 1.1 0.9 2.6 3.4 5.1 8.1 23.4 35.6 91.5 76.2 81.5 68.9 52.2 Norway 1.4 2.9 3.5 6.9 10.8 17.2 14.5 25.7 30.7 35.1 44.2 40.1 51.9 Sweden 2.1 3.2 9.1 17.5 29.5 54.0 51.2 57.7 114.4 93.4 78.1 76.5 73.8 Switzerland 0.4 0.2 0.2 0.5 0.3 0.4 0.6 0.7 2.0 3.1 14.0 5.5 5.2 United Kingdom 4.3 3.4 3.5 3.4 3.1 8.5 10.4 11.7 19.8 46.5 74.5 56.0 41.2 United States 5.0 4.0 4.0 5.0 7.0 17.0 24.0 17.0 11.0 11.0 21.0 25.0 20.0 World Bank 0 0 0 0 0 0 0 0 0 0 0 0 0 Total 23.5 26.1 39.4 58 89.1 163.2 173.4 200.1 433.7 456.3 620.7 461.3 437 (continued on next page) TABLE 6.9-continued Creditor 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Canada 33.9 24.7 30.4 27.2 36.0 42.7 37.1 34.7 29.1 31.7 20.2 10.7 25.4 9.2 Denmark 30.2 29.1 40.1 132.5 49.2 77.8 78.7 78.6 89.1 94.8 80.9 76.6 59.1 91.1 European Union 19.2 28.0 28.0 36.7 34.2 34.2 52.7 41.5 37.7 110.8 65.5 85.7 59.2 40.6 Finland 16.9 21.1 16.6 25.0 33.2 67.2 55.4 51.0 40.5 34.8 15.7 22.5 9.4 9.1 France 2.9 1.5 1.3 2.0 2.4 2.6 16.7 3.2 3.0 27.7 20.4 12.2 22.7 5.9 Germany 36.3 49.8 32.4 44.9 60.2 68.1 51.9 61.0 66.0 68.5 72.4 64.4 67.3 58.7 Italy 7.6 9.4 13.7 17.6 76.5 27.5 13.4 16.1 9.1 19.0 266.6 10.2 3.7 3.2 Japan 12.6 18.3 22.9 30.4 40.9 82.6 56.7 43.4 61.1 79.3 99.6 106.7 125.9 109.5 Netherlands 34.7 40.4 36.3 60.5 73.9 78.9 71.8 94.6 63.7 50.5 55.1 57.8 78.6 74.9 Norway 54.8 45.7 45.4 71.8 75.4 79.2 57.6 103.0 85.7 81.9 68.6 50.5 52.3 53.2 Sweden 69.3 55.1 49.0 106.4 76.5 103.6 90.4 149.6 143 93.1 91.0 50.8 45.3 65.2 Switzerland 5.8 6.1 6.1 19.2 17.0 14.5 23.5 19.0 13.9 29.2 11.9 20.0 19.0 15.6 United Kingdom 43.9 29.7 23.0 17.5 38.6 58.7 60.2 30.6 61.8 100.6 37.0 40.4 30.3 66.0 United States 14.0 20.0 19.0 7.0 9.0 14.0 6.0 78.0 133.0 27.0 24.0 24.0 18.0 13.0 World Bank 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Total 382.1 378.9 364.2 598.7 623.0 751.6 672.1 804.3 836.7 848.9 928.9 632.5 616.2 615.2 Note: Includes technical assistance. Source: OECD-DAC. TABLE 6.io TANZANIA: TOTAL DISBURSEMENTS ON ALL EDA LOANS (millions of U.S. dollars) Disbursements 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 All EDA loans 43.354 50.170 92.049 162.551 174.964 254.353 133.353 197.779 162.307 162.466 240.136 554.34 261.041 Disbursements 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 All EDA loans 350.819 358.528 160.872 227.458 205.825 336.728 194.158 287.737 276.387 350.987 200.445 243.845 218.090 202.769 Source: World Bank EDA database. TABLE 6.ii TANZANIA: TOTAL DISBURSEMENTS ON EDA LOANS BY SECTOR (millions of U.S. dollars) Sector Creditor type 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Agriculture, forestry, fishing Multilateral 1.280 1.211 0.997 1.868 2.372 18.432 13.905 15.965 15.872 27.009 13.804 27.424 11.892 Agriculture, forestry, fishing Bilateral 0.968 1.459 1.052 1.142 2.186 10.070 5.221 12.165 8.274 6.013 11.458 13.881 11.440 Balance of payments support Multilateral 0 0 0 0 0 30.000 0 20.450 0 0 0 10.000 0 Balance of payments support Bilateral 0 0 0 0 0.968 2.310 2.090 1.512 0 0 0 0 0 Communication Multilateral 0 0 0 0 0 0 0 0 0 0 0 0 0 Communication Bilateral 0 0 0 0 0 0 0 0 1.031 6.418 5.310 4.936 3.271 Community, social, personal, and environmental services Multilateral 0.024 0.483 1.265 3.962 1.952 2.857 4.964 9.650 5.458 9.666 7.741 20.076 41.269 Community, social, personal, and environmental services Bilateral 0.747 0.568 1.403 2.863 6.058 5.573 8.214 5.391 8.006 5.642 4.619 4.470 2.781 Construction Bilateral 0 0 0 0 0 0 0 0 1.817 1.259 0.524 0.205 0 Contributions not recorded as DAC Multilateral 0 0 0 0 0 0 0 0 0 0 0 0 0 Contributions not recorded as DAC Bilateral 0 0 0 0 0 0 0 0 0 3.173 0 0 0 Contributions to finance current imports Multilateral 0 0 0 0 0 0 0 0 0 0 0 0 0 Contributions to finance current imports Bilateral 0 0 0.0670 1.818 2.592 6.928 9.225 6.789 5.634 3.239 16.863 20.383 21.307 Debt reorganization Bilateral 0 0 0 0 0 0 0 0 0 0 0 0 0 Electricity, gas, and water production Multilateral 1.242 7.245 9.333 10.012 8.944 3.422 1.814 3.591 6.232 12.316 40.607 14.907 2.730 Electricity, gas, and water production Bilateral 4.238 7.070 11.321 39.634 21.743 10.690 10.033 10.550 14.342 0.623 10.897 14.827 20.615 Financial, insurance, real estate, business Multilateral 0 0 1.928 0.817 3.827 6.433 4.818 4.367 8.282 11.136 16.325 10.399 14.050 Financial, insurance, real estate, business Bilateral 0 0 0.404 0.620 2.401 6.312 6.455 4.533 2.129 1.832 1.602 1.661 1.265 General purpose contributions Multilateral 0 0 0.223 0.424 0.239 0.713 1.670 10.518 2.699 4.188 3.201 14.224 11.581 General purpose contributions Bilateral 2.656 3.214 4.320 6.799 8.845 8.371 5.623 5.683 5.785 8.544 14.090 11.545 11.553 Manufacturing Multilateral 0 0 0 0 0 1.603 15.660 18.090 23.269 14.966 17.399 27.859 43.619 Manufacturing Bilateral 3.650 2.359 2.556 31.961 40.690 56.571 29.021 45.148 30.314 18.958 34.279 15.987 11.245 Mining, quarrying Multilateral 0 0 0 0 0 0 0 0 0 0 5.513 23.090 15.005 Mining, quarrying Bilateral 0 0.017 0.028 0.021 0.016 0.010 0.006 0.003 0.004 0 0 0 0 Not applicable Bilateral 0 0 0 0 1.259 2.099 1.775 1.356 0.907 0.546 0.325273.621 0.988 Trade, restaurants, lodging Multilateral 0 0 0 0 0 0 0 0 0 0 0.732 0.567 2.305 Trade, restaurants, lodging Bilateral 2.543 1.959 0.792 0.007 0.002 0 0 0 0 0 0 0 0 Transport and storage Multilateral 8.115 6.384 4.490 0.507 2.300 6.611 2.812 3.689 7.547 3.669 12.060 2.407 4.320 Transport and storage Bilateral 17.891 18.201 51.870 60.096 68.570 75.348 10.047 18.329 14.705 23.269 22.787 41.871 29.805 (continuted oJI next page) TABLE 6.1i -continued Sector Creditor type 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Agriculture, forestry, fishing Multilateral 11.086 16.053 10.668 11.142 3.544 23.400 15.359 88.887 48.920 126.743 32.766 37.582 29.138 37.519 Agriculture, forestry, fishing Bilateral 9.364 6.060 6.315 5.323 0.217 3.523 0.685 2.081 1.975 0.187 3.304 3.551 2.817 0 Balance of payments support Multilateral 0 0 0 0 0 0 1.414 2.919 3.094 0 0 0 0 0 Balance of payments support Bilateral 0 0 0 0 0 0 0 16.902 0.012 0 0 0 0 0 Communication Multilateral 10.600 5.008 1.767 3.386 5.586 3.385 7.680 5.915 2.611 1.850 1.338 5.100 6.307 27.885 Communication Bilateral 2.417 1.066 0.406 0.227 0.113 0.047 9.919 11.631 6.329 2.938 0 3.565 1.520 0.069 Community, social, personal, and environmental services Multilateral 15.934 15.977 14.529 60.670 54.460 192.083 25,324 12.658 25.806 9.920 14.085 35.411 27.664 22.952 Community, social, personal, and environmental services Bilateral 66.679 8.700 9.527 1.950 1.228 17.144 2.007 0.259 0.066 0.001 0.150 0 0.744 1.074 Construction Bilateral 9.441 0 1.241 2.462 1.399 0 0 0 0 0.357 0 0 0 0 Contributions not recorded as DAC Multilateral 0 0 0 0 0 0 0 0 0 0 0 0 1.114 0 Contributions not recorded as DAC Bilateral 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Contributions to finance current imports Multilateral 0 0 0 0.329 0.711 1.510 0.185 2.984 7.500 0 0 0 0 0 Contributions to finance current imports Bilateral 15.116 1.684 0.010 11.241 0.638 0.506 0.349 0.216 0.484 0 0 8 0 0 Debt reorganization Bilateral 0 6.123 0.750 1.699 1.801 1.329 16.817 5.506 8.249 5.687 7.493 0 0 0 Electricity, gas, and water production Multilateral 3.339 14.298 9.430 18.409 8.464 26.440 15.488 13.218 5.635 5.643 4.454 13.954 57.877 51.547 Electricity, gas, and water production Bilateral 34.455 26.055 14.870 26.487 2.299 7.687 11.481 12.296 16.125 69.748 4.369 1.350 23.753 2.612 Financial, insurance, real estate, business Multilateral 8.430 5.510 0.326 2.830 0.001 0 0 0 49.947 46.783 68.679 61.414 14.027 16.941 Financial, insurance, real estate, business Bilateral 1.238 1.001 2.805 0.383 0 0.068 0.027 0.018 0 0 0 1.103 0 0 General purpose contributions Multilateral 0.355 2.854 7.860 3.499 4.011 2.168 1.804 0.211 0 0.132 0 0 0 0 General purpose contributions Bilateral 6.972 4.101 2.447 2.335 4.896 6.673 4.230 7.787 7.113 6.021 0.921 0 0.873 2.947 Manufacturing Multilateral 44.590 31.453 15.827 12.373 7.836 4.634 52.574 50.576 35.853 21.720 7.152 3.472 1.974 0 Manufacturing Bilateral 23.279 6.600 0.946 25.068 0.024 1.219 0 1.069 1.787 3.167 0 0 0 0 Mining, quarrying Multilateral 4.646 0.741 0.573 3.674 1.863 0.661 2.668 4.556 0.342 0.601 0.411 0 1.535 1.623 Mining, quarrying Bilateral 34.201 180.532 10.346 5.939 2.374 1.583 0 0 0 0.011 7.935 0 0 0 Not applicable Bilateral 0.005 0.096 19.167 0.067 75.707 0 0 0 0 0 0 0 0 0 Trade, restaurants, lodging Multilateral 2.441 1.194 0.002 0.180 0 0 0 0.654 5.209 0 8.038 0 0 0 Trade, restaurants, lodging Bilateral 0 0 0 0 0 0 0 0 0 0 0 6.203 0 0 Transport and storage Multilateral 13.566 6.784 17.469 20.908 15.025 24.322 9.928 22.658 41.687 46.203 38.541 57.941 47.617 37.600 Transport and storage Bilateral 32.665 16.638 13.591 6.877 13.628 18.346 16.219 24.736 7.643 3.275 0.809 5.199 1.130 0 Source: World Bank EDA database. TABLE 6.12 TANZANIA: TOTAL DISBURSEMENTS BY SECTOR (millions of U.S. dollars) Sector 1970 1971 1972 1973 1974 1975 7976 1977 1978 1979 1980 1981 1982 Agriculture, forestry, fishing 2.248 2.670 2.049 3.010 4.558 28.502 19.126 28.130 24.146 33.022 25.262 41.305 23.332 Balance of payments support 0 0 0 0 0.968 32.310 2.090 21.962 0 0 0 10 0 Communication 0 0 0 0 0 0 0 0 1.031 6.418 5.310 4.936 3.271 Community, social, personal, and environmental services 0.771 1.051 2.668 6.825 8.010 8.430 13.178 15.041 13.464 15.308 12.360 24.546 44.050 Construction 0 0 0 0 0 0 0 0 1.817 1.259 0.524 0.205 0 Contributions not recorded as DAC 0 0 0 0 0 0 0 0 0 3.173 0 0 0 Contributions to finance current imports 0 0 0.067 1.818 2.592 6.928 9.225 6.789 5.634 3.239 16.863 20.383 21.307 Debt reorganization 0 0 0 0 0 0 0 0 0 0 0 0 0 Electricity, gas, and water production 5.480 14.315 20.654 49.646 30.687 14.112 11.847 14.141 20.574 12.939 51.504 29.734 23.345 Financial, insurance, real estate, business 0 0 2.332 1.437 6.228 12.745 11.273 8.900 10.411 12.968 17.927 12.060 15.315 General purpose contributions 2.656 3.214 4.543 7.223 9.084 9.084 7.293 16.201 8.484 12.732 17.291 25.769 23.134 Manufacturing 3.650 2.359 2.556 31.961 40.690 58.174 44.681 63.238 53.583 33.924 51.678 43.846 54.864 Mining, quarrying 0 0.017 0.028 0.021 0.016 0.010 0.006 0.003 0.004 0 5.513 23.090 15.005 Not applicable 0 0 0 0 1.259 2.099 1.775 1.356 0.907 0.546 0.325 273.621 0.988 Trade, restaurants, lodging 2.543 1.959 0.792 0.007 0.002 0 0 0 0 0 0.732 0.567 2.305 Transport and storage 26.006 24.585 56.360 60.603 70.870 81.959 12.859 22.018 22.252 26.938 34.847 44.278 34.125 Sector 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Agriculture, forestry, fishing 20.450 22.113 16.983 16.465 3.761 26.923 16.044 90.968 50.895 126.93 36.070 41.133 31.955 37.519 Balance of payments support 0 0 0 0 0 0 1.414 19.821 3.106 0 0 0 0 0 Communication 13.017 6.074 2.173 3.613 5.699 3.432 17.599 17.546 8.940 4.788 1.338 8.665 7.827 27.954 Community, social, personal, and environmental services 82.613 24.677 24.056 62.62 55.688 209.227 27.331 12.917 25.872 9.921 14.235 35.411 28.408 24.026 Construction 9.441 0 1.241 2.462 1.399 0 0 0 0 0.357 0 0 0 0 Contributions not recorded as DAC 0 0 0 0 0 0 0 0 0 0 0 0 1.114 0 Contributions to finance current imports 15.116 1.684 0.010 11.570 1.349 2.016 0.534 3.200 7.984 0 0 8 0 0 Debt reorganization 0 6.123 0.750 1.699 1.801 1.329 16.817 5.506 8.249 5.687 7.493 0 0 0 Electricity, gas, and water production 37.794 40.353 24.300 44.896 10.763 34.127 26.969 25.514 21.760 75.391 8.823 15.304 81.63 54.159 Financial, insurance, real estate, business 9.668 6.511 3.131 3.213 0.001 0.068 0.027 0.018 49.947 46.783 68.679 62.517 14.027 16.941 General purpose contributions 7.327 6.955 10.307 5.834 8.907 8.841 6.034 7.998 7.113 6.153 0.921 0 0.873 2.947 Manufacturing 67.869 38.053 16.773 37.441 7.860 5.853 52.574 51.645 37.640 24.887 7.152 3.472 1.974 0 Mining, quarrying 38.847 181.273 10.919 9.613 4.237 2.244 2.668 4.556 0.342 0.612 8.346 0 1.535 1.623 Not applicable 0.005 0.096 19.167 0.067 75.707 0 0 0 0 0 0 0 0 0 Trade, restaurants, lodging 2.441 1.194 0.002 0.180 0 0 0 0.654 5.209 0 8.038 6.203 0 0 Transport and storage 46.231 23.422 31.060 27.785 28.653 42.668 26.147 47.394 49.330 49.478 39.350 63.140 48.747 37.600 Source: World Bank EDA database. TABLE 6.13 TANZANIA: DISBURSEMENTS ON EDA LOANS BY CREDITOR (millions of U.S. dollars) Creditor 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 African Develop- ment Bank 0 0 1.928 4.850 3.983 6.524 1.038 0.574 10.605 1.018 36.233 22.215 8.224 Canada 1.158 1.460 4.646 37.740 17.989 4.871 5.764 1.121 0 0 0 0 0 Denmark 2.368 2.026 1.373 1.067 8.858 8.268 13.426 24.189 21.296 8.632 15.061 2.174 0.524 Finland 0 0 0.481 0.853 0.869 7.212 5.965 3.538 1.846 0 0 0 8.520 France 0 0 0 0 0 0 0 0 0 7.714 25.094 25.571 20.805 Germany 0.042 0.746 1.264 3.322 7.010 10.518 9.480 12.965 23.308 7.689 7.289 3.948 2.239 Italy 2.120 1.142 1.562 1.380 0.601 0.263 0.082 0.019 0 0 4.307 3.899 3.646 Japan 0 0 0 22.385 38.797 31.762 22.377 13.606 15.066 16.556 17.846 36.262 32.107 Netherlands 0 0.206 0.722 2.325 5.055 8.135 7.180 6.207 0.0540 1.578 1.353 1.322 1.467 Norway 0.013 0.013 0 0 0 0 0 0 0 0 1.232 1.396 0.416 Sweden 7.923 10.254 10.833 6.893 4.219 2.223 1.129 0.390 0 0 0 0 0 Switzerland 0 0 0 0 0 0 0 0 0 0 0 0 0 United Kingdom 0 0 0 0.235 0.253 0.129 0.042 0.021 0.565 0.694 0.412 1.123 1.552 United States 4.370 9.472 4.213 2.998 3.261 12.233 5.023 11.820 6.555 9.013 8.471 8.892 8.163 World Bank 10.661 15.323 16.085 12.075 14.161 59.804 39.383 63.367 52.061 72.690 70.104 98.086 107.238 Total 28.655 40.642 43.107 96.123 105.056 151.942 110.889 137.817 131.356 125.584 187.402 204.888 194.901 Creditor 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 African Develop- ment Bank 7.012 10.407 22.772 34.381 0.290 133.204 4.215 0.052 24.402 1.515 13.114 10.908 6.548 51.644 Canada 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Denmark 3.207 2.274 0.034 22.975 0 0 0 0 0 0 0 0 0 0 Finland 0 5.835 0 0.333 0 0 0 0 0 0 0 0 0 0 France 11.627 5.171 1.902 1.507 1.176 0.919 2.216 4.091 4.501 1.450 0 0 0 0 Germany 1.512 0.985 0.654 0 0 0 0.107 0.124 0 0 0 0 0 0 Italy 10.234 24.960 15.168 29.231 20.152 30.813 30.363 36.466 22.503 77.671 5.368 0.214 23.753 2.681 Japan 17.185 4.641 8.899 5.351 0.247 0.102 0.051 5.139 7.912 5.622 7.493 0 0 0 Netherlands 1.320 0.676 0.399 0.141 0.057 0.023 0.014 0 0 0 0 0 0 0 Norway 0.157 0 0 0 0 0 0 0 0 0 0 0 0 0 Sweden 0 0.521 0 0 0 0 0 0 0 0 0 0 0 0 Switzerland 0 0 0 0 0 0 0 0 0 0.357 0 0 0 0 United Kingdom 0.560 0.369 0.231 11.101 0.345 0.180 0.782 18.722 2.147 0.067 3.454 10.599 2.817 0 United States 6.676 4.325 1.859 0.575 0.217 0.052 0 0.261 0 0 0 0 0 0 World Bank 90.251 74.345 45.903 92.730 95.004 137.771 116.284 189.297 180.664 237.796 155.891 183.203 159.774 136.111 Total 149.741 134.509 97.821 198.325 117.488 303.064 154.032 254.152 242.129 324.478 185.320 204.924 192.892 190.436 Source: World Bank EDA database. TABLE 6.14 TANZANIA: TOTAL GRANT ELEMENT BY CREDITOR Creditor 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 African Develop- mentBank 0 0 0.059068 -0.219959 -0.008669 0.672122 -0.00086 -0.00022 4.095906 -0.014364 15.44780 8.960966 1.009146 Canada 0.428885 0.847759 4.077151 32.15810 15.70584 4.312462 4.921095 0.954465 0 0 0 0 0 Denmark 0 0.077785 0.349752 0.605820 6.142921 6.691987 10.60145 19.36068 17.95983 7.178247 13.30686 1.891968 0.453129 Finland 0 0 0.235896 0.428941 0.436987 4.073034 2.070643 0.534894 0.120133 0 0 0 1.176416 France 0 0 0 0 0 0 0 0 0 0.799027 7.499406 6.939984 6.985513 Germany 0 0.358599 0.607601 2.301310 5.388103 8.352197 7.589165 10.29757 17.78176 5.504168 5.298418 2.870716 1.633721 Italy 0 0 0.074119 0.083336 0.040274 0.016187 0.006350 0.001471 0 0 0.114621 0.103763 0.545133 Japan 0 0 0 2.673332 4.663534 3.839221 2.704419 1.650193 3.937501 6.266033 9.210448 22.39592 19.25626 Netherlands 0 0.084988 0.279146 1.112096 3.182486 5.739715 5.200835 4.433109 0.004265 0.960408 0.926012 0.913500 0.920965 Norway 0 0 0 0 0 0 0 0 0 0 0.460238 0.521503 0.155405 Sweden 0.226757 0.182973 -0.029109 0.008076 -0.015288 -0.018836 -0.024991 0.000010 0 0 0 0 0 Switzerland 0 0 0 0 0 0 0 0 0 0 0 0 0 United Kingdom 0 0 0 0.000675 0.000726 0.000370 0.000121 0.000060 0.073153 0.090798 0.053903 0.156812 0.248030 United States 1.510034 2.166728 1.575552 1.336404 1.499387 6.546919 2.712198 6.481516 3.486323 2.944059 2.583836 4.078003 4.577506 World Bank 0 -0.024220 0.530551 1.117831 3.375268 14.71110 21.49838 30.52487 20,46963 31.78776 29.62509 63.81241 81.77445 Total 2.165676 3.694615 7.759730 41.60596 40.41157 54.93648 57.27881 74.23863 67.92851 55.51614 84.52665 112.6455 118.7356 Creditor 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 African Develop- mentBank 5.931847 9.040213 17.23824 28,35543 0.111428 103.7071 3.255719 0.008935 18.89661 1.145093 9.513868 7.717076 3.927583 35.31172 Canada 0 0 0 0 0 0 0 0 0 0 0 0 0 Denmark 2.858558 2.023903 0.029039 20.23753 0 0 0 0 0 0 0 0 0 0 Finland 0 1.887951 0 0.045979 0 0 0 0 0 0 0 0 0 0 France 4.799597 2.375770 1.127927 0.696489 0.342149 0.214480 0.695216 1.283022 1.398497 0.454902 0 0 0 0 Germany 0.968862 0.533921 0.351928 0 0 0 -0.034962 -0.040517 0 0 0 0 0 0 Italy 3.813431 12.21083 6.770602 11.90796 9.694800 15.00536 14.58652 17.50653 10.96460 33.01884 3.532701 0.148883 16.52540 1.593231 Japan 10.03737 2.654794 4.777057 2.451754 0.156750 0.064755 0.032782 0.926615 1.524406 1.123954 1.646417 0 0 0 Netherlands 0.792835 0.395292 0.228568 0.075769 0.034948 0.014101 0.008583 0 0 0 0 0 0 0 Norway 0.058650 0 0 0 0 0 0 0 0 0 0 0 0 0 Sweden 0 -0.096700 0 0 0 0 0 0 0 0 0 0 0 0 Switzerland 0 0 0 0 0 0 0 0 0 0.086695 0 0 0 0 United Kingdom 0.107953 0.053687 0.025175 6.596000 -0.013186 -0.008057 -0.037344 1.835822 -0.140262 0.003590 0.116446 -1.062441 0.270265 0 United States 4.284486 2.907917 1.276863 0.396112 0.149002 0.035562 0 0.037709 0 0 0 0 0 0 World Bank 54.76558 42.79851 28.81201 69.67844 72.43180 107.0718 87.70944 142.7214 136.3662 179.5619 106.7344 132.7891 110.0874 93.21205 Total 88.41918 76.78610 60.63742 140.4414 82.90769 226.1052 106.2159 164.2795 169.0100 215.3950 121.5438 139.5926 130.8107 130.1170 Source: World Bank EDA database. TABLE 6.15 TECHNICAL ASSISTANCE GRANTS TO TANZANIA AS A SHARE OF TOTAL TA GRANTS (millions of U.S. dollars) Recipient 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Tanzania 21.4 24.0 34.8 41.2 48.6 60.2 77.0 80.2 106.5 138.6 172.9 176.8 181.2 Total 437.6 503.5 594.5 693.9 786.4 1,015.4 1,038.6 1,106.9 1,378.2 1,687.6 2,206.7 2,206.0 2,072.3 Recipient 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 Tanzania 173.9 138.7 141.4 163.4 186.7 210.2 202.4 209.4 209.0 236.5 238.9 211.9 266.7 254.2 Total 2,071.5 2,010.0 2,227.4 2,769.6 2,989.1 3,331.1 3,371.6 3,667.6 3,767.8 4,160.8 4,185.5 3,651.7 4,178.1 3,915.8 Source: World Bank EDA database. Tanzania 357 NOTES Thanks for comments are due to Shanta Devarajan and Benno Ndulu. 1. Donors drew this negative conclusion about aid effectiveness even though they were seldom able to establish a causal link between aid and economic out- comes. 2. The ESRF study found that 17 items whose implementation either had not started or was unsatisfactory according a 1996 study have since then been imple- mented satisfactorily. However, six issues whose implementation was unsatis- factory then remain unsatisfactory to date. The issues that have been satisfactorily addressed include control of leakage through bonded warehouses; determina- tion of management/operations establishing criteria for bonded warehouses; improved supervision of transit trade (including tracking down goods and con- ducting audits); Zanzibar leakage (tariffs have been harmonized and Zanzibar is now under the same preshipment inspection system); exemptions (rampant exemptions have been reduced and the audit of IPC exemptions has been un- dertaken, and the audit report is now with Tanzania Revenue Authority for follow-up action). The issues on which little or no progress has been made in- clude CIS debt recovery, enforcement of tax laws and other legal provisions, and prosecutions and recovery of losses through exemptions. The CIS debts amount to about Tshs. 115.6 billion, owed by the government (Tshs. 12.5 billion), parastatals (Tshs. 72.6 billion), and the private sector (Tshs. 30.0 billion). Except for debts owed by government, many of the outstanding CIS debts have been disputed. The amount of debt that is not in dispute is in the region of Tshs. 15.2 billion. Even this can hardly be collected as most debtors have tended to plead hardship. The study examined the performance in CIS debt re- covery and observed that the report submitted by Tanzania Revenue Authority to the Treasury showed some progress, albeit modest. Until December 1997, a total of Tshs. 15.1 billion had been collected. The collection represents 9 percent of the total amount owed. The study could not establish any substantial collections made during the last six months of 1998. Reasons for this poor performance include shortcomings in the legal framework and inadequate information and documen- tation. Disputes relating to interest rates and exchange rates adopted in the recov- ery of the debts have also contributed to delays in payments. 3. It is hard to assess the significance of this trend as World Bank procedures for estimating cost only recently moved to dollar budgeting. 4. Raikes and Gibbon (1996) provide an extended discussion of the policy changes during this period. 5. This section draws heavily on Bigsten and Danielsson (1999). Bevan and others (1990) provide an extensive analysis of macroeconomic developments during the crisis period. 6. The government issued a progress report on the implementation of the Warioba report in February 1999 in response to critics. Also, in March 1999 a high-level, three-day retreat on good govemance was held. Participants included 358 Post-Socialist Reformers senior government officials (permanent secretaries and the heads of governance institutions, such as the accountant and auditor general, and anti-corruption institutions); resource persons (drawn from ESAMI, ESRF, and the University of Dar es Salaam, and selected public institutions such as TRA and the Corrup- tion Prevention Bureau); and a few donors (UNDP, USAID, Sweden, and the Netherlands). 7. Eriksson (1998) provides a detailed discussion of the impact of parastatals on the Tanzanian economy. Moshi (1998) offers evidence of the extent of the budgetary burden that the public enterprises have constituted. REFERENCES Adam, C., A. Bigsten, P. Collier, E. Julin, and S. O'Connell. 1994. "Evaluation of Swedish Development Cooperation with Tanzania." Secretariat for Analysis of Swedish Development Assistance, Ministry of Foreign Affairs, Stockholm. Bagachwa, M., A. Mbelle, and B. van Arkadie, eds. 1992. Market Reforms and Parastatal Restructuring in Tanzania. Economics Department and Economic Research Bureau, University of Dar es Salaam. Bevan, D., A. Bigsten, P. Collier, and J. W. Gunning. 1988. "The Decline in Tan- zanian Incomes during the Nyerere Experiment: Evidence from Household Budget Surveys." In W. van Ginneken, ed., Trends in Employment and Labour Incomes: Case Studies on Developing Countries. Geneva: International Labour Organization. Bevan, D., P. Collier, J. W. Gunning, A. Bigsten, and P. Horsnell. 1990. Controlled Open Economies. New York: Oxford University Press. Bigsten, A., and A. Danielsson. 1999. "Is Tanzania an Emerging Economy?" Report to the OECD. Gothenburg and Lund, Sweden. Processed. Chang, C., E. Fernandez-Arias, and L. Serven. 1999. "Measuring Aid Flows: A New Approach." Policy Research Working Paper 2050. World Bank, Devel- opment Research Group, Washington, D.C. Collier, P. 1991. "Aid and Economic Performance in Tanzania." In U. Lele and I. Nabi, eds., Transitions in Development: Aid Organizations and the Effectiveness of Aid. New York: St. Martin's Press. . 1997. "The Failure of Conditionality." In C. Gwyn and J. Nelson, eds., Perspectives on Aid and Development. Washington, D.C.: Overseas Develop- ment Council. Eriksson, G. 1998. "The Soft Budget Constraint: The Emergence, Persistence and Logic of an Institution: The Case of Tanzania 1967-1992." Ph.D. diss. Stockholm School of Economics. ESRF (Economic and Social Research Foundation). 1998. "SPA/JEM Project: Analysis of Progress and Relevance." Dar es Salaam. Embassy of Sweden. 1998. "Partnership in Development: Sweden and Tanza- nia." Dar es Salaam. Tanzania 359 . 1999a. "Informal follow-up of partnership commitments from the Helleiner Workshop in January 1997." Dar es Salaam. .* 1999b. "PFP, SAC, PER, CEM and CAS in Tanzania." Dar es Salaam. Feyzioglu, T., V. Swaroop, and M. Zhu. 1998. "A Panel Data Analysis of the Fungibility of Foreign Aid." World Bank Economic Review 12 (1):29-58. Helleiner, G. 1997. "Changing Aid Relationships in Tanzania? A Progress Re- port, Year-End 1997." Department of Economics, University of Toronto. . 1999. "Changing Aid Relationships in Tanzania." Report to the Govern- ment of the Republic of Tanzania. Dar es Salaam. Helleiner, G., T. Killick, N. Lipumba, B. Ndulu, and K-E. Svendsen. 1995. "Re- port of the Group of Independent Advisers on Development Cooperation Issues Between Tanzania and Its Aid Donors." Royal Danish Ministry of Foreign Affairs, Copenhagen. IMF (International Monetary Fund). 1997. "The ESAF at Ten Years: Economic Adjustment and Reform in Low-Income Countries." Occasional Paper 156. Washington, D.C. . 1998. "Report of the Group of Independent Persons Appointed to Con- duct an Evaluation of Certain Aspects of the Enhanced Structural Adjust- ment Facility." Washington, D.C. Kaufmann, D., and S. O'Connell. 1991. "The Macroeconomics of the Parallel Foreign Exchange Market in Tanzania." In A. Chibber and S. Fischer, eds., Economiic Reform in Sub-Saharan Africa. Washington, D.C.: World Bank. Likwelile, S. 1998. "Private Investment, Macroeconomic Environment and Eco- nomic Growth in Tanzania, 1967-1995." African Economic Research Consor- tium, Nairobi. Processed. Moshi, H. 1998. "Fiscal and Monetary Burden of Tanzania's Corporate Bodies: The Case of Public Enterprises." Research paper 75. African Economic Re- search Consortium, Nairobi. Moshi, H., and A. Kilindo. 1995. "The Impact of Government Policy on Macro- economic Variables: A Case Study of Private Investment in Tanzania." Afri- can Economic Research Consortium, Nairobi. Mutalemwa, D., P. Noni, and S. Wangwe. 1998. "Managing the Transition from Aid Dependence: The Case of Tanzania." Economic and Social Research Foun- dation, Dar es Salaam. Processed. Ndulu, B. 1993. "Exchange Rate Policy and Management in the Context of Eco- nomic Reforms in Sub-Saharan Africa." In G. Hansson, ed., Trade, Growth and Deevelopinent: The Role of Politics and Institutions. London: Routledge. Ndulu, B., and M. Hyuha. 1984. "Investment Patterns and Resource Gaps in the Tanzanian Economy, 1970-82." In N. Lipumba and others, eds., Economic Stabilisation Policies in Tanzania. Economics Department and Economic Re- search Bureau, University of Dar es Salaam. Ndulu, B., and S. Wangwe. 1997. "Managing Tanzania's Economy in Transition to Sustained Development." Discussion paper 14. Economic and Social Re- search Foundation, Dar es Salaam. 360 Post-Socialist Reformers Nyoni, T. 1997. "Foreign Aid and Economic Performance in Tanzania." Research Paper 61. African Economic Research Consortium, Nairobi. PricewaterhouseCoopers. 1999. "Tanzania Public Expenditure Review: Health and Education. Summary of Preliminary Findings." Dar es Salaam. Raikes, P., and P. Gibbon. 1996. "Tanzania: 1986-1994." In P. Engberg-Pedersen, P. Gibbon, P. Raikes, and L. Udsholt, eds., Limits of Adjustment in Africa: The Effects of Economic Liberalization, 1986-1994. Oxford: James Currey; Ports- mouth, N.H.: Heinemann. Stewart, F., J. Klugman, and B. Neyapti. 1999. "Socio-Economic Causes of Con- flict: A Comparative Study of Kenya, Uganda and Tanzania." Organisation for Economic Co-operation and Development, Paris. Processed. Tanzania. 1998a. "Public Sector Reform Program-me: Strategy and Action Plan, 1998-2003." Draft. - 1998b. "Guidelines for the Rationalisation of Donor's Local Cost Com- pensation." President's office, Civil Service Department, Dar es Salaam. Draft. - 1999. "The Education Sector Reform and Development Programme." Ministry of Education, Dar es Salaam. Wangwe, S. M., and W. Lyakurwa. 1998. "Aid Dependency: The Phasing Out of Projects. A Macro Perspective." Economic and Social Research Foundation, Dar es Salaam, and African Economic Research Consortium, Nairobi. Wangwe, S. M., and Y. M. Tsikata. 1999. "Macroeconomic Developments and Employment in Tanzania. Report for the ILO." Economic and Social Research Foundation, Dar es Salaam. World Bank. 1997a. "Tanzania: Policy Framework Paper, 1997/98-1999/2000." Washington, D.C. 1997b. "Tanzania: Country Assistance Strategy." Washington, D.C. 1999a. "Country Policy and Institutional Assessments: Report on the 1998 Ratings." Washington, D.C. Processed. - 1999b. "Public Expenditure Review: Tanzania Mission. Draft Aide Memoire." Dar es Salaam. - 1999c. World Development Indicators 1999. Washington, D.C. Mixed Reformers C6te d'Ivoire Elliot Berg Patrick Guillaumont Jacky Amprou Auvergne University Clermont-Ferrand, France Jacques Pegatienan University of Abidjan, Cote d'Ivoire ACRONYMS AND ABBREVIATIONS AFD Agence Francaise de D6veloppement AGEPE Agence d'Etudes de Promotion d'Emploi APEXCI Cote d'Ivoire Export Promotion Agency BCEAO Banque Centrale des Etats d'Afrique de l'Ouest CAISTAB Caisse de Stabilisation CFAF CFA franc DAC Development Assistance Committee (of the OECD) DCGTx Direction de Control des Grands Travaux EDA Effective development assistance ESAF Enhanced Structural Adjustment Facility (of the IMF) EU European Union FDI Foreign direct investment FSAL Financial Sector Adjustment Loan HRDP Human Resources Development Project IDA International Development Association (of the World Bank Group) IMF Intemational Monetary Fund L/J Legal/judicial MOJ Ministry of Justice NPV Net present value ODA Official development assistance ODF Official development finance OECD Organisation for Economic Co-operation and Development OHADA Organisation pour la Harmonisation des Droits d'Affaires en Afrique (Organization for the Harmonization of Business Law in Africa) OMOCI Office de Main d'Oeuvre de la C6te d'Ivoire PAGE Projet d'Appui a la Gestion Economique (Economic Management Support Project) PIP Public Investment Program PAGRH Projet d'Appui a la Gestion des Resources Humaines (Human Resources Management Support Project) PASCO Competitiveness and Regulatory Reform Adjustment Program PDSSIR Programme de Developpement de Secteur Sante Integr6 (Inte- grated Health Services Project) PSD Private sector development PMA Minimum activities package RCI R6publique de Cbte d'Ivoire SAL Structural Adjustment Loan SECAL Sectoral Adjustment Loan SOE State-owned enterprise TA Technical assistance TOF Total gross official flows (from the public sector) TOR Terms of reference UMEOA Union Monetaire Economique Ouest Africaine (West African Monetary and Economic Union) The Republique de C6te d'Ivoire (RCI) presents an unusual case in many respects. Few countries have experienced so many reforms over so long a period-22 World Bank adjustment loans since 1980, nine IMF programs, numerous bilaterally financed reform operations; no sec- tor seems to have escaped reformer attention. Between 1994 and 1997 the RCI was the World Bank's biggest adjustment loan customer, the recipi- ent of more than a third of all the Bank's adjustment lending to Africa. The RCI experience is special also in that it has been marked by a highly contentious relationship with the Bretton Woods institutions. Unlike most countries, whose reform efforts follow acknowledged fail- ures of development policies, the RCI political leadership retained its belief in the validity of its own development model, which had for de- cades been widely regarded as successful-indeed as an "economic miracle"; it was thought to be only temporarily derailed after 1980. The reform model of the Bretton Woods institutions was viewed to be ideo- logically based and thus unsuitable. Two other aspects of the Ivoirian experience are of special interest. To an uncommon extent, the reform process was externally driven. Until the early 1990s, the role of foreign technical assistance, mainly from France, was greater than anywhere else in the world. Reform programs were thus externally conceived in a multiple sense-via French techni- cal assistance working in the RCI government, via membership in the Franc Zone, and via the Bretton Woods institutions that were so strongly present after 1990. Among other effects, this gave rise to especially seri- ous problems of stunted local ownership of the reforms. Moreover, the RCI is often cited as a case where foreign aid retarded reform, notably by allowing delays in needed exchange rate adjustment after 1986. In this paper we consider sectoral as well as macroeconomic reform. Four areas are considered: trade liberalization, exchange rate policy, health and education sector public expenditures, and legal and judicial reforms. We describe the reform programs and analyze their origins- and to some extent their effectiveness. We assess the extent to which this country experience confirms the hypotheses that inspired the general research on aid and reform in Africa. The paper is a joint product of the four authors. Each had responsibil- ity for specific sections, working from an outline prepared by the lead author, Elliot Berg. He edited and modified portions of the individual section drafts, but these are presented more or less as drafted by the individual authors. Readers thus will not find compromise or consensus views on contentious issues (though we agree on most of the analysis). Rather, they are allowed to see the diversity of views and approaches among the co-authors. The paper is organized in seven sections. Section 2 describes aid pat- terns in C6te d'Ivoire-amounts, types, and origins. This chapter is the 365 366 Mixed Reformers work mainly of Jacky Amprou, research associate at the Centre de Re- cherche sur le Developpement International (CERDI), Universite de l'Auvergne, Clermont-Ferrand, France. He put the basic tables and graphs together. Jacques Pegatienan of Universite d'Abidjan drafted sections on bilateral aid, nonfinancial flows, and aid coordination. The other au- thors had some input in this section as well. Section 3 gives an overview of the economic and institutional back- ground to reform, the evolution of policy, and governance characteris- tics. It broadly follows the outline suggested in the terms of reference (TOR) for the Aid and Reform in Africa project. The main author of this chapter is Jacques Pegatienan. Sections 4 through 7 consider specific reform experiences and the role of aid. We focus on the four areas suggested in the TOR: macroeconomic policy, trade liberalization, public expenditure reforms in health and education, and legal/judicial system reforms. We try to treat each of these as an integrated reform story. We describe the reform agenda, comment on effectiveness of the reforms, and analyze the role of aid and aid do- nors in reform conception and processes. The generalizations and hy- potheses that are set out in the TOR are addressed in the context of each of the specific reform areas. This seemed the best way to reduce overgeneralization in analyzing the aid-reform relationship. Main con- clusions are set out in section 8. Section 4 focuses on the 1994 devaluation decision. This is a crucially important episode in the RCI reform experience, central to the assess- ment of aid-reform relationships in that country, and still controversial. We therefore concentrate on it and little else in the area of macroeco- nomic policy reform. The author of this section is Patrick Guillaumont, director of CERDI at the Universit6 de l'Auvergne. Section 5, on trade liberalization, is mainly the work of Jacques Pegatienan. Section 6, on public expenditure reforms in education and health, and section 7, on legal/judicial reform, were written by Elliot Berg, as was the concluding chapter. The research team was in Abidjan between July 31 and August 12, 1999. This timing was not ideal: many relevant people were away on vacation. But we had no other feasible option. The main consequence was that we were unable to see some key actors in the private sector, leaving us with a sketchy understanding of some issues. Mainly for this reason we have little to say about two important reform areas: the re- structuring of cocoa and coffee marketing and the privatization/deregu- lation of maritime transport. This is unfortunate, since the suitability of these reforms is contested by numerous Ivoirians. The World Bank office in Abidjan provided indispensable support. Special thanks are due to Chantal Dejeu, Acting Country Director, who guided us with a sure hand, and to Mme. Attobra, who organized our Cte d'Ivoire 367 complicated schedule of meetings. Thanks are also due of course to all those who took time in August to meet with us. A list of persons inter- viewed is attached as appendix 7.1. Appendix 7.2 includes the statistical tables. AID DEFINITIONS,VOLUMES, AND ISSUES Concepts of external aid abound. The standard or conventional concept is official development assistance (ODA), which is defined by the De- velopment Assistance Committee (DAC) of the Organisation for Eco- nomic Co-operation and Development (OECD) as financial flows to developing countries and multilateral institutions provided by official agencies, which meet the following criteria: their main objective is the promotion of economic development and welfare of recipient countries, and they are concessional and convey a grant element of at least 25 per- cent. ODA includes technical assistance grants and is usually calculated net of amortization (see figure 7.1 and tables 7.1 and 7.2, and table 7.8 in Appendix 7.2). FIGURE 7.1 COTE D'IVOIRE: TOTAL NET ODA, 1970-96 Millions of 1996 constant U.S. dollars 1,800 Multilateral 1,600 - _ DAC countries 1,400 - 1,200 1 000 800 600 400 i I 2000rI Source: OECD, Geographical Distribution of Financial Flows. 368 Mixed Reformers TABLE 7.1 ODA COMMITMENTS TO C6TE D'IVOIRE, 1975-98 (annual average in thousands of 1996 constant U.S. dollars) Donors 1975-79 1980-84 1985-89 1990-93 1994-97 Bilateral 147,407 264,242 273,895 839,755 537,281 France 47,575 96,342 111,093 348,263 201,932 European Union members 77,175 132,065 143,116 393,856 267,471 North America 22,657 24,220 9,434 56,651 22,477 Japan 0 11,615 10,252 40,985 45,401 Multilateral 22,018 39,638 3,129 75,305 397,689 IDA 0 0 0 39,647 317,142 European Union 22,018 39,638 479 25,357 68,844 African Development Fund/ African Development Bank 0 0 2,650 10,301 11,703 Source: OECD Creditor Reporting System. The main disadvantage of this indicator is that it does not take into account nonconcessional flows. This is an important omission since it means that sizeable financial flows from major multilateral donors are not included, such as regular loans of the World Bank and the African Development Bank and the standby agreements and extended facility loans of the International Monetary Fund (IMF). Another weakness of the net ODA concept is that it does not take account of interest payments on external debt. TABLE 7.2 C6TE D'IVOIRE: TOTAL ODA COMMITMENTS BY SECTOR, 1975-97 (annual average in thousands of 1996 constant U.S. dollars) 1975-79 1980-84 1985-89 1990-93 1994-97 Social infrastructure and services 25,939 29,069 10,856 63,797 150,779 Economic infrastructure and services 28,200 58,517 13,053 60,211 115,478 Production services 40,102 70,616 61,430 75,249 134,043 Commodity aid/general program assistance 15,514 11,178 49,876 314,382 235,751 Action relating to debt 0 18,746 24,743 35,034 137,863 Emergency assistance 0 355 0 1,563 643 Other' 20,752 20,264 6,133 28,182 38,266 a. Includes multisector/crosscutting, support to NGOs, and unallocated/unspecified. Source: OECD Creditor Reporting System. Cote d'lvoire 369 FIGURE 7.2 C6TE D'IVOIRE: OFFICIAL DEVELOPMENT FINANCE FLOWS (DISBURSEMENTS), 1970-96 Millions of 1996 constant U.S. dollars 2,000 2,000 - Nongrant component of official loans 1,800 = Grant equivalent of official loans 1,600 - Technical assistance grants 1 Non-technical assistance grants 1,200 ~ J\ 1,000 60°0° / __F 400 200 O . . . . . . . . . . . . . . . . . . . . . ld C ,b 9&o 1ct 9 #bP 9 b0t 992 P9, 9bc Note: Official loans include IMF concessional loans (SAF and ESAF) and noncon- cessional loans (enhanced facility and standby agreement). Source: Chang, Fernandez-Arias, and Serven 1999. A second concept, developed by World Bank researchers, is effective development assistance (EDA). It defines aid as the sum of the grant equivalent of all official loans plus transfers that are full grants, exclud- ing technical assistance (Chang, Femdndez-Arias, and Serv6n 1999).1 This adjusted measure uses the conventional grant data (from the OECD's DAC), but aggregates grant equivalents of loans rather than the full face value of all loans, even those with less than a 25 percent grant element. This indicator does not consider debt relief, whether it is in the form of debt forgiveness or debt rescheduling, as new aid.2 It retains only the concessional-window IMF flows (Structural Adjustment Facility loans and Enhanced Structural Adjustment Facility loans) and discards the nonconcessional windows loans (standby agreements and Extended Fund Facility loans). We present data on EDA as such (figure 7.3 and table 7.9 in Appendix 7.2), and include it in a broader concept: official development finance (ODF), which includes not only EDA (grants and the grant element of official loans), but also the nonconcessional part of the loans, technical 370 Mixed Reformers FIGURE 7.3 C6TE D'IVOIRE: EDA FLOWS BY TYPE, 1975-96 Millions of 1996 constant U.S. dollars 2,000 2 Grants (including technical assistance) 1,800 _ Loans 1,600 1,400 1,200 1 000 800 600 400 200 ~~~ N~~~ Source: Chang, Fernandez-Arias, and Serven 1999. TABLE 7.3 C6TE D'IVOIRE: TOTAL DISBURSEMENTS ON OFFICIAL LOANS BY SECTOR, 1975-96 (millions of 1996 constant U.S. dollars) 1975-79 1980-84 1985-89 1990-93 1994-96 Agriculture, forestry, fishing 503.1 382.5 553.7 644.8 374.3 Balance of payments support 0.0 820.1 578.2 969.3 783.7 Contribution to finance current imports 0.0 9.3 0.0 51.1 25.9 Social sectors 209.8 380.4 123.7 269.1 493.0 Manufacturing, trade, mining, finance 465.6 493.3 150.0 393.4 246.7 Infrastructure 860.6 796.8 935.6 701.6 133.5 Other 66.2 208.6 63.6 61.5 42.5 Note: Does not include IMF nonconcessional disbursements. Source: Chang, Femrnndez-Arias, and Serv6n 1999. C6te d'lvoire 371 TABLE 7.4 C6TE D'IVOIRE: OFFICIAL DEVELOPMENT FINANCE AND TOTAL DISBURSEMENTS ON LOANS BY CREDITORS, 1975-96 (annual average in millions of 1996 constant U.S. dollars) 1975-79 1980-84 1985-89 1990-93 1994-96 Total official developmentfinancea Bilateral 418.8 382.8 296.2 633.3 827.7 Multilateral 236.9 690.3 531.0 611.6 547.3 Total 655.7 1073.1 827.3 1245.0 1375.1 Total disbursements on loan Bilateral 254.2 229.0 164.8 371.9 274.8 Multilateral 199.0 673.2 403.3 459.8 425.1 Total 453.2 902.3 568.0 831.7 699.9 Total grants by creditor (inclusive of technical assistance) Bilateral 164.6 153.7 131.5 261.4 552.9 Multilateral 37.9 17.1 127.7 151.8 122.3 Total 202.5 170.8 259.2 413.3 675.2 Total non-technical assistance grants Bilateral 21.7 18.7 27.5 132.7 443.4 Multilateral 22.3 8.4 118.2 130.4 114.3 Total 44.0 27.1 145.7 263.2 557.7 Sources: Chang, Femendez-Arias, and Serv6n 1999; OECD-DAC. assistance grants, and IMF nonconcessional loans (see table 7.4 and fig- ures 7.2, 7.7, and 7.8). Another concept is total gross official flows (TOF), which is derived from the OECD-DAC database and is defined as ODA plus "other official flows." These other flows are transactions by the of- ficial sector whose main objective is other than development-motivated, or whose grant element is below the 25 percent threshold that would make them eligible to be recorded as ODA. The main classes of transac- tions included are official export credits, official sector equity and port- folio investment, and IMF and World Bank nonconcessional loans. ODF and TOF can be useful for examining the relation between policy reform and external assistance, assuming that the behavior of the gov- ernment is more influenced by the amount of total cash inflows than by the pattern or composition of those flows (see figure 7.4). We do not rely on them in this paper. Still another relevant concept is net transfers on aid, which is total net transfers on public and publicly guaranteed long-term debt from official creditors (including net transfers with the IMF) plus disbursements of grants. A related concept, aggregate net transfers, is net transfers on aid 372 Mixed Reformers FIGURE 7.4 C6TE D'IVOIRE: COMPARISON OF ODF AND TOF, 1970-97 Millions of 1996 constant U.S. dollars 2,500 Total gross official flows 2,000 (OECD) Total official development flows 1,500 - (Chang, Femnadez-Arias, and Serven) M 1,000 E _ l 500 o l l l l l o l lp l l l l l l l l l l l l 4) N Sources: OECD-DAC; Chang, Fernandez-Arias, and Serven 1999. plus private flows. The World Bank defines it in Global Development Fi- nance as net transfers on debt plus foreign direct investment (FDI), plus portfolio equity flows and official grants, minus FDI profits. The basic aid data are set out in the figures and tables in the text and in appendix 7.3. Stylized Facts The following main points emerge from these data and those provided in the statistical appendix. * The RCI relied relatively little on aid until recent years. Up to 1979 the country had access to international money centers, and private capi- tal inflows were substantial (figure 7.9). Modest aid inflows came mainly from France and other European countries; the World Bank had a small presence, and the World Bank Group's International Development Association (IDA) none at all until the 1990s. This changed in the 1980s, though the RCI was not a sizeable aid recipi- CMte d'Ivoire 373 ent until the late 1980s (see figure 7.1 and tables 7.1, 7.2, 7.4, and 7.8). From the mid-1960s to the mid-1970s, it received ODA amount- ing to US$200 million-US$300 million a year (constant 1996 dol- lars), though total development financial flows were much larger (see tables 7.4, 7.8, and 7.10). Until the late 1980s, the annual ODA inflow was less than US$300 million in most years (constant 1996 dollars). The amount began to rise in 1988 and 1989, averaged US$800 million annually between 1990 and 1993, then rose dramatically in the two post-devaluation years (1994 and 1995). It reached a peak of almost US$1.8 billion in 1994 before falling back to about US$500 million in 1997. In terms of aid intensity, ODA until the late 1980s was less than 2 percent of GDP, imports, and gross domestic investment. It rose to 10-20 per- cent in the early and mid-1990s, but returned to under 5 percent with respect to all these indicators in 1997 (figure 7.5). The largest part of official development flows to RCI was on nonconcessional terms until 1993. Thus while grants amounted to US$200 million or less (1996 dollars) from 1975 to 1987, nongrant equivalents of official loans were two to six times greater (see fig- ures 7.2, 7.7, and 7.8, and table 7.4). It was not until RCI became FIGURE 7.5 C6TE D'IVOIRE: AID INTENSITY, 1975-97 Percent Percent 250 25 Aid as a percentage of GNP (right axis) 200- 20 Aid as a percentage of gross domestic investment 150 (left axis) j 15 Aid as a percentage 100 - of imports of goods 10 and services \. (left axis) , 50 /-5 0 0 Source: World Bank, World Development Indicators 1999. 374 Mixed Reformers IDA-eligible in 1994 that grants exceeded nongrants. In 1996 grants made up 80 percent of ODF. • The heavy borrowing after 1980 put RCI in the heavily indebted cat- egory. In 1996 the net present value (NPV) of public debt was more than four times the recent average value of exports. Debt service was over 35 percent of exports. The NPV of public debt was 145 percent of GDP and 640 percent of government revenue. Public debt service claimed 52 percent of government revenue.3 Signifi- cant debt relief has been obtained from the Paris Club (table 7.13 in Appendix 7.2). And in 1997 a debt restructuring agreement was signed with commercial creditors to restructure commercial debt that had not been serviced since 1987. Virtually all outstanding com- mercial commitments were covered. The amount involved is al- most US$6.8 billion, of which US$2.4 billion is principal and US$4.4 billion is interest, including arrears and penalties. * A dramatic shift took place in the composition of aid after 1990: adjust- ment loans and grants became dominant, accounting for almost half the total in 1990-93 (see tables 7.2 and 7.3). The World Bank's shift was particularly strong. In the period 1988-97, adjustment lending accounted for 79 percent of total Bank commitments to the RCI, compared with only 33 percent for the rest of Africa.4 Project lend- ing suffered. Before 1981, most Bank lending went to agriculture, transport, urban development, and water and sanitation. But the share of these four sectors fell from 76 percent of Bank commit- ments in 1968-79 to 26 percent in 1994-96. * Not only did project lending decline, it also shifted toward projects in support of policy reform-market liberalization and institutional or capacity development, for example. (See appendix table 7.11 show- ing World Bank policy loans.) * Aggregate aid flows show a sharp rise in the three years immediately following the devaluation of the CFAfranc (1994-96), averaging some US$1.25 billion in constant dollars annually. They declined to less than half a billion dollars in 1997, though this level of concessional assistance was higher than it had been during most of the pre-re- form period. * Technical assistance (TA) continues to make up a large share of total aid. Actually, there is some dissonance in the data on technical assis- tance. Table 7.5 shows an increase in TA grants between 1985-89 and 1994-96. But this seems inconsistent with casual observation- very few expatriate TA providers are seen in government offices- and with trends in TA from France, surely the main provider. The number of French cooperants fell from an average of over 2,100 in 1985-89 to 600 in 1994-96 and some 250 in 1998. Part of the expla- nation is probably found in the financing arrangements for French CBte d'lvoire 375 TABLE 7.5 C6TE D'IVOIRE: TECHNICAL ASSISTANCE GRANTS BY CREDITOR, 198o-96 (annual average in milhons of 1996 constant U.S. dollars) 1975-79 1980-84 1985-89 1990-93 1994-96 Bilateral 142.9 135.1 103.9 128.7 109.5 Multilateral 15.5 8.7 9.6 21.4 8.0 Total 158.5 143.7 113.5 150.1 117.5 Sources: Chang, Fernandez-Arias, and Serv6n 1999; OECD-DAC. TA, a large part of whose cost was paid for by the RCI itself, which means that the decline in TA numbers would not be matched by a proportional decline in its external financing. * The aggregate inflow data show a rapid decline in the second part of the 1990s. The numbers derived from World Bank/OECD data (see above) show "net transfers on aid," as set out in figure 7.6. These indicate that net transfers on aid were positive in most years dur- ing the 1990s, including 1995 and 1996. But after reaching a peak in FIGURE 7.6 C6TE D'IVOIRE: NET TRANSFERS ON AID, 1975-97 Millions of 1996 constant U.S. dollars 1,200 _ Bilateral donors 1,000 = Multilateral creditors - Total 800- 600 400- 200 -200 W -400 -6 0 0 . . . . . . . . . . . . . . . . . . . . . . . Souirces: OECD-DAC; World Bank, Global Development Finance 1999. 376 Mixed Reformers 1994 (the year of the devaluation) they decreased sharply in the following years, becoming negative in 1997. Table 7.6 (from official Ivoirian sources), which refers to net transfers on debt (without grants), shows negative transfers as early as 1995. Nonfinancial Aid We turn now to some key issues. Nonfinancial aid was especially impor- tant in the earlier phases of Ivoirian development. It took the form of technical assistance from France, primarily secondary school and uni- versity teachers. Other TA was in health and social development, infra- structures, culture, and scientific research. A total of 31,000 cooperants were present during 1981-95, more than 80 percent of them professors. The number decreased drastically over the years. The rate of decline reached a peak in 1993 (-30.8 percent). As noted above, the number of French cooperants today is a mere 250, of whom 140 are professors, a 58 percent drop from 598 in 1995.5 Nonetheless, technical assistance overall still seems to comprise almost a quarter of total ODA. In the 1960s and 1970s one of C6te d'Ivoire's distinctive features was the influence of French TA over the system of decisionmaking in techni- cal administrations as well as in the president's office. In those years, one could find almost no Ivoirian nationals in the position of adviser; France provided most advisers in this pre-structural adjustment period, when the Bretton Woods institutions were absent. President Felix Houphouet-Boigny wanted a gradual replacement of the French, but he wanted good teaching and good administration too, so he continued to employ French technical assistants, paying much of the cost out of local resources. More important, he failed to put in place meaningful and re- liable arrangements for the real transfer of technical skills and compe- tencies to Ivoirians. Resident expatriate technical assistants everywhere tend to develop an informal network of information and influence to which local people have no access. Where there are many expatriate TA people this is exag- gerated, and where there are many from one country it is more pro- nounced still. This was particularly evident in the RCI. Aside from its negative effects on transfer of knowledge, this situation put French cooperants in a particularly influential position; dialogue between do- nors and the government was filtered to a significant extent through the screen of French technicians. To the extent that foreigners served as technical advisers to the presi- dent, to all ministers, and to senior management of state-owned enter- prises, ideas generated in the country over a long period could not be of local vintage. French TA deserves credit for bringing about some good results-for example, the vigorous economic growth recorded until the TABLE 7.6 C6TE D'IVOIRE: AGGREGATE NET TRANSFERS BY DONOR, 1995-97 (billions of current CFA francs) Disbursements Repayments Net Transfers 1995 1996 1997 Total 1995 1996 1997 Total 1995 1996 1997 Total IMF 175.1 138.9 0.0 314.0 99.2 58.5 28.4 186.1 75.9 80.4 -28.4 127.9 IBRD and IDA 244.8 253.6 152.7 651.0 305.5 300.6 272.7 878.8 -60.7 -47.0 -120.0 -227.7 UNDP 0.8 1.2 4.6 6.6 0.0 0.0 0.0 0.0 0.8 1.2 4.6 6.6 EU 92.3 74.5 40.3 207.1 17.0 18.1 16.8 52.0 75.2 56.4 23.5 155.1 IFAD 1.7 0.4 1.3 3.4 0.9 1.3 1.9 4.1 0.8 -1.0 -0.6 -0.7 AfDB 45.9 64.1 44.7 154.7 119.2 141.4 131.6 392.3 -73.3 -77.3 -87.0 -237.6 BOAD 19.3 6.5 7.4 33.1 1.2 3.4 3.5 8.0 18.1 3.1 3.9 25.1 Paris Club members 345.9 163.9 71.7 581.5 303.3 237.9 147.6 688.8 42.6 -74.0 -75.9 -107.3 London Club members 0.0 0.0 0.0 0.0 0.0 0.7 6.0 6.7 0.0 -0.7 -6.0 -6.7 Other 31.6 28.5 67.1 127.2 133.6 78.0 71.5 283.1 -102.0 -49.5 -4.4 -155.9 Total 957.2 731.6 389.8 2078.6 979.9 840.0 680.0 2499.9 -22.7 -108.4 -290.2 -421.3 Source: C6te d'Ivoire 1998b: 25. 378 Mixed Reformers late 1970s and the relatively strong public administration. By the same token, they have to bear some responsibility for the failures-inadequate attention to knowledge transfer, reliance on approaches and institutions that would prove unsustainable without lavish use of imported man- power, and the maintenance or adoption of inappropriate policies in key sectors like education and health. Education was the single most impor- tant area of intervention for French technical cooperation, yet reform made slow progress during their tenure in that sector and is today in much disarray. After the mid-1980s government officials changed their attitudes to- ward the use of local human resources. Nowadays, in C6te d'Ivoire, the norm is a local technical adviser. Expatriate, including French, technical advisers are rare, in ministries as well as in the president's and the prime minister's offices. Academics and university professors are more and more associated with the analytical and technical work of the adminis- tration. One testimony to the new situation is the fact that there is not a single expatriate technical adviser in the prime minister's office. This is now the rule. Outside technical influence remains strong, though its origins have shifted. In the era of structural adjustment, the IMF and the World Bank have become powerful advisers to the government. The Bretton Woods institutions' advice is built on an analytical and technical foundation that is usually very impressive, and hence influential in all circles, in- cluding that of local technical advisers. Little analytical and technical work is conducted by locals that is not subject to the intellectual influ- ence of the Bretton Woods institutions. But while ideas in general and reform ideas in particular continue to be externally generated, local coun- terparts are more and more equipped with a capacity to develop home- grown reform ideas, and able to engage in real dialogue with their aid partners. Bilateral Financial Aid With respect to bilateral aid, we concentrate on France. Because of that country's close relationship with the RCI, its aid connection merits spe- cial attention. France is also a major donor, the third-largest ODA con- tributor behind the IMF and the World Bank; it provided two-thirds of bilateral aid in 1996, well ahead of Japan's 26 percent and Germany's 5 percent. French financial assistance to C6te d'Ivoire is accounted for by his- tory, commercial and financial interests, and political strategy. Colonial history and cultural links strengthened the ties with the former colo- nizer, as did the strong pro-French sentiments of the RCI's first and long- time head of state, President Felix Houphouet-Boigny. The president not Cote d'lvoire 379 only maintained French influence in C6te d'Ivoire, but also championed it in other francophone countries. France has commercial and financial interests in Cote d'Ivoire, which is its second-largest West African mar- ket (after Nigeria). The equity capital of Ivoirian modern sector firms is largely French. Large-scale trading businesses are French or French-con- trolled. Finally, Cote d'lvoire is a political asset as it is a heavyweight partner in the region. France provided financial support for the economic development of C6te d'Ivoire. This financial assistance was vital, especially during the difficult years starting in the mid-1980s and in 1990-93, when the coun- try was trying to avoid bankruptcy. The former colonizer paid its ex- colony's debts to save it from default. It can be argued that by so doing, it relieved the pressure for structural adjustments, thereby retarding re- form. On the other hand, pressures remained extremely strong, and the country eventually did pursue many reforms. Debate continues on this question, which we review in section 4. As pointed out there, the fact that President Houphouet-Boigny strongly opposed devaluation could not be overlooked. In the years just before the devaluation, France's influence on policy was diluted by the strong presence of the Bretton Woods institutions. Its influence by no means disappeared, however. Aside from other doors open to it, the strategy of cofinancing most reform programs provided continuing entr6e. The mechanism of consultation of cofinanciers was used by France to press its views, dissent from policies, and occasionally provide the political backing that CBte d'Ivoire needed in the negotia- tions with these institutions. After the devaluation, the role of cofinancier of structural adjustment programs was strengthened. Indeed, the Agence Francaise de D6veloppement (AFD) spent 600 billion CFA francs in C6te d'Ivoire in the period 1994-98, of which more than half (CFAF 315 bil- lion) was for structural adjustment loans. These resources financed ex- ternal debt service and internal debt arrears to the private sector. One of the key implications of the liberalization reforms was a wider opening of the local market to non-French businesses. This challenge was foreseen by France and was met by the extension of CFAF 150 bil- lion in credits to the private sector-25 percent of total AFD assistance between 1994 and 1998. Almost three-fourths of this assistance was in the form of loans to finance the participation of French-controlled firms in the provision of public services-electric power, gas, water, transport (rail and airport), seaport, and toll bridges. (It is not irrelevant that a French firm won the bid privatizing the national telecommunications network.) There is some feeling among Ivoirians that French aid "unlevels" the private sector playing field. Although a major problem of local compa- nies has always been difficulty of access to credit, French firms engaged 380 Mixed Reformers in the local market benefit from guarantees to facilitate borrowing. These amounted to CFAF 38 billion in 1994-98, that is 25 percent of AFD assis- tance to the private sector. Thus it can be argued that the French aid inflows associated with the liberalization reforms reinforced the already strong position of French firms in the economy, especially in key strate- gic sectors. It should be noted, however, that a somewhat contrary view was expressed by one private sector spokesman, who believed that Levantines and others have won market share from the big French trad- ing houses since the devaluation. Aid Coordination In 1991 the government created an institution specialized in the coor- dination of aid (grants and loans). The objectives of this body were extended in 1994 to cover the monitoring of public investments as well. The institution is the "Comite de mobilisation des financements exterieurs et de suivi des investissements publics," or COMFESIP. Its objectives are to speed up the preparation of funding requests, sched- ule expected disbursements of exceptional funding, compare sched- uled disbursements with liquidity needs of the Treasury, and speed up the implementation of public investment projects. Its main activities are periodic review of the project portfolio, periodic publication of sta- tus reports on project implementation, and capacity building in aid management and coordination. It produces a major annual report on aid inflows and disbursements. COMFESIP is composed of representatives of the main economic and financial units involved in aid management. It is administered by a tech- nical group and a permanent technical secretariat. The technical group includes the prime minister's office and directors of the budget, the Trea- sury, the planning agency, and the Banque Centrale des Etats d'Afrique de l'Ouest (BCEAO). The permanent technical secretariat is Bureau Na- tional d'Etudes de Developpement (BNETD). That comprises the main formal structure on the government side. It produces useful documen- tation, but it is not clear how effective it is in operational terms-that is, in coordinating actual donor programs. Donors, on their side, have a variety of consultative arrangements, mostly at the sectoral level. These meet informally and produce no reports. In 1997 there were three such groups, each chaired by a lead donor. These groups met monthly in the first semester of 1997, but since then seem to have met sporadically or not at all. There are informal relations between donors and the COMFESIP sec- retariat, but roundtables on sectoral issues have been organized to de- termine strategies and to pledge funding budgets. By mid-1999 four such roundtables had been organized: for environment, health, regional plan- C6te d'Ivoire 381 ning and decentralization, and education. Donors use common letters to government to express their concerns and opinions on policy issues. Despite the fairly extensive formal structure of consultative mecha- nisms, it is not clear how much genuine consultation takes place. Some incidents suggest that it is highly sporadic and imperfect. In the health sector, some donors believe true consultation has been minimal, and there does appear to be considerable cross-purpose ac- tivity. In agriculture there has not been wholehearted donor consensus with respect to the nature and pace of state withdrawal from cotton, coffee, and cocoa marketing, and this has limited the extent of mean- ingful coordination. How intensive coordination may be in other sec- tors is not clear. POLICY REFORM AND INSTITUTIONAL DEVELOPMENT The history of C6te d'Ivoire's economic performance can be divided into five periods. The years between 1960 and 1977 were marked by high GDP growth (6-7 percent a year) and general prosperity. Terms of trade and were favorable and stable, and surpluses derived from export rev- enues financed a high level of public investment. The 1973 oil shock did not slow down growth because the prices of cocoa and coffee rose si- multaneously. They achieved record prices in 1976-77, which led to a surge in the rate of public investment financed by the increased rev- enues of the Caisse de Stabilization (CAISTAB) as well as by foreign borrowing. This period was remarkable for the big size of public invest- ment programs, and their high unit cost and low economic returns (World Bank 1998b). In 1978-80 financial crisis hit. The situation deteriorated in 1978 with a significant slowdown of the economy as a result of the decline in the prices of cocoa and coffee and consequent fall in public investment. In- flation remained high because of higher import prices caused by the 1979 oil shock and a parallel expansion in the money supply. The rise of interest rates on world capital markets and the sharp deterioration of the terms of trade between 1978 and 1980 induced an increased deficit of the current account in that period. The rise in domestic inflation con- tributed to the appreciation of the real exchange rate. The demand-driven growth pattern came to an abrupt end (Hopkins 1989). Debt service be- came so heavy that the government was forced to call on the Bretton Woods institutions for financial assistance. Between 1981 and 1986 the economy was stabilized, but with reces- sion. The major economic problems facing the country were overexpan- sion of the public sector with the consequence of declining public savings and rising external debt, a large and ill-conceived Public Investment 382 Mixed Reformers Program (PIP), distortions in the structure of price incentives for agri- cultural products, and excessive protection in industry with induced high costs and low export incentives. These trends led to the erosion of the competitiveness of the economy over the 1970s. These problems were addressed by a succession of IMF and World Bank loans: an Extended Fund Facility (EFF) agreement with the IMF in 1981 and Fund standby agreements (SBAs) in 1984, 1985, and 1986; and three successive World Bank structural adjustment loans (SALs) in 1981, 1983, and 1986. The stabilization and adjustment measures contained in these loans are described below. These adjustment programs and policies, coupled with favorable changes in the external environment, had positive results. Despite terms of trade deterioration early in the decade and a drought in 1983, internal balance was restored and competitiveness improved as a result of de- preciation of the real exchange rate (the devaluation of the French franc), and tighter fiscal and monetary policies. In 1984 exports increased, the current account moved into surplus, and the growth rate picked up. It increased further in 1985. However, the debt burden increased also: the ratio of foreign debt to GDP that year increased to 84 percent from 41 percent in 1980. In 1986, the real growth of the economy was still posi- tive though lower than population growth. This improvement was short-lived. From 1987 to 1994 the RCI struggled to avoid bankruptcy. The terms of trade deteriorated in 1987 due to a falling dollar and declining world prices of cocoa and coffee. The growth rate fell, the public investment-to-GDP ratio fell, and the current account was in deficit because of the large interest payments on public debt. The government was forced to suspend its debt service pay- ments in May 1987. No financial assistance came from the World Bank in 1988 because of the absence of an agreement on a macroeconomic framework (Ouayogode and Pegatienan 1994). Despite the deterioration of the terms of trade and the fall of CAISTAB's revenues after 1986, government maintained the producer price of cocoa and coffee and tried to stock cocoa in an attempt to influ- ence the world market. The result was increased fiscal deficit, accumu- lation of domestic arrears to the private sector, and an illiquid banking system. During this period intense reform activities were initiated. Two standby agreements were signed with the IMF, one for a year in 1988, a second for two years in 1989. The 1989 program had two main objec- tives: to reduce producer prices of cocoa and coffee and to cut civil ser- vants' salaries. The producer price measures were implemented; they were halved in September 1989. But social resistance prevented the sec- ond objective from being realized. Nonetheless, another standby agree- ment was signed in 1991 for one year. Between 1989 and 1991 the World Bank supported the government's reform efforts with six sectoral adjustment loans or credits (SECALs) in Cote d'Ivoire 383 the following areas: human resources (the Human Resources Develop- ment Project, HRDP), competitiveness (Competitiveness and Regulatory Reform Adjustment Program, PASCO), banking (Financial Sector Ad- justment Loan, FSAL), agriculture, water, and energy. Most disburse- ment on these loans, however, did not occur until 1994, after devaluation. Notwithstanding the signing of the EFF, SBAs, SALs, and SECALs, and some implementation, such as cutting producer prices and some civil service salaries, financial crisis dominated the early 1990s. The main policy concern was to avoid falling into arrears with the multilateral lenders. In the face of the structural disequilibria caused partly by external shocks and partly by domestic policies, it appeared that the adjustment effort was too little because it concentrated on internal disequilibrium. It also appeared too late, probably because of its political and social costs. It was under these circumstances that the CFA franc was devalued in January 1994. The post-devaluation years, 1994-98, have been marked by economic recovery. The 50 percent devaluation was aimed at improving the inter- nal and external competitiveness of the economy. It was coupled with measures designed to contain inflationary pressures. The implementa- tion of this policy package improved external competitiveness and eco- nomic performance. IMF and World Bank support played a big role in bringing about these results. An Enhanced Structural Adjustment Facil- ity (ESAF) agreement was signed in 1994 and another in 1998 for three years each. Large inflows came from the World Bank via loans signed earlier (the Economic Recovery Credit, HRDP, PASCO, and FSAL) as well as programs in agriculture and transport. Along with and within the framework set by these lead donors, other multilaterals and bilaterals extended their financial assistance to the country as well. As noted in section 1, aid inflows in 1994 amounted to US$1.7 billion, more than twice the pre-devaluation levels, and remained high until 1997. The positive real growth obtained after devaluation was a dramatic change in comparison with the desperate pre-devaluation situation. (See the economic indicators in table 7.7.) But the recovery is still fragile and poverty issues have only begun to be addressed. According to a World Bank poverty assessment, the rate of headcount poverty increased from 11 percent of the population in 1985 to 31 percent in 1993 and 37 percent in 1995 (World Bank 1997). The last figure is contested, and the four years of 5-6 percent growth since 1995 undoubtedly improved the situ- ation. But there is no room for complacency. Policy Reforms Since the inception of the economic and financial crisis in the late 1970s, the reforms undertaken were of different types: macroeconomic, struc- tural, sectoral, and social. The reform in macroeconomic policies was TABLE 7.7 C6TE D'IVOIRE: SELECTED ECONOMIC AND FINANCIAL INDICATORS, 1987-96 Annual percentage changes 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 GDP in real terms -0.4 1.1 2.9 -1.1 0.0 -0.2 -0.3 2.0 7.1 6.8 GDP per capita -4.1 -2.3 -1.1 -5.0 -3.8 -4.0 -4.0 -1.9 3.2 3.1 Consumer price index (annual average) 7.0 6.9 1.0 -0.7 1.6 4.2 2.4 32.2 7.7 3.5 External sector on the basis of CFAF Export of goods f.o.b. -20.6 -14.6 5.1 -1.9 -6.6 2.2 -8.5 123.3 20.3 15.0 Import of goods f.o.b. -3.2 -9.4 8.8 -15.4 10.9 2.0 -3.8 76.7 35.3 20.6 Export volume -6.0 -4.1 9.4 11.0 -6.9 7.8 -11.0 4.9 70.1 20.3 Import volume -6.2 -10.2 -0.6 -13.4 9.6 6.8 -8.0 -9.5 37.6 12.4 Terms of trade (deterioration) -18.1 -11.7 -12.3 -9.5 -1.0 -0.8 - 9.0 14.2 -10.9 Real effective exchange rate (depreciation) 8.3 0.7 -6.8 2.4 -2.8 3.3 -3.1 -34.5 8.3 -2.5 Monetary Net domestic assets 5.2 10.1 -8.0 -3.7 -2.1 -0.1 -7.1 -20.0 5.8 0.7 Of which: Government -2.5 90.1 -1.8 -4.1 2.6 12.9 28.8 -15.9 -1.8 -1.6 Private sector 5.7 -1.6 -5.0 -3.3 -4.0 -9.0 -18.3 10.2 19.9 2.2 Money and quasi-money M2 -1.4 0.2 -7.4 -3.1 0.0 -1.9 -4.3 47.0 17.0 3.0 Percent of GDP Public finances Total revenues and grants - 25.7 21.8 21.5 19.9 20.2 18.1 20.6 22.8 23.2 Total expenditures - 40.3 38.5 33.4 32.9 31.9 30.0 27.1 26.5 25.3 Overall deficit (-) payment order basis - -14.6 -16.6 -12.0 -13.0 -11.7 -11.9 -6.5 -3.7 -2.1 Primary balance - -3.1 -5.8 -1.1 -0.8 -0.3 -2.2 3.0 4.9 50.6 Gross domestic investments 11.8 11.5 10.3 8.5 8.6 8.5 7.8 11.1 12.9 13.9 Of which: Central government 5.1 4.7 4.4 3.6 3.4 3.8 3.7 4.1 4.2 4.3 Nongovernment 6.7 6.8 5.9 4.9 5.1 4.7 4.1 7.0 8.7 9.6 Current account balance -8.6 -11.5 -12.5 -12.2 -11.5 -11.4 -11.0 -1.0 -5.0 -4.8 Extemal public debt 77.0 83.1 93.3 107.3 116.6 127.0 141.9 183.9 157.9 153.5 Government domestic debt 44.6 27.7 25.1 20.9 Specified units GDP at market prices (billions of CFAF) 3,031.7 3,054.5 3,112.8 2,939.3 2,960.0 2,952.1 2,946.2 4,256.0 4,987.7 5,473.6 CFAF per U.S. dollar 300.5 297.9 319.0 272.3 282.1 264.7 283.2 555.2 499.1 511.6 Population (millions) 10.6 11.0 11.4 11.8 12.2 12.7 13.2 13.7 14.2 14.7 Source: IMF. 386 Mixed Reformers initiated through the IMF's Extended Facility agreement in 1981 and by six SBAs between 1984 and 1991. These concerned fiscal issues (govern- ment revenue recovery and expenditure reduction, rationalization of public investment programs, control of domestic debt, and restructur- ing of public enterprises), and monetary and financial problems. Over the years, the main objectives of fiscal policy reforms were the elimina- tion of budget deficits and the generation of surplus in the primary bud- get, in order to pay an increasing share of debt interest and to cover expenditures on physical and social infrastructure. The monetary and financial reforms aimed at reducing domestic credit, maintaining the liquidity of the banking system, and improving the mobilization of do- mestic savings. The structural adjustment loans were initiated in the early 1980s. SAL I, in 1981, focused mainly on agricultural pricing and investment policy. SAL II, in 1983, continued to focus on agriculture and added industrial and housing policy. The reform in the industrial sector dealt with export incentives, tariff and nontariff barriers, and the revision of the invest- ment code to include measures favorable to small and medium enter- prises. The new housing policy favored lower income housing, reduced the housing benefits for civil servants, and designed mechanisms to cut recurrent housing costs. SAL III, initiated in 1986, aimed at reinforcing the measures of the first two SALs and improving the relative price of tradeables in order to stimulate exports, correcting the urban/rural terms of trade, increasing rural incomes, and encouraging private investment. After the first three SALs, which were a mixture of structural and sectoral reforms, the first ESAF loan was granted by the IMF in 1994. It had the following objectives: improvement of competition and the regu- latory framework, liberalization of the labor market, price decontrol, civil service reduction, accelerated privatization and private sector develop- ment, and deregulation of export crop marketing. Structural adjustment became more specific with sectoral reforms. Between 1991 and 1993 the World Bank supported the three SECALs mentioned earlier, in regulatory reform, financial sector reform, and edu- cation and health. The reforms in these sectors were deepened after the 1994 devaluation, and reinforced by a private sector development (PSD) program. Post-devaluation reforms in agriculture aimed at improving farmer incentives and removing the state from export crop production and marketing. Specific measures concerned the following items: the lib- eralization of cocoa and coffee marketing and the restructuring of CAISTAB, the launching of a new system to allocate export rights by auction, organization of a closer association of the private sector with the management of the cocoa and coffee trade, and liberalization of rice imports. Cote d'Ivoire 387 Far-reaching social sector reforms were introduced in the Human Resources Development Project, the SECAL mentioned earlier. These were reinforced in 1997 by the adoption of a national program for eradi- cating poverty, with measures designed to increase the rate of school enrollment and literacy, improve basic health care, increase women's participation in development activities, curb rural exodus, and control population growth. It is clear that the RCI undertook, over a 20-year period, a vast program of reforms involving virtually every area of eco- nomic policy and most economic institutions. Table 7.11 suggests the substantive range of the reform agendas over the years. Some of these reforms were soundly conceived, others less so. In some-particularly during more recent years-the Ivoirian input was substantial. In most it was slight. Aid and aid donors, and particularly the Bretton Woods in- stitutions, clearly played a significant role in reform conception and de- sign. Despite uneven implementation, the economy is more open than before and more liberal. Instances of unequal or discriminatory treat- ment persist. The effects of price liberalization in particular are dampened by in- formation retention by some technical administrations. The perceptible improvements at the macroeconomic level have not yet materialized at the microeconomic level among the urban population, whereas the ru- ral population is probably better off. Significant institutional improve- ments have also occurred, in the budget process for example, and in export and investment promotion agencies. These favorable changes, however, are balanced by rampant corrupt practices in the administra- tion and by the lack of responsiveness of the political system. Institutions and Governance Two institutions play a critical role in the working of the economy and in the reform process: CAISTAB and the CFA monetary union. These insti- tutions were the pillars of a system that worked with remarkable suc- cess for two decades. They were supported by France and by the Bretton Woods institutions, and they were deeply anchored in the thinking of President Houphouet-Boigny. All of this made it especially difficult to adapt them when changed circumstances called for reform. CAISTAB is the marketing board for cocoa and coffee crops. Until September 1989 it applied the principle of never reducing producer prices and of increasing them less than proportionally to rising world prices. The consequence of that asymmetrical behavior was to tax peasants and transfer to the govemment budget resources generated by the positive difference between world and producer prices. It also financed budget- ary losses in cases of negative differentials between the two prices. 388 Mixed Reformers CAISTAB was instrumental in the overexpansion of public investment and in the acceleration of state participation in productive and market- ing activities. For decades, CAISTAB was a bone of contention in the policy dialogue between Cote d'Ivoire's government and the Bretton Woods institutions. C6te d'Ivoire's membership in the West African Economic and Mon- etary Union (Union Monetaire Economique Ouest Africaine, UMEOA) is a determining factor in its exchange rate and fiscal and monetary poli- cies. The UMEOA maintains a fixed rate of exchange between the French franc and the CFA franc. This institutional feature is critical to the evolu- tion of external competitiveness and to adjustment to external shocks. Since for many years a nominal devaluation of the CFAF could not be implemented to reduce real wages and thereby restore external com- petitiveness, the only way to have that happen was to reduce nominal wages. But opposition to the reduction of civil servants' nominal sala- ries in 1989 left the government with the option of restoring competi- tiveness through cuts in public expenditure other than salaries. After a floor was reached in cuts of current expenditure, the government turned to large reductions of public investment, which accelerated the decline of the economy. The 14 countries of the franc zone accepted devaluation of the CFAF in January 1994 to restore external competitiveness. That decision is discussed in detail in section 4. The government demonstrated a strong political commitment to un- dertake appropriate development programs that would transform the economy and create new activities. This commitment was expressed by an investment code that granted many incentives to investors, and by the direct participation of the state in productive activities through state- owned enterprises and the creation of physical and human capital. The state intervention and the explosion of the number of parastatals tempo- rarily stimulated economic growth. Unfortunately, this state involvement increased the bureaucratization and the inefficiencies of the economy Stability has been an important feature of the RCI political landscape. It was preserved through the autocratic regime of President Houphouet- Boigny, who tolerated no political dissent. Unlike neighboring countries, the RCI lived for decades without a single coup d'etat and there was almost no political violence. This stability built trust and confidence con- ducive to private investment, especially foreign direct investment. Houphouet also controlled personally the financial resources of the CAISTAB and the decision mechanism on the selection of public invest- ment projects. His control over these resources was a source of consider- able economic and political power that he used to consolidate his regime. In 1990, in the depths of the economic and financial crisis, constitu- tional changes created a new prime minister position that brought to the forefront BCEAO Governor Alassane Dramane Ouattara. He ruled the CBte d'Ivoire 389 country during the illness and until the death of President Houphouet- Boigny in December 1993. It is fair to say that he saved Houphouet's political regime from falling apart. The premiership was a significant political innovation because it guaranteed the continuity of power dur- ing the illness of the president. It also had important economic policy implications. The policy dialogue with the Bretton Woods institutions was eased by the positioning of an actor who could be trusted for his technocratic skills and his commitment to reform. To some extent this prevented the collapse of the economy. The 1995 presidential elections ended the transitional period and conferred legitimacy on President Henri Konan Bedie's regime. President Houphouet-Boigny stayed in power for 40 years. This long political tenure had advantages and disadvantages. The advantages were linked to the fact that his policies were deemed appropriate, given the structure of the economy. Agriculture was made a priority activity, peas- ants received substantial shares of world prices, and the public invest- ment financed by agricultural revenues created the basic physical and human capital necessary to support productive activities. These policies were implemented for a sufficiently long period with a strong political will and commitment that gave them credibility. The disadvantage of the long political tenure lay in the fact that the government overextended itself and public investment programs were too ambitious for too long. For decades, President Houphouet, who con- trolled every economic decision, resisted and postponed needed reforms. He was a fierce opponent of devaluation and there is no question that his long illness and death facilitated the decision to devalue the CFAF, reform the CAISTAB, and launch the privatization spree. The first attempt at political democratization was the institution in 1980 of municipal and legislative elections within a one-party system. This did not have much impact on the political and economic evolution of the country because political dissent was still not tolerated. The 1990 democratization was prompted by the reaction of civil servants in 1989 to the cutting of their salaries. This contentious salary cut was a critical feature of the "internal adjustment" effort needed to restore fiscal stabil- ity and competitiveness. Between 1990 and 1993 scores of political par- ties were born, a free press appeared, and President Houphouet had in 1990, for the first time, a challenger in an electoral contest. Since 1990 civil society has been organizing itself to voice its concerns and give expression to political dissent. The large number of political parties, the diversified media, and the professional organizations and independent trade unions together act on the government as agents of restraint, in the sense that they critically watch the behavior of political figures and those in public office. So far, this surveillance has been only partially success- ful in inducing more responsible behavior among government officials. 390 Mixed Reformers The multiple parties, independent media, and numerous professional organizations have had only limited influence on the process of policy formulation. Thus public opinion does not influence the content of policies in deci- sive ways, because consultation by the government is far from being the rule and those who happen to be consulted rarely find their opinions taken into account. Civil society has little political voice. It lacks cohe- siveness. For decades, ethnic and regional cohesion was skillfully pre- served by President Houphouet-Boigny. This cohesion was not furthered by President Bedi6's policies on ethnicity-his concept of ivoirite. The overthrow of Bedie and the popular revolt against military rule late in 1999 were remarkable events, but subsequent ethnic turmoil has para- lyzed the economy and put economic reform on hold. THE CFA PARITY ISSUE C6te d'Ivoire has undertaken extensive macroeconomic and fiscal policy reforms since 1980, as indicated in the previous section. But by far the most important policy issue concerned the exchange rate-the appro- priate relationship between the CFA franc and the French franc. This question dominated policy debate and shaped the economic environ- ment for a decade at least. The parity issue was central both to aid management and to the pace of reform. So it is essential to look carefully at the donor role in the Janu- ary 1994 decision to change the parity of the CFAF from one CFAF equals two French centimes, which had been the prevailing rate since 1948, to one CFAF equals one French centime. In addition to being important, this question has also been contentious. It is often cited as a case where aid retarded reform, because aid inflows allowed the RCI government to postpone the inevitable exchange rate devaluation. This section there- fore focuses on the devaluation decision. The decision to devalue the CFAF obviously did not depend on RCI alone, since the CFAF was the common currency of all member coun- tries of the Union Monetaire Ouest Africaine (since 1994, the Union Monetaire et Economique Ouest Africaine part of the franc zone), and the Union statutes required unanimity in such decisions.6 The French Treasury guarantees the convertibility of the CFAF, and the rules of the monetary arrangement require that the French government be consulted as far as possible about any contemplated change in CFA parity. Moreover, given the economic weight of C6te d'Ivoire in the Union and the influence of France on the other member countries, it was essential that both C6te d'Ivoire and France agree to the devaluation if the other countries were to go along. If C6te d'Ivoire had opposed Cote d'Ivoire 391 devaluation-as did some other member states until the meeting of heads of state in Dakar on January 10-11, 1994, where the decision was made-devaluation would have been impossible. The decision to devalue was very different from other economic policy measures. Its economic and political significance was considerable, which helps explain why it was so difficult a step to take. Once the decision was taken, implementation had to be immediate and irreversible. Also, it had to be taken at the highest political levels. In the case of C6te d'Ivoire this meant that President Houphouet-Boigny, who had been for years an ardent opponent of devaluation, had to be brought around. As things turned out, due to the illness of the president, the decision to devalue was made by his prime minister, who was then Alassane Ouattara. C6te d'Ivoire began its adjustment program in 1980, and the parity issue arose at the outset. But the policy context varied substantially over the ensuing 15 years, and with it the attitudes and roles of aid donors. Three periods can be distinguished: an initial period of donor agree- ment on maintaining parity; a second period when they differed in their recommendations; and a third period when they came together again, in favor of devaluation. Temporary Consensus on Maintaining Parity: 1980-85 The period of temporary consensus on maintaining the existing parity was 1980-85. Cote d'Ivoire's initial adjustment efforts, at the beginning of the 1980s, were undertaken with the support of the Bretton Woods institutions (see section 1). The Ivoirian authorities, the Bretton Woods institutions, and France agreed on the adjustment strategy to be pur- sued, which did not include devaluation. One indicator of this consen- sus is that the adjustment loans given by France through the Caisse Franqaise de Coop6ration Economique explicitly required a previous agreement with the IMF. I As for the consensus between Cote d'Jvoire and the aid donors, one manifestation is the fact that the agreements were successfully implemented. The implementation of a consensus adjustment program was facili- tated by the nature of the monetary cooperation agreement with France, and by favorable conditions in international currency markets. The mon- etary cooperation with France was based on the existence of the opera- tions account maintained by the French Treasury at the Banque Centrale des Etats d'Afrique de l'Ouest (BCEAO), which deposits in this account at least 65 percent of its reserves and can run a deficit without a fixed limit. For the first time since its establishment, the account of the BCEAO was in deficit in 1980, mainly because of the situation of Cote d'Ivoire. The operations account was thus fulfilling the role for which it had been created-that is, to allow franc zone countries to withstand a temporary 392 Mixed Reformers balance of payments disequilibrium without being forced to abandon convertibility or parity. The operations account gave countries time to implement corrective measures and await a return to more favorable external conditions. The crisis then facing C6te d'Ivoire was mainly due to falling prices of coffee and cocoa, which had reached exceptionally high levels in 1976- 77. The temporary nature of this price boom was not taken account of in macroeconomic policy. A rapid rise in civil service salaries occurred along with an even greater expansion of public investment and external debt. The result was a rapid rise in the public sector deficit, which reached 12 percent of GDP in 1982; it had been zero in 1977. This deficit was not sustainable, given that the regulations of the BCEAO imposed strict lim- its on its advances to governments. At the same time, C6te d'Ivoire lost its competitiveness, since the real effective exchange rate rose by 30 per- cent between 1975 and 1980, when it reached its peak (up 24 percent) compared with 1970.8 Because of the budgetary roots of the disequilibrium, and because the trade balance remained positive, the stabilization program implemented with the help of the IMF focused, properly, on reform of public finances. Macroeconomic stabilization would allow, by "competitive disinflation," a depreciation of the real exchange rate. As things turned out, the return to competitiveness was smoothed by the devaluations of the French franc and the appreciation of the dollar. And rises in cocoa and coffee prices in 1985 made the budgetary adjustment easier. The macroeconomic results seemed to validate the policy path adopted. The operations account became positive again in 1985. The public sector deficit was cut beginning in 1983 and turned into a surplus in 1985. The real effective exchange rate depreciated by 34 percent be- tween 1980 and 1985, returning to its 1975 level, while the nominal ef- fective exchange rate only depreciated by about 14 percent. This exchange rate decline allowed initial moves toward reducing protection, which were incorporated in the second SAL (1983). Quanti- tative restrictions and reference prices were eliminated and a tariff re- form was introduced, which was to reduce the effective protection rate to 40 percent. To this was to be added an export subsidy of 40 percent of value added (for further details see Geourjon 1996: 89). This was to be a mock or indirect devaluation. Doubts existed about whether this mea- sure could be implemented-reasonable doubts as it turned out. That it was proposed suggests that a problem of parity was seen to exist. Overall, in 1985, C6te d'Ivoire seemed to have returned to a positive growth path after a period of recession. One noteworthy fact during this period is that the real prices paid to producers of coffee and especially of cocoa were maintained, even increased, which brought about a realloca- tion of income in favor of rural households. CMte d'Ivoire 393 The donor community seemed at the time satisfied with these out- comes. Senior officials of the World Bank even cited the adjustment ex- perience of C6te d'Ivoire up to 1983 as a success story (see, for example, Serageldin 1988). Breakdown of Consensus: 1986-92 Beginning in 1986 the economic environment turned sharply negative for C6te d'Ivoire. Several external factors contributed, as noted in the previous chapter. The terms of trade became once again unfavorable; and as it turned out this deterioration was the deepest and most durable the country had ever known.9 The French franc strengthened, the dollar fell, and currencies of big neighboring countries (Ghana and Nigeria, in particular) were devalued. Confronted with this change in economic conditions the position of the World Bank, and then of the IMF, gradually shifted in favor of devalua- tion. No official position was announced publicly, but in private, and some- times even in public, World Bank staff expressed their doubts about the prospects of this new phase of adjustment without devaluation.'" These publicly expressed opinions were regarded by French officials as untimely and imprudent, certain to exacerbate the loss of confidence and capital flight already observable among Ivoirian economic actors. During this time, the French authorities never missed an occasion to emphasize their attachment, and that of African governments, to parity maintenance. This was the message given especially during meetings of finance ministers of the franc zone. And, in private, French spokesmen let World Bank management know of their displeasure every time they felt that Bank staff committed an incartade-an ill-considered statement or action that was damaging to their official doctrine. These incidents influenced how the devaluation was prepared: French experts worked mainly with those of the IMF. The Bank, which had done so much to bring about acceptance of the idea that devaluation was necessary, was not much involved in preparing the decision. This discord between the main donors made it very difficult for a long time to engage in quiet, scientific analysis of the costs and benefits of a devaluation for a country like C6te d'lvoire. Political factors seemed most fundamental in determining the positions adopted. The French government was concerned about the social and political turmoil that might accompany devaluation in countries where France had a military presence. French authorities feared that they would lose the political confidence of African governments and African elites, who would be directly and negatively affected. They feared that the lack of consensus among the countries concerned might lead to a breakup of the franc zone. Indeed many suspected that some partisans of devaluation in the 394 Mixed Reformers international institutions were using the devaluation issue to undermine the institutions of the franc zone. The French government, finally, was sensitive to the anti-devaluation arguments of some groups of French companies active in Africa, although other enterprises were favorable to parity change. It is important to recall that on the French side, as early as the end of the 1980s, policy positions were neither uniform nor rigid. Policy dis- cussions were held, secretly, in both the Ministry of Cooperation and the Ministry of Finance. Devaluation had its proponents in various corners of the administration. But none of this was observable; the official posi- tion remained firm and clear. Moreover, French academics said little on the subject and rarely fa- vored devaluation in their writing. In part this was because some of them remained skeptical about the efficacy of adjustment via changes in prices. In part it was because they shared the official concern about disruptive sociopolitical fallout from inflation. Some also continued to believe in the possible success of "competitive disinflation" (or internal adjustment), by which price stability coupled with productivity increases would re- establish equilibrium. Many, finally, convinced about or resigned to the need to devalue, worried about the effects on speculation and capital flight that might flow from pro-devaluation public statements. Thus the advice they might have given on the likely consequences of a devalua- tion remained unpublished. Within the World Bank there were also perhaps noneconomic factors underlying their stance. For one thing, Bank staff were not comfortable with the franc zone arrangements and the policies followed by member countries. Many Bank staff members were skeptical about their impact on growth. They believed that the franc zone monetary arrangements lessened adaptability, mainly by removing the exchange rate instrument from the policy arsenal, and that their instruments of monetary and fis- cal discipline were often ineffective. They also considered that member countries were exposed to different, not necessarily correlated, external shocks. Finally, they were concerned that C6te d'Ivoire's heavy indebt- edness in relation to the Bretton Woods institutions had become unsus- tainable in the absence of IDA credits. With respect to Cote d'Ivoire, urban opinion, that of wage earners, was against devaluation. Intellectuals had various, often nuanced, views about the devaluation, and about the franc zone regime. The authorities, until 1992, were strongly opposed to devaluation. Their stance reflected the views of President Houphouet-Boigny, who believed that the de- clines in cocoa and coffee prices on world markets sprang from a funda- mental injustice imposed on his country by the international economic system. Himself a planter, he attached great value to the maintenance of prices paid to producers. He also was a firm believer in maintaining Cote d'Ivoire 395 parity of the CFAF. These two policies were to him symbols of the suc- cess of Ivoirian development, not to be abandoned. Given the lack of adequate reserves at the Caisse de Stabilisation, the simultaneous maintenance of coffee and cocoa prices proved impossible. The grave financial crisis that occurred in 1987-88 led to a break be- tween the RCI and the Bretton Woods institutions. The two agreements signed with the IMF in 1987 and 1988 were cancelled, and World Bank lending was suspended until producer prices for the two commodities were reduced. In 1989 President Houphouet-Boigny tried to influence world prices by keeping crops off the market. This did not work. The president had to admit that he had lost the "cocoa battle," and he had to cut producer prices by half. From his perspective, devaluation of the CFAF would have then marked the total victory of international market forces over C6te d'lvoire. Despite these conditions, the IMF and the World Bank continued to provide help for stabilization and structural adjustment programs with- out devaluation. But the implementation of these programs proved very difficult. The IMF loan that emerged from a 1989 agreement was not fully disbursed, nor was the loan attached to a second agreement in 1991. World Bank sectoral loans in 1989-90 were fully disbursed, but not the adjustment loans granted in 1991. The 1989 Fund program envisaged a vigorous effort to put public finances in order, generating positive budget savings. The Ivoirian gov- ernment tried in 1990 to impose substantial cuts in civil service salaries. But the announcement of these measures gave rise to serious civil pro- tests, which forced the government to give up on direct salary reduc- tions. Another approach, introduced by Ouattara-first as president of the Interministerial Coordinating Committee and later (after illness dis- abled President Houphouet-Boigny) as prime minister-resorted to other ways of reducing public expenditure. The social disruption-including violence-that occurred in 1990 when the government tried to cut salaries served to reinforce doubts, notably in the French aid community, about the right policy to follow. A growing number of observers doubted that factor costs in Cote d'Ivoire could be lowered without devaluation, given the rigidity of nominal wages. At the same time, the violent reaction to the attempt to lower the already low standards of living in urban areas reinforced fears about the political consequences of devaluation. The conditionality attached to adjustment loans then became ambigu- ous, as, for example, in the PASCO competitiveness loan granted in 1991. Its purpose was to support a series of reforms aimed at making the Ivoirian economy more competitive, according to the logic of adjust- ment without devaluation: priority measures dealt with raising produc- tivity. At the same time, it carried as a second-tranche disbursement 396 Mixed Reformers condition the lowering of the real effective exchange rate to its 1985 level, as calculated by the Fund. However, the real depreciation level thus pro- grammed was such (about 25 percent) that it was not realistically pos- sible without devaluation. In addition, the way the exchange rate was calculated was disputable and risky because it gave too much weight to some primary product exporters that were in situations of extreme mon- etary disequilibrium, such as Brazil. Consequently, it became clear that the Fund would not disburse the subsequent tranches, whatever the magnitude of the reforms under- taken elsewhere. The very useful measures agreed to in the sectoral loans were thus diminished in importance. As was predictable, the Bank and the Fund suspended aid to the RCI in 1992, their actions inducing similar responses by other donors. France, on the other hand, contin- ued its adjustment aid (although suspending other aid from the AFD because of payment arrears). However, French adjustment aid was used by C6te d'lvoire to service its debt to the World Bank.11 During these years foreign aid defined as net transfers decreased significantly (see figure 7.6). Did the discordant points of view of donors affect the pace of reform during this period? Looking at the devaluation reform alone, it is clear that the French position postponed its implementation. On the other hand, the decision not to devalue pushed the Ivoirian authorities to introduce, in cooperation with the donors, productivity-raising reforms and reforms in public finance. These were in place and helpful after the devaluation. Examples are the dismantling of quantitative restrictions on imports, the liberalization of most domestic prices, and labor market and judicial system reforms (discussed further below). Even regional integration was advanced by the pre-devaluation mea- sures-those aimed, for example, at strengthening the monetary union (formation of a multinational banking commission) and at transforming it into an economic union (multilateral monitoring of public finance and gradual moves toward a common market). These changes could be in- stitutionalized when the devaluation took place. Despite the desirability of these (nondevaluation) reforms, it should be noted that their implementation was surely slowed down when the Fund and the Bank suspended aid after 1992 because of nonperformance on macroeconomic conditionality The suspension reduced pressures to implement these reforms. The Decision to Devalue: 1993-94 A whole set of factors came together to bring about the Ivoirian decision to devalue. The analytic and empirical work undertaken in the early 1990s played only a minor role. The various estimates of the apprecia- Cate d'Ivoire 397 tion of the real exchange rate were so sensitive to choice of base year and to weighting that they were unconvincing to many decisionmakers. In the C6te d'Jvoire case in particular, the evolution of real and nominal effective exchange rates, as published by the IMF, were highly contest- able, since they were so dependent on results of the chaotic evolution of other countries."2 In addition, the overvaluation of the CFAF arose less from changes in the real exchange rate than from deterioration of the terms of trade, which in 1992 was in its sixth consecutive year, and from the rising debt burden, which was particularly heavy between 1987 and 1991. Directly and via the budget, these two factors contributed to the pro- nounced decline of per capita income. Moreover, devaluation rumors forced the Central Bank to maintain a high real rate of interest, which discouraged investment. As for the many estimates of the degree of overvaluation, they suf- fered from the same uncertainties as those for real exchange rates, and also from hypotheses about the acceptable size of the current account deficit that were politically based and difficult to justify. These calcula- tions did help give credence to the idea that the parity had become un- sustainable. But the decision to devalue was not based directly on these various calculations purporting to estimate the extent of the country's lack of competitiveness. The decisive factor that led the Ivoirian prime minister and the French authorities to support devaluation was the government's financial con- dition. Experience had shown the enormous difficulty of improving public finances in a stagnant economic environment. No matter how much effort C6te d'Ivoire made, it could not do without international aid. Confronted with the suspension of adjustment lending by the Bretton Woods institutions, France changed its policy Whereas French adjust- ment loans formerly were made only after the aided country had signed an accord with the IMF, in 1991 and 1992 France continued its support, despite the lack of an Ivoirian agreement with the Fund. C6te d'Ivoire was in effect given time to renew its dialogue with its external partners. The French clearly wished to avoid a rupture of relations of franc zone countries with the international economic community. The decision was prepared, with great discretion, by a small group of civil servants. The political authorities mostly supported the change by this time, although their official posture in defense of the existing parity remained unchanged. The shift in the position of the French gov- ernment was brought about in part by the power of persuasion of the responsible civil servants and in part by the deteriorating situation within the franc zone and the resulting financial costs being incurred by France. There were also other factors-a complex mixture of a desire to transform French policy toward Africa, fatigue and resignation, and 398 Mixed Reformers in some cases a sentiment that the devaluation was a kind of deserved punishment for past errors. In September 1993 Edouard Balladur, then French prime minister, sent a letter to the heads of state of franc zone countries stating that begin- ning in 1994 France would revert to its former policy on adjustment lend- ing: that is, it would make no such loans in the absence of an agreement with the IMF. This could be interpreted as a signal that France was ready to accept CFAF devaluation. The Ivoirian prime minister, Alassane Ouattara, who had been director of the Africa department of the Fund, and who had lived through the earlier experience with the failed "inter- nal adjustment" effort, was also on board. His agreement was encour- aged by the indications that aid donors would be ready to recommence lending at high levels after devaluation. Ouattara's adherence was officially announced during the meeting of the franc zone heads of state in Dakar in January 1994. It helped win the agreement of all the other states, some of which had been strongly op- posed to the change in parity. C6te d'Ivoire used the months before the devaluation to win over other countries of the zone to the devaluation cause, in particular Burkina Faso. The Ivoirian authorities also used the time to prepare an economic policy program to accompany the devalu- ation. In 1994, net transfers from multilateral institutions to the franc zone countries, especially C6te d'Ivoire, and from bilateral donors as well, reached exceptionally high levels. But it was a transitory gain from the devaluation. As noted in section 2, these flows sharply declined in the following years, with net transfers from multilateral institutions com- ing close to zero in 1995 and 1996 and becoming negative in 1997 (see figure 7.6). Three Major Hypotheses C6te d'Ivoire's choice first not to devalue, then to devalue, can be looked at in light of the three major hypotheses presented in the terms of refer- ence. First, did the Ivoirian government decide against devaluation, then in its favor, independently of aid relationships? The answer is clearly no. It was external assistance that allowed retention of the existing parity until 1994, and it was the suspension of aid that provoked the devalua- tion. The data given in section 2 illustrate the precipitous fall in financial inflows on the eve of the devaluation (notably in 1993), and their rise afterward. Nonetheless, another question has to be asked. In allowing resistance to devaluation and in provoking it later, did aid influence the pace of other reforms? If we consider the first hypothesis from this perspective the response is more complicated. As we saw, some reforms were adopted C6te d'Ivoire 399 because of the political option taken to adjust without devaluation, and because this political posture was credible and supported by external donors. But it was no longer credible when these reforms were intro- duced in a particularly difficult context of bad relations with donors, marked by several suspensions of aid. It is difficult to say, in these cir- cumstances, what was cause and effect in each instance of reluctance to reform and withdrawal of aid. Consider next the second hypothesis. Did nonfinancial aid have a stronger impact than conditioned financial aid on the decision to de- value? The devaluation decision was made at the highest political lev- els. It surely could have been influenced by economic ideas, analysis, or the simulations done in Washington or in France on the anticipated ef- fects in the RCI. But given the nature of the decision, its close link to fiscal imperatives, and its eminently political dimension, financial aid surely played the essential role. In other areas (incentive structures, ju- dicial system, and so forth) studies and technical assistance had an im- portant role, but these were for the most part packaged with financial assistance-notably in the framework of structural adjustment loans. What about the third hypothesis-that financial aid was effective only when reforms were underway? In the present context this means after the devaluation, but not before. This question has meaning only for the periods when aid was not suspended. Certainly economic growth was more rapid after the devaluation than before, but this could be the com- bined result of the devaluation, the improved terms of trade after the second half of 1993, and the recovery of aid inflow. It is difficult to attack this question. Multiple elements of economic policy determine the efficiency of aid, not just the devaluation. And nonpolicy factors, notably the external environment, condition aid ef- fectiveness. Before the devaluation, the decline in the terms of trade prob- ably raised the marginal efficiency of an aid dollar in the C6te d'Ivoire, given the strong financial constraints on growth prevailing at that time. TRADE LIBERALIZATION Evolution of Trade Policy Initiation of Industrial Protection: 1960-73 At independence in 1960, Cote d'Ivoire had no industrial base (Cote d'Ivoire 1989). The colonial power neglected the interests of the private sector. The incentive system in place after 1956 favored cheap imports of basic foodstuffs for Europeans and discouraged local transformation of products by imposing high tariffs on raw materials, semifinished inputs, 400 Mixed Reformers and equipment. In 1959 the RCI government put in place an investment code (law 59-134, September 3, 1959) with a generous incentive system aimed at attracting foreign capital on a selective basis. If granted prior- ity status, investors were exempt from payment of duties on exports and on imports of raw materials entering into exports of finished goods. In addition to these exemptions, priority firms could benefit from a sta- bilized fiscal regime and an establishment convention for 25 years that granted exoneration from import taxes on raw materials. Quantitative restrictions were used as an additional instrument to help the installation of industrial firms. They included bans, quotas, and im- port licensing. Products subject to import licensing had to be screened by a special committee on import control. The range of government interventions included a complex system of legal price controls."3 In case of "emergencies" the government could (and did) intervene to block prices of any good or service. Prices were blocked in 1967 when the value added tax increased, in 1968 when the minimum wage increased, and in 1969 because the French franc was devalued. Thus, 1966-73 was a period marked by price controls and intermittent imposition of general price ceilings. Reinforcement of Industrial Protection: 1973-84 Industrial policy was reinforced along two lines: implementation of a strategy of industrial development based on the transformation of local and imported raw materials and protection of the industrial base built on the incentive system of the investment code. A reform was imple- mented in 1973 that included incentives favoring the import of a wide range of machines and production equipment, incentives to transform timber by taxing its export in raw form, and high import taxes and pro- hibitions that were used as incentives to promote textile industries. New duties were introduced on exports and imports; the export tax was spe- cifically designed to stimulate domestic transformation. The use of quantitative restrictions, common during the first episode, was extended after 1973. The instruments used were prohibitions, quo- tas, and import licensing. The number of products subject to import li- censing increased from 86 in 1973 to 310 in 1976 and to 427 in 1982. As a result of these protective measures, the rate of nominal protection in- creased from 30 percent in 1971 to 32 percent in 1978. The rate of effec- tive protection increased sharply from 39 percent to 76 percent in the same period. Rates of implicit protection were even higher, though they declined during the period (Bohoun 1996). The government intervened directly in productive activities by invest- ing in state-owned enterprises (SOEs) in order to process agricultural raw materials. The creation of SOEs also was intended to increase local control Cote d'Ivoire 401 over industrial activities dominated by French capital. More than 250 SOEs out of 400 listed in 1981 were formed between 1971 and 1980.14 Trade Liberalization: 1984-88 In 1984, the investment code and the trade tariff were reformed to re- duce the level of industrial protection and to promote industrial exports. The investment code eliminated the exoneration of taxes on imported raw materials granted to priority firms. It also favored location outside of the Abidjan area and facilitated the emergence of small and medium- scale enterprises. The objective was to reduce the rate of nominal protec- tion to a maximum of 40 percent for finished products and 25 percent for raw materials, and the rate of effective protection to a maximum of 40 percent. Thus, low tariffs on equipment and raw materials offered better protection to industry while selectively penalizing the competing foreign products. Quantitative restrictions were eliminated for products subject to im- port licensing. In compensation, an import surtax and an entry fiscal surcharge were imposed. An export subsidy was put in place in 1984 to compensate for the overvaluation of the CFAF and to improve the com- petitiveness of firms. Despite the general mood favorable to trade liberalization, some in- creases of tariffs and fiscal taxes on certain products occurred. In 1987 a general tariff increase was imposed. In 1988 the payment of the export subsidy was discontinued. New waves of exonerations and grants of priority status to industrial firms marked the years 1985 to 1988; 54 firms were granted priority status, 30 of them in 1986 alone (Bohoun 1996). Reversal of Protection: 1988-90 During 1988-90, the deterioration of government revenues prompted a policy shift toward increased protection. In 1989 tariff rates on im- ports and exports increased from 5 to 10 percent and a new tax (statis- tical tax) of 2 percent was created. It was estimated that the average tariff rate by 1990 had increased up to its 1983 level (Contamin 1997). Nontariff barriers were reinforced through import controls with the following modalities: creation of a customs value committee, quantita- tive and qualitative inspection with price comparison, and some pro- hibitions and import quotas. Reversal of Liberalization: 1990-93 Starting in 1991, four Sectoral Adjustment Loans (SECALS) were imple- mented, addressing competitiveness (PASCO), human resources de- 402 Mixed Reformers velopment (HRDP), the financial sector (PASFI), and economic man- agement (PAGE). All to some extent were aimed at increasing indus- trial competitiveness, though PASCO was most directly targeted that way. The others aimed at improvements in the general institutional environment. New waves of selective trade liberalization occurred. Some tariffs and taxes were eliminated, and others were reduced (those on agricultural and industrial inputs and also on finished products, textiles in particular). Devaluation and Deepening of Trade Liberalization The devaluation of the CFAF in 1994 improved significantly the rate of effective protection of the economy This helped the reduction of nominal protection also needed to limit the inflation generated by devaluation. In 1995, on the basis of 350 tariff positions, the weighted average tariff rate decreased to 29 percent from 37 percent in 1990 (Contamin 1997). Export promotion was a prominent item in the reform initiated by the devaluation package, with the private sector playing the main role in the new export promotion agency, APEXCI. Critical reforms were also implemented in the agricultural sector. The role of CAISTAB in internal and external marketing of cocoa and coffee was drastically reduced to increase private sector participation and improve the transparency of the external marketing system. The significant change was an auction system for export rights. Some critics claim that this reform gave rise to unanticipated problems involving overbidding and concentration of export rights on a few exporters. The rice marketing system was also liberalized after a difficult dialogue about imports of broken rice. (Many in government feared that unfettered import of broken rice would un- dermine the domestic rice sector.) Nontariff barriers were eliminated, prices of agricultural products and equipment were liberalized, and ef- fective protection of local production had to be limited to 40 percent. State-owned enterprises in oil palm (PALMINDUSTRIE), sugar (SODESUCRE), and cotton (CIDT) were privatized. Local opinion acknowledges that trade and price liberalization has been effective, but with some reservations. The playing field is not even; there is some discriminatory and unequal treatment linked to political and family connections. Also there are cases where insufficient informa- tion about decontrolled prices subjects small private actors in trading activities to the corrupt practices of civil servants. The business commu- nity is favorable to liberalization with equal treatment for everybody. To them tariff reduction appears to be a successful reform area, probably because it was well prepared. But this opinion is not shared by selected influential nonbusiness representatives. The private sector welcomes less government intervention but only a few firms interviewed in a recent survey favored zero public intervention. At the time of the survey (May- Cote d'Ivoire 403 June 1998) spokespersons for 31 percent of the firms questioned thought that liberalization had had a positive impact, 55 percent reported no impact, and only a few (3 percent) indicated a negative impact. The trade and industrial policy reforms were implemented gradu- ally. The building of the industrial base took place incrementally. The investment code was created in 1959 while the tariff reform was imple- mented in 1973. These two policy instruments were relied on for many years. They were replaced by a new investment code and a new tariff regime in 1984. The first set of moves from protectionist policies to liberalization took many years-from the 1960s to 1984. Implementation was gradual be- cause the Ivoirian government was reluctant and resisted the reforms. The government feared a hemorrhage of customs receipts and was pre- occupied with finding ways to compensate for the losses of trade-re- lated revenues. There was also the reluctance to make doctrinal jumps, from a controlled economy to a liberalized one, and concern about the costs that might arise. From the field interviews it appears that implementation is a major problem on the government's end, especially for recent liberalization reforms of cocoa and coffee marketing and CAISTAB. These reforms are widely thought to have been brutal, that is introduced too fast. It will, according to some, leave an unorganized rural society facing a volatile world market and oligopsonistic purchasers/exporters. Peasants have not been properly prepared, in this view, and farmer organizations are weak. As will be noted in the concluding section, the experience with coffee marketing liberalization has not confirmed these fears. Scheduled reform measures were not always taken. In particular, the 1984 liberalization reform featured a simulated devaluation of the CFAF by combining an increase of import tariff and an export subsidy, with the expectation that the first measure would generate sufficient resources to finance the subsidy. The payment of the export subsidy was halted for lack of resources. Overall, the simulated devaluation was credible neither to the Ivoirian administration nor to the private sector. A second reason for this failure had to do with the institutional base: the administration of tax collection in general and the customs administration in particular was weak and of limited efficiency. Even the World Bank management thought that this measure was hastily prepared by its own staff. The Role of Aid Trade liberalization was not adopted in reaction to political demands by political elites or local interest groups. The prevailing political sentiment among decisionmakers was to resist these reforms. They were adopted under the pressures of a severe financial crisis caused by a decline of 404 Mixed Reformers world cocoa and coffee prices and of world demand for exports, losses of foreign exchange and fiscal revenues, and heavy debt burden. Two types of policy reversals were observed that were also caused by the financial crisis. The first type, reversal to protection, was explained by the losses of foreign exchange and of fiscal revenues due to the fall of prices of cocoa and coffee. This policy reversal highlighted the primacy of short-run financial problems over structural ones; it was adopted under the pressure of the IMF to the dismay of the World Bank, an example of possible conflicts between these two institutions. The reversal to liberalization, in contrast, was accounted for by the primacy of external competitiveness concerns over worry about short- term revenue losses. These losses were duly recognized as the price to pay to improve income in the medium term because there was no alter- native. But selectivity was the rule; no across-the-board tariff reduction was implemented. Many in government find it difficult to acknowledge that a reform idea is not domestic. The debate over who initiated the policy dialogue is ambiguous and can best be summarized by this phrase, gathered from field interviews: the initiative comes from the government but the gov- ernment has advisers; indeed the IMF and the World Bank are advisers to the government. It is clear that trade liberalization was initiated by the World Bank. It was generally opposed by the administration and by a business com- munity used to the shelter of protection and fearful of external competi- tion. At the time of the reform and beyond, the administration was weakened by successive institutional changes and by the departure of highly qualified senior civil servants for positions in the private sector after the merger in 1977 of the Ministry of Planning with the Ministry of Economy and Finance. A weakened administration could not be expected to generate a homegrown reform package. Recently, the World Bank also initiated the reforms of cocoa and coffee marketing and the restructur- ing of CAISTAB. The internal dialogue within government about trade liberalization was also difficult as DCGTx (Direction de Control des Grands Travaux), a French-managed parallel administration, overpowered the technical ministries (Ouayogode and Pegatienan 1994). Between 1986 and 1991, DCGTx was formally responsible for economic policies and programs, including the adjustment programs. This agency was generally suspi- cious of trade liberalization, reflecting French dirigiste traditions (and the sentiments of President Houphouet) in this matter. Parallel markets did exist during episodes of liberalization and policy reversals. Traders, especially women, made frequent trips across the Ghana border and to Nigeria, where prices of the traded goods were competitive. This informal trade in cheap textiles and other manufac- tured goods harmed the interests of enterprises in the formal sector. But Cote d'Ivoire 405 the complaints from formal business organizations did not, apparently, have much impact on the reforms because of the limited dialogue be- tween the administration and the private sector. In matters of industrial and trade policies, the main actors in the policy dialogue were the government and the World Bank. The World Bank had difficulty convincing President Houphouet-Boigny, the sole decisionmaker in the government, of the necessity to liberalize. He thought that the RCI was sufficiently open and did not need to further open up to the world market. His technical advisers were amused by the effective protection and domestic resource cost calculations, and the con- clusions and policy recommendations derived therefrom. They believed them to be too academic to be convincing to either officials or business- men. This basic skepticism helps explain why the simulated devalua- tion of 1985-86 lacked credibility and failed. Since the early 1990s, the influence on economic policymaking of ex- ternal donors in general and of the Bretton Woods institutions in par- ticular increased and widened because of the deep financial crisis, the illness and death of President Houphouet-Boigny, and the weakness of the present political leadership. The two policy reversals and the deep- ened post-devaluation liberalization are the net results of interactions between these influences and government's reactions to them and to its own constraints. The private sector did not play a significant role in the policy dia- logue before the privatization program, but it had sufficient influence to cause, in part, the failure of the 1985-86 simulated devaluation. Since the 1994 devaluation and the acceleration of the privatization program, the private sector carries more weight. One example of that is its pres- sure to eliminate the public export promotion agency (CCIA) and to substitute the privately managed APEXCI. Nonetheless, even in this era of private sector and public sector partnership, government con- sultation with the private sector, let alone other elements in civil soci- ety, is not the norm. Since 1990, after President Houphouet-Boigny was forced by social and political unrest to open up politically, civil society has been free to voice any concerns, opinions, and dissenting views about government policies. However, the use of these new opportunities by a large number of professional organizations and independent media has not yet proven to have critically influenced the process of policy formulation. EDUCATION AND HEALTH SECTOR REFORMS Until the 1980s, reform in the social sectors was not a live issue. The central concern was expansion: providing more schooling, more health care. This reflected in part the buoyant economy of the two post- 406 Mixed Reformers independence decades, which generated growth in revenues. It also re- flected President Houphouet's awareness of education's role in economic growth and his view that a well-educated Ivoirian elite was essential to reduce dependence on expatriate skills. In this period, donors financed mainly infrastructure investments, with some effort at quality improvement; France assisted by providing thou- sands of teachers. Thus the World Bank financed only three education projects prior to 1980, for teacher training and vocational education fa- cilities. The Bank financed no health projects until 1986 and no addi- tional education (or health) project until the 1991 Human Resources Development Project. By the mid-1980s the education system clearly exhibited the pattern of problems common in Africa, especially francophone Africa: high unit and aggregate costs, weak budget and personnel management, internal inefficiency (distorted budgets and high dropout or repeater rates), ex- ternal inefficiency (limited relevance and connection to the job market), and inequitable allocation of benefits. The health system shared many of these characteristics-notably weak management, inefficient allocation of resources, and concentration of benefits on the relatively well-off. The Ivoirian problem is particularly severe. The RCI government spent (and spends) relatively more on education and health than most other countries; the social sector took more than half of budget outlays in the 1980s, and its share remained high at the end of the 1990s. Though social indicators improved, they remain uninspiring, and are below indicators in other countries at the same income level. Even in 1996 only half the primary school-age children were in school and only one-third of the population had access to health care. By some measures, even the urban health care system is less effective than it is in the much poorer Sahelian countries.'5 These problems have absorbed the attention of reformers for more than a decade. They are considered here in two parts. First we describe briefly the main reforms and comment on their effectiveness. Education and health are treated separately. In the second part we analyze process issues and the role of aid. The main emphasis is on the dimension of public expenditure reform. Reforms and Their Implementation A major theme in education reform has been the imperative need to cut costs and in particular salary costs. Government had to pay high salaries to recruit the large number of teachers required by the expan- sion of the education system in the 1960s and 1970s, and to replace during the ensuing years the numerous expatriates present in the sec- tor.16 The result was that by the mid-1980s the cost of a high school C6te d'Ivoire 407 teacher was higher in Abidjan than in Madrid. In 1993 a primary school teacher was paid 13 times the country's average per capita GDP, five times the ratio in East Asia and 10 times that in industrial countries (World Bank 1999a: 73).7 Even after some cutting, in the early 1990s education accounted for half the total wage bill, which in tum claimed 60 percent of government revenues. Any policy to contain government expenditures (a key objec- tive of macroeconomic reform) thus had to confront the issue of teacher salaries. It is no surprise that the education reform program was domi- nated by measures for cost-cutting, though in both health and educa- tion, resource reallocation and management strengthening were also priorities.18 An early measure was the reduction of salaries for new teachers. Teacher salary scales were reattached to the civil service scales, from which they had earlier been detached. (The premium over the civil ser- vice salary scale was 40 percent.) Salary rates were to be frozen at 1991 levels and the aggregate salary bill for teachers was to be limited to a 1 percent a year annual rise for the period of the HRDP (1991-94). The limited rise in the salary bill was to be accomplished by relying more heavily on teaching assistants, reducing the number and cost of French TA providers, increasing workloads of teachers, and redeploying administrators to classrooms. Also, growth in numbers of secondary and university students and teachers was to be controlled. Transfer payments through the university support center and fellowships and transport subsidies were to be reduced and boarding facilities in Abidjan cut back. Numerous other measures were introduced in both sectors-cost re- covery, freezes on new construction, and introduction of generics, for example. But heavy emphasis was given to two sets of reforms of public expenditures, one substantive and the other procedural. The substan- tive reforms concerned raising the level of expenditures on primary edu- cation and primary health care and increasing the primary subsectoral shares of total sectoral spending. In education the primary enrollment rate was to be increased; in health, the share going to preventive, rural activities was to be increased. The procedural reforms aimed at stronger management, especially financial and personnel management. To strengthen the budget process and also monitor implementation of reforms, annual programs and bud- gets were prepared, which brought together in a transparent way indi- cators of past budget performance, and established annual targets based on HRDP objectives. A lengthy set of budget performance indicators was established. This was to supplement the parallel general budget reforms being undertaken at the Ministry of Finance and the Ministry of Plan- ning, such as nomenclature changes, some of them conceived with social sector budgeting deficiencies in mind. 408 Mixed Reformers Many of the individual reforms that were part of HRDP conditional- ity were implemented (World Bank 1995). For example, some fringe ben- efits were cut and salary rate increases were contained; expatriate TA teachers virtually disappeared; 1,100 assistant teachers were hired in 1996; aggregate transfer payments (scholarships, food, and transport subsidies) were reduced; 2,500 administrators were redeployed to class- room teaching at the primary level; locally financed construction costs in both education and health were put on hold; and sector programs were elaborated for both health and education. Health sector changes included the extension of cost recovery for medicines and consultations, encouragement of the use of generic drugs, and a start in defining a decentralized sector strategy based on primary care. (Most of the rec- ommendations made in a joint donor health sector program in 1990 were also implemented.)19 Some of the HRDP spending targets were achieved.20 In health, the ratio of nonsalary operating expenditures (mainly medicines) to salaries rose, from 39 to 45 percent between 1991 and 1995, and this trend con- tinues.2" Some cost recovery took place, though well below the program targets. And though the primary subsector's share in total health spend- ing did not rise during the HRDP period (1991-95), it did increase (by 16 percent) in 1996 and 1997. The tertiary subsector, however, grew by twice as much. Expenditure reforms also are visible in the education sector. Despite a rise of 10 percent in primary enrollments and a soaring university popu- lation (it rose from 26,000 in 1991 to 58,000 in 1995), the total education wage bill stayed constant between 1991 and 1994. The unit cost of pri- mary school teachers fell from 13 times per capita GDP in 1993 to 9 times by 1995. Subsidies to university students (housing, scholarships, trans- port, and so on) were cut by 20 percent in real terms and by half in terms of budget share between 1992 and 1996. At the same time, nonsalary operating costs as a percentage of the overall education budget rose from 4 percent in 1992 to 13 percent in 1996. Unit education subsidies have become more efficient and equitable in RCI than in most other African countries.22 Investment in education as a share of total public investment more than doubled to an average of 7 percent between 1994 and 1997. In both sectors, however, achievements have fallen far short of objectives: * Social sector budget sharesfell. Education spending as a share of the total budget fell from 34 percent in 1992 to 27 percent in 1996. The health budget as a share of current expenditure fell from 7.4 per- cent to 6.7 percent between 1991-93 and 1994-97. * Social sector investment did poorly. Investment in health declined in the same period from 7.7 percent of total public investment to un- Cote d'Ivoire 409 der 5 percent (World Bank 1999b, vol. 2: 27-28). More than 40 per- cent of the investment during 1992-94 went to administration; the share of the primary subsector fell from 17 to 14 percent during those years. A shift occurred in 1996 and 1997, but the primary subsector share of the 1997-99 PIP was only 20 percent compared with 26 percent for the tertiary.23 Investment in education as a share of the total investment budget increased only slightly between 1991- 93 and 1994-7, from 9 to 10 percent. * Reallocation in favor of primary education and primary health care did not occur. Instead of rising, primary school expenditures fell from 52 percent of total education expenditures in 1991 to 47 percent in 1995. The trend continued at least through 1997. In health the pri- mary subsector share of recurrent expenditures was supposed to rise from 35 percent in 1991 to 42 percent in 1995. Instead, it de- clined to 32 percent in 1995.24 The gross enrollment rate in primary education was supposed to rise from 74 percent in 1991 to 77 per- cent in 1995. Instead it fell to 67 percent in 1994, from which it increased little by 1997. * The benefits ofpublic spending on health and education remain unequally distributed. The primary school gross enrollment rate is only 51 per- cent among the poorest 20 percent of households, while it is 99 percent among the richest 20 percent. Only 12 percent of eligible children in the poorest quintile of households attend secondary school, compared with 65 percent for the best-off 20 percent. Al- though 19 percent of the primary school subsidy goes to the poor- est 20 percent (and 14 percent to the best-off), only 7 percent of the secondary school subsidy goes to the poorest quintile (37 percent to the richest). Only 12 percent of the tertiary subsidy goes to the poorest 20 percent, while 71 percent goes to the top quintile. Of the total education subsidy, the poorest quintile gets 13 percent, the richest 35 percent.25 * Internal efficiency of the school system improved little. The share of nonsalary recurrent costs in education and health budgets in- creased. In principle, this should be an indicator of greater effi- ciency; it would indicate that teachers and medical staff have more supplies to work with. In fact, however, the post-devaluation in- flation has to be taken into account. Prices of imported goods rose sharply, and of import competing goods also; the consumer price index rose by 50 percent between 1993 and 1997. Deflated figures for nonsalary operating costs between 1994 and 1996 indicate that these fell by 50 percent in real terms at the primary level, 38 per- cent at the secondary level, and 20 percent at the university level. Moreover, the basic indicator of internal efficiency, repetition rates, shows that these have increased in most levels of the primary 410 Mixed Reformers subsector and at many levels within the secondary school subsector between 1991 and 1995 (World Bank 1996). * External efficiency probably worsened. The number of students in the public sector technical and vocational training schools declined dur- ing much of the 1990s, while private participation grew. The pub- lic sector had only 10,000 students, 5 percent of all secondary students, and less than the 15,000 in private professional schools in 1995. The subsector suffers from recruitment of uncommitted stu- dents. (Technical and professional training is a second choice for pupils, or a step to general secondary and then to university.) The subsector trains for nonexistent formal sector jobs. It is heavily underutilized, hence even more expensive than is standard for tech- nical/professional systems. In 1993 there were five students per teacher. * Strengthening of sectoral management was uneven. The reform pro- grams of the 1990s contained important components aimed at stronger social sector administrative capacity. But budgets were poorly prepared and defended, and project development for in- vestment, rehabilitation, and maintenance was especially weak. Inputs were the sole or main preoccupation; little monitoring or evaluation of outputs occurred. Staffing was poorly monitored, with inconsistencies between ministry payrolls and civil service rosters. Some improvement has occurred, particularly in education. Officials interviewed say that personnel management is surer and budget prepa- ration and implementation significantly better. There is some evidence for the latter. Implementation rates are much higher now than in the early 1990s, both for the overall budget and the Budget Special d'Investisse- ment et Equippement (BSIE). The volume of unpaid engagements (depenses engagees non-ordonnances) is much lower than it was until 1995, and the resulting unpaid bills problem is minimal. That is a strong indicator of better budgeting. Knowledge of aid flows and their management is much better.26 Nonetheless, the depth and sustainability of the budget reform re- mains to be demonstrated. Much of it is new. For example, the reform of budget nomenclature, which absorbed many months of TA time, was rejected by IMF experts as inadequate. This change, which should have been introduced in 1992 or 1993, did not arrive until 1998. The elaborate system of budget indicators worked out under HRDP seems to have been a one-shot exercise. It may have left some traces, but these were not readily visible in mid-1999.27 This means that monitoring and evalua- tion capacity remains embryonic or nonexistent. Project preparation and screening capacities are also uncertain.28 Cote d'lvoire 411 Concrete evidence that gains in capacity in the health sector are not extensive can be found in the recent suspension of the World Bank's major integrated health project (PDSSIR) and in the story of misappro- priation of European Union (EU) assistance, which became headline news early in 1999, and of which more will be said below. Growth in capacity appears to have been more substantial in the edu- cation sector. There is a newly installed planning unit in the Ministere d'Enseignement Nationale et Formation de Base. But project prepara- tion capacity is said to be much stronger at the tertiary level, at the Ministere d'Enseignement Superieur et Recherche Scientifique; they apparently prepare more and better projects for inclusion in the PIP than do the other education ministries. The dispersion of decision units makes sectoral planning and coordination difficult, a difficulty exacerbated by the frequent restructurings and shuffling of units between ministries. Three factors seem to be most relevant in explaining the patchy imple- mentation and unsatisfactory outcomes of Ivoirian social sector reform programs. First, the major donor efforts made early in the 1990s, notably the HRDP, were really part of the larger macroeconomic drama then in process. Everybody's focus was on the macro issues. It is no surprise that the HRDP was managed by the Ministry of Finance. It was intended to provide quick-disbursing, financial gap-closing assistance to the RCI. For its first two years little happened while the devaluation decision was pondered. Right after devaluation, disbursement occurred: 80 per- cent of the US$360 million credit was disbursed in 1994. The sideshow character of the health and education reform agenda meant that Ivoirian decisionmakers gave it limited time and energy. The Bank and other donors had the same hierarchy of priorities as the government, and this is the second reason for poor performance. Macro- economic stabilization was the highest priority by far. Supervision and concern over the substantive sectoral issues were not negligible, but the Bank side was preoccupied with keeping the money flowing. Attention to the conditionality in the social sectors was secondary, and assessments of performance perfunctory. Finally, local decisionmakers believe only weakly in the major donor objective of shifting resources from tertiary levels to primary. Believers among politically aware elites are a distinct minority. Nor are all bilat- eral donors fully persuaded. On the Ivoirian side the argument is partly political. The enormous growth in numbers of university students was a potential time bomb. It had to be confronted. Moreover, there are eco- nomic rationales for continuing attention to the tertiary sector. It would be disastrous, they believe, to let the existing social infrastructure at that level continue its deterioration. The quality of university outputs has been declining, at great social cost.29 It cannot be improved without ad- ditional spending. 412 Mixed Reformers Officials and other Ivoirians strongly disagree with judgments such as those incorporated in a recent Bank-financed health sector PIP re- view, which criticized six projects in the 1997-99 sectoral PIP, mainly renovation and re-equipment of university hospital centers in Abidjan (Shephard 1999). These six projects absorb 33 percent of the PIP and 40 percent of donor PIP financing. Almost 60 percent of their cost is to be financed by donor grants (Japanese and Spanish), a much higher ratio than other projects. Concerned officials and other Ivoirians question the validity of Bank recommendations to cut back on these projects (which are after all heavily grant-financed and likely to produce improved ser- vices) in favor of primary care-focused projects that are more depen- dent on local financing and may have uncertain outcomes. Aid and Reform Education reforms that began in the late 1980s in RCI were not entirely without precedent. Substantive reforms and strategy statements were drafted by Ivoirians and their advisers as early as 1972, and there were many internally generated ideas for change in both education and health in the 1980s, before there was a strong donor presence. So some reform proposals, for example in curriculum, teacher training, and nurse train- ing, were already in the air, and provided some of the elements for the later reform proposals. Part of the reform program of the 1990s can thus be said to have been internally generated. Moreover, Ivoirians provided inputs during the preparation of major projects like the Human Resources Development Program, which was almost four years in the making. Extensive consultations, formal and informal, marked this process. The companion TA project, the Human Resources Management Support Project (Projet d'Appui a la Gestion des Resources Humaines, PAGRH), had an even longer gestation period, involving even more local consultation. It was framed with great sensi- tivity to the problems of technical assistance delivery, which were then (in 1993) very much on the table. It emphasized use of local rather than resident expatriate TA. So even the first round of social sector reforms, during the first half of the 1990s, had some domestic input. But it seems clear that the main trigger for the major reforms of the 1990s was the financial crisis and the need to contain social sector expenditures if macroeconomic stabil- ity was to be attained. Donors were the main bearers of this message. The significant planks of the reform program were focused on cost- cutting and on equity, donor preoccupations much more than local. The specific reforms were crafted mainly by donor staff, who also acted as their chief champions. Local technocrats in the core ministries, and some politicians, realized that the social sector budgets had to be con- C8te d'Ivoire 413 strained, and that this would involve strong salary restraint and sharp cutbacks in subsidies to students, among other changes. But the major impetus for these reforms came from outside, particularly the Bretton Woods institutions. The outside impetus is clearest with respect to the idea that expendi- tures should be restructured in favor of primary education and primary health care. As noted above, the Ivoirian leadership was preoccupied until 1994 with debt crises and the devaluation issue. There was not much residual energy for, and only modest interest in, social sector expendi- ture reallocations. The Finance and Planning ministries negotiated ad- justment credits and signed conditional agreements with detailed letters of development policy. The subject matter was education and health sec- tor reform, but the pressure was to bring in the money. The Education and Health ministries participated relatively little in negotiations on the US$360 million adjustment credit begun in 1991, the HRDP; substantive concerns were in the back seat during its implementation. As noted earlier also, very few Ivoirians had much conviction that the reallocation of spending in favor of primary subsectors was sensible, especially since it was easier to raise grant money for tertiary activities. This lack of conviction coexisted uneasily with the feeling, widespread at least since the mid-1990s, that more attention to primary schooling and health care was politically correct, the right thing to do. These were not the only factors that made the donors the dominant players. There was also the fact that capacity in the government was too limited to enable it to take the lead. This changed in the late 1990s, when a stronger, more experienced corps of local technicians began to assert itself, drafting documents like the Policy Framework Paper and the com- prehensive strategy statement called the "African Elephant." Local tech- nicians also played important roles in the conception and implementation of sectoral investment or development programs in health and educa- tion after the mid-1990s. But for the first generation of sweeping reforms, the domestic contribution was much more limited. Donors, then, bore the major responsibility for the adoption of the major reforms, from the expenditure reallocation objectives to strength- ening of budget and personnel management. Local support existed in the core economic ministries for the management reforms and, less clearly, for the budgetary shift to primary subsectors. But few other local cham- pions were apparent. Most of the actors-Education and Health minis- try personnel, teachers, doctors, nurses, pharmacists, and others-were neutral or opposed. Their professional associations were in any event not vocal. Students at the secondary and tertiary levels were opposed. One other environmental factor contributed in a basic way to the domi- nant donor role in the reform process. It is not clear that users of health and education services expressed at any stage great dissatisfaction with 414 Mixed Reformers the status quo. A recent study surveyed 700 rural users of health facili- ties. It found that general satisfaction with respect to the care provided was rather good. The main complaint was lack of medicines (Cote d'Ivoire 1998a). This same lack of demand for change from the affected popula- tion has been noted in studies in Southem and East Africa (Mogedal, Hodne Steen, and Mpellmbe 1995). In the education sector user dissatis- faction is evident mainly in the technical and vocational subsector, where it finds expression in stagnant student numbers and movement to pri- vate providers. The Major Hypotheses What does the education and health reform experience in RCI tell us about the general hypotheses that inspire the research project of which this case study is part? Take first the hypothesis that nonfinancial aidgenerates more reform than money aid. There is little or no evidence in support of this proposition in the institutional reform experience of the RCI. Two aspects of the educa- tion and health sector reforms in RCI raise doubts about it. First, "nonfi- nancial aid" in this study is defined as freestanding technical assistance, donor economic and sector research, or advisory services. But in the case at hand, technical assistance projects have been handmaidens of finan- cial aid projects. The Human Resources Management Support Project, for example, was really a part of the HRDP, an adjustment credit. Nonfi- nancial aid that is untied to money seems to have been rare. Secondly, financial aid projects themselves carry a great deal of non- financial baggage. In the RCI case many of the adjustment credits have had study components, training tours, training seminars, and so forth. More important is the long gestation period that has been typical for RCI projects. The HRDP was more than three years in preparation, and took several years to get off the ground after approval. Many studies and much dialogue occurred during this period. It is not only the Bank that has such long identification and appraisal periods for financial aid projects. The European Union Fond Europ6en de Developpement (FED), the European community development financing agency, does exten- sive program preparation that entails research and analysis and dialogue. For example, a number of basic studies and surveys formed the back- ground for the health sector component of the upcoming Eighth FED. Ivoirians participated in and reviewed these studies, which generated intensive dialogue. Studies and dialogue attached to money are a priori likely to be more effective thanfreestanding studies, which probably include most World Bank economic and sector work. The money is almost sure to make re- search and dialogue more action-oriented. It claims attention and con- Cote d'Ivoire 415 centrates minds in a way that freestanding studies do not. It is not clear, for example, that the 1996 Bank poverty assessment, which found sharp increases in the extent of poverty, had any effect in shaping education and health policy. Aside from the fact that the results have been con- tested, the political class is not unaware that many Ivoirians are very poor (Jones and Ye 1997). In any case the importance of a primary-level social sector focus in combating poverty has been part of public dis- course for many years. A second hypothesis is thatfinancial aid works when policy reforms and institution building are underway. This is not an operational hypothesis. In the 1990s at least, reforms and institution building of some sort are al- most everywhere and always "underway." In the RCI, reform of some kind was going on most years in the 1980s and every year in the 1990s. The proposition is ambiguous in other respects. It can be taken as a sequel to the proposition that financial aid does not work when policy/ institutional reform is not underway, which is unclear but probably in- correct in its extreme form, when applied to social sector-type reforms. Even in the worst policy/institutional environment, financial aid can have some positive effects if appropriately delivered, and in any case may be a prerequisite for getting a seat at the policy table. If the argu- ment is that aid works better in good environments, that is surely true. In the RCI case a government more convinced and committed to expendi- ture reallocation in favor of primary education and health care would have gone further in that direction in the past decade. Does the RCI social sector reform experience confirm the hypothesis that governments choose to reform independently of the aid relationship? This can be and has been expressed in various forms: that aid has no impact on policy, or that aid cannot buy reform. This view, a major conclusion of the World Bank's influential study Assessing Aid (1998a), is shared by some Ivoirian officials. One said that the Bank and others who believe that their efforts have induced reform are fooling themselves just like those who accepted Chanticleer's claim that his crowing makes the sun come up. But it is not tenable to argue, in the case of institutional reforms for education and health, that aid has had no positive effects on reform. Even though implementation has been partial and some of the institu- tional reforms (performance budgeting in education, for example) did not "take," the reform program would have been different in content (less good), slower in implementation, probably less effective in results. The action-oriented studies, training, and study tours would have surely been fewer in number and probably less good had external financing been absent. The message about reallocation of spending to primary lev- els would have been much more muted. Some of the least popular measures, such as salary cost containment and reductions in student subsidies, would have been less effective. The 416 Mixed Reformers harmed interest groups were told that it was essential to follow this path because "the World Bank insists on it" and large aid inflows depended on it. The debt relief initiative is now cited by the government in all austerity-related discussions with stakeholders. It seems no exaggeration to say that aid brought about such reforms in public expenditure management as have occurred. This was true of the process reforms in budget-making, cost-cutting efforts in student subsidies, staff redeployment, and, above all, in bringing formal accep- tance of the idea of reallocation of spending in favor of primary educa- tion and primary health care. Probably the most effective vehicle for aid influence was the dialogue, formal and informal, surrounding program preparation and supervi- sion. When, during project preparation or supervision, the Health Min- istry proposes that more nurses be recruited, Bank staff suggest concentrating more resources on training and supervision. When the ministry requests help for building up in-house maintenance capacity for biomedical equipment, the donor counterpart suggests contracting out that function. A chef de service says he needs an administrative secre- tary who knows all about the Ivoirian legal code. The donor spokesman asks why he needs that rare and costly competence. Of course, not all donor staff suggestions are good. Not all donor project designs are well suited to the local environment. This raises the broader question of whether aid can retard reform. The aid presence does seem to have had some negative effects in the health sector. One is a distortion of priorities. Primary-level spending by the Ministry of Health has lagged badly behind plans and needs. The donor community has pushed hard to have the ministry spend faster and spend particularly on building up rural health infrastructure. A Bank review of the impor- tant Integrated Health Services Project (PDSSIR) comments as follows: The project undoubtedly has some perverse effects. Because of the large amounts of money available, and the resulting pressure to disburse, it shapes the priorities of responsible officials. It pushes them to invest in health facility construction and equipment, while the basic problems lie elsewhere, notably in the poor definition of responsibilities and the weakness of capacities to manage health facilities and properly organize work. (C6te d'Ivoire 1998c) Those interviewed in our field visit took up this same theme: that while the Ministry of Health has much stronger intellectual capacity than it had a few years ago and many more cadres with training in public health, its implementing capacity remains extremely weak. Donor rep- resentatives pointed out, for example, that while the National Health Development Plan is a good document, it has been around for three years and little has been done to implement it. This is due in part to systemic Cote d'Ivoire 417 problems that destroy incentives, such as low civil service pay and poor working conditions. And partly it is due to high turnover at senior lev- els within the ministry. Overfull plates are fundamental. More dramatic evidence of implementing deficiencies is at hand. An audit of European Union programs in health at the end of 1998 discov- ered a large volume of improper transactions, spending that had not been adequately justified. The audit estimated that CFAF 18 billion-23 billion out of total program expenditures of CFAF 53 billion was involved. The government took no disciplinary action for some months, but fi- nally six ministry officials were indicted and, after a further delay, the minister was sacked. Two reasons are given to explain this unhappy incident. One is that the EU was pressed to spend, to speed up construction of rural facilities; otherwise the slow-moving administration would fail to meet the ambi- tious goals that had been established for health service provision. They urged government to loosen budget procedures, notably to raise the level of payment that could be made without going through all the laborious steps prescribed in the budget laws. The second is that too many goals were set for the program, not only building large numbers of health posts but also using small and medium-sized contractors to do the job. Capac- ity for managing all this was simply not there. Similarly, after many warnings, the World Bank in August 1999 sus- pended its main health project, the PDSSIR. Two years after project effectiveness, with its completion date not far off, only 16 percent of the project budget had been disbursed. The underlying reason here was again excess optimism about local capacity to implement. But there were other explanations. Some critics say that in addition to being too ambitious, the project was put together too fast and had too short a time horizon. Other critics, more severe, argue that the project was poorly conceived. It had two pillars, both new in RCI and both complicated: the focus on a "minimum activities package" (PMA in French), and the reorganization of health service structures along "district" lines. Definition of a PMA turned out to be far more difficult than anticipated. And more impor- tant, the district concept did not fit easily into existing structures, or in- deed into existing decentralization programs supported by other donors. The ministry could not or would not implement. Since 1993-that is, well before the Bank introduced its own decen- tralization approach in the PDSSIR-other donors had been pursuing another approach to deconcentration or decentralization. The EU, the French, and perhaps others, were trying to strengthen the Regional Health Directorates. The Bank introduced the different concept of "dis- tricts." It organized them quickly, ignoring the regionalization efforts of the other donors.30 The Bank also pushed hard on the PMA approach, apparently assuming that it could be easily and quickly adopted. But 418 Mixed Reformers it raised enormous problems of fit with existing structures, just as the district concept did. All kinds of fundamental issues had to be resolved. How to reconcile existing regional structures with the district, which has no traditional counterpart in RCI? How to work out PMA delivery when the tradi- tional system is vertical-a family planning unit, a maternal and child care unit, and so on. How to build sustainable capacity when different donors take over particular vertical slices. United Nations programs are completely vertical: UNICEF does vaccinations, the WHO information transfer, the UNFPA something else. How to bring about integrated ser- vice provision for primary health care when vertical structures remain dominant? There was little coordination, little real dialogue to permit resolution of these basic matters of strategy and approach. Some donors appear to harbor strong resentment of the World Bank role. One view is that the Bank, seeking "leadership," tends to monopolize the dialogue. The ex- ample is given of a Bank project that contains a component on reform of medical school curriculum, which leads to a squeezing out of other do- nors. Dissatisfaction with the World Bank role in the health sector seems to be surprisingly lively, greater than is indicated in client survey data (World Bank 1998c). LEGAL AND JUDICIAL REFORM A predictable, efficient, and transparent judicial system is a cornerstone of good economic management, and critical for social equity, private sector development, and good governance in general. That sound legal and judicial institutions make these vital contributions has been recog- nized for decades. Because these institutions are weak in many develop- ing and transition economies, aid donors have financed programs of legal/judicial (L/J) reform in many developing countries. In the RCI, as in many low-income countries, legal reforms during the past decade have had a twin focus. One concern has been to remove legal obstacles to labor market flexibility, frequently regarded as a major reason for slow supply responses to stabilization and liberalization. The second concern is much broader: reform of the overall judicial system, strengthening its operational capacity and adapting it to the needs of a modern market economy. Labor Market Reforms The pre-reform situation (the late 1980s) was that the Labor Code and negotiated collective agreements made it difficult for employers to ad- Cote d'Ivoire 419 just work forces to fluctuating needs. Temporary workers were consid- ered permanent after 90 days on the job. Overtime regulations were strict, and required prior authorization from the Ministry of Labor. It was very difficult to sack anyone; charges of "abusive" firing were common, and could lead to legal damages and reinstatement. Employers had to list staff by seniority, and unless they followed a last-in first-out policy in staff reductions their action could be considered "abusive." The state-run labor exchange (Office de Main d'Oeuvre de la C6te d'Ivoire, OMOCI), a monopoly hiring agency, was another source of prob- lems. Although in practice most firms did their own hiring, their choices needed OMOCI approval. Moreover, prior authorization by OMOCI was necessary for staff reductions. This gave rise to hassles and delays and opened opportunities for corruption. Labor market reform was a major component of the Bank's 1991 Com- petitiveness and Regulatory Reform Adjustment Program (PASCO). Elimination of the OMOCI monopoly, elimination of constraints on hir- ing temporary or casual labor and on collective layoffs, and dropping prior administrative authorization for overtime were conditions of ef- fectiveness for PASCO. Revision of the Labor Code and encouraging renegotiation of the central collective agreement were conditions of sec- ond-tranche release. Many of these reforms were implemented. OMOCI's monopoly was abolished. Restrictions on temporary workers were lifted. Layoffs for economic reasons were made easier and the Labor Ministry's prior au- thorization was no longer needed for overtime. The Labor Code was liberalized and approved by the National Assembly in 1992. Many of these changes had been introduced only under extreme Bank pressure, and with long delays that indicated obvious reluctance. The decrees implementing the Labor Code (21 of them) did not appear until 1996, four years after the National Assembly vote, and only after their issuance was made a tranche release condition of another adjustment loan (the Private Sector Development Adjustment Credit of 1996). Also, while OMOCI's legal monopoly over hiring and firing was elimi- nated, the government created a new public placement agency, Agence d'Etudes de Promotion d'Emploi (AGEPE), which had many of the same employees as OMOCI.31 AGEPE was supposed to monitor employment trends, but it has come to act as a regulatory body, and is said to be an unequal competitor with private recruitment agencies. It delivers work permits and operates as a recruitment agency itself for foreigners. More- over, it charges a large fee, particularly for non-Africans. It is an expen- sive agency, and apparently inefficient.32 Despite revision of the Labor Code and institutional changes in the labor market, it is not certain that the actual functioning of these mar- kets is much more flexible than in pre-reform days. One reason is that 420 Mixed Reformers the extent of employer inability to adjust to fluctuating needs for labor was almost surely exaggerated. Subcontracting of parts of company op- erations was always possible; this was one way to meet seasonal needs, and was probably widespread. Also, some of the forces that stayed em- ployer hands are still present. Collective bargaining agreements reflect (indirectly) government policies and concerns. The general judicial en- vironment remains uncongenial. Employers are still vulnerable to legal challenges by fired workers claiming abusive action and to uncertain outcomes of court proceedings. Nonetheless, new elements of flexibility have certainly been intro- duced into Ivoirian labor markets as a result of the reforms of the past decade. Employers appear to be more comfortable in this new environ- ment. They say so, and the buoyancy of private foreign direct invest- ment tends to confirm it. General System Reform Concerned donors and the RCI government turned to general judicial reform in the early 1990s. The prevailing diagnosis was that the quality of the L/J system had declined since the 1970s, and that by 1990 it was very weak, distrusted by the private sector, and a serious obstacle to market-oriented economic growth. The main elements of the common diagnosis were as follows: * Antiquated legal code. Most of the laws relating to business dated from the nineteenth century. They were complex, anachronistic, and seriously deficient in such areas as contracts and credit, corpo- rate financial and accounting regulations, and bankruptcy and liq- uidation. * Inadequate resources. The Ministry of Justice (MOJ) budget was small, less than 2 percent of the state budget in the early 1990s, with al- most no provision for nonsalary recurrent costs. Courts were too few in number: three-quarters of cases were handled by the Abidjan Court, inevitably clogging the system; only 6 first-level courts were in operation (of 10 that had been created), and 26 court sections out of 39. Judges were also too few. They numbered 253 for a popu- lation of 14 million, about one-twentieth of international standards. And the greffiers (clerks of court), who were responsible for admin- istering the court's activity, were poorly trained. Only 47 of the 467 greffiers were attaches (had postsecondary education); the rest were assistants (301) and secretaries (120). * Inefficient and inequitable functioning. Access to the courts was ex- pensive and procedures burdensome. Respect for claims was ir- regular and contracts were frequently not enforced. Corruption was CBte d'lvoire 421 said to be widespread, encouraged by the fact that judges frequently and without penalty failed to state the grounds for rulings, and by the fact that judgments were not published. Judges' decisions were often of low quality.33 The poor administration of judicial arrange- ments was attributed to a shortage of judges, their weak general training, and their lack of specialized knowledge (in financial and accounting matters in particular), as well as to complex procedures and provisions that allowed defendants to adopt delaying tactics. Corruption, archaic working conditions, and inefficiency in the office of the clerk of the court (the greffier) was a major source of bottlenecks in the system.34 * Lack of transparency. Information about applicable regulations and procedures was dispersed and inaccessible, and procedures lacked transparency. This led to inconsistencies in treatment of claimants; uneven treatment of private sector agents was particularly harm- ful to business confidence. The first major move in judicial reform in the RCI began in 1991 with the preparation of the World Bank's Economic Management Support Project (Projet d'Appui a la Gestion Economique, PAGE). The project was designed jointly with the French Ministry of Cooperation and emerged from a decision of the two aid agencies to divide responsibility in this reform area. The French aid agency (Cooperation Franqaise) and the World Bank worked together through most of the 1990s, the Bank through judicial reform components in a succession of multi-component projects, the French by a US$1.3 million grant from the French Ministry of Coopera- tion.3" The Ministry of Justice was to be strengthened. Better budgeting, organizational changes, computerization, and tighter inspection/con- trol were to increase transparency and raise the productivity of existing resources. The court system would be expanded by hiring more judges and other staff and creating new courts in secondary cities. Training of judges and others would be intensified. An arbitration court for com- mercial dispute settlement would be created. Management and dissemi- nation of legal data would be improved by creating a documentation center. Offices of the clerks of the court (greffiers) would be strength- ened by modernizing offices in Abidjan and Dabou, training staff, ac- quiring computers, giving greater incentives to personnel, and simplifying legal forms. Major changes in the nature and management of the fees paid to the greffes would reduce corruption and raise money for court use (see below). While these internal reform efforts were underway, the RCI govern- ment and the French governmen