2 0 0 6 A B C D E 35518 Annual World Bank Conference on Development Economics Growth and Integration Edited by François Bourguignon and Boris Pleskovic Themes and Participants for the ANNUAL WORLD BANK CONFERENCE ON DEVELOPMENT ECONOMICS (ABCDE) St. Petersburg, Russia "BEYOND TRANSITION" JANUARY 18­19, 2006 Growth after Transition: Is Rising Inequality Inevitable? Economic Space Governance Judicial Foundations of a Market System François Bourguignon · James Anderson · Anders Aslund · Chong-en Bai · Erik Berglof · Patrick Bolton · Karolina Ekholm · Yegor Gaidar · Kiran Gajwani · Alan Gelb · Cheryl Gray · Irena Grosfeld · Sergei Guriev · Nazgul Jenish · Ravi Kanbur · Thierry Mayer · Bill Megginson · Pradeep Mitra · Gur Ofer · Marcelo Olarreaga · Ugo Panizza · Guillermo Perry · Andres Solimano · Ernesto Stein · Matthew Stephenson · Jan Svejnar · Mariano Tommasi · Stefan Voigt · Ruslan Yemtsov · Xiaobo Zhang · Ekatherina Zhuravskaya Growth and Integration Annual World Bank Conference on Development Economics 2006 Growth and Integration Edited by François Bourguignon and Boris Pleskovic THE WORLD BANK Washington, D.C. ©2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 09 08 07 06 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6093-0 ISBN-13: 978-0-8213-6093-4 eISBN: 0-8213-6094-9 DOI: 10.1596/978-0-8213-6093-4 ISSN: 1020-4407 Contents ABOUT THIS BOOK vii INTRODUCTION AND SUMMARY 1 François Bourguignon and Boris Pleskovic WELCOME ADDRESS Africa's Growth Challenge in a Globalized World 9 His Excellency Abdoulaye Wade, President of the Republic of Senegal OPENING ADDRESS Equity and Growth in Africa 17 François Bourguignon KEYNOTE ADDRESS The Importance of Research for Economic Policy: The Case of Senegal 25 Abdoulaye Diop Growth and Integration Explaining African Economic Growth: The Role of Anti-Growth Syndromes 31 Augustin Kwasi Fosu and Stephen A. O'Connell COMMENTS Jean-Paul Azam 67 Abdoulaye Diagne 71 Jean-Paul Azam 77 Financial Reforms Financial Sector Reforms in Africa: Perspectives on Issues and Policies 81 Lemma W. Senbet and Isaac Otchere COMMENT Mthuli Ncube 121 VI | CONTENTS Economic Development Convergence and Development Traps: How Did Emerging Economies Escape the Underdevelopment Trap? 127 Jean-Claude Berthélemy COMMENT Mthuli Ncube 157 Trade and Development Risks and Rewards of Regional Trading Arrangements in Africa: Economic Partnership Agreements between the European Union and Sub-Saharan Africa 163 Lawrence E. Hinkle and Richard S. Newfarmer COMMENT Léonce Ndikumana 189 Investment Climate Business Environment and Comparative Advantage in Africa: Evidence from the Investment Climate Data 195 Benn Eifert, Alan Gelb, and Vijaya Ramachandran COMMENT Léonce Ndikumana 235 About This Book The Annual World Bank Conference on Development Economics is a forum for dis- cussion and debate of important policy issues facing developing countries. The con- ferences emphasize the contribution that empirical and basic economic research can make to understanding development processes and to formulating sound develop- ment policies. Conference papers are written by researchers in and outside the World Bank. The conference series was started in 1989. Conference papers are reviewed by the editors and are also subject to internal and external peer review. Some papers were revised after the conference, sometimes to reflect the comments by discussants or from the floor. Most discussants' comments were not revised. As a result, discus- sants' comments may refer to elements of the paper that no longer exist in their orig- inal form. Participants' affiliations identified in this volume are as of the time of the conference, January 27, 2005. François Bourguignon and Boris Pleskovic edited this volume. As in previous years, the planning and organization of the 2005 conference was a joint effort. Spe- cial thanks are due to Alan Gelb for overall guidance. We thank several anonymous reviewers for their comments and Aehyung Kim, Célestin Monga, and Benno Ndulu for their useful suggestions and advice. We also thank the conference coordinator, Leita Jones, whose excellent organizational skills helped ensure a successful confer- ence. Finally, for pulling this volume together, we thank the editorial staff, especially Stuart Tucker, Aziz Gokdemir, and Mark Ingebretsen from the Office of the Pub- lisher; Barbara Karni, who edited the manuscript; Sherrie Brown, who proofread the book; and Theresa Bampoe, who reviewed the proofs. The book was designed by Naylor Design, Inc. Book production and dissemination were coordinated by the World Bank Office of the Publisher. Introduction and Summary FRANÇOIS BOURGUIGNON AND BORIS PLESKOVIC The Annual World Bank Conference on Development Economics (ABCDE) is one of the best-known conferences on development economics. The purpose of the ABCDE is to expand the flow of ideas among scholars and practitioners of development policy from academia, government, and the private sector around the world. By fostering a better understanding of development and the problems developing countries face, the conference aims to enhance policy making at the World Bank and at its partner institu- tions. The ABCDE also provides a forum for exposition by academics and practitioners as they seek to identify and elaborate on new ideas and issues pertinent to development. The 17th conference, held in Dakar, Senegal, January 27­28, 2005, was the sec- ond such conference to be held in a developing country. The change in locale reflects the growing importance of research done in developing countries and the desire to bring such conferences closer to participants in the developing world. The first day of the conference opened with remarks by His Excellency Abdoulaye Wade, President, the Republic of Senegal; and by François J. Bourguignon, senior vice president and chief economist, World Bank; and a keynote address by Abdoulaye Diop, Minister of Finance and Economy, Senegal. Their remarks were followed by five papers on the theme of growth and integration, addressing growth and integration, financial reforms, economic development, trade and development, and the investment climate. Opening Addresses: Africa's Growth Challenge in a Globalized World In his opening address, His Excellency President Abdoulaye Wade asserts that Africans want to be active participants in shaping the world and in debating economic matters at international forums in order to tackle Africa's economic François Bourguignon is senior vice president, Development Economics, and chief economist at the World Bank. Boris Pleskovic is research manager, Development Economics, at the World Bank. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 1 2 | FRANÇOIS BOURGUIGNON AND BORIS PLESKOVIC development challenges. He is hopeful that new emerging coalitions will integrate Africa's voice into the globalization debate to be heard. Many global issues affect Africa directly. Free and fair trade are very important, because products such as cotton are the main source of livelihood for millions of peo- ple across the continent. Technology and knowledge transfer to Africa would improve the productivity of agricultural production. External resources, including international aid, have promoted economic growth in Africa, although in some coun- tries they have created a debt overhang that has impeded economic progress. The deterioration of the terms of trade has worsened the financial condition of Africa. Economic growth and regional integration are part of Africa's efforts to improve its capacity to deal with the challenges of a globalized world. Recent efforts at regional integration have included the establishment of the New Partnership for Africa's Development (NEPAD), the West African Economic and Monetary Union (WAEMU), and the Economic Community of West African States (ECOWAS). NEPAD seeks to foster good governance, private sector development, and capacity building. It has identified eight priority sectors: agriculture, education, energy, envi- ronment, health, information and communication technologies, infrastructure, and access to markets of developed countries. WAEMU provides a framework for finan- cial and economic integration. ECOWAS functions as an effective political instru- ment for regional integration. It is hoped that these efforts, together with other North-South initiatives, such as the African Growth and Opportunity Act (AGOA) and the Millennium Challenge Account (MCA), will spur economic growth and development in Africa. In his opening address, François Bourguignon offers a nonlinear view of growth. He argues that there is complementarity among the determinants of growth, partic- ularly in Sub-Saharan Africa and other low-income regions. Investing in new capital cannot be done without an adequately trained labor force, agricultural productivity cannot be raised without investing in rural infrastructure, private investments cannot be dynamic without the appropriate institutional environment, and so forth. From both an analytical and a policy point of view, complementarity makes it essential to identify what may be the constraining factor in a given economy at a given time. These constraints tend to change over time; it may be necessary to deal today with constraints likely to be binding tomorrow to avoid creating growth that is a succes- sion of short-lived surges and stagnation periods. Another condition for sustained growth is equity, the second pillar of the World Bank's strategy. Here, too, Bourguignon suggests that there is strong complementar- ity between long-run growth and equity. Economists tend to stress the conflict between efficiency and equality of income or consumption in a well-functioning economy. But efficiency and inequality need not work at cross-purposes: in imper- fect markets, there may be situations in which redistributive policies can increase effi- ciency and growth. Bourguignon concludes that there is no universal recipe for growth: designing growth strategies should be a highly context-specific exercise. Macroeconomic aggre- gates and some characterizations of institutions are not the only variables that mat- INTRODUCTION AND SUMMARY | 3 ter; micro information is equally important, especially in the African context. If well- crafted growth strategies are to lead to sustained growth--and not to the ephemeral growth episodes observed in many African countries and elsewhere--parallel progress must also be made on the equity side of the two-pillar strategy. Empower- ing poor people in all dimensions of their social life is a necessary condition for sus- tainable growth in the long run. This empowerment is possible only if economic growth creates enough resources. Keynote Address: The Importance of Research for Economic Policy in Senegal Abdoulaye Diop, Minister of Finance and Economy, Senegal, discusses two issues: recent developments in Senegal's economic situation and its prospects and the impor- tance of research for economic policy. Senegal experienced economic growth of 6.5 percent in 2003 and 6.1 percent in 2004. Although poverty still affects more than half the country's population, the percentage of people living in poverty fell from 67.9 percent in 1994­5 to 57.1 percent in 2001­2. This improvement has been fos- tered by economic growth since the January 1994 devaluation. Structural reforms, including implementation of the national good governance program, a healthy macroeconomic framework, and the creation of an incentive-based environment, have collectively contributed to recent economic growth. Economic research has become an integral part of policy making in Senegal, con- tributing, for example, to the drafting of its Poverty Reduction Strategy Paper. The government has begun to support various research centers studying development problems in Senegal and in Africa. As a result of policy exchanges among local par- ticipants, the Poverty Reduction Strategy--that is, the notion that to reduce poverty, growth must be brought to the people who are affected by it--has become broadly accepted by the population and decision makers. Research includes the creation of a database, the collection of statistics, and the execution of surveys on living conditions, health, household spending budgets, demographic studies, and other development-related issues. The availability of a database enables researchers to respond to requests by public officials in a timely manner. But the process of knowledge accumulation is often slowed by the "brain drain" caused by the lure of more favorable conditions elsewhere. Building research capacity in formulating economic policies in Senegal is thus challenging. Growth and Integration Augustin Kwasi Fosu and Stephen A. O'Connell examine the episodic nature of African growth. They describe four "anti-growth syndromes" (regulatory regimes, redistribution regimes, intertemporal regimes, and state breakdown), tracking their evolution over time and assessing their impact on predicted growth within a sample 4 | FRANÇOIS BOURGUIGNON AND BORIS PLESKOVIC of African countries. They draw on case study evidence to illustrate the syndromes and to study the forces behind their adoption and abandonment. Fosu and O'Connell argue that the central puzzle of the 1960­2000 growth record is the failure of the vast majority of African economies to achieve sustained increases in per capita incomes. This failure is due largely to some combination of adverse growth opportunities and unsuccessful policy choices. What the African evidence suggests, however, is that being syndrome free is virtually a sufficient condition for avoiding the short-run growth collapses that have so often undermined long-run growth performance on the continent. While an absence of syndromes does not guar- antee rapid growth, it is virtually a necessary condition for sustaining rapid growth in the medium term. Where strong growth emerges in the midst of a syndrome, it tends to do so either temporarily, as a result of unsustainable dynamics, or both tem- porarily and fortuitously, as a response to favorable shocks. The most daunting puzzle in the African growth experience is the failure of its coastal, resource-poor economies to participate in the explosion of manufactured exports from developing countries that occurred in the last quarter of the twentieth century. The global evidence suggests that countries such as Kenya, Mauritius, and Senegal faced very favorable growth opportunities during this period. But only Mau- ritius managed to capitalize on its low transport costs and abundant labor by develop- ing a strategy to attract domestic and foreign investment into export-oriented manu- facturing. At the beginning of its successful export push, Mauritius had a very substantial head start over the rest of coastal Africa in terms of human development. A critical and unresolved question is whether an export-led growth strategy is subject to major threshold effects in human development. If it is, achieving success may require a massive initial impetus to increase the quantity and quality of health and education. Financial Reforms Lemma W. Senbet and Isaac Otchere analyze recent financial sector reforms in Africa and provide some policy guides to building the capacity of financial systems. The financial sector does much more than just serve as a conduit for mobilizing savings. It also produces information, reveals prices, facilitates risk sharing, provides liquid- ity, promotes contractual efficiency, improves governance, and facilitates global inte- gration. These vital functions are at the heart of a well-functioning system. Various measures can improve the functioning of the banking sector. These include well-designed deposit insurance schemes and market- and incentive-based regulation of capital markets. With respect to capital market development, Senbet and Otchere focus on the fledgling stock markets and supporting institutions in Sub-Saharan Africa. Their rec- ommendations include measures for developing nonbanking institutions, which fos- ter competition in the financial system and mitigate the adverse consequences of oli- gopolistic banking systems in Africa. The development agenda for stock markets INTRODUCTION AND SUMMARY | 5 should consider measures that increase public confidence, improve informational efficiency, provide liquidity, create synergy among regional stock markets, and facil- itate global integration. Senbet and Otchere take a functional perspective with respect to financial sector reforms, recognizing that African financial systems must be designed to serve multi- ple functions if they are to contribute to economic development. They examine the prospects for financial globalization of Africa based on mutual gains for global investors and the region at large. They then investigate the features of African finan- cial systems, describing packages of reforms that have been adopted and identifying the constraints faced by these systems in achieving the desired outcomes of savings mobilization and financial stability. Based on this assessment, they suggest develop- ing the domestic financial sector, designing efficient regulatory schemes and safety nets, implementing measures that increase the depth and liquidity of African stock markets, and reforming the financial sector. Economic Development Jean-Claude Berthélemy examines how emerging economies managed to escape underdevelopment traps by reviewing the empirical research on convergence clubs. Twin peaks in the international distribution of incomes, the quadratic relation between initial income and future growth, and the dependence of structural param- eters of conditional convergence equations suggest that multiple equilibria may exist. A poor country cannot grow out of poverty unless policy initiatives are taken to change initial conditions in such a way that it can "jump" from its low, stable, ini- tial equilibrium to another, higher-income equilibrium. In the presence of multiple equilibria, policies, not just initial conditions, matter. Berthélemy considers the following policy issues: demography, savings behaviors, and human capital accumulation; qualitative aspects of the economic growth process, such as financial development or diversification of the economy; and insti- tutional issues, such as corruption and armed conflicts, which may trap a nation in a destructive dynamic in which poverty and poor institutions reinforce each other. He observes that several of these mechanisms may be at work in different phases of the economic development process, so that the development process may involve multiple equilibria. To lift a poor country out of its underdevelopment trap, economic reforms would be more effective than "big push" strategies that seek to increase income in poor countries through large transfers of assistance. He asserts that the jump should lead to multiple growth peaks, or a "growth acceleration cycle" pattern. Education policies may have been a key factor in determining economic perform- ance in developing countries during the past 50 years. Many economies in Southeast Asia implemented ambitious education policies early on, initially aiming at achieving universal primary education. Most African economies have lagged behind. 6 | FRANÇOIS BOURGUIGNON AND BORIS PLESKOVIC Trade and Development Lawrence E. Hinkle and Richard S. Newfarmer review the Economic Partnership Agreement (EPA) process from a development perspective. The primary objective of EPAs between the European Union and Sub-Saharan Africa is to spur economic development in Sub-Saharan Africa. The authors examine the key issues raised by the merchandise trade and regional integration aspects of the planned EPAs, including their potential effects on regional integration in Sub-Saharan Africa; the relationship between EPAs and the European Union's Everything But Arms (EBA) Initiative and the access of Sub-Saharan African exports to EU market; the reciprocal opening of Sub-Saharan African and EU markets and the parallel Most Favored Nation (MFN) liberalization and restructuring of tax systems in Sub-Saharan Africa that would need to complement it; and the relationship between the EPAs and the World Trade Orga- nization's Doha Round. Hinkle and Newfarmer argue that EPAs present an opportunity to accelerate global and regional trade integration in Sub-Saharan Africa, if two conditions are met. First, the European Union must, as it has promised, truly treat EPAs as instruments of development, subordinating its commercial interests to the development needs of its African partners and effectively coordinating the EPA trade and development- assistance components. Second, Sub-Saharan African countries need to use the EPAs to accelerate trade and investment climate reforms that are necessary to raise growth rates and integrate African countries into regional and global economies. EPAs have much promise, but they also pose some development risks for Sub- Saharan Africa. A number of steps will need to be taken to limit these risks. These include creating attractive incentives for trade and investment climate reforms in Sub- Saharan Africa; reducing MFN tariffs and improving domestic tax systems; liberal- izing intra-African trade, both within and between regional economic communities; designing development-friendly provisions governing trade in services, investment, and competition; allowing enough flexibility to accommodate substantial variations in regional and country conditions; and timing and sequencing a complex series of reforms in an appropriate manner. Hinkle and Newfarmer conclude that if key issues are not resolved, the EPA process could end up being replaced by improved prefer- ences ("super EBA"), as proposed by the Blair Commission for Africa, or abandoned. Investment Climate Benn P. Eifert, Alan H. Gelb, and Vijaya Ramachandran consider two issues. The first is the impact of high indirect costs and losses in reducing the productivity and competitiveness of African manufacturing. The second is the complications for reforms caused by the fact that business sectors in Africa are heavily segmented by size, productivity, and ethnicity. Their conceptual framework is based on trade the- ories that see the evolution of comparative advantage as influenced by the business climate--a key public good--and by external economies between clusters of firms INTRODUCTION AND SUMMARY | 7 entering related sectors. Macroeconomic data from purchasing power parity esti- mates, though imprecisely measured, confirm that Africa is high cost relative to its levels of income and productivity. Firm-level estimates using data from surveys undertaken for Investment Climate Assessments between 2000 and 2004 indicate the importance of high indirect costs and business environment­related losses in depressing the performance of African firms relative to those in other countries when performance is expanded from a "factory-floor" concept to a broader "net productivity" measure. There are differ- ences across African countries, however, with some showing evidence of a stronger business community and better business climate. Eifert, Gelb, and Ramachandran argue that African business sectors are sparse and fractured along ethnic dimensions (indigenous, minority, and foreign investors). There are also wide productivity differences between large and small firms. Minority and foreign investors have much higher productivity and a greater propensity to export, but Africa's difficult business climate and the tendency to overcome it by working in ethnic networks slows new entry and may decrease the incentives of key parts of the business community to form an aggressive pressure group for reform. This increases the possibility that countries settle into a low-productivity equilibrium. Eifert, Gelb, and Ramachandran conclude by discussing reforms, such as creating new opportunities for the private sector, that can diversify and boost the competi- tiveness of African economies. Welcome Address Africa's Growth Challenge in a Globalized World HIS EXCELLENCY ABDOULAYE WADE PRESIDENT OF THE REPUBLIC OF SENEGAL It is my privilege to welcome you here in Dakar for this year's Annual Bank Confer- ence on Development Economics (ABCDE). It is indeed a great honor for Senegal to host this important gathering. The topic you have chosen for this meeting--growth and integration--is very important for Africa. We therefore expect a great deal from your discussions. Eco- nomic growth is one parameter of development. In college courses, we try to explain what it is and how it comes about. We provide students with extremely precise the- ories of economic and quantitative growth, of the sources of increases in income (the so-called Y), and we insist that the development process includes long-term struc- tural, institutional, and cultural change. As policy makers we understand that things are slightly more complicated. One aspect of the complexity is the appropriate focus of growth: should economic development be primarily about producing more, with all the consequences this may imply for the environment for instance? Or should the main objective be to share the benefits of growth? These questions are not addressed in the same manner in all countries, because they depend on philosophical doctrines. Another aspect of the complexity of the issue is the so-called tradeoff between growth and poverty reduction. I have some strong views on this. I find a lot of confu- sion in the proposition that one can separate poverty reduction strategies from the quest for growth. How can we sustain our fight against poverty (even defined, too simplistically for my own taste, as the proportion of people living on less than US$1 a day) if total income is not increased? It seems to me that there cannot be a long-term increase in per capita incomes if growth is not the centerpiece of the poverty reduction strategy. The consensus at both the World Bank and the International Monetary Fund seems to be that far from being at odds with each other, growth and poverty reduction are complementary objectives of public policy. They should be pursued simultaneously. A more challenging aspect of the issue, at least in my opinion, is the fact that the fight against poverty is not simply an economic matter but also a moral challenge. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 9 10 | ABDOULAYE WADE Africa is lagging behind on all indicators of economic growth and development. Yet it is being forced to operate in an international economic environment in which the rules governing economic competition are unfair. One example of such unfair- ness is the harmful effects on African economies of agricultural subsidies granted by developed countries to their farmers. I recently wrote an op-ed piece in Le Monde (September 10, 2003) in which I argued that Africa should not be considered a mere "adjustment variable" in the globalization equation. We Africans are not willing to be passive actors in the globalization debate. We want to be active participants, to shape the world. We would like to be able to express our own views in international forums and to influence the outcome of the discussion. We have done so in recent years thanks to our strong alliance with other Southern countries. A good illustration of our ability to influence the debate occurred during the development discussions in Cancun, which may not have been a shining success but at least did not boil down to the traditional business-as-usual talk on globalization and economic development. Africa's willingness to be more engaged in the debate in Cancun marked a new era of more intense participation by our policy makers in international negotiations on economic matters. In my September 2003 op-ed piece, I wrote, "It is my deep conviction that improper subsidies to the most competitive agricultural productions [in industrial countries] ruin most of the farmers in the so-called Third World." I also said that I support free and fair trade. That is where a major problem with globalization lies. I would even call it the great tragedy: policy makers in developed countries do not believe that what is good for them can also be good for us! We are being told to adjust to arbitrary decisions presented as generous principles of global free trade and free exchange. Unfortunately, these principles are too often dictated by great powers and violate our right to bring our own ideas to the table and be part of the global conversation. Such an attitude from developed countries leaves no prospect for Africa's economic development. Fortunately, there is now a large coalition of people of good will from all backgrounds (North, South, development experts, policy mak- ers, academics, and so forth) working hard to ensure that Africa's voice in the glob- alization debate is heard and taken into consideration. The issue of free and fair trade is extremely important to African countries, espe- cially for products such as cotton, the main source of livelihood for millions of peo- ple across the continent. Despite their importance, the current difficulties should not discourage those of us who still believe in globalization. I say this because there is a large group of anti-globalization protesters out there who have lost faith in international cooperation. While I respect them very much and even share their goals, I disagree with their methods and strategies. I used the following metaphor at the G-8 summits in Italy and in Evian: globalization is like a moving global train. We must get on it, even in the last car, even in third class. It would still be better than being left on the sidelines. That would be unacceptable. We must face the globalization challenge and fight for our interests. Objecting to globalization is a purely intellectual position, which a responsible policy maker in charge of resolving concrete development problems cannot adopt. It would be suicidal. The appropriate stance is to be part of the world we are trying to AFRICA'S GROWTH CHALLENGE IN A GLOBALIZED WORLD | 11 transform. And as I indicated earlier, we have some strong partners in this endeavor. I hope we find additional partners among participants to this meeting who can help us refine our strategies and approaches for diversifying our economies and for get- ting paid a fair price for our commodities in international markets. This being said, we must acknowledge that some serious impediments--wars and conflicts, disease, the lack of democracy--make it difficult for Africa to seize the eco- nomic opportunities of globalization. These factors prevent the continent from reach- ing its production potential. They also explain, at least partly, Africa's pervasive poverty. Fortunately, there are some positive signs. The fragile peace accord reached in Sudan after a 20-year civil war is one such sign. That accord, together with oth- ers, can lead to a major expansion of production possibilities in Africa. Indeed, we could draw on the development experience of industrial countries, which used to rely on agriculture. It is important to bridge what I call the agricultural divide. That is the topic of a conference to be held in Dakar beginning February 4. I am convinced that science and technology and knowledge transfer could be adjusted to Africa's needs and circumstances in a way that would help raise the productivity of the African farmer. Globalization implies competition, which in turn implies productivity gains. This can be obtained only through knowledge and technology. This brings me to the challenge of financing economic growth in Africa. So far, African countries have relied on their own domestic efforts and on external aid to finance growth. Unfortunately, external aid has often led to large, unsustainable indebtedness. It often happens that human endeavors based on good intentions bring about negative effects in the long run. We are now trying to address the debt issue through various instruments, such as the Heavily Indebted Poor Countries (HIPC) initiative. There is now a global consensus on the need for debt cancellation. Prime Minister Tony Blair of the United Kingdom and President Jacques Chirac of France, among other world leaders, have publicly supported the idea. I have been given the honor of preparing a major African seminar on debt issues. This meeting will pro- vide a good opportunity for us to express our views on that topic. I know there is some debate over the whole issue of aid and its effectiveness for growth in Africa. I do not want to start that discussion now, but my own view is that international aid and external resources in general (from public or private sources) have been very important to support growth on the continent. At the same time, however, in some countries external aid has created debt overhang that clearly impedes economic progress. When debt levels reach 80 percent or more of gross domestic product, there is little room for economic growth--or for anything else for that matter. Of course, the debt issue can be put in the context of the general pattern of asymmetric North- South relationships and the deterioration of the terms of trade. How do we approach economic growth in Senegal? We try to ensure that the ben- efits of prosperity are shared by all groups of the population. Focusing on the social aspects of the growth challenge, we have decided, for instance, to increase wages in the public sector twice in recent years. The private sector has followed the move, vol- untarily increasing salaries by 8 percent. This demonstrates good cooperation on eco- nomic policy between the government and the business community. It also confirms that the private sector is not necessarily opposed to improving workers' conditions. 12 | ABDOULAYE WADE To remain in business, they must find the right balance between the demands of a very competitive world and the need to maintain the welfare of their workforce. For- tunately, here in Senegal there is a clear harmony between the public and the private sectors on matters of economic policy. We are devoting a significant share of our financial resources to what is often called social benefits--namely, education. To me, education is not a social benefit but an investment. Senegal is currently one of the few countries in the world in which 40 per- cent of the budget is allocated to education. And we would like to do even more. We believe that we should give priority to education and training for a certain period of time, just like Japan and other Asian countries did. Once we have developed a strong human resources base, we will move to other policy areas, such as agriculture or infra- structure, which are not neglected today but are not the number one priority. Let me now turn to the second topic on the agenda of the conference: regional integration as a way of fostering growth and taking full advantage of globalization. It is a major preoccupation in Africa. Fortunately, we now have the New Partnership for Africa's Development (NEPAD), which is the appropriate instrument to deal with regional integration. It has little to do with nation-states and is designed on a regional or continental rationale. In my mind, it should deal only with regional projects. I know that this initiative is not always well understood, even among African leaders. It sometimes happens that an African head of state is asked about NEPAD and makes the following answer: "President Wade is pushing us through NEPAD to give priority to education. Yet I would like to give preeminence to a different sector." Well, such answers are beside the point. Each country should feel free to focus and sequence its development priorities the way it sees them. Let me stress the fact that NEPAD is not intended to become a substitute for national projects. The vision from its initiators was to design and implement cross-country, transnational projects. NEPAD suggests that Africa should focus on three major parameters. The first is good political governance. There is no need for me to insist on this point. You all know that it has become indispensable not only for the management of our own resources but also to create the domestic conditions needed to attract foreign capital. I would simply break it down into good public governance on the one hand, which is mainly about democracy, human rights, and everything else within the power of the state, and good private governance on the other hand, which is about creating a business-friendly environment. On this point, I would emphasize the need for a well- functioning judicial system and the fight against corruption as prerequisites for secure investment conditions. The second parameter in the NEPAD framework is the role of the private sector, which should be given priority in any development strategy. In all sectors of the Sene- galese economy, we are in the process of transferring managerial responsibilities from the public sector to the private sector. I am fully aware that these privatizations do not always occur under the best conditions, even though we do our best to have the most open and transparent process. The underlying problem here--and this brings me to the third parameter of NEPAD--is the weak human resources base on the con- tinent, or to put it differently, the lack of adequate training. We observe it every- AFRICA'S GROWTH CHALLENGE IN A GLOBALIZED WORLD | 13 where, including in the smallest market transactions: people buy and sell goods with- out good knowledge of cost and value. Check this out the next time you go to almost any store here in Dakar. Unfortunately, those are the current conditions and "rules" under which we must operate. On the basis of these three parameters--good governance, private sector develop- ment, and capacity building--NEPAD has offered an action plan, which covers eight priority sectors: infrastructure, agriculture, education, health, information and com- munication technologies, energy, access to markets of developed countries, and the environment. Of course, we could have added other important ones, but we chose to focus on these eight. We have started implementing regional projects in these sectors. In agriculture, for instance, Senegal and Mali are building on their common history, the Niger River, and their common land to implement a project that would reduce the cost of some imports currently paid in dollars, a step that will help our balance of payments. This is an example of NEPAD as a true instrument of integration. If African leaders are serious about advancing the regional integration agenda, they should develop cross-national projects in the eight NEPAD sectors. Since 2002, several conferences have been held throughout the continent to iden- tify interesting opportunities. A list of potential projects currently exists, covering all regions. I believe, wrongly or rightly, that the financing has been secured. Our devel- opment partners are committed to financing these projects. Unfortunately, as always, the major challenge is at the implementation level. I seize this opportunity to thank the World Bank for putting at our disposal a team of experts that is helping us imple- ment some of the projects. But at the continental level, the problem has not yet been resolved. Let me be clear: there are good prospects for economic integration in Africa. The challenge is to work simultaneously at several levels, starting at the subregional level. The African Union and the United Nations have divided the continent into five major regions: Central Africa, East Africa, Southern Africa, West Africa, and North Africa. In each of these five regions, there are interesting initiatives for regional integration. Examples include the Organisation de mise en valeur du fleuve Sénégal (OMVFS), a good example of cooperation in building hydraulic infrastructure; the road from Casablanca to Dakar through Mauritania, an outcome of cooperation by Maurita- nia, Morocco, and Senegal; the Algiers-Bamako road; and the Dakar-Djibouti rail- road. One former U.S. official said that "transportation is not only about concrete, asphalt, and steel. It is about connecting people to people and people to services. It is about creating links between communities and promoting the integration of regional blocs and markets. It is about creating jobs and economic opportunities. Transportation is thus the tie that binds, links, and integrates." I read this quotation to make you aware of my own faith in the importance of infrastructure in economic development. Indeed, development is impossible without roads, railroads, and air- ports. In countries where there is no transportation infrastructure, farmers cannot bring their agricultural products to domestic and foreign markets. It may sound triv- ial, but bumpy roads affect the quality and price of export products, which in turn affects competitiveness. 14 | ABDOULAYE WADE I should also mention as good examples of regional integration the West African Economic and Monetary Union (WAEMU), a framework for financial integration of independent states. It was an extraordinary gamble, and I congratulate its initiators-- my predecessors--for making that bet. Despite slow progress on economic and finan- cial integration, it is a great success. On the political front, we have organizations such as the Economic Community of West African States (ECOWAS), which are effective instruments for regional integration. This reinforces my point that the key constraint on African development today is not the lack of ideas. It is not even the lack of financial resources. This may seem like a paradox, but that is my view. I almost had a heart attack when I learned that Sene- gal uses only about 20 percent of the World Bank credits for which it is eligible. I thought the number was much higher than that. While I understand that there are other, more significant indicators to take into account when assessing our relation- ship with the Bank, I believe that this is an illustration of a larger problem: resources are often available, but the absorption capacity in Africa is insufficient. It is impor- tant to explore the reasons behind this constraint. They vary from weak project analysis to delays in administrative and disbursement procedures to our inability to negotiate with donors the optimal level of conditionality for the quick use of exter- nal financing. Moreover, even when financial resources are available, we do not always manage them efficiently. This is also true in the private sector: the banking sector in most WAEMU countries has a surplus of liquidity but few investment opportunities. Another paradox is that African institutions such as insurance compa- nies are investing billions of dollars in developed countries while we are looking for money here at home. That is the real problem of development financing. African economies are in need of a small group of visionary people who can find creative ways of attracting and using these large, dormant savings. I would like to end my remarks by linking African growth and regional integra- tion in the context of a globalized world. There is a price to be paid for regional inte- gration. Efforts should be made in each country to shift from a national vision of eco- nomic development to a regional or continental vision. I have proposed launching projects that would connect Africa to the developed world, in particular in the area of information and communication technology. One example of such a project is the Basic Digital Solidarity project, designed in the framework of NEPAD, to be initiated March 14, 2005. The idea is to try to connect countries of the South with those in the North so that the former can benefit from knowledge transfer. Specifically, the project will equip African countries with various new technology instruments for self-education, e-learning, capacity building in the public sector, and training mod- ules for schools and colleges. This project may be the most important for spurring economic growth and development in Africa. It is a win-win venture for everyone involved: developed countries will sell their hardware to African countries, and we will benefit from it. I take this opportunity to thank all the Latin American, Asian, and Arab countries that supported this idea from the beginning. I am very much encouraged by other exciting North-South initiatives, such as the African Growth and Opportunity Act (AGOA), which provides unlimited and cus- AFRICA'S GROWTH CHALLENGE IN A GLOBALIZED WORLD | 15 toms-free exportations from certain African countries to the United States. I should point out that Senegal has not been able to take advantage of this opportunity. Here again, money is not everything: in order to export, you have to be able to produce, but producing requires infrastructure and investments. There is also the Millennium Challenge Account, which makes financing available for certain countries. I am pleased to note that Senegal is eligible for this funding for 2004 and 2005, which will allows us to invest in major infrastructure projects. I hope my message has been clear: the problems of economic growth and regional integration--to put it differently, the problem of African development--must be seen as one of shared responsibility. People of good will from the North and the South should come together and rise to the challenge of fighting poverty. I thank you for your attention and wish you great success in your work. Opening Address Equity and Growth in Africa FRANÇOIS BOURGUIGNON Your Excellencies President Abdoulaye Wade and Minister Abdoulaye Diop, dear colleagues. First, let me thank the government of Senegal for hosting this year's regional Annual Bank Conference on Development Economics (ABCDE), our Global Development Network colleagues for allowing us to free ride on their annual con- gress, and all the speakers and participants at this conference. I am sure you would agree that the ABCDE conferences provide unique opportunities for a closer dialogue between development economists and policy makers in developing countries and their colleagues in industrial countries and international organizations, including the World Bank. The great idea to have these encounters in developing regions rather than in Washington came from my predecessor, Nick Stern. It allows us to involve thinkers and policy makers in the various developing regions more deeply in the global conversation about development, and also to ensure that region specificity is taken into account in our discussions. The first of these regional ABCDEs took place two years ago, in India. I am happy that this year's conference is taking place in Sene- gal, a country that is exemplary in many respects, with its strong and ambitious lead- ership, which has being playing a critical role in the New Partnership for African Development (NEPAD); strong institutions; solid academic structures; good develop- ment policies; and very encouraging performance during the past decade. The theme for this conference, growth and integration, should not come as a sur- prise to anyone. The evolution of the global distribution of income is characterized by two important trends: convergence and divergence. A few countries that were initially low-income countries have converged toward the standards of living of industrial regions to become middle-income countries. But many low-income countries, partic- ularly in Sub-Saharan Africa, have diverged from the mean global income. Such a divergent evolution is clearly not sustainable in the long run: it goes against the most basic equity principles and, as many political leaders recognize, is bound to jeopard- ize in one way or another political stability in the world. The case of Sub-Saharan Africa is particularly worrisome: between 1970 and 2003, the average annual rate of François Bourguignon is senior vice president, Development Economics, and chief economist at the World Bank. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 17 18 | FRANÇOIS BOURGUIGNON growth of real income per capita was ­0.12 percent, compared with 2.6 percent in South Asia and 5.6 percent in East Asia. Several factors--including devastating conflicts and the deterioration and instabil- ity of the region's terms of trade since the early 1980s--may explain the slow growth of Sub-Saharan Africa (see Collier and Gunning 1999). But we must also acknowl- edge that even in the more stable Sub-Saharan African countries, and in the rest of the developing world as well, growth is something of a mystery. The hopes expressed by some researchers a few years ago have not been realized: we have not yet discov- ered the magic recipe for growth. One can even wonder if it is realistic to believe that such a recipe exists. The rather simplistic idea from various linear economic models according to which additional labor and physical capital are enough to generate growth in income and productivity has proved wrong, especially in Africa. The good news is that economic analysis and research on growth over the past 10­15 years have yielded important lessons. One is that successful growth strategies must be highly context specific and analysis intensive. Drawing on lessons from a recent World Bank report (2005) and from studies by Rodrik (2004a, 2004b) and others, I would say that we now have some good ideas on general principles to guide growth strategies, but the prioritization of these principles and the way in which they may be practically implemented require difficult and essentially country-specific analyses, which we have not yet mastered. This new approach should be the main theme of our collective reflection today. With your permission, I will elaborate a lit- tle more on this during the remainder of my time this morning. Let me suggest two main research paths that, interestingly enough, follow closely the two-pillar poverty reduction strategy advocated by the World Bank over the past few years: growth and the investment climate on the one hand; empowerment of all, or equity, on the other. I would like first to emphasize the complexity of the relationship between the characteristics of the investment climate and growth and to stress the need to iden- tify in each country the true constraining factors to growth. Clearly, there is comple- mentarity among the determinants of growth, particularly in Sub-Saharan Africa and other low-income regions. Investing in new capital cannot be done without an ade- quately trained labor force, agricultural productivity cannot be raised without invest- ing in rural infrastructure, private investments cannot be dynamic without the appro- priate institutional environment, and so forth. From an analytical perspective, I consider this complementarity to be a very serious difficulty in understanding the roles of the various potential determinants of growth. From both an analytical and a policy point of view, complementarity makes it essential to identify what may be the binding factor of growth in a given economy at a given point of time. This identifi- cation process is obviously not easy, and analytical efforts should seek to facilitate the process. However, the same complementarity also makes it necessary to consider evolving growth strategies. The relevant binding constraints tend to change over time; it may be necessary to deal today with constraints likely to be binding tomor- row to avoid creating growth that is a succession of alternate short-lived surges and stagnation periods. EQUITY AND GROWTH IN AFRICA | 19 Another condition for sustained growth is equity, the second pillar of the World Bank's strategy. Here again I believe that there is strong complementarity in the long run. For well-crafted growth strategies to yield sustained growth--and not the ephemeral growth "episodes" observed in many African countries and elsewhere-- parallel progress must be made on the equity side of the two-pillar strategy. Empowering poor people in all dimensions of their social life is a necessary condi- tion for sustainable growth in the long run. At the same time, this empowerment is possible only if enough resources are available through economic growth. Let me briefly elaborate on these two propositions. First, it is important to empha- size the complex interactions among growth factors and to understand the need to identify binding constraints when designing growth strategies. A large body of research has been produced in recent years on growth determinants. On the theoret- ical side, growth mechanisms and channels through which various factors may affect growth have been extensively analyzed. This has led to a much better understanding of several important growth issues, particularly in situations that are inconsistent with the basic assumptions behind the textbook neoclassical growth model, à la Solow. Much attention has been devoted in particular to economies of scale, techno- logical changes, the role of trade, institutions, inequality of wealth, and so forth. On the empirical side, efforts have focused on measuring the role of various growth fac- tors, mostly by comparing growth performance and growth strategies across coun- tries, through linear econometric models estimated on a cross-section of countries, possibly observed in various time periods. The results have been disappointing: African policy makers looking for concrete prescriptions on how to generate economic growth in their countries would be sur- prised to learn that very few factors have been found to be systematically significant. Moreover, for those factors, the relationships obtained from econometric analysis are usually not strong enough to imply general validity. In an interesting paper published in 1997 under the provocative title "I Just Ran Two Million Regressions," Xavier Sala-i-Martin identified at least 60 variables that have been found to be significant in some part of the literature on growth. Would this imply that any growth strategy designed in Africa (or elsewhere) should include each of these variables, as well as those not yet found by economic analysis? Some potential growth determinants are absent from the list of so-called signifi- cant variables, or their effect is estimated imprecisely, sometimes even with a coun- terintuitive sign. Components of infrastructure capital fall in this category. Several factors deemed to be important, including property rights, are found to have only limited and generally very imprecise effects, although they are sometimes significant. Several factors explain why econometric analysis in this field is likely to yield weak results. The most important factor is the small number of observations. If long- run growth is the subject of analysis, the number of independent observations is essentially the number of countries, or the number of periods for each country, for which the appropriate data are available. This number does not permit a very sophis- ticated search for complex relationships between potential growth determinants and growth outcomes. 20 | FRANÇOIS BOURGUIGNON The most obvious reason for the relative weakness of econometric work on growth is probably the difficulty of using truly structural models to explain why some factors are strong determinants of growth in some instances but not in others. The challenge is to introduce such a complex interactive relationship in the analysis. The complementarity between growth factors provides a good illustration of such a complex interaction. A case in point is the controversy among economists on the importance of educa- tion for growth. Clearly, there is something odd in studies that do not find education to have been a key factor of growth. Does anyone believe, for instance, that indus- trial countries could have developed over the past century or so without any change in the level of education of their population? Some countries have overinvested in education without stimulating growth until a later stage of the growth process. The best-known example of this phenomenon is the Republic of Korea in the early 1960s. Some researchers suggest that African countries such as Ghana or even Cameroon in the 1970s have gone through the same process. To understand such a puzzle, one can think of a two-factor world in which education is complementary to capital. If we were able to identify specific moments in history when this ratio was too high or too low--and if we had good measures of these two types of capital--the rule would be straightforward: in countries in which the physical-to-human capital ratio is too high, education should be a highly significant determinant of growth and vice versa. Unfortunately, the value of the ratio is unknown and probably depends on countries and development stages. Of course, estimating the appropriate nonlinear specification of the growth model that this criteria suggests should allow us to see whether this interpretation of the weakness of econometric work on growth is valid. But there is a problem: because of data limitations and the large number of ways in which human-to-physical capital ratios may differ across countries, such estimation is practically impossible to con- duct. Against the background of these technical difficulties, this explanation, which is consistent with the results of simple linear econometric models, may help us dimin- ish our expections. After all, if education does not emerge as a significant factor in a cross-country regression, it may be because physical capital is more frequently the binding factor, not necessarily that education plays no role. If education does not emerge in a time series analysis, it may be because it has been less frequently binding than capital over time. One can easily generalize this argument to other growth determinants. For instance, infrastructures or institutions of various kinds could very well enhance or weaken the role of any factor of growth. In fact, contemporaneous development experiences clearly suggest that such complex interactions are very often at work. A good example is the role of property rights, correctly thought to be an important determinant of investment decisions all over the world. Yet the absence of formal property rights did not prevent the spectacular take-off of the Chinese economy, probably because the overall incen- tive system in place was more important. Likewise, one could argue that the mere exis- tence of a relatively well-functioning property rights system in countries such as Benin or Senegal has not been enough to accelerate growth, probably because other con- straints, yet to be identified, have not been removed. EQUITY AND GROWTH IN AFRICA | 21 The nonlinear view of growth that I am describing leads to one important conclu- sion: various factors of growth play different roles at different moments. Therefore, as in the simple complementarity case, some factors will be binding at a given time, whereas changes in some others will have a negligible impact on growth. Identifying those binding constraints, and anticipating some of them, should be the main objec- tive of growth analysis. Finding ways to remove constraints is the main challenge in designing growth strategies. Clearly, this should be a highly context-specific exercise, very different from a universal recipe for growth. A general roadmap for implementing this approach would start with a complete diagnosis of an economy in the light of the basic theoretical models of growth determinants: physical bottlenecks, multiple equi- librium situations, government failures, market failures, and ways to remedy them.1 One should remember that macroeconomic aggregates and some characterizations of institutions are not the only variables that matter; micro information is equally important, especially in the African context. Conventional analyses of growth are not to be ruled out. For instance, it is cer- tainly useful to know the rate of growth of the various factors of production so that one can observe whether one factor grew much less than the others, which makes it likely to become a binding constraint, depending on its initial ratio. But this is not enough, especially if one wants to take into account nonphysical binding constraints, which arise so often in Africa, in particular government or market failures. The approach I suggest is not necessarily inconsistent with universal recipes, it simply points to the fact that, in African countries and in any specific environment, some of these recipes may be more relevant to the quest for growth than others. I rec- ognize that my proposal is much more demanding analytically, but there are reasons to expect that its return will be higher. Let me now turn to my second proposition, that there is a strong complementar- ity between long-run growth and equity. This is one of the central messages of the 2006 World Development Report. Economists tend to stress the conflict between efficiency (growth) and equality of income or consumption in a well-functioning economy. They argue (often rightly) that redistributive policies that aim to decrease inequality are harmful to economic efficiency and growth. Typical examples include a tax policy that distorts prices and reduces investment incentives or cash transfers that reduce labor incentives. African countries have long suffered from such misguided fiscal policies, especially during the 1960s and 1970s, when hard socialism was in vogue in countries such as Benin, Ethiopia, Guinea, and the Republic of Congo. Efficiency and inequality need not work at cross-purposes: in the case of imper- fect markets, there may be situations in which redistribution increases efficiency. Consider, for instance, an imperfect credit market in which many people lack access to credit. This is the case in many African countries, where large sections of the busi- ness community lack access to banking services. Eifert, Gelb, and Ramachandran (2005) find that only 10 percent of microenterprises in Senegal have access to loans and overdraft--a much lower figure than in Bolivia (75 percent), India (42 percent), or Nicaragua (53 percent). Access to credit is low even in good performers such as 22 | FRANÇOIS BOURGUIGNON Tanzania (8 percent) and Uganda (7 percent). Because they do not have enough col- lateral, people who are excluded from the credit market cannot undertake investment projects, even though these projects may be highly profitable. By contrast, the lucky few who have access to credit or enough personal wealth will invest in projects whose return covers the rate of interest. The net effect of such a situation is overinvestment in less profitable projects. Broadening the credit market increases efficiency, even after factoring in the risk of default. Two policy options can help solve the problem: improving the functioning of the credit market (through subsidized microcredit, for instance) and providing more col- lateral to those excluded from the market by favoring capital accumulation in that group. The first solution is more feasible than the second. Both options would trans- fer wealth (over some period of time) from the well-off to the disadvantaged and increase the output of the economy. The general proposition that redistribution may be efficient when markets are imperfect could have a broad application. Yet the kind of redistribution involved in the preceding example is different by nature from standard income redistribution. Efficiency gains are obtained by equalizing access to economic options, either by guar- anteeing the same access to markets to all or by giving all people the means they need to access those markets. This is why I referred to "equity" rather than to "equality" in my statement of the general principles for development. I define equity as equality of opportunities rather than equality of results, such as income or consumption, as in the standard economic literature. More equity means more investment options avail- able to large sections of the population that were initially kept outside the economic sphere. With more and better investments of all kinds undertaken, more equity leads to faster, sustained growth. Access to infrastructure, for instance, is more skewed in Sub-Saharan Africa than in any other region, with the 30 percent of people living in urban areas receiving 80 percent of services. Yet we know that there is a strong rela- tionship between income and access to roads. In Niger, for example, the already very low road-to-population ratio declined from 2,800 kilometers per 1 million people in 1980 to an estimated 1,100 in 2000. If more people had access to roads (say, if this ratio had doubled or tripled, as in some other developing countries), it is likely that, all other things equal, output would have been higher in Niger. From both an equity and an efficiency point of view, it might thus be better to allocate marginal public spending to rural roads rather than to urban infrastructure. One can make the same argument about the potential output gains that African countries could obtain by reducing inequalities in the education and health sectors by facilitating access to these services for a larger share of the population. Tax policies designed to expand access to these sectors could very well be distortive in the short run, but this distortion may be compensated by efficiency gains in the long run, due to more equal access to the accumulation of human capital in the population. In this regard, Sub-Saharan Africa has some of the most disturbing social indicators in the world, some of them revealing both low average per capita income and a high level of inequity. Sub-Saharan Africa's Gini coefficient of consumption--an imperfect measure of income inequality--of 46 is higher than that of South Asia (31), despite comparable levels of average per capita income (US$500 in Sub-Saharan Africa EQUITY AND GROWTH IN AFRICA | 23 versus $510 in South Asia at official exchange rates; $1,750 versus $2,640 after cor- recting for purchasing power parity). Data from the World Development Indicators indicate that the gross secondary school enrollment ratio in Sub-Saharan Africa was half that of South Asia (22 per- cent versus 40 percent) in 1990. The immunization rate for DPT for children under one year was only 46 percent in Sub-Saharan Africa in 2000, much lower than the 57 percent in South Asia. Infant mortality, a good proxy for the health status of young children, was 50 percent higher in Sub-Saharan Africa than in South Asia (10.1 percent versus 6.6 percent). These statistics show that a large proportion of people in Africa are prevented from building their human capital. Other things equal, this will make them less productive when they join the workforce and could exclude them from exploiting investment and growth opportunities. In short, unequal access to human capital in Sub-Saharan Africa leaves many people there unable to con- tribute effectively to the growth of their economies. Both the general level of devel- opment and equity explain differences in social indicators. Another example of the efficiency loss due to inequity is property rights. Ambigu- ous property rights for land are bad for investment and productivity, and they steer economic benefits away from the rural poor. Evidence from studies conducted in Ethiopia by Deininger and others (2003) indicates that only 4 percent of landowners believe they have the right to sell their land and that eliminating the threat of redis- tribution of land and improving tenure rights and transferability could raise produc- tivity by 5.6 percent. Conflicts or political instability possibly resulting from inequity in a society, including access to land, may be a major obstacle to growth. Empirical evidence from Uganda suggests that productivity is higher in plots located in areas without conflict (Castagnini and Deininger 2004). One would expect the same con- clusion from economic studies on the recent conflict over land tenure among large communities in Côte d'Ivoire. Last but not least are gender inequalities. Women provide most of the labor in many countries, but unequal access to land, credit, and political rights severely reduce productivity and growth. Some studies suggest that if Africa's women had equal access, growth across the continent would rise 0.8 percent. Other studies point to women's ability to achieve much higher values of agricultural output per hectare than men, on much smaller plots (Udry 1996). Let me briefly conclude by reiterating that economists know more or less the basic mechanisms that may generate growth in the short term, especially in countries in which binding constraints can be clearly identified. This does not mean that they know which policy instruments can slacken these constraints. This does not mean that they have the solution to the more serious problem of how to sustain growth in the long run. Increasing inequity may not necessarily affect short-term growth spurs, but it is incompatible with sustained long-term growth. There is a synergy between the two pillars of the strategy of poverty reduction advocated by the World Bank: growth through reforms to improve the investment climate and equity. Sustaining higher rates of growth in African countries--a prereq- uisite for poverty reduction--will require that the large segments of the population currently outside the development process be adequately equipped to participate in 24 | FRANÇOIS BOURGUIGNON public policies in a meaningful way and to seize the economic opportunities available to them. This requires a more equal distribution of existing economic opportunities and an expansion of those opportunities thanks to the additional resources created by growth. Failing to meet this important condition will maintain Sub-Saharan Africa in a low-income or slow-growth trap. Note 1. Since presenting this lecture, I became aware of the growth diagnostics methodology pro- posed by Ricardo Hausmann, Dani Rodrik, and Andres Velasco (2004). This methodol- ogy, based on identifying binding constraints through their shadow prices, represents a step in the right direction, although estimating shadow prices is likely to be difficult. Pilot studies using this methodology are currently under way at the World Bank. References Castagnini, Raffaella, and Klaus Deininger. 2004. "Incidence and Impact of Land Conflict in Uganda." Policy Research Working Paper 3248, World Bank, Washington, DC. Collier, Paul, and Jan Willem Gunning. 1999. "Why Has Africa Grown Slowly?" Journal of Economic Perspectives 13 (3): 3­22. Deininger, Klaus, S. Jin, B. Adenew, Hsg Selassie, and B. Nega. 2003. "Tenure Security and Land-Related Investment: Evidence from Ethiopia." Policy Research Working Paper 2991,World Bank, Washington, DC. Eifert, Benn, Alan Gelb, and Vijaya Ramachandran. 2005. "Business Environment and Com- parative Advantage in Africa: Evidence from the Investment Climate Data." World Bank, Research Department, Washington, DC. Hausmann, Ricardo, Dani Rodrik, and Andres Velasco. 2004. "Growth Diagnostics." John F. Kennedy School of Government, Harvard University, Cambridge, MA. Rodrik, Dani. 2004a. Growth Experience: Lessons from the 1990s. Washington, DC: World Bank. ------. 2004b. "Growth Strategies." John F. Kennedy School of Government, Harvard Uni- versity, Cambridge, MA. Udry, Christopher. 1996. "Gender, Agricultural Production and the Theory of the House- hold." Journal of Political Economy 104 (5): 1010­46. World Bank. 2005. Economic Growth in the 1990s: Learning from a Decade of Reform. Washington, DC. Xavier, Sala-i-Martin. 1997. "I Just Ran Two Million Regressions." American Economic Review 87 (2): 178­83. Keynote Address The Importance of Research for Economic Policy: The Case of Senegal ABDOULAYE DIOP Thank you, Mr. President, ladies, gentlemen. I will articulate my remarks on two issues rather quickly, because the time allotted to me is rather short. I will first bring you up to date on recent developments in Senegal's economic situation and prospects. Then I will come to the theme of my speech. Over the past few years, Senegal has recorded very significant macroeconomic results, which have been recognized by the international community. The country experienced a growth rate of about 6.5 percent in 2003 and 6.1 percent in 2004, while price stability and the fiscal deficit have been kept at sustainable levels. In addi- tion, the return to growth has enabled us to reduce the percentage of the population living in poverty from 67.9 percent in 1994­5 to 57.1 percent in 2001­2--a decrease of about 10.8 percent, if the poverty indicators and results from the recent house- holds censuses can be trusted. The magnitude of poverty still remains very high, how- ever, affecting more than half the population. At the very least, the continued growth that the economy has enjoyed since the January 1994 devaluation allows Senegal to view wealth creation as the first lever in its poverty reduction strategy. Thus, you can see why particular emphasis is placed on Senegal for the implementation of a growth acceleration strategy. This strategy will certainly benefit from the positive results of policies aimed at creating a healthy macroeconomic framework and an incentive-based environment. Structural reforms, notably the implementation of the national good governance program and the imple- mentation of reforms in public finance and public contracts, seek to increase effi- ciency, effectiveness, and transparency of public actions to facilitate the implementa- tion of the private sector development strategy and the process of wealth creation. That said, the importance given to promoting the private sector in economic development forces us to refine our macroeconomic management practices and develop our knowledge of the microeconomic dimensions of growth. This means that this conference in Senegal is well timed. Because public finances are becoming viable Abdoulaye Diop is the Minister of Finance and Economy of Senegal. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 25 26 | ABDOULAYE DIOP again and a lasting environment for growth is being maintained, the state needs to know more and more about the evolution of inequality, about operators' reactions to the various policies that have been put in place, starting with microeconomic fac- tors and hindrances to growth. It is no surprise that these questions have already been raised by economic researchers in Senegal, in universities and economic admin- istrations. Research centers have been created within economics departments and in public agencies. These centers have contributed to the development of programs like the Poverty Reduction Strategy Paper (PRSP), the program for education and train- ing, and the private sector development strategy. The state has assigned these centers the mission of performing economic and social research with the goal of increasing our understanding of the problems of development in Senegal and in Africa and of clarifying the decision-making process. Previously, support from development part- ners paved the way for significant improvements in the allocation of public invest- ments, the development of macroeconomic and macro-financial model management, and the promotion of information technology tools within public and economic administrations. The scope of the work undertaken by the centers has been enlarged to include the PRSP and the national good governance program. Among the insti- tutes and other research centers that are most prominent are those that have a regional or continental mission as well as projects based in Dakar or in other African capitals where the work program covers Senegal or requires Senegalese expertise. The sources of inspiration for participants are multiplying, which could be an advan- tage at a time when dialogue and participation occupy an ever more important place in the process of public decision making. By way of illustration, I will highlight the many exchanges between local partici- pants that have allowed the poverty reduction strategy to become an approach largely shared by the population and decision makers. Among the many points of consensus that these exchanges have established is the idea that growth is necessary but not sufficient to reduce poverty. To reduce poverty, growth must be brought to the people who are affected by poverty. Growth must reach all social classes, which must have the ability to participate in the creation of the country's wealth. As obvi- ous as these points may seem, there still has not been a meeting of the minds on them, which may delay the achievement of expected results. To be armed against these risks, it is important that we include activities and safeguards in these research and training programs to ensure, among other things, the development of specific abili- ties for strategy implementation development in the fight against poverty and inequality. It is also important to create a national database of social indicators based on statistics and surveys on living conditions, health, household spending budgets, demographic studies, and so forth. Doing so will allow researchers to respond more easily to requests by decision makers and to validate the results of their research. We must also strengthen collaborative ties between researchers and government agencies in charge of data production, thereby improving the relevance of economic and social information and facilitating its circulation and use. The process of accu- mulation is often disrupted by the "brain drain" phenomenon. We are experiencing this in Senegal, where many researchers have emigrated to other countries. Existence THE IMPORTANCE OF RESEARCH FOR ECONOMIC POLICY | 27 in the name of sufficient capacity seems to be the nexus of the problem of questions faced by many developing countries, especially in Africa. If a country on the list does not have such capacity, I believe that adopting policies of reform or creating them will always prove difficult. Growth and Integration Explaining African Economic Growth: The Role of Anti-Growth Syndromes AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL African governments actively influenced the incentive environment for economic growth between 1960 and 2000. Distortionary control regimes, inefficient ethno-regional redis- tribution, unsustainable spending booms, and state failure feature prominently in case evidence on 27 countries, as does the absence of these excesses, a situation referred to as "syndrome free." Policy matters tremendously: separately or in combination, anti- growth syndromes reduce predicted growth by 1.0­2.5 percentage points a year. Syndrome-free status emerges as virtually a necessary condition for sustained rapid growth in Africa and virtually a sufficient condition for avoiding short-run growth col- lapses. The political and economic reforms put in place by the mid-1990s bode well for Africa's long-run growth. Africa's growth record after 1960 confounded the early expectations of many in the development community. Little was expected in the 1960s of tiny Botswana and Mauritius, the continent's great performers over the 1960­2000 period; Mauritius was gently ushered into a disastrous phase of import-substituting industrialization by the Fabian socialist James Meade--later a Nobel laureate in economics--as the only plausible means of employment creation in a poor, remote, ethnically fractious, and human-capital scarce economy (Meade and others 1961). Hopes were much greater for Nigeria, the mineral-rich countries of Southern Africa, and countries with high agricultural potential, such as Kenya, Sudan, and Uganda, where the continent's resource abundance was expected to be harnessed to achieve national development aspirations. It was even hoped that growth could accelerate in the landlocked coun- tries with low agricultural potential, where aid inflows, the falling costs of interna- tional travel, and the possibility of labor out-migration appeared to offer promising new directions for growth strategy (Karmarck 1971). Augustin Kwasi Fosu is director of the Economic and Social Policy Division at the UN Economic Commission for Africa, in Addis Ababa; the views expressed here are not attributable to the UN. Stephen A. O'Connell is professor of econom- ics at Swarthmore College, in Swarthmore, Pennsylvania. The authors thank Matt Meltzer for research assistance, Jean- Paul Azam for his remarks, and Benno Ndulu, Patrick Guillaumont, Cathy Pattillo, and Jeffrey Williamson for their insightful comments. They also gratefully acknowledge the influence of Growth Project researchers and steering and edi- torial committee members Olusanya Ajakaiye, Jean-Paul Azam, Bob Bates, Paul Collier, Jan Gunning, Benno Ndulu, Dominique Njinkeu, and Charles Soludo on the ideas presented here. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 31 32 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL Many African economies experienced at least moderate growth in the 1960s, but by the late 1970s the promise of political independence had given way to economic crisis throughout most of the continent (see annex table A1). While modest cumula- tive progress occurred in human development, per capita incomes stagnated or fell in much of Africa starting in the early 1970s, reviving on a continentwide basis only in the second half of the 1990s. Poverty headcount ratios worsened over the full 1960­2000 period. Africa's long-run growth shortfall--relative not just to expectations but also to the experience of other developing regions--remains a central preoccupation of the cross-country growth literature. Regression evidence confirms the critical importance of policy and institutional variables that capture the quality of governance, as well as of variables such as landlocked status, tropical location, and disease burden that cap- ture the hostility of Africa's geography (Collier and Gunning 1999). This paper focuses on patterns of governance. Drawing on 27 detailed country studies of the political economy of African growth, it characterizes the dominant pat- terns of African economic policy between 1960 and 2000, explores their implications for growth, and examines the dynamics of their adoption and abandonment. The study follows the suggestion of Temple (1999), who outlines the limitations of cross- country growth econometrics and makes a convincing case for the complementary role of intensive case analysis. The next section describes the Explaining African Economic Growth Project, designed to produce the first major, comprehensive assessment by African research economists of the continent's growth experience since independence. Section 2 begins by underscoring the episodic nature of African growth. It then introduces the four anti-growth syndromes, tracks their evolution over time, and assesses their impact on predicted growth within the African sample. The third and fourth sections draw directly on the case study evidence, first to illustrate the syndromes and then to study the forces behind their adoption and abandonment. The last section provides some observations on complementary elements of pro-growth strategy in Africa. The AERC Growth Project The Explaining African Economic Growth Project (henceforth simply "the Growth Project") was conceived in 1997 as a collaborative effort of Harvard University, Oxford University, and the African Economic Research Consortium (AERC). It was formally launched with the presentation of framework papers at a Harvard Univer- sity workshop in March 1999. Until the mid-1970s, many African economies grew rapidly. Growth decelerated substantially thereafter, both in absolute terms and relative to other regions. Although reforms produced a rebound in some countries as early as the mid-1980s, sustained growth has yet to sweep the continent or to make decisive inroads in reversing a legacy of stagnation. The Growth Project seeks to explain this growth record by combining global evi- dence on the determinants of growth with country-based work on the microeco- nomic behavior of firms and households, the organization of markets, and the polit- EXPLAINING AFRICAN ECONOMIC GROWTH | 33 ical economy of policy. In a two-stage approach, research teams first used cross- country regression models to place each country's growth experience in comparative perspective. Separately, each team then divided the 1960­2000 period into country- specific episodes--periods of a few years' duration to a decade or longer during which the incentive environment for growth had an identifiable and reasonably con- sistent character. The bulk of the research occurred at the country level, where the task was to marshal evidence on why the proximate determinants of growth evolved as they did. Growth episodes that were poorly captured by the first stage motivated a search for country-specific mechanisms omitted from the cross-country model. The two stages of analysis disciplined and informed each other, producing unified and comparable accounts of individual-country experience and generating new insights at both the country and cross-country levels. Guided by this methodology, a committee of African and non-African scholars has interacted with 27 country-based research teams since December 1999, in a cumulative process of technical training, country research, and peer review. Coun- tries were chosen with a view to achieving a structurally diverse and regionally bal- anced sample, with some emphasis given to economic size (table 1). While the sam- ple covers the bulk of Sub-Saharan African countries, some bias likely remains due to the under-representation of countries in which armed conflict or state collapse have left major gaps in the historical record. The Growth Project has established a network of African growth scholars and generated a rich store of analytically comparable country material. The challenge in synthesizing this material is to identify lessons of broad and fundamental relevance while doing justice to the diversity of country experience. To this end the editorial committee developed a two-way taxonomy of opportunities by choices. For the region as a whole and country by country, what were the key growth opportunities between 1960 and 2000? What decisions led to their so often being missed? Why were these decisions, rather than more growth-enhancing ones, made? Where do the main growth opportunities and constraints now lie? This paper focuses on choices, in particular on ways in which African govern- ments have actively shaped the incentive environment for private economic activity. TABLE 1. Country Cases in the Growth Project, by Subregion and Opportunity Group Coastal, resource- Landlocked, resource- Resource-rich Subregion poor countries poor countriesa countriesb Eastern and Kenya Burundi, Chad, Ethiopia, Cameroon, Rep. of Central Africa Niger, Sudan, Uganda Congo Southern Africa Mauritius, Mozambique, Malawi Botswana, Namibia, South Africa, Tanzania Zambia West Africa Benin, Côte d'Ivoire, Burkina Faso, Mali Guinea, Nigeria, Sierra Ghana, Senegal, Togo Leone Source: AERC Growth Project. a. Ethiopia and Sudan are classified as landlocked owing to their geographies even though they are not explicitly landlocked, except for Ethiopia after 1993. b. Includes all countries classified by Collier and O'Connell (2005) as resource rich for some portion of the 1960­2000 period. 34 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL The message is neither novel nor easy, but it is grounded both in the cross-country literature and in the case study evidence. It is that the quality of government policy-- or more broadly, governance--mattered a great deal for African economic growth over the post-independence period. Avoiding a well-defined set of anti-growth syndromes-- control regimes, adverse redistribution, unsustainable spending booms, and state failure--increased growth by 1.0­2.5 percentage points in African countries. Main- taining "syndrome-free" status was virtually a necessary condition for rapid growth, and it was virtually sufficient for avoiding the growth collapses that so often under- mined sustained progress in Africa. Syndrome-free status was not sufficient, however, for achieving rapid growth. Indeed, governments managed to choose largely pro-growth policies in roughly a third of the episodes identified in the country studies. This pattern contrasts with the pejorative characterization of African governance that is so common in the literature. While pro-growth policies tended to secure moderate growth, sustained rapid growth was rare. The challenge, in these cases, is to explain the relative importance of struc- tural constraints, external shocks, and country-specific features in accounting for the lack of rapid growth in most African countries. Where growth opportunities were favorable but countries nonetheless slipped into anti-growth syndromes, the evidence often suggests a role for imperfect information with respect to policy choice, in conjunction with the poor quality of domestic polit- ical institutions. State failure is, of course, not a policy choice, but even here in many cases the choices of political incumbents exerted a powerful effect on the degree to which patterns of ethno-regional polarization inherited from the colonial period were defused or emerged in open conflict. The case studies provide support for the view that the market-based reforms and improvements in public sector management that got underway in many countries in the mid-1980s established a strong basis for renewed growth among politically sta- ble African countries during the turbulent 1990s. Where reforms or political stabil- ity failed, however, so did growth. Looking ahead, the key to achieving sustained economic growth and poverty reduction is to develop a political economy capable of preserving political stability while maintaining a credible commitment to market- based economic reforms. Anti-Growth Syndromes The Growth Project and this study focus on Sub-Saharan Africa,1 which has per- formed differently from North Africa.2 The stylized facts of African growth are well known (see, for example, Collier and Gunning 1999). Although there is considerable cross-country diversity, certain trends can be observed (Ndulu and O'Connell 2005): · Slow long-run growth in per capita income by global standards, with roughly equal contributions from relatively low investment rates and slow growth in total factor productivity. EXPLAINING AFRICAN ECONOMIC GROWTH | 35 · Extremely high fertility rates and population growth rates, resulting in high age- dependency ratios. · A deterioration in relative growth performance in the mid-1970s, with some rebound relative to the world as a whole after the mid-1990s. · Slightly better performance of human development indicators compared with real incomes over the full period, but with insufficient progress in these categories to prevent divergence relative to other low-income regions, such as South Asia. · Limited structural transformation, with the bulk of the labor force remaining in agriculture and exports tending to remain concentrated in a few primary commodi- ties. The focus on primary commodities was heightened by the emergence of new mineral exporters starting in the early 1970s and continuing through the late 1990s. · A persistence of the medium-term volatility characteristic of low-income countries worldwide. As indicated in annex table A1, periods of rapid growth were not uncommon in Sub-Saharan Africa, and rapid growth was sustained over long periods in Botswana and Mauritius. Sustained moderate growth was observed in a number of countries, including Ghana and Uganda, since the mid-1980s. But the central puzzle of the 1960­2000 growth record is the failure of the vast majority of African economies to achieve sustained increases in incomes.3 This failure must ultimately be traced to some combination of adverse growth opportunities and unsuccessful policy choices, where choices refers broadly to the full set of economic roles undertaken by the state, as producer, consumer, and regulator of economic activity. These categories form the basis of the two-way taxonomy developed by the project editors after an intensive review of the country evidence. On the opportunities dimension, the study emphasizes aspects of location and resource endowment that distinguish strong-growth and weak-growth countries throughout the world. As Collier and O'Connell (2005) note, resource-rich economies are those in which the rents from natural resource exports are sufficiently large that success or failure in managing these rents plays a decisive role in overall growth per- formance. This is a time-varying category: economies such as Zambia were resource- rich at independence, but Cameroon, Equatorial Guinea, and Nigeria joined this cate- gory well after independence, and this group will soon include Chad and Sudan. Among the remaining economies, the study emphasizes the importance for growth opportunities of transport costs, political borders, and distance to markets. The key distinction is between coastal, resource-poor economies, such as Madagascar, Senegal, and Tanzania, and landlocked, resource-poor economies, such as Burundi and Malawi. The opportunity classification is therefore a threefold distinction between resource- rich, landlocked resource-poor, and coastal resource-poor economies. Placing geography and resource endowments at the core of this taxonomy is sub- jective. The cross-country growth literature includes a long list of alternative vari- ables that condition growth outcomes on a global basis and may be substantially 36 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL predetermined with respect to policy choices in the medium term. These include exposure to external shocks (Guillaumont, Guillaumont Jeanneny, and Brun 1999; Blattman, Hwang, and Williamson 2004; and Easterly and Levine 1998 on neighbor- hood effects); the quality of inherited public institutions (Acemoglu, Johnson, and Robinson 2001); the level of human development (Glaeser and others 2004); and the degree of ethno-linguistic fractionalization (Easterly and Levine 1997). Aspects of these variables are captured by the classification used here: resource-rich economies have greater exposure to terms of trade shocks, landlocked resource-poor economies are more dependent on neighbor effects, and the resource-rich criterion implicitly benchmarks for relative endowment of natural and human resources. But there is no need to force experience through this taxonomy. Other predetermined influences on growth operate as cross-cutting variables that may or may not be correlated with the opportunity categories. The country evidence reveals patterns of ethno-regional polarization inherited from the colonial period. As argued below, these patterns appear to have been an important force behind observed policy choices. Ultimately, it may be political geography rather than solely economic geography that will most effectively capture the opportunity structure confronting African countries (Azam 2005; Bates 2005a, b). On the choices dimension, the narratives contained in the country studies organize experience by growth episodes. Country authors were asked to characterize the nature of that environment for each episode, to assess its impacts on investment and growth, and to study the internal political economy of its evolution over time. While no single account can capture the complexity of African experience, this episodic structure lends itself to the identification of recurring patterns in the country evidence. Four broad anti-growth syndromes emerged repeatedly in the case studies. These syndromes are not exhaustive of the ways in which African governments have actively shaped the growth environment, and in a substantial proportion of episodes, countries avoided all four syndromes. The patterns outlined here do not constitute a complete account of growth outcomes, which requires controlling for exogenous shocks and underlying growth opportunities. Three of the four syndromes--control or regulatory regimes that distort produc- tive activity and reward rent-seeking, redistributive regimes that compromise effi- ciency for the sake of redistribution, and intertemporal regimes that produce macro- economic instability by systematically undervaluing the future--can usefully be thought of as directly reflecting the choices of incumbent state actors. The fourth-- state breakdown--refers to situations of civil war or intense political instability in which a government fails to provide security or to project a coherent identity in a substantial portion of the country. The policy patterns characteristic of each syn- drome are described in full in Collier and O'Connell (2005). This paper draws on that discussion before exploring the links of these syndromes with growth. Based on a detailed reading of the country evidence, the editorial committee and country authors classified each country's situation as displaying one or more of the syndromes or, in the many cases in which syndromes were absent, as syndrome free (table 2). At a subsequent stage, the editorial committee extended the classification to the remaining African countries, based on a reading of the relevant literature.4 The EXPLAINING AFRICAN ECONOMIC GROWTH | 37 TABLE 2. Frequency of Anti-Growth Syndromes between Independence and 2000, by Opportunity Group (percent) Distribution Anti-growth syndrome of annual Opportunity observations by Inter- State Syndrome group opportunity group Regulatory Redistributive temporal breakdown free Proportion of country/years after independence Unweighted displaying each syndrome or syndrome free Coastal, 50.7 45.0 14.3 6.4 12.1 33.8 resource-poor Landlocked, 30.3 50.7 17.2 21.0 14.4 28.3 resource-poor Resource- 19.0 44.4 24.1 23.8 10.6 34.4 rich Total 100.0 46.6 16.9 13.9 12.5 32.3 Population- Proportion of person-years after independence weighted displaying each syndrome or syndrome free Coastal, 52.5 40.4 22.8 4.1 11.1 37.8 resource-poor Landlocked, 26.1 40.2 23.3 30.4 17.0 14.2 resource-poor Resource- 21.4 15.2 73.2 63.2 6.5 7.9 rich Total 100.0 35.1 33.5 23.3 11.7 25.4 Source: Judgments by the AERC Growth Project editorial committee. Note: Columns add up to more than 100 percent because a given country/year may be characterized by more than one anti-growth syndrome. Opportunity groups vary over time, as explained in the text. data are reported both by country/years and by person/years, with each country/year weighted by the mid-sample population. Differences reflect the wide distribution of population sizes in Sub-Saharan Africa and particularly the influence of Nigeria (reclassified from coastal to resource rich in 1971). The classification is based on policies, not growth outcomes. For each syndrome, theoretical considerations and global evidence suggest potentially substantial effects on growth, holding other influences constant. But a country exhibiting an anti- growth syndrome may nonetheless grow rapidly, as Sudan did in the late 1990s. Conversely, a country that steers clear of syndromes may stagnate, as Malawi did in the 1980s and Côte d'Ivoire did in the early 1990s. What the African evidence sug- gests, however, is that being syndrome free is virtually a sufficient condition for avoiding the short-run growth collapses that have so often undermined long-run growth performance on the continent. At the same time, while an absence of syn- dromes does not guarantee rapid growth, it is virtually a necessary condition for sus- taining rapid growth in the medium term. Where strong growth emerges in the midst 38 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL of a syndrome, it tends to do so either temporarily, as a result of unsustainable dynamics, or fortuitously, as a response to favorable shocks. Regulatory Syndromes African countries gained independence in the wake of World War II, in an economic and intellectual environment transformed by the Great Depression and the emer- gence of the Soviet Union as an industrial power. Founding governments had highly ambitious views of what governments could and should accomplish in the service of modernization (a theme developed in detail by Ndulu 2004). Botswana, Côte d'Ivoire, and Kenya (under Kenyatta) differed qualitatively from Ghana (under Nkrumah), Tanzania, and Zambia in the status they attached to markets and private accumulation as opposed to direct controls and government regulation in the devel- opment process. They also differed in terms of the importance they attributed to international markets and foreign private capital in the development process. These differences continue to distinguish episodes both within and across countries throughout the post-independence period. Regulatory regimes are those in which governments heavily distorted major eco- nomic markets (labor, finance, domestic and international trade, and production) in service of state-led and inward-looking development strategies. These regimes often emerged under the banner of socialist or communist ideology. Within this category, soft controls are characterized by aggressive import substitution, financial repression, and rapid expansion of the state enterprise sector, as in Zambia under Kenneth Kaunda. Hard controls imply deeper and more widespread repudiation of markets and Soviet-style planning. These measures are often, though not always, identified with revolutionary Marxist ideology, as in Ethiopia under the Derg regime. Formal adoption of a socialist orientation came to Tanzania in 1967, with the Arusha Decla- ration, and to Zambia in 1969, with the Mulungushi Declaration. Such pronounce- ments often provide key markers, with the distinction between soft and hard controls a matter of judgment. By the early 1970s, for example, Tanzania appeared to have adopted hard controls, with extremely tight exchange controls, price controls, and government marketing monopolies covering major portions of internal and external trade; a monopolistic nationalized banking sector; and the forcible relocation of smallholders into collective ujamaa villages. The regulatory regime in Zambia, by con- trast, remained soft until the adoption of market-based reforms in the early 1990s. The hallmark of regulatory regimes is a dirigiste mentality that tolerates substan- tial market distortions in an attempt to rapidly alter historical patterns of resource allocation. No single indicator captures these regimes, because patterns of interven- tion are country and institution specific. In countries with independent national cur- rencies, for example, where exchange controls were widely used to direct cheap for- eign exchange to favored uses, the size of the black market premium provides a measure of how far policy makers were willing to depart from market pricing in one key arena. This instrument was absent in the Communauté Financière d'Afrique (CFA) zone, where convertibility and an open capital account were enforced at the communitywide level and supported by a convertibility guarantee from France. Within the CFA zone, indicators such as the steepness of effective protection or the EXPLAINING AFRICAN ECONOMIC GROWTH | 39 ratio of public sector wages to national incomes provide better indicators of how far governments were willing to tilt the incentive environment. Real interest rates on banking deposits provide a useful indicator of the degree of financial repression imposed or tolerated by countries with national currencies. For the CFA countries or members of the Rand Monetary Area (Lesotho, Namibia, and Swaziland), real inter- est rates say less about national policy orientations; other measures of financial repres- sion, including subsidies granted to loss-making development banks, are more useful. Among countries with national currencies, black market premiums are distinctly higher in hard control regimes than in soft control regimes; they are virtually absent in syndrome-free countries (table 3). A measure of ex ante real interest rates is signif- icantly negative only in regulatory regimes, and the median rather than the less reli- able mean distortion in hard control regimes is roughly twice that in the soft regimes. Highly distorting regulatory regimes undermine growth both directly, through the deadweight efficiency losses associated with driving a wedge between prices and social opportunity costs, and indirectly, through the diversion of private activity from socially productive arenas to rent-seeking. The classic treatment in the African con- text is Bates (1981), who argues that the stagnation of both industry and agriculture was rooted in the use of marketing boards, import quotas, and other direct interven- tions to generate massive resource transfers from agriculture to industry and govern- ment. Agriculture stagnated because low producer prices and inadequate public investment undermined private incentives to invest and produce; industry stagnated for lack of foreign and domestic competition and because controls invited the diver- sion of resources from production to rent-seeking. What supported the resource transfer out of agriculture, in Bates' analysis, was the greater political coherence, and therefore political influence, of the largely urban interest groups that gained from the transfer. While Batesian interest-based accounts help explain the persistence of control regimes in the country studies, what looms larger in their adoption is the ideological TABLE 3. Regulatory Distortions in African Countries with National Currencies, from Independence to 2000 (percent, except where otherwise indicated) Variable Hard controls Soft controls Syndrome free Black market premium (to 1998) Mean 196.0 148.9 23.8 Median 79.6 41.0 9.6 Number of observations 93 242 316 Ex ante real deposit interest rate Mean -19.6 -76.1 -5.2 Median -7.6 -4.7 -0.8 Number of observations 49 127 147 Sources: Black market premium: Global Development Network database. Deposit and inflation rates: International Financial Statistics online. Note: The ex ante real deposit rate is the nominal deposit rate minus the expected inflation rate, with expected infla- tion proxied by a weighted average of once-lagged (0.7) and twice-lagged (0.3) inflation. 40 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL background and political origins of leaders. Intellectual fashions go farther than urban bias, for example, in explaining the strands of Fabian socialism adopted from Senegal to Tanzania to Mauritius; the agricultural origins of early leaders in Côte d'Ivoire, Kenya, and Malawi help explain the more supportive stance toward agri- culture in these countries. Redistributive Syndromes Redistributive syndromes are those in which efficiency-compromising resource trans- fers played a dominant role in defining the growth environment. Progressive, or ver- tical, redistribution syndromes aggressively seek to equalize the size distribution of household assets or income. Redistribution from rich to poor households has ambiguous effects on growth--deadweight efficiency losses are unavoidable, but if redistribution dramatically alters the investment constraints facing poor households, the net impact may promote growth. Of course, such redistribution may be highly desirable in its own right. But vertical redistribution does not feature prominently in the case material--in very few cases is it a defining characteristic of the growth envi- ronment. Only perhaps in Namibia in the 1990s and currently, much more damag- ingly, in Zimbabwe is vertical redistribution an important factor. Regional redistribution syndromes often respond to regionally polarized political conflict. This kind of redistribution is much more common than vertical redistribu- tion. This syndrome reflects pre-existing economic and political cleavages that were in some cases developed and more often articulated during the colonial period. Nige- ria is a prime example. An uneasy federal accommodation of the poorer, larger, and more politically coherent Muslim north with the Yoruba- and Ibo-dominated coastal regions produced a civil war in 1967 and a succession of northern military govern- ments intent on maintaining control over coastal oil resources. Since investment is forward looking, a redistribution that is anticipated by asset owners and potential investors can undermine growth well before it actually occurs. The anticipation is typically rooted in the prospect of a political transition that will remove the current elite from power and remove the protections afforded by that elite to favored groups. The leading example in the case material is South Africa, where the apartheid system became increasingly isolated politically beginning in the early 1980s, undermining the investment incentives of the white minority. This situ- ation persisted at least until the transition to majority rule and the adoption a few years later of a federal structure. In the authors' view it continues, to a substantial degree, under the rule of the African National Congress. Redistribution implies the appropriation and transfer of private incomes or assets, thereby drawing the state into market interventions that may compromise growth by distorting the incentive to invest. But as Azam (1995, 2001) notes, redistribution may be growth enhancing in an environment of deep political polarization. Figure 1 (developed in detail by Azam 2005) shows the essence of this argument. It shows two regions, whose relative income endowments are as given at point A. Region 1 is in control of the national government. By virtue of its low income, however, region 2 has a low opportunity cost of conflict. In this situation the option of armed rebellion EXPLAINING AFRICAN ECONOMIC GROWTH | 41 FIGURE 1. Buying the Peace Income in region 2 D F C B E A 45% Income in region 1 (incumbent) Source: Azam 2005. may appear attractive to political leaders in region 2, even if it is inefficient for the country as a whole. Indeed, if the likely result of a (costly) civil war, at point B, is anywhere north of point A, then region 2 is strictly better off under rebellion than under the status quo. The threat of armed rebellion is then credible. If transfers were costless, the contract curve of efficient bilateral bargains between the regions would be the segment CD. Taking the efficiency costs of redistribution into account, the contract curve may be closer to EF, with point E being the minimal interregional transfer capable of "buying the peace." To identify regional transfers as an anti- growth syndrome would reflect, in this case, the adoption of an irrelevant counter- factual. Put more simply, the "no redistribution" point (A) is not an equilibrium. In Azam's analysis, deeply polarized countries that reach efficient political bar- gains avoid a counterfactual of state breakdown. Growth is slower than in struc- turally similar countries in which polarization is absent but faster than in polarized countries in which institutional weaknesses, collective action problems, and oppor- tunistic behavior by regional elites prevent the striking of durable and intertempo- rally efficient bargains. The redistributive syndrome is constructed to exclude cases in which redistribution appears to be a reasonably efficient response to deep ex ante polarization, recognizing that such distinctions of degree are contentious. Intertemporal Syndromes Intertemporal syndromes redistribute resources from the future to the present. The primary form of this syndrome is unsustainable public spending booms, often precip- itated by the arrival or prospect of large commodity export incomes, with the 42 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL attendant loosening of financial constraints. Policy errors are of two fundamental types, sometimes occurring in combination. Less subtle is the failure to keep consump- tion spending in line with a reasonable assessment of permanent income, as in the dra- matic expansion of public sector employment during the bauxite boom in Guinea and the uranium boom in Niger during the 1970s. More subtle is the rush to expand the public investment budget in the absence of institutional mechanisms to ensure adequate economic returns, as in the proliferation of new state capitals and politically located infrastructure investments in Nigeria during the oil booms of the 1970s. In each case, the full boom/bust period, rather than just the boom phase, is examined as a syndrome. This forces a reinterpretation of the bust phase as in large part a consequence of previ- ous policy choices, rather than a response to intervening policy reforms. Within the intertemporal syndrome the term looting is used for the limited set of episodes in which a narrow political elite subordinates any coherent economic pro- gram to its own subsistence on what the authors of the Burundi study call the "rents to sovereignty"--mineral resources, foreign borrowing, aid, tax revenues. Almost all such episodes are associated with dictators--Idi Amin in Uganda, Charles Taylor in Liberia--though in the Burundi case the cohesion of the Bururi faction survived a number of internal transitions. By some arguments these syndromes belong more properly in the state breakdown category. The hallmarks of intertemporal syndromes are debt accumulation and other indicators of macroeconomic imbalance (table 4). State Breakdown The state breakdown syndrome differs qualitatively from the other syndromes in that it constitutes an outcome rather than a program. Examples of state breakdowns are the large-scale armed rebellions in Sierra Leone in the 1990s and the civil war in Chad in 1979­80. Also included are a few brief episodes of acute political crisis, such as Niger's failed democratization during the 1990s, when a sequence of constitu- tional crises undermined any projection by the government of a coherent incentive environment for private economic activity. A fundamental intellectual challenge for interpreters of Africa's growth record is to assess the relative importance of exposure to shocks and actual experience of shocks. Dehn (2000) undertakes this exercise for commodity price collapses and argues that what matters for growth is the actual sample path taken by commodity prices rather than the ex ante volatility of that path. Restricting attention to the actual occurrence of state breakdown means that the potentially powerful effect on growth of a high ex ante risk of civil war is not captured. This underscores the broad nature of the syndrome-free category. It certifies countries as free of dramatic errors of commission but, as shown below, does not constitute a sufficient condition for rapid growth. Patterns over Time The evolution of these four broad syndromes was tracked for 46 African countries between the date of political independence (or 1960, whichever is later) to 2000 (fig- EXPLAINING AFRICAN ECONOMIC GROWTH | 43 TABLE 4. Correlates of Intertemporal Syndromes, from Independence to 1997 Intertemporal Syndrome- All country/year Variable syndrome free observations Fiscal deficit (percent of GDP, from early 1970s) Mean 4.2 2.3 3.1 Median 3.4 2.5 2.9 Number of observations 108 159 267 Overvaluation index (from early 1960s) Mean 173.5 133.6 144.5 Median 168.1 124.3 134.2 Number of observations 151 402 553 Increase in external debt-to-GDP ratio (percentage points of GDP, from 1972) Mean 5.6 1.3 3.0 Median 3.1 0.2 1.0 Number of observations 182 280 462 Inflation ratea (percent, from early 1960s, ending 2000) Mean 452.2 11.8 111.5 Median 21.2 6.5 8.5 Number of observations 197 503 700 Sources: Fiscal deficit excluding capital grants; overvaluation index (ratio of real effective exchange rate to the coun- try's average overvaluation index for 1970­85, as calculated by Dollar 1992, normalized to indicate overvaluation for values of more than 100); and increase in external debt: Global Development Network database. CPI inflation rate: IMF, International Financial Statistics. a. Reported for countries with national currencies only. ure 2). The most dramatic change over time is the displacement of syndrome-free cases by control regimes between the late 1960s and the mid-1980s and their subse- quent restoration between the mid-1980s and the early 1990s. The period of democ- ratization, which occurred between 1988 and 1994 (Bratton and van de Walle 1997), was one of improvements in the growth environment, reflecting widespread liberal- ization of controls and the resolution of some redistributive and intertemporal syn- dromes. This period also saw, however, a substantial increase in state breakdown. Syndromes and Growth The adverse impact of economic policy on growth became a central theme of the African development literature with the 1981 Berg Report (World Bank 1981) and the global onset of the structural adjustment era. The syndrome classification is consistent with many of the central themes of the subsequent cross-country growth literature, including the well-known finding that while individual policy measures rarely gener- ate highly robust impacts, groups of policy measures are invariably jointly significant (Levine and Renelt 1993). It covers virtually all of Sub-Saharan Africa. 44 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL FIGURE 2. Policy Syndromes in 46 Countries in Sub-Saharan Africa since Independence 0.8 All syndromes 0.5 Regulatory syndromes 0.6 0.4 0.3 0.4 0.2 0.2 0.1 0 0 1960 1970 1980 1990 2000 1960 1970 1980 1990 2000 Syndrome-free Regulatory Soft controls Redistributive Breakdown Hard controls Intertemporal syndromes country-years of 0.20 Redistributive syndromes 0.20 Intertemporal syndromes 0.15 0.15 0.10 0.10 oportion 0.05 0.05 Pr 0 0 1960 1970 1980 1990 2000 1960 1970 1980 1990 2000 Vertical Anticipated Looting Anticipated Regional Unsustainable spending Source: Subjective classification by Growth Project editorial committee based on country studies and broader literature. The four anti-growth syndromes are combined into a single dichotomous variable indicating whether a country is syndrome-free in a particular year or displays one or more anti-growth syndromes (table 5). Both mean and median growth rates (the lat- ter considerably more robust to outliers) are substantially higher in the absence of syndromes. An absence of syndromes is strongly associated with avoiding a growth collapse (defined as a year in which a three-year centered moving average of growth is negative): less than 20 percent of syndrome-free countries experienced collapses. By the same token, syndromes exert a powerfully negative impact on the prospects for strong growth over the medium term. Just 25 percent of countries that were not syndrome-free enjoyed five-year average growth of more than 2.5 percent (roughly the developing country median for 1960­2000), and less than 20 percent grew more than 3.5 percent a year. All of these differences are highly statistically significant. The implied correlations reported in table 5 are descriptive rather than structural; causality may run in both directions. Nonetheless, these data have the advantage of mobilizing a significantly larger sample of observations than is available in any cross- country growth regression (samples shrink drastically when individual measures of policy are used or other variables are present). Moreover, the coarseness of the syn- drome classification protects the sample from some types of reverse causality--for example, exogenous shocks that simultaneously affect growth and standard mea- sures of policy (such as black market premium, or the fiscal deficit) without induc- ing a major shift in policy regime. A regression framework can be used to assess the impact of syndromes after condi- tioning exogenous shocks and unobserved country effects (table 6). The African growth data have extraordinarily high dispersion. To limit the leverage of outlying observations, EXPLAINING AFRICAN ECONOMIC GROWTH | 45 TABLE 5. Relationship between Syndromes and Growth, from Independence to 2000 Observations displaying Syndrome-free Indicator one or more syndromes observations All observations Growth rate of real GDP per capita (percent) Mean -0.37 2.33 0.49 Median 0.02 1.82 0.78 Number of 1,109 515 1,624 observations Frequency of growth collapses and surges (percent of country/year observations) Three-year moving 49.3 19.1 39.5 average of growth less than zero Three-year moving 50.7 80.9 60.5 average of growth zero or positive Total 100.0 100.0 100.0 Five-year moving 75.9 47.5 66.7 average of growth less than 2.5 percent Five-year moving 24.1 52.6 33.3 average of growth 2.5 percent or more Total 100.0 100.0 100.0 Five-year moving 80.7 55.8 72.6 average of growth less than 3.5 percent Five-year moving 19.3 44.2 27.4 average of growth 3.5 percent or more Total 100.0 100.0 100.0 Source: AERC Growth Project. these regressions are estimated using a least absolute deviation criterion rather than the standard least-squares criterion. Estimated coefficients refer to the impact of marginal changes on the predicted median, rather than the mean, of the dependent variable. In columns 1 and 2, exogenous shocks are conditioned on export markets (through partner growth rates) and agricultural supply (through normalized rainfall levels). In the rainfall-only case, a set of fixed-year effects is included to allow for time-specific influences that are invariant across countries. Columns 3 and 4 include location and endowment effects as proxied by the opportunity variables. The omit- ted geographical category is landlocked, resource-poor economies, so the coefficients on coastal and resource-rich economies estimate the growth impact of these cate- gories relative to the landlocked countries. The economic return to coastal location is decidedly smaller within Africa than in a global sample of developing countries, an issue discussed in detail in Collier and O'Connell (2005). Columns 5 and 6 repeat the 46 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL TABLE 6. Robust Regressions Controlling for Shocks (dependent variable: growth in real GDP per capita) Variable (1) (2)a (3) (4)a (5)b (6)a,b Syndrome free 1.747 1.650 1.672 1.816 2.120 2.086 (0.315)*** (0.294)*** (0.291)*** (0.293)*** (0.292)*** (0.080)*** Partner growth 0.309 0.309 0.296 (0.092) (0.084) (0.072)*** Rainfall 0.165 0.167 0.135 0.171 0.178 0.177 (0.144) (0.151) 0.131 (0.150) (0.112) (0.031)*** Coastal 0.502 0.104 (0.298)* (0.299) Resource rich 0.451 0.595 (0.375) (0.378) Number of 1,492 1,795 1,492 1,795 1,492 1,795 observations Source: AERC Growth Project. Note: All regressions are estimated using the least absolute deviation method. The excluded category in regressions 3 and 4 is landlocked, resource-poor economies. Standard errors are in parentheses. a. Regression includes yearly fixed effects. b. Regression includes country fixed effects. * p < .1, *** p < .01. specifications of columns 1 and 2 but incorporate country fixed effects. In these columns any spurious correlation arising from unobserved country-specific factors is swept out, and the results assume only that the differential effect of syndromes is the same across countries. Table 7 repeats this exercise for the full set of anti-growth syndromes. The omit- ted category is now "syndrome free." Although the syndromes are classified by ref- erence to policy regimes and not growth outcomes, their individual links to growth are powerful and in each case highly statistically significant. The marginal impacts of individual syndromes are strikingly consistent: the regulatory, redistributive, and intertemporal syndromes shift predicted median growth by a full percentage point, while state breakdown reduces predicted median growth by roughly 2.5 percentage points a year. The impacts of regulatory regimes and state breakdown remain strong in column 5, where unobserved country heterogeneity is controlled for, confirming the salience of these regimes within the country narratives. The impact of redistribu- tive regimes, in contrast, is less robust to controlling for unobserved heterogeneity. This is consistent with the analysis by Azam; ethno-regional polarization may be operating here as a slowly moving unobserved variable whose presence is proxied by episodes in which aggressive redistribution dominates the growth environment. These are large effects. While these results are descriptive rather than structural, at least one major source of spurious correlation was eliminated by controlling for exogenous shocks that may simultaneously lower growth and induce the adoption or abandonment of syndromes. EXPLAINING AFRICAN ECONOMIC GROWTH | 47 TABLE 7. Robust Regressions Controlling for Shocks: All Syndromes (dependent variable: growth in real GDP per capita) Variable (1) (2)a (3) (4)a (5)a,b Regulatory -0.987 -0.914 -0.908 -0.860 -1.73 (0.290)*** (0.196)*** (0.253)*** (0.219)*** (0.308)*** Redistributive -0.997 -1.158 -0.946 -1.123 -0.652 (0.366)*** (0.244)*** (0.319)*** (0.271)*** (0.434) Intertemporal -0.878 -1.090 -0.891 -1.044 -0.335 (0.409)** (0.291)*** (0.367)** (0.333)*** (0.425) Breakdown -2.320 -2.482 -2.138 -2.612 -2.59 (0.432)*** (0.300)*** (0.378)*** (0.334)*** (0.439)*** Partner growth 0.292 0.280 0.317 (0.093)*** (0.081)*** (0.079) Rainfall 0.102 0.157 0.076 0.141 0.179 (0.146) (0.106) (0.126) (0.118) (0.122) Coastal 0.330 -0.062 (0.292) (0.240) Resource rich 0.757 0.605 (0.360)** (0.298)** Number of 1,492 1,795 1,492 1,795 1,492 observations Source: AERC Growth Project. Note: Standard errors are in parentheses. a. Regression includes yearly fixed effects. b. Regression includes country fixed effects. ** p < .05, *** p < .01. Table 5 suggests that being syndrome free was (nearly) a necessary condition for sustained growth in the medium term. The traction from this measure declined over time. For a syndrome-free country, the probabilities of surpassing the modest hurdles of 2.5 or 3.5 percent annual growth declined by nearly half between the 1960s and the 1990s (table 8). Such a decline is plausible. Alternative hypotheses consistent with it might include uncertainties introduced by the wave of democratic reforms in Africa; "first mover" agglomeration benefits accruing to the dynamic emerging economies in Asia and the emergence of a more competitive international market for African exporters after the early 1980s; and lagged effects of previous syndromes, including deterioration of institutions and infrastructure. But the result itself is of questionable robustness: a test for period effects after controlling for global growth rates and other shocks revealed mixed results. It is nevertheless clear that some por- tion, perhaps most, of the apparent decline in the payoff from avoiding syndromes is associated with a deteriorating external growth environment. Controlling for shocks by including partner growth rates or period fixed effects, the growth differentials associated with avoiding syndromes are not significantly smaller in the 1990s than in earlier decades. 48 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL TABLE 8. Percentage of Syndrome-Free Countries Experiencing Rapid Growth, by Decade Annual per capita growth Annual per capita growth Decade exceeding 2.5 percent exceeding 3.5 percent 1960s 73 60 1970s 52 41 1980s 44 42 1990s 40 35 Sources: World Bank World Development Indicators and AERC Growth Project. Case Study Evidence Case studies of 27 Sub-Saharan Africa countries shed light on the dynamics involved in the occurrence and resolution of each syndrome. Some of that evidence, as well as examples of syndrome-free regimes, is presented here. Regulatory Syndromes Regulatory syndromes involve the heavy distortion of major economic markets by the government, usually in the service of state-led and inward-looking development strategies. The severity of such distortions differs. Soft controls entailed a substantial expansion of the state enterprise sector, accompanied by considerable financial repression and foreign exchange controls, for example. Hard controls involved more severe distortion. The state became the main agent for resource allocation, often under a politically radical mantra of Marxist socialism. Burkina Faso, 1960­82 (Soft Controls), 1983­90 (Hard Controls) Burkina Faso became independent in 1960, along with most of francophone Africa. Savadogo, Coulibaly, and McCracken (2003) characterize the period from 1960 to 1990 as one of state interventionism in the absence of a dynamic private sector. The degree of intervention evolved between 1960 and 1982. In the cereals (sorghum, mil- let, and maize) markets, for example, intervention was tightened considerably follow- ing the drought of the mid-1970s. Private trade became restricted, and OFNACER (the Cereal Office), originally set up to handle food aid, became a bilateral monopoly on cereal trade. Classification of this period as characterized by soft rather than hard controls reflects the determination of a relatively politically conservative leadership (President Yameogo, 1960­6; General Lamizana, 1966­80; Colonel Zerbo, 1980­2) to leave considerable, if uneven, latitude for private sector activity. The period from 1960 to 1982 was one of substantial political instability, as reflected in the coups of 1966, 1980, and 1982 (a legacy repeated in 1983 and 1987). The first two coups appear to have arisen, at least in part, from popular dissatisfac- tion with economic conditions; annual per capita GDP growth averaged less than 1 percent until the early 1980s.5 The era of soft controls ended in August 1983, when a group of officers led by the charismatic radical Thomas Sankara took over the gov- EXPLAINING AFRICAN ECONOMIC GROWTH | 49 ernment by coup and imposed a severe form of "state capitalism." The resulting hard control episode lasted until 1990. The centerpiece of the control program was a repu- diation of the legitimacy of private property. For example, a rental law was levied that required landlords to transfer a whole year of rent to the government; for 1984 rent was declared free. The hard control period was punctuated by a bloody coup in October 1987, when Sankara was killed as part of a power play among the radicals and replaced by Blaise Compaore. As economic stagnation continued, increasing financial imbalances and economic difficulties forced the Compaore government to seek salvation from the Bretton Woods institutions (Savadogo, Coulibaly, and McCracken 2003). The abandonment of controls and the initiation of a syndrome- free episode (see below) dates from 1991, the year of Compaore's election as a civil- ian president under a new constitution. Cameroon, 1960­77 (Soft Controls) Cameroon's first president, Ahmadou Ahidjo, pursued soft control policies in an indus- trialization drive based on the import substitution paradigm that was widely adopted in Africa and elsewhere (Ndulu 2004). His adherence to soft rather than hard controls appears to have reflected his relatively conservative political background. An element that may have favored state controls over a more market-friendly approach was the high priority accorded by the government to peace building (Kobou and Njinkeu 2004); Cameroon had experienced a serious ethno-regional-based insurrection in the pre-independence period; neighboring Nigeria, with similar ethno-regional cleavages, experienced a civil war between 1967 and 1970. The end of Cameroon's soft control period was marked by an investment push in the oil sector and the onset of major oil revenues in 1978, which initiated an unsustainable spending boom (see below). Ghana, 1960­8 and 1972­83 (Hard Controls) At the time of independence, in 1957, the clear choice was between J. B. Danquah, a politically conservative politician from the majority Akan ethnic group who pre- ferred a go-slow, market-friendly approach, and Kwame Nkrumah, from the minor- ity Nzimma ethnic group, who favored a socialistic modality with a strong role for government. Nkrumah won the election based partly on the appeal of the promises of faster economic progress through the government's active role in the economy and partly out of fear of possible Akan domination. Nkrumah transformed the country to a one-party state by 1960, as the Republic of Ghana, assuming a strong role as president from his initial position as prime minister from the time of independence. The government embarked on a radical path of indus- trialization, with the state playing the leading role. The socialistic ideology of collec- tive ownership, later dubbed "Nkrumahism," reigned supreme. Nearly all large-scale operations were owned by the state. In the process, private enterprises, especially those of medium to large sizes, were squeezed out, through political intimidation or the diminished availability of finance. Huge spending on nationally unproductive projects, given Nkrumah's vision of a total liberation of the African continent from colonialism and imperialism, led the country rapidly toward macroeconomic difficul- ties. By 1966 the country had a deficit of US$391 million in its net international 50 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL reserves, down from a surplus of US$269 million at the time of independence. Real per capita GDP fell from US$500 in 1960 to about US$470 by 1966 (in 1987 dollars), as inflation rose from virtually zero to 22.7 percent. Nkrumah's overthrow by a mil- itary coup in February 1966 was greeted with great popular enthusiasm. Policy nonetheless changed little until the brief episode of market-based reforms instituted by Prime Minister Kofi Busia in 1969. The period from 1972 to 1983 was one of "muddling through" (Aryeetey and Fosu 2003). A series of five governments came to power, mainly through military coups and countercoups (the civilian government of President Hilla Limann, 1979­81 was a brief exception). Governments in this period responded to economic difficulties by imposing further constraints. Price and import controls were imposed in response to the inability to contain macroeconomic imbalances and the unwillingness to devalue the local currency.6 Except for the explicitly ideological tinge of the Rawlings regime, which came to power in a coup in December 1981 (following a brief coup by Rawl- ings himself in 1979), these governments were not particularly radical; however, they viewed control policies as the way out. The military government of Ignatius Acheam- pong (1972­9), for example, engaged in schemes such as Operation Feed Yourself and flirted with the concept of a "uni-government" that would include both the military and civilians. Severe fiscal difficulties were apparent in Ghana during the early 1980s. By 1983 (a year of very severe drought), central government revenues had shrunk to just 5 percent of an already diminished GDP, down from 20 percent in 1970. Inflation had risen from 18 percent in 1974 to 116 percent in 1981, even though prices were supposed to be controlled; domestic investment had fallen to less than 4 percent of GDP, from 14 percent in 1974; and the current account balance was US$175 million in deficit, with no loans or grants to finance any of it and the country already expe- riencing major arrears. The radical-leaning Rawlings appeared to have little choice but to succumb to the IMF/World Bank­sponsored Economic Recovery Program, ushered in during April 1983, followed by the structural adjustment program in 1986. The period from 1983 to 2000 is characterized as syndrome free (see below). Sierra Leone, 1973­89 (Hard Controls) Sierra Leone gained independence in 1961. In 1968 the All People's Congress (APC) assumed power, ending an initial period of conservative politics but leaving in place a largely market-friendly policy regime, characterized here as syndrome free. Follow- ing a period of acute political instability, the oil shocks of the 1970s propelled the implementation of aggressive price and exchange rate controls and financial repres- sion. The 1973­89 period was one of systematic erosion of pro-growth political and economic institutions. Export crops--palm kernel, cocoa, and coffee--were sub- jected to explicit and, perhaps more significantly, implicit taxation by the monopson- istic Sierra Leone Produce Marketing Board. Meanwhile, import subsidies and price controls became sources of rents and mechanisms for redistribution to favored north- ern interests. The government borrowed heavily during this period to finance grandiose public projects; its hosting of the 1980 Organization of African Unity sum- mit, for example, consumed more than 60 percent of government revenues in 1980 EXPLAINING AFRICAN ECONOMIC GROWTH | 51 (Davies 2005). The economic situation deteriorated steadily, reaching crisis propor- tions by the mid- to late 1980s: by 1987 the poverty rate stood at 80 percent, infla- tion was 180 percent, and tax revenues had fallen by half, to less than 10 percent of GDP. The government sought relief from the IMF in the form of a structural adjust- ment program late in 1989, but economic reform was interrupted by civil war. The 1990­2000 period is characterized as one of state breakdown (see below). Adverse Redistribution Redistribution that "buys the peace" can be favorable to growth, as can pro-poor redistribution, if its net effect is to crowd in investment in physical and human capi- tal. The focus here is on adverse redistribution, in which growth-reducing instru- ments are used to redistribute resources to groups favored by the incumbent politi- cal elites. Burundi, 1972­88 The Kingdom of Burundi gained independence in 1962, with the chaotic departure of the Belgian colonial regime and the assassination of Prince Louis Rwagasore a month after his pluralist party won national legislative elections in September 1961. The period from 1962 to 1972 was one of acute political instability, including large- scale political violence in 1965. This period is characterized in this study as a period of state breakdown. A sharply redistributive phase got underway with the accession to power of Cap- tain Michel Micombero, a young Tutsi officer from the Bururi province in southern Burundi (Nkurunziza and Ngaruko 2003). The period of adverse redistribution dates from 1972, when a bloody Hutu rebellion initiated a period of systematic redistribution and repression by the Micombero government and Tutsi-dominated army. During the 1972­87 period, the government created a large number of public corporations that distributed rents to members of the Tutsi political elite, mostly Bururi-Tutsis, who formed the political base for the ruling elite. Meanwhile, the army-led Tutsis perpetrated severe political repression against the majority Hutus, following a massacre of the Tutsi minority by Hutus in 1972. The rationale for this combination of redistribution and political repression emanated from the perceived fear of domination by the majority Hutus and may well have constituted an optimal political strategy from the narrow and contingent perspective of incumbent Tutsi elites (Nkurunziza and Ngaruko 2003). The pro-Tutsi pattern of ethno-regional redistribution ended with the replacement of Colonel Jean-Baptiste Bagaza by Major Pierre Buyoya in a coup in 1987. A renewed period of civil war and state breakdown ensued in 1988 (see below). Sierra Leone, 1970­89 The redistributive behavior of the government during the 1970­89 period was based on interethnic rivalry between the northern-based Temnes and southeastern Mende groups, each of which commanded roughly 30 percent of the population. The Temnes dominated the APC party, which took over the government in 1968. To con- solidate its political base, the new government engaged in regional redistribution in 52 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL favor of the Temnes. This redistribution disenfranchised the Mendes and increased interethnic polarization, sowing seeds for the armed conflict and state breakdown of the 1990s (see below). Togo, 1975­2000 Early Togolese governments differed much less over economic policy than over the continuation of political and economic ties with France, with the country's first presi- dent, Sylvanus Olympio, seeking to cut the country's colonial ties. Sub-Saharan Africa's first military coup, organized by a northern-based military junta, brought the pro-French southerner Nicolas Grunzitzky to power in 1963. The junta replaced Grun- zitzky in a bloodless coup in 1967, initiating a 37-year period of military dictatorship under Gnassingbe Eyadema. A distinctly ethno-regional bias of economic policy began to emerge in the early years of the Eyadema government, with commodity export taxes channeled increasingly into development programs favoring the Kara region of the president's Kabye ethnic group (Gogué and Evlo 2004). This ethno-regional bias became a dominant theme of policy in the mid-1970s, as a major expansion in public sector employment, financed initially by windfall increases in phosphate prices, was directed disproportionately to the Kabye. The regional dimension of redistribution continued through the 1990s, as abortive democratic reforms produced a hardening of repression and political control by the Eyadema government. Intertemporally Unsustainable Spending Intertemporally unsustainable spending booms are most often tied to temporary rev- enue bonanzas from increases in commodity export prices. Additional resources are often required as borrowing becomes cheaper and governments underestimate the resources required to realize overambitious public spending targets. When the revenue boom ends, fiscal difficulties force a reduction in public spending. This often means a sharp reduction in public investment spending, as current spending entitlements cre- ated during the boom phase--including expansions in public employment--prove politically difficult to reverse. The upshot is that many projects remain uncompleted, and given the lumpiness of investment, the value of the marginal product, which might be low to begin with, is seldom realized. The post-boom period is usually one of sharply reduced growth, as the cases that follow illustrate. Cameroon, 1978­93 Oil revenues rose sharply after the initiation of its production and exports in 1978. The revenues were removed from the normal budget process and placed in a special account managed directly by President Paul Biya, who replaced President El Hajj Ahmadou Ahidjo in 1982 upon Ahidjo's voluntary retirement. A failed coup attempt in 1984 led Biya to solidify his grip on power. This entailed major spending, thanks in part to the oil revenues directly under his direct control, geared toward the con- struction of an ethnically based political alliance. A sharp overvaluation of the real exchange rate emerged in the mid-1980s, as franc zone rules prevented a deprecia- tion to match the collapse of world oil prices. The introduction of multiparty politics EXPLAINING AFRICAN ECONOMIC GROWTH | 53 in 1990 further fueled political rivalry and spending intended to win electoral author- ity. The full oil-financed boom/bust episode ended with the 1994 devaluation of the CFA franc. Togo, 1974­89 Togo suffered from the negative oil price shock of 1973, but the phosphate boom of 1974­5, coupled with a dramatic increase in coffee prices in 1977, produced a wind- fall in public revenues. The government consolidated its control over phosphate rev- enues by nationalizing the sector. As a percentage of GDP, public investment increased from 13.4 percent in 1973 to 47.0 percent by the late 1970s, matched by more than a doubling of the workforce in the formal industrial sector, mostly pub- lic, between 1973 and 1979 (Gogué and Evlo 2004). By the late 1970s, however, Togo had begun to experience fiscal difficulties. These difficulties led to major increases in external borrowing, with external debt rising from 15.1 percent of GDP in 1970 to 116.4 percent in 1978. With crisis looming, the government negotiated an IMF-administered Financial Stabilization Program in 1979, and a series of Structural Adjustment Programs began in 1982. Government investment shrank from nearly 50 percent of GDP in the late 1970s to 20 percent by 1989; an employment freeze shrank the civil service by 13 percent between 1985 and 1988. By 1990 the unwind- ing of the unsustainable spending boom was essentially complete. The growth envi- ronment was subsequently dominated by the emergence of civil strife in 1990 and a temporary, abortive political reform and restoration of repressive military dictator- ship in 1991. State Breakdown State breakdown occurred when domestic law and order broke down, preventing the government from carrying out its basic functions of providing security and other essential services. It usually occurred during a civil war. It also occurred as a result of political instability, such as that caused by multiple coups d'état within a relatively short time or constitutional crises. Burundi, 1988­2000 Following the history of pro-Tutsi redistribution and the harsh political repression of the majority Hutus, civil war broke out in 1988. A second civil war occurred in 1993, following the assassination by the army of Melchior Ndadaye, Burundi's first democratically elected president. Pierre Buyoya, the Bururi Tutsi who seized power in a military coup in 1987 and then lost the 1993 election, regained power in a sec- ond military coup in 1996. Nkuruziza and Ngaruko (2003) characterize the entire period from 1988 to 2000 as one of war and unprecedented economic crisis. Chad, 1979­84 Chad gained independence in 1960 as a highly polarized country, with the cotton- producing south successfully leveraging its higher income and superior educational endowment into control of the national government. For the first decade and a half 54 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL of independence, the Tombalbaye government failed to sufficiently share southern wealth with the north, resulting in the emergence of low-intensity insurrection in the north (Azam and Djimtoingar 2003). Following a 1975 coup, General Félix Mal- loum, another southerner, attempted to bridge the political divide by appointing the northern rebel leader Hissène Habré as prime minister. Habré's aggressive anti- corruption drive alienated southern elites, leading to the 1979­84 civil war that cul- minated in northern-based rule. Following the 1979­84 war, the north severely repressed the south politically, a pattern that was attenuated only after Idriss Déby took over the government in 1990 and reinstated a regionally based power-sharing approach, appointing General Kamougue, who formerly led the resistance in the south, as president of the national assembly. The devaluation of the CFA in 1994 helped stabilize the economic envi- ronment as well, especially for cotton production. Sierra Leone, 1967­8, 1991­2000 State breakdown occurred in Sierra Leone in two periods. The first was a brief period of acute political instability following the closely fought election of March 1967. The election was won by the APC opposition party, dominated by the northern-based Temne group, but a coup prevented the victors from taking power. Within the span of a year, two other coups occurred, the second of which brought the APC to power and provided the basis for the redistributive episode of 1970­89. State breakdown reemerged with the outbreak of civil war in 1991. Though multiple reasons are usu- ally cited, the primary cause of the war appears to be the autocratic rule of the APC, which prevented any smooth political changeover and thus caused a rebel movement to take up arms against the government. The war was fueled by discontent, with eth- nic rivalry mainly between the Temnes of the APC and the Mendes as a driving force, together with readily available diamonds controlled by the rebels. The war was offi- cially ended in January 2002, after the defeat of rebel forces by external intervention. Syndrome-Free Cases Syndrome-free status typically denotes a combination of political stability with rea- sonably market-friendly policies. About the time of independence, the combination tended to be associated with the ascendancy of politically conservative governments-- as a result of political accident, replacement of an authoritarian government with a disastrous economic record by a military coup, or major fiscal difficulties that forced the government to seek fiscal space by submitting to a structural adjustment program administered by the Bretton Woods Institutions. Botswana, 1960­2000 Botswana is the shining example of a syndrome-free country in Africa. This outcome emanates from the democratic multiparty political arrangement based on the Tswana traditional political culture, which also served to protect the interest of minority groups. Unlike in most other African countries, state-led development in Botswana was based on strategic facilitation of the private sector rather than state suppression. The government of Botswana has also been democratic. EXPLAINING AFRICAN ECONOMIC GROWTH | 55 Why did Botswana succeed in establishing such a market-friendly democratic gov- ernment when so many countries failed or never even attempted to do so? One pos- sible answer is the decision by the time of independence in 1966 to base the govern- ment on pre-colonial traditional conservative Tswana culture, involving widespread political participation. This decision helped create "an indigenous developmental state" (Maundeni 2001; Maipose and Matsheka 2004). Botswana's success was also facilitated by the fact that it has a relatively small (less than 2 million even today), homogeneous population, as a result of which it did not suffer as much pre-independence polarization as many other African coun- tries. In addition, the interests of the members of the government, mostly cattle raisers, were consistent with the pursuit of market-friendly policies favoring the rural sector. Thus, urban bias policies, such as overtaxing the rural sector through the use of state marketing boards or overvaluating the exchange rate, were avoided. Botswana's natural resource endowment also gave it an advantage. Unlike the alluvial diamonds of Sierra Leone, for example, Botswana's diamond deposit requires deep-earth mining. Thus, the marginal cost of mining without appropriate state sanctioning was relatively high, attenuating the use of diamonds as a financing tool for insurgency. Another geographical benefit is Botswana's proximity to South Africa. The coun- try's membership in both the Southern African Customs Union and the Rand Mon- etary Area monetary union helped maintain exchange rate and monetary stability. Botswana was also the beneficiary of generous donor support based on its status as a democratic state at the doorstep of apartheid South Africa. Burkina Faso, 1991­2000 Burkina Faso's economy was relatively market driven between 1991 and 2000, with a new constitution adopted in June 1991 that ushered in political liberalization and a devaluation of the CFA franc in 1994. The genesis of this regime was the set of fiscal difficulties faced by the military regime in the late 1980s. Captain Blaise Compaore of the Front Populaire came to power in October 1987 following a coup. The Front Populaire "first attempted to continue with the revolutionary mood" (Savadogo, Coulibaly, and McCracken 2003, p. 53 ), but in the face of major fiscal and economic difficulties was compelled to seek assistance through the structural adjustment program of the IMF and World Bank. Ghana, 1984­2000 The fiscal and balance of payments crises of the late 1970s and early 1980s forced the radical-leaning Rawlings government to accept the bitter medicine of deregula- tion and economic liberalization in 1984 and structural adjustment in 1986. After years of muddling through, with the economy on its knees, little revenue to turn the situation around, and no hope for external financial support, the radicals had to bow to reason and accept market-based economic reforms. In addition, it was clear that after years of coups and countercoups and nothing to show for changes in regimes, another coup would not necessarily be the solution.7 56 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL Togo, 1960­73 Market-friendly policies survived the coups of 1963 and 1967 and the assassination during the 1963 coup of the politically conservative but economically liberal Presi- dent Sylvanus Olympio. The genesis of market-friendly policy reflected Olympio's background as a former executive of an international trading company, as well as the influence of the business community, which constituted his support base. The syndrome-free period ended with the phosphate booms of the mid-1970s. Sierra Leone, 1961­6 The politically conservative leadership of the protectorate-based Sierra Leone Peo- ple's Party (SLPP), comprising chiefs and people from the hinterland, facilitated a market-friendly set of policies. The SLPP maintained political stability by forging an alliance with the relatively educated Freetown-based Creoles. The stability ended when the closely fought election of March 1967 resulted in a successful coup that prevented the winner, the APC, from taking office. The APC was finally able to assume office following a third successful coup following the election. The more socialistic APC began to put in place relatively market-distorting and adverse redis- tributive measures of state controls. Endogenizing Policy Choices The case study examples demonstrate that policy choices are generally endogenous to both domestic and external factors. These include initial conditions; supply shocks; growth opportunities; institutions, especially the military; and economically driven political expediency.8 Nation-Building within the Constraints of Post-Independence Realities The nature of policies adopted in the period immediately following independence appears to have been very much determined by the coalition that won. The more politically conservative the winning coalition, the more likely was the adoption of relatively market-friendly policies. To a great extent such political leaning was influ- enced significantly by the experience of the leadership with respect to the competing development paradigms at the time: capitalist versus socialist.9 Most African leaders found the socialistic direction attractive, for several reasons. First, they were strong believers in the need for equitable growth and viewed capitalism as a mechanism for a few capitalists to become rich at the expense of the masses. Second, the private sec- tor was largely nonexistent in many countries; the state was thus seen as the primary agent for economic growth and development. Third, the government's role was seen as preserving the state, which normally comprised adversarial ethnic groups. This objective was seen as requiring a strong central authority with the power for resource redistribution to be targeted at attenuating centrifugal forces. The need to preserve the noncohesive embryonic nation following independence was the preoccupation of the vast majority of African leaders. Such preoccupation would dominate most of the policies pursued, including the adoption of controls, including those that augmented the power of the state, for redistribution purposes.10 EXPLAINING AFRICAN ECONOMIC GROWTH | 57 Time Preference and the Role of Government In many cases in which the leadership at the time of independence was politically conservative (supporting market-friendly policies), the governments did not last long.11 There was much political agitation in favor of the government leading the charge to improve the lot of the people, especially given the rather slow pace at which the relatively undeveloped private sector was able to deliver the expected development outcomes. Sooner or later, such governments were supplanted by more politically radical leaders, who preached the need for governments to quickly improve the lot of the masses through more active intervention. These relatively interventionist, but initially popular, governments also tended to resort to autocratic measures, as well as redistribution (regional or vertical), to solidify their political position. Resorting to controls is consistent with such a strategy for two reasons: government control of resources through marketing boards is intended to provide monopolistic/monoposonistic rents for redistribution, and price and foreign exchange controls that result from the rationing of resource shortages by the state or policies to reward the elite constituency (overvaluation of foreign currency, for example) provide opportunities for economic rent for the political coalition. Thus it is not surprising to observe that the incidence of syndromes, particularly controls and intertemporally unsustainable spending, increased into the 1970s (Fosu 2005). The Role of the Military In most cases where radical governments won out at the time of independence, the economic outcome was eventually dismal. Such regimes usually became authoritar- ian, making the military the only option for change. In many cases attempted coups did not succeed, leading to even harsher measures by the targeted government and further deterioration in economic conditions (Fosu 2002).12 The inability to ascend to power through the ballot box led to the tendency to usurp power unconstitution- ally, through the military in the form of coups and through armed rebellions when the coup route was foreclosed.13 Political instability and state failure were the likely outcomes. Importance of Supply Shocks The role of supply shocks in the adoption of control regimes or intertemporal unsus- tainable spending booms should not be underestimated. In many cases, droughts, negative international supply shocks, or both led to the adoption or strengthening of government controls. Governments usually misinterpreted positive supply shocks as permanent rather than transitory and engaged in spending booms that could not be sustained.14 Fiscal difficulties often resulted, forcing governments to seek assistance through some form of structural adjustment measures administered by the IMF/ World Bank, thus beginning the process of adopting market-friendly policies. Rational Political Choice Adopting policies that are inconsistent with market forces--and usually lead to syndromes--is often consistent with the political interests of the ruling elite. In the absence of general elections, the most important political coalitions in support of the 58 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL government are the elites, most of whom reside in urban areas. Thus, overtaxing rural production, such as cash crops, is politically optimal (Bates 1981). So is over- valuing the domestic currency (Fosu 2003), which would provide the basis for elite political instability in the form of coups. A notable exception to the African norm of the urban elite constituting the elec- toral coalition for governments is the case of Botswana, where the ruling party drew its support from the rural sector and therefore reflected rural interests. It avoided urban-bias policies, such as overtaxing rural production through state marketing boards and overvaluing the exchange rate. By not going the way of one-party states, Botswana also succeeded in allowing the voices of various interest groups to be heard. This direction probably helped provide room for addressing grievances, obvi- ating the need for coups or armed conflicts. Rational-actor models suggest that policy choices may be heavily conditioned by growth opportunities. Gallup and Sachs (1999) argue, for example, that political elites are less likely to adopt outward-oriented growth strategies in landlocked than in coastal countries, because the economic benefits from openness are too small or too far in the future to justify the up-front costs of promoting exports. The distribu- tion of syndromes by opportunities (table 2) provides modest support for this view. Individually and in combination, anti-growth syndromes were slightly more likely to emerge in the low-opportunity landlocked and resource-poor group than in the high- opportunity coastal or resource-rich group. Much stronger patterns emerge when the population is weighted (bottom panel of table 2). Large countries in the coastal and resource-poor group did decidedly better at avoiding all syndromes, while large resource-rich countries were particularly prone to inefficient redistribution and intertemporal errors. Achieving a Syndrome-Free State How was Botswana able to maintain its multiparty democracy? The homogeneity of its electoral coalition and the weakness of the opposition parties meant that the rul- ing party had little threat of being replaced. Thus, multiparty politics could be accommodated without the usual interethnic rivalries holding sway. Such rivalries led many African leaders at independence to view the creation of one-party states with strong central governments as the solution; they were concerned that multiparty politics would exacerbate ethnic polarization. Unfortunately, the authoritarian nature of these governments, coupled with latent ethnic rivalries, led to competition for economic rents created by government controls of the economy. Ethnic polariza- tion motivated many African leaders at independence to pursue policies meant to hold the newly created states together. Furthermore, those policies were consistent with a major paradigm in favor of the important role of government in the develop- ment process. Unfortunately, these policies also created anti-growth syndromes. The more recent syndrome-free regimes in Africa can be attributed to the fiscal and other economic difficulties in which many of these countries found themselves. Governments throughout Sub-Saharan Africa sought assistance to escape these diffi- culties. Ultimately, they turned to the IMF and World Bank, which prescribed pro- grams that required them to adopt market-friendly policies. The ultimate test is the EXPLAINING AFRICAN ECONOMIC GROWTH | 59 extent to which these policies are maintained, especially in the light of negative shocks, such as severe droughts and other natural calamities, deterioration in the international terms of trade, and violent conflicts. Unfortunately, as the number of syndrome-free cases has increased since the late 1980s, so has the frequency of state breakdowns. The challenge is to attenuate this syndrome without a return to the other policy syndromes. Concluding Observations The ability of African governments to continue to intervene in markets on a large scale began to decline in the late 1970s, as the continent's fiscal health and overall performance deteriorated in the face of external shocks and, in some cases, dwindling external support. By the early 1980s, structural adjustment programs dominated the policy debate in many African countries. The liberalizing agenda of these programs reflected a variety of influences, including the unsustainability of increasingly distor- tionary control regimes, the rise of conservative governments in Europe and the United States, the elimination of the Soviet Union as the exemplar and main patron of central planning, and the spectacular success of outward-oriented strategies among the East Asian manufacturing exporters. A decade of cumulative reforms had brought most African governments to syndrome-free status by the mid-1990s. African growth rebounded strongly in the late 1990s. This is consistent with the empirical results presented here, which suggest that syndrome-free status was worth some 2 percentage points of predicted growth during the 1960­2000 period. This is not a small impact: other things equal, avoiding syndromes over the full period would have supported continent-wide growth at roughly the rate attained by the industrial countries. Under these conditions, average GDP per capita in Sub-Saharan Africa today would be more than twice its current magnitude. What are the prospects for truly rapid growth in Sub-Saharan Africa--that is, annual per capita growth of 4 percent or more? Of the countries studied here, only Botswana, Mauritius, and Uganda achieved this outcome for extended periods. They did so, in large part, by steering clear of syndromes--by guaranteeing "peace, easy taxes, and a tolerable administration of justice," identified centuries ago by Adam Smith as the core functions of government (Cannan 1904, p. xxxv). Extrapolating from these findings, avoiding anti-growth syndromes will remain a necessary condi- tion for rapid growth in Africa. The manifest difficulties of maintaining such status, and the implications of failure to do so, are well illustrated by the contemporary cases of Côte d'Ivoire and Zimbabwe. Two important questions arise from the results presented here. First, what accounts for the variation in growth outcomes among the syndrome-free episodes? Global econometric evidence suggests important roles for geography and resource endowments, luck (contrast copper and oil), and gradations of leadership and insti- tutional quality that are not captured by the dichotomous syndrome/syndrome-free distinction. Much can be learned from the country studies about how these factors affect the growth environment in particular contexts. The growth literature also iden- tifies ethno-linguistic fractionalization as an anti-growth factor, one that operates 60 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL through the ability of governments to maintain growth-oriented policies (Easterly and Levine 1997). O'Connell and Ndulu (2000) and O'Connell (2004) do not find significant effects of ethno-linguistic diversity on growth once adequate controls were made for Africa's geography and initial conditions. The case study evidence suggests a potentially powerful role for ethno-regional polarization--a situation in which a few large, regionally based groups dominate the political loyalties of a substantial fraction of the population. Further work is required to develop adequate measures of ex ante polarization and assess their impact on growth through conflict, economic policy, or both (O'Connell 2004 makes a start). Second, are there arenas in which failures to undertake proactive intervention, especially in response to opportunities--errors of omission, in effect, rather than the errors of commission emphasized here (Collier and Gunning 1999)--represent bind- ing constraints on rapid growth? Here the case study evidence is necessarily less informative, given the limited experience of sustained rapid growth in Africa. Uganda's experience from 1986 to 2000, for example, was primarily one of recov- ery from the depths of Amin and post-Amin breakdown (1971­86). The Ugandan strategy during the post-1986 period represents as clear a case of prioritizing core functions--in effect, of achieving syndrome-free status--as can be found in the development literature (Reinikka and Collier 1999; Kasekende and Atingi-Ego 2004). To date, therefore, Uganda's lessons have to do with the power of syn- dromes, not with the details of growth strategy for landlocked and resource-poor countries. Globally, landlocked countries have significantly lower incomes than their coastal neighbors, on which they rely disproportionately not just for transport services but also for markets, including markets for labor services. Within Africa the relatively strong long-run growth performance of Lesotho and Swaziland within the landlocked group may in part reflect their close integration with the South African economy. The integration of regional markets, both for goods and for transport and labor services, should be a high priority for Africa's landlocked economies. Botswana's growth has been driven by the development of very favorable mineral potential. Globally, mineral rents constitute a low-mean, high-variance growth oppor- tunity, with favorable outcomes dependent on good luck with commodity prices and on avoiding the intertemporal and state breakdown syndromes. Transparent and pru- dent management of rents is therefore at the heart of what Botswana has to teach other resource-rich African countries. In the longer run, success requires economic diversification, an area in which Botswana's achievements have been limited both by its tiny population and, perhaps, by its membership in a customs union dominated by South African manufacturing firms. Africa's larger resource-rich economies will have to look to examples from outside Africa for lessons in economic diversification.15 Perhaps the most daunting and important puzzle in the African growth experience is the failure of its coastal, resource-poor economies to participate in the explosion of manufactured exports from developing countries that occurred in the last quarter of the 20th century (Collier and O'Connell 2005). The global evidence suggests that countries such as Kenya, Mauritius, and Senegal faced very favorable growth oppor- tunities during this period. But only Mauritius--after a disastrous experiment with EXPLAINING AFRICAN ECONOMIC GROWTH | 61 import-substituting industrialization--managed to capitalize on its low transport costs and abundant labor by developing a strategy to attract domestic and foreign investment into export-oriented manufacturing. In the early 1970s, Mauritius adopted an outward-oriented, shared-growth strategy that focused on attracting new investment and creating jobs (Madhoo and Nath 2003). This required keeping labor costs low, which Mauritius achieved partly through a rapid drive toward universal primary education and partly through a combination of wage restraints, a competi- tive exchange rate, and the provision of nonwage benefits (for example, housing sub- sidies) to low-income workers. For coastal, resource-poor African economies, a central focus of a growth strategy should continue to be the achievement of competitiveness in nontraditional export mar- kets, particularly in manufacturing but also in services. Avoiding syndromes will play a key and cumulative role here; Collier and O'Connell (2005) find that diversification into nontraditional export markets depends in part on the duration of syndrome-free status, so that countries, such as Senegal, that achieved such status in the early 1990s tended to see substantial diversification later in the decade. Such diversification bodes well for sustained growth and development, as the Asian and other examples have clearly demonstrated. A critical and unresolved question is whether an export-led growth strategy is sub- ject to major threshold effects in the area of human development. If it is, achieving success may require a massive initial impetus to increase the quantity and quality of investment in health and education. At the beginning of its successful export push, Mauritius did indeed have a very substantial head start in terms of human develop- ment relative to the rest of coastal Africa. Life expectancy at birth, for example, was 62 years in Mauritius in 1970--significantly higher than the 48 year average in Ghana, Kenya, Senegal, and South Africa. Mauritius was already well into the demo- graphic transition, with a total fertility rate of 3.7, far lower than the 6.9 in the com- parison group. Enrollment rates in primary, secondary, and tertiary education were also much higher in Mauritius.16 Other things equal, these advantages increase the short-run economic returns to a labor-intensive export-led growth strategy, thereby raising the political returns to such a choice relative to a policy of predation on exist- ing trade (Gallup and Sachs 1999). Indivisibilities in human resources investment, if present, would only strengthen the leverage of this argument in explaining the differ- ent choices of Mauritius and the rest of coastal Africa during the critical years of the 1970s and early 1980s. By 2000 slow but steady progress in raising literacy rates and educational enroll- ments had put many African coastal economies into a position comparable to that of Mauritius on the eve of its export drive, suggesting perhaps that remaining thresh- olds, if any, are modest. But demographic indicators continue to lag and, in some cases, such as life expectancy, show a dramatic reversal of progress since the mid- 1990s, particularly in Southern Africa. Moreover, the environment for manufactured exports from poor countries is more difficult now than it was in the early 1980s, as major new exporters (such as China and India) reap the benefits of early policy reforms, subsequent agglomeration, and most recently, improved access to industrial 62 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL country markets for textiles and apparel. Public investments in education and espe- cially health are therefore likely to play a particularly critical role in successful growth strategies among Africa's coastal economies. The fundamental challenge for growth researchers is to understand the processes through which durable and pro-growth bargains are identified and sustained among political elites and between political elites and the populations that sustain them in power. Sub-Saharan Africa's experience between 1960 and 2000 suggests that the achievement and maintenance of syndrome-free status must be a central operational objective. Annex TABLE A1. Real GDP Growth in Sub-Saharan Africa, 1961­2000 (percent) Country 1961­5 1966­70 1971­5 1976­80 1981­5 1986­90 1991­5 1996­2000 Angola -- -- -- -- -- 3.3 -3.8 6.5 Benin 3.3 2.7 1.4 4.1 4.7 0.9 4.2 5.3 Botswana 6.3 11.0 18.2 12.2 10.0 11.9 4.1 6.3 Burkina Faso 3.0 2.9 3.1 3.6 4.2 2.6 3.8 4.3 Burundi 1.9 7.6 0.6 4.2 5.4 3.7 -2.2 -1.0 Cameroon 2.7 1.6 6.7 6.9 9.4 -2.2 -1.9 4.7 Cape Verde -- -- -- -- 8.6 3.5 5.2 6.4 Central African 0.7 3.2 2.0 0.7 2.3 0.0 1.1 2.4 Republic Chad 0.7 1.4 0.9 -4.5 9.2 1.9 2.4 2.3 Comoros -- -- -- -- 4.3 1.6 0.9 1.0 Congo, Dem. 2.8 3.8 2.5 -1.5 1.9 0.0 -7.1 -3.9 Rep. of Congo, Rep. of 3.4 5.0 8.0 5.2 10.6 -0.3 0.7 2.5 Côte d'Ivoire 8.0 9.7 6.4 4.5 0.3 1.2 1.5 3.5 Djibouti -- -- -- -- -- -0.7 -1.8 -0.2 Equatorial -- -- -- -- -- 1.4 7.0 35.7 Guinea Eritrea -- -- -- -- -- -- 12.5 1.2 Ethiopia -- -- -- -- -0.9 5.3 1.5 5.3 Gabon 8.2 5.6 18.1 0.4 2.6 1.7 3.1 1.8 Gambia, The -- 4.5 5.5 4.4 3.2 4.1 2.1 4.8 Ghana 3.1 3.0 0.0 1.0 -0.3 4.8 4.3 4.3 Guinea -- -- -- -- -- 4.5 3.7 4.2 Guinea-Bissau -- -- 3.2 -0.6 6.4 3.8 3.2 1.1 Kenya 3.5 5.9 10.0 6.3 2.5 5.6 1.6 1.8 Lesotho 7.6 2.8 8.6 12.3 3.2 5.9 4.0 3.0 Liberia 3.2 6.6 1.6 2.2 -1.9 -16.5 -21.7 38.3 Madagascar 1.4 4.7 0.7 1.5 -1.5 2.7 -0.3 3.8 Malawi 4.6 5.0 7.6 4.9 2.2 2.3 3.5 3.9 Mali -- 3.4 3.4 4.9 -2.2 3.9 3.0 3.9 Mauritius -- -- -- -- 4.3 7.4 5.1 5.3 EXPLAINING AFRICAN ECONOMIC GROWTH | 63 TABLE A1. continued (percent) Country 1961­5 1966­70 1971­5 1976­80 1981­5 1986­90 1991­5 1996­2000 Mozambique -- -- -- -- -4.6 5.6 3.5 8.0 Namibia -- -- -- -- -0.2 2.7 5.0 3.5 Niger 6.3 -0.5 -2.1 5.4 -2.3 2.6 0.8 2.9 Nigeria 4.5 5.6 5.8 4.1 -2.8 5.4 2.5 2.8 Rwanda -1.6 7.6 0.8 10.3 2.7 1.5 -4.0 9.8 São Tomé -- -- -- -- -- 1.8 1.6 2.1 and Principe Senegal 2.0 2.0 2.5 1.2 3.2 3.2 1.5 5.3 Seychelles 3.7 3.8 7.1 8.6 0.9 5.6 2.9 6.4 Sierra Leone 4.4 4.2 2.4 2.3 0.9 1.1 -5.0 -3.3 South Africa 6.8 5.8 4.6 2.5 0.9 1.8 0.9 2.6 Sudan 1.9 1.4 5.0 2.7 0.8 4.6 5.1 6.1 Swaziland -- -- 9.6 3.2 2.6 11.2 2.8 3.3 Tanzania -- -- -- -- -- -- 1.8 4.2 Togo 10.1 6.7 3.7 5.1 -0.2 2.5 0.6 2.3 Uganda -- -- -- -- 0.7 5.1 7.0 6.5 Zambia 6.2 1.6 2.5 0.4 0.5 1.6 -1.3 2.8 Zimbabwe 3.6 9.4 4.9 1.7 4.4 4.6 1.4 2.1 Sub-Saharan 5.1 5.0 4.8 2.8 1.2 2.7 1.6 3.6 Africa Sub-Saharan 3.5 4.1 4.9 3.2 1.5 3.4 2.2 4.2 Africa excluding South Africa Source: World Development Indicators CDROM (World Bank 2004). -- Not available. Note: Overall means are weighted by GDP. Notes 1. This section draws on Collier and O'Connell (2005) and O'Connell (2004). The first com- prehensive presentation of the synthesis taxonomy was written by Collier as a background note for the Growth Project conference in November 2003. 2. North Africa is often treated separately from Sub-Saharan Africa in the development lit- erature. Fosu (2001) emphasizes sharp differences in economic growth between the two regions. The precipitous decline in African growth as of the mid-1970s, for example, appears to be a phenomenon of Sub-Saharan Africa rather than of Africa as a whole. Between the 1960­74 and the post-1974 periods, average annual GDP growth of Sub- Saharan Africa fell by more than half, from 5.4 percent to 2.0 percent, while growth in North Africa declined only slightly, from 5.1 percent to 4.4 percent. 3. See the annex for an indication of the nonsustainability of GDP growth for many African countries based on five years of data. 4. The first stage of this exercise was undertaken by Jean-Paul Azam, Robert Bates, Paul Col- lier, Augustin Fosu, Jan Gunning, Benno Ndulu, and Stephen O'Connell in August 2003, based on draft versions of the country studies. The classification was assessed by country authors at a November 2003 conference and refined in response to their comments. In 64 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL August 2004 the editorial committee (including Dominique Njinkeu) undertook a similar exercise to extend the sample to most of Africa. 5. The first coup was actually "staged from a popular uprising demanding the departure of the president and not by a conventional coup d'état carried out by the army" (Savadogo, Coulibaly, and McCracken 2003, p. 6). Labor unions played a key role in organizing the uprising; the excesses of the first president, Maurice Yameogo (1960­6), also played a catalytic role. 6. Fosu (1992) attributes the unwillingness to devalue to fears of precipitating a military coup. 7. Fosu (2002) shows that a successful coup may raise growth when economic performance is extremely poor and investment is low. But there is no guarantee that a coup attempt suc- ceeds. An abortive coup is likely to exacerbate the growth plight. 8. This section draws on Fosu (2005). 9. See, for example, Ndulu (2004) for a discussion of the influence of international para- digms on African leaders' policy choices. 10. The political bent of the early leadership, in contrast, appears to a large extent to be due to chance. The case study sample reveals about the same number of syndrome-free cases as syndromes in 1960 (Fosu 2005). 11. There were several notable exceptions. El Hajj Ahmadou Ahidjo of Cameroon voluntar- ily retired in 1982, Félix Houphouet-Boigny of Côte d'Ivoire and Hastings Banda of Malawi continued in power until they retired, and Botswana was able to maintain its multiparty democracy without military intervention. One reason behind Botswana's achievement might be that there was little ethnic rivalry to be taken advantage of by coali- tions of a fractionalized military. Another, perhaps more significant, is that the military was not an important institution in Botswana. 12. See, for example, the case of Cameroon, where following the failed coup of 1984, Presi- dent Paul Biya became more authoritarian and engaged in a strong ethnically redistribu- tive process to shore up his political support. 13. The genesis of the Liberian civil war, for example, is traced in large part to the inability of the regular military to overthrow Samuel Doe (Davies 2005). 14. Alternatively, governments may undertake short-term projects, especially if they have a relatively short expected political lifespan due to a high risk of government overthrow. 15. Indonesia leveraged its 1974­9 oil boom into a doubling of the number of primary schools, strengthening the platform for its subsequent diversification both within the pri- mary commodity sector and into manufactured exports (Duflo 2001). 16. The human development data, including enrollment rates, are from the World Bank's World Development Indicators. References Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. "The Colonial Origins of Comparative Development: An Empirical Investigation." American Economic Review 91 (5): 1369­1401. Aryeetey, Ernest, and Augustin K. Fosu. 2003. "Economic Growth in Ghana: 1996­2000." AERC Growth Project, Nairobi. Azam, Jean-Paul. 1995. "How to Pay for the Peace? A Theoretical Framework with Reference to African Countries." Public Choice 83 (1­2): 173­84. ------. 2001. "The Redistributive State and Conflicts in Africa." Journal of Peace Research 38 (4): 429­44. ------. 2005. "Political Geography and Redistribution." Paper presented at the AERC/Weatherhead Center Workshop on Explaining African Economic Growth, Harvard University, Cambridge, MA, March 18­19. EXPLAINING AFRICAN ECONOMIC GROWTH | 65 Azam, Jean-Paul, and Nadjiounom Djimtoingar. 2004. "Cotton, War and Growth in Chad, 1996­2000." AERC Growth Project, Nairobi. Bates, Robert H. 1981. Markets and States in Tropical Africa. Berkeley: University of Califor- nia Press. ------. 2005a. "Political Insecurity and State Failure in Contemporary Africa." CID Work- ing Paper No. 115, Center for International Development, Harvard University, Cam- bridge, MA. ------. 2005b. "Political Reform." CID Working Paper No. 114, Center for International Development, Harvard University, Cambridge, MA. Blattman, C., J. Hwang, and J. G. Williamson. 2004. "The Impact of the Terms of Trade on Development in the Periphery, 1870­1939: Volatility and Secular Change." NBER Work- ing Paper No. 10600, National Bureau of Economic Research, New York. Bratton, Michael, and Nicolas van de Walle. 1997. Democratic Experiments in Africa. Cam- bridge: Cambridge University Press. Cannan, Edwin. 1904. Editor's Introduction to Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations. New York: G.P. Putnam's Sons. Collier, Paul, and Jan Willem Gunning. 1999. "Explaining African Economic Performance." Journal of Economic Literature 37 (1): 64­111. Collier, Paul, and Stephen A. O'Connell. 2005. "Chapter 2: Opportunities, Episodes, and Syn- dromes." Paper presented at the AERC/Weatherhead Center Workshop on Explaining African Economic Growth, Harvard University, March 18­19. Davies, Victor A. B. 2005. "Sierra Leone's Growth Performance: 1961­2000." Paper pre- sented at the AERC/Weatherhead Center Workshop on Explaining African Economic Growth, Harvard University, Cambridge, MA, March 18­19. Dehn, J. 2000. "The Effects on Growth of Commodity Price Uncertainty and Shocks." World Bank Policy Research Working Paper Series No. 2455, Washington, DC. Dollar, David. 1992. "Outward-Oriented Developing Countries Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976­85." Economic Development and Cultural Change 40 (3): 523­44. Duflo, Esther. 2001. "Schooling and Labor Market Consequences of School Construction in Indonesia: Evidence from an Unusual Policy Experiment." American Economic Review 91(4): 795­813. Easterly, William, and Ross Levine. 1997. "Africa's Growth Tragedy: Policies and Ethnic Divi- sions." Quarterly Journal of Economics 112 (4): 1203­50. ------. 1998. "Troubles with the Neighbours: Africa's Problem, Africa's Opportunity." Jour- nal of African Economies 7 (1): 120­42. Fosu, Augustin K. 1992. "Political Instability and Economic Growth: Evidence from Sub- Saharan Africa." Economic Development and Cultural Change 40 (4): 829­41. ------. 2001. "The Global Setting and African Economic Growth." Journal of African Economies 10 (3): 282­310. ------. 2002. "Political Instability and Economic Growth: Implications of Coup Events in Sub-Saharan Africa." American Journal of Economics and Sociology 61 (1): 329­48. ------. 2003. "Political Instability and Export Performance in Sub-Saharan Africa." Journal of Development Studies 39 (4): 68­82. ------. 2005. "Anti-Growth Syndromes in Africa: A Synthesis of the Case Studies." Paper presented at the AERC/Weatherhead Center Workshop on Explaining African Economic Growth, Harvard University, Cambridge, MA, March 18­19. 66 | AUGUSTIN KWASI FOSU AND STEPHEN A. O'CONNELL Gallup, John L., and Jeffrey D. Sachs. 1999. "Geography and Economic Development." In Annual World Bank Conference on Development Economics 1998, ed. Boris Pleskovic and Joseph E. Stiglitz, 127­70. Washington, DC: World Bank. Glaeser, Edward L., Rafael La Porta, F. Lopez-de-Silanes, and Andrei Shleifer. 2004. "Explaining Growth: Institutions, Human Capital, and Leaders." Paper presented at the Brookings Institution Panel on Economic Activity, Washington, DC, March. Gogué, Tchabouré A., and Kodjo Evlo. 2004. "Togo: Lost Opportunities for Growth." AERC Growth Project, Nairobi. Guillaumont, Patrick, Sylviane Guillaumont Jeanneney, and Jean-François Brun. 1999. "How Instability Lowers African Growth." Journal of African Economies 8 (1): 87­107. Karmarck, Andrew. 1971. The Economics of African Development, Revised Edition. New York: Praeger Publishers. Kasekende, Louis A., and Michael Atingi-Ego. 2004. "Uganda Country Case Study." AERC Growth Project, Nairobi. Kobou, Georges, and Dominique Njinkeu. 2004. "Political Economy of Cameroon Post- Independence Growth Experiences." AERC Growth Project, Nairobi. Levine, Ross, and David Renelt. 1993. "A Sensitivity Analysis of Cross-Country Growth Regressions." American Economic Review 82 (4): 942­63. Madhoo, Yeti N., and Shyam Nath. 2003. "Explaining African Economic Growth Perfor- mance: The Case of Mauritius." AERC Growth Project, Nairobi. Maipose, Gervase S., and T. C. Matsheka. 2004. "Explaining African Growth Performance: The Case of Botswana." AERC Growth Project, Nairobi. Maundeni, Z. 2001. "State Culture and Development in Botswana and Zimbabwe." Journal of Modern African Studies 40 (1): 105­32. Meade, James E. 1961. The Economic and Social Structure of Mauritius: Report to the Gov- ernor of Mauritius. London: Methuen. Ndulu, Benno J. 2004. "African Growth and Development Paradigms." Draft chapter of Growth Project synthesis volume, Nairobi. Ndulu, Benno J., and Stephen A. O'Connell. 2005. "Policy Plus: African Growth Performance 1960­2000." Paper presented at the AERC Plenary Session on Explaining African Eco- nomic Growth, Nairobi, May. Nkurunziza, Janvier D., and Floribert Ngaruko. 2003. "Economic Growth in Burundi from 1960­2000." AERC Growth Project, Nairobi. O'Connell, Stephen A. 2004. "Africa's Growth Experience: Emerging Lessons from the Growth Project." Paper presented at the AERC Plenary Session, Nairobi, May 29. O'Connell, Stephen A., and Benno J. Ndulu. 2000. "Africa's Growth Experience: A Focus on Sources of Growth." Framework paper for the AERC Growth Project, Nairobi. Reinikka, Ritva, and Paul Collier, eds. 1999. Uganda's Recovery: The Role of Farms, Firms and Households. Washington, DC: World Bank. Savadogo, Kimseyinga, Siaka Coulibaly, and Coleen A. McCracken. 2003. "Analyzing Growth in Burkina Faso over the Last Four Decades." AERC Growth Project, Nairobi. Temple, Jonathan. 1999. "The New Growth Evidence." Journal of Economic Literature 37 (1): 112­56. World Bank. 1981. Accelerated Development in Sub-Saharan Africa: An Agenda for Action. Washington, DC: World Bank. ------. Various years. World Development Indicators. Washington, DC: World Bank. Comment on "Explaining African Economic Growth: The Role of Anti- Growth Syndromes," by Augustin Kwasi Fosu and Stephen A. O'Connell JEAN-PAUL AZAM This paper provides a very useful and exciting account of the outcome of the research performed within the AERC collaborative project on the African Growth Experience. It brings out nicely the idea that there is no point in talking about Africa as a homo- geneous group of countries and that most countries have experienced several growth regimes. It is thus useful to take the growth episode as the unit of observation. Some typology is required for getting the right focus on the problems that are retarding growth in these countries in some of these episodes. Here the criterion chosen for pin- ning down the right kind of heterogeneity is a combination of two variables: captur- ing opportunities and suffering from syndromes. Regarding opportunities, much is made of the simple three-way classification of countries as resource-rich, coastal, or landlocked. The coastal and landlocked categories classify the resource-poor countries according to a rough measure of the transportation costs they face. It is obvious why these three types of countries face different sets of opportunities for growing. A strong point emerges from the analysis of growth performance according to this classifica- tion: Africa's comparatively bad growth performance is due mainly to the fact that coastal countries have not performed as well as coastal countries on other continents. Resource-rich and landlocked countries in Africa are performing roughly the same as developing countries in the same categories on other continents. This is briefly men- tioned in this paper and developed at length elsewhere (Collier and O'Connell 2004). Africa's generally disappointing growth performance in the 1980s and 1990s compared with other developing countries can be explained to a large extent by just four syndromes. Their impact depends on the opportunities each country faced. The first syndrome refers to a particularly disruptive way of extracting resources-- by imposing controls on markets and firms and more generally distorting the func- tioning of markets. Hard controls refer to the type of economic systems modeled more or less after the Soviet system; soft controls refer to milder socialistic systems. In both cases the damage is done by extracting rents by the worst method--by distorting and disturbing markets. Jean-Paul Azam is professor of economics at Université Toulouse 1 Sciences Sociales. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 67 68 | JEAN-PAUL AZAM The next two syndromes place more emphasis on the use made of those resources than on the way they are extracted. They may be viewed as the social causes behind the latter. In the redistributive syndrome, actual or anticipated, what matters is that the initial distribution of wealth across groups is too unequal for peace to be sustain- able. In this case, some resources have to be redistributed to reduce polarization. Var- ious transactions costs and collective action problems may make such redistribution very costly socially. A very special method for redistribution occurs when a public investment boom is used by the government to transfer resources from the future to the present. This is labeled the intertemporal syndrome. It is often triggered by a commodity boom, which allows the ruling group to get hold of the windfall, often amplified by the addi- tional foreign borrowing that is made easier by the initial boom. When the boom busts, the subsequent adjustment cost is often passed on to some other groups. The last syndrome in the list is state failure. It includes both civil war and extreme political instability. The reference group--countries affected by none of these syndromes--is called syndrome free. Very few countries escape from these syndromes in the resource-rich and the land- locked categories, as shown in table 2, in which incidence is measured as a percent- age of the population rather than relative to the number of countries. The fact that performance in these countries is similar to that of comparable countries on other continents suggests that the occurrence of these syndromes is to some extent endoge- nous to the type of opportunities open to them. This is particularly clear for the resource-rich countries, which are especially exposed to the intertemporal syndrome. These countries can be labeled Hotelling countries. Exhaustible resources have long been known to raise fascinating problems of intertemporal allocation of resources. They are also highly exposed to the static redistribution syndrome. What is striking in table 2 is that the most problematic countries, from the viewpoint of explaining Africa's poor growth performance--that is, the coastal countries--are predomi- nantly affected by the regulatory syndrome. They are much less prone to the other syndromes. The econometric results show the usefulness of this syndrome approach. All are significant with the predicted sign, at least in the equations without country fixed effects. This is an interesting result, which should be emphasized in the paper: only the most political syndromes--namely, the regulatory and state collapse syndromes-- survive the inclusion of country fixed effects. This point comes out in comparing the last column to the other columns in table 7. This suggests that the other two syn- dromes are much more determined by geography and somehow reflect deeper prob- lems, mainly related to the presence of resources. They do not really belong to the realm of policy issues, at least in the short run. Combining this insight with the observations made about table 2 suggests that countries are either exogenously exposed to redistribution problems (by geography or ethnic division) or they are not. If they are, they have to choose between a political strategy of redistribution and one of state collapse. This choice was brought out in the conflict literature by Azam and Mesnard (2003). COMMENT ON FOSU AND O'CONNELL | 69 Table 2 suggests that resource-rich countries have generally avoided state collapse by choosing redistribution. This conclusion is strikingly at variance with the strand of the literature on conflict that regards state collapse as a consequence of the so- called resource curse. Landlocked countries seem to be much more exposed to state collapse, perhaps because rulers have fewer resources with which to buy peace. Moreover, for these countries the opportunity cost of going to war is low, because of the low returns to labor. Peaceful countries have to choose a method for extracting resources. Doing so by imposing controls creates both rents and distortions, which are socially more costly than orderly taxation. Depending on whether population weights are used, table 2 shows that between a third and a half of the countries chose controls. The economet- ric results in table 7 suggest that this choice is not fully determined by country char- acteristics, as this syndrome survives the inclusion of country fixed effects. The emerging picture suggests that politics matters greatly in explaining why African performance--particularly the performance of coastal countries, which are affected by the most political syndromes--differs from that of other developing countries. The paper presents two views explaining why this is the case. One is the power of ideas: the crippling regulatory regimes were imposed because wrong ideas--Fabian socialism, for most of the affected countries, with some deeper Marx- ian influences for the most extreme cases--were adopted wholesale from Europe (Ndulu 2004). Many early rulers in independent African countries referred to these European models in their speeches and writings, despite the fact that the type of social structure prevailing in Africa is completely different from that in Europe. Since the mid-19th century, what has mattered in Europe has been the division of society along class lines. In contrast, in Africa ethnic divisions are paramount. This simple view is only partially satisfactory, because it does not explain why African leaders were fooled by these ideas. Perhaps there was a fashion effect, as many of the early post-independence leaders had been trained in Europe or the United States. However, it is also clear that these ideas were exceedingly convenient for political window dressing. We know from Marx that interest groups are prone to creating a good ideo- logical image for themselves and that "history comes behind a mask." An alternative view is that these sets of ideas were ideal for hiding the social and political progression of the urban elite groups. The state-sponsored, educated elite needed some strong justification for taking over these countries; the type of socialist thinking that prevailed in many of these countries was ideal for providing the ideo- logical underpinning they needed. Let me suggest the type of scheme we need for understanding more deeply why Africa grows slowly--the "triangular redistribution game." This game allows one to articulate the two driving forces: ethnicity and the rise of the urban educated elites. Ethnic groups are based on traditional social structures, with their power base centered on the village society. These groups invest in the migration of their most promising offspring, which involves both a geographical migration to the city and a crucial cultural mutation through the schooling system. There is a well-known edu- cation bias among migrants. All groups send "delegates" to the urban scrum, where 70 | JEAN-PAUL AZAM the postcolonial state provides many opportunities for acquiring wealth and political influence. Throughout their lives, these migrants will have to send remittances back to the village, especially in times of hardship. Their fundamental role is to provide some insurance services, to shield the standard of living of those left behind from the usual shocks, including droughts or disease. They will also support their parents dur- ing their old age. The resources used for this redistribution from migrants to villagers can best be accumulated by working in the urban formal sector. In many poor coun- tries, this sector consisted largely of the public sector, where regular, high-paying jobs could be found. A very important part of migrants' duties is to serve as bridgeheads for helping sub- sequent generations of migrants find jobs and housing in the urban sector. Migrants thus create social networks, which provide various services to their members, who share the duty of supporting those who stay behind in the village. The most powerful of these "delegates" will be able to influence the allocation of public investment across regions, either as politicians in democratic systems or as prominent officials in the public sector. These people are the main channels through which redistribution across groups takes place, when it is needed. The urban elite that emerges naturally has a clear incentive to expand its influence, to distribute patronage within its social net- works. In the short run, a state-centered development pattern, as advocated by the various brands of socialist doctrines, provides the most convenient medium for such a patronage system. Development of the oversized parastatals that were so common in many African countries finds its roots in this system. It is much more difficult to develop these networks across the private sector. Foreign firms, in particular, are espe- cially difficult to control by these social networks. It takes a farsighted political elite to understand that such a system is self-defeating. These unproductive units are not self-sustaining and require an ever-expanding extrac- tion of resources from the productive sector. The system gradually destroys the incen- tives to produce, crippling the economy. In the best cases, the elite understand this and work collectively as an Olsonian "stationary bandit." In the long run, there is more for the elite in a thriving economy, where incentives for production and investment are well protected. References Azam, Jean-Paul, and Alice Mesnard. 2003. "Civil War and the Social Contract." Public Choice 115 (3): 455­75. Collier, Paul, and Stephen O'Connell. 2004. "Opportunities, Episodes, and Syndromes." Department of Economics, Swarthmore College, Swarthmore, PA. Ndulu, Benno. 2004. "African Growth and Development Paradigm." World Bank, Washing- ton, DC. Comment on "Explaining African Economic Growth: The Role of Anti-Growth Syndromes," by Augustin Kwasi Fosu and Stephen A. O'Connell ABDOULAYE DIAGNE Since the end of the 1970s, African societies have been confronted with a profound crisis in the growth of their economies. Development economists have tried to find the key to restarting the engine of African growth and to keep it running under a sys- tem capable of significantly reducing poverty as quickly as possible. If the Berg Report of 1981 constitutes an obligatory reference (World Bank 1981), other works have also left their mark on this search for the crucial factor needed to increase growth above the levels of the 1960s and 1970s. Collier and Gunning (1999) propose a synthesis of both microeconomic and macroeconomic research aimed at taking into account the changes in African economies. Particular interest has always been paid to the choice of public policies, undoubtedly because this is the variable that is most under the control of African decision makers. The paper by Augustin Fosu and Stephen O'Connell is part of this tradition of researching the role that economic policies have played in accounting for growth in Africa. It makes an original contribution in several respects. First, it synthesizes work conducted by African economists themselves on the problems of growth in their con- tinent. Second, it draws on a large number of geographically diverse case studies (27), allowing conclusions to be drawn about Sub-Saharan Africa as a whole. Third, until now researchers have tried to show that policies have exercised more influence on African economies than other variables (aid, investment, the international environ- ment, and so forth). Fosu and O'Connell go farther by placing the emphasis on fac- tors that were at the origin of the choice of such policies. What does the paper teach us that is new about African economic policy? To respond, I examine the major features of African growth identified by the authors and the conditions that govern the choice of the policies they identify through the case studies. Abdoulaye Diagne is the director of the Consortium pour la Recherche Economique et Sociale (CRES) and professor of economics at the Université Cheikh Anta Diop, Dakar. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 71 72 | ABDOULAYE DIAGNE The Stylized Facts of Economic Growth in Africa The paper underscores the increasing diversity of economic conditions since the mid- dle of the 1970s, the great volatility of economic growth in each country and between countries, the near absence of countries that have maintained a place among the best performers with respect to growth over a long period, and the fact that periods of growth within countries are of varying intensity. These stylized facts have endowed African economic growth with an erratic character that could be the direct cause of the decline it has undergone since before the mid-1970s. The authors try to identify the types of policies and syndromes that were hostile to growth. These included interventionist regimes that introduced major distortions in production and that favored the rent, redistribution regimes that imposed a new allocation of assets and revenues to the detriment of economic efficiency, intertem- poral regimes that showed a systematic preference for the present to the detriment of the future, and the breakdown of the state as a result of civil war or serious political instability that prevents the normal functioning of public institutions. Can these syndromes account for weak growth in Africa? Fosu and O'Connell show that economic growth (measured by the mean or the median) is stronger in the absence of syndromes, that the probability of posting negative growth over three consecutive years is less than 20 percent in their absence, and that the probability of observing average annual growth of more than 2.5 percent (the median value of developing countries between 1960 and 2000) for five years is less than 25 percent in the presence of syndromes. This descriptive analysis is supplemented by economet- ric analysis, which consists of regressing the variation in the rates of growth on a dichotomous variable indicating, for each year, whether a country presents one or more of the syndromes. The authors control for exogenous shocks, resource endow- ments, and geographic location and its specific effects. Their estimates, which cover a sample of 46 countries, confirm the results of the descriptive statistics. First, each of the syndromes has an impact that is at once considerable and significant on the median growth rates predicted. Second, the absence of syndromes is not a sufficient condition for achieving sustained growth over the medium term. Third, the "output" achieved by refraining from adopting anti-growth policies did not fall during the 1990s relative to its level in the previous decades, even if the international environ- ment of the 1990s was unfavorable to Sub-Saharan Africa. It is worth asking whether the exogenous shocks arising from international mar- kets or natural phenomena, as well as the permanent effects specific to each country, are the only variables to be controlled for to identify the true effect of the syndromes on growth. It may be, for example, that growth in a particular country depends at least partially on favorable economic conditions in another African country on which it is dependent. It would be relevant to try to explain the growth observed in Burk- ina Faso and Mali by that of Côte d'Ivoire. The economic situation in Nigeria must surely influence the economic growth in many countries in West Africa. The influ- ence over the past 10 years of the South African economy on the economies of South- ern Africa would appear to be worthy of consideration. Just as the effects of interna- tional or natural shocks must be investigated, the effects of bordering countries must COMMENT ON FOSU AND O'CONNELL | 73 also be taken into account; the degree of interdependence of African economies is certainly greater than what is shown by the official statistics on intra-African commerce. On the econometric level, it is surprising that the authors did not attempt to ver- ify whether the causality goes only from syndromes to growth, or whether growth also influences the adoption or abandonment of syndromes that are unfavorable. One could hypothesize that a major crisis or the absence of growth over a long period reduces the ability of interest groups to defend their rent positions and favors the adoption of policies that stimulate growth. Conditions for the Appearance of Syndromes The paper tries to explain why anti-growth syndromes were adopted and why they were abandoned. Analysis of the 27 case studies in the African Economic Research Consortium (AERC) project suggests the following factors: · Most African countries opted for socialist political regimes. This ideological ori- entation favored the adoption of policies that proved unfavorable to growth. · In the absence of a democratic means of changing governments when living con- ditions deteriorate as a result of the application of economic policies that are unfa- vorable to growth, coups d'état and armed rebellions are used to change political regimes. These solutions have often created an environment that favors the adop- tion of anti-growth syndromes, such as the breakdown of the state, and the appli- cation of a redistribution policy that does not favor efficiency. · Exogenous shocks have played an important role. For example, an increase in the price of raw materials exported by a country may lead its government to signifi- cantly increase its outlays (intertemporal syndrome). When prices fall, the outlays are not reduced. The budgetary and foreign deficits that result can compel coun- tries to seek assistance from the Bretton Woods institutions, which oblige them to adopt less interventionist policies. · The sociological base of the ruling political class may explain the choice (or lack of choice) of anti-growth syndromes. If this base is mainly urban, for example, the chances of adopting measures that are unfavorable to the rural sector are extremely high. What can be said about the conditions governing the choice of policies? It is regrettable that the qualitative analysis was not supplemented by a more formal analysis evaluating the chances of adopting or abandoning a syndrome. Use of mod- els of discrete choices (for example, multinomial models) would have allowed the authors to estimate the impact of each of the factors cited above on the probability of choosing a particular syndrome. One might draw up a classification based on the relative influence of each factor. A precondition to this exercise is a complete inven- tory, beginning with the case studies, of factors specific to the appearance of a syn- drome in a country at a particular moment in its history. 74 | ABDOULAYE DIAGNE One might ask whether the democratization of political life observed in a grow- ing number of African countries since the beginning of the 1990s favors the applica- tion of policies that favor growth. Neither the case studies presented nor the synthe- sis allows us to answer the question of whether a democratic system has a greater chance of applying policies favorable to growth. More exhaustive examinations should have been carried out to better discern the impact of democratization of polit- ical life on the choice or rejection of anti-growth syndromes. Many studies of growth in Africa have underlined the crucial importance of adherence by public decision makers to good policies. It is surprising that the case studies did not try to measure the degree of internalization of these policies, espe- cially in the countries that decided to apply policies favorable to market mechanisms. Identification of a group of indicators that would allow the evaluation of the inter- nal support for policies (not only by national authorities but also by interest groups) would tell a great deal about the conditions needed to ensure the success of a pro- gram of reforms. The internalization of the reforms is not limited to public decision makers. The presence, or lack thereof, of suitable capacities to conceive and imple- ment good policies has been shown to be crucial. In the episodes of strong growth, has this condition always been confirmed in Sub-Saharan Africa? The level of political and social instability--which may manifest itself in conflicts or in factors that lead to political crisis and conflict--must play a role in the choice of economic policies. A measure of the intensity of the conflicts would have allowed the authors to better explain the adoption or abandonment of a syndrome. Toward an African Economic Policy? The AERC research project represents an important effort in attempting to under- stand African economic growth. It makes clear the need for further research to iden- tify the conditions that have been important for implementing reforms favorable to strong and sustainable economic growth. It seems necessary to construct and test hypotheses about the following questions: · Is there a political leadership with a sufficiently long time horizon to persevere in applying difficult reforms? · At the base of the conception and application of reforms, is there a team with coherent views on the economic policy to be pursued, a team on which the polit- ical authorities can depend? · Are democratic regimes better positioned than authoritarian ones to produce strong economic growth? · Where has the mobilization of public support been sought and obtained to oppose interest groups hostile to growth? · Has the average citizen noticed the economic benefits of policies that have gener- ated strong growth? If so, how long did it take? COMMENT ON FOSU AND O'CONNELL | 75 Emphasis should be placed on episodes of significant increases in production, to better understand the conditions under which growth occurs and the reasons why periods of growth have not been sustained. In analyzing the economic growth of Sene- gal from 1960 to 2000, Diagne and Daffé (2002) have shown the important role that the political economy of reforms plays in mobilizing the resources needed to sustain growth and development. This conclusion may apply to all of Sub-Saharan Africa. References Collier, P., and J. W. Gunning. 1999. "Explaining African Economic Performance." Journal of Economic Literature 38 (March): 64­111. Diagne A., and G. Daffé, eds. 2002. Le Sénégal en quête d'une croissance durable. Paris: CREA et Editions Khartala. World Bank. 1981. "Accelerated Development in Sub-Saharan Africa: An Agenda for Action." The Berg Report. Washington, DC. Comment on "Growth in Senegal: Features and Trends," by Gaye Daffé JEAN-PAUL AZAM Gaye Daffé's paper* on growth in Senegal has a different tone from the remarks by President Abdoulaye Wade and Minister Abdoulaye Diop that preceded it. While they discussed the exceptional stretch of growth enjoyed by Senegal since the 1994 devaluation of the CFA franc, Daffé looks at the long term, emphasizing the disap- pointing performance of Senegal since independence. This is a useful contrast that brings out nicely how exceptional the growth performance achieved by Senegal over the past decade has been, with a growth rate consistently exceeding 5 percent a year. Moreover, the two large-scale household surveys performed in 1994 and in 2001 (ESAM 1 and 2) show that this long episode of sustained growth entailed a massive reduction in poverty. Hence we have a good example of pro-poor growth, which seems quite encouraging for Senegal and from which some lessons are probably worth drawing for other similar countries. The devaluation cut the real wages of civil servants and public sector employees about 40 percent. At first, this provided a brake on the wages paid in the rest of the formal economy. Although the formal economy accounts for less than 5 percent of the Senegalese labor force, the cut in the real cost of formal sector labor was the key to the relatively rapid growth that followed. It affected growth through two chan- nels. First, the cut in the real wage bill of the government freed up some resources, which the government used efficiently to boost public investment after 1996. This was a crucial time to provide some help to the economy, to tide it over the recession that plagued the whole area around the turn of the century. But more than the boost in effective demand, which Keynesians would probably emphasize, it was the com- plementarity between public and private capital that most probably did the work. This is suggested by the fact that other countries in the area also boosted their pub- lic expenditures in real terms following the devaluation, with a different impact (neighboring Mali, for example, expanded its much needed public expenditures in education, without such a positive impact on growth). Second, the cut in real wage costs in the formal sector enhanced the profitability of the entire modern sector, with Jean-Paul Azam is professor of economics at Université Toulouse 1 Sciences Sociales. *The paper by Gaye Daffé was unavailable for publication at press time. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 77 78 | JEAN-PAUL AZAM both manufacturing and services enjoying rapid growth after 1996. The agricultural sector ended the 1990s with a much smaller share of GDP and of the labor force. This important transfer of labor to the urban economy and the modern sector played a decisive part in reducing poverty. Two lessons are worth drawing from the Senegalese experience. The first is that growth can be boosted even in a traditionally slow-growing economy, provided the wage bill of the modern sector is kept in check. This was the key lesson of the seminal paper by Lewis (1954), and it found a clear vindication with the Senegalese experience. The second is that poverty falls significantly if growth is sustained long enough. Pro- poor growth thus walked on two legs in Senegal: redistribution of income not from the rich to the poor but from wages to profits in the formal sector and redistribution of labor from stagnant agriculture to growing manufacturing and services. Reference Lewis, Arthur. 1954. "Economic Development with Unlimited Supplies of Labour." Reprinted in The Economics of Underdevelopment (1958), ed. A. N. Agarwala and S. P. Singh, 400­49. Delhi: Oxford University Press. Financial Reforms Financial Sector Reforms in Africa: Perspectives on Issues and Policies LEMMA W. SENBET AND ISAAC OTCHERE The financial system does more than merely mobilize savings. Its other tasks--producing information, revealing prices, sharing risk, providing liquidity, promoting contractual effi- ciency, promoting governance, and facilitating global integration--are vital to a well- functioning system. Understanding these functions is crucial in diagnosing financial sys- tems and serving as guides for policy prescriptions. This paper adopts a holistic approach to analyzing financial sector reforms in Africa. The approach recognizes the important complementarities between the development of banking systems and capital markets. The paper provides a catalogue of policy guides for developing and building the capacity of the financial sector in Africa, par- ticularly Sub-Saharan Africa. The policy prescriptions pertain to both the banking sector and capital markets, viewed broadly to include markets for stocks, bonds, and other financial instruments. The paper is organized as follows. The next section describes financial sector reforms in Africa. Section 2 examines the foundations of financial sector reforms, based on par- adigms in finance. It takes a functional perspective, recognizing that African financial systems should be designed to provide multiple functions. These functions provide a channel for linking financial sector and economic development; they constitute a basis for diagnosing the features of African financial systems. Using this functional perspec- tive, the section assesses the prospects for financial globalization of Africa based on mutual gains for global investors and the region at large. Section 3 outlines the main features of African financial systems. It describes the packages of financial sector reforms that have taken place and identifies constraints these systems face in achieving the desired outcomes of savings mobilization and financial stability. Section 4 provides a series of policy prescriptions. Section 5 identifies areas for future research. Lemma W. Senbet is the William E. Mayer Professor of Finance and chair of the Finance Department, Robert H. Smith School of Business at the University of Maryland, College Park. Isaac Otchere is associate professor of finance at the University of New Brunswick, Canada. This paper has benefited from prior work--Senbet (2001) and Aryeetey and Senbet (1999). It provides an extensive data update, particularly regarding African stock markets and financial globalization, as well as new analytical perspec- tives and updated policy prescriptions. We acknowledge comments from an anonymous referee and Mthuli Ncube. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 81 82 | LEMMA W. SENBET AND ISAAC OTCHERE Description of Financial Sector Reforms Most African countries, particularly those in Sub-Saharan Africa, have recently undergone extensive financial sector reforms. The reform package includes interest rate liberalization; removal of credit ceilings; restructuring and privatization of state- owned banks; the introduction of a variety of measures to promote the development of financial markets, including money and stock markets; and the introduction of private banking systems, along with bank supervisory and regulatory schemes. An outgrowth of these financial sector reforms has been a surge of interest in the establishment of stock exchanges and their rapid proliferation in recent years. The number of stock markets in Africa currently exceeds 20 (table 1).1 Between 1992 and 2002, most countries experienced growth in the number of firms listed on the exchanges (an exception is South Africa, which had 683 firms listed in 1992 but only 472 firms in 2002, a decline of 31 percent). An important byproduct of these devel- opments was the emergence in 1998 of a regional market, domiciled in Abidjan, Côte d'Ivoire, the Bourse Régionale des Valeurs Mobiliéres. This market serves as an anchor for the CFA countries (Benin, Burkina Faso, Côte d'Ivoire, Guinea-Bissau, Mali, Niger, Senegal, and Togo). TABLE 1. Number of Firms Listed on African Stock Exchanges, by Country, 1992 and 2002 Number of listed firms Growth 1992­2002 Annualized growth Country 1992 2002 (percent) rate (percent) Algeria 2 3 50.0 10.67 Botswana 11 19 72.7 5.09 Côte d'Ivoire 27 38 40.7 3.16 Egypt, Arab Rep. of 656 1151 75.5 5.24 Ghana 15 24 60.0 4.37 Kenya 57 50 -12.3 -1.18 Malawi 1 8 700.0 34.59 Mauritius 22 40 81.8 5.59 Morocco 62 56 -9.7 -0.92 Namibia 3 13 333.3 14.26 Nigeria 153 195 27.5 2.23 South Africa 683 472 -30.9 -3.30 Swaziland 3 5 66.7 4.75 Tanzania 2 5 150.0 20.11 Tunisia 17 46 170.6 9.47 Uganda 2 3 50.0 14.47 Zambia 2 11 450.0 23.75 Zimbabwe 62 77 24.12 1.99 Total 1,780 2,216 24.49 2.22 Sources: UNDP 2003 and authors' calculations. FINANCIAL SECTOR REFORMS IN AFRICA | 83 Regionalization is an important and welcome development in view of the thinness and illiquidity of individual stock markets in Africa. The Abidjan-based market is bound to serve as a positive role model for other regions of Africa. It is encouraging that already the Anglophone countries of West Africa are contemplating forming a regional stock exchange under the umbrella of the Economic Community of West African States (ECOWAS). Kenya, Tanzania, and Uganda could be natural partners in forming a regional stock exchange in East Africa. The Southern African Development Commu- nity stock exchanges have proposed forming a regional stock exchange. Increasing development of capital markets and accelerated financial sector reforms are vital for integrating Africa into the global financial economy and attract- ing international capital. A recent encouraging development has been the growing, albeit still low-level, attention Africa has received from international investors. Eigh- teen Africa-oriented investment funds are now trading in New York and Europe. The attention Africa has recently attracted has been driven primarily by economic fundamentals and economic systems that increasingly empower private initiative. It is tempting to view Africa's embrace of globalization and openness as merely the fashion of the day, which will dissipate. This view is misguided. Africa has little choice but to integrate into the global economy. Financial sector reforms and devel- opments are a crucial channel for integrating globally and keeping Africa at the cut- ting edge of best international practices. Serious financial sector reforms, the development of capital markets, and regional integration are welcome developments. Performance outcomes have not been encourag- ing, however. The deeper and politically sensitive issues of institutional development-- contractual and legal systems, accounting and disclosure rules, regulatory and supervi- sory mechanisms--are still inadequate. Thus, despite the extensive financial sector reforms that have taken place in terms of interest rate and price liberalization, financial systems in Sub-Saharan Africa face severe inefficiency, illiquidity, and thinness. Moreover, despite a surge of global investor interest in the 1990s, Africa has been bypassed by the massive international capital flowing to developing economies. Aggre- gate private capital flows to developing countries have been rapidly exceeding official development assistance flows since the 1980s. By contrast, Africa remains the only developing region in which development assistance flows exceed private capital flows. Malfunctioning and poorly developed financial systems impede the mobilization of both global capital and domestic savings. Africa continues to experience high levels of capital flight. Sub-Saharan Africa has the highest proportion of private wealth held abroad of any region (an estimated 30­40 percent versus 8 percent for Latin America and 3 percent for East Asia), with the estimated stock of capital flight exceeding the stock of Africa's external debt (Boyce and Ndikumana 2001; Ndikumana and Senbet 2005). To compensate for the grossly inadequate formal financial systems, informal savings channels are prevalent. All of these factors--the dearth of private international capital, the low level of domestic resource mobiliza- tion, capital flight, and untapped resources in the informal sectors--lead to a con- siderable financing gap, with adverse consequences for growth and poverty allevi- ation in Africa. 84 | LEMMA W. SENBET AND ISAAC OTCHERE Foundations of Financial Sector Reforms: A Functional Framework The dominant development thinking, which views the financial sector as a mere con- duit for mobilizing capital, has inspired financial liberalization and reform programs in developing countries that unduly emphasize the development of the banking sec- tor. Focusing on the savings mobilization role of the financial system is shortsighted. In an economic environment characterized by uncertainty, capital markets do more than mobilize capital and allow for risk allocation and risk sharing among market participants. Risk sharing, for example, allows high-risk, high-return projects to be undertaken; without such a mechanism, such projects would be rationed out of the market. This section identifies the vital link between financial sector development and eco- nomic development based on the evidence. It stresses the need to gain a deeper appre- ciation of the multiple functions of a financial system, because functions serve as a channel for the system's impact on economic growth. These functions are described in the context of the agency-based paradigm in finance. The ultimate consequence of a well-functioning financial system is economic development. The empirical evidence suggests that well-functioning financial mar- kets, along with well-designed institutions and regulatory systems, foster economic development. The functions the financial system performs link financial system development and economic development. For instance, studies have found that func- tional features of stock markets, such as liquidity, turnover, and efficiency of pricing of risk, are positively correlated with current and future economic growth and pro- ductivity improvements (Levine 1997). A liquid financial market, characterized by active trading among a large number of investors and firms, provides an exit strategy for both investors and issuing firms. Liquidity is a crucial feature of financial sector development. It also provides a channel for more efficient corporate governance and resource allocation, whereby resources are allocated to the most productive and innovative firms. The link between liquidity and economic growth is supported by the empirical evidence: countries with liquid markets experience faster rates of capital accumulation and subsequently greater productivity gains (Levine 1997; Levine and Zervos 1998) (table 2).2 The positive link between financial sector development and economic develop- ment provides a strong case for the development of a well-functioning financial sec- tor. The implication for African economies is particularly encouraging, because it suggests a link between financial sector development on the one hand and poverty alleviation and employment creation on the other. The central question is how to develop a well-functioning financial sector and build its capacity to exploit its poten- tial contribution to economic development. The Multiple Functions of a Financial System The functional perspective departs from the dominant development thinking on the role of a financial system, which focuses on savings mobilization (McKinnon 1973; Shaw 1973). It recognizes a more comprehensive and holistic approach. FINANCIAL SECTOR REFORMS IN AFRICA | 85 TABLE 2. Stock Market Liquidity Measures in Selected Countries, 1980­95 Country Turnover ratio Log (per capita GDP) Low-income countries Bangladesh 0.03 5.23 India 0.43 5.78 Indonesia 0.19 6.32 Pakistan 0.14 5.79 Zimbabwe 0.07 7.88 Average 0.17 6.20 Middle-income countries Chile 0.07 6.09 Colombia 0.09 7.71 Egypt, Arab Rep. of 0.06 10.09 Jordan 0.16 7.01 Malaysia 0.24 7.73 Mexico 0.54 7.98 Peru 0.16 7.52 Philippines 0.22 6.57 Sri Lanka 0.07 9.34 Turkey 0.50 7.88 Venezuela, R. B. de 0.13 9.95 Average 0.20 7.99 High-income countries Australia 0.29 9.70 Austria 0.44 9.86 Belgium 0.12 9.79 Canada 0.31 9.90 Denmark 0.21 7.10 Finland 0.20 10.08 Germany 1.04 9.96 Greece 0.12 8.97 Israel 0.65 9.29 Italy 0.30 9.76 Japan 0.43 9.97 Korea, Rep. of 0.85 8.53 Kuwait 0.24 9.63 Netherlands 0.37 9.79 New Zealand 0.19 9.44 Norway 0.33 10.18 Portugal 0.15 8.69 Singapore 0.33 9.42 Spain 0.27 6.50 Sweden 0.30 10.12 United Kingdom 0.38 6.98 United States 0.54 9.65 Average 0.37 9.24 Sources: International Monetary Fund Emerging Market Database and Tadesse (2004). 86 | LEMMA W. SENBET AND ISAAC OTCHERE First, an environment characterized by risk and uncertainty calls for a financial system that provides efficient risk sharing and diversification. Under uncertainty, risk allocation and sharing are vital functions of financial markets. Consequently, both the quantity and quality of capital need to be considered in the functioning of finan- cial markets. The traditional view is guided primarily by the quantity of capital mobi- lized in an economy (Haque, Hauswald, and Senbet 1997). The second major dimension to consider in studying a financial system is its role in an environment characterized by imperfect information and agency problems, such as those prevailing in Africa. In such an environment, there are incentive problems and potential conflicts between capital contributors and the society at large (including nonfinancial claimholders, such as employees and customers) (Barnea, Haugen, and Senbet 1985). A well-functioning and liquid financial system serves as a vehicle for efficient contracting among conflicting parties and for disciplining corporate insiders. It allows for active contests for corporate control, for example, so that resources are controlled by those who create the most value for the stakeholders--and ultimately for the society at large. Policy makers need to be aware of the multiple functions of capital markets in designing mechanisms for efficient functioning of these markets. The depth of the capital market reform and development must be judged on the basis of the efficiency with which the various functions of capital markets are carried out. For instance, the mere establishment of stock exchanges is inconsequential if the environment is hos- tile to risk-sharing opportunities and liquidity provision. Stock markets that are devoid of liquidity or produce no information do not function efficiently. By the same token, the mere existence of banks is of little value if they do nothing more than mobilize deposits and use them primarily to purchase government securities. The dearth of commercial and private lending prevents banks from serving as informed agents or intermediaries and building vital information capital for the efficient allo- cation of resources. This pattern of financial disintermediation or dysfunctional intermediation is widely observed in Sub-Saharan Africa. Thus, it is imperative to use the multiple functions of capital markets as a guiding principle in building the capac- ity of African financial systems and developing measures for integrating the region into the global financial economy. Prospects for and Benefits of Financial Globalization Attracting global capital is an important function of a well-functioning domestic financial system, for a variety of reasons described below. What are the essential ingredients for financial globalization in Africa? Since the opening up of the world economy in the 1980s, massive flows of private international capital have flowed into developing countries. Accompanying this has been a radical shift in the pattern of capital flows, with portfolio flows (stocks and bonds) growing fastest. Sub-Saharan Africa was left out of these portfolio flows. Dur- ing the same period, international development finance flows declined. Sub-Saharan Africa continues to account for the largest proportion of development finance. In fact, since the mid-1980s, aggregate private capital flows to developing nations have rap- idly outpaced official development assistance flows. This is a welcome development FINANCIAL SECTOR REFORMS IN AFRICA | 87 and a positive trend in view of shrinking development flows. Sub-Saharan Africa (excluding South Africa) is the only region in which development assistance exceeded private capital flows (figure 1). Foreign direct investment grew slightly during the 1980s and 1990s. It is highly concentrated, however, with the bulk channeled to the four resource-rich countries-- South Africa, Nigeria, Angola, and Ghana. FIGURE 1. Official Development Assistance and Private Capital Flows, by Region, 2000 and 2001 2000 100 80 60 cent Per 40 Official Private 20 0 Sub-Saharan East South Latin Asia Asia America Europe Middle and and East Africa Pacific and and (South the Central North Caribb Asia Africa Africa ean excluded) 2001 100 80 60 Official cent Private 40 Per 20 0 Sub-SaharanEast South Lati Eur Middle Asia nAmerica ope East and Asia and Africa Pacific and Central and North (Sout the h CaribbeanAsia Africa excluded) Sources: World Bank 2003, 2004. 88 | LEMMA W. SENBET AND ISAAC OTCHERE Portfolio equity flows to Sub-Saharan Africa were nonexistent before 1992. They have emerged slowly over the years, although they remain small relative to other emerging markets. Thus, with guarded optimism for further development, it is now appropriate to look into prospects for globalization of African financial markets. Financial globalization is a two-way street; successful globalization is character- ized by gains for both global investors and emerging economies. The competitiveness of Africa in attracting international capital depends on its ability to improve the global risk-reward ratio. Why is Africa of potential interest to global investors? What does Africa gain from financial globalization? Potential Benefits for Global Investors in Africa The benefits from global risk diversification arise from diversity in the economic cycles of countries. U.S. investors, for example, benefit from diversifying their invest- ment portfolios globally, because the U.S. economy does not move in tandem with the economies of the rest of the world. Optimal investment strategies are therefore global. Growing evidence suggests that global strategies should include emerging markets, and even pre-emerging mar- kets, such as those in Africa (see, for example, Diwan, Errunza, and Senbet 1994). To the extent that Africa's economies do not move in tandem with those of the devel- oped world, there are opportunities for international investors to benefit from includ- ing African financial markets in their global portfolios.3 Actual portfolio holdings are far from the optimal global strategy, suggesting con- siderable potential for further globalization of emerging markets as the world becomes more integrated. Emerging markets are underrepresented in the global port- folios, and Sub-Saharan Africa is underrepresented in emerging market portfolios. With increasing economic and financial reforms, Africa is bound to participate in the growing allocation of global investments to emerging markets. The potential benefits of the Africa portfolio go beyond its contribution to global risk diversification. The other dimension of global strategy is investment reward for a given level of risk. In the 1990s, very low price-earnings multiples were observed on African stock markets, suggesting that these stocks were undervalued. In 1997, for instance, shares of some well-established companies, such as Zambia Sugar, Stan- dard Chartered Bank (Ghana), and Delta Corporation of Zimbabwe, were trading at 2.5 times prospective earnings (Gopinath 1998). (To put things in perspective, the U.S. broad market index traded at about 28 times earnings in 1997.) By and large, the low price-earnings ratios reflected an extraordinary level of perceived risk. Price-earn- ings multiples have picked up in recent years. Another indicator pointing to untapped value potentials for global investors is the recent performance of various stock markets in Sub-Saharan Africa, which have out- performed the emerging indices of the International Finance Corporation (IFC) by a substantial margin. (See the discussion of the outlook for financial globalization of Africa below.) Potential Benefits to Africa of Integrating into the Global Financial Economy Attracting global capital and integrating its economies into the global economy would benefit Africa in several ways--by diversifying sources of external finance, distribut- FINANCIAL SECTOR REFORMS IN AFRICA | 89 ing local equity risks globally, reducing the cost of capital for local companies, revers- ing capital flight, and promoting and validating capital market institutions. Heavy reliance on sovereign debt is associated with crises, and official aid flows are already shrinking. Financial globalization allows access to more diversified sources of external finance. In contrast to the syndicated bank lending of the 1970s, international investors can share in the riskiness of local capital markets, especially through equity investments. Financial globalization allows for the market risks of local securities to be shared internationally. Greater sharing of risks in the local cap- ital markets through global holdings of local shares reduces the cost of capital for local firms, enhancing the liquidity of the local market and mobilizing capital by firms. The effect is improved economic performance, as projects that had been for- gone due to excessive risk exposure are adopted (see Errunza, Senbet, and Hogan 1998). Thus the potential benefits of financial globalization to Africa go beyond the mobilization of international capital flows to include improvement in the liquidity and functioning of the local market and the growth of local firms. Evidence from Latin America and East Asia suggests that financial globalization leads to large reversals of capital. Measures that retain domestic capital also help reverse capital flight. Given the scale of capital flight from Africa, there is a potential for significant reversal if the region is integrated into the global financial economy. Because foreign investors demand world-class services, financial globalization enhances the credibility of domestic capital market institutions (custodial, clearing, settlement, and brokerage services; information and accounting disclosures; regula- tions). Globalization puts governments under greater pressure to strengthen the rule of law, enforce contracts, and increase the availability of information in response to international investor demands. Globalization thus exposes African stock markets to best practices and standards and increases pressure to reform local stock markets. Focusing on the banking sec- tor precludes opportunities for building up informational technology that is unique to risk capital (for example, disclosure and accounting standards). This informa- tional discipline has a positive externality for the entire financial sector, including the banking sector. It is therefore important that African countries not put counterpro- ductive restrictions in place by stacking the odds against outside investors. Outlook for Financial Globalization of Africa The prospects for Africa's integration into the global financial economy are encour- aging. There is growing integration of world capital markets, including those in emerging economies, with increasing capital mobility. Barriers to international capi- tal flows have been reduced (by eliminating fixed commissions and deregulating financial markets, for example). Moreover, rapid advances in information technol- ogy facilitate capital flows, research, and international investing. On the real sector side, multinational corporations have expanded global operations and taken advan- tage of global capital markets. Thus, investors seeking the benefits of global diversi- fication are now better able to access markets. African markets are joining this wave of global integration, as Lamba and Otchere (2001) show. The evidence suggests that African markets are increasingly integrated with capital markets in other regions. Financial sector reforms have contributed to 90 | LEMMA W. SENBET AND ISAAC OTCHERE this development. Globalization of capital flows has increased the relevance of emerging capital markets. Increasing integration of African markets with other cap- ital markets will encourage the flow of investments into these countries as investors seek to capitalize on the potential diversification benefits. As these markets become increasingly integrated with world capital markets, they may be able to provide sig- nificant diversification benefits and allow Africa to achieve the benefits of globaliza- tion listed above. The recent performance of African stock markets also bodes well for financial globalization. From 1998 to 2002, the cumulative returns in U.S. dollar terms for Sub-Saharan Africa were 4.3 percent--significantly higher than the ­3.9 percent of the S&P 500 and the ­15.5 percent of the U.K. FTSE 100 index (UNDP 2003). African stock markets were among the best performers in 2003. According to Databank Financial Services, the average African stock market return in 2003 was 44 percent. This compares favorably with the 30 percent return on the Mor- gan Stanley Capital International (MSCI) Global index, the 32 percent on the MSCI-Europe index, the 26 percent for the S&P index, and the 36 percent on the Nikkei index. Stock markets in the Arab Republic of Egypt, Ghana, Kenya, Mau- ritius, Nigeria, and Uganda performed extraordinarily well, earning returns of more than 50 percent. The Ghana Stock Market Exchange was the best performer, earning more than 144 percent in 2002, and outperforming 61 markets surveyed by Databank. In 2002­3 the Ghana stock market led the world, with a compound return of 256 percent. Databank attributes this performance to improving macro- economic conditions, better corporate sector performance, and perhaps the low baseline equity valuations (table 3). These data suggest that African equity markets represent largely unexploited opportunities for international investors. Moreover, investing in these markets could help international investors diversify their portfolios. Despite this potential, Africa has received a scant portion of capital flows to emerging markets, as global equity funds have maintained low exposure in Africa. Africa's portion of global emerging equity portfolios is about 8 percent, but the bulk of this exposure is in South Africa (Mensah 2003). Why is Sub-Saharan Africa so marginalized in the global financial economy? The following sections deal with the challenges and constraints facing African financial systems and offer some policy pre- scriptions to help ease those constraints. Financial Sector Reforms in Africa: Constraints and Opportunities for Further Reforms Financial sector reforms have been and continue to be among the key pillars of struc- tural adjustment programs in Africa. In the past two decades, African countries have embarked on financial sector restructuring involving deregulation and the gradual opening up of the financial sector to foreign participation. They have liberalized inter- est rates and exchange rates, removed credit ceilings, restructured and privatized banks, and introduced measures to promote capital market development. These meas- FINANCIAL SECTOR REFORMS IN AFRICA | 91 TABLE 3. Risk-Adjusted Performance of African Stock Markets, 1999­2002 (percent) Country Mean return Risk-adjusted return (Sharpe measurea) Botswana 28 0.53 Côte d'Ivoire -6 -1.21 Egypt, Arab Rep. of 20 0.36 Ghana 41 0.14 Kenya 9 -0.21 Malawi 15 -0.77 Mauritius 11 0.05 Morocco 12 0.28 Namibia 1 -0.26 Nigeria 32 0.39 South Africa 13 0.04 Swaziland 8 -0.08 Tanzania 52b -- Tunisia 4 -0.36 Uganda 3b -- Zambia 13 -1.20 Zimbabwe 75 0.57 Source: Databank Financial Services. -- Not available. a. Sharpe measure is based on mean stock return and risk-free (Treasury-bill) rates of return from 1992 to 2002. b. Returns are available for one year only. ures represent a major reversal of policies of the post-independence era, during which African governments intervened in the financial sector. The menu of intervention has been extensive and included nationalization of private banks, establishment of entirely new state banks and nonbank financial institutions, the imposition of quantitative restrictions on the allocation of credit, and restrictions on external capital flows. Although the intentions of government intervention in the financial sector may have been to mobilize capital for investments and allocate capital to priority sectors, its actions were counterproductive and produced utterly dysfunctional financial sys- tems. Moreover, the intended goals of capital mobilization and allocation of capital to growth areas were not realized. In fact, the repressive era produced a lost decade for Sub-Saharan Africa (the 1980s), during which the region marched backward while the rest of the world, particularly the emerging countries in East Asia and Latin America, moved forward. Recent financial sector reforms were intended to reverse the adverse consequences of the repressive financial policies of the post-independence era, although the pace and extent of policy reform varied across countries. The performance outcomes of the financial sector reforms have not been encouraging. Moreover, results have been similar across countries, even where the pace of reform has varied. For instance, although Ghana's removal of lending restrictions was much more rapid than that of Tanzania, the evolution in interest rates and spreads has been similar (Nissanke and Aryeetey 1998). 92 | LEMMA W. SENBET AND ISAAC OTCHERE Elements of Financial Sector Reform in Africa Although the type and extent of financial sector reform in Africa has varied across countries (table 4), a more liberalized financial environment has emerged as a result of these reforms. A major feature of interest rate and credit market liberalization was elimination of monetary policies that involved direct determination of lending and deposit rates. These rates are now by and large left to market conditions. Moreover, indirect monetary control policy emerged as a viable option over its direct counter- part. Thus, interest rate liberalization, along with removal of credit ceilings and quantitative controls, constituted an important element of financial sector reforms. Restructuring and recapitalization of banks was another major element of finan- cial sector reforms. Bank restructuring included privatization of state-owned banks. Regulatory and supervisory schemes were introduced along with bank restructurings and privatizations. Challenges to Reforms of Credit and Banking Systems This paper takes a functional perspective in studying the characteristics of African financial systems. In particular, it looks at the effectiveness of financial intermedia- TABLE 4. Dates of Financial Sector Reforms in Selected Countries in Sub-Saharan Africa Structural Credit Stock adjustment market Banks Banks exchange Country program adopted liberalized restructured privatized established Benin 1989 Botswana 1991 1991 1989 Cameroon 1990 1998 Côte d'Ivoire 1989 1999 1976 Gambia, The 1986 Ghana 1983 1988 1989 1997 1996 Kenya 1989 1991 1989 1954a Madagascar 1994 1999 Malawi 1987 1988 1990 1996 Mauritania 1990 Mauritius 1983 1993 1988 Namibia 1992 1991 1992 Nigeria 1987 1987 1990 1992 Tanzania 1985 1991 1991 1994 1998 Uganda 1987 1988 1996 1998 Zambia 1991 1992 1994 Zimbabwe 1991 1991 1997 1946b Sources: Inanga and Ekpenyong 2004; Chirwa and Mlachila 2004; World Bank privatization database. a. Kenya and some other Sub-Saharan African countries had stock markets before they embarked on structural adjust- ment programs, although they were relatively inactive. b. A new exchange was established in 1994. FINANCIAL SECTOR REFORMS IN AFRICA | 93 tion, banking regulation and supervision, the diversity of financial instruments, and the depth and capacity of stock markets. While there is some diversity in the outcomes of the restructuring efforts across countries, the differences are not very significant. In most countries, the functions of savings mobilization and financial intermediation have not been fully achieved since reforms were initiated (Nissanke and Aryeetey 1998; Haque, Hauswald, and Senbet 1997). Even in countries such as Ghana and Malawi where reforms were relatively orderly, most banking institutions have not developed the capacity for efficient risk management and control. In Ghana and Uganda, among the most progressive countries in Sub-Saharan Africa, the allocation of bank lending portfolios targeted either the lowest risk category (for example, gov- ernment securities) or excessively high-risk clients (Aryeetey and Senbet 1999). In Ghana, 60 percent of bank assets have government exposure, including state-owned enterprises. In Uganda, a third of bank assets are held in government securities. Lack of Financial Deepening and Credit to the Private Sector The performance outcomes of financial sector reforms have been discouraging and at times perverse. The desired effects on savings mobilization and credit allocation have not materialized. The financial deepening measures, the M2/GDP ratio, and measures of private credit (the private credit/GDP ratios) do not show clear upward trends in most countries (tables 5 and 6). On both indicators, African countries (with the exception of South Africa) generally lag behind their Asian counterparts. Banking Instability and Dysfunctional Intermediation The problems of the banking systems in Africa remain severe, even after financial sec- tor reforms. Banking system assets are heavily concentrated at the short end of the market and are excessively liquid (Nissanke and Aryeetey 1998). Moreover, the port- folios of banking institutions continue to be dominated by an extremely high inci- dence of nonperforming loans and excess liquidity. Information and enforceability problems, along with financial distress and bank failure, have serious consequences. In particular, they result in adverse selection, where weak banks attract dispropor- tionately high-risk and weak enterprises (for example, state-owned companies), which see no downside in borrowing at high interest rates and bet on a small prob- ability of good outcomes. Such a weak banking system breeds instability. Inappropriate Banking Incentives and Poorly Designed Safety Nets A major objective of monetary and financial policy makers is to stabilize a nation's payments system. Explicit or implicit deposit insurance is used to reduce the risk of systemic failure of banks and stabilize the payments and financial system. Deposit insurance is socially counterproductive, however, if the system is not structured appropriately. When deposits are guaranteed, depositors face no risk; risk is transferred to an insuring agency. Bank owners have an incentive to gain by choosing excessively risky asset portfolios and engaging in high-risk lending. Thus, in an ill-designed deposit insurance system, public mismanagement of the system and private incentive incom- patibility problems can actually increase the systemic risk and instability of a 94 TABLE 5. M2/GDP Ratio in Selected Countries in Sub-Saharan Africa, 1992­2003 Annualized Growth growth Country 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 (percent) (percent) Algeria 0.49 0.55 0.49 0.40 0.36 0.39 0.46 0.45 0.41 0.49 -- 0.65 32.46 2.59 Botswana 0.30 0.23 0.22 0.22 0.22 0.23 0.29 0.34 0.29 0.34 0.30 0.30 1.97 0.16 Côte d'Ivoire 0.28 0.28 0.28 0.29 0.27 0.24 0.23 0.22 0.22 0.23 0.29 -- 3.24 0.29 Egypt, Arab 0.85 0.85 0.85 0.80 0.79 0.78 0.77 0.76 0.77 0.82 0.88 0.97 15.05 1.18 Rep. of Ghana 0.21 0.20 0.23 0.22 0.21 0.24 -- -- -- -- -- -- 16.16 2.53 Kenya 0.37 0.37 0.41 0.44 0.48 0.48 0.45 0.44 0.43 0.40 0.41 0.40 10.23 0.82 Malawi 0.21 0.22 0.26 0.19 0.16 0.14 0.17 0.16 0.19 -- -- -- -11.11 -1.30 Mauritius 0.69 0.70 0.71 0.77 0.74 0.77 0.76 0.81 0.79 0.80 0.83 0.84 21.07 1.61 Morocco 0.60 0.63 0.62 0.66 0.62 0.73 0.71 0.78 0.83 0.87 0.89 0.92 53.00 3.61 Namibia 0.28 0.32 0.33 0.37 0.40 0.38 0.38 0.41 0.41 0.37 -- -- 31.77 2.80 Nigeria 0.23 0.28 0.29 0.16 0.13 0.15 0.18 0.21 0.21 0.23 2.68 0.32 35.23 2.55 South Africa 0.50 0.47 0.49 0.50 0.51 0.54 0.57 0.58 0.56 0.59 0.62 0.64 27.41 2.04 Swaziland 0.34 0.32 0.29 0.25 0.25 0.26 0.26 0.26 0.22 0.21 0.21 -- -36.83 -4.09 Tanzania 0.22 0.24 0.19 0.20 0.17 0.17 0.16 0.17 0.17 0.18 0.22 -- -0.92 -0.08 Tunisia 0.47 0.46 0.46 0.46 0.46 0.49 0.48 0.52 0.55 0.57 0.57 0.56 20.65 1.58 Uganda 0.07 0.10 0.11 0.11 0.12 0.13 0.14 0.14 0.16 0.16 0.19 0.19 165.87 8.49 Zambia -- 0.14 0.15 0.17 0.18 0.17 -- -- -- -- -- -- 20.10 3.73 Zimbabwe 0.17 0.23 0.23 0.27 0.25 0.29 0.24 0.21 0.25 -- -- -- 54.06 4.92 Sources: World Bank World Development Indicators (various years) and authors' calculations. -- Not available. FINANCIAL SECTOR REFORMS IN AFRICA | 95 TABLE 6. Credit to the Private Sector in Selected Sub-Saharan African Countries, 1995­2002 (percent of GDP, except where otherwise indicated) Annualized Growth growth Country 1995 1996 1997 1998 1999 2000 2001 2002 (percent) (percent) Algeria 5.2 5.5 4.0 4.6 5.5 6.1 8.0 6.8 30.8 3.4 Botswana 13.2 11.1 9.9 11.9 15.2 16.1 16.2 18.4 39.4 4.2 Côte d'Ivoire 20.2 19.7 19.4 18.6 16.0 17.2 15.9 14.8 -26.7 -3.8 Egypt, Arab 47.1 41.5 46.6 54.0 59.7 59.3 61.6 60.6 28.7 3.2 Rep. of Ghana 5.2 6.6 8.2 9.4 12.0 14.1 14.1 12.0 130.8 11.0 Kenya 33.8 34.6 34.9 30.3 30.5 30.1 24.6 23.4 -30.8 -4.5 Malawi 7.8 4.2 3.9 6.1 5.6 6.2 6.8 4.1 -47.4 -7.7 Mauritius 48.3 44.7 50.5 59.2 58.5 26.7 62.7 61.3 26.9 3.0 Morocco 48.9 45.8 33.5 50.4 54.6 58.6 54.0 54.4 11.3 1.3 Namibia 56.9 50.5 52.6 51.2 49.2 44.7 47.3 48.4 -14.9 -2.0 Nigeria 7.8 10.6 8.2 9.1 13.8 13.9 17.8 17.8 128.2 10.9 South Africa 130.6 137.1 135.7 118.9 136.3 141.9 148.5 131.7 0.8 0.1 Swaziland -- -- -- -- -- 14.2 13.1 14.3 0.7 0.2 Tanzania 8.9 3.4 3.9 4.7 4.6 4.6 4.9 6.3 -29.2 -4.2 Tunisia 68.4 63.5 64.5 50.8 65.1 66.2 67.9 68.6 0.3 0.1 Uganda 4.1 4.7 4.3 5.2 5.9 6.3 5.9 6.7 63.4 6.3 Zambia 7.2 9.3 8.0 6.8 7.4 9.5 7.2 6.2 -13.9 -1.9 Zimbabwe 35.3 35.4 37.6 38.8 27.2 25.2 25.8 37.0 4.8 0.6 Sub-Saharan -- 68.1 65.1 57.9 66.2 66.0 65.2 53.5 -21.4 -3.4 Africa Developing -- 18.7 47.2 50.2 52.7 55.3 52.1 55.9 198.9 16.9 countries World -- 107.7 109.7 103.1 109.0 119.5 120.7 118.1 9.7 1.3 Sources: World Bank World Development Indicators 1997­2004 and authors' calculations. -- Not available. banking system. Cull, Senbet, and Sorge (2005) study the impact of deposit insurance schemes on financial development and stability of a variety of countries around the world. Desired outcomes are achieved in countries with high-quality regulation; deposit insurance schemes turn out to be counterproductive in countries with weak legal and regulatory regimes. Weak financial systems become even more unstable. The lesson for Africa is to devise efficient and incentive-compatible regulatory sys- tems in the face of safety nets such as explicit or implicit deposit insurance schemes. The basic idea of moral hazard and excessive bank risk taking can be formalized along the lines of John, Saunders, and Senbet (2000). Left alone and inefficiently reg- ulated, bank owners would adopt excessively high-risk portfolios (by engaging in speculative real estate lending, for example), in the hopes of reaping big payoffs under favorable economic conditions and transferring losses to the insuring agency (and society at large) when projects fail. Figure 2 characterizes the investment incen- tives of a banking system. A bank is endowed with an opportunity to hold a 96 | LEMMA W. SENBET AND ISAAC OTCHERE portfolio of risky assets (loans). These investment opportunities can be depicted by their rewards, shown along the vertical axis in figure 2 (a schedule of bank values), and their risks (volatilities) of terminal cash flows, shown along the horizontal axis. Following the agency tradition, assume that bank asset or lending risk choices by cor- porate insiders (bank managers) are imperfectly observed by outsiders (depositors and regulators). The agency problem arises because bank owners make investment and risk choices to maximize the value of the bank equity currently outstanding rather than maximizing the total value of the bank assets. In the absence of bank leverage (or bank deposits), bank equity and bank value maximization decisions are identical. The first-best solution is represented by the value-maximizing level of risk for the investment opportunity. There is a unique value-maximizing level of risk for unimodal structures, at the apex of the concave production curve in figure 2. With undercapitalization or sufficiently high levels of debt (deposit financing), bank insiders (deciding on behalf of bank owners) depart from the first-best risk out- come. Investment will be affected by the amount of capital in place and its comple- ment, the level of debt financing. Bank owners will invest at a point of asset risk choice above the first-best level of risk (along the right-hand side of the value maxi- mizing point), so as to maximize the value of their equity holdings. The distortion in risk choice depends on the level of bank capital as well as the investment risk char- acteristics. However, the risk distortion effect decreases with the level of bank capi- tal. This provides a rationale for capital adequacy requirements, which minimize the instability brought about by moral hazard and excessive bank risk taking. When cap- ital adequacy measures are weak, banks are more likely to become distressed. FIGURE 2. Moral Hazard and Excessive Risk-Taking Bank value V1* Value under bank V1 asset risk shifting Bank 2 Bank 1 1* 1 Risk Source: Cull, Senbet, and Sorge 2005. FINANCIAL SECTOR REFORMS IN AFRICA | 97 Supervisory and Regulatory Failure Serious efforts have been made to improve banking regulation and supervision in several African countries to achieve financial stability and high-quality intermedia- tion. Supervision and regulation have remained grossly inadequate, however, and the quality of financial intermediation is either reflective of excessively high risk (leading to bank distress) or is excessively conservative (leading to a shortage of credit to the private sector). Even when the regulatory structure looks efficient on paper, it may not be properly implemented. Effective implementation requires a well-developed and coherent legal environment that forces regulators to faithfully and obediently implement and enforce regulation. In the absence of such support- ing institutional arrangements, the consequences of conflicts between society and regulators can lead to suboptimal regulatory outcomes (Hauswald and Senbet 1999). Nigeria is often cited as an example of this type of regulatory failure. Regu- latory failure may also occur at the level of managing bank insolvency. For instance, it is in the best interests of bank regulators to dissolve an insolvent bank whose assets are insufficient to meet its debt obligations (Kane and Rice 2001). Regulators facing short decision horizons find it in their interest to "recycle" bad loans of trou- bled banks. In this environment, an insolvent bank can stay in business, and even raise more deposits, as long as it does not face severe liquidity problems that would eventually lead to disclosure of its insolvency. Bank Concentration The size and scope of financial service activities in most African financial systems have often been limited by policy. State-owned banks dominate, and the banking sec- tor is oligopolistic. Bank oligopoly is particularly worrisome. In Uganda, four for- eign-owned banks account for 75 percent of deposits and assets. Four banks account for 65 percent of market share in Ghana. In Tanzania, four foreign banks account for 49 percent and three local banks for 49 percent of market share. This highly con- centrated banking structure (four-bank syndrome) allows for only very limited deposit and lending competition, reinforcing high lending and deposit spreads (Haque, Hauswald, and Senbet 1997). Governments that issue large quantities of high-yielding securities or bills to meet their fiscal requirements exacerbate the prob- lem. As shown in figure 3, financial sector reforms have led to a decline in the pub- lic share of domestic credit, but government and public enterprises continue to enjoy the largest proportion of bank credit in many countries (see table 6). High Interest Rate Spreads Financial liberalization could increase the gross interest margin, as competition in the credit market resulting from the reduction in barriers to entry of foreign and domestic competitors increases slowly relative to that of the deposit market. The readjustment of the bank cost structure may materialize only slowly. Thus, under the reform programs, an initial increase in the spread between lending and deposit rates would be expected. With more efficient banking practices, this spread should narrow. Moreover, financial liberalization should be accompanied by a reduction in the reserve requirement, which in turn should increase loanable funds, putting 98 | LEMMA W. SENBET AND ISAAC OTCHERE FIGURE 3. Financial Deepening and Credit to the Private Sector in Sub-Saharan Africa, 1995­2002 120 100 GDP 80 of cent 60 Per 40 20 0 1995 1996 1997 1998 1999 2000 2001 2002 Credit to private sector (excluding Mauritius and South Africa) Financial deepening (excluding Mauritius and South Africa) Credit to private sector (Mauritius and South Africa) Financial deepening (Mauritius and South Africa) Sources: World Bank World Development Indicators 1997­2004 and authors' calculations. downward pressure on the cost of funds. These developments should reduce the interest rate margin (the gap between the lending rate and the deposit rate). How- ever, in most countries, lending rates rose sharply during the reform years, rising much faster than deposit rates. In many cases, the spread has continued to widen (table 7). The good news is that, as expected, financial sector reforms have yielded the desirable outcome of positive real interest rates. However, the exorbitantly high real interest spreads are disappointing and remain a puzzle. Coupled with the failure to produce the desired results of increased investments and savings, high spreads con- stitute a devastatingly negative outcome of financial sector reforms in Africa. This phenomenon points to dysfunctional financial intermediation as well as grossly inadequate competition in African financial systems. In particular, these systems are characterized by state dominance and the oligopolistic behavior of a few commer- cial banks. Financial sector competition remains very limited in a large number of countries, including Botswana, Malawi, Mauritius, Morocco, South Africa, Swazi- land, Zambia, and Zimbabwe, where gross interest rate margins have increased. Most of the countries in which financial sector competition has been limited are in Southern Africa. Interest rate spreads have narrowed in recent years, suggesting growing competi- tion in the financial system as well as enhancement of contract enforceability in the credit markets. Margins have been declining for most countries, including Algeria, Egypt, Kenya, Mozambique, Namibia, Nigeria, Tanzania, and Uganda. Over the 1992­2003 period, for instance, the gross interest rate margin in Tanzania declined by more than 37 percent, with an annualized decline in the gross interest rate mar- gin of 5 percent. Looking ahead, countries that build institutional capacity with mul- tiple financial institutions and increase the availability of nonbank finance can foster competition in the financial systems and experience declining spreads as well as bet- ter financial intermediation. TABLE 7. Interest Rate Spreads in Selected African Countries, 1992­2003 (percent) Change Annualized 1992­2003 change Country 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 (percent) (percent) Algeria -- -- 4.00 3.00 4.50 2.75 2.50 2.50 2.50 3.25 3.25 2.75 -31.25 -3.68 Botswana 1.50 1.43 3.49 4.31 4.07 4.83 4.81 5.17 5.24 5.60 5.66 6.10 306.67 12.40 Egypt, Arab 8.30 6.30 4.70 5.60 5.10 4.00 3.60 3.80 3.70 3.80 4.50 5.30 -36.14 -3.67 Rep. of Kenya -- -- -- 15.20 16.20 13.53 11.09 12.83 14.24 13.03 12.96 12.44 -18.16 -2.20 Malawi 5.50 7.75 7.00 10.06 19.00 18.04 18.61 20.37 19.88 21.21 -- -- 285.64 14.45 Mauritius 7.06 8.18 7.88 8.58 10.04 9.84 10.64 10.71 11.16 11.32 11.12 11.47 62.46 4.13 Morocco -- -- -- -- -- -- 6.20 7.10 8.10 8.30 8.60 8.80 41.94 6.01 Mozambique -- -- -- -- -- -- 16.13 11.77 9.34 7.63 8.72 12.54 -22.26 -4.11 Namibia 8.85 8.41 7.87 7.67 6.60 7.48 7.78 7.66 7.89 7.74 6.03 5.94 -32.88 -3.27 Nigeria 6.72 8.41 7.39 6.70 6.78 10.63 8.07 7.48 9.58 8.18 8.10 6.49 -3.42 -0.29 South Africa 5.13 4.66 4.47 4.36 4.61 4.62 5.29 5.76 5.30 4.40 4.98 5.20 1.36 0.11 Swaziland 5.43 6.46 6.71 7.61 7.59 7.50 7.58 7.56 7.47 7.10 7.23 7.04 29.65 2.19 Tanzania -- -- -- 18.20 20.38 18.44 15.41 14.14 14.19 15.45 13.14 11.43 -37.20 -5.04 Tunisia -- -- -- -- -- -- -- -- -- -- -- -- -- -- Uganda -- -- -- 12.00 9.00 9.53 9.50 12.82 13.08 14.19 13.54 9.09 -24.25 -3.04 Zambia 6.07 -- 24.42 15.29 11.65 12.21 18.72 20.25 18.56 22.82 21.87 18.62 206.75 9.79 Zimbabwe -- 6.88 8.11 8.81 12.65 13.95 13.00 16.88 18.04 24.07 18.10 61.37 792.01 22.01 Sources: World Bank World Development Indicators (various years) and authors' calculations. -- Not available. 99 100 | LEMMA W. SENBET AND ISAAC OTCHERE Bank Privatization and Performance Widespread privatization of state-owned banks took place in Sub-Saharan Africa in the wake of the extensive financial sector reforms in the region. This is a wel- come move, given the evidence that privatized firms generally outperform state- owned enterprises. The conventional justification for state ownership of banks-- that state-owned banks serve underserved markets, particularly rural areas and small enterprises--is being challenged. It is often alleged that private banks tend to concentrate their lending in urban areas and on large enterprises while simultaneously being prone to costly banking crises. In fact, growing evidence suggests that state ownership is associated with less financial development, less growth, worse bank performance, and less financial sta- bility (Barth, Caprio, and Levine 2004). In recent years there has been sharp decline in the state ownership of banks in developing countries, including Africa, although some countries (such as Ethiopia) continue to resist bank privatization. Africa experienced the sharpest decline in state ownership between 1999 and 2002 (Clarke, Cull, and Shirley 2004). The pre- and postprivatization operating performance of the privatized banks relative to rival banks is compared here, using data from the World Bank privatization database and the appendix to Megginson (2002). The measures of performance are asset quality, man- agement efficiency, earnings ability, and labor (employment levels and productivity). The data reveal a deterioration in the asset quality of the privatized banks (table 8). Loan loss provision also increased significantly in the postprivatization period. Prof- itability of the privatized banks declined after privatization, as did the return on equity. The privatized banks were more profitable than their rivals, although the difference in performance was not statistically significant. There was no perceptible change in the efficiency of the privatized firms with respect to the rivals. Overall, the postprivatiza- tion operating performance of the privatized banks in these countries worsened. The evidence from capital market data is also consistent with the worsening per- formance effect of bank privatization in Africa. The results presented in table 9 show that the privatized banks realized abnormal negative returns. Investors who bought the shares of the privatized banks on the first day of trading and held them for five years lost 27 percent of their wealth. The performance of privatized banks worsened after year four, although the returns are not statistically different from those of rival banks. The poor operating and stock market performance of the privatized banks in Africa suggests that there have not been gains from bank privatizations--a finding that is at odds with the evidence on other regions (Clarke, Cull, and Shirley 2004). Two factors may account for the difference. First, the privatization data are based on share issuance. African countries have malfunctioning stock markets with limited capacity for monitoring (see below), suggesting that stock market based privatization did not work. Megginson, Nash, and van Randenborgh (1994) suggest that capital market monitoring that accompanies privatization elicits postprivatization perform- ance improvements. Moreover, a well-developed and active capital market allows the newly privatized firms greater access to capital needed to finance profitable projects. Second, about 89 percent of the privatized banks in Africa were only partially privatized, with the governments owning about 65 percent of the "privatized" banks. TABLE 8. Performance of Privatized Banks in Africa Relative to Rivals Preprivatization Postprivatization Difference Item Year -1 Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 mean mean (pre-post) Provision/loan ratio Privatized 2.38 3.06 1.97* 3.69* 2.00** 2.54** 1.04* 2.83 3.05* 2.38 Rival 0.27 1.43*** 1.42*** 0.94*** 1.31*** 1.21*** -- 1.44*** -- 0.27 Difference 2.79* 0.54 2.27 1.06 1.23 -0.17 -- 1.61 -- 2.79* Gross impaired assets Privatized -- -- 5.73 31.61 6.33 20.25 15.26 --00 21.64* -- Rival -- -- 7.33* 6.45*** 7.29*** 7.25** 6.08* --00 9.34*** -- Difference -- -- -1.60 25.16 -0.96 13.00 9.18 --00 12.30 -- Return on assets (percent) Privatized 3.53 4.23 3.56 1.24 2.22 1.49 2.22 3.23 2.55 -0.68 Rival -- 1.89* 2.43** 2.05*** 3.23*** 1.88*** 1.70** --00 1.84*** -- Difference -- 2.34 1.13 -0.81 -0.10 -0.39 0.52 --00 0.71 -- Return on equity (percent) Privatized 12.52 16.25 16.52* 13.94 12.58 15.35 21.97 21.47 10.45 -11.02 Rival -- 33.25 29.24 21.77 16.86 16.19 15.89 --00 10.60*** -- Difference -- 17.00 -12.72 -7.83 -4.28 -0.84 6.08 --00 -0.15 -- Cost-to-income Privatized 80.85 70.03* 73.17* 83.04* 81.03** 85.04** 70.55* 80.35*** 80.61*** 0.26 Rival 73.92* 65.66* 66.20*** 68.54*** 71.85*** 72.08*** 73.46*** 75.64*** 83.65*** 8.01 Difference 6.93 4.37 6.97 14.50 9.18 12.96** -2.91 4.71 -3.04 7.75 101 (Continues on next page) 102 TABLE 8. continued Preprivatization Postprivatization Difference Item Year -1 Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 mean mean (pre-post) Net interest margin Privatized 3.10 12.36 7.09 2.67* 3.52* 2.32** 3.10* 3.10 4.10*** 1.00 Rival -- 4.40* 3.20* 7.05*** 3.20*** 4.10*** 3.03*** -- 5.08*** -- Difference -- 7.96* 3.89 -4.38 0.32 -1.78 0.07 -- -0.98 -- Growth in staff levels Privatized -- -- -1.21 0.70 -1.22 1.54 -2.06 -- -4.67 -- Rival -- -- 0.12 2.66* 2.66*** 0.87* 1.10*** -- 1.33** -- Difference -- -- -1.33 -1.96 -3.88 -0.67 0.32 -- -6.00*** -- Sources: World Bank privatization database; Megginson (2002). -- Not available. *Significant at the 10 percent level. **Significant at the 5 percent level. ***Significant at the 1 percent level. FINANCIAL SECTOR REFORMS IN AFRICA | 103 TABLE 9. Market-Adjusted Buy and Hold Returns of Privatized and Rival Banks (percent) Privatized banks (N = 7) Rival banks (N = 27) Difference Period of stock ownership Mean Percent Median Mean Percent Median Mean Median (months) return positive return return positive return return return 1-12 -0.08 2.5 -0.08 -0.01 50 -0.01 -0.07 -0.07 (-1.16) (-1.41) (-0.89) (-0.50) (0.00) (0.55) (-0.96) (-0.64) 1-24 -0.06 50 -0.02 -0.01 50 -0.01 -0.05 -0.01 (-0.82) (0.00) (-0.55) (-0.30) (0.00) (-0.44) (0.60) (-0.16) 1-36 -0.02 57 0.09 0.11 69 0.15 -0.13 ­0.06 (-0.13) (0.38) (0.08) (2.58)* (1.96) (2.29)* (-0.74) (0.55) 1-48 -0.05 56 0.04 0.09 64 0.13 -0.14 ­0.09 (-0.30) (0.33) (0.01) (1.31) (1.57) (1.38) (-0.75) (0.67) 1-60 -0.27 40 -0.26 -0.17 38 -0.17 -0.10 -0.09 (-1.68) (-0.63) (1.43) (-2.57)* (-1.37) (-2.19)* (-0.57) (0.71) Sources: Datastream International; Megginson (2002). Note: Market-adjusted buy and hold returns are the returns accruing to investors who bought the shares on the first day of trading and held the stock for 12, 24, 36, 48, and 60 months. Figures in parentheses are t-statistics for mean returns and z-statistics for the median returns or percentage positive. *Significant at the 5 percent level. This leaves banks vulnerable to government intervention. The evidence suggests that performance improves as the percentage of government ownership declines (D'Souza, Megginson, and Nash 2000). Clarke, Cull, and Shirley (2004) find that very limited gains accrue from privatization even when the government holds a minority stake. This is not surprising given the state's multiple objectives, which are politically driven and detract from efficiency. The moral of the bank privatization experience in Africa is thus twofold. First, government intervention and partial ownership of privatized banks needs to be elim- inated. Second, there is a need to cultivate a deep and well-functioning financial sys- tem that can lead to gains from bank privatization. Functionality of African Stock Markets Beyond mere capitalization, it is the functional efficiency of financial systems that contributes significantly to economic growth. African stock markets should be judged on the basis of their efficiency in carrying out these functions. Stock Market Liquidity and Depth The growth in the number of stock exchanges--from 10 in 1992 to 24 in 2004--has been impressive. But these markets have been ineffective. Except for the South African stock market, pre-emerging stock markets in Africa are by far the smallest of any region, in terms of both the number of listed companies and market capitaliza- tion. Moreover, trading activity is minimal (table 10). Stock turnover is extremely low in most countries, suggesting very low liquidity and exit strategies. International investors are not investing in Africa. Most of the capital raised through initial public offerings or seasoned offerings comes from African institutions and individual 104 | LEMMA W. SENBET AND ISAAC OTCHERE investors (UNDP 2003). The lack of liquidity should be of particular concern to pol- icy makers given the evidence linking liquidity to economic growth (figure 4). The inefficiency of the secondary market is a barrier to raising capital in the pri- mary market when companies seek to make initial public offerings. Companies find it too costly to offer new issues of seasoned shares in malfunctioning stock markets. The dysfunctional nature of the stock markets is also reflected in their operational inefficiency. Brokerage services are poor, and settlement and operational procedures are slow (in some countries it takes months to execute a single transaction). The fail- ures are due partly to problems in the design of the market microstructure and to lack of trained personnel. Sources of Macro Risks: Macroeconomic and Political Instability Country risk is the predominant source of variation in stock returns across countries. Macroeconomic and political instability lead to high volatility on financial markets. Investors are concerned about adverse changes in government policies. TABLE 10. Liquidity of African Stock Markets, 1992­2002 (percent turnover) Country 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Average Algeria -- -- -- -- -- -- -- 0.3 1.7 1.5 -- 1.2 Botswana 5.1 7.7 8.2 9.5 9.5 9.6 9.7 3.6 4.8 5.1 3.6 7.0 Côte d'Ivoire 0.8 1.4 2.8 1.6 2.1 1.9 2.1 5.6 2.8 0.7 1.2 2.1 Egypt, Arab 6.0 4.5 17.8 8.4 16.7 28.1 20.6 27.5 38.7 16.0 28.1 19.3 Rep. of Ghana -- 4.2 4.0 1.3 1.1 4.3 4.3 2.7 2.0 2.5 2.9 2.9 Kenya 1.9 1.3 2.0 3.4 3.6 5.8 3.9 5.3 3.7 3.8 2.1 3.4 Malawi -- -- -- -- -- -- 6.8 3.7 4.2 13.8 2.8 6.3 Mauritius 2.4 4.6 5.4 4.4 4.8 8.1 5.6 4.7 1.8 10.3 4.5 5.1 Morocco 3.7 18.8 18.0 40.8 5.0 8.6 8.9 18.5 10.0 10.7 11.1 14.0 Namibia -- -- 9.0 1.6 8.7 3.5 3.0 3.2 7.1 5.3 64.2 11.7 Nigeria 1.1 1.0 0.7 0.7 2.0 3.6 5.5 4.9 6.2 9.2 8.1 3.9 South Africa 7.5 7.6 6.9 6.1 11.3 19.3 34.3 27.8 34.0 49.9 42.1 22.4 Swaziland -- -- 0.6 0 0.4 5.4 0 0 0 7.9 0 33.5 Tanzania -- -- -- -- -- -- -- 3.9 17.2 2.0 2.7 6.4 Tunisia 4.1 4.8 11.6 16.9 6.6 11.2 8.3 1.6 22.1 13.7 38.9 12.7 Uganda -- -- -- -- -- -- -- -- -- -- 1.9 1.9 Zambia -- -- -- -- 1.5 1.1 1.0 4.3 3.4 24.4 0.9 5.2 Zimbabwe 3.2 3.7 9.6 7.4 7.0 27.4 14.2 9.0 11.5 19.2 1.1 10.3 Total Africa 7.2 7.5 7.2 6.8 10.9 19.0 29.1 26.3 35.0 39.6 35.1 20.3 Sub-Saharan -- -- -- -- -- 23.1 8.0 19.9 23.0 22.5 23.8 23.1 Africa Low- and -- -- -- -- -- 85.1 46.3 67.9 87.6 122.3 58.0 85.1 middle- income countries World -- -- -- -- -- 77.1 -- 86.8 87.6 74.7 123.6 77.1 Sources: IFC 1997; World Bank World Development Indicators 2001­4. -- Not available. FINANCIAL SECTOR REFORMS IN AFRICA | 105 Abrupt changes in government policies and political climate are common in Africa. These changes have adverse effects on financial markets. A case in point is the dramatic price swing in the Zimbabwe stock market, which rose 90 percent in 1996 and then fell more than 50 percent during the last quarter of 1997 in the wake of dramatic government farm and pension policies (Gopinath 1998).4 High currency exchange volatility is endemic to African economies, creating an impediment to foreign investments. In view of the dearth of hedging mechanisms through derivative markets (forward, futures, and options), an indirect approach would be to increase the number of export-oriented companies on the stock exchanges. In particular, exchanges with exposure to hard currency exports should be targeted, to provide substantial hedging against local currency devaluation. The preliminary regression results in the appendix show the adverse impact of currency depreciation on the performance of African stock markets. Afro-pessimism/pooling. Despite the substantial political, economic, and financial sector reforms that have taken place in the region, Sub-Saharan Africa still conjures up images of war, famine, massive corruption, project failure, poor governance, and gross violations of human rights. This information gap has serious consequences for pre-emerging African stock markets and African financial systems. "Afro- pessimism" leads to perceptions by potential investors of high political and invest- ment risks, as reflected in the low creditworthiness ratings of Sub-Saharan African countries relative to other regions (see table 11). Although these ratings rose signifi- cantly in the 1990s (from 18.1 in 1997 to 28.7 in 2003), they remain very low, sug- gesting high country risk. The perception of risk is excessive given the underlying fundamentals. Unfortu- nately, perception is reality in an environment characterized by grossly imperfect information, making it difficult to distinguish good prospects from bad ones. The average quality of the Africa pool may mask the high quality of genuinely reforming countries. Africa is a continent of much diversity in terms of financial sector reforms. FIGURE 4. Stock Market Liquidity (Turnover) in Sub-Saharan Africa, 1997­2003 1.4 1.2 1.0 0.8 cent Per 0.6 0.4 0.2 0 1997 1998 1999 2000 2001 2002 2003 Sub-Saharan Africa Developing countries World Sources: IFC 1997; World Bank World Development Indicators 2001­4. 106 | LEMMA W. SENBET AND ISAAC OTCHERE TABLE 11. Country Risk Ratings of Selected African Countries, 1996­2003 Percentage Annualized change percentage Country 1996 1997 1998 1999 2000 2001 2002 2003 since 1996 change Algeria 22.8 24.5 25.8 26.5 33.1 30.6 31.5 41.6 82.5 7.8 Botswana 49.8 51.2 51.9 36.5 56.1 56.7 59.0 62.2 24.9 2.8 Côte d'Ivoire 18.5 20.1 22.2 25.5 24.1 18.5 18.5 15.7 -15.1 -2.0 Egypt, Arab 35.1 39.7 43.2 45.4 51.0 47.1 45.5 41.1 17.1 2.0 Rep. of Ghana 29.6 31.5 30.3 30.7 29.5 25.4 25.7 25.8 -12.8 -1.7 Kenya 27.9 28.6 25.9 24.8 25.0 21.7 22.9 24.6 -11.8 -1.6 Malawi 19.7 21.0 19.8 19.5 19.6 17.9 19.6 18.8 -4.6 -0.6 Mauritius 50.8 51.9 53.0 53.9 54.6 53.9 53.5 53.9 6.1 0.7 Morocco 39.3 40.9 42.4 44.3 47.3 45.4 48.2 49.4 25.7 2.9 Namibia -- -- -- 38.0 37.4 39.1 40.8 39.8 4.7 0.9 Nigeria 15.2 15.3 16.4 17.9 18.1 18.3 17.6 20.2 32.9 3.6 South Africa 46.3 46.4 46.6 45.6 55.1 49.5 52.7 54.6 17.9 2.1 Swaziland -- -- -- -- -- 27.0 28.2 30.7 13.7 4.4 Tanzania -- 18.7 19.9 19.5 20.3 20.6 21.3 21.8 16.6 2.2 Tunisia 45.5 47.9 49.0 50.3 54.5 50.8 53.7 52.6 15.6 1.8 Uganda 16.1 20.1 19.9 21.7 22.7 21.4 20.0 20.1 24.8 2.8 Zambia 16.5 16.0 17.2 14.9 15.5 16.0 15.8 15.3 -7.3 -0.9 Zimbabwe 32.5 33.8 29.8 25.1 17.1 13.0 11.9 11.0 -66.2 -12.7 Sub-Saharan -- 18.1 18.5 19.3 17.9 17.5 19.0 28.7 58.6 6.8 Africa Low- and -- 28.6 28.6 28.5 28.0 27.0 26.8 25.3 -11.5 -1.7 middle- income countries World -- 38.6 35.3 36.0 37.0 33.9 33.5 30.4 -21.2 -3.4 Sources: World Bank World Development Indicators 2002­4; IFC 1997. -- Not available. This information gap can be minimized by providing potential investors with timely and reliable data with which to estimate investment risks in Africa. Consequently, there is a need for more extensive, detailed, and reliable data that capture the diver- sity of Africa, along with data describing the financial circumstances of private insti- tutions within the formal financial system. Globalization and capital flow volatility. Financial globalization engenders consid- erable risks and associated crises. The global risks stemming from volatility of global financial markets, large unfavorable international exchange rate fluctuations, and large unfavorable international interest rate movements manifest themselves in large unfavorable swings in international capital flows. For instance, countries that expe- rienced large capital flows suffered commensurately large and sudden outflows. The sudden and large collapse of capital inflows can be enormously costly. Depositors could withdraw their deposits, leading to a credit crunch. Creditors could refuse to FINANCIAL SECTOR REFORMS IN AFRICA | 107 supply short-term credit, leading to default and contagion of the type experienced in East Asia (and earlier in Mexico). The damage associated with financial crisis is evi- dent in the dramatic declines in asset and currency values.5 Currency, stock market, and property values declined by 30 to 50 percent in Indonesia, the Republic of Korea, Malaysia, and Thailand during June­December 1997 (Goldstein 2001). Financial instability is accompanied by poor or declining economic performance. GDP declined in the first quarter of 1998 in Hong Kong (China), Indonesia, Japan, the Republic of Korea, Malaysia, and Thailand. The unfavorable economic performance of the East Asian region was partly attributable to the negative wealth effect arising from dramatic declines in currency and asset values. In addition, IMF austerity meas- ures typically called for budget discipline (that is, reductions in government spending and budget deficits) and increases in interest rates to avert the further outflow of cap- ital. The spillover effect of the Asian crisis was worldwide, affecting even the United States, where Asia-dependent companies experienced declining profits. The East Asian countries have been back in recovery and those that genuinely reformed their financial systems in responding to financial crisis are particularly performing well. Lack of risk control mechanisms and trained personnel. Increasingly sophisticated financial systems require well-trained personnel. Inadequately trained staff, along with inefficient management, contributes to the dysfunction of financial systems. This problem is evident in Africa, despite serious attempts to improve conditions through training programs. However, these programs have been limited and inade- quate (Aryeetey and Senbet 1999). The investment climate in most African countries has improved.6 Institutional Investor rates a country's chance of default between 0 (highest risk) and 100 (lowest risk). Between 1997 and 2003, the indicator rose from 18.1 to almost 29.0, an improvement of more than 50 percent. Only Côte d'Ivoire, Ghana, Kenya, Malawi, Zambia, and Zimbabwe experienced deterioration in the investment/country risk rat- ing (table 11). Algeria, Nigeria, and Swaziland recorded significant improvement. During this period, the investor risk rating of low- and middle-income countries and the world as a whole worsened. Policy Recommendations: Building the Capacity of African Financial Systems This section puts forth several policy prescriptions to help develop and enhance the capacity of African financial systems to fill the severe financing gap in the region. Financial sector development affects economic development through the multiple functions the systems perform (among them the provision of liquidity). Promoting economic development has long-run implications for poverty reduction in Africa. Developing the Domestic Financial Sector It would be tempting to think that African countries could bypass the immense chal- lenges facing their domestic financial systems and access global financial markets to 108 | LEMMA W. SENBET AND ISAAC OTCHERE meet the severe financing gap. This temptation is fueled by the prospect of domestic companies cross-listing on foreign stock exchanges or issuing depository right securities traded on foreign exchanges. This view is misguided. While financial glob- alization can be an important complement, it cannot be a substitute for domestic financial sector development. In fact, a vital source of integrating Africa into the global financial economy is the existence of deep and well-functioning domestic finan- cial systems, including local banking systems and local stock markets. Well- functioning domestic financial systems are vital for domestic resource mobilization, because they provide incentives and profitable options for domestic capital to be retained. Retention of domestic capital is crucial in view of the massive financial cap- ital flight from Africa. Financial sector development goes beyond developing measures for financial global- ization as a means of attracting foreign capital inflows. It also seeks to mobilize domes- tic investor interest by cultivating attractive investment opportunities to channel savings within the domestic economy. Africa has potential channels for investment opportuni- ties with the potential for wider local participation in the stock markets, credit markets, and mortgage markets. As in other countries, investor participation is possible directly or indirectly, through such vehicles as pension plans, insurance policies, and mutual funds. Domestic investors are also more likely to be better informed about economic opportunities in their country than investors in other countries; development of the domestic financial sector is one mechanism of empowering domestic investors. Thus, it is crucial to create incentives for developing domestic channels for attractive investment opportunities as part of a broad agenda for financial sector development. Designing Efficient Regulatory Schemes and Safety Nets Regulatory and supervisory systems have not been effective in most African countries. This is in part reflected in the prevalence of banking failure and distress even in those countries that have reformed extensively. Some regulatory regimes have been too strict, reducing the bank risk taking that is vital for efficiently allocating resources in the economy. The purpose of banking regulation is not to eliminate banking failure but to curtail general, systemic banking crises. This is done with an appropriate reg- ulatory and supervisory scheme that relies on capital adequacy requirements, surveil- lance on asset risk choices, and rapid resolution of crises.7 Restricting Bank Asset Choices One way to regulate bank behavior is to impose mandatory restrictions on bank asset choices to limit bank risk taking. The regulatory restriction of bank asset risk choices can be counterproductive, however. Often such restrictions boil down to limiting banks to a common pool of risk. Such restrictions on asset choices may lead to socially counterproductive outcomes. Different banks face different investment opportunity sets, with different risk choices and hence different value creation strate- gies. Forcing banks to choose from a common pool of low-risk investments leads to the inefficient allocation of resources. In this sense, capital regulation, in which the bank's own incentives limit risk shifting, is superior to asset regulation. FINANCIAL SECTOR REFORMS IN AFRICA | 109 Regulating through Capital Adequacy Requirements Capital regulation is a useful tool for reducing the risk-shifting incentives of banks. Because the degree of risk distortion is larger with greater leverage (lower bank cap- ital), regulatory measures designed to move a bank that is critically undercapitalized to a higher level of capitalization will reduce incentives for excessive risk taking. However, capital regulation has only limited effectiveness in controlling risk incen- tives. Because banks may have different opportunities to exploit risk within a capital zone, placing all banks within the same zone into the same risk classification is inap- propriate. In this sense, capital ratios are poor proxies for measuring bank safety (see John, Saunders, and Senbet 2000). This point highlights the danger of standardizing capital adequacy requirements across countries (through the Basel Accord, for exam- ple); capital regulatory rules cannot be applied to all banks even in a single country. If such schemes are adopted, they should at least be country specific (see Kane and Rice 2001 for alternative views). Designing Appropriate Deposit Insurance Schemes A debate continues on the relative efficacy of explicit and implicit insurance of bank deposits. The debate is not resolved by abolishing deposit insurance altogether. Even in countries that lack formal or explicit deposit insurance schemes (that is, most African countries), deposits are implicitly insured. Often banks that are presumed to be too big to fail are bailed out by governments to minimize the adverse impact on society. Deposit insurance can play a useful function in enhancing the stability of financial institutions and helping reinforce their monitoring functions. However, given the already shaky state of African financial institutions, it is important that any such schemes be incentive compatible and not increase the risk-taking incentives of finan- cial institutions (and hence enhance financial sector instability) (see Cull, Senbet, and Sorge 2005). This suggests that rather than importing deposit insurance systems, with all their imperfections, from developed countries, such as the United States, African schemes should be designed. One clear advantage of explicit insurance is that the regulatory agency can charge an insurance premium, thereby reducing pressure on the government budget. This is not possible with implicit insurance. Another, indirect advantage of explicit deposit insurance is that the pricing of the insurance can be used in the regulatory process, since it has implications for bank incentives (see below). Regulating through Incentive Features of Management Compensation The regulatory schemes involving bank capital and assets can be improved upon by explicit consideration of the incentives of bank decision makers (or top manage- ment). The extent to which bank management is aligned with shareholder interests depends on the structure of compensation in place. Bank regulation can be more effi- cient if it takes account of these incentive features of compensation in pricing deposit insurance and disciplining bank risk behavior. John, Saunders, and Senbet (2000) show that managerial compensation interacts with capital regulation to induce bank risk choices that are optimal from the standpoint of society. The principal reason is that the risk-shifting incentives embodied in bank equity can be mitigated through 110 | LEMMA W. SENBET AND ISAAC OTCHERE appropriate design of managerial compensation contracts. Consider, for instance, the incentive features of management compensation involving fixed salary, bonus, and equity participation. Designing an optimal structure can produce risk incentives that induce bank management to be neither overly risk aggressive nor overly conservative. This calls for fair pricing of deposit insurance premiums that reflects the incentive features of bank management compensation. If deposit insurance is fairly priced, it pays bank owners to precommit to socially optimal risk levels by designing (declar- ing to the government regulator) an optimal compensation structure with the appro- priate incentive features. African bank regulators should seriously consider adopting an incentivized regulatory system. Fostering Bank Competition and Accelerating Bank Privatization Financial activity in Sub-Saharan Africa is characterized largely by the oligopolistic behavior of a few commercial banks (in several cases, government owned). The absence of adequate competition is reflected in the large gap between deposit rates for savers, which tend to be very low, and interest rates for borrowers, which tend to be very high. There is also evidence that return on bank equity is higher in Africa than in other developing regions (Applegarth 2004). By cultivating channels through which firms can issue various debt instruments and raise equity while simultaneously provid- ing more long-term options for saving and asset management for investors, financial activity can become more market efficient and conducive to economic growth. Privatization is one way to curtail state dominance of the banking system and fos- ter banking competition. However, the performance of African bank privatizations has not been encouraging, from the standpoint of operating performance or effi- ciency. The continuing partial state ownership of these banks interferes with the full achievement of gains from bank privatizations; as a policy matter, privatization pro- grams should be genuine and devoid of government intervention. Partial privatiza- tions reduce the performance and efficiency gains from privatization. The emergence of nonbank financial institutions and activities would help foster competition in African financial systems. The development and building capacity of African stock markets can be viewed in this same framework, because in addition to injecting competition into the system, stock markets allow for mobilization of capi- tal beyond traditional bank savings accounts. Moreover, the nonbanking systems can allow for large-scale privatizations of state-owned enterprises (see the discussion below of privatization through stock markets). African financial systems should cre- ate an environment that also enables the development of mutual funds or unit trusts as well as pension funds and insurance companies. Crafting Efficient Capital Market Regulation Capital markets cannot be expected to develop without credible legal and regulatory schemes that promote, rather than inhibit, private initiative. These schemes are also a foundation for fostering regional stock markets and attracting international invest- ments, ultimately integrating Africa into the global economy. Policy makers need to take care, however, not to overregulate capital markets or take a paternalistic view of investors. It is not the job of the regulator to determine what is best for investors but to create an environment in which investors make FINANCIAL SECTOR REFORMS IN AFRICA | 111 informed decisions. Regulatory emphasis should therefore be on fairness, full disclo- sure, and transparency. Government regulation of securities markets, particularly stock markets, should involve oversight of self-regulatory agencies, such as the stock exchanges and brokerage industry. Self-regulatory organizations design rules for business operations and professional conduct of properly licensed members. The oversight function itself is typically done by securities and exchange commissions, which are organs of the government. Self-regulation builds on the capacity and wis- dom of people inside member firms rather than on government bureaucrats, who lack intimate knowledge of the day-to-day functions of markets. Increasing the Depth and Liquidity of African Stock Markets Several measures can enhance the liquidity and market depths of African stock markets. Fostering Public Confidence and Informational Efficiency Maintaining public confidence and informational efficiency through transparent and credible financial disclosure rules is central to the functioning of financial markets. Transparent and trusted financial statements are a fundamental element of informa- tion about public and private firms. Confidence is particularly crucial in Africa. Public confidence is fostered by the creation of an even playing field, with strict enforcement of existing rules. The mere existence of legislation, which declares and grants inalienable property rights, is inconsequential. What is needed is an independent judiciary that strongly enforces and protects property rights. The government's role is vital in ensuring the enforce- ability of private contracts, the use of accounting procedures, and the upholding of legal standards. Capital market development and functioning require regulatory schemes that promote, rather than inhibit, private initiative and build investors' and savers' confidence in the functioning of markets. Developing Incentives for Listing Some African governments already give tax incentives to listed companies. Ghana's regular corporate tax rate of 32.5 percent is reduced to 30 percent for listed compa- nies. Zambia's regular corporate tax rate of 35 percent is reduced to 30 percent for listed companies (Mensah 2003). It would also be useful for large multinational com- panies already operating in Africa to consider mobilizing financing locally, by issu- ing debt or equity in domestic markets. Their participation would help deepen domestic capital markets. Their involvement in these markets requires partnership with the home countries of these multinational corporations, because companies ben- efit from donor-subsidized financing or risk insurance mechanisms, such as the Over- seas Private Investment Corporation (OPIC).8 Strengthening local insurance, pen- sion, and social security institutions can increase domestic savings and improve the quality of investment decision making. Listing is, of course, necessary but not suffi- cient for capital market development. Trading and liquidity are also needed to link stock market development and economic growth. Another impediment to listing is limited knowledge of accounting and disclosure requirements. It is important that stock exchange authorities design a program to 112 | LEMMA W. SENBET AND ISAAC OTCHERE help firms meet these requirements. These programs could be provided in-house or with technical assistance from advanced countries with well-established stock mar- kets. The level of awareness should be such that companies are able not only to list locally but also to cross-list globally. African firms, stock exchanges, market partic- ipants, and securities regulators should be exposed to best international practices in listing requirements. Consolidating African Stock Markets Stock market consolidation and regional cooperation is an important vehicle for pooling resources and building the capacity of illiquid and thin markets. These efforts are already under way. Regional stock exchanges are being created along lin- guistic lines. As noted above, the Francophone countries of West Africa formed the first regional stock exchange, the Abidjan-based Bourse Régionale des Valeurs Mobiliéres, in 1998. The Anglophone countries of West Africa are contemplating forming a regional stock exchange under the umbrella of the ECOWAS. Kenya, Tan- zania, and Uganda have contemplated formation of a regional stock exchange, as have the Southern African Development Community (SADEC) stock exchanges. Regionalization of African stock markets should enhance the mobilization of both domestic and global financial resources and inject more liquidity into the markets (Senbet 2001). Mechanisms for regional integration include establishing regional securities and exchange commissions, regional self-regulatory organizations, regional committees to promote harmonization of legal and regulatory schemes, and coordi- nated monetary arrangements (through currency zones, for example). The tax treat- ment of investments must be harmonized, because tax policy is an important incen- tive or disincentive for both issuers and investors. Ultimately, the regulations; accounting reporting systems; and clearance, settlement, and depository systems should conform to international standards. Some crucial elements of stock market consolidation or regionalization are discussed below. Harmonizing rules and regulations. Regionalization requires a strong commitment on the part of African economies to harmonize legal and regulatory schemes, accounting and disclosure rules, tax regulations and incentives, and fiscal and mon- etary policies. Cross-border monitoring and enforcement of laws may enhance com- petition among member countries in the region and enhance public confidence in the markets. Thus, as a prerequisite to large-scale cross-border trading, the infrastructure for the domestic capital markets and the regulatory regimes needs to be strengthened. Synergizing development efforts in human resources. Regional cooperation in cap- ital market development calls for a regional approach for skills development, train- ing programs, and research and information collaboration. Regionalization or sub- regionalization is an essential element facing the continent in its effort to develop and build capacity of capital markets, as well as to integrate into the global economy. Pooling resources for the production of capital market data and research. Global and local investors are poorly informed about African financial systems. A compre- hensive effort to develop and build the capacity of African stock markets and consol- idate these markets requires establishment of a research and information arm. This is FINANCIAL SECTOR REFORMS IN AFRICA | 113 another area of synergy and team effort calling for regional cooperation. In the short- run it would be desirable to leverage the activities of some important institutions that are already in place. Such institutions include the African Economic Research Con- sortium, the United Nations Economic Commission for Africa, the African Develop- ment Bank, the International Development Research Centre, and the New Partner- ship for African Development (NEPAD). Databases on capital markets and banking need to be created to allow first-rate research to be conducted by African financial economists as well as researchers outside Africa. This research should eventually lead to the establishment of a frontier Africa index or regional indices monitored both by practitioners and researchers. Governments may play a role in facilitating research coverage of companies on other African exchanges and developing a frontier Africa index or regional indices. Privatizing through African Stock Markets Stock markets can be an important avenue for privatization. A growing number of state-owned enterprises--such as Kenya Airways--are using this vehicle for privati- zation. These stock market­based privatization programs contribute directly to the depth of stock markets by increasing the number of listed companies. Privatizing state-owned enterprises through the stock market also brings the discipline of stock markets to these firms. Privatization is thus one way to take advantage of the multi- ple functions stock markets provide. The stock market is also an important means of depoliticizing privatization pro- grams, because it makes it possible to privatize large-scale enterprises at fair prices. Local stock markets also provide an opportunity for local investors to purchase shares. This helps allay some concerns about foreign control of privatized assets. Such privatization programs also foster diversity of ownership of the economy's resources and help dispel concerns that the stock market is just for the elite. Privati- zation through the stock market, as opposed to outright sale to an individual or a favored group, promotes distribution of ownership. It also increases public aware- ness about the market by attracting first-time buyers of market instruments. Institu- tional funds or unit trusts are effective means of involving small local investors in large-scale privatization.9 Building Capacity in Human Resources and Training Programs Global financial markets have become highly sophisticated in recent years. They are increasingly characterized by advanced and exotic securities, including a variety of derivative securities. Derivates are useful mechanisms for hedging risk, but if mis- managed they can lead to financial disaster. For this reason, training of financial managers should be at the forefront of financial market development in Africa. It can be done through improved business school curricula at universities and training pro- grams at capital market institutions, including securities and exchange commissions, central banks, and stock exchanges. Well-functioning markets have well-informed participants: investors, investment advisors, brokers, accountants, government regulators, and self-regulators. Such 114 | LEMMA W. SENBET AND ISAAC OTCHERE market participants and regulators are in short supply in Sub-Saharan Africa. In addition to university programs, specialized training programs are needed to train financial managers and regulators. This is an area calling for regional cooperation and approach. Indeed, a regional approach for skills development, training pro- grams, and research and information collaboration facilitates the development of regional markets. Human capital development should be thought of more comprehensively to include the banking sector. Capacity building in risk assessment and control is vital for banks to carry out their functions in building information capital. Banks that fail to develop the capacity to assess risk and monitor the optimal management of their loan portfolios are not interested in investing in information capital. This capital is crucial for the development and functioning of financial systems and the integration of otherwise locally fragmented markets. Fostering an Environment for Good Corporate Governance Different stakeholders in a firm have different interests. Appropriate mechanisms need to be available to stakeholders of corporations to exercise control over corpo- rate insiders and management so that their interests are protected. Corporate gover- nance provides such mechanisms. One familiar element of corporate governance is the board of directors. In recent years its effectiveness has been called into question. The growing consensus is that the board has to be independent of the chief executive officer, through appointments of directors who are outsiders with no serious business interests with the firm. The audit committee of the board should be composed entirely of outside directors. It is also often suggested that the audit committee appoint an external auditor (see John and Senbet 1998 for the specific elements of board effectiveness). Corporate scandals in the United States involving Enron and WorldCom are attributable to the failure of corpo- rate governance. Board members and the external auditor were held hostage by cor- porate insiders and did not carry out their functions properly. Given the crucial role of transparent financial statements in fostering confidence in financial markets, African governments can play a vital role in improving corporate governance by establishing institutions that will ensure that firms adopt globally accepted accounting and disclo- sure rules and regulations.10 Thus, corporate governance and an appropriately designed bankruptcy code that provides sufficient creditor and debtor rights should be key ingredients of any agenda for stock market development. Well-functioning corpo- rate governance will also facilitate the globalization of African financial markets. Promoting the Development of Regional Credit Rating Agencies A viable and credible regional credit rating agency is needed in Africa. Its establish- ment could be facilitated by pooling resources. Thus, regional cooperation can go beyond stock markets into the development of debt or bond markets. A regional credit rating agency is vital for developing secondary markets in pri- vate bonds. While individual markets themselves may be too small to support rating agencies, a sufficient number of debt instruments available in the region may support the establishment of a regional credit rating agency. Such an agency would be a cat- FINANCIAL SECTOR REFORMS IN AFRICA | 115 alyst for cross-border trading in debt securities by providing an assessment of sover- eign and credit risk for investors with limited knowledge of debt in other countries. The rating agency could rate both corporate and government debt. Creation of such an agency should be an important step in facilitating the issuance of long-term gov- ernment bonds, which can then serve as a benchmark for the issuance of long-term private corporate debt. Promoting the Financial Globalization of Africa Africa has been left out of the massive international capital that flowed to develop- ing economies with the opening up of the world economy in the 1980s. As a result, it has also been insulated from the adverse shocks of the crises in Mexico and East Asia. Negative wealth shocks and global contagion can be enormously costly to the countries affected and potentially destabilizing to the entire global economy. This does not mean that Africa should continue to remain insulated from finan- cial globalization. Avoiding globalization's risks marginalizes Africa. The East Asian financial crisis erupted in the wake of Africa's gradual embrace of financial global- ization and its slow march toward the emerging markets club. The lessons from the crisis, if drawn correctly, can help Africa adopt sustainable financial globalization (see Senbet 2001 for details on globalization). Among these lessons is the need to build information capacity that allows for more extensive, detailed, and reliable data capturing the diversity of Africa, along with data capturing the financial circumstances of private institutions, listed companies, and banks. The timeliness and reliability of financial data are crucial for assessing invest- ment risk. The measures discussed above to develop and build the capacity of African financial systems would help achieve sustainable financial globalization in Africa. Financial crises do occur, of course, even when reform measures are in place. Therefore, African financial systems should increase the capacity to resolve or con- trol crises in an efficient fashion. Controlling financial crisis calls for speedy mea- sures, including rapid closure of failed banks or restructuring of their balance sheets and quick restructuring of insolvent firms. The tenure of failed banks should not be prolonged, because it simply transfers private losses to taxpayers. Moreover, in a cri- sis environment, failed banks have incentives to take even greater risks. Future Research This paper takes a functional perspective in diagnosing African financial systems, but the analysis is based on indirect indicators, such as liquidity provision. As data on direct indicators (such as corporate governance) are made available, a more direct diagnosis can be made and modern analytics brought to bear on the capacity of African financial systems. This paper uses evidence from other emerging markets to study the link between financial sector development and economic performance. As more cross-sectional and time series data on the channels for this link (for example, liquidity) become available, the issue can be addressed directly based on African data. 116 | LEMMA W. SENBET AND ISAAC OTCHERE Better data will also allow researchers to analyze the contribution of African stock markets to global markets. They will allow researchers to expand performance analysis to fully capture the risk dimensions affecting African financial systems. Measures that help retain domestic capital help attract international capital and reverse capital flight. The gap in financial globalization can fuel capital flight. It would be interesting to study the issue of capital flight in the context of reforming African financial systems. Doing so requires a capital market database commensu- rate with the data already available on African capital flight. Financial sector reforms have not brought about the desired improvements in per- formance in the sector; informal and microfinance institutions continue to respond to the financing gaps faced by small borrowers in both rural and urban sectors. The credit flow through informal finance has not increased significantly, however (Nis- sanke and Aryeetey 1998). The microfinance sector faces several challenges, includ- ing the lack of transparency, the price of credit, and the limited integration between the informal and formal sectors. Future research should study financial integration and the introduction of appropriate legal and regulatory regimes. Linking microfi- nance with the formal sector would be one way to get savings mobilization into the open at an integrated market price. The research agenda should also examine the extent to which measures to develop and build the capacity of the formal financial sectors contribute to the "formalization" of microfinance. Appendix The factors shown below constitute only a partial list of relevant factors. A more complete analysis, which includes such factors as investor protection, legal origin, and capitalization, is under way. These preliminary results are nevertheless indicative of the predicted signs. Stock market performance (the dependent variable) is the aver- age three-year stock market return in the African country. Liquidity is measured by the value of the shares traded as a ratio of market capitalization. The exchange rate is proxied by the three-year exchange rate between the domestic currency and the U.S. dollar. APPENDIX TABLE 1. Factors Affecting Stock Market Performance in Africa Factor Effect Explanation Liquidity Positive The higher the liquidity, the lower the returns. Change in exchange rate Negative The higher the depreciation of the African currency relative to the U.S. dollar, the lower the returns. Country risk Negative The higher the institutional credit rating (the lower the country risk), the better the stock market performance. Source: Authors. FINANCIAL SECTOR REFORMS IN AFRICA | 117 Notes 1. Three other exchanges not shown in table 1 (the Khartoum [Sudan] Stock Exchange, the Maputo [Mozambique] Stock Exchange, and the Cameroon Stock Exchange) have only recently been established or have been inactive and are therefore not included in the analysis. 2. The stylized facts in table 2 are based on the evidence provided by Tadesse (2004) on the link between liquidity and economic performance. 3. The evidence points to a low correlation between Africa and developed countries, gener- ating potential for beneficial global diversification. Data on stock return correlations among selected African and non-African countries are available from the authors upon request. 4. The two policy changes were land reform to take over 1,500 mostly white-owned com- mercial farms and a decision to pay US$240 million in pensions to disgruntled veterans of the Zimbabwe independence war. 5. See Gande, John, and Senbet (2004) for the role of distorted incentives in making emerg- ing economies more vulnerable to financial crisis and the proposed mechanisms to pre- vent financial crisis. See the African Economic Research Consortium plenary synthesis by Senbet (2001) for more detailed analysis of global financial crisis and its implications for Africa. 6. Country risk rating of African countries by Moodys and Standard & Poor's is a recent phenomenon. 7. Financial globalization can have a spillover effect on bank behavior and the associated regulation. As Kane and Rice (2001) suggest, the discipline coming from globalization brings banking crises more into the open. They argue that the severity of such crises is diminished, because globalization increases the social cost on taxpayers and makes it less likely that regulators keep banking insolvency hidden. The policy implication is straight- forward: encourage entry by foreign banks into Africa as a mechanism for pressure and market discipline. 8. Chevron/Texaco recently raised financing in Angola (Applegarth 2004). The U.S. Export- Import Bank recently provided a bond guarantee on South Africa's bond exchange in con- nection with financing for the purchase of U.S.­manufactured aircraft. The bond is sup- ported by the full faith and credit of the U.S. government. It provides an opportunity for local investors to access an investment-grade security while simultaneously increasing U.S. exports to Africa. This is another example of mutual gains from globalization. 9. Privatization should also include financial institutions, so that they can perform their function of delegated monitoring and help achieve the efficient allocation of resources. 10. A welcome recent development is NEPAD's identification of economic and corporate governance as one of its priority areas. Governance has been tasked to a committee of finance ministers for developing action plans. References Applegarth, P. 2004. "Capital Market and Financial Sector Development in Sub-Saharan Africa." Report of the Africa Policy Advisory Panel, Center for Strategic and International Studies, Washington, DC. Aryeetey, E., and L. W. Senbet. 1999. "Essential Financial Market Reforms in Africa." Paper presented at World Bank Conference on Twenty-First Century Africa, Abidjan. Barnea, A., R. A. Haugen, and L. W. Senbet. 1985. Agency Problems and Financial Contract- ing. Englewood Cliffs, NJ: Prentice-Hall, Inc. 118 | LEMMA W. SENBET AND ISAAC OTCHERE Barth, J. R., G. Caprio, and R. Levine 2004. "Bank Regulation and Supervision: What Works Best." Journal of Financial Intermediation 13 (2): 205­48. Boyce, J. K., and L. Ndikumana. 2001. "Is Africa a Net Creditor? New Estimates of Capital Flight from Severely Indebted Sub-Saharan African Countries 1970­1996." Journal of Development Studies 38 (2): 27­56. Chirwa, E., and M. Mlachila. 2004. "Financial Reforms and Interest Rate Spreads in the Banking System in Malawi." IMF Staff Papers 51 (1): 96­122. Clarke, G., R. Cull, and M. Shirley 2004. "Empirical Studies of Bank Privatization: Some Lessons." World Bank, Development Economics Research Group, Washington, DC. Cull, R., L. W. Senbet, and M. Sorge. 2005. "Deposit Insurance and Financial Development." Journal of Money, Credit and Banking 37 (1): 43­82. Diwan, I., V. Errunza, and L. Senbet, 1994. "Diversification Benefits of Country Funds." In Investing in Emerging Markets, ed. M. Howell and S. Claessens, 199­214. Washington, DC: World Bank. D'Souza, J., W. Megginson, and R. Nash. 2000. "Determinants of Performance Improvements in Privatized Firms: The Role of Restructuring and Corporate Governance." Mimeo, Uni- versity of Oklahoma. Errunza, V., L. W. Senbet, and K. Hogan. 1998. "The Pricing of Country Funds from Emerg- ing Markets: Theory and Evidence." International Journal of Theoretical and Applied Finance 1 (1): 111­43. Gande, A., K. John, and L. Senbet. 2004. "The Role of Incentives in the Prevention of Finan- cial Crisis in Emerging Economies." Mimeo, University of Maryland. Goldstein, M. 2001. "The Asian Financial Crisis: Origins, Policy Prescriptions and Lessons." Journal of African Economies 10 (Supplement): 72­103. Gopinath, D. 1998. "Taking the Road Less Traveled." Institutional Investor (March): 173­4. Haque, N., R. Hauswald, and L. W. Senbet. 1997. "Financial Market Development in Emerg- ing Economies: A Functional Approach." Working paper, Robert H. Smith School of Busi- ness, University of Maryland, College Park. Hauswald, R., and L. W. Senbet. 1999. "Public and Private Agency Conflict in Banking Reg- ulation." Working paper, Robert H. Smith School of Business, University of Maryland, College Park. IFC (International Finance Corporation). 1997. Emerging Stock Markets Factbook. Washing- ton, DC: IFC. Inanga, E., and D. Ekpenyong. 2004. "Financial Liberalization in Africa: Legal and Institu- tional Frameworks and Lessons from Other Less Developed Countries." International Development Research Centre, Ottawa, Canada. John, K., and L. W. Senbet. 1998. "Corporate Governance and Board Effectiveness." Journal of Banking and Finance 22 (4): 371­403. John, K., A. Saunders, and L. W. Senbet. 2000. "A Theory of Banking Regulation and Man- agement Compensation." Review of Financial Studies 13 (Spring): 95­125. Kane, E., and T. Rice. 2001. "Bank Runs and Banking Policies: Lessons for African Policy- makers." Journal of African Economies 10 (Supplement): 36­71. Lamba, A., and I. Otchere. 2001. "Analysis of the Linkages among the African Equity Mar- kets and the World's Major Equity Markets." African Finance Journal 3 (2): 1­25. Levine, R. 1997. "Financial Development and Economic Growth: Views and Agenda." Jour- nal of Economic Literature 35 (2): 688­726. FINANCIAL SECTOR REFORMS IN AFRICA | 119 Levine, R., and S. Zervos. 1998. "Stock Markets, Banks, and Growth." American Economic Review 88 (3): 537­58. McKinnon, R. I. 1973. Money and Capital Market in Economic Development. Washington, DC: Brookings Institution. Megginson, W. 2002. "Sample Firms Privatized through Public Share Offerings, 1961­November 2002." Working paper, Michael F. Price College of Business, University of Oklahoma, Norman. Megginson, W., R. C. Nash, and M. van Randenborgh. 1994. "Financial and Operating Per- formance of Newly Privatized Firms: An International Empirical Analysis." Journal of Finance 49 (2): 403­52. Mensah, S., 2003. African Capital Market Forum Quarterly Newsletter. Accra. Ndikumana, L., and L. W. Senbet. 2005. "Capital Flight from Africa: Strategies for Preven- tion and Reversal." Paper prepared for the World Bank workshop "Challenges of African Growth," Dakar, January. Nissanke, M., and E. Aryeetey. 1998. Financial Integration and Development in Sub-Saharan Africa. London and New York: Overseas Development Institute and Routledge. Senbet, L. 2001. "Global Financial Crisis: Implications for Africa." Journal of African Economies 10 (Supplement): 104­40. Shaw, E. 1973. Financial Deepening in Economic Development. New York: Oxford Univer- sity Press. Tadesse, S. 2004. "The Allocation and Monitoring Role of Capital Markets." Journal of Financial and Quantitative Analysis 39 (4): 701­30. UNDP (United Nations Development Programme). 2003. Africa Stock Markets Handbook. New York: UNDP. World Bank. 2003. Global Economic Prospects. Washington, DC: World Bank. ------. 2004. Global Economic Prospects. Washington, DC: World Bank. ------. Various years. World Development Indicators. Washington, DC. Comment on "Financial Sector Reforms in Africa: Perspectives on Issues and Policies," by Lemma W. Senbet and Isaac Otchere MTHULI NCUBE The paper, according to the authors, takes "a functional perspective and holistic approach" to analyzing financial sector reforms in Africa. The savings mobilization (quantity of capital) function, prevalent in the development literature (McKinnon 1973; Shaw 1973) is incomplete. Other functions of the financial system, which includes both the banking sector and capital markets, include information produc- tion, price discovery, risk-sharing, liquidity provision, promotion of contractual effi- ciency, promotion of governance, and global integration. This approach is motivated by recognizing that the environment is characterized by risk and uncertainty as well as agency problems generated by imperfect informa- tion flow. Overall, there is a high correlation between development and the liquidity of markets and between productivity gains and economic growth. I like this general approach of the paper. The paper does not discuss the markets for fixed income securities. Eskom, the electricity utility in South Africa, borrows from international capital markets by issu- ing corporate debt. Foreign investors hold South African government bonds, which they trade on the bond exchange. In the rest of Sub-Saharan Africa, only the primary market for Treasury bills exists as a monetary policy instrument. Secondary markets for government bonds do not exist, despite liberalization initiatives. The banking sector is the largest part of the financial system in most African coun- tries. In most countries, foreign banks operate alongside local banks. The foreign banks bring the benefits of global best practice and foreign credit lines, and they have the balance sheets to finance large infrastructure projects. The paper omits the glob- alization of the banking sector. The paper asserts that globalization of African markets may reverse capital flight. Instruments of globalization African companies have used to raise foreign capital and achieve a public listing in global equity markets are American Depositary Receipts and Global Depositary Receipts. A significant driver of capital flight is governance at the national level. Capital flight reversal is driven largely by good governance, which often delivers political and economic stability and economic growth. Perhaps globalized financial markets enhance only the fundamental impact of governance. Mthuli Ncube is professor of finance at the University of Witwatersrand Wits Business School, Johannesburg, South Africa. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 121 122 | MTHULI NCUBE FIGURE 1. Index of Financial Infrastructure in Selected African Countries, 2003 4.0 3.5 3.0 2.5 2.0 Index 1.5 1.0 0.5 0 BotswanaCam Ghana Kenya Les Malawi MozambiqueNamibiaNigeri South TanzaniaUganda Zambia Zimbabwe eroon oth o a Africa Source: Commonwealth Business Council 2003. The paper notes that the level of financial deepening, as measured by M2/GDP, has not improved significantly in Africa over the years. Perhaps broader indicators of the financial infrastructure should be used. A financial infrastructure index extracted from the Commonwealth Business Council index shows a different picture for the M2/GDP ratios (figure 1). For example, Botswana, which has a lower M2/GDP ratio than South Africa, has a higher financial infrastructure indicator. This measure could complement a ratio such as M2/GDP to capture the quality of financial institutions. The paper observes that banks in Africa hold a large part of their assets in the form of Treasury bills. They do so partly because they perceive other investments in Africa as highly risky and because lending opportunities are not well developed, due to the small private sector in most countries. Another reason for holding T-bills is that the banks are designated by the central banks as primary dealers in T-bills and are therefore obliged to participate in T-bill auctions conducted by the central bank. The central bank uses T-bills as a monetary policy tool for controlling liquidity in the economy. Therefore, banks end up with a large holding of T-bills for unavoidable monetary policy reasons. Yet another reason for holding T-bills is the weak banking environment. T-bills act as insurance against potential systemic risk. In a weak banking system, banks demand T-bills from one another as security for interbank deposits. T-bills then become the guarantor for the interbank market and insurance against systemic risk. Deposit insurance schemes are set up so that insurance premiums paid by each bank reflect the bank's risk class. The risk is that once the class is known to be unfavorable, bank failure and systemic problems could ensue. However, while deposit insurance schemes are poorly designed in Africa, as the paper correctly points out, designing deposit insurance schemes is difficult universally, even for developed countries. Good quality supervision is perhaps what matters in managing potential bank failures. The paper notes that the supervisory capacity of the central banks in Africa is weak, resulting in weak banking systems. In some cases supervision is excessive, crip- pling banking institutions for purely political reasons. Most African countries have just a few banks. Privatization of state banks is slow. One of the reasons for the slow privatization is that the banks provide loans to politi- COMMENT ON SENBET AND OTCHERE | 123 cians and government activities on favorable terms. State banks have a history of having bad loan books. The interest rate spread between deposit and lending or asset rates is wide in most African countries, for several reasons. First, the wide spread is a way of man- aging credit risk by maximizing profits on loans that do not default. Second, the conduct of monetary policy by the central bank may require that a portion of deposits be deposited with the central bank for no interest, in the form of a reserve requirement. The banks would have to make up for the forgone interest by raising lending rates. Third, innovative instruments for taking risk off the banks' balance sheets do not exist, forcing banks to manage credit risk by increasing the interest rate spread. The paper correctly points out that stock markets in Africa are generally illiquid. This partly reflects the lack of comprehensive pension fund reform and subsequent reform of the investment management institutions. If pension funds are changed from defined benefit to defined contribution schemes, the risk shifts to the pensioner, who begins to demand good performance from fund managers. The fund manage- ment market becomes more competitive, spurring more active investment in the stock market. In addition, large state-controlled provident funds have crowded out the private pension market. Furthermore, the national provident funds invest heav- ily in property, not the stock markets. The paper presents country ratings on the basis of the Institutional Investor rat- ings. These ratings are useful for investment in fixed income securities but not equity investments. It might make sense to use the country allocation models used by global fund managers, which are based on four broad factors: · Growth factors: GDP growth and export growth, inflation levels · Risk factors: the beta of the equity index of a country relative to the Morgan Stan- ley Capital International (MSCI) Emerging Markets Index, exchange rate risk · Liquidity factors: money supply growth, the level of real interest rates · Valuation factors: price-earnings ratios relative to the MSCI Emerging Markets price-earnings ratio, dividend yield, price-to-book ratio of each equity market The paper argues for the consolidation of African stock markets to create critical mass. This recommendation will work only if there is a monetary union, currency convertibility, or one regional currency, all of which are hard to achieve. A way for- ward would be to encourage cross-listing of companies across countries, which could increase liquidity across stock markets. The paper argues for the improvement of corporate governance standards in Africa in line with best international practice. One way to move ahead would be to monitor the quality of governance across countries to show progress. Figure 2 shows corpo- rate governance indicators for various African countries. The paper discusses the importance of the reliability of the justice system and gen- eral enforceability of contracts. We need to be able to track the reliability of the jus- tice system over time and across countries. Figure 3 is an index of reliable justice in selected African countries. 124 | MTHULI NCUBE FIGURE 2. Index of Corporate Governance in Selected African Countries, 2003 3.5 3.0 2.5 2.0 Index1.5 1.0 0.5 0 BotswanaCamer Ghana Kenya Lesotho Malawi MozambiqueNamibiaNigeria South Tanzan Uganda Zambia Zimbabwe oon Africa ia Source: Commonwealth Business Council 2003. FIGURE 3. Index of Reliable Justice in Selected African Countries, 2003 4.0 3.5 3.0 2.5 2.0 Index 1.5 1.0 0.5 0 BotswanaCamer Ghana Kenya Lesoth Malawi MozambiqueNamibiaNige Sout Ugan Zam Zimb h Africa ania Tanz bia oon ria o da abwe Source: Commonwealth Business Council 2003. The paper omits an important element in financial sector reforms in Africa: the link to monetary policy. Financial sector reforms in Africa have occurred alongside moves from direct to indirect monetary policy. Effective indirect monetary policy requires efficient capital markets and financial systems, which emit the right infor- mation signals for decision making. For instance, a well-developed bond market could produce information on future inflation through the shape of the yield curve. The information on inflation is then used for monetary policy decisions, particularly in setting the level of short-term interest rates. This important link should have been discussed in the paper for completeness. The informal sector is a large but unmeasured part of any African economy. Its size in Africa reflects the failure of the formal financial system. The informal finan- cial sector is vital in the economy and could also be a reason for the prevalence of disintermediation. The paper should have discussed the informal financial sector briefly, as both a symptom of formal market failure and a way of complementing the formal financial sector. References Commonwealth Business Council. 2003. "Business Environment Survey." London. McKinnon, R.I. 1973. Money and Capital in Economic Development. Washington, DC: Brookings Institution. Shaw, E.S. 1973. Financial Deepening in Economic Development. New York: Oxford Univer- sity Press. Economic Development Convergence and Development Traps: How Did Emerging Economies Escape the Underdevelopment Trap? JEAN-CLAUDE BERTHÉLEMY The underdevelopment trap hypothesis postulates that poor countries are locked in a low equilibrium and that big-push policies, involving massive external assistance, are necessary to lift them out of poverty. In fact, unless they are accompanied by structural change, transfers do not trigger successful take-off. Empirical examination of develop- ing countries reveals that initial education policies have played a critical role in allow- ing some countries to jump out of their underdevelopment trap. External assistance has not played a significant role. The idea of underdevelopment traps, fundamentally linked to the notion of multiple equilibria, is not new. It emerged at the very early stages of the development econom- ics literature and is associated in particular with seminal contributions by Young (1928), Rosenstein-Rodan (1943), and Nurkse (1953). It has been revisited by growth analysts since the mid-1980s, following empirical contributions by Abramovitz (1986) and Baumol (1986), who associated multiple equilibria with the notion of convergence clubs. This paper first reviews the empirical research results that provide some evidence of the existence of convergence clubs. Several complementary pieces of evidence are available: the appearance of twin peaks in the international distribution of incomes; the observation of a quadratic relation between initial income and future growth, observed both on a cross-section and on a time-series basis; and the dependence of structural parameters of conditional convergence equations on the initial level of some of the conditioning variables. A consequence of multiple equilibria is that a poor country cannot grow out of poverty unless some policy initiative is taken to change initial conditions in such a way that the country can "jump" from its initial low-level, stable equilibrium to another equilibrium that is equally stable but characterized by a higher level of income. Hence policies, not only initial conditions, matter very much in the discus- sion of convergence clubs and multiple equilibria. Jean-Claude Berthélemy is professor of economics at the University of Paris 1. He wishes to thank Christian Morrisson, Mthuli Ncube, Erick Thorbecke, and an anonymous referee for helpful comments. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 127 128 | JEAN-CLAUDE BERTHÉLEMY It has taken some time for new findings on multiple equilibria to be translated into practical policy recommendations. This approach has recently taken the form of sev- eral "big push" proposals, such as those by Sachs and others (2004), in their work for the United Nations Millennium Project; by Collier (2004), in the context of the discussions initiated by the Blair Commission for Africa; and by Radermacher (2004), in the context of the Global Marshall Plan Initiative supported by the Club of Rome. The idea of multiple equilibria is also very much behind the British pro- posal of a new International Finance Facility. Under that proposal, much larger aid budgets would be allocated up-front, financed by long-term borrowings by donor countries, so that the outcome of a given level of fiscal effort could be concentrated in a relatively short period, giving the necessary impetus to lift poor countries out of their underdevelopment trap. These ideas are intellectually appealing. The question, however, is whether they can lead to policy recommendations that make sense. This inquiry should start with an identification of factors that may be responsible for the underdevelopment trap in which poor countries may be locked. A first group of mechanisms may be related to factor accumulation in a broad sense: demography, savings behaviors, and human capital accumulation associated with education. A second group concerns more qual- itative aspects of the economic growth process, such as financial development or diversification of the economy. A third group has to do with political institutions, for example, corruption and conflicts, which may trap a nation in destructive dynamics in which poverty and poor institutions reinforce each other. Several of these mecha- nisms may be at work in different phases of the economic development process, cre- ating "multiple multiple equilibria." In the presence of multiple equilibria, two broad complementary strategies may be conceived to lift a poor country out of its underdevelopment trap. The pure "big push" strategy consists of artificially increasing the available income of such coun- tries through large transfers of assistance, under the assumption that such transfers will be sufficient to initiate a self-reinforcing economic growth process. This paper suggests that this strategy may very often be ineffective. A second, possibly more effective, strategy consists of promoting economic reforms in poor countries, which would increase their rate of economic growth for a while. This paper shows analyt- ically that this strategy has a better chance to be effective. The analytical discussion makes some predictions on the dynamics of the time series of a country that jumps from one stable equilibrium to another. In a nutshell, the argument is that this jump should lead to multiple growth peaks, or a "growth acceleration cycle" pattern. The most critical question is whether one may detect in the recent economic his- tory any evidence of a successful escape from the underdevelopment trap consistent with the previous analytical prediction. To answer this question, this paper uses Maddison's (2003) database, which provides, for a large number of countries, a com- plete series of GDP per capita measured in purchasing power parity, from 1950 to 2001 and sometimes to 2002 or 2003. These series are transformed using a Hodrick- Prescott filter, to eliminate the short-run cyclical component. In a majority of cases, the transformed series suggests the occurrence of growth peaks, or growth accelera- tion episodes. Single growth accelerations, however, cannot be interpreted simply as CONVERGENCE AND DEVELOPMENT TRAPS | 129 jumps from a stable low equilibrium to a stable higher equilibrium: there is no rea- son to believe that such shifts are not simply the result of a standard convergence process, from a relatively low starting point, below a stable equilibrium, to this sta- ble equilibrium. Nevertheless, in a variety of instances the observed time series dis- play the dynamics that would characterize a jump from a stable equilibrium to another--that is, they are characterized by multiple growth peaks. These cases are the most interesting ones from which to derive conclusions on the policies that could lift a country out of an underdevelopment trap. The relevant policy question is why some countries, starting at similar levels of development, have apparently jumped out of their underdevelopment trap and oth- ers have not. This question is germane to discussing the relative economic perform- ance over the past 50 years of African countries and other least developed economies on the one hand and Southeast Asian countries and other emerging economies on the other. Answering this question in a systematic way is not easy, given the dearth of comparative economic data for the period during which such economies diverged, in the early 1960s. One plausible explanation is that many Southeast Asian economies implemented early ambitious education policies, initially aiming at achieving univer- sal primary education, while most African economies have lagged behind in this respect. This is not due principally to low education expenditure in Africa but rather to very different education strategies. The paper is organized as follows. The first section reviews the stylized facts back- ing the convergence club hypothesis. The second section reviews possible explana- tions for multiple equilibria. The third section provides an analytical discussion of conditions under which a country can jump out of its underdevelopment trap. The fourth section studies the dynamics of per capita GDP of developing countries, in search of some evidence of an escape out of the underdevelopment trap. The fifth sec- tion examines different possible explanations of such successful take-offs. The last section draws some policy conclusions. Stylized Facts on Convergence Clubs The notion of multiple equilibria has become fashionable and is now known as the "poverty trap" hypothesis (Kraay and Radatz 2005). This expression is misleading, because it mixes two aspects: a macroeconomic analysis of factors that may lock a country in a low-level equilibrium and a microeconomic theory explaining why a poor individual stays poor. Obviously, the two notions are not independent, but they are not synonymous either: an underdevelopment trap may exist due to externalities, for instance. This paper focuses only on the macroeconomic aspect. To avoid confu- sion, the term underdevelopment trap is used. Before it become fashionable, the underdevelopment trap hypothesis was explored by growth analysts studying the notion of "convergence clubs." This notion is based on the idea that, although no absolute convergence of economies toward a similar level of development is observable, some local convergence properties can be observed. In other words, convergence clubs exist if countries have globally heterogeneous growth 130 | JEAN-CLAUDE BERTHÉLEMY dynamics, but countries may be grouped in subsets that exhibit homogeneous growth dynamics. Each "club" can then be considered as a group of countries in the same kind of equilibrium, in a multiple equilibria setting. Early evidence of underdevelopment traps can be found in the empirical literature on convergence clubs, following empirical contributions by Abramovitz (1986) and Baumol (1986). Baumol defines convergence clubs principally by policy regimes. He identifies three convergence clubs: OECD developed market economies, centrally planned economies, and a third club, less precisely defined, grouping middle-income countries. Later contributions insisted on a characterization of convergence clubs by initial conditions, in the tradition initiated by the early development economics literature. Chatterji (1992) proposes a very simple but empirically powerful approach link- ing the definition of convergence clubs to initial conditions. This approach consists of observing, on a cross-sectional comparative basis, that the rate of growth of GDP per capita depends quadratically, not linearly, on its initial level. In the growth equa- tions estimated by Chatterji, the initial GDP per capita typically has a positive parameter and its square a negative one, implying that at low levels of development, economic growth depends positively on initial income, while the standard absolute convergence relation (a negative relation between the initial income and successive growth) holds at higher levels of development. Hausmann, Pritchett, and Rodrik (2004) uncover a similar nonmonotonic con- cave relationship between the initial level of output per capita and economic growth, using a time-series approach rather than cross-sectional comparisons. Their results are germane to those proposed, in a comparative growth framework, by Thorbecke and Wan (2004), on East Asian emerging economies, and Berthélemy and Söderling (2001), on African countries. Quah (1997) shows that although the international partial distribution function of incomes was unimodal in 1960, it was bimodal some 40 years later. Therefore, countries have diverged, into at least two different groups, which may be viewed as convergence clubs. The notion of conditional convergence introduced by Barro (1991) runs counter to the assumption of convergence clubs. According to Barro and his numerous fol- lowers, the observed divergence of economies could be explained by the fact that the convergence process depends on a limited number of independent conditioning vari- ables. This would imply that, although economies converge to different steady states, their growth processes can be represented using the same model: instead of proper multiple equilibria, one would observe multiple variants of the same equilibrium, parameterized by the conditioning variables. Durlauf and Johnson (1995) challenge this conclusion. They show that a standard growth regression à la Barro has parameters that are significantly different in coun- tries at low levels and at higher levels of development. This suggests that, notwith- standing its relevance, the notion of conditional convergence does not preclude the existence of convergence clubs. Such convergence clubs are associated with different initial conditions, leading to different growth regimes, and not merely to different growth rates. Durlauf and Johnson identify two potential threshold points, defined CONVERGENCE AND DEVELOPMENT TRAPS | 131 by the initial level of education or the initial level of income per capita. Each of these thresholds separates countries into two groups, defining convergence clubs. Hansen (2000) provides a way to estimate a confidence interval for such thresholds. Berthélemy and Varoudakis (1996) suggest that at least one other factor, initial financial develop- ment, could be responsible for multiple equilibria. The Multiple Sources of Multiple Equilibria The possible existence of multiple equilibria was recognized very early on in the the- ory of economic growth. The standard argument is that cumulative processes lead to an economic decline when the economy is initially below a certain threshold of eco- nomic development, while economic progress is possible when this threshold has been passed. The principal arguments of this kind that have been considered in the literature are summarized here. To simplify the analysis, the various arguments are considered in isolation from one another. In each of these arguments, the variable driving the development level of the econ- omy plays the role of a state variable, z. The law of motion of this variable has a par- ticular shape, leading to cumulative processes. This state variable can be conceived as a factor of production, such as physical or human capital; as a structural economic characteristic, such as financial depth or diversification; or as an institutional feature, such as the proclivity of the economy toward corruption or civil strife. The law of motion of z can be defined as z. = h(z) (1) If h(z) is a monotonically decreasing function, there is only one dynamically stable steady state, which defines a common development level toward which all economies should converge. If it is not monotonic, there are possibly several steady states, as shown in figure 1. FIGURE 1. Multiple Equilibria and Growth Cycles z z1 z2 z3 z4 z Source: Author. 132 | JEAN-CLAUDE BERTHÉLEMY In figure 1 the boldfaced curve describes a situation in which there are two stable steady states, z2 and z4. In such a framework, unstable equilibria (z1 and z3) inevitably alternate with stable equilibria. To obtain multiple equilibria properties, the assump- tion needs to be made that the h(z) curve crosses the horizontal axis several times. Otherwise, the considered model would predict growth cycles rather than multiple equilibria. This kind of alternative outcome is depicted by the dashed curve in figure 1. What are the theoretical insights suggesting that such multiple equilibria may exist? The oldest ones are related to the analysis of the dynamics of the accumulation of factors of production. Nelson (1956) provides a formal framework in which cap- ital accumulation could be characterized by a cumulative process, due to the absence of savings capacity when incomes are very low. As a consequence of this assumption, at low levels of capital stock, savings and investment are not large enough to cover capital depletion and demographic growth, so that the growth rate of the capital/ labor ratio of the economy declines when its initial level is below a certain threshold and increases immediately above this threshold. This argument is germane to the idea that there is an incompressible minimum level of consumption. In countries approaching this extreme absolute poverty level, the demographic growth rate may also be affected. This argument, initially intro- duced by Leibenstein (1954), is also discussed by Solow (1956), to suggest the pos- sible existence of several steady state equilibria in his growth model. Another possibility of multiple equilibria associated with the process of capital accu- mulation may be related to the nonconvexity of the production function linking out- put to capital. More precisely, in such an analytical framework, one could observe increasing returns to capital at low levels of the capital stock and therefore a positive relation between the output to capital ratio--and hence the growth rate of the capital stock, assuming a constant marginal savings rate--and the capital stock. This assump- tion, which is similar to the increasing returns to scale assumption made by Rosenstein- Rodan (1943), was also recognized by Solow (1956) as possibly creating multiple equi- libria in his growth model. More recently, the possibility of multiple equilibria has been reconsidered in the course of discussions on the role of human capital, particularly education, in the development process. The theoretical model of Azariadis and Drazen (1990) clearly shows that a low level of educational development could lock an economy into a sit- uation of underdevelopment. The paucity of human capital resources initially avail- able considerably reduces the effectiveness of the education system and the return on education, obstructing the process of human capital accumulation, because the pri- vate return on human capital falls so low that parents do not invest in the education of their children. In this analysis, the education sector has a property similar to that attributed to the research and development (R&D) sector in standard endogenous growth models, namely, a dynamic externality. When the stock of knowledge avail- able within the population is insufficient, the gains from this externality cannot mate- rialize. As a result, growth is hampered, unless the state implements a strongly proac- tive education policy. Following this argument, one could well observe dynamically stable situations in which education is poorly developed and investing in education is not profitable for individuals. CONVERGENCE AND DEVELOPMENT TRAPS | 133 In developed economies, economic growth may be associated to a large extent with productivity gains, related in particular to the results of R&D activity and not only with factor accumulation. This may create nonstandard growth patterns given the dynamic externalities characterizing the R&D sector. In a developing country, some of these productivity gains may be imported from more advanced economies. But structural changes in the domestic economy may lead to productivity gains as well, with nontrivial consequences for the dynamics of the economy. Berthélemy and Varoudakis (1996) explore this line of argument regarding the financial deepening process. In a poor country, the initially weak state of the financial system and the low income level of the population may persist and reinforce each other as a result of a cumulative process: low incomes imply that the amount of savings to be intermedi- ated is small, which leads to high unit costs and weak competition in the financial sector. The result is a sluggish and inefficient capital accumulation process--owing to both the insufficient size of the financial sector and its imperfect competition--that prevents economic growth and locks the economy into a low-equilibrium state. Another structural change in the economy that may lead to multiple equilibria is related to the diversification process. Several researchers have suggested that the level of diversification of an economy could have a positive effect on the growth process. In a contribution to the analysis of the role of diversification in economic growth in the Republic of Korea and Taiwan (China), Feenstra and others (1999) suggest that out- put diversification could play a role in the growth process that is very similar to the role of input diversification considered in the seminal endogenous growth model of Romer (1990). Diversification may also reduce the vulnerability of an economy to external shocks, creating conditions favorable to economic growth, if only because the economy can then invest in higher risk, higher payoff projects. Several empirical papers suggest that this positive link between diversification and growth was observable in a number of economies in addition to the Republic of Korea and Taiwan (China).1 In parallel, there are some reasons to believe that economic diversification is influ- enced by the level of development attained by the economy. Imbs and Wacziarg (2003) show empirically that the diversification of an economy could be related to its development level, measured by GDP per capita, through an inverted U-shaped rela- tion. Berthélemy (2005) suggests that one possible explanation of this diversification dynamic could be related to the development of intra-industry trade: economic diver- sification could not be profitable in a standard neoclassical model in which an econ- omy should specialize on the basis of its comparative advantage, but the development of intra-industry trade has shown that other sources of gain from trade can be found. The theory of intra-industry trade that underlines these new sources of gain from trade can also be used to explain economic diversification. Given that intra-industry trade is greatest when trade partners have similar levels of development, this may explain why Imbs and Wacziarg observed a decline of diversification at very high lev- els of development and hence an inverted U-shaped relation between development and diversification. Whatever the theoretical reason for this inverted U-shaped rela- tion, it may lead to multiple equilibria when combined with the usual arguments that a more diversified economy has greater growth potential. The institutional framework may also be characterized by dynamic cumulative processes, which may draw an economy toward an underdevelopment trap under 134 | JEAN-CLAUDE BERTHÉLEMY some circumstances. Collier (2004) discusses this possibility in some detail. Accord- ing to his analysis, two phenomena--corruption and civil strife--are particularly rel- evant to explaining observed political and economic cumulative crisis situations in developing countries, particularly in Sub-Saharan Africa. Tirole (1996) discusses vicious circles of corruption in a game theory framework. Where corruption is high, individuals have no incentive to invest in reputation and therefore remain corrupt. In contrast, in a low corruption equilibrium, a bad reputa- tion has a high cost, and being corrupt is not profitable. Krueger (1993) discusses a related phenomenon of pervasive rent-seeking, in the framework of political economy models. Political economy models can also be used, as suggested by Collier (2004), to explain situations in which civil strife and poverty reinforce each other, leading to the possible existence of multiple equilibria. A large number of cumulative processes may be at work in which the growth rate of a state variable characterizing an economy decreases with its initial level when this level is below a threshold and increases, at least for a while, when it is higher. Each of these mechanisms can lead to the existence of multiple equilibria or to growth cycles. These mechanisms are mutually interdependent and should be combined to describe potentially "multiple multiple equilibria" or complex growth cycle paths. The Dynamics of the Jump Out of an Underdevelopment Trap Are examples of countries jumping from one equilibrium to another observed in the real world? To examine this question, consider a highly stylized analytical frame- work, in which one or several cumulative processes are possibly at work. This "mul- tiple multiple equilibria" structure cannot be fully described in a single-dimension framework. In theoretical terms, the economy can be described by a vector Z of state variables, which may be conceived, for instance, as several factors of production. The dynamics of this system are described by a differential equation system: . Z = H(Z) (2) To simplify the framework, exogenous technical progress is ignored, so that the H function is time invariant. Under the multiple multiple equilibria assumption, the condition H(Z)=0, which describes steady states, has n solutions, Z1 to Zn. Assuming that GDP per capita is a function of Z: y = f(Z) (3) it follows immediately that there are at least n solutions to the equation y. = 0. (4) This multiplicity of solutions can be described in a graph similar to that of figure 1, but in which the z state variable is replaced by y. This graph hides part of the complex- ity of the model. A combination of equations (1) and (2) leads to: y& = fziZ&i = fziHi(Z) (5) i i CONVERGENCE AND DEVELOPMENT TRAPS | 135 Usually, this cannot be collapsed into an equation like: y. = g(y) (6) because the growth rate of y does not depend only on y but on the whole structure of Z. This is a standard aggregation problem. To assume a growth equation such as equation (6), it is necessary to ignore changes in structure of the Z vector, which is a strong assumption. For a given country, assuming that there are no exogenous shocks, only a portion of figure 1 (after replacing z by y) will be observed, describing a convergence path toward a stable equilibrium. For instance, if the initial income per capita is between y1 and y2, it will converge toward y2; if it is between y3 and y4, it will converge toward y4. Only exogenous shocks can lead to a durable shift from an initial convergence path toward a given stable equilibrium to a convergence path toward a different sta- ble equilibrium (in figure 1 from convergence toward y2 to convergence toward y4). A permanent shock, defined as an irreversible change in the curve depicted in figure 1, would obviously lead to a change in the long-term equilibrium. But the relevant question here is whether after a temporary shock a country could move from one long-term equilibrium to another. Such exogenous shocks can be defined, in the framework depicted by figure 1, as a combination of two kinds of shocks: a shock equivalent to a temporary transfer, modifying the initial income, and a temporary productivity shock, modifying its growth rate. A temporary positive transfer of income would have the same effect as a tempo- rary leftward translation of the curve depicting g(y), with a reverse shift of equal magnitude to the right when the transfer is ended. Assume that at the time of the shock the economy is initially in the neighborhood of its steady state, at a point such as a0 in figure 2. Then, if the transfert is such as: t y3 ­ y2, (7) this shock will have an adverse impact on the growth performance of the economy, which will initially have a negative growth rate, as shown in figure 2, where the econ- omy moves to point a1. This initial negative growth rate is due to the definition of y2 as a stable equilibrium, implying that immediately to the right of y2 the growth rate of the economy is negative. Later the growth rate of the economy increases, while staying negative for a while. If the transfer is discontinued when the economy reaches a2, the growth rate again becomes positive. In the end the economy converges toward y2. In summary, although it temporarily increases the available income (defined by y +t), a positive transfer initially has a negative impact on growth performance; after the end of this temporary shock, the income of the economy converges again toward its initial long-term equilibrium y2. The shock leads only to a circular movement of the economy around its initial equilibrium. However, if the transfer is sufficiently large, that is, if t > y3 ­ y2, (8) the growth rate will initially remain positive, and even possibly increase, as illustrated by the shift from a0 to a1, and then to a2, in figure 3. However, if the transfer is only 136 | JEAN-CLAUDE BERTHÉLEMY FIGURE 2. Effect of Small Temporary Transfer on Growth y a3 a0 y1a2 y2 y3 y4 y a1 Source: Author. FIGURE 3. Effect of Large Temporary Transfer on Growth y a1 a2 a0 y1 y2 y3 y4 y a3 Source: Author. temporary, after a while it will be reversed, as depicted in figure 3. There is still a positive probability that the income of the economy converges toward y4. More pre- cisely, a necessary condition to obtain eventually a convergence toward y4 is: y4 ­ y3 > t > y3 ­ y2. (9) To obtain such a result, it is necessary to assume that the periodicity of the cycle characterizing the function g(y) increases with y. If this periodicity is nearly constant, the previous condition cannot be met and the economy converges again toward y2, as depicted in figure 3. To summarize, only a large transfer could possibly drive the economy from a path of convergence toward a low equilibrium to convergence toward a higher equilib- rium. But, even if the transfer is large enough to initially prevent the growth rate from becoming negative, there is no guarantee that after termination of the transfer the economy will continue converging toward the higher equilibrium. CONVERGENCE AND DEVELOPMENT TRAPS | 137 FIGURE 4. Effect of Temporary Productivity Shock on Growth y& a2 a1 a0 a3 y1 y2 y3 y4 y Source: Author. A temporary productivity shock would have potentially quite different conse- quences on the economy, as depicted in figure 4. This kind of shock could be an increase in total factor productivity or the result of a growth-enhancing change in the structure of factors. If a positive productivity shock temporarily lifts the growth rate of the economy to high enough levels, so that the U-shaped portion of the g(y) curve remains above the horizontal axis, the economy will converge toward the higher equilibrium depicted by y4 in figure 4. To obtain such a result, a necessary and suf- ficient condition is that when the productivity shock ends, the income of the econ- omy is above y3. This discussion leads to three useful conclusions. First, an attempt to lift an econ- omy out of a low equilibrium trap through a transfer policy is doomed to failure if the transfer is small, even if the transfer is made for a very long period. This seems to justify "big push" approaches, which call for large transfers. However, the prob- ability that such a policy succeeds does not depend only on the size of the initial transfer but also on conditions on the shape of g(y), which are not trivial. To be fair to the "big push" proposals, a transfer can change the structure of factors Z, given that the factor accumulation that it permits is usually not spread evenly over all fac- tors. This means that the g(y) curve may be also shifted upward thanks to the trans- fer policy, if, for instance, it is designed to promote the accumulation of factors that are particularly productive or scarce. But this argument points to the absolute neces- sity of avoiding focusing discussions on the amount of transfer that is requested. Second, an attempt to lift an economy out of its low equilibrium through reforms leading to a positive productivity shock, whether through a change in the factor mix or through an increase in total factor productivity, has a better chance of succeeding. This suggests that discussing how such a structural change can be obtained should form the core of discussions on Marshall Plan­type proposals. Third, jumping from a low equilibrium to a higher equilibrium implies a nonmonot- onic evolution of the growth rate of the economy, with peaks and troughs before the economy stabilizes on its new convergence path (see figure 4). In other words, the dynam- ics of the economy would be characterized by growth cycles featuring multiple peaks. 138 | JEAN-CLAUDE BERTHÉLEMY Stylized Facts on Multiple Growth Peaks Most of the empirical literature on convergence clubs has concentrated on cross- country growth comparisons aimed at identifying convergence clubs. This literature is useful, but more is needed to draw policy-relevant conclusions. Is there evidence of a jump of one or several economies from a low-level equilibrium to a higher level equilibrium? In the absence of such empirical evidence, advocating a Marshall Plan strategy would lack empirical foundations. As shown in the previous section, if there are such phenomena, one should observe rather unusual dynamics in the concerned economies. As illustrated in figure 4, their growth patterns should be characterized by several successive growth peaks, not by a single growth acceleration episode. The observed growth dynamics of a number of developing countries are analyzed here, using the annual time series built by Maddison (2003), some of which have been updated in the database recently released by the Groningen Growth and Devel- opment Centre and the Conference Board (henceforth GGDCCB). This dataset pro- vides series from 1950 to 2001 (to 2002 or 2003 for the GGDCCB update); it is thus more complete than the more often used Heston, Summers, and Aten (2002) series, which begin only later (in the 1950s or in 1960 in many instances) and end at best in 2000. To observe jumps from one equilibrium to another, one needs to compare growth performance of countries that were initially at similar levels of development but later diverged. Using as long a time series as possible is an advantage. Another reason to use Maddison's data is that these data are better documented. They are based on a critical comparison of all possible data sources, including Heston, Sum- mers, and Aten when no better source is available. The data set used here includes 99 developing countries with complete purchasing power parity GDP per capita series available for the whole period 1950­2001 or beyond: 49 in Africa, 29 in Asia and the Middle East, and 21 in Latin America. East- ern Europe and Central Asia transition economies are excluded, because the changes there have occurred only recently. Cuba and a few African countries for which the data quality was very poor are also excluded.2 To eliminate the short-run cyclical component of these time series, the data were transformed with an appropriate Hodrick-Prescott filter.3 The Hodrick-Prescott approach (1997) proposes a decomposition of a series yt into two components: a "growth" component, gt, and a cyclical component, ct. The growth component is defined so that it varies "smoothly" over time, the measure of its smoothness being the sum of the squares of its second differences. In this frame- work the growth component series is computed so as to minimize T T t + [(t - t )-(t 2 -1 -1 - t )]2, -2 t=1 t=1 where T is the number of observations and l is called a "smoothing factor." The larger l, the smoother the solution series gt. When l approaches infinity, gt approaches the least square fit of a linear time trend model. This method of decomposition is usu- ally applied to quarterly series rather than annual series. Applying the original CONVERGENCE AND DEVELOPMENT TRAPS | 139 Hodrick-Prescott filter to yearly data would be misleading, because it would smooth the series much too heavily. It would conceal a number of medium-term acceleration episodes, because such accelerations would be treated as short-term cyclical move- ments, as Baxter and King (1999) and Ravn and Uhlig (2002) have noted. Both papers show that the smoothing parameter used in the Hodrick-Prescott filter must be adapted for use in annual series. Following their approach, a value of 10 for the smoothing parameter was used here (Hodrick and Prescott recommend a value of 1,600 for quarterly series). Observing the relation between the transformed GDP per capita series and their growth rates provides the empirical basis for drawing the equivalent of the g(y) curves and possibly for uncovering multiple growth peak patterns associated with equilibrium jumps. Given that the g(y) function is presumably only an imperfect aggregation function, a different pattern for each country is used here, instead of using a panel data approach. In the case of a single growth peak, which may characterize the standard dynamic path of convergence to a stable steady state equilibrium, the observed g(y) curve has an inverted U shape. This shape is observed in a number of cases, related in some instances with significant economic progress since 1950. The example of Paraguay is provided in figure 5. In several other instances, the observed g(y) curve depicts a cyclical growth pat- tern instead of a single growth acceleration. The example of Malaysia is provided in figure 5. With a few exceptions, situations in which the economy has not achieved progress since 1950 are associated with the absence of any significant growth peak. In most FIGURE 5. Growth Patterns in Malaysia, Paraguay, and Senegal 0.08 Senegal Paraguay 0.06 Malaysia 0.04 capita per 0.02 GDP of owth 0 Gr 1,000 10,000 ­0.02 ­0.04 GDP per capita (logarithmic scale) Source: Author's computation based on Maddison (2003) and Groningen Growth and Development Centre and Conference Board databases. 140 | JEAN-CLAUDE BERTHÉLEMY such cases, the economy is cyclically stagnating around its initial level, as in the case of Senegal, reported in figure 5. Using a parametric method to detect single and multiple growth peaks would pre- sumably be inappropriate, given the peculiar shape of a g(y) curve featuring multiple peaks. A simple nonparametric method is proposed here, based principally on the observation of the curves drawn for the different countries. To be more precise, a growth peak is defined as an episode in which growth is accelerating for at least seven consecutive years, is positive throughout the same period, and peaks at a pace of at least 3.5 percentage points. Given the elimination of the short-run cyclical component of the time series, they display very few short-term fluctuations, allowing many cases of continuous growth acceleration phases for long periods to be identified. There are, however, a few instances in which such fluctuations have not been eliminated, leading to relatively frequent ups and downs instead of smooth acceleration and deceleration patterns. Such inconvenience is minor when as a result the count of peaks drops from one to zero, given that the aim here is to detect multiple growth peak patterns. This is a draw- back, however, in the cases of Hong Kong (China), Lesotho, the Seychelles, and Tai- wan (China), where the growth graphs suggest multiple growth peaks that do not pass the test because of their relatively short duration (figure 6). In the case of Hong Kong (China) and Taiwan (China), there are so many peaks in the period (despite the FIGURE 6. Growth Patterns in Hong Kong (China), Lesotho, the Seychelles, and Taiwan (China) 0.10 Hong Kong (China) Taiwan (China) Lesotho 0.08 Seychelles 0.06 capita per 0.04 GDP of owth 0.02 Gr 0 100 1,000 10,000 100,000 ­0.02 GDP per capita (logarithmic scale) Source: Author's computation based on Maddison (2003) and Groningen Growth and Development Centre and Conference Board databases. CONVERGENCE AND DEVELOPMENT TRAPS | 141 Hodrick-Prescott smoothing) that each of them could be only of a rather short duration. Ignoring the intermediate small fluctuations between the start and the end of the growth peak, the data suggest that multiple peaks existed in these four economies. A synthesis of the analysis of the growth dynamics of countries under review is pro- vided in tables 1­3. The results show that about half the countries (54 out of 99) expe- rienced at least one growth peak since 1950. This observation is globally consistent with the findings of Hausmann, Pritchett, and Rodrik (2004), who detect many growth acceleration episodes in the Heston and Summers time series, although their approach is somewhat different. Hausmann, Pritchett, and Rodrik use raw GDP per capita series, not series transformed with the Hodrick-Prescott filter. They use a definition of acceleration based on the average acceleration over a given period of time (of seven years), which means that the smoothing procedure is part and parcel of the method of detection of accelerations. Searching for steady accelerations for a relatively long period of time--as I do-- would lead to no result with raw data, because in most instances short-term cyclical fluctuations would prevent the observation of a steady acceleration. I prefer to first smooth out the series, with a procedure that is independent of the search of acceler- ation episodes, before identifying such episodes.4 Hausmann, Pritchett, and Rodrik find accelerations in some 51 developing countries, which is close to my own count. The difference stems partly from the fact that they do not consider small countries (such as Lesotho, the Seychelles, or Swaziland). They find more accelerations in Latin America and fewer in Asia than I do, but most of the countries in which they observe accelerations appear on my list as well. With respect to twin peaks, Hausmann, Pritchett, and Rodrik observe 13 cases of multiple growth accelerations, a majority of which are also twin-peak cases in my own exercise (the Dominican Republic, Indonesia, the Republic of Korea, Malaysia, Mauritius, Pakistan, and Thailand). In contrast to Hausmann, Pritchett, and Rodrik, who look only for growth acceleration and are not interested in clearly separating single growth acceleration cases from multiple growth accelerations cases, I am inter- ested in identifying only multiple growth accelerations cases. In my analytical frame- work, these are the only interesting cases from the point of view of underdevelop- ment trap analysis.5 Many countries that have experienced single growth peaks experienced only mod- est average growth performance since 1950. Seventy percent of these countries (26 out of 37) grew less rapidly than the United States (2.1 percent a year), implying that their backwardness with respect to the developed economies deepened over time. Multiple peaks are detected here for only 17 countries (15 when assuming a stricter selection parameter). As a result of their prolonged growth cycle patterns, these countries usually performed much better than others. All of them grew faster than the United States. Occurrences of high growth performance over the past five decades may be asso- ciated with the observation of multiple growth peaks, insofar as they are much less frequent in the single or no growth peak cases. However, a few countries grew faster 142 | JEAN-CLAUDE BERTHÉLEMY TABLE 1. Growth Patterns in Africa Per capita GDP (1990 purchasing power parity dollars) Number of growth peaks Average annual Based on Based on growth since 1950 3.5 percent 4.0 percent Country (percent) 1950s 1990s growth filter growth filter Botswana 5.0 400 3,500 2 2 Lesotho 3.1 400 1,400 2 2 Mauritius 3.0 2,600 8,700 2 2 Tunisia 2.8 1,200 3,800 2 1 Cape Verde 2.7 500 1,300 0 0 Swaziland 2.5 800 2,500 1 1 Egypt, Arab Rep. of 2.3 900 2,500 1 1 Seychelles 2.2 2,100 5,800 2 2 Mauritania 1.6 500 900 1 1 Guinea-Bissau 1.6 400 800 0 0 Algeria 1.5 1,500 2,700 0 0 Malawi 1.4 400 600 0 0 Guinea 1.3 300 500 0 0 Morocco 1.3 1,400 2,600 0 0 Mali 1.2 500 800 1 1 Burkina Faso 1.2 500 800 0 0 Cameroon 1.1 700 1,100 1 1 Congo, Rep. of 1.1 1,400 2,300 1 1 Namibia 1.1 2,300 3,600 1 1 Zimbabwe 1.0 800 1,300 1 1 South Africa 1.0 2,800 3,900 0 0 Burundi 0.9 400 700 1 1 Rwanda 0.9 600 700 1 1 Ethiopia 0.9 400 500 0 0 Kenya 0.9 700 1,000 0 0 Gambia, The 0.8 700 900 0 0 Mozambique 0.7 1,200 1,200 1 1 Nigeria 0.7 800 1,100 0 0 Tanzania 0.7 500 500 0 0 Côte d'Ivoire 0.4 1,100 1,300 1 1 Gabon 0.4 3,600 4,800 1 1 Benin 0.4 1,000 1,200 0 0 Ghana 0.4 1,200 1,200 0 0 Sudan 0.4 900 800 0 0 Uganda 0.3 700 700 1 1 Senegal 0.3 1,400 1,300 0 0 Zambia 0.1 800 700 0 0 Comoros 0.0 600 600 1 1 Togo 0.0 600 700 1 1 Chad -0.1 500 400 0 0 Central African -0.4 900 600 0 0 Republic (Continues on the next page) CONVERGENCE AND DEVELOPMENT TRAPS | 143 TABLE 1. continued Per capita GDP (1990 purchasing power parity dollars) Number of growth peaks Average annual Based on Based on growth since 1950 3.5 percent 4.0 percent Country (percent) 1950s 1990s growth filter growth filter Liberia -0.4 1,100 1,100 0 0 Somalia -0.4 1,200 900 0 0 Angola -0.5 1,100 700 1 1 Madagascar -0.5 1,000 700 0 0 Djibouti -0.6 1,600 1,200 0 0 Niger -0.9 900 500 0 0 Sierra Leone -1.0 700 700 0 0 Congo, Dem. Rep. of -1.7 700 300 0 0 Source: Author's computation based on the Maddison (2003) and Groningen Growth and Development Centre and the Conference Board databases. TABLE 2. Growth Patterns in Asia and the Middle East Per capita GDP (1990 purchasing power parity dollars) Number of growth peaks Average annual Based on Based on growth since 1950 3.5 percent 4.0 percent Country (percent) 1950s 1990s growth filter growth filter Korea, Rep. of 5.8 1,000 11,100 2 2 Taiwan (China) 5.6 1,200 13,000 2 2 Oman 4.7 800 7,000 1 1 Hong Kong (China) 4.5 2,600 20,000 2 2 Singapore 4.4 2,300 18,100 2 2 China 4.3 600 2,600 1 1 Thailand 4.0 900 5,900 2 2 Israel 3.5 3,500 14,700 0 0 Malaysia 3.2 1,500 6,800 2 2 Myanmar 2.8 500 1,000 2 2 Indonesia 2.7 1,000 3,100 2 2 Saudi Arabia 2.5 2,800 8,900 1 1 Syrian Arab 2.3 3,000 7,200 1 1 Republic Yemen, Rep. of 2.3 900 2,500 1 1 Pakistan 2.2 600 1,800 2 1 India 2.2 700 1,500 1 0 Vietnam 2.2 700 1,400 1 1 Sri Lanka 2.1 1,300 2,900 2 2 Iran, Islamic Rep. of 2.0 1,700 4,100 1 1 (Continues on the next page) TABLE 2. continued Per capita GDP (1990 purchasing power parity dollars) Number of growth peaks Average annual Based on Based on growth since 1950 3.5 percent 4.0 percent Country (percent) 1950s 1990s growth filter growth filter Bahrain 1.8 2,500 4,500 1 1 Jordan 1.7 1,900 3,900 1 1 Philippines 1.6 1,300 2,200 0 0 Cambodia 1.5 600 1,000 0 0 Nepal 1.4 500 900 0 0 Lao PDR 1.3 600 1,000 0 0 Bangladesh 1.0 500 700 0 0 Lebanon 0.7 2,500 3,000 0 0 Iraq -0.1 2,100 1,200 0 0 Afghanistan -0.7 700 500 0 0 Source: Author's computation based on the Maddison (2003) and Groningen Growth and Development Centre and the Conference Board databases. TABLE 3. Growth Patterns in Latin America Per capita GDP (1990 purchasing power parity dollars) Number of growth peaks Average annual Based on Based on growth since 1950 3.5 percent 4.0 percent Country (percent) 1950s 1990s growth filter growth filter Trinidad and Tobago 2.7 4,500 10,600 1 1 Dominican Republic 2.5 1,200 2,800 2 2 Brazil 2.3 1,900 5,100 2 2 Costa Rica 2.2 2,300 5,200 1 1 Panama 2.1 2,100 5,300 1 1 Mexico 2.1 2,700 6,400 0 0 Jamaica 2.0 1,900 3,600 1 1 Chile 1.9 4,100 8,300 1 1 Colombia 1.7 2,300 5,200 1 1 Ecuador 1.5 2,100 4,100 1 1 Paraguay 1.2 1,600 3,300 1 1 El Salvador 1.2 1,600 2,400 0 0 Guatemala 0.9 2,100 3,200 0 0 Peru 0.9 2,600 3,300 0 0 Uruguay 0.9 5,100 7,400 0 0 Honduras 0.8 1,300 1,900 0 0 Argentina 0.7 5,200 8,000 1 1 Bolivia 0.6 1,800 2,400 1 1 Venezuela, R. B. de 0.1 8,700 8,800 0 0 Nicaragua -0.1 1,900 1,400 0 0 Haiti -0.6 1,100 900 0 0 Source: Author's computation based on the Maddison (2003) and Groningen Growth and Development Centre and the Conference Board databases. CONVERGENCE AND DEVELOPMENT TRAPS | 145 than the United States but without multiple growth peaks. These cases deserve care- ful analysis. Some of these countries successfully managed their oil or mineral booms. These include Oman and Saudi Arabia (and to some extent the Arab Republic of Egypt, the Syrian Arab Republic, and the Republic of Yemen) in the Middle East; Swaziland in Africa; and Trinidad and Tobago (and to some extent Mexico) in Latin America and the Caribbean. These countries avoided a reversal of the growth process originating in the mineral/oil boom by cautiously managing their windfall gains and in some cases diversifying their economies.6 In China, and to some extent India and Vietnam, growth performance was enhanced by transition reforms. The observation period is not long enough to observe a proper growth cycle pattern, however. Cape Verde, which implemented a central planning economic policy after independence in 1975 until the end of the 1980s, can be also considered a transition case. Israel performed rather well on average, but its performance has been too much affected, positively and negatively, by the consequences of its conflicts with neighbor- ing countries to be interpretable in simple economic terms. Costa Rica has grown slightly faster than the United States and belongs to the sin- gle growth peak category, but it is a borderline case: it experienced a second growth peak in the 1990s, culminating at only 3.3 percent a year, instead of the 3.5 percent threshold used here. Finally, an objection to this interpretation of multiple growth peak cases could be that for some countries, multiple growth peaks have been observed because these countries have been subject to multiple positive shocks. This may be true in some cases. But the main objective of this test is to eliminate some accelerating countries from the list of countries that may have experienced a jump from a low equilibrium to a higher equilibrium. In this sense, this test is rather powerful, because it eliminates many candidates (37 out of 54). The fact that I observe only a limited number of multiple growth peak cases suggests that such countries have indeed experienced peculiar dynamics. The multiplicity of shocks is common to all countries, although for most countries it has not led to multiple growth peaks but merely to alternating phases of progress and decline, leading in the long run to rather modest economic performance. The cyclical component of per capita GDP series, which results from the Hodrick- Prescott filtering procedure, may be analyzed to measure the size of shocks (positive or negative) that have affected developing countries. A simple approach is to com- pute the standard deviation of this cyclical component. The standard deviation is smaller for economies that have experienced twin growth accelerations (2.7 percent) than for other economies (3.5 percent), and the difference is statistically significant at the 9 percent level. This suggests that countries that have emerged have experi- enced smaller shocks than other countries. It would therefore be dubious to explain their emergence by external shocks. It may be that only positive shocks, rather than all kinds of shocks, explain jumps out of underdevelopment traps. If shocks are not normally distributed, comparing standard deviations does not provide the right information on the size of such 146 | JEAN-CLAUDE BERTHÉLEMY positive shocks. To double-check my results, I compared the size of the maximum values of the cyclical components for the two groups of countries. Again, countries with twin growth accelerations have smaller shocks than other countries (with a maximum value of the cyclical component equal, on average, to 7.4 percent, against 9.6 percent for the other group of countries), and the difference is significant at the 12 percent level. If external shocks cannot be considered as generally responsible for jumps out of the underdevelopment trap, it is necessary to examine the economic factors underly- ing the dynamics of multiple growth peak countries compared with other countries with similar initial levels of development. Only an analysis of factors possibly affect- ing growth potentials in these two groups of countries can provide more substantive arguments to strengthen the interpretation of the occurrences of multiple growth peaks as nonstochastic jumps out of an underdevelopment trap. The next section is devoted to this exercise. Possible Explanations of Successful Take-Offs Why did some countries with low levels of development in the 1950s later achieve very impressive progress while others experienced equally impressive economic fail- ure? To avoid comparing countries with different initial development levels, only countries in which per capita GDP was below US$1,500 in 1990 purchasing power parity terms in the 1950s are considered here. Out of the 17 multiple growth peak cases, this choice eliminates Brazil, Hong Kong (China), Malaysia, Mauritius, the Seychelles, and Singapore, leaving 11 success stories. The role of Hong Kong (China) and Singapore as business centers created specific favorable conditions that made them special cases. Brazil and Malaysia started their development before World War II, although Malaysia was severely hit in the 1930s by the collapse of the rubber market and Brazil was hit by the Great Depression and worldwide protectionist policies. In 1950 these economies were cer- tainly in a better position to develop than poor economies in Africa or Asia. Mauri- tius and the Seychelles, although less developed than these countries, enjoyed rela- tively favorable early natural conditions. All in all, it is safe to exclude these economies from a comparison of success stories with poor and poorly performing countries. The list of countries with initial per capita GDP below US$1,500 that did not later experience multiple growth peaks is a relatively long one. It includes 40 African coun- tries, 9 Asian countries, 2 Middle Eastern countries, and 2 Latin American countries. Even after eliminating countries for which good data are not available, this leaves plenty of countries to compare with the 11 selected multiple growth peak countries.7 Given that the initial growth peaks in multiple growth peak countries were gener- ally observed in the 1960s or at the latest in the early 1970s, the data examined here characterize the growth potential of the two groups of countries at the beginning of the 1960s--that is, immediately before the take-off of emerging countries. The analy- sis focuses on measurable economic variables previously identified as factors poten- tially responsible for the occurrence of a low equilibrium trap. CONVERGENCE AND DEVELOPMENT TRAPS | 147 Institutional Factors In the 1960s there were institutional differences among poor countries, but the insti- tutional framework of countries that took off in Asia was not considered at that time as particularly favorable to their development. The best-known compendium of evi- dence of institutional drawbacks in Asia in the 1960s is Myrdal's 1968 trilogy, Asian Drama. His thesis was that the institutional characteristics of these countries consti- tuted severe obstacles to their development. He wrote the trilogy on the basis of his 10 years in India, but he provided many empirical arguments in support of a gener- alization of his conclusions to Indonesia, Malaysia, Myanmar, Pakistan, the Philip- pines, Sri Lanka, and Thailand. He devoted a full chapter to corruption in Asia and its adverse consequences: "The significance of the corruption in Asia is highlighted by the fact that wherever a political regime has crumbled . . . a major and often deci- sive cause has been the prevalence of official misconduct among politicians and administrators" (Myrdal, p. 937). Myrdal's predictions concerning Asia were mistaken in a number of ways, but his work provides clear evidence that institutional issues such as corruption and non- democratic and nontransparent political governance were not confined to Africa. Quite the contrary, many African countries enjoyed a relatively stable political environment in the 1950s and the early 1960s and benefited from several policy ini- tiatives aimed at promoting their economic development. Some evidence of the devel- opmental attitude in African economies in the decade before independence can be found in the Economic Survey of Africa Since 1950, conducted by the United Nations in 1959. In former French colonies, France's 1946 Constitution and its new political orientation toward colonial territories gave an impetus to development, leading, for instance, to the creation of the Fonds d'Investissement pour le Développe- ment Economique et Social (FIDES), which played a major role by financing economic infrastructure building (Berthélemy 1980). Côte d'Ivoire was on a rapid growth track in the 1950s, particularly following the opening of the port of Abidjan. A similar trend was observable in British colonies after the Colonial Development and Welfare Act was passed, in 1945. Although more historical material would be necessary to fully support this view, looking for quantitative factors rather than for purely institutional factors to explain the low equilibrium traps in which many poor countries, particu- larly in Africa, remained locked in the 1950s and the 1960s, seems warranted. Investment Data on investment suggest that the two groups of countries were quite similar in the 1960s (table 4). Countries in the first group, with multiple growth peaks, did not have impressive investment ratios until the end of the 1960s; until then these ratios were comparable to those in other poor countries. The external financing of invest- ment, approximated here by the difference between gross fixed capital formation and gross domestic savings, was on average more favorable in the first group of countries than in the second. This difference is, however, due only to the Republic of Korea, which received significant external assistance from the United States until the early 1970s, and to Botswana, which had low savings but significant foreign assistance as well as investments in the mining industry from South Africa. No evidence suggests 148 | JEAN-CLAUDE BERTHÉLEMY TABLE 4. Investment and Savings Ratio, 1960-70, by Country Group (percent) Gross domestic Gross domestic fixed investment fixed investment/GDP minus gross domestic savings/GDP Group 1960 1965 1970 1960 1965 1970 Multiple growth 11.9 (1.2) 17.7 (6.4) 20.1 (6.7) 0.5 (7.1) 5.0 (10.9) 6.2 (7.3) peaks group Single or no growth 11.8 (4.6) 13.4 (4.7) 14.8 (5.9) -0.3 (1.1) -0.6 (2.7) -2.0 (15.1) peak group Source: Author's computation based on World Bank World Development Indicators database. Note: Standard deviations are shown in parentheses. Lesotho is excluded from the multiple growth peaks group because of its very high negative savings. TABLE 5. Demographic Indicators, 1960-70, by Country Group (percent) Age dependency ratio Population growth Group 1960 1965 1970 1960 1965 1970 Multiple growth 0.9 (0.1) 0.9 (0.1) 0.9 (0.1) 2.4 (0.6) 2.5 (0.4) 2.5 (0.5) peaks group Single or no growth 0.9 (0.1) 0.9 (0.1) 0.9 (0.1) 2.3 (2.3) 2.5 (0.6) 2.5 (0.5) peak group Source: Author's computation based on World Bank World Development Indicators database. Note: Standard deviations are shown in parentheses. that the high performance of the countries that took off in the 1960s was due ini- tially to a jump out of a low savings trap: although very high savings and investment rates occurred in these countries later on, this was not the case in the 1960s. Demographics Demographic data suggest very similar initial conditions in the two groups of countries (table 5). In both groups the annual population growth rate was about 2.5 percent, and the age dependency ratio was about 88 percent. One cannot explain the difference in growth dynamics of the two countries by a demographic trap hypothesis. Education Education data reveal a more complex picture (table 6). In terms of education expen- diture as a percentage of GDP, the two groups of countries were similar in 1960, although on average countries in the first group spent a little more on education. However, education achievements are dramatically different. In 1960 the proportion of the adult population (people over the age of 15) having attended school was already twice as high in the first group of countries as in the second group, and the proportion of adults having completed primary school was three times as high. Part of this difference was inherited, as suggested by ratios computed on the population CONVERGENCE AND DEVELOPMENT TRAPS | 149 TABLE 6. Education Indicators, 1960-70, by Country Group (percent) Proportion of population Proportion of population Education over 15 with over 15 with complete expenditure/GDP primary education primary education Group 1960 1965 1970 1960 1960 Multiple growth 2.5 (1.0) 3.3 (1.4) 3.6 (1.4) 46.0 (22.6) 23.8 (16.8) peaks group Single or 2.2 (1.1) 2.9 (1.4) 3.2 (1.5) 25.0 (19.3) 9.4 (7.5) no growth peak group Source: Author's computation based on World Bank World Development Indicators database and Barro and Lee 1996 database. Note: Standard deviations are shown in parentheses. over 25--that is, people who were of primary school age in the early 1950s or earlier. However, in most countries in the first group, ambitious education policies were implemented only after World War II. This is the case, for instance, in the Republic of Korea. According to Lee, "In 1945, when the Korean peninsula was liberated from Japanese colonial rule, the country was left with widespread illiteracy and an ill- trained labor force" (1995, p. 15). The period from 1945 to 1970 witnessed a dra- matic expansion of education, thanks to public policies targeting the elimination of illiteracy. In 1960 only Pakistan and Tunisia in the first group had very low education performance, comparable to those of countries in the second group--but these are also the countries whose classification in the first group is the least plausible, as sug- gested in the previous section. In Tunisia the relatively low school attainment figures reported by Barro and Lee (1996) are underestimated, because they do not take into account Islamic schools. According to data reported by Morrisson and Talbi (1996), close to 10 percent of the active population in 1966 had attended an Islamic school, and 24 percent had completed the classical primary school curriculum. Moreover, the Tunisian government made a major investment in primary education in the 1960s, which shows up in school attainment figures only toward the end of the 1960s. Significant progress in literacy was made in the first group of countries, despite education budgets that were not significantly higher than in the second group. A plausible explanation for this phenomenon is the very different distributional char- acteristics of educational policies in emerging Asian economies and in African coun- tries (Berthélemy forthcoming): African countries on average implemented relatively inequitable education policies, spending significant resources on secondary and higher education, leaving few resources available for primary schools. Berthélemy's (forthcoming) findings can be complemented by an analysis of the ratio of the num- ber of people who attended secondary school to the number of people who com- pleted primary education. The usual expectation is that this ratio, the probability of attending secondary school conditional on primary school completion, should be higher in countries with higher average education. In fact, in 1960 the nonemerging countries had the same ratio as the emerging countries, close to 40 percent. Relative 150 | JEAN-CLAUDE BERTHÉLEMY to their meager achievements with respect to primary enrollment, the nonemerging countries performed rather well on secondary schooling. By comparison, UNESCO data for the early 1950s suggest that on average a majority of children in emerging countries were already able to attend primary school by 1951­3--performance that is much higher than that attained by nonemerging countries in the early 1960s-- while the ratio of the number of people who attended secondary school to the num- ber of people who completed primary education in these countries was only about 13 percent.8 Recent data suggest that the relative cost of secondary to primary edu- cation is twice as high in Africa--where most nonemerging countries are--as in Asia (UNESCO 2004). Moreover, most emerging countries in Asia have delegated a sig- nificant part of the secondary school system to the private sector (33­40 percent in the Republic of Korea, the Philippines, and Thailand), reducing the relative cost of secondary education to the government budget. All these observations help explain why relatively poor education achievements in Africa, compared with Asia, are con- sistent with comparable education expenditure. In implementing education policies that are biased against primary education, the newly independent African states extended the policies that had been applied by their colonial rulers. With the exception of territories in which missionaries took full responsibility for education (such as the Belgian colonies), education under the colo- nial regimes was reserved for a small elite, trained to become civil servants in the colonial administration or employees in trading companies. As a consequence of such education policies, many countries made little progress in the 1950s and the 1960s toward increasing literacy, despite significant public expenditure in the education system. This may have contributed to locking them into a low equilibrium trap, if only because without literacy it is extremely difficult to take advantage of technical progress produced elsewhere. Financial Development and Diversification Financial development and diversification may also have contributed to multiple equi- libria (table 7). Countries in the first category had greater financial depth (measured by M3/GDP) than nonemerging countries in the early 1960s, albeit with much varia- tion across countries. The average is distorted by Pakistan, which in 1960 already had an M3/GDP ratio of 0.44. The Dominican Republic, Indonesia, the Republic of Korea, Sri Lanka, and Thailand had financial depth ratios of 0.20 or less, compara- ble to the threshold identified by Berthélemy and Varoudakis (1996) for separating convergence clubs based on financial depth. Moreover, the low average in the second group of countries, particularly in 1960, was due to very low financial depth in sev- eral African countries, where the ratio was less than 0.05. At least seven countries in this group (the Republic of Congo, Egypt, Ghana, Honduras, India, Morocco, and the Philippines) had financial depth ratios comparable to countries in the first group but nevertheless remained poor. At this stage of development, financial depth may have played only a minor role; it may have played a more significant role at a later stage, as suggested by the sharp improvement in financial depth observed in the Republic of Korea and Thailand--and also, according to Dessus, Shea, and Shi (1995), in Taiwan (China)--in the course of the 1960s. CONVERGENCE AND DEVELOPMENT TRAPS | 151 TABLE 7. Financial Depth and Export Diversification Indicators,1960­70, by Country Group (percent) Share of manufactured exports M3/GDP on total merchandise exports Group 1960 1965 1970 1960 1965 1970 Multiple growth 24.2 (11.0) 25.8 (11.0) 28.1 (11.0) -- 19.5 (24.0) 23.1 (31.1) peaks group Single or no growth 10.6 (9.9) 14.3 (7.4) 17.5 (7.9) -- 9.9 (15.5) 11.0 (13.0) peak group Source: Author's computation based on World Bank World Development Indicators database. -- Not available. Note: Standard deviations are shown in parentheses. On diversification very detailed trade statistics would be necessary to produce an accurate indicator (see Berthélemy 2005). The measure considered here is the share of manufactures in total exports of merchandise. This is a very crude indicator, although it makes sense given that diversification is highly dependent on the devel- opment of a manufacturing industrial base. Even this simple indicator is unavailable in a large number of countries in 1960. Only very hypothetical conclusions can thus be made. On average, the first group of countries had a higher share of manufactures in total exports than the second. In this first group, the Republic of Korea is an exceptional case. Thanks to a successful export promotion policy initiated in 1961, manufactured exports rose sharply in the early 1960s, from 20 percent in 1962 to 59 percent in 1965. Undoubtedly, this export promotion policy played a significant role in Korea's success. Data on Taiwan (China) suggest a similar evolution, with manufactured goods accounting for 32 percent of exports in 1960 and 79 percent in 1970 (Dessus, Shea, and Shi 1995). Thailand, which took off almost at the same time, diversified much later; Indonesia, Myanmar, and Sri Lanka were equally nondiversified in the 1960s. There is therefore no clear-cut empirical evidence supporting a causal relation from the diversification impulse initiated by export promotion policies (as in the Republic of Korea) to the initial economic take-off. This does not mean that diversi- fication did not play a role at later stages: economies in the first group that have devel- oped the most--the Republic of Korea and Taiwan (China)--were the earliest diver- sifiers. As Feenstra and others (1999) suggest, this may have contributed to their economic performance. Regression Results To synthesize the foregoing discussion, a probit model is estimated in which the dependent variable is a dummy variable equal to 1 for the first group of countries and 0 for other countries (table 8). Unsurprisingly, a number of variables--savings and investment ratios, demographic data, education expenditure, and the ratio of manufactured exports--have no significant impact on the probability of belonging to 152 | JEAN-CLAUDE BERTHÉLEMY the multiple growth peaks group. In addition, the indicators considered earlier to characterize the size of short-term fluctuations do not significantly influence the probability of emergence. Conversely, the education outcome variable, defined here as the proportion of adults who had completed primary school in 1960, is highly sig- nificant, as reported in column 1 of table 8. Very similar results (not shown) are obtained when using the proportion of adults with some schooling. The M3/GDP ratio measured in 1960 (column 2) also yields significant results. It may be that successful countries were successful because of a neighboring effect: the fact that most of these countries are located in Asia is presumably not just by chance. Unsurprisingly, an Asian dummy is also highly significant, as shown in col- umn 3 of table 8. The schooling variable remains significant in column 4 when asso- ciated in the same model with the Asian dummy. This is not the case, however, of M3/GDP (column 5). An attempt to put all three variables in the model does not lead to significant results, due to the high number of missing observations, although the education variable scores best here. This simple econometric exercise confirms the previous analysis: ambitious pri- mary education policies played a role in the take-off of countries experiencing mul- tiple growth peaks. Financial depth could have played some role, but the evidence is not quite clear at the initial low level of development observed in the early 1960s in such countries. These conclusions are consistent with work by Berthélemy and Varoudakis (1996), who identify nested convergence clubs, defined first by a thresh- old based on education and second by a threshold based on financial development. It would be interesting to pursue the analysis one step farther and study, in a dynamic framework, the determinants of further growth accelerations of countries that started jumping out of their underdevelopment trap in the 1960s. However, TABLE 8. Probit Regression for the Occurrence of Multiple Growth Peaks Item (1) (2) (3) (4) (5) (6) Proportion of adult 0.051*** 0.042** 0.083 population having (2.789) (2.277) (1.586) completed primary school M3/GDP 0.072** 0.055 0.229 (2.353) (1.412) (1.381) Asian dummy 1.214*** 0.891* 1.0646*** -1.228 variable (2.968) (1.694) (2.327) (0.703) Intercept -1.378*** -2.021*** -1.372*** -1.573*** -2.459*** -5.681 (3.669) (3.153) (5.248) (3.853) (2.872) (1.614) Pseudo R2 0.228 0.248 0.155 0.292 0.450 0.618 Number of 37 29 63 37 29 20 observations Source: Author's computation based on World Bank World Development Indicators database and Barro and Lee 1996 database. Note: Pseudo-Student statistics are shown in parentheses. * Significant at the 10 percent level. ** Significant at the 5 percent level. *** Significant at the 1 percent level. CONVERGENCE AND DEVELOPMENT TRAPS | 153 given the small number of such countries, the econometric results would be doomed to failure for lack of significant counterfactual evidence. A different approach, com- bining case studies and comparative econometrics, is probably needed. Conclusion The notion of convergence clubs based on multiple equilibria is intellectually appeal- ing and provides a plausible interpretation of the persistence of average poverty in a number of developing countries while the rest of the world economy has achieved dramatic economic progress. It may also constitute a good starting point to legiti- mate "big push" plans. However, the most relevant question for policy is not to observe the existence of convergence clubs but to obtain evidence of actual jumps of some countries out of their initial underdevelopment trap. Such jumps should lead to growth acceleration cycles--that is, a sequence of multiple growth peaks. These dynamics are observable in several economies that are considered success stories. Lessons learned from this analysis of the divergence among developing countries over the past 50 years suggest a very cautious approach to "big push" plans for least developed countries. Only policies able to promote growth through different kinds of productivity-enhancing mechanisms, rather than mere financial assistance policies, can make a difference. This conclusion is consistent with findings by Berthélemy and Söderling (2001), who show that the rare successful take-offs in Sub-Saharan Africa have been due to increases in total factor productivity rather than to accelerated investments, which were often associated with short-lived commodity booms. Education policies, particularly policies aimed at improving the literacy of the pop- ulation, may have played an important role in allowing countries to jump out of their underdevelopment traps. This conclusion is very close to the one offered by Myrdal (1968) in the third volume of his Asian Drama trilogy, where he forcefully advocated policies aimed at improving "population quality" to lift Asia out of its poverty trap. This important role of education may appear at odds with recent research show- ing that schooling has low macroeconomic returns, particularly when tested on inter- national panel data sets (see, for example, Pritchett 2001).9 One possible explanation is linked to the multiple equilibria analysis. If dynamic externalities in the education sector create multiple equilibria, the relationship between human capital achieve- ments and growth performance will display a highly nonlinear stepwise pattern. Under such circumstances, linear estimations cannot be robust, because the underly- ing equation that is tested is strongly misspecified. Linear estimations can perform relatively well, as in the early papers by Barro (1991) and Mankiw, Romer, and Weil (1992), or badly, as in Pritchett (2001). Moreover, when positive results are obtained, they are doomed to collapse when tested on panel data sets (see, for exam- ple, Islam 1995), where the stepwise shape is inevitably mixed with country fixed effects (see Berthélemy 2002). By definition the probit estimation shown here escapes this problem, by testing the equivalent of a stepwise pattern. The conclusions of this analysis should not be applied mechanically. The quality of education, not only its quantity, matters very much. It is likely that emerging economies have succeeded because they were serious about education, not only 154 | JEAN-CLAUDE BERTHÉLEMY because they dramatically improved their quantitative indicators of schooling in the 1950s. Least developed countries today face an environment that is different from that facing emerging countries in the 1960s, if only because of the globalization process. Nevertheless, it is unlikely that attempts to lift some of these countries out of their underdevelopment traps will have any chance of succeeding if ambitious plans are not designed to eradicate illiteracy, which is possibly even more of a strategic asset in today's globalized world economy than it was in the 1950s and the 1960s. Notes 1. Berthélemy (2005) provides a brief review. 2. Results for countries that were members of OECD before 1994 are not reported, although in many instances their growth dynamics have properties that are similar to those reported for developing countries. 3. The series transformed with the Hodrick-Prescott filter is the natural logarithm of GDP per capita. 4. Another difference is that Hausmann, Pritchett, and Rodrik use Heston, Summers, and Aten's data instead of Maddison's data. 5. A frequent mistake is to interpret the growth acceleration episodes observed by Haus- mann, Pritchett, and Rodrik as consistent with escapes from underdevelopment traps, when they could also be consistent with standard convergence, for example, from point z1 to z2 in figure 1 (see Kraay and Raddatz 2005). 6. In Swaziland the exploitation of iron ore mines has coincided with railway construction, needed for transportation of the mineral. Construction of railroads has contributed to economic development. 7. In the first group of 11 countries, data availability is better. Because most of my data are extracted from the World Bank's World Development Indicators online database, some data are missing for Taiwan (China). 8. Only an approximation of this ratio, computed on enrollment flows rather than on cap- ital stock, can be computed. Raw data used to compute the enrollment rates come from UNESCO (1955). 9. Previous comparative studies (for example, Barro 1991) concluded that schooling had a positive effect on growth. References Abramovitz, Moses. 1986. "Catching Up, Forging Ahead, and Falling Behind." Journal of Economic History 46 (2): 385­406. Azariadis, Costas, and Allan Drazen. 1990. "Threshold Externalities in Economic Develop- ment." Quarterly Journal of Economics 105 (2): 501­26. Barro, Robert J. 1991. "Economic Growth in a Cross-Section of Countries." Quarterly Jour- nal of Economics 106 (2): 407­43. Barro, Robert J., and Jong Wha Lee. 1996. "International Measures of Schooling Years and Schooling Quality." American Economic Review 86 (2): 218­23. Baumol, William J. 1986. "Productivity Growth, Convergence, and Welfare: What the Long- Run Data Show." American Economic Review 76 (5): 1073­85. CONVERGENCE AND DEVELOPMENT TRAPS | 155 Baxter, Marianne, and Robert G. King. 1999. "Measuring Business Cycles: Approximate Band- Pass Filters for Economic Time Series." Review of Economics and Statistics 81 (4): 575­93. Berthélemy, Jean-Claude. 1980. "L'économie de l'Afrique occidentale française et du Togo: 1946­1960." Revue française d'histoire d'outre-mer 67 (248­249): 302­37. ------. 2002. "Convergence Clubs and Underdevelopment Traps." In Development Is Back, ed. Jorge Braga de Macedo, Colm Foy, and Charles Oman, 61­76. Paris: OECD. ------. 2005. "Commerce international et déterminants de la diversification économique." Revue d'Economie Politique 115 (5): 591­611. ------. Forthcoming. "To What Extent Are African Education Policies Pro-Poor?" Journal of African Economies. Berthélemy, Jean-Claude, and Ludvig Söderling. 2001. "The Role of Capital Accumulation, Adjustment and Structural Change in Economic Take-off: Empirical Evidence from African Countries." World Development 29 (2): 323­43. Berthélemy, Jean-Claude, and Aristomene Varoudakis. 1996. "Economic Growth, Convergence Clubs, and the Role of Financial Development." Oxford Economic Papers 48 (2): 300­28. Chatterji, Monojit. 1992. "Convergence Clubs and Endogenous Growth." Oxford Review of Economic Policy 8 (4): 57­69. Collier, Paul. 2004. "African Growth: Why a `Big Push?' " Paper presented at the plenary ses- sion of the biannual workshop of the African Economic Research Consortium, Nairobi, December 5. Dessus, Sébastien, Jia-Dong Shea, and Mau-Shan Shi. 1995. Chinese Taipei: The Origins of the Economic "Miracle." Paris: OECD. Durlauf, Steven N., and Paul A. Johnson. 1995. Multiple Regimes and Cross-Country Growth Behavior." Journal of Applied Econometrics 10 (4): 365­84. Feenstra, Robert C., Dorsati Madani, Tzu-Han Yang, and Chi-Yuan Liang. 1999. "Testing Endogenous Growth in South Korea and Taiwan." Journal of Development Economics 60 (2): 317­41. Groningen Growth and Development Centre, and the Conference Board. 2004. "Total Econ- omy Database." Available at http://www.ggdc.net. Hansen, Bruce E. 2000. "Sample Splitting and Threshold Estimation." Econometrica 68 (3): 575­603. Hausmann, Ricardo, Lant Pritchett, and Dani Rodrik. 2004. "Growth Accelerations." NBER Working Paper 10566, National Bureau of Economic Research, New York. Heston, Alan, Robert Summers, and Bettina Aten. 2002. "Penn World Table Version 6.1." Center for International Comparisons at the University of Pennsylvania (CICUP). Avail- able at http://pwt.econ.upenn.edu/php_site/pwt_index.php. Hodrick, Robert J., and Edward C. Prescott. 1997. "Postwar U.S. Business Cycles: An Empir- ical Investigation." Journal of Money, Credit and Banking 29 (1): 1­16. Imbs, Jean, and Romain Wacziarg. 2003. "Stages of Diversification." American Economic Review 93 (1): 63­86. Islam, Nazrul. 1995. "Growth Empirics: A Panel Data Approach." Quarterly Journal of Eco- nomics 110 (4): 1127­70. Kraay, Aart, and Claudio Raddatz. 2005. "Poverty Traps, Aid and Growth." Background paper for the 2005 Global Monitoring Report. World Bank, Washington, DC. Krueger, Anne O. 1993. "Virtuous and Vicious Circles in Economic Development." American Economic Review 83 (2): 351­5. Lee, Chung H. 1995. The Economic Transformation of South Korea: Lessons for the Transi- tion Economies. Paris: OECD. 156 | JEAN-CLAUDE BERTHÉLEMY Leibenstein, Harvey. 1954. Theory of Economic-Demographic Development. Princeton, NJ: Princeton University Press. Maddison, Angus. 2003. The World Economy: Historical Statistics. Paris: OECD. Mankiw, N. Gregory, David Romer, and David N. Weil. 1992. "A Contribution to the Empir- ics of Economic Growth." Quarterly Journal of Economics 107 (2): 407­37. Morrisson, Christian, and Béchir Talbi. 1996. Long-Term Growth in Tunisia. Paris: OECD. Myrdal, Gunnar. 1968. Asian Drama: An Inquiry into the Poverty of Nations. New York: Twentieth Century Fund. Nelson, Richard. 1956. "A Theory of the Low-Level Equilibrium Trap in Underdeveloped Economies." American Economic Review 46 (5): 894­908. Nurkse, Ragnar. 1953. Problems of Capital Formation in Underdeveloped Countries. Oxford: Basil Blackwell. Pritchett, Lant. 2001. "Where Has All the Education Gone?" World Bank Economic Review 15 (3): 367­91. Quah, Danny T. 1997. "Empirics for Growth and Distribution: Stratification, Polarization, and Convergence Clubs." Journal of Economic Growth 2 (1): 27­59. Radermacher, Franz Josef. 2004. Global Marshall Plan: A Planetary Contract. Hamburg: Global Marshall Plan Foundation. Ravn, Morten O., and Harald Uhlig. 2002. "On Adjusting the Hodrick-Prescott Filter for the Frequency of Observations." Review of Economics and Statistics 84 (2): 371­80. Romer, Paul. 1990. "Endogenous Technological Change." Journal of Political Economy 98 (5): S71­102. Rosenstein-Rodan, Paul. 1943. "Problems of Industrialization of Eastern and South-Eastern Europe." Economic Journal 53 (210/211) : 202­11. Sachs, Jeffrey D., John W. McArthur, Guido Schmidt-Traub, Margaret Kruk, Chandrika Bahadur, Michael Faye, and Gordon McCord. 2004. "Ending Africa's Poverty Trap." Brookings Papers on Economic Activity 2004 (1) : 117­240. Solow, Robert M. 1956. "A Contribution to the Theory of Economic Growth." Quarterly Journal of Economics 70 (1): 65­94. Thorbecke, Erik, and Henry Wan, Jr. 2004. "Revisiting East (and South East) Asia's Develop- ment Model." Paper presented at the Cornell Conference on "Seventy-Five Years of Devel- opment," Ithaca, NY, May 7­9. Tirole, Jean. 1996. "A Theory of Collective Reputations (with Applications to the Persistence of Corruption and to Firm Quality)." Review of Economic Studies 63 (1): 1­22. UNESCO (United Nations Educational, Scientific and Cultural Organization). 1955. World Survey on Education. Paris. ------. 2004. Global Education Digest. Paris. United Nations. 1959. Economic Survey of Africa Since 1950. New York: United Nations. Young, Allyn. 1928. "Increasing Returns and the Big Push." Economic Journal 38 (152): 527­42. Comment on "Convergence and Development Traps: How Did Emerging Economies Escape the Underdevelopment Trap?" by Jean-Claude Berthélemy MTHULI NCUBE Jean-Claude Berthélemy's paper tests the factors driving economic growth and what are termed convergence clubs around the world. It presents an excellent discussion of the literature to set the context for the paper's contribution. The author notes that two groups of countries exist, defined by the initial level of income and initial level of education, as observed by Durlauf and Johnson (1995). Another stylized fact is that GDP per capita is a quadratic function of its initial level, as Quah (1997) and Hausmann, Pritchett, and Rodrik (2004) observe. The paper presents growth as characterized by differential equations driven by states of nature. The equations capture the notions of peaks and growth clubs that have emerged in the global economy. The paper also identifies some of the factors driving economic growth, including demography, savings behavior, human capital accumula- tion, financial sector development, financial institutions, and export diversification. Berthélemy analyzes growth patterns of 99 countries between 1950 and 2001. He defines a growth peak as an episode in which growth increased for at least seven consec- utive years, growth was positive throughout the same period, and growth peaked at a pace of at least 3.5 percent. On the basis of this definition he finds that 54 of the 99 countries studied experienced at least one peak and 17 countries had multiple peaks. Only five African countries--Botswana, Lesotho, Mauritius, the Seychelles, and Tunisia-- had two or more growth peaks; the majority of African countries had no peak at all. The paper estimates a probit regression model in which the dependent variable is a binary variable for the growth peaks. The independent variables in the analysis are education levels (the proportion of adult population having attended school), the ratio of M3 to GDP (a measure of financial deepening), and an Asian dummy. The results have strong implications for the notion of a "big push"--that is, mas- sive increases in financial aid to poor countries, as embodied in the United Nation's Millennium Development Goals and the United Kingdom's African Commission Report. The mere transfer of resources will not be effective in producing sustainable Mthuli Ncube is professor of finance at the University of Witwatersrand Wits Business School, Johannesburg, South Africa. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 157 158 | MTHULI NCUBE FIGURE 1. African Human Resources Development Index, 2003 3.5 3.0 2.5 2.0 Index1.5 1.0 0.5 0 BotswanaCamer Ghana Kenya Lesotho Malawi MozambiqueNami Nigeria South TanzaniaUganda Zamb Zimbabwe oon bia Africa ia Source: Commonwealth Business Council 2003. economic growth unless the efforts target education and domestic financial sector growth. Institutional development and economic and structural reforms are critical to the success of "big push" initiatives. While the availability of educated human resources is important, as the paper sug- gests, the management of these resources is equally important. Africa has been plagued by the problem of human capital flight to more stable or richer countries. Education statistics may not capture such human capital flight, which seems to have contributed to low economic growth. A human capital indicator is needed that goes beyond mere education levels (figure 1). The Human Resources Development Index, which measures human capital investment in general, could be used. Berthélemy identifies the quality of political institutions as critical in determining economic growth patterns. However, he does not attempt to use institutional indica- tors to explain growth peaks in the regression analysis. Indicators of institutional quality are, of course, hard to come by. But given the importance of corruption and civil strife in reinforcing Africa's underdevelopment trap, some indicators of corrup- tion should have been utilized as explanatory institutional variables. While the paper rightly identifies financial deepening as a driver of economic growth, it should have discussed how the domestic financial systems could be strengthened. One area is the reform of pension systems, which can lead to vibrant and deeper banking systems, capital markets, and investment institutions. Unlike the literature, which tests a large number of variables, Berthélemy tests the importance of just three variables (education, financial deepening, and an Asian dummy). It is not clear how he chose these variables. Sala-i-Martin, Doppelhofer, and Miller (2004) test 67 factors across 88 countries, using annualized GDP per capita growth rates for the period 1960­96. They find that 18 of 67 factors are sig- nificant in explaining economic growth. The 10 most significant factors are the dummy for East Asian countries, primary schooling enrollment in 1960, the average price of investment goods between 1960 and 1964, the initial level of GDP per capita in 1960, the fraction of tropical area, population densities in coastal areas in the 1960s, the prevalence of malaria in the 1960s, life expectancy in the 1960s, the frac- tion of the population that is Confucian, and the African dummy. The author could have tested more factors, as these studies suggest. COMMENT ON BERTHÉLEMY | 159 FIGURE 2. Institutional Index for Selected Countries in Sub-Saharan Africa, 2003 3.5 3.0 2.5 2.0 Index1.5 1.0 0.5 0 BotswanaCamer Ghana Kenya Les Malawi MozambiqueNamibiaNig South otho eria oon Afr TanzaniaUganda Zambia Zimbabwe ica Source: Commonwealth Business Council 2003. The paper correctly identifies institutional factors as important in determining the pattern of economic growth. Some of the factors I consider important are the avail- ability of free media, a reliable justice system, efficient administration, effective gov- ernment, the quality of corporate governance, human resources development, finan- cial infrastructure and framework, corruption reduction, and consistency of policymaking. I computed these factors from the Commonwealth Business Council literature and averaged them across each country to come up with what I call an institutional index for each country (figure 2). Botswana, which has had multiple growth peaks, has the highest level of institu- tional quality. Zimbabwe has the lowest institutional index, in line with its weak growth pattern. The paper uses ordinary differential equations to characterize economic growth dynamics. This "big-bang" approach, which says that growth begins somewhere in time and proceeds in different directions for different regions and countries, can be captured by more robust processes. If economic growth is considered stochastic, stochastic differential equations should be used to capture growth dynamics. An example of a stochastic differential equation with mean reversion is dG = b( ­ G)dt + dW(t), (1) where G is GDP per capita, is the mean GDP per capita growth rate, is the volatility of GDP per capita growth rate, and W(t) is the Weiner process. The growth rate is "pulled" to at the rate b. Equation (1) is an Ornstein-Uhlenbeck process in which growth reverts to some mean over time. Jumps in growth could perhaps also be captured using diffusion- jump processes. A stochastic differential equation process such as equation (1) has 160 | MTHULI NCUBE the added advantage that it captures the impact of volatility of GDP per capita growth, which is an important information variable. References Commonwealth Business Council. 2003. "Business Environment Survey." London. Durlauf, S.N., and P.A. Johnson. 1995. "Multiple Regimes and Cross-Country Growth Behav- iour." Journal of Applied Econometrics 10 (4): 365­84. Hausmann, R., L. Pritchett, and D. Rodrik. 2004. "Growth Accelerations." NBER Working Paper 10566, National Bureau of Economic Research, New York. Quah, D.T. 1997. "Empirics for Growth and Distribution: Stratification, Polarization, and Convergence Clubs." Journal of Economic Growth 2: 27­59. Sala-i-Martin, X., G. Doppelhofer, and R.I. Miller. 2004. "Determinants of Long-Term Growth: A Bayesian Averaging of Classical Estimates (BACE) Approach." American Eco- nomic Review 94 (4): 813­35. Trade and Development Risks and Rewards of Regional Trading Arrangements in Africa: Economic Partnership Agreements between the European Union and Sub-Saharan Africa LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER Economic Partnership Agreements (EPAs) present an opportunity to accelerate global and regional trade integration in Sub-Saharan Africa. To realize their potential develop- ment benefits, the European Union must truly treat EPAs as instruments of develop- ment, subordinating its commercial interests to Africa's development needs and effec- tively coordinating trade and development assistance. The African countries need to use EPAs to accelerate the trade and investment climate reforms necessary to raise growth rates and integrate their economies regionally and globally. A number of important issues will need to be addressed to limit the development risks associated with EPAs. If these issues cannot be satisfactorily resolved, the EPA process could end up being replaced by improved preferences or even abandoned. Africa's Trade with the European Union Trade is vitally important for Sub-Saharan Africa's small economies. On average, exports of goods and nonfactor services account for 29 percent of GDP in these countries and imports of goods and nonfactor services account for 34 percent of GDP. Partly for historical reasons, the European Union is Sub-Saharan Africa's (hereafter "Africa") largest single trading partner, buying 31 percent of Africa's mer- chandise exports and providing 40 percent of its merchandise imports (Hinkle and Schiff 2004).1 Since the 1970s, the European Union has provided unilateral preferential access to its market to 77 African, Caribbean, and Pacific (ACP) countries under the Lomé Conventions (I--IV). The Cotonou Agreement, signed in June 2000, provides for a continuation of this preferential access until 2008. Under the Cotonou trade prefer- ences, all imports of manufactured goods from the ACP countries enter the European Union duty free, although they are still restricted by rules of origin that are very Larry Hinkle is the Lead International Economist in the World Bank's Africa Region. Richard Newfarmer is economic advisor in the World Bank's International Trade Department. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 163 164 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER demanding for small low-income economies. With the exception of 990 tariff lines covering agricultural and processed agricultural products produced in the European Union, ACP agricultural products also enter the European Union duty free. The most valuable of the European Union's preferences for Africa have been those extended to a few traditional primary exports (sugar, meat, fish), some of which are governed by separate commodity protocols. In 2002 the European Union unilaterally began pro- viding complete tariff-free, quota-free market access for all imports other than arms from the 49 least developed countries--33 of which are in Africa--under its Every- thing But Arms (EBA) initiative. Despite the trade preferences accorded to Africa under the Lomé and Cotonou Agreements, Africa's share of world exports to the European Union has declined in parallel with the decline in Africa's share of total world exports. The share of the African Economic Partnership Agreement (EPA) countries' exports in the EU market has fallen steadily, from 3 percent in 1985 to 0.9 percent in 2003, a reflection of competitiveness problems and supply constraints as well as declining real prices of some primary commodities and restrictive rules of origin. In addition, Africa's exports to the European Union are heavily concentrated in primary products and petroleum; there has been little diversification of exports to the European Union under the Lomé and Cotonou Agreements. Trade between Africa and the European Union is much less important for the European Union than it is for Africa, because the European Union's economy is much larger. Africa accounts for only 1.4 percent of the European Union's total mer- chandise exports and 1.7 percent of its merchandise imports. Thus, the impact on the European Union of free trade arrangements with Africa under the planned EPAs is likely to be limited, and the adjustment is likely to be much easier for the European Union than for Africa. Furthermore, if EPA negotiations were conducted in commer- cial terms, the differences in economic size and in the relative importance of Euro- pean Union­Africa trade to the two sides would give the European Union a much stronger bargaining position. Objectives and Main Features of Economic Partnership Agreements The Cotonou Agreement, signed in June 2000 between the European Union and the ACP countries, provides for negotiation of EPAs between them to strengthen this trade relationship. EPAs are intended to reformulate the trade preferences accorded to the ACP countries under the Cotonou and Lomé Conventions to make them more supportive of broader development goals, more effective in promoting trade between the ACP and the European Union, and more compatible with World Trade Organi- zation (WTO) rules. EPA negotiations were launched at the all-ACP level in September 2002. According to the Cotonou Agreement's timetable, the negotiations are to be completed by Decem- ber 2007, and EPAs are to become effective as of January 2008, although these dead- lines could be modified by mutual agreement. Implementation is to take place gradually over 10­12 years, but this time limit could also be extended by mutual agreement. RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 165 The trade relationship between Africa and the European Union is thus important for Africa's development. EPAs could offer considerable potential benefits for Africa. This paper examines the planned EPAs from a development perspective. It seeks to clarify some of the issues faced by the European Union and African countries and to suggest changes needed to make EPAs internally consistent and development friendly.2 Objectives of Economic Partnership Agreements The overarching objective of EPAs is to support the development of African countries. Both the European Union and the ACP countries have repeatedly emphasized their desire to use EPAs as instruments of development. EPAs are expected to support eco- nomic development in the ACP countries by establishing free trade agreements with them, strengthening regional integration among them, and increasing "aid for trade" under the European Union's technical and financial cooperation programs. The second central objective of the EPA process is to promote outward-oriented regional integration among the ACP countries. Increasing trade integration in Africa's regional economic communities (RECs) is a longstanding concern of African coun- tries.3 Traditionally, the 77 ACP countries have been divided into six broad regional groupings: the Caribbean, the Pacific, and four loosely defined African subregions-- western, central, eastern, and southern Africa. EPAs are intended to establish free trade areas between the European Union and each of these regional groupings, the precise composition of which is to be determined by the ACP countries themselves. The Cotonou Agreement envisages a model of outward-oriented parallel North- South-South integration, regionally within the EPA groups and bilaterally between each of the EPA groups and the European Union. In addition, the "hub-and-spoke" effect that a series of bilateral free trade agreements between the European Union and individual ACP countries could have would be mitigated by the regional integration aspects of EPAs.4 Main Features of Economic Partnership Agreements EPAs are intended to have four key features. They are supposed to replace unilateral preferences with reciprocal free trade arrangements to make EPAs compatible with WTO rules, to provide comprehensive coverage of trade and trade-related measures, to treat lesser developed countries (LDCs) differently from other ACP countries, and to increase "aid for trade." Reciprocal Free Trade and WTO Compatibility To make trade relations between the European Union and the ACP countries com- patible with WTO rules, the Cotonou Agreement provides for replacing the existing relationship of unilateral preferential access by ACP countries to EU markets with reciprocal free trade agreements between the European Union and ACP countries. Under such reciprocal free trade, the European Union would provide tariff-free access to its markets for ACP exports and ACP countries would reciprocate by pro- viding tariff-free access to their own markets for EU exports. 166 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER The WTO compatibility problem arises because the European Union's special uni- lateral preferences for ACP countries under the Cotonou Agreement, like those under the Lomé Conventions, are inconsistent with the WTO's "enabling clause." This clause permits industrial countries to give unilateral preferential treatment to only two groups of countries: LDCs or all developing countries. Because the Cotonou preferences, like the earlier Lomé ones, are not part of the European Union's general system of preferences extended to all developing countries and some ACP countries are not LDCs, these preferences do not conform with the WTO's enabling clause. Hence the European Union needed to obtain waivers from the WTO, first for the Lomé Convention in 1994 and then for the Cotonou Agreement at the Doha Minis- terial in 2001. Even with the waivers, the European Union's commodity protocols governing preferential trade in bananas and sugar with the ACP countries have already been successfully challenged in the WTO. To bring the European Union's trade relations with the ACP countries into line with WTO rules, the Cotonou Agreement provides for replacing the unilateral trade preferences that the European Union currently accords to the ACP countries with economic partnership agreements involving reciprocal obligations. Like other free trade agreements between developed and developing countries, EPAs would be gov- erned by Article XXIV rather than by the enabling clause, which applies only to uni- lateral preferences granted to developing countries and preferential trade agreements between developing countries. Article XXIV requires that countries entering into a reciprocal free trade agreement liberalize "substantially all trade" within a "reason- able length of time," without distinguishing between developed and developing countries.5 WTO compatibility is an important consideration for the European Union; it weighed heavily in its launching of the EPA process. In contrast, because of their small size, many African countries have a low profile in the WTO, have been exempted de facto from many WTO disciplines, and have not been involved directly in the disputes over commodity protocols. Many African countries have argued that they should be exempted from Article XXIV or that it should be amended in the Doha Round to permit less than full reciprocity by developing countries participat- ing in a free trade agreement with industrial countries. Comprehensive Coverage and the Implicit Reform Model The Cotonou Agreement envisages that EPAs will include comprehensive coverage of trade in services, investment, competition, trade facilitation, and aid for trade as well as merchandise trade. The European Union sees EPAs as instruments for addressing many of the supply-side constraints that have limited the expansion and diversification of African exports to the European Union despite the Lomé-Cotonou trade preferences. EPAs would thus be used to leverage and lock in broad programs of trade and investment climate reforms. The European Union envisages EPAs as broadly similar in this respect to the accession agreements between the European Union and Central and Eastern Europe, to the European Union­Mexico free trade agreement and NAFTA, and to Chile's free trade agreements with the European Union and the United States. In contrast, despite having signed the Cotonou Agree- RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 167 ment, many African countries are hostile to the inclusion of the "Singapore issues" (investment, competition, trade facilitation, and government procurement) in EPAs, which they perceive as a "divide and conquer" strategy through which the European Union can achieve using EPAs what it could not in the WTO. Differentiation between LDCs and Non-LDCs The Cotonou Agreement is supposed to differentiate between LDCs and non-LDCs. The last part of Article 2 of the Agreement states: "Differentiation and regionaliza- tion: cooperation arrangements and priorities shall vary according to a partner's level of development, its needs, its performance and its long-term development strategy. Particular emphasis shall be placed on the regional dimension." Part 1 of Article 85 adds that "the least-developed ACP States shall be accorded a special treatment in order to enable them to overcome the serious economic and social difficulties hinder- ing their development so as to step up their respective rates of development." For example, LDCs may not be expected to open their markets to EU exports as rapidly or as much as non-LDCs to maintain their preferential access to EU markets as long as the differential pace of market opening does not undermine the overriding develop- ment and regional integration objectives. As explained below, the differentiation prin- ciple creates some complications with respect to regional integration within Africa, and differentiation between LDCs and non-LDCs could end up being a temporary fea- ture of EPAs. Aid for Trade A fourth intended feature of the EPAs is more effective coordination of trade and aid. The European Union is one of Africa's largest aid donors. In principle, its financial and technical cooperation programs could provide substantial assistance for overcoming problems and taking advantage of the opportunities created by improved market access and trade liberalization under EPAs. The opportunity to effectively coordinate trade and aid with a major trading partner and aid donor is a distinct advantage of the EPA process relative both to multilateral trade negotiations and to most other bilateral trade negotiations, where the link between aid and trade is tenuous or nonexistent. Coordinating the programs of the European Commission's two large general direc- torates for trade and aid and the finance and trade ministries of African countries, all of which have different constituencies and perspectives, is essential to achieving a pro- development outcome. But this task will be a challenge for all concerned. One issue with particularly important implications for the EPA negotiations is the level of financial resources that will be allocated to support trade liberalization in Africa. The level of development finance available from the European Union under the Ninth European Development Fund has already been set through the end of 2007, and its regional integration support programs have been worked out with the various country groupings. The ACP countries contend that these resources are insuf- ficient to support trade liberalization and expansion as well as their other develop- ment needs. Any effects from the changes in market access and liberalization under the EPAs will be felt only after 2007, potentially long thereafter, depending on the liberalization 168 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER schedules. Because the financial envelopes for the 10th European Development Fund and subsequent funds for the period after 2007, when the EPAs will be implemented, have not yet been determined, their size and operational features presumably could still be designed to support the EPA process. Because of the limited scope that the European Union has under EPAs to offer additional trade preferences to African countries, particularly LDCs, aid for trade will have to play a central role in creating a favorable economic and political envi- ronment for the comprehensive reforms envisaged under the EPAs. As part of the G-8 effort to mobilize increased funding for development in Africa, the European Union has pledged to double its aid levels by 2012. Trade Commissioner Peter Man- delson has indicated that a "substantial portion" of this increase will go to aid for trade and that a mechanism will be put in place to strengthen coordination between the relevant directorates of the European Commission. Economic Partnership Agreements and Regional Trade Groups in Africa Intraregional trade within Africa is limited, amounting to only about 12 percent of the average African country's merchandise exports and 7 percent of its merchandise imports. Intra-African and transit trade is, however, very important for landlocked African countries and for a handful of coastal countries (Côte d'Ivoire, Kenya, Nige- ria, and South Africa) that have some manufacturing capacity. Significant barriers to trade integration remain within free trade areas in Africa and even within customs unions. An important objective of the EPA process is to promote outward-oriented regional integration among African countries and to limit the hub- and-spoke effect that bilateral free trade between the European Union and individual African countries could have. However, determining the nature of the EPA groupings in Africa has been no simple matter, because of the multiplicity and heterogeneity of regional trade agreements, which include a number of overlapping preferential trade areas, free trade agreements, and customs unions with different structures, operational rules, and implementation levels. The interaction of the EPA process with political support for regional integration in Africa creates an impetus for rationalizing this sit- uation and an opportunity for removing intraregional trade barriers. Negotiation of the Economic Partnership Agreements Phase I of the EPA negotiations, during which ACP-wide issues were to be addressed by the group as a whole, ran for one year, from September 2002 through October 2003. The ACP group and the European Community then issued a joint ministerial statement declaring that Phase I of the EPA negotiations had proceeded "satisfacto- rily" and that there was a high degree of "convergence" on matters of principle. Both sides agreed that discussions on common ACP-wide issues would continue during Phase II, in parallel with EPA negotiations at the regional level. In fact, there appears to have been little agreement on many key issues between the ACP group and the European Union and even disagreement on some issues RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 169 within the ACP group itself. Issues that were still under discussion at the end of Phase I included WTO compatibility; treatment of non-LDCs not entering into EPAs; rules of origin; technical barriers to trade and sanitary and phytosanitary issues; safe- guards, antidumping, and dispute settlement; commodity protocols; an ACP­ European Union framework agreement on fisheries; the fiscal, economic, balance of payments, and social implications of EPAs; and implementation mechanisms. In parallel with Phase I, the European Union initiated discussions of regional EPA groupings with Africa. The African countries subsequently formed the following four regional EPA groups for negotiating EPAs with the European Union: · The Economic Community of West African States (ECOWAS) plus Mauritania form the ECOWAS EPA group. · The Central African Economic and Monetary Community (CEMAC) plus São Tomé and Principe form the CEMAC EPA group. · Sixteen countries, including most of the members of the Community of Eastern and Southern African (COMESA), form the Eastern and Southern Africa (ESA) EPA group, for which COMESA serves as Secretariat. · Four Southern African Customs Union (SACU) members (all but South Africa) and three neighboring countries (Angola, Mozambique, and Tanzania) form the Southern Africa Development Community (SADC) EPA group. In October 2003 the CEMAC and the ECOWAS EPA groups launched Phase II negotiations. The Eastern and Southern Africa group began negotiations in February 2004, and negotiations with the SADC EPA group began in July 2004. All of the groups have agreed on broad "roadmaps" setting out the organizational arrange- ments for the negotiations. The first year of Phase II was devoted to two tasks. The first was establishing the status of regional integration within the four negotiating groups to address measures for integrating and strengthening regional markets, including their outward orientation, before considering steps to liberalize European Union­Africa trade. The second task was discussing selected questions of priority interest to individual EPA groups, such as fisheries, sanitary and phytosanitary regu- lations, and technical barriers to trade. General Issues Concerning Regional Negotiating Groups in Africa Although the composition of the four regional EPA negotiating groups in Africa has now largely been determined, a number of important regional trade integration issues remain to be addressed. Role of Customs Unions Of the four EPA negotiating blocs in Africa, three (the ECOWAS group, the Eastern and Southern Africa group, and the SADC group) are free trade areas that contain smaller customs unions (the Union Economique et Monétaire Oueste Africaine [UEMOA] in ECOWAS, the East African Community [EAC] in Eastern and South- ern Africa, and the Southern African Customs Union [SACU] in SADC). The fourth 170 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER negotiating group, CEMAC is a customs union. The members of each customs union share a common external tariff. The existence of a common external trade policy per- mits the members of customs unions (that is, UEMOA, CEMAC, SACU, and EAC) to enter into trade agreements, including EPAs, as a group, because all trade agree- ments apply equally to all member countries. The European Commission believes that, ideally, it would be desirable for all of the EPA regions to implement common external tariffs before the entry into force of the EPAs and that it is desirable to pur- sue EPA negotiations with a customs union first, to determine how much can be accomplished in this format and establish a good precedent for addressing regional integration issues. Customs Unions with LDC and Non-LDC Members A complication in negotiating EPAs with customs unions is the potential inconsis- tency between the Cotonou Agreement's differentiation in the treatment of LDCs and non-LDCs and its regional integration objective. All four of the customs unions in Africa include both LDCs and non-LDCs. Under its Everything But Arms (EBA) initiative, the European Union unilaterally granted full duty-free and quota-free sta- tus to LDCs, without requiring preferential access to the LDCs' markets in return. Having already obtained EBA preferences, LDCs now have little incentive to open their markets to the European Union's exports on a preferential basis under EPAs, unless EPAs provide other important benefits for LDCs, as discussed below. In contrast, after the Cotonou Agreement trade preferences expire in 2007, non- LDCs in Africa must agree to provide preferential access to EU exports to maintain preferential access for their own exports to the European Union. This difference in negotiating incentives is particularly serious for customs unions, which cannot main- tain their common external tariffs if preferential access to imports from the European Union is granted by their non-LDC members but not by their LDC members. Given the importance of regional integration to the African countries, differentiation could end up being a temporary feature of the EPA process. Role of Free Trade Agreements in EPAs The situation is different in free trade areas, within which each country is free to maintain its own external trade policy. On the one hand, because of differences in tariffs and other trade policies in free trade agreement members, negotiating a single, unique EPA that applies identical provisions to all African members of a free trade agreement is not likely to be feasible; instead, separate (similar) EPAs will need to be negotiated by each of the free trade agreement's member countries. On the other hand, a free trade agreement makes it easier to provide for differential treatment of member countries, because they do not all have to follow a common external trade policy. In particular, the tariffs of LDCs and non-LDCs on imports from the Euro- pean Union could differ. Because all four of the EPA groups are combinations of cus- toms unions, free trade areas, and countries with independent trade policies, it may not be feasible to establish customs unions with common external tariffs including all countries in each EPA group, as the European Commission envisages. Instead, a flexible variable geometry approach may be needed, at least in the near term, until regional integration is further advanced in Africa. RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 171 Barriers to Intraregional Trade in Africa Despite the proliferation of preferential trade agreements in Africa, significant barri- ers to intraregional trade still remain within "free" trade areas, and even within cus- toms "unions," in all four of the regional EPA negotiating groups. Intraregional free trade, where it has been effectively implemented, is generally limited to a small group of selected products.6 None of the three free trade areas involved in the EPA negotiations (ECOWAS, COMESA, and SADC) yet has even relatively free internal trade among its members. The ECOWAS free trade agreement has not yet been effectively implemented. Twelve of COMESA's members have only recently started implementing its free trade agree- ment, and some countries that are doing so still maintain substantial barriers to free intraregional trade. Internal trade among the SADC group is even more restricted, except within SACU. Intraregional trade is not much freer within Africa's four customs unions. With the exception of SACU, these are all partial customs unions, in the sense that, although they have common external tariffs, there is no pooling of customs revenue, member countries maintain customs barriers within the union, and trade within the customs unions is subject to restrictive rules of origin. In both CEMAC and UEMOA, substan- tial obstacles to internal free trade and country deviations from the common external tariff remain. The EAC customs union is just becoming operational and faces similar problems, because its design is similar. Only SACU is a fully functioning customs union with relatively free internal trade, a common external tariff observed by all of its members, common administration of the external tariff, and pooling of the rev- enues from it. The advantages of a "customs union" in which barriers to internal trade remain in place or members do not fully implement the common external tariff are of course limited. Rules of origin that constrain intraregional trade in Africa are a problem in both free trade agreements and customs unions. Within all of the free trade areas and all of the customs unions except SACU, internal trade is subject to rules of origin, most of which are quite restrictive. Protectionist interests constantly lobby to increase the restrictiveness of these rules of origin as well as to maintain high tariff barriers. Lib- eralization of rules of origin within regional economic communities is thus an impor- tant regional integration issue in Africa. Inefficiencies and corruption in customs administration, cumbersome and costly transit arrangements, and informal road blocks and "tolls" on key transit routes are also major obstacles that need to be addressed under the rubric of intraregional trade facilitation. In addition, some natural trading partners with substantial cross-border trade belong to different regional economic communities and EPA groups. Cross-border trade between the members of different RECs is not covered by liberalization meas- ures governing trade within the separate regional economic communities. For exam- ple, Nigeria belongs to ECOWAS, whereas Cameroon, on Nigeria's southern border, belongs to CEMAC. The Republic of Congo belongs to CEMAC, whereas the Demo- cratic Republic of Congo belongs to the Eastern and Southern Africa group. The sub- stantial cross-border trade between these countries thus does not benefit from meas- ures designed to liberalize intra-REC trade within ECOWAS and CEMAC. 172 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER Measures to Increase Intra-African Trade Although political support for regional integration is currently greater than sup- port for the liberalization of external trade. African countries will still need to over- come political and administrative obstacles to liberalization of intra-African trade. In the EPA process, several steps could be taken to promote intra-African trade in goods and services and minimize the hub-and-spoke effect of EPAs: · Full free intra-regional trade in goods and services and related trade facilitation measures could be implemented within customs unions and free trade agreements in all EPA groups. Free movement of labor within regional economic communi- ties also needs to be encouraged. These reforms are particularly important for landlocked and poorer countries. · EPA groups could implement free trade with one another at the same time that they implement free trade with the European Union. Doing so would promote cross-border trade between members of different regional economic communities. It would also limit the potential hub-and-spoke effect that could result from the European Union's establishing separate free trade areas with each regional EPA group in Africa. · African countries could take advantage of the EPA process to adopt simple, stan- dardized, liberal rules of origin for free trade agreements and customs unions in Africa and address the other trade facilitation issues noted above. · Members of customs unions that have not fully implemented their common exter- nal tariff could do so while implementing other trade liberalization measures under their EPAs. Implementation of the above reforms is likely to be difficult, however, as some of them have existed on paper for years in some RECs but little progress has been made in implementing them. Establishment of free trade with the European Union could significantly reduce protectionist opposition to effective implementation of free trade within Sub-Saharan Africa and liberalization of most favored nation (MFN) trade. Once a product can be freely imported from the European Union, there will be less incentive for protectionist interests to block similar imports from other REC mem- bers or the rest of the world. Region-Specific Issues Three important region-specific issues affecting EAC, SACU-SADC, and ECOWAS have not yet been resolved. One important issue concerns the EAC's customs union, three of whose members (Kenya, Rwanda, and Uganda) are participating in the East- ern and Southern Africa group while the fourth (Tanzania) belongs to the SADC group. Because the four EAC countries are implementing a common external trade policy, they will need to negotiate a common EPA as a group. Consequently, some geographic realignment of the ESA and SADC EPA negotiating groups may be required. A second issue is the complex, confused situation with the EPAs for SADC and the SACU countries. Four of the five SACU countries (Botswana, Lesotho, Namibia, RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 173 and Swaziland) are ACP countries and have joined the SADC EPA group, along with Angola, Mozambique, and Tanzania. The fifth and by far largest SACU member, South Africa, which has a central role in determining SACU's trade policies, is eligi- ble neither for the trade benefits of the Cotonou Agreement nor for an EPA. Further- more, South Africa has already signed a separate free trade agreement with the Euro- pean Union (the Trade Development and Cooperation Agreement), many aspects of which apply de facto to the other four SACU countries. In addition, SACU is cur- rently negotiating a free trade agreement with the United States. More thought needs to be given to the role of SACU, the EU­South Africa free trade agreement, and mem- bership in the SADC and ESA EPA groups to create a sound basis for the regional and global integration of the economies of the subregion. The third issue concerns the planned ECOWAS common external tariff and Nige- ria. ECOWAS comprises the eight members of the UEMOA customs union and seven other countries. The ECOWAS members have adopted a free trade agreement, but its implementation has been very limited. ECOWAS plans to create a customs union, with the non-UEMOA countries adopting a common external tariff similar to UEMOA's common external tariff by 2008, when EPAs are supposed to come into effect. In principle, Nigeria and the other ECOWAS countries thus may be able to negotiate a common EPA as a customs union. However, Nigeria's economy is larger than that of all of the other ECOWAS countries put together, its trade policy is cur- rently much more restrictive than that of the UEMOA, and it is an open question whether Nigeria will implement the ECOWAS common external tariff in the near future or demand modifications to it. If full implementation of the ECOWAS cus- toms union is delayed significantly, negotiation of the ECOWAS EPA could be more complicated than currently envisaged, involving variable geometry arrangements for some countries that have implemented the ECOWAS customs union arrangements and others that participate only in the ECOWAS free trade agreement. Africa's Access to the EU Market under Everything But Arms In addition to giving trade preferences to ACP countries under the Cotonou Agree- ment, the European Union provides preferences to LDCs under the EBA initiative, adopted in 2001. EBA has complicated the EPA process by creating different market access conditions and negotiating incentives for the 33 LDCs and 13 non-LDCs in Africa eligible for EPAs. Implications of the Everything But Arms Initiative for LDCs in Africa Under the EBA initiative, the European Union unilaterally grants to all 49 LDCs-- including non-ACP LDCs, such as Bangladesh--quota-free and tariff-free access to its market for all products except arms.7 The 33 LDCs in Africa thus now have access to the EU market under the EBA initiative as well as under the Cotonou Agreement. EBA is part of the European Union's General System of Preferences (GSP). It is compatible with the WTO's enabling clause, as it grants special preferences to a 174 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER permissible grouping of developing countries (LDCs). In contrast to the European Union's broader GSP, which is revised every three years, EBA runs for an unlimited period and is not subject to periodic reviews. Market access is thus more secure under EBA than under GSP. For this reason, it is presumably more likely to encourage investment in new exports. However, as countries develop, they can graduate from LDC status (determined by the United Nations Conference on Trade and Develop- ment and the WTO), thereby losing their eligibility under the EBA initiative. EBA has also eliminated the preference margin of ACP countries relative to non-ACP LDCs, although the Cotonou rules of origin for the ACP countries are less restrictive than the EBA's rules of origin. As the EBA preferences were provided unilaterally by the European Union without a quid pro quo from LDCs, LDCs have little to gain in terms of lower tariffs on their exports from entering into reciprocal free trade with the European Union under EPAs. For LDCs to obtain something in return for giving EU imports tariff-free access to their markets, EPAs need to offer benefits beyond those provided by the EBA initiative.8 Market Access for Non-LDCs in Africa Access to the EU market by the 33 LDCs in Africa that benefit from EBA would pre- sumably not be greatly reduced if they decided not to enter into EPAs (except possi- bly as a result of EBA's somewhat more restrictive rules of origin than the Cotonou Agreement, discussed below). In contrast, when the Cotonou Agreement's trade pref- erences expire in 2008, any of the 13 non-LDCs that do not sign EPAs will presum- ably revert to GSP status. The European Union's GSP provides less favorable prefer- ential access than the Cotonou Agreement--narrower product coverage, smaller margins of preference, and more restrictive rules of origin.9 The value of Cotonou preferences, and hence the impact of reverting to the Euro- pean Union's less favorable GSP, varies considerably among the 13 non-LDCs in Africa. For three of these countries, the value of the tariff savings under the Cotonou preferences in 2002 exceeded that under the GSP preferences by 16­39 percent of the value of their total exports to the European Union. Swaziland benefited from tariff sav- ings on meat equivalent to 39 percent of its total merchandise exports to the European Union, Mauritius from tariff savings on sugar and knitted garments equivalent to 20 percent of its total EU merchandise exports, and the Seychelles from tariff savings on fish equivalent to 16 percent of its exports to the European Union. In contrast, for the 10 other non-LDCs, the value of the tariff savings under the Cotonou preferences for current exports is small relative to that under the GSP. For five non-LDCs (Cameroon, Côte d'Ivoire, Kenya, Namibia, and Zimbabwe) the loss would be 3­6 percent of the value of their total merchandise exports to the European Union. For the remaining five non-LDCs (Botswana, the Republic of Congo, Gabon, Ghana, and Nigeria), the loss would be 2 percent or less. Thus in the medium term, these 10 coun- tries would not lose much in terms of reduced market access by reverting to the Euro- pean Union's GSP. The immediate effect of moving from the Cotonou to the EBA regime under an EPA would also vary significantly among the 13 non-LDCs in Africa. For four coun- RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 175 tries (Botswana, Cameroon, Côte d'Ivoire, and Namibia), moving from Cotonou to EBA preferences would be worth 3­8 percent of the value of their current exports to the European Union. For the other nine non-LDCs, the immediate gain from moving from Cotonou to EBA would be less than 1 percent of the value of their current exports to the European Union. Thus, in the near term, the gain solely from moving from Cotonou to EBA access is likely to be small.10 However, these figures cover only current export products and quantities--they do not take into account the effect of improved prices or market access on either the expan- sion of existing exports or the diversification into new products that would be possible with EBA access to the EU market. The figures also underline the importance of address- ing other constraints to export diversification, such as the European Union's restrictive rules of origin and competitiveness and supply problems in the African countries. The non-LDCs in Africa thus need to obtain improved market access through EPAs to create opportunities to diversify and expand their exports in the long term. Discriminating under EPAs between the exports of LDC and non-LDC members of the same REC could slow the growth of the REC's exports and the development of the REC as a whole. Hence the most development-friendly option would be for the European Union to provide EBA market access to all African countries--LDCs and non-LDCs--that sign EPAs. The European Union's Rules of Origin The rules of origin under both the Cotonou Agreement and the EBA initiative are complex and constrain Africa's exports. The rules of origin under EBA are the same as those under the European Union's GSP and are less favorable than those under the Cotonou Agreement. In fact, the rules of origin that apply under Cotonou are suffi- ciently less restrictive to make continuing to export under them more attractive for some African exporters of some products that face low tariffs under Cotonou than switching to EBA, with its zero tariffs but more restrictive rules of origin. As a result, African LDCs make little use of EBA preferences and continue to export to the Euro- pean Union under Cotonou preferences (Brenton 2003). Two examples illustrate the potential benefits from liberalizing restrictive rules of origin. The first is the success of garment exports from Africa under the provisions of the U.S. Africa Growth and Opportunity Act's (AGOA) preferences, which permit African garment exporters to use, for a temporary period, fabrics imported from third countries outside the region instead of having to source the fabrics from within the region or the United States. The second example is Mauritius's exports of knit products to the European Union, which get around the European Union's two-stage processing rule of origin governing textiles and garments by producing knit garments directly from imported yarn. Unless the European Union's rules of origin are liberalized and simplified under EPAs, the benefits of even full tariff-free, quota-free market access will be limited. Liberalization of these rules of origin is particularly important for the diversification of Africa's exports to the European Union, because restrictive rules of origin typically prevent the development of simple nontraditional exports. 176 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER The Commission for Africa (the Blair Commission) has recommended that the European Union's complex rules of origin be replaced by a simple, uniform 10 per- cent value-added requirement in EPAs. An even less restrictive option would be to allow African exporters to choose between a "change of tariff heading" rule and a uniform 10 percent value-added rule.11 Liberalization of African Imports from the European Union, Parallel Reductions in MFN Tariffs, and Complementary Reforms in Africa In return for secure preferential access to the EU market under the EPAs, African countries are supposed to reciprocate by making preferential reductions in tariffs on Africa's imports from the European Union. Designing the required liberalization of African imports in a pro-development fashion is a complex and controversial aspect of the EPA process.12 Two important complementary reforms in Africa will be required to achieve pro-development outcomes: reducing MFN tariffs and restructur- ing indirect tax systems. Likely Extent and Timing of Preferential Tariff Reductions under EPAs Article XXIV of the WTO, which governs all free trade agreements involving indus- trial countries, requires that tariffs be completely eliminated on "substantially all trade" by participants in free trade agreements within a "reasonable length of time." Although precise definitions of the two phrases in quotations have not been estab- lished, "substantially all trade" has been interpreted in previous EU free trade agree- ments to mean 90 percent or more. The 90 percent figure refers to the existing level of trade (which is constrained by existing tariffs and other trade restrictions) rather than to the potential trade level once trade liberalization has taken place. Under the free trade agreement between the European Union and South Africa, the European Union will eliminate tariffs on about 95 percent of its imports from South Africa, and South Africa will eliminate tariffs on about 86 percent of its imports from the European Union to reach an average coverage of about 90 percent of the trade between the two. If, as seems likely, a similar approach is followed in EPAs, the European Union would probably liberalize 100 percent of its imports from Africa (that is, grant EBA access to all imports from Africa), and Africa would liber- alize 80­90 percent of its imports from the European Union to arrive at a target aver- age liberalization of 90 percent or more. The structure of protection differs significantly from country to country in Africa. Hence, agreeing on a common list of sensitive imports to be excluded from preferen- tial liberalization with the European Union by all the members of an EPA group is likely to involve complex negotiations, unless common external tariffs are adopted before negotiations over EPA tariff reductions commence. For example, Stevens and Kennan (2005) find that within the four African EPA groupings there is only a 1­12 percent overlap of the high tariff products that member countries may wish to exclude from liberalization for protective reasons.13 Furthermore, the imports that generate RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 177 the most tariff revenues often differ significantly from those competing with domestic industry. If confirmed, these findings would suggest that countries in free trade areas that do not have a common external tariff may wish to exclude very different prod- ucts from preferential free trade. Thus, in EPA groups that do not adopt a common external tariff before EPA tariff negotiations, regional harmonization of exclusions may be necessary to avoid creating different tariff regimes in neighboring countries and to foster regional integration. It is also possible that, depending on the magnitude of likely revenue losses and the fiscal options for replacing them, EPA groups may need to make tradeoffs between revenue and protective objectives in determining a common list of imports to be excluded from preferential free trade by all members. Ten years has been generally considered to be a "reasonable" length of time for implementation of EU free trade agreements, except in special cases. A 12-year implementation period was allowed in the Trade Development and Cooperation Agreement between the European Union and South Africa, and the Blair Commis- sion for Africa has suggested extending this period to 20 years in EPAs. Thus, the implementation period for EPAs is likely to fall in the 12- to 20-year range. Problematic Effects of Preferential Tariff Reductions Many African countries still have high and distorted MFN tariffs. Even in countries in which peak tariffs have been reduced to the 20­25 percent range, tariff exemp- tions and special protective measures are often a problem. The existing structure of tariffs within the various EPA groups increases the likelihood that powerful protec- tionist interests, including foreign firms producing behind high tariffs, will resist adoption of an efficient common external tariff. Preferential bilateral tariff reduc- tions under EPAs could, in the presence of high MFN tariffs, lead to costly diversion of trade from low-cost non-EU to high-cost EU suppliers as well as to some monop- oly pricing by EU exporters and implicit transfers of forgone tariff revenues from African governments to them. Little empirical work has been done on the risk of trade diversion and monopoly pricing by EU suppliers in Africa. Advocates of EPAs tend to assume that the EU economy is sufficiently large and competitive that trade diversion and monopoly pricing are not likely to be serious problems. Observers familiar with supply condi- tions in small, isolated African markets that are costly to serve and with historical patterns of market dominance in the region tend to be more skeptical about the exis- tence of effective competition among EU suppliers. As explained above, it is likely that the preferential elimination of African tariffs on imports from the European Union will cover only 80­90 percent of current imports and completely exempt some sectors and products. Line-by-line negotiation of which imports are to be excluded from free trade with the European Union is likely to be dominated by protectionist interests and will not lead to pro-development outcomes. Preferential liberalization may well be restricted to sectors with little or no domestic production, while the main import-competing industries (where the largest efficiency gains might occur) may be excluded because of opposition from currently protected firms, as was the case in the European Union­South Africa Trade 178 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER Development and Cooperation Agreement. Limiting liberalization to 80 percent of Africa's current imports would probably allow most of the important import-com- peting sectors in Africa to be exempted because, with relatively small import-compet- ing manufacturing sectors and current high protective tariffs (and quantitative restrictions in a few cases), imports of products that compete with domestic produc- tions are likely to represent a small percentage of total imports (see, for example, Stevens and Kennan 2005). A tariff structure resulting from selective, ad hoc preferential reductions in tariffs on imports from the European Union would be problematic both for the sectors for which imports are liberalized and for those for which they are not. Sectors for which tariffs on EU imports are eliminated may experience trade diversion or monopoly pricing by EU suppliers; sectors for which current high tariff rates are maintained would continue to have little incentive to increase their efficiency or reduce monop- olistic profits. In addition, eliminating tariffs on inputs imported from the European Union would raise effective protection rates for import-competing industries and reduce the incentive to export, even to the European Union. Thus, EPAs could fur- ther distort existing tariff regimes by increasing the already high effective protection rates for import-competing domestic industries while eliminating revenues from tar- iffs on imports from the European Union that do not compete with domestic produc- tion. Subsequent negotiation of similar such selective partial "free" trade agreements with additional OECD countries (as, for example, SACU is currently doing with the United States) would exacerbate these distortions. Need for Accompanying MFN Tariff Reductions Because of the distortions that can result from ad hoc, selective preferential liberal- ization, a number of proactive developing countries (Bolivia, Chile, and Mexico) lowered their MFN tariffs to a uniform rate of 10 percent or so on a comprehensive basis before selectively eliminating tariffs on a preferential basis under free trade agreements with the European Union and the United States. They have followed up on these initial cuts with additional MFN liberalization and adoption of free trade agreements with other large trading partners. The most successful regional trade agreements (the European Union itself, NAFTA) have also used low MFN tariffs to stimulate trade (World Bank 2005). Initial MFN Tariff Reductions Parallel MFN and preferential liberalizations are beneficial because of both the stan- dard static benefits from increases in trade and the much larger dynamic gains from a more open economy and because of the reduction in trade diversion and the implicit transfer of forgone tariff revenues to foreign suppliers benefiting from pref- erential tariff reductions. African countries would need to complete such an MFN liberalization before the preferential tariff reductions under EPAs with the European Union take place, so that trade diversion would be minimized and EU suppliers would not have the chance to sell in highly protected domestic markets in Africa, thereby obtaining large implicit transfers of forgone tariff revenues, before the MFN RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 179 tariff is reduced. To increase competition and lower the prices of tradable goods in Africa, the liberalization would need to cover all sectors, including the main domes- tic import-competing industries. However desirable the adoption of a low uniform tariff of 10 percent or so may be economically, because of political or revenue constraints it may not be possible to lower the MFN tariff this much in all EPA groups before implementing preferential tariff reductions under the EPA. In those EPA groups (ECOWAS and CEMAC) in which establishment of a common external tariff precedes implementation of the EPA, adopting or lowering the common external tariff will offer an opportunity to start the necessary MFN tariff reductions in a coordinated way and to reduce the magnitude of the subsequent changes needed before implementing preferential liber- alization of imports from the European Union under an EPA. For example, the ECOWAS/UEMOA common external tariff with rates of 0, 5, 10, and 20 percent is currently the most open tariff structure in Africa. Its implementation by the ECOWAS countries that have not already done so would put them in a good start- ing position for the subsequent MFN reductions needed for an EPA. ECOWAS's common external tariff, with its maximum rate of 20 percent and simple four-rate structure, is also a good model for other EPA groups to follow in adopting (or liber- alizing) their own common external tariffs before preferentially reducing tariffs from the European Union under EPAs. In the case of any of the free trade areas (ESA or SADC) in which it turns out not to be feasible to adopt an efficient common external tariff before implementing an EPA, a well-designed MFN liberalization would make it easier for the countries in the free trade agreement to move toward a common external tariff and create a favorable environment for further liberalization of trade within regional economic communities. As new countries adopt common external tariffs, there will inevitably be protectionist pressures for high tariffs (as is possible, for instance, in the case of Nigeria's adopting the ECOWAS common tariff) that will need to be resisted. Subsequent Parallel Two-Track MFN and Preferential Tariff Reductions The initial MFN tariff reductions in the EPA groups may not go as far as ultimately needed. However, the selective, phased nature of the likely preferential liberalization of imports from the European Union, with 80­90 percent of EU imports becoming tariff free over 12 years or more and the remaining 10­20 percent continuing to face current tariffs, offers the possibility of a parallel two-track approach to MFN trade liberalization, with different schedules of reform for the two groups of products. In the case of the 80­90 percent of imports from the European Union that will have no tariffs, the simplest approach for avoiding trade diversion and monopoly pricing resulting in implicit transfers of forgone tariff revenues to EU suppliers is for African countries to lower their MFN tariffs on products that will be freely imported from the European Union to no more than 5 percent. From a development perspective, the only rationale for maintaining even such a small preference margin for imports from the European Union would be to use the availability of this margin as an incentive to encourage other countries to enter into similar free trade agreements, an unlikely 180 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER prospect in current circumstances. A zero preference margin (that is, lowering the MFN tariff to the same level as the preferential EU tariff) would be the most development-friendly outcome, because it would minimize the risk of trade diversion or monopoly pricing by EU suppliers. In addition, to limit revenue losses during the implementation period, instead of directly lowering to zero the tariffs on imports from the European Union that are to be fully liberalized under EPAs, the preferential and MFN tariff reductions on these products could be carried out in steps, by first gradu- ally lowering the preferential European Union and MFN tariffs to a revenue tariff of 5 percent or so on these imports and then maintaining the two tariffs at this level for as long as possible for revenue purposes. In the case of the 10­20 percent of imports on which there would be no prefer- ential reductions in the tariffs facing the European Union, EPAs by themselves would create no immediate urgency to reduce MFN tariffs, except to offset the increases in effective protection that would result from eliminating the tariffs on many imported inputs. However, MFN liberalizations are often hard to engineer politically, and adoption of common external tariffs by the African EPA groups will complicate future liberalizations, because their members will not be able to unilat- erally reduce their tariffs and all (or a decisive majority) of an EPA group's mem- bers will need to support future liberalizations. Hence it would be desirable to take advantage of the reform dynamics of the EPA process to continue the trade liberal- ization effort. To avoid increasing antiexport bias when tariffs on inputs imported from the European Union are lowered and to put competitive pressure on import- competing sectors to operate more efficiently, the maximum tariff in the nonliber- alized sectors should be gradually lowered to 15 percent or less, with the lower tar- iff bands adjusted as necessary.14 In cases in which such MFN reductions are likely to cause particularly difficult adjustment problems, it would be better to allow more time to phase in the tariff reductions (as the Commission for Africa has recom- mended) and to provide more adjustment assistance under the aid for trade program than to postpone indefinitely the required structural adjustment by exempting some industries from the liberalization. The maximum 15 percent MFN tariff on the 10­20 percent of trade not subject to preferential liberalization could subsequently be further reduced to 10 percent or so in accordance with the needs of individual customs unions and countries. MFN tariff reductions are essential if EPAs are to achieve their development objective. The European Union therefore needs to support well-designed parallel MFN tariff reductions rather than just push for preferential reductions in tariffs on EU imports. If the MFN reductions cannot be formally included in the EPAs, the nec- essary MFN tariff reductions will need to be made by African customs unions and countries unilaterally lowering their MFN tariffs at the same time as they lower their tariffs on imports from the European Union under EPAs. Potential Revenue Losses and the Restructuring of Indirect Tax Systems Import tariffs remain an important source of government revenue in many African countries. Preferential reductions in tariffs on merchandise imports from the Euro- RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 181 pean Union could have significant fiscal implications for some of these countries. Rev- enues from import tariffs amount to about 2 percent of GDP and 15 percent of gov- ernment revenues in the median African country. In the one-third of African countries that are most dependent on tariff revenues, import tariffs amount to 3­6 percent of GDP and 25 percent or more of government revenues, reaching as much as 40­50 per- cent of government revenues net of grants in a few extreme cases.15 Because the European Union is the largest source of imports for most African countries, supplying 40 percent of total imports in the average country, some coun- tries could lose significant tariff revenue from reducing tariffs on imports from the European Union. Even assuming no trade diversion, an average country, in which tariff revenues are 2 percent of GDP and 40 percent of imports come from the Euro- pean Union, would lose tariff revenues equivalent to 0.8 percent of GDP (7­10 per- cent of government revenues) from eliminating tariffs on all imports from the Euro- pean Union if the same average tariff were paid on imports from the European Union as on those from the rest of the world.16 The revenue losses could be significantly greater in countries that are highly dependent on tariff revenues: countries in which tariffs account for 4 percent or more of GDP could lose revenues worth 1.5 percent or more of GDP (15­20 percent of government revenues) from eliminating tariffs on imports from the European Union. Any trade diversion from non-EU to EU suppliers because of the elimination of tariffs would lead to further loss of revenues. Busse, Borrmann, and Grossmann (2004) analyze the effects on the ECOWAS countries of preferential elimination of tariffs on all imports from the European Union using a partial equilibrium model. They find that imports from the European Union would increase by 5 percent (Guinea-Bissau) to 21 percent (Nigeria), rising about 9 percent in the median coun- try. The loss of government tariff revenues would range from a low of 0.3 percent of GDP (3.6 percent of government revenues) in Niger to a high of 4.1 percent of GDP (19.8 percent of government revenues) in Cape Verde. Mid-range countries such as Ghana and Senegal would experience revenue losses of 1.8­1.9 percent of GDP (10­11 percent of government revenues). These figures capture the ultimate effect of a fully phased-in elimination of duty on all imports from the European Union and thus are indicative of the upper limit of the magnitude of possible revenue losses. Excluding 10­20 percent of imports from the European Union from preferential liberalization could ease the fiscal shock, particu- larly if tariffs on imports that are important sources of revenue are not reduced. If, moreover, tariff reductions are phased in gradually over a 10­20 year period as envis- aged in the EPAs, the adjustment should be manageable (0.05­0.30 percent of GDP a year), except possibly in a few extreme cases that may require special attention. To protect their fiscal positions and maintain macroeconomic stability, African countries will need to reform their indirect tax systems so that revenues from value- added and nondiscriminatory excise taxes levied at equal rates on imports and domestic products replace forgone tariff revenues. Improvements in tax administra- tion could also yield large revenue gains in some countries. Countries that face par- ticularly large revenue losses or that have already successfully implemented a 182 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER value-added tax and have little scope for increasing the revenues from it may need to consider strengthening other components of their tax and revenue systems. In view of the unmet expenditure needs for government spending on health, edu- cation, and infrastructure and the debt burden and structural constraints to increas- ing government revenues, many African countries may also need to design the pref- erential reductions in tariffs on imports from the European Union in a way that limits the loss of revenues. Stevens and Kennan (2005) estimate that following a strategy of maximizing protection for domestic producers by excluding imports with the high- est tariff rates from the preferential liberalization would cause one-quarter of the ACP countries to lose less than 40 percent of their tariff revenues on imports from the European Union; three-fifths of the ACP countries would lose less than 60 per- cent of these revenues. The products that African governments may need to exclude to maintain govern- ment revenues may differ from those that would need to be excluded to protect domestic producers. Hence, in deciding which products to exclude, giving priority to protecting fiscal revenues, together with limiting the extent of the exclusions, could provide a healthy counterbalance to the demands of vested interests for continued protection. In addition, the loss of revenues from completely eliminating tariffs on 80­90 percent of imports from the European Union could be limited in the medium term by gradually lowering the preferential EU and MFN tariffs on these imports to a revenue tariff of 5 percent or so and maintaining the revenue tariff at this level for as long as possible while the domestic tax system is restructured and the tax adminis- tration improved. Restructuring indirect tax systems and strengthening tax administration are com- plex legal and administrative undertakings. They require some of the most demand- ing and time-consuming reforms facing African countries. Indirect taxes will also need to be harmonized within regional economic communities; and variations in rev- enue impacts within the same community will create additional complications. Restructuring of indirect tax systems is necessary in the long term to modernize rev- enue systems; but all countries seeking to integrate into the world trading system will need to do so sooner or later.17 Detailed country by country analysis of the revenue implications of preferential tariff reductions and parallel MFN liberalization will be essential for planning complementary reforms in domestic tax reforms. The required strengthening of tax administration and the restructuring of the indirect tax system needs to be started before tariff reductions on EU imports are implemented. In addition to having enough time to plan and implement the required reforms, African countries, par- ticularly the least developed ones, may require substantial technical and financial assistance from the European Union during the implementation period if elimina- tion of tariffs on imports from the European Union is not to undermine their fiscal positions. One particularly useful confidence-building measure would be to pro- vide back-up grant financing for tariff losses and related revenue shortfalls until the tax reform is completed, tax administration improved, and forgone tariff revenues replaced. RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 183 Designing EPAs for Development The fundamental condition for realizing the potential development benefits of EPAs is to use them as instruments for development rather than for commercial gain. Doing so poses challenges for both the European Union and Africa. For the European Union, the challenge will be to maintain the development focus of the EPA process in the face of pressures from business and political interests--a difficult task for trade negotiators accustomed to thinking in term of mercantilist commercial advantage and political deals. The European Commission will also need to simultaneously address complex development issues in multiple diverse regions and to effectively coordinate trade and development assistance. Failure to meet these challenges could lead to EPAs being abandoned or causing more harm than good, at least until the African countries are able to implement the necessary complementary reforms on their own. For the African countries, the key challenge is to use EPAs to accelerate reforms that are necessary in the long term for integrating their economies regionally and globally. Implementing free trade with the European Union will require a number of difficult policy and administrative steps by African countries, including undertaking parallel MFN tariff reductions to avoid costly trade diversion and monopoly pricing by EU suppliers, replacing forgone tariff revenues, and liberalizing internal trade within and among Africa's regional economic communities. There are many obstacles to expanding the production of tradable goods in Africa in addition to the merchandise trade­related reforms discussed here. The investment and supply response to merchandise trade­related reforms will be much greater if these reforms are followed by other investment climate reforms. The readiness of individual countries to use the EPA process to leverage a wide range of investment- climate reforms will be critical in determining the success of the EPAs.18 Four design issues are critical to the success of the EPA process. These are creating effective incentives to reform; allowing enough flexibility to deal with wide variations in regional and country conditions appropriately; phasing and sequencing a long list of complex policy actions; and designing, if feasible, pro-development provisions con- cerning trade in services, investment, and competition Incentives to Reform and Aid for Trade Many African countries are reluctant to undertake needed trade-related reforms. It is unclear whether the incentives in the current EPA design are sufficient to induce them to do so. Where trade agreements have been used to leverage a wide range of reforms (as in the accession of the Central and Eastern European countries to the European Union), the reforming countries perceived the trade agreement as attractive both eco- nomically and politically. Unfortunately, EPAs have become controversial, and many African countries are wary of their potential political costs. For EPAs to be politically attractive, they will need to include measures to accelerate growth by diversifying and expanding exports and raising investment levels. The European Union will need to include generous treat- ment of African exports in EPAs by liberalizing its restrictive rules of origin, extending 184 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER market access under EBA to non-LDCs in Africa, eliminating agricultural export sub- sidies and decoupling agricultural production support in products of particular inter- est to Africa in the Doha Round, and in general facilitating exports from Africa in whatever way possible. Furthermore, if EPAs are to be credible instruments for lever- aging reforms, they will need to be enforced. A development-friendly approach will thus need to be found for handling enforcement and dispute settlement problems. Providing attractive levels of aid for trade to support implementation of EPAs is also likely to be a critical incentive, particularly for LDCs. To reverse the current neg- ative perception of EPAs, resources provided as aid for trade will need to be per- ceived as additional and substantial. The additional aid for trade will need to be sig- nificantly greater in real terms than the amount needed to finance the transitional costs associated with EPAs and related reforms and to cover both budgetary assis- tance and financing of new investment. If the European Union were to offer an ambi- tious program of aid for trade as well as the trade and related benefits noted above, it might tip the balance of political economy incentives in favor of EPAs. Careful thought therefore needs to be given to the design of the aid for trade package as nego- tiations unfurl. Flexibility in Regional and Country Design EPAs were originally conceived as agreements between the European Union and cus- toms unions that have common external tariffs. However, the regional EPA negoti- ating groups in Africa are diverse combinations of customs unions, free trade areas, and independent countries with wide variations in trade policies and commitment to outward-oriented regional integration. Except for SACU, few, if any, are capable in the medium term of reducing internal border barriers or coordinating their external trade regimes to the extent required in a full customs union with revenue pooling and no customs barriers at intraregional borders. Hence, EPAs will probably have to per- mit flexible combinations of customs unions and free trade agreements. Country conditions within the regional groupings are also diverse. All four EPA groups include both LDCs and non-LDCs, as well as small economies alongside one or two dominant large ones. Commitment to both external and intraregional liberal- ization differs similarly. The political obstacles to reform (such as civil unrest or strong vested interests) may be so great that some countries, or even subregions, may have to fall back on the GSP or EBA initiative for access to the EU market until they are ready to implement EPA-related reforms. Progress in these diverse country circumstances is likely to require that EPAs allow for some differences in the trade regimes and the pace of reform in the same EPA group, at least in the initial phases. Timing and Sequencing of EPA-Related Reforms The EPA design envisages addressing a wide range of difficult development issues in multiple sectors in 44 countries in four diverse regional groups over a period of 12­20 years. The 2007 deadline for completing the EPA negotiations is rapidly approaching. Even for countries with substantial institutional capacity, dynamic RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 185 leadership, and strong political support for reforms, the design and implementation of such comprehensive programs will be a challenge. For countries with serious capacity constraints and tepid political support for reform, the list of reforms envis- aged is daunting. Hence a single mega-EPA agreement with a 12- to 20-year imple- mentation period may prove too demanding to negotiate and implement even for the most committed regions. A series of agreements, with careful attention paid to the timing and sequencing of reforms, may be more realistic than a single compre- hensive EPA. The reform process will have to be carefully steered to take advantage of the dynamism provided by the EPA negotiations and their deadlines while allowing real- istic amounts of time for policy design and implementation and avoiding initiatives that are not institutionally or politically sustainable. With appropriate program design, the envisaged 12- to 20-year implementation period appears reasonable. However, given the depth of some of the controversies and the need for more preparatory work on some key issues, the 2007 deadline for completing negotiations may eventually need to be extended for some countries, particularly if trade in serv- ices, investment, and competition policy are to be included. Designing Development-Friendly Provisions on Trade in Services, Investment, and Competition Policy The investment and supply responses to the merchandise trade reforms will be much greater if these are followed by actions in other areas. It may be possible to address some supply-side and behind-the-border reforms by including trade facilitation, trade in services, investment, and competition policy in the EPA process. Despite the con- troversy surrounding the Singapore issues, this possibility merits serious examination. In-depth analytical work on how to design such measures in a development-friendly way needs to be undertaken with some urgency if they are to be included in EPAs. Conclusion: Rewards, Risks, and Alternatives In many African countries the political economy environment for trade liberalization is difficult, because protectionist interests are powerful and well-organized. Political enthusiasm for unilateral liberalization is limited. Hoping for a "round for free," African countries have yet to really engage in the multilateral WTO discussions. Hence, by itself, the Doha Round may not lead to any reductions in applied MFN tariff rates in Africa. The prospects for regional integration efforts in Africa are also uncertain. Thus in the absence of EPAs, progress in the medium term is likely to be confined to a handful of countries. EPAs and the current regional trade areas in Africa may not be optimal arrange- ments from an economic point of view, but strong political forces support their con- tinuation. Although controversial, the EPA process has already gathered substantial momentum in both the European Union and Africa. It is catalyzing an active debate over, and some actions on, trade reform and regional integration in Africa. 186 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER Implementation of EPAs involving liberalization of both intraregional and external trade together with investment climate reforms may be more feasible to engineer polit- ically than unilateral or intraregional trade liberalization by themselves, because a quid pro quo, in the form of improved access to the EU market and aid for trade, would be involved. EPAs could thus offer Africa an important opportunity to accel- erate trade and growth. EPAs could be beneficial economically for Africa if they can be used to leverage important policy reforms. The European Union hopes that the EPA process can rein- force current reform efforts by strengthening public and private sector reform advo- cates and deepening the regional integration process. The need for African countries to liberalize imports from the European Union to improve their access to the EU mar- ket provides an opportunity for them to integrate into the global economy, strengthen regional integration in Africa, accelerate trade and related reforms under potentially favorable conditions, and lock in these reforms in a way that makes them credible to the rest of the world, investors, and donors alike.19 The EPAs' negotiat- ing schedules and deadlines create an impetus for global and regional integration. Technical and financial assistance from the European Union could provide more gen- erous support than is often available for countries undertaking trade reforms. The EPA process also entails some serious risks for Africa. EPAs could go astray because, despite the European Union's good intentions, lobbying by politically pow- erful business interests at critical moments could lead to commercially advantageous decisions for the European Union rather than decisions that are conducive to devel- opment. Initial negative reactions from the European Commission's trade directorate to liberalization of EU rules of origin and complementary MFN tariff reductions by African countries are not good omens. The management of a truly development- oriented EPA process effectively combining trade and aid could prove too complex. Partial preferential liberalization of imports from the European Union under EPAs could hurt Africa, because of the distorted structure of protection and the increased antiexport bias it would create and the likelihood of significant costly revenue losses, trade diversion, monopoly pricing by EU suppliers, and continued inefficiency of import-competing industries. Some African countries may not be able to implement the necessary parallel MFN liberalization or the required accompanying domestic reforms as a result of lack of political support or limited institutional capacity. These may suffer from losses of fiscal revenue, trade diversion, and monopolistic pricing until they do so. The EPA process could also end up being abandoned, with the acri- mony surrounding its abandonment poisoning the environment for trade and invest- ment-climate reforms, at least temporarily. The Cotonou Agreement envisages a fairly uniform approach to EPAs, with broadly similar agreements by six regions more or less simultaneously. However, because of the diversity of initial conditions, considerable flexibility and selectivity is likely to be needed in the treatment of both the EPA regional groups and the countries within them. Moreover, to avoid being limited to the minimum level of progress that is pos- sible in the most difficult subregions, it may be desirable to confine EPAs to those regions and countries that are most committed to the EPA process and can use it effec- tively to leverage their reform programs, letting others fall back on the GSP and EBA RISKS AND REWARDS OF REGIONAL TRADING ARRANGEMENTS IN AFRICA | 187 initiative for access to the EU market until they are ready to reform. Bilateral free trade agreements, or other temporary arrangements, with individual non-LDCs may also need to be considered for countries that would be most adversely affected by dropping back from Cotonou preferences to GSP. Notes 1. Figures for the European Union are for the EU­15. 2. In this paper the terms pro-development and development friendly refer to policies that accelerate the growth of real GDP and global and regional trade integration in Africa. 3. Regional economic community (REC) is a term used in Sub-Saharan Africa to describe customs unions, free trade areas, and other regional integration arrangements. In other regions, the term regional trade area/agreement (RTA) is applied to the same types of organizations. The two terms are used interchangeably here. 4. When a hub country or region (such as the European Union) signs a free trade agreement with small countries (the spokes) that do not sign free trade agreements among themselves, the hub country tends to benefit more, because it has free access to all markets whereas the spokes have free access only to the hub market. This "hub-and-spoke effect" increases the incentive for exporters to invest in the hub country rather than in the spokes in order to serve all of the markets. 5. The General Agreement on Tariffs and Trade (GATT) and the WTO have been notified by their members of more than 100 free and preferential trade agreements under the enabling clause and Article XXIV. None of these agreements has been found to violate GATT­WTO rules. 6. See Yang and Gupta (2005) for a recent review of regional trade initiatives in Sub- Saharan Africa. 7. Implementation of full market access was immediate, except for transition periods for bananas, rice, and sugar. Tariff-rate quotas restricting LDC exports of these products to the European Union are being phased out over eight years. 8. Such benefits could come from a combination of liberalization of the European Union's rules of origin, aid for trade to deal with adjustment problems and support the growth and diversification of exports, trade in services, and establishment of a favorable stable long-term trade relationship with the European Union. 9. The European Union has recently adopted a GSP+ scheme, which provides additional preferences for countries that comply with a list of international conventions on human rights, the environment, and other issues. The GSP+ preferences are not as attractive as EBA preferences. 10. This paper uses the term "EBA access" to refer to tariff-free, quota-free market access for everything but arms. The European Union's EBA initiative is one way of providing EBA access to its markets. Such access could also be provided by EPAs, as recommended here. 11. A "change of tariff heading" rule would allow garment exporters to use third-country fabrics, which have been a key to the expansion of African garment exports under AGOA. 12. The requirement to liberalize African imports from the European Union is referred to in the popular debate over EPAs as the "reciprocity issue." 13. This finding is somewhat surprising given the existence in Africa of three customs unions with common external tariffs. 14. Other strategies for minimizing the distortions caused by excluding some products from full liberalization are limiting the extent of the exclusions and focusing the exclusions on imports generating the most revenues, as discussed below. For a discussion of additional 188 | LAWRENCE E. HINKLE AND RICHARD S. NEWFARMER measures for reducing the distortion effects of preferential liberalizations, see Schiff and Winters (2003). 15. See Hinkle, Herrou-Aragon, and Kubota (2003) for an analysis of tariffs in Sub-Saharan Africa. 16. It is possible that in some countries the average tariff paid on imports from the European Union is lower than that paid on imports from the rest of the world, because aid-financed imports are typically exempt from import duties and the European Union and its mem- ber countries are Sub-Saharan Africa's largest source of aid. 17. Some African countries (such as those in UEMOA) have already embarked on this process in the course of ongoing reform programs. 18. For example, the improvement in market access under the EBA initiative does not yet appear to have led to substantial increases in African exports, in part because of supply constraints to export diversification and expansion. 19. In addition, inclusion of trade in services, investment, trade facilitation, and aid for trade in EPAs, if feasible, could provide an opportunity to use the EPA process to address a broad range of obstacles to a supply response in trade-related sectors. References Brenton, Paul. 2003. "Integrating the Least Developed Countries into the World Trading Sys- tem: The Current Impact of European Union Preferences under Everything But Arms." World Bank Policy Research Working Paper 3018, World Bank, Washington, DC. Busse, Mathias, Axel Borrmann, and Harold Grossman. 2004. "The Impact of ACP/European Union Economic Partnership Agreements on ECOWAS Countries: An Empirical Analysis of Trade and Budget Effects." Hamburg Institute of International Economics, Hamburg, Germany. Commission for Africa. 2005. Our Common Interest: Report of the Commission for Africa. London. Available at www.commissionforafrica.org. Hinkle, L., and M. Schiff. 2004. "Economic Partnership Agreements between Sub-Saharan Africa and the EU: A Development Perspective on their Trade Components." World Bank, Africa Region, Washington, DC. Hinkle, L., A. Herrou-Aragon, and K. Kubota. 2003. "How Far Did Africa's First Generation Trade Reforms Go? An Intermediate Methodology for Comparative Analysis of Trade Policies." Africa Region Working Paper No. 58, World Bank, Washington, DC. Schiff, M., and L. Alan Winters. 2003. Regional Integration and Development. Oxford: Oxford University Press. Stevens, C., and J. Kennan. 2005. "EU­ACP Economic Partnership Agreements: The Effects of Reciprocity." Institute of Development Studies Briefing Paper, Sussex, United Kingdom. World Bank. 2005. Global Economic Prospects 2005: Trade, Regionalism, and Development. Washington, DC. Yang, Y., and S. Gupta. 2005. "Regional Trade Arrangements in Africa: Past Performance and the Way Forward." IMF Working Paper WP/05/36, Washington, DC. Comment on "Risks and Rewards of Regional Trading Arrangements in Africa: Economic Partnership Agreements between the European Union and Africa," by Lawrence E. Hinkle and Richard S. Newfarmer LÉONCE NDIKUMANA Since the independence era of the 1960s, African policy makers have regarded regional integration as an essential component of strategies for achieving fast and sustainable economic development. Intra-African regional integration and trade arrangements with other regions are seen as a means of addressing two important structural constraints to economic growth faced by all the economies on the conti- nent: the small size of domestic markets and the limited access of African products to developed countries' markets. As a result, subregional trade arrangements have proliferated over the past decades, while the continent has aggressively pursued trade arrangements with Europe, the United States, and Japan. The objective of these ini- tiatives is to accelerate economic growth, increase the continent's attractiveness to foreign investors, and ultimately allow African economies to maximize the benefits of globalization. However, as the paper by Hinkle and Newfarmer indicates, impor- tant structural and organizational issues need to be addressed before the continent can reap the benefits of economic integration and trade liberalization. The proliferation of regional trade arrangements has not been accompanied by meaningful gains in increased intra-African trade (Yeats 1998). Although intra-African trade has grown since the 1990s, this trend cannot be attributed to regional arrange- ments, as trade within regional entities remains low. Intraregional trade is also uneven, in two respects. First, it tends to be dominated by a few large economies, such as South Africa in the Southern African Region, Kenya in the East, and Côte d'Ivoire, Ghana, and Nigeria in the West. Second, there is an imbalance between imports and exports. For example, Côte d'Ivoire's intraregional imports are only half its intraregional exports; for Nigeria the ratio is six to one in favor of exports (Yeats 1998). The implication is that losses in tariff revenue due to liberalization will be unevenly distributed, which may con- stitute another hurdle in the trade negotiation process. Intra-Africa trade remains dom- inated by primary commodities, with petroleum products representing one-third to one- half of total trade, followed by agricultural products (18 percent). Manufactured products, machinery, and equipment represent a minute fraction of intra-African trade. Léonce Ndikumana is associate professor of economics at the University of Massachusetts, Amherst. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 189 190 | LÉONCE NDIKUMANA These patterns are the result of several structural features of production systems in Africa, the most prominent being the lack of diversification and limited product differentiation across countries. Intra-African trade has not expanded mainly because African countries tend to produce the same products, generally unprocessed primary goods. Do Economic Partnership Agreements (EPAs) offer a solution to the chronic prob- lems faced by African countries in their quest for integration in the global economy? The objectives of EPAs are commendable. They seek to help African, Caribbean, and Pacific countries integrate into the world economy and promote sustainable develop- ment through increased market access. Unless these initiatives are properly designed and executed, however, their benefits may fall short of expectations, for several rea- sons. First, there is a risk that implementation of EPAs may deepen the continent's dependence on primary product exports while reducing the chances of diversification of the production systems. The European Union accounts for almost two-thirds of Africa's total exports (Ianchovichina, Matto, and Olarreaga 2001). The bulk of these exports are oil and agricultural products. The EPAs are a welcome development in that they allow African producers to tap the traditionally protected Western markets for agricultural products. However, there are reasons to doubt the long-term bene- fits of this development. The ability of African economies to raise their agricultural production is limited. Moreover, these economies have no control over the prices of these products in international markets. The scope for expansion of quantitative gains is limited, and the costs arising from instability in export revenues may be con- siderable. Therefore, in addition to negotiations over access to markets in Europe, African countries need to emphasize mechanisms for hedging or smoothing fluctua- tions in revenues from primary commodity exports. Second, implementation of EPAs will result in losses of tariff revenues. These losses are likely to be uneven. An important question is whether African countries will be able to compensate for these losses through gains from increased trade. The bulk of African countries' exports are in primary commodities. As tariffs on these products are removed, the prices for African products in EU markets will decline. It is not clear whether the increase in demand in response to price reductions will be enough to com- pensate for price decline. As they remove tariffs on EU products, African countries will lose tax revenue, which will need to be made up with gains from other sources of tax revenues. The empirical literature is not particularly encouraging: low-income countries are generally unable to recover from other sources the revenues lost through liberalization (Baunsgaard and Keen 2005). If countries are not able to recover the revenues lost to liberalization, they will have to undertake fiscal adjustments that are both economically painful and potentially politically costly. Another issue raised by EPAs is the distribution of welfare gains between Africa and the European Union. Removal of tariffs on EU exports to Africa will result in direct gains for European exporters. However, given that EU products compete with products from other regions that may not benefit from similar preferential tariff treatment, market prices for imported goods in Africa may not decline substantially. Therefore the gains for African consumers may be limited. COMMENT ON HINKLE AND NEWFARMER | 191 As the paper by Hinkle and Newfarmer and other studies (de la Rocha 2003; Babarinde 2003) in this area point out, EPAs may be inconsistent with existing regional arrangements. A myriad of regional initiatives already exists in Africa, where trade arrangements overlap and often conflict. It is not always clear whether the marginal gains from membership in an additional regional arrangement are sig- nificant enough to justify the economic and administrative costs. Nonetheless, the implementation of EPAs has to contend with this reality of African regional integra- tion. If implementation of EPAs could induce African countries to harmonize and streamline conventions in regional arrangements, this would indeed be a major bonus from this initiative. Harmonizing conventions across regional entities would also help reduce the unevenness in gains from integration. The success of EPAs in integrating African economies into global markets depends partly on the supply response in member countries. The gains from EPAs will be very limited if the initiative results in locking African economies into their traditional role as suppliers of low-value-added primary products. Implementation of EPAs will pro- mote growth in Africa only if they induce greater diversification of production, higher investment, and increased productivity. African countries cannot hope to out- source domestic economic development policy through integration with the Euro- pean Union. The main challenge remains the ability of governments to create a sta- ble macroeconomic environment and sound regulatory framework that facilitates improvement of the investment climate. Policy makers cannot be distracted from the basic condition for sustained economic development, which is the ability to achieve a high and sustained level of domestic investment and savings as well as improved pro- ductivity of human and physical capital. Countries that can achieve these goals will reap higher gains from EPAs, because they will have more to offer to global markets. References Babarinde, F. 2003. "Regionalism and Economic Development." In African Economic Devel- opment, ed. Emmanuel Nnadozie, 473­97. New York: Academic Press. Baunsgaard, T., and M. Keen. 2005. Tax Revenue and (or?) Trade Liberalization. Interna- tional Monetary Fund, Working Paper WP05/112, Washington, DC. De la Rocha, M. 2003. The Cotonou Agreement and Its Implications for the Regional Trade Agenda in Eastern and Southern Africa. World Bank Policy Research Paper 3090, Wash- ington, DC. Ianchovichina, E., A. Mattoo, and M. Olarreaga. 2001. Unrestricted Market Access for Sub- Saharan Africa: How Much Is it Worth and Who Pays? World Bank Policy Research Paper 2595, Washington, DC. Yeats, A. J. 1998. What Can Be Expected from African Regional Trade Arrangements? Some Empirical Evidence. World Bank Working Paper 2004, Washington, DC. Investment Climate Business Environment and Comparative Advantage in Africa: Evidence from the Investment Climate Data BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN Macroeconomic and microeconomic data highlight the impact of high costs in reducing the productivity and competitiveness of African manufacturing. Data on relative prices suggest that African economies are relatively high cost. Firm-level data from Investment Climate Assessments conducted between 2000 and 2004 corroborate this finding, highlighting the importance of high indirect costs in depressing the performance of African firms relative to those in other countries when performance is expanded from a "factory-floor" concept to a broader "net productivity" measure. Business sectors heavily segmented by size, produc- tivity, and ethnicity potentially pose problems for reforms. Potential approaches to reform and development assistance must take these complications into account. Developing countries, traditionally exporters of primary goods, have increased their exports of manufactured goods in recent decades: in 2002, 60 percent of their exports were manufactured goods. The technology content of developing country exports has also been rising rapidly, especially in Asia, which has seen the emergence of dynamic regional trading networks. Sub-Saharan Africa (hereafter Africa) has lagged in this process of economic diver- sification; except in Mauritius and South Africa, manufacturing and processing capac- ity remains modest. Slow progress in economic diversification and technological upgrading has been associated with weak private sector development, lagging incomes and development outcomes, and marginalization of Africa on the world trading stage. This article draws on a number of firm surveys undertaken for the World Bank's Investment Climate Assessments to better understand some of the factors underlying Benn Eifert is a PhD student in the department of economics at the University of California­Berkeley. Alan Gelb is Director of Development Policy at the World Bank. Vijaya Ramachandran is assistant professor in the department of public policy at Georgetown University and visiting fellow at the Center for Global Development. Data used in this paper were collected by the Regional Program on Enterprise Development in the Africa Private Sector Group of the World Bank. The authors would like to thank Demba Ba, François Bourguignon, George Clarke, Mary Hallward- Driemeier, Lawrence Hinkle, Giuseppe Iarossi, Phil Keefer, Michael Klein, Taye Mengistae, Todd Moss, Léonce Ndiku- mana, Guy Pfeffermann, Steven Radelet, Manju Shah, Gaiv Tata, an anonymous referee, and seminar participants at the World Bank, the Center for Global Development, and Cornell University for excellent comments and suggestions. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 195 196 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN Africa's slow industrial growth. A number of studies have used firm surveys to ana- lyze productivity determinants in Africa (Biggs, Shah, and Srivastava 1995; Bigsten and others 2000; Fafchamps 2004; Mengistae and Pattillo 2002; see also the exten- sive Regional Program on Enterprise Development studies that have been conducted since 1995, available at www.worldbank.org/rped). It is now possible to combine expanded coverage in Africa with comparisons with other low-income countries that have managed to effect the transition to manufacturing exporter status. This study does not cover all factors that affect competitiveness. It attempts to analyze the cost structure of firms in Africa relative to those in competitor regions. It also interprets comparative firm data to consider why reform has been difficult and slow in Africa. The first section of the paper reviews three theories of comparative advantage that offer different perspectives on Africa's slow diversification. The first stresses the role of factor endowments in shaping economic structure and the composition of trade. The second focuses on the ability of countries to provide public goods to the invest- ment community in the form of a stable and low-cost business climate. The third emphasizes the important role of firm entry in creating a critical mass of industries able to reap knowledge and other dynamic scale externalities. Taken together, the second and third theories suggest powerful externalities associated with conglomer- ation effects, but they also suggest that these externalities will be hard to achieve in the absence of a low-cost business environment. Deviations of countries' price levels relative to those derived from predicted purchasing power parity (PPP) conversion factors confirm that Africa indeed appears to be a high-cost region. Section 2 turns to the firm-level investment climate surveys to assess the perform- ance gap between Africa and its competitors. As well as estimating conventional total factor productivity, it defines a concept of net total factor productivity (TFP) and con- siders the contributions of factory-floor productivity, indirect costs, and certain busi- ness environment­related losses to overall differentials in competitiveness. The data suggest that these losses and indirect costs are crucial determinants of competitiveness. Section 3 considers the implications of ethnic and size/efficiency cleavages in Africa's business sectors for the political economy of business-related reforms. Doing Business 2005 (World Bank 2004a) gives Africa low ratings on business climate indi- cators and reform relative to other regions. If, as suggested by the World Develop- ment Report 2005 (World Bank 2004b), the business climate is so important for growth and development, why have African countries been so slow to improve it? Part of the reason is the severity of physical constraints in Africa, but equally impor- tant is the configuration of political interests for and against reforms in areas related to the business climate. Section 3 considers survey evidence from the perspective of ethnicity, highlighting the fracturing of business interests along ethnic and domestic and foreign lines, as well as the tradeoffs for firms (especially larger ones) between the gains from a better business climate and the losses from the competitive entry that improvements are likely to encourage. These factors weaken the ability, and perhaps also the motivation, of the business community to press for a better business climate, compounding the ambivalent attitude toward business expressed in Afrobarometer surveys and reflected in statements by many prominent African officials. These fac- tors raise the question of whether to attempt reforms across the board or to sequence BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 197 them to create pockets of opportunity that in turn can build constituencies for wider reforms. Section 4 concludes with a summary of policy implications as well as thoughts on areas for further research. Comparative Advantage and Costs in Sub-Saharan Africa Three Approaches to Comparative Advantage Development and structural change are closely associated. As Chenery and Syrquin (1975) and others note, growth involves introducing new, higher value-added activ- ities and products rather than simply expanding production of old ones. In the ini- tial stages, this process involves the relative contraction of low-productivity agricul- ture and the rise in the share of industry. Overall growth within the industrial sector is the aggregation of repeated industry cycles of takeoff, maturation, and stagnation (or migration to less advanced countries), with more productive economies advanc- ing up the technological ladder. Trade theory is central to understanding economic structure and structural change, because countries tend to export goods they can pro- duce most cheaply and efficiently relative to other countries. Because sustained eco- nomic growth is driven by the emergence of new economic activities--rather than the perpetual scaling up of old activities--trade theory is key to understanding growth. Wood and Berge (1997) and Wood and Mayer (2001) compare Africa's factor endowments with those of other regions. With capital assumed mobile in the long run, relative endowments of skills and land (resources) per capita are shown to have a strong relationship with the composition of exports. Countries higher up the skills and land spectrum export more manufactures relative to processed or primary goods and a larger proportion of their manufactures are higher technology. A pessimistic view based on these results would argue that Africa's scant human capital and rich natural resource base ensure that manufactured exports will always be unprofitable. These theories do not, however, fully account for Africa's low income level despite its resource abundance, nor do they explain the dynamic path of factor accumulation (pervasive financial and human capital flight) that has shaped Africa today.1 Krug- man (1980, 1981, 1983) shows that comparative advantage is also a function of dif- ferences in productivity and costs that do not derive from relative factor abundance. The main effects here can be expressed through two approaches, one relating to busi- ness environment factors and the other to dynamic economies of scale. The business environment may be defined as the nexus of policies, institutions, physical infrastructure, human resources, and geographic features that influence the efficiency with which firms and industries operate.2 At the firm level, it directly influ- ences the costs of production; at the industry level, it often relates to market structure and competition. Its impact is felt more heavily in traded sectors that are not partic- ularly intensive in natural resources (manufacturing, high-value services) than in pri- mary production and extractive resource sectors, because the former tend to require more intensive use of "inputs" of logistics, infrastructure, and regulation (Collier 2000). The combination of macroeconomic instability, crime and poor security, a weak and 198 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN politicized financial system, shoddy local roads and electricity systems, high trans- port costs, and predatory local officials will have relatively little influence on the pro- ductivity and costs of offshore oil industries, but it will be devastating for small- and medium-scale manufacturing. Even efficient firms able to transform inputs into out- puts with high efficiency and low "factory-floor" costs can be driven out of business by a poor business environment. Dynamic economies of scale generated by learning processes, network effects, and industry-specific spillovers represent a further elaboration beyond classical produc- tion and trade theory (Krugman 1980, 1991). Evidence suggests that dynamic scale economies play a considerable role in shaping the structure of production, as illus- trated by path dependence in the development of individual industries, the "lumpy" nature of growth in a particular product across countries (for example, high degrees of specialization in narrow industrial lines; see Burgess and Venables 2004) and within countries (for example, urbanization, industrial clusters, and path dependence in the development of individual industries; see Krugman 1991). But individual firms do not internalize the social value of the potential economies of scale from their entry into a particular industry. Thus entry, investment, and the development of new industries still depend on the quality of the business environment, good policies, and sound infrastructure (Collier 2000); incentives provided by competition in an appro- priate institutional setting (Grossman and Helpman 1990); and geographic advan- tages and disadvantages (Krugman 1991).3 Business environments do not have to be perfect, but they have to be good enough on a number of crucial dimensions to stim- ulate investment and competition sufficient to launch the self-reinforcing process of industrial growth. This theoretical framework offers several insights that go beyond those of classical trade theory for understanding patterns of trade and industrialization. First, while many resource-rich countries have been unable to move past primary products, other countries with good policies (Australia, Chile, Malaysia, the United States) have built high-value-added resource processing industries in the early stages of industrialization, using these as a springboard to even higher value activities. Second, the broad factor- based specialization predicted by classical trade theory does not map well onto reality. Countries with similar factor endowments often export different products, often to one another. Studying U.S.-bound exports from Bangladesh, the Dominican Republic, Honduras, the Republic of Korea, and Taiwan (China) at a very fine level of disaggre- gation, Hausmann and Rodrik (2002) find that exports are characterized by special- ization in a narrow range of activities with surprisingly little overlap across countries. African examples of new industries such as Kenya's horticulture-floriculture sector and the garment sectors of Madagascar and Lesotho also suggest the importance of indus- trial clustering. Pessimistic evaluations of the prospects of diversification and growth in resource- rich countries miss part of the story.4 African countries often suffer from poor poli- cies, weak institutions, and shoddy infrastructure (see Collier and Gunning 1997; Eifert and Ramachandran 2004). High transport costs and sparseness are also impor- tant (Venables and Limao 1994; Winters and Martins 2004): gross domestic prod- uct (GDP) per square kilometer in Africa (excluding South Africa) is one-tenth that BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 199 of Latin America and one-twentieth that of India. Manufacturing value-added per hectare (excluding South Africa) is only 1.2 percent that of China. Moreover, the GDP of the median country in Africa is barely US$3 billion, suggesting that efforts to overcome high regulatory costs will not be rewarded by large market potential. These factors increase costs, depress productivity, discourage investment, and obstruct the self-reinforcing processes of growth, clustering, and dynamic economies of scale. Within Africa productivity is strongly related to exports, both as a cause and as a consequence (Bigsten and others 2000; Söderbom and Teal 2003). But most African firms are simply not productive enough to export manufactures. Africa's factor endowment may be consistent with competitiveness in a variety of labor- intensive natural resource processing industries, but most African countries have been unable to take even this step toward higher value-added processing. Are African Countries Really High Cost? The Macroeconomic Evidence PPP conversion factors--expressed as the ratio of a country's GDP measured in mar- ket prices to its income measured in PPP prices--provide an estimate of a country's aggregate price level relative to those of other countries. This ratio ranges from less than 0.2 in some poor countries to 1.0 or higher in OECD countries (table 1). Unfor- tunately, the price deflators for PPP calculations were last updated for the 1993­6 period; some countries may look different today. Furthermore, although survey cov- erage was widespread in Africa and other regions, global links were weak, as were links for some important comparators, notably China and India. The variance in PPP deflators is therefore subject to considerable error and potential bias, in unknown directions. A new round of data collection is under way, but it will take some time for this effort to be completed.5 PPP conversion factors are closely related to income levels due to the "Balassa effect"--the fact that productivity gaps between rich and poor countries are larger in tradable sectors than in nontradables, while rich countries also have higher demand for nontraded goods and services. Nontraded goods and services therefore tend to be relatively more costly in rich countries. International trade tends to equalize prices of traded goods, so that aggregate price-level differences tend to be driven by the prices of nontraded goods, although the final prices of most tradable goods will also be affected by trade restrictions and the prices of inputs such as port services and domes- tic transport. For manufacturing firms, higher traded goods prices affect competitive- ness through the cost of imported capital equipment and raw materials; higher non- traded goods prices do so through a wide range of indirect costs (transport, logistics, electricity, telecommunications, rent, security, and so forth). With per capita incomes averaging roughly US$300 in 2005, Africa's poor economies have only four-fifths the income level of South Asia and half that of East Asia. But using PPP conversion factors reveals that their costs are 75 percent higher than in South Asia and 35 percent higher than in East Asia. Costs in Africa's poor countries are 31 percent higher than predicted, whereas costs in China are 20 percent below and costs in South Asia (primarily India) are 13 percent below their predicted levels (table 2).6 These results are broadly compatible with the estimates from 200 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 1. Ratio of Exchange Rate to Purchasing Power Parity Conversion Factor, by Region, 1993­6 Exchange rate to Actual: Income per capita purchasing power parity predicted ratio at market prices Region conversion factor (Balassa gap) (dollars) OECD 1.19 1.07 26,500 Latin America and the Caribbean South America 0.64 1.16 4,000 Central America 0.46 0.93 2,850 Caribbean 0.55 1.07 3,200 Middle East and North Africa 0.42 0.93 2,200 Europe and Central Asia 0.42 0.90 2,450 East Asia and the Pacific All 0.29 0.91 750 China 0.23 0.80 550 South Asia 0.22 0.87 375 Sub-Saharan Africa All 0.37 1.07 550 Poor countriesa 0.31 1.28 300 Source: World Bank World Development Indicators, 2004. Note: A value of 1.0 implies that cost levels are equal to those predicted by the Balassa trend line relating income level to purchasing power parity ratios. Regions with costs or prices higher than predicted have values above 1.0. a. Excludes Botswana, Cape Verde, Mauritius, Namibia, and South Africa, all of which are well-managed middle- income countries. Sala-i-Martin and Artadi (2003), which show that capital costs are one-third higher than world levels in Africa and one-third lower than world levels in Asia. These results may also reflect the price of land, which is often surprisingly high in Africa. But they suggest that cost divergences extend beyond capital goods to encompass a wide range of goods and services. Moving to the country level, while income explains 90 percent of the cross- country variation in price levels, some countries lie substantially above or below the regression line. Even the noisy PPP data show some systematic patterns. The most deviant outlier is the Republic of Congo, an oil-producing country with a record of political instability, poor governance and economic management, low capacity, and poverty and high inequality, located relatively far from major inter- national markets. Despite the country's modest per capita income (US$750 at mar- ket prices), market prices for goods and services are close to OECD levels (the exchange rate to purchasing power parity ratio is 0.72). Many strong performers lie well below the regression line. Most of these countries have effected the transition to manufactured exporter status. They have created a critical mass of industrial activities able to take advantage of cheap local inputs to TABLE 2. Costs, Export Structures, and the Ratio of Exchange Rates to Purchasing Power Parity in Countries with per Capita GDP of Less Than US$1,000, 1993­6 Manufacturing Exchange rate/ Balassa Country (percent of exports) purchasing power parity gap Major exporters 69.0 0.25 0.78 Bangladesh 87.2 0.26 0.86 China 84.4 0.24 0.73 Haiti 76.7 0.26 0.76 India 72.4 0.21 0.68 Kyrgyz Rep. 38.4 0.33 1.06 Nicaragua 33.7 0.19 0.60 Pakistan 83.8 0.27 0.96 Sri Lanka 72.5 0.29 0.85 Ukraine 67.8 0.24 0.65 Vietnam -- 0.20 0.67 Moderate exporters 19.0 0.30 1.01 Azerbaijan 20.0 0.25 0.77 Comoros 33.4 0.29 0.91 Ethiopia 11.2 0.18 1.18 Gambia, The 19.6 0.23 0.78 Ghana 13.2 0.23 0.75 Guinea 20.1 0.32 1.06 Honduras 23.6 0.29 0.91 Kenya 26.4 0.34 1.12 Madagascar 15.1 0.29 1.08 Moldova 20.3 0.30 1.09 Mongolia 10.2 0.30 0.97 Mozambique 16.7 0.25 0.97 Rwanda 13.8 0.24 0.92 Uganda 13.0 0.27 1.03 Negligible exporters 4.0 0.51 1.48 Angola -- 0.31 0.79 Benin 3.7 0.47 1.54 Burkina Faso 4.4 0.30 1.02 Burundi 1.3 0.24 0.96 Cameroon 8.0 0.41 1.24 Congo, Rep. of 2.7 0.72 2.30 Côte d'Ivoire 6.1 0.56 1.64 Guinea-Bissau 0.2 0.23 0.95 Malawi 8.6 0.48 1.21 Mali 2.1 0.40 1.36 Mauritania 0.4 0.30 0.72 Niger 0.5 0.29 1.02 Nigeria 1.1 0.37 1.19 Papua New Guinea 4.0 0.45 1.17 Sudan 2.8 0.23 0.66 Togo 5.8 0.25 0.80 Yemen, Rep. of 0.6 0.71 1.38 Zambia 7.0 0.48 1.79 Source: World Bank World Development Indicators, 2004. -- Not available. Note: Export data are not available for Angola or Vietnam. Angola's manufacturing sector accounts for 4 percent of GDP. Vietnam's manufacturing sector accounts for 20 percent of GDP. A value of 1.0 implies that a country's cost or price level is as predicted by the regression line relating price and income levels. 201 202 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN reduce costs for other firms and consumers alike. This pattern also holds within Africa: price levels in Africa's better performing countries, including Mauritius and South Africa, which have shifted from primary to manufactured exports, are also close to predicted values. Countries above the line are typically weak performers, and most are still at the primary exporting stage. The pattern that emerges from table 2 is quite strong. It suggests that countries with high costs have low efficiency in producing a wide range of nontraded goods and services that serve directly as intermediate inputs to production or that underpin the efficient provision of services, such as finance, which are essential for production. As countries move down the efficiency frontier, the costs to manufacturing firms of obtaining these inputs rise, squeezing their value-added between rising overall costs and the price at which their products can be imported. Few firms can insulate them- selves from high domestic costs. In extreme cases, the economy retreats into a com- bination of subsistence agriculture and concentrated hydrocarbon or hard-mining activities that are able to shield themselves from economywide effects.7 If the costs facing many African firms are even close to these estimates, they will affect their competitiveness. In addition, to the extent that households and workers also face high prices, the market value of their wages and incomes overstates their purchasing power relative to households in other poor, but low-cost, countries. African firms may therefore face relatively high wage costs for firms, but African workers have relatively low purchasing power. Of course, given the limited accuracy of the PPP data, this highly aggregated exercise is only indicative. The next section, which analyzes microeconomic evidence at the firm level, throws more light on the nature of high business costs in Africa. Firm Costs and Productivity: Evidence from Investment Climate Surveys This section examines the microeconomic data gathered by the World Bank's invest- ment climate firm surveys between 2001 and 2004.8 It shows that the losses and indirect costs estimated in these surveys represent a significant drag on manufactur- ing competitiveness that often escapes attention in the literature on growth and firm performance.9 The Survey Data The cross-sectional data cover nine African countries (Eritrea 2002, Ethiopia 2002, Kenya 2003, Mozambique 2001, Nigeria 1999, Senegal 2003, Tanzania 2003, Uganda 2003, and Zambia 2002) and six comparators (Bangladesh 2002, Bolivia 2000, China 2003, India 2002, Morocco 2000, and Nicaragua 2003). The data cover about 7,000 firms in six industry categories (textiles, garments, and leather; food and beverage pro- cessing; metals and machinery; chemicals and paints; wood and furniture; and other). About 2,700 of the firms are in Africa, 1,800 of them in Sub-Saharan Africa. The sam- ple includes firms of various sizes, with more micro, small, and medium enterprises in Africa and Latin America than in Bangladesh and China. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 203 The Sub-Saharan African countries are small and generally poorer than the com- parator countries, and, like Bangladesh and India, they tend to be more agrarian (table 3). Investment rates also tend to be lower, although Eritrea and Mozambique recently benefited from large investments. Manufacturing sectors in African coun- tries tend to be modest, with very low exports: the manufacturing share of merchan- dise exports is two-thirds or more in Bangladesh, China, and Morocco but averages just 15 percent in Sub-Saharan Africa. There are also important differences within Sub-Saharan countries. The surveys in Eritrea and Ethiopia took place in the aftermath of a conflict that had a particularly detrimental effect on Eritrea's economy, as continuing conscription created severe labor shortages. By closing off access to Eritrean ports, the conflict also exacerbated the long-standing isolation of Ethiopia's economy, in which state control of private activity is pervasive, foreign direct investment (FDI) is low, "party-statal" firms are common, and tension exists between the government and the traditionally Amharic investment community. Nigeria has also been subject to considerable instability. Its oil-dominated economy has suffered from extremely poor governance and has not yet seen major opening to trade and investment. Eritrea, Ethiopia, and Nigeria are distinctive enough that it would be surprising to find "normal" results there. In contrast, Kenya, Mozambique, Senegal, Tanza- nia, Uganda, and Zambia share a recent legacy of wide-ranging policies to open their economies to trade and foreign investment. Of these, only Kenya and Senegal have avoided severe disruption to their established business communities since independence, through revolutions and civil conflict (Mozambique and Uganda), or socialist development and widespread nationalization (Mozambique, Tanzania, and Zambia). Within this group, Mozambique, Senegal, Tanzania, and Uganda are the best managed.10 Kenya suffered an extended period of poor governance. Zambia experienced an extended period of inconsistent reforms, macro instability, and a series of controversial privatizations that strained relations among the government, donors, and a business sector traditionally dependent on mining-related activities.11 Gross and Net Factor Productivity in Africa This section compares the performance of African firms with those in selected com- parator countries. It begins with technical efficiency, or (gross) productivity, a common focus in the literature. It then examines the effect of business environment­related losses that depress productivity and considers the cost of energy as well as a range of indirect costs, such as transport, telecom, security, land, bribes, and marketing, which are not often considered in the analysis of TFP. Netting out these costs from value- added yields a concept of "net value-added" and a corresponding measure of "net total factor productivity" that is conceptually closer to profitability. This broader view of firm performance, which extends beyond the traditional emphasis on factory-floor pro- ductivity and labor costs, is important to understanding economic outcomes in Africa. Together with the losses that depress (gross) productivity, indirect costs associated with operating expenses represent a heavy drag on net productivity and profitability in most African countries in the sample and reduce competitiveness. 204 TABLE 3. Economic Indicators for Selected Sub-Saharan African and Comparator Countries, 2000­2 Median capital GNI Trade Agriculture Investment FDI Manufacturing Manufacturing per worker in per capita (percent (percent (percent (percent (percent GDP (percent manufacturing Country (dollars) of GDP) of GDP) of GDP) of GDP) annual growth) of exports) (dollars) Sub-Saharan Africa Eritrea 160 111 21 39 5.3 85.4 -- 20,600 Ethiopia 100 49 52 18 1.2 75.0 9.8 2,350 Nigeria 290 81 35 20 2.4 43.7 0.2 20,200 Kenya 360 57 19 14 0.4 131.0 22.0 9,150 Mozambique 210 79 23 40 8.6 139.2 7.5 5,400 Senegal 480 38 18 18 1.3 137.3 37.0 8,900 Tanzania 280 71 45 17 3.7 85.9 18.0 3,350 Uganda 250 40 31 20 2.6 102.9 6.5 1,800 Zambia 330 75 22 18 2.9 114.5 17.0 8,000 Other Bangladesh 380 33 23 23 0.3 165.6 92.0 1,050 Bolivia 900 49 15 16 9.3 151.9 17.0 5,650 China 940 52 15 37 3.7 388.7 88.0 6,700 India 480 31 23 22 0.6 1,55.6 77.0 2,050 Morocco 1,190 66 16 25 4.2 1,74.0 64.0 8,050 Nicaragua 720 73 18 29 5.0 1,41.2 13.0 2,450 Sources: Investment Climate Surveys, 2000­2; World Bank World Development Indicators, 2004. -- Not available. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 205 Earlier analyses of firm survey data from Africa also attempted to account for indirect costs in measuring value-added of firms (Biggs, Shah, and Srivastava 1995) However, due to lack of availability of comparator data outside Africa, these studies were not able to place Africa in a global perspective. This study uses direct cost accounting as well as econometric techniques to investigate productivity and losses and costs, across countries in Sub-Saharan Africa and elsewhere. The authors are aware of no previous dataset that provides the level of detail on sales, costs, and inputs to reliably document these issues and study their implications. Gross (Factory-Floor) Productivity Much firm-level research focuses on productivity, examining differences in physical output produced for a given quantity of inputs. Econometric analyses of productiv- ity often use data on the value of sales and inputs to estimate TFP, a "factory-floor" concept associated with firms' capacity to translate inputs into outputs. In the Cobb- Douglass form, the natural log of TFP is often estimated in the following manner: ­ ln(Ai) = ln(Yi ­ Mi) ­ aln(Ki) ­ b ln(Li) ­ cZi, (1) where A is gross TFP, Y is sales revenue, M is raw materials, K is capital, L is labor, a is the capital share, b is the labor share, Z is a vector of sector and country dum- mies and interaction effects, and c is the corresponding vector of parameters.12 In equation (1), Y ­ M is (gross) value-added. This approach, sometimes estimated in constant elasticity of substitution (CES) or translog form or augmented to address endogeneity concerns, is the classic approach to firm performance at the micro level. Many analyses of African industry, including the World Bank's Investment Cli- mate Assessments and several analyses carried out using African firm survey data from the 1990s, have focused on TFP (Biggs, Shah, and Srivastava 1995; Söderbom and Teal 2003). These analyses suggest that average TFP tends to be low in African firms. Skills and human capital shortages and technology gaps are possible reasons for this problem. The Investment Climate Surveys suggest that hostile business envi- ronments depress firm sales due to losses related to infrastructure and service short- comings, as discussed below. Our estimates of gross TFP for the pooled sample of 15 countries strongly support the proposition that average gross TFP is lower in most African countries than in developing countries elsewhere in the world.13 We used a number of techniques to deal with some of the estimation and robustness issues. One of the main estimation issues is relative prices. Firms in different countries (and even different sectors or regions within a country) likely face different prices for their outputs and for capital and intermediate inputs. Firms in remote areas may receive rents from natural pro- tection and market domination and pay high prices for capital equipment and raw materials. Productivity will appear higher where output prices are inflated and lower where capital goods prices are inflated. To enable sensitivity analyses of the impact of pricing differences, the data on aggregate price levels presented above are com- bined with information on the relative prices of investment and consumption goods from Sala-i-Martin and Artadi (2003). Capital inputs are adjusted using investment good prices, and outputs are adjusted using consumption good prices.14 206 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN Production function estimates from the gross TFP estimations indicate that the shares of capital across sectors ranges from 0.26 to 0.40 and the share of labor ranges from 0.58 to 0.86 (table 4). Constant returns to scale cannot be rejected at the 5 per- cent level in any sector. Productivity differentials among sectors are large and in some cases significant; food and beverages and wood and furniture firms appear particu- larly productive, metals and machinery firms appear less so. Alternative estimations TABLE 4. Regression (OLS) Results for Equation 1, Adjusted Prices Item Coefficient Standard error Constant 4.38 0.23 Log capital 0.40 0.02 Log labor 0.66 0.04 Bangladesh 1.11 0.11 Bolivia 0.96 0.13 China 1.53 0.11 Eritrea 0.43 0.19 Ethiopia 0.69 0.13 India 1.62 0.11 Kenya 1.20 0.14 Morocco 1.46 0.11 Mozambique 0.38 0.18 Nicaragua 1.21 0.12 Nigeria 0.67 0.15 Senegal 1.30 0.15 Tanzania 1.08 0.14 Uganda 0.80 0.14 Chemicals 0.27 0.35 Food and beverage 0.67 0.29 Metals and machinery -0.06 0.31 Textiles, garments, and leather 0.21 0.28 Wood and furniture 0.74 0.40 L*ch (interaction) 0.11 0.06 L*fb 0.07 0.05 L*m 0.04 0.05 L*tgl -0.12 0.04 L*w 0.20 0.08 K*ch -0.04 0.03 K*fb -0.07 0.03 K*m -0.02 0.03 K*tgl -0.14 0.03 K*w -0.14 0.04 Number of observations 7,011 R2 0.65 Source: Authors' calculations. Note: Omitted country: Zambia. Omitted sector: other. Re-estimation without "other" firms has little effect on the coefficients on factors or country dummy variables. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 207 FIGURE 1. Average Gross Total Factor Productivity of Selected Countries Relative to China 1.2 1.0 China to 0.8 0.6 elativer 0.4 TFP oss 0.2 Gr 0 ea occo India Zambia Eritr Nigeria Kenya EthiopiaUganda BoliviaTanzania SenegalMor Bangladesh Nicaragua Source: Authors' calculations. (available on request) using translog production functions and stochastic frontier methods produce very similar results. Equation 1 was used to convert the residuals to an index of TFP relative to China (figure 1). African countries exhibit a wide range of productivity relate to the aver- age TFP of China.15 Indian firms are slightly more productive than Chinese firms, and Moroccan firms are about 90 percent as productive. Senegalese, Nicaraguan, and Kenyan firms are about 75­80 percent as productive as their Chinese counter- parts; firms in Ethiopia, Uganda, Tanzania, Nigeria, and Bolivia are about 45­60 percent as productive; and firms in Zambia, Eritrea, and Mozambique are just 30­35 percent as productive as Chinese firms. These results are in line with most previous findings. Results of Previous Analyses of African Productivity Previous studies suggest that while factory-floor productivity is relatively low in many African countries, it is not low enough (relative to wages) to explain the con- tinent's weak manufacturing competitiveness. For instance, in a study of garment industries, Cadot and Nasir (2001) find that the countries with the lowest factory- floor labor productivity (Mozambique and Ghana) are about half as productive as China but that this differential is more than made up for by lower wages (table 5). If factory-floor productivity is the bottom line for competitiveness, garment firms in Madagascar, Kenya, Ghana, Mozambique, and Lesotho, where unit labor costs per men's casual shirt are just 40­60 percent of those in China, should be dominating those in Chinese export-processing zones. These findings mirror work by Biggs, Srivastava, and Shah (1995), who suggest that African firms are well placed to compete on labor costs. Gelb and Tidrick (2000) cite evidence on the cost structures of African firms in the 1990s suggesting that labor costs are a relatively small share of total costs (less than 20 percent) and that other types of costs may be more important. Eifert and Ramachandran (2004) note that 208 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 5. Factory-Floor Productivity and Labor Costs in Garment Assembly in Selected Countries Number of men's casual Monthly wage of Labor cost shirts produced per semi-skilled machine per shirt Country machine operator per day operator (dollars) (dollars) Chinese export- 8­22 150 0.29 processing zone Lesotho 18 82­95 0.19 Kenya 12­15 60­65 0.18 India 16 70­75 0.17 Madagascar 14­15 55­65 0.16 Mozambique 10­11 40­50 0.16 Ghana 12 30­45 0.12 South Africa 15 255 0.65 Source: Cadot and Nasir 2001. the ratio of labor costs to value-added at the firm level (a common proxy for unit labor costs) has less predictive power than previously suggested with respect to man- ufacturing performance at the country level within Africa.16 This finding indicates that the focus on factory-floor productivity and labor costs may be too narrow. Much of the literature on the business environment spurred by the Investment Cli- mate Surveys has focused on explaining variation in (gross) TFP using "hard" (non- perceptions-based) indicators in areas such as infrastructure quality, regulatory bur- den, and product market competition.17 Studies that exclude country fixed effects find large effects of business environment variables on TFP (Batsos and Nasir 2004). However, the problem of unobserved variables is vast: any explanatory variable that differs enough between Africa and its higher performing comparators produces large, significant effects. The indirect nature of potential links between business environ- ment variables and TFP further compounds the omitted variables problem. Studies that include fixed effects find less of a role for business environment vari- ables (Dollar, Hallward-Driemeier, and Mengistae 2003). In the case of variables that are essentially cross-country in nature (such as port quality), identification is very difficult. Pooled multicountry estimations--even those that include country dummies--are limited in their ability to shed light on the complex interactions of the explanatory variables. If the binding constraints on firm performance vary across countries, there is no reason to expect similar magnitude effects of individual busi- ness environment components across countries.18 While further econometric analysis on firm-level TFP with a set of business envi- ronment variables on the right-hand side may advance the state of knowledge, a dif- ferent approach is taken here. The analysis first considers the impact of losses in reducing TFP and then includes indirect costs, as a percentages of sales. This analy- sis does not include all business environment­related costs, risks, and losses, but it provides some idea of how aspects of the environment affect competitiveness.19 BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 209 FIGURE 2. Losses from Power Outages in Selected Countries, by Percentiles 16 14 sales 12 of 10 8 6 centage 4 2 Per 0 na ea bia Chi India India Eritr SenegalUganda Zam Kenya Ethiopia Nigeria Nicaragua Tanzania Mozambique Bangladesh Ahemdabad, 25th percentile 37th 50th 63rd 75th Source: Authors' calculations based on Investment Climate Surveys, 2000­4. Impact of Direct Losses on Factory-Floor Productivity The Investment Climate Surveys provide information on a range of business losses that are helpful in understanding productivity shortfalls. One example is losses due to power outages. Variants of the following question were asked: "What percent of annual sales did you lose last year due to power outages or surges from the public grid? Please include losses due to lost production time from the outage, time needed to reset machines, and production ruined due to processes being interrupted." Simi- lar questions were asked about problems with infrastructure-related issues, such as delays in logistics and transport failures--areas that may be plausibly interpreted as costs to the firm. The data indicate that losses due to shipping delays or hold-ups at ports are significant in some countries. Unfortunately, only data on losses due to power failures are available for all the countries in the sample. African firms report substantially higher losses than their counterparts in higher performing countries (figure 2), which translates into a corresponding decline in meas- ured productivity. This result also holds up when the log of gross TFP is regressed on losses from power outages. The coefficient is roughly 0.01, so a reported loss of 1 per- cent of sales is statistically associated with 1 percent lower gross TFP. A substantial portion of the variance in measured productivity between China and several African countries (especially Zambia, Ethiopia, Kenya, Nigeria, and Tanzania) can be attrib- uted to infrastructure and logistics-related losses. In Kenya losses from power failure amount to 6 percent of sales for the median firm; in China they are only 1 percent of sales. Interestingly, power failure is the one variable that Dollar, Hallward-Driemeier, and Mengistae (2003) find to be robustly associated with TFP.20 Impact of Indirect Costs on Productivity Energy is consistently the largest component of indirect costs, averaging about one- third of the total (table 6). Transport tends to follow at about 5­15 percent, land costs cluster at about 10 percent, telecommunications and security fall in the range of 2­8 percent, and water represents about 2 percent.21 Marketing is also often a sig- nificant cost (8­16 percent). A range of items fall under the heading "other costs." 210 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 6. Composition of Indirect Costs, by Country (percent) Category Ban Bol Chi Eri Eth Ind Ken Mor Moz Nic Nig Tza Sen Uga Zam Energy 18 16 -- 26 45 37 35 51 20 32 28 59 59 21 32 Land rent 8 0 32 -- 5 1 5 10 5 7 5 9 10 12 2 Transport 6 15 16 4 5 21 16 9 6 -- 6 -- -- -- -- Telecom 2 2 -- 5 1 8 8 3 2 -- 5 -- -- -- -- Royalties 2 2 -- -- 0 0 1 2 -- -- -- 0 -- -- 1 Water -- 5 -- 2 -- -- 2 -- 1 -- 2 -- -- -- -- Subcontracting -- 5 18 -- -- -- -- -- -- -- -- -- -- -- -- Security -- -- -- 0 2 -- 7 3 2 -- -- -- -- -- 4 Maintenance -- 4 -- 9 -- -- -- -- 8 -- 32 -- -- -- -- Spare parts -- 6 -- -- -- -- -- -- -- -- -- -- -- -- -- Insurance -- 2 3 -- -- -- -- -- 2 -- -- -- -- -- -- Marketing -- 8 21 1 -- -- -- -- 1 -- 16 -- -- -- -- Independent -- 3 -- -- -- -- -- -- -- -- -- -- -- -- -- professionals Office supplies -- 1 -- -- -- -- -- -- 1 -- -- -- -- -- -- Tickets, travel -- 2 -- -- -- -- -- -- -- -- -- -- -- -- -- Export expenses -- 3 1 -- -- -- -- -- -- -- -- -- -- -- -- Accounting -- -- -- 2 -- -- -- -- -- -- 1 -- -- -- -- Other 64 27 10 52 41 32 27 22 53 61 6 32 31 67 62 Source: Authors' calculations based on Investment Climate Surveys, 2000­4. -- Not available. Note: The China Investment Climate Survey included energy costs as part of total raw materials costs; energy costs are assumed here to account for one-third of total indirect costs in China, equal to the average across the other 14 countries. These costs typically include insurance, office supplies, travel costs, accounting, maintenance, and spare parts. Infrastructure and public services--energy, transport, telecom, water, and security costs--together account for more than half of all indi- rect costs. In strong performers, such as China, India, Nicaragua, Bangladesh, Morocco, and Senegal, energy and indirect costs represent 13­15 percent of total costs, about half the level of labor costs (figure 3). In contrast, in most African countries these costs account for 20­30 percent of total costs, often dwarfing labor costs. Capital costs, which are related to the business environment, appear to be a major component of costs in Ethiopia, Nigeria, and Zambia. These cost breakdowns do not capture all factors that affect competitiveness. For instance, transport costs are much higher in Africa than in Asia or Latin America. To the extent that part of the excess is incurred indirectly in the form of higher prices for raw materials, figure 3 underestimates the magnitude of indirect costs in Africa, and the productivity gaps shown are biased upward, because African firms facing high transport costs may be using less physical raw material than the dollar values suggest. Similarly, if particular services are complementary to capital and labor and firms choose to use less of these services because of their high prices, there may well BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 211 FIGURE 3. Cost Structures in Selected Countries (firm-level averages) Mozambique Eritrea Kenya Tanzania Zambia Uganda costs Bolivia total Nigeria of e Ethiopia China Shar Nicaragua Morocco India Senegal Bangladesh 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Materials Labor Capital Indirect Source: Authors' calculations based on Investment Climate Surveys, 2000­4. FIGURE 4. Gross and Net Value-Added per Worker (adjusted dollars) 25,000 20,000 dollars 15,000 10,000 5,000 Adjusted 0 ea a occo India ZambiaEthiopia Bolivia Eritr Nigeri Kenya TanzaniaUgandaMor adesh China NicaraguaSenegal Mozambique Bangl Gross value-added per worker Net value-added per worker Source: Authors' calculations based on Investment Climate Surveys, 2000­4. Note: Dollars are adjusted for purchasing power parity and cost of consumption versus investment goods. be a negative impact on sales and measured productivity. If transport costs also raise the prices of outputs, the bias goes the other way. Purchasing price parity­adjusted exchange rates are a poor attempt to capture such subtle effects. Net value-added is defined as gross value-added less indirect costs.22 It is a broader indicator of firm performance than gross value-added. Figure 4 compares these two concepts of value-added in per-worker terms. In high-performing economies with rel- atively low indirect costs, median net value-added is a high share of gross value-added: 67 percent in Morocco, 71 percent in India, 74 percent in Bangladesh, 76 percent in 212 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 7. Estimation of Equation (2): Net Productivity Regression Variable Coefficient Standard error Constant 3.62 0.26 Log capital 0.41 0.02 Log labor 0.62 0.04 Bangladesh 1.70 0.14 Bolivia 1.42 0.15 China 2.27 0.14 Eritrea 1.16 0.23 Ethiopia 1.41 0.16 India 2.24 0.13 Kenya 1.61 0.17 Morocco 2.04 0.14 Mozambique 1.11 0.21 Nicaragua 1.79 0.15 Nigeria 1.18 0.18 Senegal 1.98 0.18 Tanzania 1.64 0.17 Uganda 1.36 0.16 Source: Authors' calculations based on Investment Climate Surveys, 2000­4. Nicaragua, 82 percent in Senegal, and 85 percent in China. For Bolivia, Eritrea, Mozambique, Nigeria, Tanzania, and Kenya, the range is 42­51 percent, suggesting the significant impact of cost disadvantages in these countries. For Zambia gross value-added is already a small percentage of sales (22 percent), so indirect costs (including the cost of energy) badly squeeze the viability of manufacturing firms. Net TFP is then estimated as ln(Âi) = ln(Yi ­ Mi ­ ICi) ­ aln(Ki) ­ b ln(Li) ­ cZi (2) where Âi is net TFP, IC is indirect costs, and (Y ­ M ­ IC) is net value-added. Country averages of firm-level net TFP are estimated using country dummy vari- ables to estimate gaps.23 As before, several different methods, including translog and CES functions and stochastic frontier methods, were used to test for robust- ness. Because all yielded very similar results for the country dummy variables, the simplest method--OLS estimates of a Cobb-Douglass production function--was used (table 7). The gap between African and non-African firms widens when net TFP is exam- ined, as indirect costs interact with other firm characteristics (figure 5). African coun- tries in the mid-range of 40­60 percent of Chinese gross TFP fall to 20­40 percent when net TFP is compared. Kenya, which appears relatively strong on gross TFP, falls dramatically on net TFP as a result of very high indirect costs. Zambia, the most extreme case, falls from 30 percent to 10 percent. Only in Senegal--the strongest African performer on both gross and net TFP--is the effect of indirect costs relatively BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 213 FIGURE 5. Net and Gross Total Factor Productivity of Selected Countries Relative to China (adjusted prices) 1.2 1.0 China to 0.8 0.6 elativer 0.4 0.2 Index, 0 ea occo India ZambiaEritr Kenya China ambiqueEthiopiaNigeriaBoliviaUganda TanzaniaNicaragua SenegalMor Moz Bangladesh Gross TFP Net TFP Source: Authors' calculations based on Investment Climate Surveys, 2000­4. low. African countries have low factory-floor productivity, and high indirect costs further weakens their relative performance. Profitability and Returns to Capital Few African surveys reveal manufacturing sectors with high profit margins or high returns to capital. Senegal, Tanzania, and Uganda are the strongest African perform- ers, Senegal because of its high productivity and low indirect costs, Tanzania and Uganda because of low capital intensity and labor costs. The three countries compare favorably with Morocco and Nicaragua on profitability, because firms in the latter countries face extremely high labor costs. China and Morocco have relatively low labor costs, high productivity, and low indirect costs, and firm profitability in these countries reflects these factors. Other than Senegal, Tanzania, and Uganda, returns in Africa, particularly in Zambia, are low. The three African countries with high cap- ital costs as a share of total costs (Ethiopia, Nigeria, and Zambia) are also the three least profitable countries. Indirect costs squeeze African firms tightly (figure 6). Moderate reductions in these costs would sharply increase the viability of African manufacturing enterprises, pushing many firms out of the red and making just-profitable firms much more lucra- tive (figure 7 and table 8). For most African countries, even reducing indirect costs as a share of total costs to the level of Senegal would have a greater effect on profit margins than would halving labor costs. Reducing indirect costs might be expected to boost profitability by an equivalent amount, but the complex interactions between costs and firm behavior suggest that simple arithmetic is not always appropriate. For instance, firms facing lower indirect costs and more reliable power supplies may use different technologies and more busi- ness services, further increasing their productivity and profitability. Part of the increases in profit margins may accrue to workers in the form of rent-sharing and higher wages. This suggests that further work should study the role of different types of costs in more sophisticated models of firm behavior. 214 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN FIGURE 6. Cost Structures and Profitability in Manufacturing in Selected Countries 1.4 1.2 1.0 0.8 costs 0.6 ect 0.4 0.2 Indir 0 ­0.2 ­0.4 ivia ea Bol China India Eritr Kenya occo bique Ethiopia Mor NigeriaSenegalanzaniaUgandaZambia Nicaragua T Bangladesh Mozam Materials Labor Capital Indirect Profit Source: Authors' calculations based on Investment Climate Surveys, 2000­4. FIGURE 7. Actual and Estimated Profitability in Selected Countries under Alternative Scenarios 0.3 0.2 sales) of 0.1 Zambia centage 0 per ea ­0.1 occo India (as Nigeria Bolivia Eritr Kenya China Ethiopia Mor Uganda anzaniaT Senegal ofits ­0.2 Nicaragua Bangladesh Pr Mozambique ­0.3 Actual profits Profitability if indirect costs fell to 13 percent of total Profitability if labor costs fell by one-third Source: Authors' calculations based on Investment Climate Surveys, 2000­4. Note: Means of the 5th­95th percentiles of profitability. The Political Economy of Business Climate Reforms Why should we open our economy all for the benefit of South African and Asian business? --Comment by high government official during the discussion of an Investment Climate Assessment in Africa The third theory of comparative advantage discussed above stresses the value of hav- ing dense networks of firms operating in a competitive environment and being able to generate "thick markets" and learning externalities. This highlights the impor- tance of entry for sparse economies. The Investment Climate Assessments indicate a long list of entry barriers; the Doing Business indicators confirm that barriers tend BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 215 TABLE 8. Actual and Estimated Return on Capital in Selected Countries under Alternative Scenarios (percent) Estimated return on Estimated return on Actual return capital if indirect costs fall capital if cost of labor Country on capital to 13 percent of total costs falls by one-third Bangladesh 0.35 0.35 0.54 Bolivia 0.00 0.10 0.07 China 0.38 0.39 0.45 Eritrea 0.04 0.07 0.05 Ethiopia 0.01 0.04 0.03 India 0.71 0.76 0.93 Kenya 0.08 0.20 0.11 Morocco 0.02 0.03 0.08 Mozambique 0.03 0.14 0.08 Nicaragua 0.17 0.19 0.40 Nigeria 0.02 0.04 0.03 Senegal 0.21 0.21 0.28 Tanzania 0.26 0.57 0.33 Uganda 0.16 0.31 0.29 Zambia -0.18 -0.08 -0.12 Source: Authors' calculations. Note: Return on capital is calculated as profit/replacement cost of capital stock. Figures represent firm-level medians. to be high in Africa and that progress in reducing them has lagged other developing regions. Recently, the World Bank's Vice President for Africa remarked in a speech in Nairobi that despite a vast amount of analytical work on the private sector, no real dialogue has emerged between the private sector and the government. Why is change not faster? Lack of money--to ease the severe infrastructure constraints identified in the surveys--is part of the problem. But many other shortcomings of the business climate reflect the political economy that underlies state performance and capacity. Reforms will need to confront the presence of long-established rent-seeking arrangements that benefit both the political and private sector elites. These arrangements are remark- ably stable, reflecting the coexistence of strong presidential systems of governance, weak administrative and technical capacity, noncredible donor conditionality, and small domestic private sectors dominated by a few large and highly profitable firms, often owned by foreigners or minorities. These structural and institutional features of African business sectors contribute to the political economy problem. Is the African Private Sector in a Low-Level Political Equilibrium? Political analyses of Africa, both old and new, shed light on the twin problems of slow growth and partially successful reforms. In his analysis of the political economy 216 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN of the African private sector, Tangri (1999) argues that the minority Asian commu- nity in East Africa, which has thrived even in difficult times, often coexists with a small, wealthy, black private sector, often closely aligned with the president or his associates, whose success is defined more by political connections and rent-sharing than by business expertise. Van de Walle (2001) reinforces this perspective, arguing that the political elite in Africa has learned to adapt to reform while finding ways to preserve rent-seeking arrangements: Leaders' notion of the political viability of reform has changed over time. Their initial reaction was almost entirely negative because they viewed rapid reform as incompatible with the methods of rule they had fashioned over several decades of rule. Over time, this has changed: from the view that reform was not viable, leaders have understood that they had no choice but to adapt their methods of rule to the evolving environment. . . over time and through experimentation, they found that their hold on power could withstand the partial implementation of adjustment programs. It remains true that polit- ical elites do not believe they can survive without recourse to a policy regime of system- atic interference, but they have learned to adapt this interference. (p. 170) Thus governments have simply changed their methods of rent-seeking in response to donor-driven reforms. This has translated into a series of partial reforms, without much change in the ability of the private sector to do business, leading to what is termed a "permanent crisis." For the private sector, it has meant keeping up with ever-changing forms of government interference, as the sources of rents and the modalities of rent-seeking have shifted with reform efforts. Van de Walle argues that there has been little policy learning in Africa relative to Asia or Latin America. Technocrats within the government have often been hostile to reforms, because their involvement and inputs are limited relative to foreign experts. Partial reforms have been largely successful at keeping donors satisfied, often leading to repeated rounds of financing to address the same issues and ultimately resulting in toothless conditionality, the preservation of rent-seeking arrangements, and little real reform despite apparent progress at the macro level.24 All of this has served to reinforce the lack of momentum on private sector development. Investment Climate Assessments are mostly technical, but some studies have begun to focus on the political feasibility of business climate reform. Reducing administra- tive barriers in Africa is enormously difficult, mainly because the state apparatus has long been used to dispense patronage (Emery 2003). Privatization programs have not been entirely successful at eliminating rent-seeking parastatals, and the privatization process itself has offered opportunities for rent-seeking and patronage. In a detailed analysis of the administrative requirements for setting up a business in Africa, Emery notes how overall complexity places a premium on means of circumventing or speed- ing up the process, which creates a flourishing environment for corruption. Most, if not all, businesses are operating outside the law in at least one or more aspects and are vulnerable to government inspectors, no matter how minor the deviance. The sur- vival of a business is consequently heavily dependent on a personal relationship with a minister or other high government official, which is often difficult to document or quantify. These relationships are crucial to firms that need to anticipate ad hoc policy or regulatory changes--a major concern of business, as shown in the Investment Cli- BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 217 mate Assessment surveys. Emery concludes that this vulnerability, combined with the arbitrary nature of enforcement arising from poor governance means that firms can be closed down or worse for operating in exactly the same way as their neighbors, their competitors, or their clients and suppliers (Emery 2003). Failure to broaden the base of the business community increases the public's skep- ticism of the private sector, particularly of foreign-owned firms. Although the World Bank and other donors focus their dialogue on technical solutions to private sector development such as better roads, more power generation, and reduction of the reg- ulatory burden, dialogues in the domestic press in Africa have focused largely on the proposition that the persistence of a private sector elite (whether foreign, ethnic minority, or black) has prevented economic empowerment of the majority of black Africans. This configuration of interests increases the likelihood of countries falling into a low-level equilibrium. With a dominant part of the business sector identified as not indigenous and shielded from "outside" entry by an adverse business environ- ment, the fractured business community has less ability--and perhaps less incentive-- to act as a powerful pressure group in favor of reform. The difficulty of shifting out of such an equilibrium is mirrored in African atti- tudes toward the private sector. While support for a market-based approach to growth and development may be growing, it is still far from widespread. The Afro- barometer surveys reveal widely differing views of the private sector in 15 African countries (Botswana, Cape Verde, Ghana, Kenya, Lesotho, Malawi, Mali, Mozam- bique, Namibia, Nigeria, Senegal, South Africa, Tanzania, Uganda, and Zambia) (Bratton, Mattes, and Gyimah-Boadi 2005). In just six countries did a majority of respondents feel that a market economy was preferable to an economy run by the government--and even in these countries the percentage of respondents averaged only 54 percent. Only 24 percent of respondents in Botswana, 26 percent in Lesotho, 37 percent in South Africa, and 39 percent in Namibia preferred a market economy over one run by the government. In most countries a majority of respondents believe that the government should bear the main responsibility for ensuring the well-being of people. In Ghana, for example, 66 percent of respondents believe that "the gov- ernment should retain ownership of its factories, businesses, and farms." Despite recognition that corruption is widespread, more than 72 percent of respondents agreed that "all civil servants should keep their jobs, even if paying their salaries is costly to the country." Fifty-eight percent of respondents indicated that the govern- ment should bear primary responsible for job creation. Most reform programs are greeted with lukewarm support or outright opposition--62 percent of respondents supported user fees and 54 percent supported market pricing, but only 35 percent supported privatization and 32 percent supported civil service reform. These results indicate that most Africans are skeptical that the private sector will deliver broad- based growth. This is consistent with discussions in the press of elite capture of the private sector.25 Beneath Aggregate Gaps: Size, Ethnicity, and Foreign Ownership in Africa African business is often segmented, with small clusters of large, foreign, and ethnic minority­owned firms that are quite different in character from their indigenous 218 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 9. Average Number of Workers in Firms in Selected Countries, by Origin of Owners Country Domestic Foreign African European Asian Arab Other Bangladesh 149 340 -- -- -- -- -- China 232 284 -- -- -- -- -- Eritrea 38 52 40 52 -- -- -- Ethiopia 22 15 -- -- -- -- -- India 27 252 -- -- -- -- -- Kenya 73 230 60 90 59 85 249 Morocco 52 100 -- -- -- -- -- Mozambique 34 45 -- -- -- -- -- Nicaragua 14 60 -- -- -- -- -- Nigeria 81 252 84 310 216 156 904 Senegal 29 63 24 65 28 35 -- Tanzania 29 159 23 84 46 43 -- Uganda 19 68 16 104 60 138 -- Source: Investment Climate Surveys, 2000­4. -- Not available. counterparts.26 These firms tend to have higher productivity than indigenous firms and to export more than their smaller indigenous counterparts They also seem to have far more market power and to be able to sustain their presence in Africa despite economic and political uncertainties. Tangri (1999) and van de Walle (2001) suggest that this group relies on trust between its members and on alliances with the politi- cal elite to generate rents on a continuous basis. This business cleavage is reflected in the surveys. In almost all countries there is a strong relationship between foreign ownership and firm size (table 9). In Africa this relationship also extends to ethnic minority ownership: the average indigenous African firm in Uganda has 16 employees, while Asian-owned firms average 60 employees, European-owned firms average 104, and Arab-owned (mostly Lebanese) firms average 138. Domestic firms have an average of 19 employees, while the average foreign firm has 68. Firms owned by foreigners or ethnic minorities outperform indigenous firms by a substantial margin in every country except Eritrea and Ethiopia. African business is thus segmented: small indigenous firms struggle to survive, while small numbers of larger firms, often owned by foreigners or ethnic minorities, have productivity levels closer to those of the average firm in high-performing economies such as China and India. The most productive private firms in Africa tend to be large. In China productiv- ity levels of small and medium-size enterprises are about 80 percent those of larger firms; in countries such as Senegal, Uganda, and Kenya, this ratio is closer to 50 per- cent (figure 8).27 Micro firms (those with fewer than 10 employees) operate at even lower levels of productivity in all countries except Ethiopia and Eritrea. Productivity differentials between small, indigenous firms and larger, foreign, or minority-owned firms have been persistent over time, and they do not seem to be driven primarily by differences in indirect or opportunity costs.28 Large firms still BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 219 FIGURE 8. Total Factor Productivity of Micro, Small, and Medium-Size Firms Relative to Large Firms in Selected Countries 1.6 1.4 ms fir 1.2 ge 1.0 lar to 0.8 0.6 elativer 0.4 TFP 0.2 0 ea India esh China Senegal Uganda Kenya Bolivia lad Eritr Ethiopia Nigeria Zambia Nicaragua Tanzania Bang Micro firms Small and medium-size enterprizes Source: Authors' calculations based on Investment Climate Surveys, 2000­4. FIGURE 9. Firms' Self-Reported Market Share, by Size and Ownership cent) 70 (per 60 e shar 50 40 market 30 20 domestic 10 0 Median India China Ethiopia Uganda Zambia Bolivia Kenya Senegal Nicaragua Tanzania Micro Small Large Source: Investment Climate Surveys, 2000­4. incur heavy costs for self-provided infrastructure, transport, logistics, security, and other items. Market power reflected in higher product prices may play a role: although large firms are more likely to be exporters, many also sell to the domestic market and enjoy very high market shares (figure 9.) Nigeria, the least competitive country in Sub-Saharan Africa, is also perhaps the country with the least support for liberalization from the business community. Although large firms face many of the same constraints as small firms, they are able to adapt in different ways. For example, unlike small firms, most large firms 220 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE 10. Share of Firms in Selected Countries Owning Generators, by Firm Size (percent) Country Micro Small and medium Large and very large Bangladesh 28 53 88 Bolivia 13 11 31 China 0 14 38 Eritrea 38 43 63 Ethiopia 3 23 43 India 23 76 91 Kenya 46 67 89 Morocco 14 15 28 Mozambique 20 23 63 Nicaragua 6 29 81 Nigeria 83 96 99 Senegal 23 19 16 Tanzania 18 60 89 Uganda 4 44 87 Zambia 30 28 61 Source: Investment Climate Surveys, 2000­4. own a generator (except in Senegal, where the power system works relatively well), and they are notified well in advance of rolling blackouts (table 10). In many coun- tries, the share of large firms with access to credit is much higher than that of small firms. Productivity differentials are in part related to access to finance; controlling for country fixed effects, firms that have a loan or overdraft account are about 6 percent more productive than other firms. Large firms spend more time than other firms dealing with regulations and regu- lators (table 11). This may be a burden but, given the persistent profitability of most large, foreign-owned, and ethnic minority firms, the finding is also consistent with the hypothesis that these firms spend more time exchanging favors with government representatives. The role of networks in the African private sector is crucial to understanding the nature of the cleavages. Biggs and Shah (2004) show that the ethnicity of firms' pro- prietors remains an important determinant of access to credit and a number of other performance variables, even when variables such as the education level of proprietors and title to marketable assets are included in regressions. Networks, usually of eth- nic minorities and based on trust among members of a relatively small group, help firms overcome the limitations of financial markets (Fafchamps 2004). They also exclude outsiders from entering certain areas of business. Networks operate in many regions, including fast-growing Asian countries, where they may have positive effects in enabling their members to compensate for dysfunctional market institutions. But their overall impact is likely to be different in Asia and Africa, because of differences in economic density and market size. In Asia their adverse effect in stifling competi- tion is likely to be small, because of the competitive pressure of many firms belong- BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 221 TABLE 11. Interaction with Government in Selected Countries, by Firm Size Median number Median percentage of senior management of inspection days time spent dealing with regulation Small and Large and Small and Large and Country Micro medium very large Micro medium very large Bangladesh 6.5 10 12 1 3 3 China -- -- -- 5 5 5 Eritrea 7 4 4 1 2 5 Ethiopia 3 7 6 1 1 1 India 3 7 12 5 10 13 Kenya 11 16 17 7.5 10 10 Morocco 0 1 3 -- -- -- Mozambique 3 3 3 10 5 8 Nicaragua 4 15 32 6 18 17 Senegal 9 13 13 -- -- -- Tanzania 12 32 62 5 10 13 Uganda 2 5 17 1 1 6 Zambia 52 64 75 15 11 13 Source: Investment Climate Surveys, 2000­4. -- Not available. ing to many networks. In Africa's very small economies, the adverse effect of a few dominant networks or firms is likely to be far larger. Firms in sparse economies are likely to place more weight on the costs and risks of encouraging entry through reforms than are firms in dense economies. Small, sparse, industrial sectors dominated by a few firms with high market share are therefore likely to see less dynamic competition. The greater access of larger, net- worked firms to technology, credit, and business expertise creates rents that, even if shared with government, would be dissipated by more open entry. This can reduce the incentive to push hard for better regulation and business services. Even though much of the benefit of reforms may accrue to small firms, these firms doubt their ability to compete and often are not operating in the dominant business sector. At the same time, the prominent role of minority and expatriate businesses increases public reser- vations over the market economy model, including over large privatizations (Nellis 2005). The danger is a low-level equilibrium, with limited pressure for reform from the business community and the public and limited response from government. Business environments usually improve slowly, but reforms seem to have occurred even more slowly in Africa than elsewhere. Between 1990 and 2002, World Bank documents claimed that Africa was about to turn the corner in terms of policy reform no fewer than 14 times (Easterly 2003). Donors have contributed billions of dollars to road construction across Africa, but the overall quality of road networks has improved little, because of poor maintenance.29 Technical recommendations for change will need to take account of this complex political economy. 222 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN Conclusion Firm-level analysis must include losses and indirect costs. A set of these costs, iden- tified in firm surveys, is examined here. Conventional, or "gross," TFP, a "factory- floor" concept, can be extended to "net" TFP by incorporating indirect costs into a net analog of value-added, including a wider span of operational costs borne by the firm. Net TFP varies much more across countries than gross TFP. Gross TFP of African firms averages 40­80 percent that of Chinese firms, depend- ing on how the estimation is done. Part of this gap is accounted for by excessive out- put losses caused by factors external to the firm. In terms of net TFP, the productiv- ity of African firms drops to just 20­40 percent of that of Chinese firms. For most African countries, reducing indirect costs even to the level of Senegal, the best per- former in the African group, would have a greater impact on profit margins than halving labor costs. Macroeconomic data confirm that Africa tends to be a "high-cost" region; the firm-level estimates confirm that high costs are reducing industrial competitiveness. Countries that have been able to diversify into manufactured exports tend to have low cost levels relative to purchasing price parity predictions and lower indirect costs. In many African countries, the problem is not so much that labor is expensive rel- ative to factory-floor productivity, it is that high indirect costs and losses reduce the return to labor in production and thus depress labor demand and real wages. This story varies considerably across and within countries. Firms in Mozambique and Zambia have relatively weak factory-floor productivity, and their value-added is heavily squeezed by high costs (three-fourths of Zambia's net TFP shortfall relative to China is explained by excess indirect costs). To a lesser extent, this is also true in Ethiopia and Nigeria. Eritrea, Tanzania, and Uganda appear to be middle-of-the- road performers, although the data on Eritrea are probably heavily influenced by state favoritism and anticompetitive rents. In Kenya a long history of entrepreneur- ship is reflected in strong potential factory-floor productivity, but high costs and losses impede competitiveness. Senegal provides an example of what an African country can achieve with a strong business community working in a relatively good business environment. African business sectors are segmented on the basis of ethnicity, ownership, and firm size. Large, foreign, and minority-owned firms typically have much higher factory-floor productivity than their smaller indigenous counterparts, although they still face high costs. These firms are more likely to export, and they also have very large shares in their domestic markets, possibly reducing competitive pressures and the drive to innovate and expand. Ethnic and structural cleavages in African busi- ness sectors also have implications for the political economy of reform. Poor busi- ness environments generate entry barriers that provide larger firms with anticom- petitive rents. Firms that might push for reform are therefore faced with a choice between a hostile business environment that they have learned to negotiate and an unknown situation with potentially large increases in entry and competition. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 223 The likelihood that this equilibrium will be sustained is buttressed by the ambiva- lent attitudes of Africans toward the business sector. The risk is that Africa will remain locked into a slowly evolving low-level equilibrium, characterized by rent- seeking behavior on the part of the public sector, quiet acquiescence on the part of the private business sector, slow entry, continuing sparseness of firms and entrepre- neurial activity, and limited gains from competition and conglomeration. In this low- level equilibrium, measures to open Africa's economies and improve regulation will have a limited effect, because of the limited incentives to focus on business services and the institutional underpinnings of competitiveness. How can the momentum of reform be accelerated? Benchmarking performance in the various areas highlighted by Investment Climate Assessments will facilitate more constructive dialogue on business climate variables in discussions among govern- ments, firms, and donors. In thinking about moving forward on the business envi- ronment agenda, however, solutions need to be framed by political economy consid- erations. Accelerating reform requires the breaking up of alliances between the private sector elite and the political elite. One approach is to pursue partial reforms that create new opportunities for the private sector while rents in established areas are slowly eroded over time. This can help convince the private sector elite that the profits in a relatively open economy may well be greater than current levels, partic- ularly if costs can be significantly reduced. Policies that encourage the arrival of new entrants into the private sector will also be very useful in increasing economic den- sity. Specific recommendations include the following: · Focus on reducing the most severe indirect costs faced by firms. In most countries the availability and reliability of power is a clear priority. Transport and logistics- related losses and costs are high in most countries; telecommunications and secu- rity costs are high in some. The Investment Climate Assessments can help bench- mark some of these costs against more competitive countries, including within Africa itself. · Level the playing field. Building strong domestic political support for private sec- tor development will require improving the performance and capacity of indige- nous firms, removing the perception that reforms benefit primarily minorities or foreigners. Programs to mitigate political risk, for example, are currently available only to foreign investors; they should be extended to domestic investors on similar terms. Tax incentives offered to large businesses should be extended to small ones. Capacity building is needed for the private sector (especially the indigenous private sector, where productivity is low) as well as the public sector. Outside of South Africa, Sub-Saharan Africa does not have a single accredited business school. · Enclave growth to increase business density. Even if political resistance and weak capacity stall sweeping countrywide reforms, it may be feasible to improve busi- ness services in limited, high-profile areas, such as export processing zones (EPZs). Within EPZs service delivery standards can be benchmarked and regularly evalu- ated. EPZs can also help address the problem of low firm density, reduce 224 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN infrastructure costs, encourage technology diffusion and knowledge spillovers, and attract new entrants.30 · Use gains to build constituencies for reform. Success in even a single enclave sector can generate demonstration effects across an economy, weakening the perception among large firms and governments that a low-level equilibrium characterized by high costs and low competition is preferable. Press accounts of the Indian experi- ence suggest that the enormous success of the high-technology sector in the early 1990s (accomplished largely without government assistance) set the standard for the rest of India's private sector; wanting to share the limelight, both firms and the government started to move away from the old model of the "license raj." In Africa examples such as Kenya's agri-business sector (which links large firms, small firms, and farms) and the EPZs in Madagascar and Mauritius offer examples that should become far better known. In China successful local development rebounded directly to the advantage of local officials, setting up strong competition among local governments to attract investments (Byrd and Gelb 1990). Top customs offi- cials could receive incentive payments based on both revenue and clearance time standards, along the lines of Tanzania's Selective Accelerated Salary Enhancement Scheme (Levy and Kpundeh 2004). Policy makers should consider how such incen- tives can be structured to boost, rather than retard, productivity. · Enhance the profile and credibility of reforms. Social sector indicators and reforms currently enjoy a higher profile in country-donor discussions than business-related reforms; the focus on performance standards has not yet carried over into business- related reforms. Innovative instruments could increase the visibility of business- related reforms. Donor-funded facilities could enable firms to buy insurance against shortfalls in service delivery from standards agreed to as part of reform programs. The objective would be less to compensate business than to ensure that business-related services receive greater attention This insurance could cover cus- toms clearance times; value-added tax rebates for exporters, initially within EPZs; and possibly power outages. Such a mechanism would provide impetus for the accurate measurement of performance (and targets to focus capacity-building efforts) and ensure that lapses in performance are speedily raised to a high policy level. Once their efficacy is confirmed, service standards and insurance programs could be implemented more broadly across the economy. · Capitalize on the concern over donor dependence. Increased donor dependence is inevitable for African countries if they are to embark on a determined push toward achieving the Millennium Development Goals. But countries such as Mozambique and Uganda, 50 percent of whose budgets come from donors, rec- ognize the political implications of donor dependence and are trying to reduce it. In a recent speech, the president of Uganda urged his countrymen to become less dependent on foreign aid. Tanzania, which hosts 1,000 donor meetings every year and prepares 2,500 donor reports every quarter, may also feel the need to reduce donor dependence (Birdsall 2004). More explicit linkage between private sector development and a reduction in donor dependence may hasten the implementa- tion of reform. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 225 Annex Data on Firm-Level Costs Tables A1­A3 contain response rates to questions about sales and costs, by share of firms. TABLE A1. Response Rates by Ownership and Exporter Status Domestic Foreign Non-exporter Exporter Country Sales Costs Sales Costs Sales Costs Sales Costs Bangladesh 0.98 0.89 1.00 0.85 0.87 0.89 1.00 0.90 Bolivia -- -- -- -- 0.61 0.43 0.67 0.49 China 1.00 0.79 1.00 0.79 1.00 0.74 1.00 0.81 Eritrea 0.52 0.40 0.59 0.53 0.56 0.45 0.29 0.29 Ethiopia 0.99 0.92 1.00 0.95 0.99 0.92 0.96 0.93 India 1.00 0.63 1.00 0.73 1.00 0.53 1.00 0.61 Kenya 0.59 0.57 0.76 0.60 0.60 0.57 0.67 0.61 Morocco 1.00 0.96 1.00 0.96 1.00 0.95 1.00 0.96 Mozambique 0.75 0.36 0.67 0.33 0.72 0.34 0.86 0.50 Nicaragua 1.00 0.60 1.00 0.64 1.00 0.59 1.00 0.65 Nigeria 0.92 0.75 0.86 0.82 0.89 0.78 1.00 0.78 Senegal 0.89 0.51 0.86 0.55 0.88 0.53 0.92 0.49 Tanzania 0.88 0.44 0.78 0.48 0.86 0.42 0.84 0.60 Uganda 0.52 0.28 0.52 0.30 0.67 0.26 0.66 0.40 Zambia 0.84 0.85 0.72 0.76 0.80 0.84 0.84 0.78 Source: Investment Climate Surveys, 2000­4. -- Not available. TABLE A2. Response Rates in Selected Countries, by Ethnicity African European Asian Middle Eastern Country Sales Costs Sales Costs Sales Costs Sales Costs Eritrea 0.53 0.43 0.50 0.50 -- -- -- -- Kenya 0.61 0.61 0.70 0.60 0.58 0.57 -- -- Mozambique 0.70 0.29 0.80 0.39 0.72 0.34 -- -- Nigeria 0.93 0.76 0.84 0.76 0.88 0.88 0.74 0.67 Senegal 0.92 0.46 0.80 0.53 0.82 0.63 -- -- Tanzania 0.86 0.37 0.75 0.44 0.81 0.54 0.90 0.20 Uganda 0.69 0.22 0.50 0.50 0.56 0.32 -- -- Source: Authors' calculations based on Investment Climate Surveys, 2000­4. -- Not available. 226 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE A3. Response Rates in Selected Countries, by Size Micro Small Medium Large Very Large Country Sales Costs Sales Costs Sales Costs Sales Costs Sales Costs Bangladesh 0.68 0.60 1.00 0.91 1.00 0.94 1.00 0.89 1.00 0.94 Bolivia 0.22 0.11 0.86 0.58 0.87 0.72 0.83 0.67 0.87 0.78 China 1.00 0.38 0.99 0.69 1.00 0.75 1.00 0.83 1.00 0.82 Eritrea 0.75 0.25 0.46 0.39 0.43 0.43 0.69 0.63 0.50 0.50 Ethiopia 0.97 0.90 1.00 0.91 1.00 0.93 1.00 1.00 1.00 0.82 India 1.00 0.65 1.00 0.53 1.00 0.80 1.00 0.86 1.00 0.82 Kenya 0.30 0.20 0.59 0.62 0.83 0.74 0.65 0.59 0.62 0.50 Morocco 1.00 1.00 1.00 0.95 1.00 0.93 1.00 0.95 1.00 0.93 Mozambique 0.73 0.29 0.84 0.42 0.57 0.35 0.65 0.31 1.00 0.50 Nicaragua 1.00 0.48 1.00 0.63 1.00 0.82 1.00 0.86 1.00 0.67 Nigeria 0.50 0.17 0.86 0.57 0.89 0.84 0.93 0.87 0.97 0.91 Senegal 0.80 0.33 0.90 0.49 0.92 0.45 0.85 0.47 0.91 0.46 Tanzania 0.85 0.10 0.86 0.33 0.81 0.48 0.90 0.66 0.82 0.29 Uganda 0.70 0.06 0.67 0.13 0.56 0.12 -- -- -- -- Zambia 0.10 0.10 0.79 0.80 0.87 0.84 0.89 0.83 0.85 0.92 Source: Investment Climate Surveys, 2000­4. -- Not available. TABLE A4. Value-Added and Capital per Worker in Selected Countries (price-adjusted dollars) Value-added/worker Capital/worker Very Very Country Micro Small Medium Large large Micro Small Medium Large large African countries Eritreaa 12,826 10,836 24,105 8,846 7,077 54,365 65,238 191,843 194,055 53,443 Ethiopia 3,422 3,136 4,277 5,988 3,707 4,515 11,643 17,820 21,860 20,909 Nigeria -- 4,118 4,413 8,384 9,119 -- 42,164 31,500 61,040 46,944 Kenya -- 5,252 9,303 18,906 6,902 31,510 17,005 29,259 25,008 17,005 Mozambique 1,431 5,112 11,451 8,997 -- 9,202 21,130 19,085 41,748 -- Senegal 14,681 17,903 40,819 37,239 34,613 18,798 13,726 22,479 23,772 1,989 Tanzania 3,548 4,862 11,037 8,935 17,870 2,299 12,921 10,402 29,017 28,798 Uganda 3,072 3,321 5,314 15,942 3,155 2,214 4,290 13,008 24,494 2,906 Zambia -- 1,493 1,773 2,333 4,666 -- 15,008 21,773 10,420 21,384 Other countries Bangladesh 3,654 5,000 6,346 4,615 4,423 1,538 5,577 6,346 3,077 4,423 Bolivia 3,023 3,953 7,093 5,698 17,093 2,791 5,000 10,930 12,442 25,698 China 6,329 8,040 14,198 13,171 14,540 6,842 4,154 6,109 6,842 20,527 India 8,810 16,667 15,238 18,095 27,381 7,143 7,143 8,810 13,571 28,571 Morocco -- 13,665 16,165 11,999 17,498 -- 12,444 9,415 4,653 5,519 Nicaragua 7,000 8,167 12,167 34,000 12,333 3,167 4,333 4,167 19,000 4,500 Source: Authors' calculations based on Investment Climate Surveys, 2000­4. a. The exceptionally high capital figures for Eritrea reflect a very low aggregate price level (probably much lower than the price of capital goods specifically) and very large stocks of very old capital equipment. -- Not available. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 227 FIGURE A1. Total Factor Productivity Index in Selected Countries 1.2 1.0 0.8 index 0.6 TFP 0.4 0.2 0 ea India occo EthiopiaNigeriaEritr UgandaZambia TanzaniaBolivia Kenya SenegalChina Mor Mozambique Bangladesh Nicaragua Source: Authors' calculations based on Investment Climate Surveys, 2000­4. Note: Index calculated using nominal prices. TABLE A5. Age Structure of Capital in Selected Countries (percent) Country Less than 10 years More than 10 years India 70 30 Morocco 62 28 Eritrea 48 52 Ethiopia 49 51 Kenya 47 53 Morocco 62 28 Mozambique 49 37 Senegal 67 33 Tanzania 47 53 Uganda 75 25 Zambia 78 22 Source: Investment Climate Surveys, 2000­4. 228 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN TABLE A6. Net Total Factor Productivity in Selected Countries, by Firm Size Country Micro Small, medium Large, very large Bangladesh 0.24 0.23 0.19 Bolivia 0.30 0.22 0.28 China 0.26 0.55 0.50 Eritrea 0.17 0.20 0.15 Ethiopia 0.07 0.10 0.06 India 0.26 0.39 0.42 Kenya 0.26 0.22 0.33 Mozambique 0.05 0.15 0.32 Morocco -- 0.55 0.53 Nicaragua 0.33 0.43 0.82 Nigeria -- 0.13 0.21 Senegal 0.26 0.64 0.61 Tanzania 0.17 0.32 0.39 Uganda 0.17 0.16 0.28 Zambia -- 0.08 0.41 Source: Authors' calculations based on Investment Climate Surveys, 2000­4. -- Not available. TABLE A7. Distribution of Indirect Costs by Type, Kenya (percent) Category Share Transport 31.9 Energy 18.5 Indirect labor costs 10.0 (payroll, administration) Security 6.4 Telecommunications 4.5 Land 2.5 Bribes 1.7 Water 0.9 Waste disposal 0.4 Other 21.5 Source: Authors' calculations based on Investment Climate Surveys, 2000­4. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 229 Notes 1. For estimates of financial and human capital flight, see Collier, Hoeffler, and Patillo (1999). While land is abundant, distortions in the regulatory and legal environments have resulted in very high prices for land in many countries. 2. See the World Development Report 2005 (World Bank 2004b) and the series of Invest- ment Climate Assessments published by the World Bank between 1999 and 2004. See also http://www.worldbank.org/rped and http://www.fias.net. 3. Montobbio (2002) analyzes structural change from the perspective of evolutionary eco- nomics. He finds that with firm-level heterogeneity in unit costs, sorting and selection driven by competition and product substitutability drive a process of structural change. Although this approach has very different analytical foundations than endogenous growth and trade theory, one of its fundamental insights--that an economy's relatively more productive firms and sectors tend to become more important over time--is similar. 4. Wood and Mayer (2001) argue that Africa's skills deficit and relative abundance in natu- ral resources condemns the continent to primary product exports for the foreseeable future. 5. Links between Africa and other regions were made by comparing prices with the United States for a limited range of products, not always perfectly matched in quality. China and India were linked through regression procedures based on income and secondary education. 6. As noted above, these estimates are subject to considerable error. 7. Technology can sometimes enable a wider range of firms to overcome high domestic cost structures. Installing their own communications systems enabled Indian software and data-processing firms to bypass ineffective and costly telecommunications systems and build on a strong base of cheap, highly trained, English-speaking labor. But such cases are likely to be rare. 8. For more information, see http://www.worldbank.org/rped or http://rru.worldbank.org. 9. Other types of costs and risks related to the business environments are beyond the scope of this paper (see World Bank 2004b). 10. The World Bank's Country Policy and Institutional Assessments (CPIA) rates Senegal, Tanzania, and Uganda in the top tercile in Africa. Mozambique also receives a high rat- ing, but it is weaker in some areas, especially the financial sector. 11. For a comparative review of some of these countries, see Devarajan, Dollar, and Holm- grin (2001). 12. The usual concern is that knowing its level of productivity Ai, firm i will choose to use more flexible inputs (for example, Li and Mi), so that the ordinary least squares estimates of a and b will be biased. Methods of dealing with the problem include obtaining panel data (a moot point for the analysis here), using instrumental variables (which has short- comings), and adopting the structural approaches taken by Olley and Pakes (1996) and Levinsohn and Petrin (2003), which are subject to substantial problems (Ackerberg, Caves, and Frazer 2005). This study does not replicate these approaches. 13. One question of potential concern is the possibility of systematic bias in the response rates to questions on sales and costs. Response rates do differ across countries, but within countries they are remarkably uniform across categories of firms (domestic/foreign, eth- nic/indigenous, exporter/nonexporter) that are known to correlate strongly with produc- tivity. The only strong pattern in response rates is that micro firms (those with fewer than 10 employees) tend to respond less often, which suggests that response rates to detailed sales and costs questions may have more to do with accounting and capacity. Field expe- rience suggests that minority firms are likely to understate sales. If true, this would accen- tuate ethnic productivity gaps shown in the data. Although selection bias is always a con- cern in any survey, it is unlikely to pose a major problem for the broad pattern of results. 14. A second and related point is that firms with substantial product or market power likely face higher output prices and thus artificially appear to be more productive. Causation also 230 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN runs in the other direction--more productive firms will likely win a larger share of their markets--so an appropriate approach for controlling for market power requires a multi- stage instrumental variables approach. Unfortunately, the quality of the market share data and the availability of instruments are poor. Production function estimations are per- formed with a very imperfect measure--self-reported market share--included directly as an independent variable. The pooled results suggest that the combined relationship (with causation in both directions) is strong, with the difference between near-zero market share and 25 percent market share associated with an output price differential of 9 percent and a 50 percent share associated with an output price differential of 13 percent. Much stronger effects were found in some African countries. 15. Nominal prices were used to re-estimate the results to understand the effect of the price corrections. The results in adjusted prices may better reflect underlying firm characteris- tics; the results in nominal prices may better reflect how firms are actually doing, in the sense that local price levels determine profits, holding firm characteristics constant. The patterns in the results are similar, but the results for countries with high price levels (espe- cially Zambia, Senegal, and Tanzania) appear somewhat stronger using nominal figures. 16. The "most competitive" countries using this benchmark appear to be Eritrea and Nige- ria; Mauritius (Africa's only major manufacturing success story) and Uganda (which experienced rapid growth in manufacturing in the 1990s) have relatively high unit labor costs. 17. Some of this literature focuses on the propensity to export, which is found to be strongly correlated with productivity measures (Clarke 2004; Dollar, Hallward-Driemeier, and Mengistae 2003). 18. In the data, productivity regressions on business environment variables such as time to enforce a contract, number of inspector visits, and the percentage of senior management time spent dealing with regulation do not produce strong results. 19. Recent work by Escribano and Guasch (2005) provides useful methodological bases that point in this direction. 20. Further econometric analysis on firm-level TFP with a set of business environment vari- ables on the right-hand side may advance the state of knowledge on productivity (Escrib- ano and Guasch 2005). 21. Transport costs are often excluded from measures of productivity, because they do not directly affect the productivity of the firm. The estimation here of net value-added lies somewhere between the strict definitions of productivity and profitability. Consequently, the full set of indirect costs faced by the firm is included. 22. Note that energy is included in our definition of net value-added by including it in indi- rect costs (rather than including it in raw materials), so our term gross value-added does not quite correspond to studies that include energy costs in raw materials. In the rest of this discussion, our use of the term "indirect costs" includes the cost of energy. 23. Some firms in the sample that have positive gross value-added have high enough indirect costs that their net value-added is negative. These firms were dropped from the net TFP regression. This biases the estimated gap between net and gross TFP downward, because firms with low gross TFP are dropped. To correct for this, the estimated average net TFP level by country is corrected for the number of firms for which net TFP is essentially zero. 24. Trade reforms in Africa, for example, have been driven by adjustment programs negoti- ated with the Bretton Woods institutions rather than a reciprocal process of negotiation with other countries to open market access. It is therefore not surprising to see the persist- ence of widespread impediments to exporting firms despite declines in levels of protection. 25. Nellis (2005) discusses the difficult political economy of privatization in Africa. 26. In some countries, ethnic fragmentation between indigenous groups is also an issue. For more on this in Ethiopia, see Mengistae (2001). 27. Stochastic frontier analyses for individual countries show that small firms in Africa are BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 231 well below the production frontier. 28. See the World Bank's Regional Program on Enterprise Development Country Studies, conducted since 1995. 29. The World Bank has made several loans to Kenya for road improvements, with little to show for it (Easterly 2003). 30. The World Development Report 2005 highlights individual country experiences in which firms in EPZs developed backward links to small suppliers. References Ackerberg, Daniel, Kevin Caves, and Garth Frazer. 2005. "Structural Identification of Production Functions." Department of Economics Seminar Series, University of California, San Diego. Batsos, Fabiano, and John Nasir. 2004. "Productivity and the Investment Climate: What Mat- ters Most?" World Bank Policy Research Working Paper 3335, Washington, DC. Biggs, Tyler, and Manju Shah. 2004. "African SMEs, Networks, and Manufacturing Perfor- mance." World Bank, Africa Region Private Sector Unit, Washington, DC. Biggs, Tyler, Manju Kedia Shah, and Pradeep Srivastava. 1995. "Technological Capabilities and Learning in African Enterprises." RPED Working Paper AFT288, World Bank, Regional Program on Enterprise Development, Washington, DC. Bigsten, Arne, Paul Collier, Stefan Dercon, Marcel Fafchamps, Bernard Gauthier, Jan Willem Gun- ning, Jean Habarurema, Abena Oduro, Remco Oostendorp, Catherine Pattillo, Mans Soder- bom, Francis Teal, and Albert Zeufack. 2000. "Exports and Firm-Level Efficiency in African Manufacturing." The Centre for the Study of African Economies Working Paper Series, Working Paper 126, September 26. Available at http://www.bepress.com/csae/paper126. Birdsall, Nancy. 2004. "New Issues in Development Assistance." Center for Global Develop- ment, Washington, DC. Bratton, Michael, Robert Mattes, and E. Gyimah-Boadi. 2005. Public Opinion, Democracy, and Market Reform in Africa. Cambridge: Cambridge University Press. Burgess, Robin, and Anthony Venables. 2004. "Towards a Microeconomics of Growth." World Bank Policy Research Working Paper 3257, Washington, DC. Byrd, William, and Alan Gelb. 1990. "Why Industrialize? The Incentives for Rural Commu- nity Governments." In China's Rural Industry: Structure, Development and Reform, ed. William Byrd and Lin Qingsong. New York: Oxford University Press. Cadot, O., and J. Nasir. 2001. "Incentives and Obstacles to Growth: Lessons from Manufac- turing Case Studies in Madagascar." RPED Working Paper No. 117, World Bank, Regional Program on Enterprise Development, Washington, DC. Chenery, Hollis, and Moshe Syrquin. 1975. Patterns of Development: 1950­1970. Oxford: Oxford University Press. Clarke, George. 2004. "Why Don't African Manufacturing Enterprises Export More? The Role of Trade Policy, Infrastructure Quality and Enterprise Characteristics." World Bank, Africa Region Private Sector Unit, Washington, DC. Collier, Paul. 2000. "Africa's Comparative Advantage." In Industrial Development and Pol- icy in Africa, ed. H. Jalilian, M. Tribe, and J. Weiss. Cheltenham, U.K.: Edward Elgar. Collier, Paul, and Jan Gunning. 1997. "Explaining African Economic Performance." CSAE WPS/97-2.2, Centre for the Study of African Economies, University of Oxford, Oxford, U.K. 232 | BENN EIFERT, ALAN GELB, AND VIJAYA RAMACHANDRAN Collier, Paul, Anke Hoeffler, and Catherine Patillo. 1999. "Flight Capital as a Portfolio Choice." World Bank Working Paper 2066, Washington, DC. Devarajan, Shanta, David Dollar, and T. Holmgren. 2001. Aid and Reform in Africa. Wash- ington, DC: World Bank. Dollar, David, Mary Hallward-Driemeier, and Taye Mengistae. 2003. "Investment Climate and Firm Performance in Developing Economies." Development Research Group, World Bank, Washington, DC. Easterly, William. 2002. "The Cartel of Good Intentions: The Problem of Bureaucracy in For- eign Aid." Journal of Policy Reform 5 (4): 223­50. ------. 2003. "Can Foreign Aid Buy Growth?" Journal of Economic Perspectives 17 (3): 23­48. Eifert, Benn, and Vijaya Ramachandran. 2004. "Competitiveness and Private Sector Develop- ment in Africa: Cross-Country Evidence from the World Bank's Investment Climate Data." Occasional Paper Series, Institute for African Studies, Cornell University, Ithaca, NY. Emery, James. 2003. "Governance and Private Investment in Africa." In Beyond Structural Adjustment: The Institutional Context of African Development, ed. N. van de Walle, N. Ball, and V. Ramachandran. New York: Palgrave Macmillan. Escribano, A., and J. L. Guasch. 2005. "Assessing the Impact of the Investment Climate on Productivity Using Firm-Level Data: Methodology and the Cases of Guatemala, Honduras and Nicaragua." Policy Research Working Paper No. 3621, World Bank, Washington, DC. Fafchamps, Marcel. 2004. Market Institutions in Sub-Saharan Africa: Theory and Evidence. Cambridge, MA: MIT Press. Gelb, Alan H., and Gene Tidrick. 2000. "Growth and Job Creation in Africa." In Strategies for Growth and Job Creation in Southern Africa, ed. Alan H. Gelb. Gaborone, Botswana: Friedrich Ebert Stiftung. Grossman, Gene, and Elhanan Helpman. 1990. "Comparative Advantage and Long-Run Growth." American Economic Review 80 (4): 796­815. Hausmann, Ricardo, and Dani Rodrik. 2002. "Economic Development as Self-Discovery." CEPR Discussion Paper No. 3356, Centre for Economic Policy Research, London. Krugman, Paul. 1980. "Scale Economies, Product Differentiation, and the Pattern of Trade." American Economic Review 70 (5): 950­9. ------. 1981. "Intraindustry Specialization and the Gains from Trade." Journal of Political Economy 89 (5): 959­73. ------. 1983. "New Theories of Trade among Industrialized Countries." American Economic Review 73 (2): 343­7. ------. 1991. "Increasing Returns and Economic Geography." Journal of Political Economy 99 (3): 483­99. Levinsohn, J., and A. Petrin. 2003. "Estimating Production Functions Using Inputs to Control for Unobservables." Review of Economic Studies 70 (2): 317­41. Levy, Brian, and Sahr Kpundeh, eds. 2004. Building State Capacity in Africa: New Approaches, Emerging Lessons. World Bank Institute Development Studies, World Bank, Washington, DC. Mengistae, Taye. 2001. "Indigenous Ethnicity and Entrepreneurial Success in Africa: Some Evidence from Ethiopia." World Bank Policy Research Working Paper 2534, Washington, DC. Mengistae, Taye, and Catherine Pattillo. 2002. "Export Orientation and Productivity in Sub- Saharan Africa." IMF Working Paper WP/02/89, International Monetary Fund, Washing- ton, DC. BUSINESS ENVIRONMENT AND COMPARATIVE ADVANTAGE IN AFRICA | 233 Montobbio, Fabio. 2002. "An Evolutionary Model of Industrial Growth and Structural Change." Structural Change and Economic Dynamics 13 (4): 387­414. Nellis, John. 2005. "Privatization in Africa: What Has Happened? What Is to Be Done?" FEEM Working Paper No. 127.05, Fondazione Eni Enrico Mattei, Milan. Available at http://ssrn.com/abstract=384421. Olley, S., and A. Pakes. 1996. "The Dynamics of Productivity in the Telecommunications Equipment Industry." Econometrica 64 (6): 1263­95. Sala-i-Martin, Xavier, and Elsa Artadi. 2003. "The Tragedy of the Twentieth Century: Growth in Africa." Columbia University Economics Discussion Paper 0203-17, New York. Söderbom, Måns, and Francis Teal. 2003. "Are Manufactured Exports the Key to Economic Success in Africa?" Journal of African Economies 12 (1): 1­29. Tangri, Roger. 1999. The Politics of Patronage in Africa. Trenton, NJ: Africa World Press. van de Walle, Nicolas. 2001. African Economies and the Politics of Permanent Crisis, 1979­1999. Cambridge: Cambridge University Press. Venables, Antony, and Nuno Limao. 1994. "Infrastructure, Geographical Disadvantage and Transport Costs." Policy Research Working Paper No. 2257, World Bank, Washington, DC. Winters, Alan, and Pedro Martins. 2004. "When Comparative Advantage Is Not Enough: Business Costs in Small, Remote Economies." World Trade Review 3 (3): 1­37. Wood, Adrian, and Kersti Berge. 1997. "Exporting Manufactures: Human Resources, Natural Resources, and Trade Policy." Journal of Development Studies 34 (October): 35­59. Wood, Adrian, and Jorg Mayer. 2001. "Africa's Export Structure in a Comparative Perspec- tive." Cambridge Journal of Economics 25 (3): 369­94. World Bank. Various years. Investment Climate Assessments. Available at www.worldbank.org/ privatesector/ic/ic_ica.htm and www.worldbank.org/rped. ------. Various years. Investment Climate Surveys. ------. Various years. World Development Indicators. Washington, DC: World Bank. ------. 2004a. Doing Business 2005. New York: Oxford University Press. ------. 2004b. World Development Report 2005: A Better Investment Climate for Everyone. New York: Oxford University Press. Comment on "Business Environment and Comparative Advantage in Africa: Evidence from the Investment Climate Data," by Benn Eifert, Alan Gelb, and Vijaya Ramachandran LÉONCE NDIKUMANA The paper by Eifert, Gelb, and Ramachandran represents a significant contribution to the research on the causes of low economic performance in Africa. Using a rich set of international survey­based data on the investment climate, the paper advances our understanding of the factors of low productivity in the private sector in Africa rela- tive to other countries. The quantitative analysis of the data offers a useful basis for discussion of policy reforms and strategies aimed at improving the business environ- ment in Africa. The paper frames the analysis of the determinants of total factor productivity in the context of three theories of comparative advantages. The first emphasizes factor endowments in shaping the patterns, composition, and evolution of trade. The sec- ond focuses on the vital role of the quantity and quality of public services and infra- structure for private sector activity. The third emphasizes the role of firm entry in reaping the benefits from scale and learning externalities that are essential for indus- trial development. The paper singles out the last two theories as more informative with regard to the causes of low productivity growth in the African private sector. The two theories highlight indirect costs in production and trade as a major determinant of total fac- tor productivity. Most important, indirect costs are a dimension of the firm's cost structure that may be more directly influenced by policy reforms. This aspect of the cost of business operations has been overlooked in traditional industrial organization research. This line of analysis is therefore most welcome not only for the innovations it brings forth in the area of research but also for the impact it may have on the debate on policy reforms aimed at boosting productivity in the private sector. The paper underlines several factors that contribute to an unfavorable investment climate in Sub-Saharan Africa relative to other regions. These factors include fixed costs related to physical constraints, such as geography, as well as constraints arising from the effects of inefficient public infrastructure, poor governance, and unsuccess- ful economic reforms. The paper raises the serious issue of the political economy of Léonce Ndikumana is associate professor of economics at the University of Massachusetts, Amherst. Annual World Bank Conference on Development Economics 2006 © 2006 The International Bank for Reconstruction and Development / The World Bank 235 236 | LÉONCE NDIKUMANA economic reforms and uncovers some troubling dilemmas and paradoxes. The data examined in the paper indicate that, contrary to what market-based theory would predict, private actors tend to oppose reforms and prefer the government to markets. This may seem inconsistent with the fact that private businesses incur high costs asso- ciated with dealing with government regulation. The key to this apparent paradox lies in what the paper calls "elite capture of the private sector," as well as the disen- chantment of the population with privatization in many countries. On the one hand, private actors and government officials have developed networks of rent-seeking that are beneficial to the direct participants (but costly to society as a whole). On the other hand, the private sector has not yet evolved as a viable substitute for the gov- ernment as a source of job creation and welfare. In a sense, the apparent preference for government over the private sector expressed in these surveys is merely a state- ment of the underdeveloped state of the private sector. Two implications follow from this analysis. First, reform programs must aim at promoting competition in the pri- vate sector, especially by removing barriers to entry. Second, the regulatory environ- ment must be simplified and made more transparent to reduce the opportunities for rent-seeking by government officials. This paper and other studies have identified high labor costs in African economies as an important constraint to business expansion and foreign investment. High labor costs are usually blamed for the relatively lower labor productivity in Africa. This paper makes two very pertinent--and novel--points with regard to the labor costs and productivity in the private sector in Africa. First, the data show that high indi- rect costs contribute more to low labor productivity than high labor costs. Most of these indirect costs are directly related to the ineffectiveness of public service delivery and the inefficiency of business regulation. The implication is that policies that emphasize labor cost compression will have limited success in increasing productiv- ity in the private sector. What is needed is a concerted effort to improve governance, the regulatory environment, public infrastructure, and public service delivery to min- imize indirect costs. The second point advanced in the paper is that while labor costs may be higher in Africa than in other regions, the purchasing power of wages is also significantly lower. Wages have not kept pace with the rising prices of goods and services, caus- ing workers' living standards to deteriorate (even as unit labor costs increased). These price dynamics hurt business operations and reduce the living standards of working people. The implication is that there is an urgent need to establish system- atic mechanisms of tripartite negotiations among governments, businesses, and workers over "incomes policies," that is, the rules of price and wage adjustments. The objective is to seek a compromise on pricing rules that protect firm's profits without eroding workers' living standards or wage rules that protect workers' living standards without prohibitively increasing production costs. An important tool for keeping labor costs low is investing in skills development through formal and informal education as well as on-the-job training to raise labor productivity. South Africa has made significant progress in tripartite nego- tiations, with the establishment of an institutional framework for multilateral bar- COMMENT ON EIFERT, GELB, AND RAMACHANDRAN | 237 gaining, coordinated by the National Economic Development and Labor Council. This approach institutionalizes a culture of private-public partnerships in policy negotiations, which promotes transparency in policy making while minimizing counterproductive labor relations disputes. Although this mechanism has its limi- tations, other African countries may draw useful lessons from the South African experience. The data examined in the paper indicate that capital intensity of production tends to be higher in Africa than in other regions. At the same time, capital is inef- ficiently utilized, as indicated by low levels of capacity utilization. The evidence underscores the two critical forms of inefficiencies in the use of resources in African economies. First, while these countries are labor surplus economies, they have dis- proportionately relied on physical capital relative to human capital. Second, poli- cies justifiably aimed at promoting technological progress have generated excessive capitalization, thereby wasting resources. As a result, economic growth has not been accompanied by significant job creation. For African countries to achieve job- creating growth, they need to promote an incentive structure that encourages labor-intensive production methods in the private sector. A strategy aimed at pro- moting labor-intensive production will not only reduce inefficiencies in the use of resources, it will also allow for a more equitable distribution of the growth divi- dends. The ultimate objective is to promote greater integration of human capital in the modernization process. The paper highlights limited access to credit as another major impediment to pro- ductivity growth in the private sector, especially for very small (micro) firms. The evi- dence suggests that credit constraints are severely binding, especially for small firms and microenterprises. Lack of access to credit by these firms is usually explained by the relatively high perceived risk associated with lending to them. However, other factors contribute to the shortage of credit to microenterprises in particular and the private sector in general. One important factor is the lack of competition in the bank- ing sector, which distorts the incentive structure. African banking systems tend to be oligopolistic, which promotes "conservative" and usurious lending practices. Lack of competition allows banks to reap high mark-ups with large interest spreads and high costs of borrowing (see the article by Senbet and Otchere in this volume). Banks tend to favor high-turnover (often speculative) activities in their credit supply while rationing investment loans to new and small firms. This state of affairs in the banking sector has major implications for private sector development, employment creation, and long-run growth. The small and microenterprise sector is vital for private sector development in Africa, because it is the most adapted to the skills endowment of the region. Large and highly capitalized firms tend to rely on sophisticated technology and require advanced managerial skills that are in short supply in Africa. In contrast, small and microenterprises can be eas- ily established and managed with the existing levels of skills. African countries should therefore promote credit allocation mechanisms that facilitate access to finance for new and existing small and microenterprises. The role of the government will be vital in this endeavor. 238 | LÉONCE NDIKUMANA One strategy that can have immediate effects on credit allocation is government loan guarantees, which reduce the effective risk faced by lenders. In advanced finan- cial systems, government-owned direct and indirect lending institutions constitute a major facilitator for access to credit for households and targeted private sector activ- ities (such as agriculture). The idea of loan guarantees in developing countries has fallen out of fashion, due mainly to concerns about inefficiencies in the operation of existing schemes and the burden on government budget. However, in countries in which loan guarantee systems operate effectively, they allow private lending institu- tions to better mobilize resources and allocate credit, which eventually improves access to credit and boosts economic activity. On its own, private banking will not be able to mobilize enough funds for private sector development. Indirect govern- ment lending is an effective and feasible means of addressing the issue of lack of access to finance in African economies. Overall, the analysis in this study suggests that the dividends from economic reforms are likely to be high, notably through reducing indirect costs of business operations, lowering uncertainty, and thus raising productivity in the private sector. The credibility of reforms and the low risk of policy reversal are critical to ensure the predictability of the investment climate (see Serven 1997; Mlambo and Oshikoya 2001). More than the level of indirect production costs, it is their predictability that matters most for private operators. The implication is that countries that are able to establish a record of macroeconomic stability and business-friendly reforms will be able to reap higher gains in productivity in the private sector. References Mlambo, K., and T. W. Oshikoya. 2001. "Macroeconomic Factors and Investment in Africa." Journal of African Economies 10 (AERC Supplement 2): 12­47. Serven, L. 1997. "Irreversibility, Uncertainty and Private Investment: Analytical Issues and Some Lessons for Africa." Journal of African Economies 6 (3) (AERC Supplement): 229­68. T he Annual World Bank Conference on Development Economics (ABCDE) brings together the world's leading scholars and development practitioners for a lively debate on state-of-the-art thinking in development policy and the implications for the global economy. The 17th conference was held in Dakar, Senegal, on January 27, 2005. The theme of the conference was growth and integration, which was divided into five topics: growth and integration, financial reforms, economic development, trade and development, and investment climate. IN THIS VOLUME Introduction by François Bourguignon and Boris Pleskovic; welcome address by His Excellency Abdoulaye Wade, President of the Republic of Senegal; opening address by François Bourguignon; keynote address by Abdoulaye Diop; papers by Augustin Fosu and Stephen O'Connell, Lemma W. Senbet and Isaac Otchere, Jean-Claude Berthélemy, Lawrence E. Hinkle and Richard S. Newfarmer, and Benn Eifert, Alan Gelb, and Vijaya Ramachandran; and comments by Jean-Paul Azam, Abdoulaye Diagne, Mthuli Ncube, and Léonce Ndikumana. ISBN 0-8213-6093-0