Post-Conflict Economies in Africa (International Economic Association)

Post-Conflict Economies in Africa Edited by Augustin Kwasi Fosu and Paul Collier Post-Conflict Economies in Africa Th...

Author: Paul Collier | Augustin Kwasi Fosu

29 downloads 1262 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Post-Conflict Economies in Africa Edited by

Augustin Kwasi Fosu and Paul Collier

Post-Conflict Economies in Africa This is IEA conference volume no. 140

This page intentionally left blank

Post-Conflict Economies in Africa Edited by

Augustin Kwasi Fosu African Economic Research Consortium, Nairobi and Economic Commission for Africa, Addis Ababa

and

Paul Collier St Antony’s College, Oxford, and Centre for the Study of African Economies University of Oxford

in association with the International Economic Association

© International Economic Association 2005 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2005 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin’s Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 1–4039–4346–X This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Post-conflict economies in Africa / edited by Augustin Kwasi Fosu and Paul Collier. p. cm. — (International Economic Association series) Originally presented at a workshop in Kampala, Uganda, in July 2002 in colloboration with the World Bank and the International Economic Association. Includes bibliographical references and index. ISBN 1–4039–4346–X (cloth) 1. Africa—Economic policy—Congresses. 2. Political stability—Africa— Congresses. 3. Africa—Politics and government—Congresses. I. Fosu, Augustin Kwasi. II. Collier, Paul. III. World Bank. IV. International Economic Association. V. Series. HC800.P677 2004 330.96—dc22 2004054466 10 9 8 7 6 5 4 3 2 1 14 13 12 11 10 09 08 07 06 05 Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

Contents

The International Economic Association

vii

Preface and Acknowledgements

ix

List of Participants and Contributors

x

List of Abbreviations and Acronyms

xii

Introduction: Post-Conflict Economies in Africa Augustin Kwasi Fosu

xv

Part I

Thematic Issues

1 Post-Conflict Reconstruction in Africa: Some Analytical Issues Tony Addison and S. Mansoob Murshed 2 The Economic and Political Consequences of Conflict and Implications for Post-Conflict Recovery in Africa Ibrahim A. Elbadawi and Njuguna S. Ndung’u

3

18

3 Economic Policy in Post-Conflict Societies Paul Collier

45

4 Ethnicity, Institutions of Governance and Conflict Avoidance Mwangi S. Kimenyi

57

Part II

Case Studies: West Africa

5 Liberia and Sierra Leone: Interwoven Civil Wars Victor A. B. Davies

77

6 The Nigerian Civil War: Causes and the Aftermath Abdul-Ganiyu Garba and P. Kassey Garba

91

7 The Economics of Civil Conflict in Africa: The Case of Chad Nadjiounoum Djimtoingar and Djona Atchénémou Avocksouma

Part III

109

Case Studies: The Horn, East and Central Africa

8 Conflict, Post-Conflict and Economic Performance in Ethiopia Alemayehu Geda and Befekadu Degefe v

125

vi

Contents

9 Prospects for Sustainable Peace and Post-Conflict Economic Growth in the Sudan Ali Abdel Gadir Ali and Ibrahim A. Elbadawi 10 Entrenching Peace in Post-Conflict Economies: The Case of Uganda Michael Atingi-Ego and Rachel Kaggwa Sebudde 11 The Challenge of Transition from War to Peace in Burundi Janvier Désiré Nkurunziza and Floribert Ngaruko

Part IV

163 180

Case Studies: Southern Africa

12 The Political Economy of Post-Conflict Economic Recovery: Making Peace Substantial in Mozambique Laudemiro Francisco and Edwin Connerley 13 Transformation for Post-Conflict Angola Fátima Moura Roque

Part V

143

197 213

Conclusion

14 Post-Conflict Economies in Africa: Synthesis and Lessons Augustin Kwasi Fosu

231

Index of Names

241

Subject Index

245

The International Economic Association A non-profit organization with purely scientific aims, the International Economic Association (IEA) was founded in 1950. It is a federation of some sixty national economic associations in all parts of the world. Its basic purpose is the development of economics as an intellectual discipline, recognizing a diversity of problems, systems and values in the world and taking note of methodological diversities. The IEA has, since its creation, sought to fulfil that purpose by promoting mutual understanding among economists through the organization of scientific meetings and common research programmes, and by means of publications on problems of fundamental as well as of current importance. Deriving from its long concern to assure professional contacts between East and West, and North and South, the IEA pays special attention to issues of economies in systemic transition and in the course of development. Since its inception it has organized more than a hundred round-table conferences for specialists on topics ranging from fundamental theories to methods and tools of analysis and major problems of the present-day world. Participation in round tables is at the invitation of a specialist programme committee, but thirteen triennial World Congresses have regularly attracted the participation of individual economists from all over the world. The Association is governed by a Council, composed of representatives of all member associations, and by a fifteen-member Executive Committee elected by the Council. The Executive Committee (2002–2005) at the time of the Lisbon Congress was: President Vice-President Treasurer Past President President-Elect Other Members

Professor János Kornai, Hungary Professor Bina Agarwal, India Professor Jacob Frenkel, Israel Professor Robert Solow, USA Professor Guillermo Calvo, Argentina Professor Maria Augusztinovics, Hungary Professor Eliana Cardoso, World Bank Professor Duardo Engel, Chile Professor Heba Handoussa, Egypt Professor Michael Hoel, Norway Professor Jean-Jacques Laffont, France Professor Andreu Mas-Colell, Spain Professor Kotaro Suzumura, Japan vii

viii

The International Economic Association

Advisers Secretary-General General Editor

Professor Alessandro Vercelli, Italy Professor Fiorella Kostoris Padoa Schioppa, Italy Professor Vitor Constancio, Portugal Professor Jean-Paul Fitoussi, France Professor Michael Kaser, UK

Sir Austin Robinson was an active Adviser on the publication of IEA Conference proceedings from 1954 until his final short illness in 1993. The Association has also been fortunate in having secured many outstanding economists to serve as President: Gottfried Haberler (1950–53), Howard S. Ellis (1953–56), Erik Lindahl (1956–59), E. A. G. Robinson (1959–62), Ugo Papi (1962–65), Paul A. Samuelson (1965–68), Erik Lundberg (1968–71), Fritz Machlup (1971–74), Edmund Malinvaud (1974–77), Shigeto Tsuru (1977–80), Victor L. Urquidi (1980–83), Kenneth J. Arrow (1983–86), Amartya Sen (1986–89), Anthony B. Atkinson (1989–92), Michael Bruno (1992–95), Jacques Drèze (1995–99) and Robert M. Solow (1999–2002). The activities of the Association are mainly funded from the subscriptions of members and grants from a number of organizations. Support from UNESCO since the Association was founded, and from its International Social Science Council, is gratefully acknowledged.

Preface and Acknowledgements In recent years there has been considerable conflict in the form of civil wars across the African continent. To inform policy-makers and analysts, the African Economic Research Consortium (AERC) has focused on post-conflict economies as part of its research portfolio under the Special Workshop (SW) modality, which provides research on a topical policy issue for dissemination in a relatively short period of time. The chapters in this volume were originally presented as papers at a Special Workshop in Kampala in July 2002, the result of collaboration with the World Bank, which has also worked extensively on conflict, and the International Economic Association (IEA), which provides a forum for the analysis and dissemination on important economic issues of international relevance. The volume consists of five parts comprising a set of thematic papers to inform the reader of the various pertinent issues on post-conflict economics; a number of country case studies drawn from all regions of sub-Saharan Africa, in order to shed light on the Africa-specific setting; and a section on conclusions. Thus the volume uniquely and richly provides in a complementary format both theoretical and empirical analyses. Many individuals and organizations have contributed to the success of the project. However, specific mentions are in order. First and foremost, the AERC, through its Advisory Committee, has been instrumental in identifying the topic and executing the research project. Second, the collaboration of the World Bank has been important in bringing the project to fruition. Third, the IEA has served as an important partner through which the quality of the product has been enhanced and the output disseminated appropriately. Much gratitude is due to Jean-Paul Fitoussi of the IEA and the Scientific Program Committee, comprising Eliana Cardoso, Paul Collier, Augustin Fosu (chair) and Mustapha Nabli. Further thanks are due to Heywood Hadfield and Michael Kaser, IEA General Editor, who supervised the editorial process. It is hoped that the present volume will contribute significantly to the salient policy dialogue on how to achieve an enduring post-conflict economy for growth and development, particularly in Africa, where the challenge seems so enormous and the imperative so real.

ix


*Professor Tony Addison, UNU/WIDER, Helsinki, Finland *Professor Alemayehu Geda, Addis Ababa University, Ethiopia, and School of Oriental and African Studies, London, UK *Dr Ali Abdel Gadir Ali, Arab planning Institute, Kuwait *Dr Michael Atingi-Ego, Bank of Uganda, Kampala, Uganda *Professor Djona Atchénémou Avocksouma, University of N’Djamena, Chad *Professor Befekadu Degefe, Centre for Economic Studies and Policy Analysis, Ethiopia Professor Eliana A. Cardoso, Georgetown University, Washington, DC, USA *Professor Paul Collier, St Antony’s College, Oxford, UK and the Centre for the Study of African Economies, Oxford, UK *Dr Edwin Connerley, United States Agency for International Development (USAID), Washington, DC, USA *Mr Victor A. B. Davies, Oxford University, UK Mr Nadjiounoum Djimtoingar, Ministère du Plan, du Développement et de la Coopération, N’Djamena, Chad *Dr Ibrahim A. Elbadawi, World Bank, Washington, DC, USA *Professor Augustin Kwasi Fosu, African Economic Research Consortium (AERC), Nairobi, Kenya, and Economic Commission for Africa, Addis Ababa, Ethiopia *Mr Laudemiro Francisco, Howard University, Washington, DC, USA *Dr Abdul-Ganiyu Garba, Ahmadu Bello University, Zaria, Nigeria *Dr P. Kassey Garba, University of Ibadan, Nigeria Professor Mukwanason A. Hyuha, African Economic Research Consortium (AERC), Nairobi, Kenya Ms Darlison Kaija, Makerere University, Kampala, Uganda Ms Rehema Kakonge, Uganda Manufacturers’ Association, Kampala, Uganda *Professor Mwangi S. Kimenyi, Kenya Institute for Public Policy Research and Analysis (KIPPRA), Nairobi, Kenya Mr Rubain Batila Kissita, Université Marien Ngouabi, Congo Brazzaville Professor William Lyakurwa, African Economic Research Consortium (AERC), Nairobi, Kenya *Professor Fátima Moura Roque, Gabinete de Estudos Africano, Lusófona University, Portugal Mr Paul Mpuga, Makerere University, Kampala, Uganda Mr Adam Mugume, Makerere University, Kampala, Uganda, and Oxford University, UK x


xi

*Dr Syed Mansoob Murshed, Institute of Social Studies (ISS), The Hague, The Netherlands Dr Ngila Mwase, UNDP, Kampala, Uganda Professor Francis Mwega, African Economic Research Consortium (AERC), Nairobi, Kenya Ms Floribert Ngaruko, Université de Nice-Sophia Antipolis, France *Professor Njuguna S. Ndung’u, International Research Centre (IDRC), Nairobi, Kenya *Mr Janvier Désiré Nkurunziza, Harvard University, USA Dr Marios Obwona, Makerere University, Kampala, Uganda Mr John Okidi, Executive Director, Economic Policy Research, Kampala, Uganda Mr Francis Okurut, Uganda Women’s Finance Trust, Kampala, Uganda Mr Nichodemus Rudaheranwa, Makerere University, Kampala, Uganda The Honourable Dr Abel Rwendeire, Minister of State for Trade and Industry, Government of Uganda *Ms Rachel Kaggwa Sebudde, Bank of Uganda, Kampala, Uganda Dr Germina Ssemogerere, Makerere University, Kampala, Uganda *Authors of chapters


ADF AERC AG AIDS AMODEG ANC APC ATU CBE CCFAN CCM CCSRP CFA cif CNDD CNPC CNR CPC CPI CPIA CSIS CSM CSNPD DDR DfID DP DP DRC DUP ECOMOG ECOWAS EPI EPLF EPRDF

Allied Democratic Forces (Uganda) African Economic Research Consortium Action Group (Nigeria) acquired immunodeficiency syndrome Association of Demobilized War Veterans (Mozambique) African National Congress (South Africa) All People’s Congress (Sierra Leone) Anti-Terrorist Unit (Liberia) Commercial Bank of Ethiopia Northern Armed Forces Command Council (Chad) Mozambican Christian Council Petroleum Revenue Control and Surveillance Assembly (Chad) Communauté Financière d’Afrique cost, insurance and freight Conseil National pour la Défense de la Démocratie (Burundi) China National Petroleum Corporation National Recovery Committee (Chad) conflict and post-conflict country Investment Promotion Centre (Mozambique) Country Policy and Institutional Assessment (World Bank) Center for Strategic and International Studies (United States) Supreme Military Council (Chad) National Committee for Peace and Democracy (Chad) disarmament, demobilization and resettlement Department for International Development (UK) Democratic Party (Kenya) Democratic Party (Uganda) Democratic Republic of the Congo Democratic Unionist Party (Sudan) Economic Community of West Africa Monitoring Group Economic Community of West African States elite political instability Eritrean People’s Liberation Front Ethiopian People’s Revolutionary Democratic Front xii


EPRP ESAF FADM FAR FDC FDD FDI FNLA FNL-PALIPEHUTU FORD FRELIMO FRODEBU FROLINAT FRONASA GDN GNPOC GPA GUNT HIV HDI HIPC IAAC ICG IDA IDP IGAD IMF KANU KIPRA KY LRA LTDS LURD MDD MDJT MPLA NAS NCNC NDP NEPAD NGO NIF NPC

xiii

Ethiopian People’s Revolutionary Party Enhanced Structural Adjustment Fund Mozambican Defence Force Forces Armées Rwandaises (Rwanda) Foundation for Democratic Cooperation Forces pour la Défense de la Démocratie (Burundi) foreign direct investment Frente Nacional de Libertação de Angola Front National de Libération-Parti pour la Libération du Peuple Hutu (Burundi) Forum for the Restoration of Democracy (Kenya) Frente de Libertação de Moçambique Front Démocratique Burundi National Liberation Front (Chad) Front for National Salvation (Uganda) Global Data Network The Great Nile Petroleum and Oil Corporation (Sudan) General Peace Agreement (Mozambique) Transitional National Union Government (Chad) human immunodeficiency virus Human Development Index Highly indebted poor country integrated approach to aid coordination (Sierra Leone) International Crisis Group international development assistance internally displaced person Inter-Governmental Agency for Development International Monetary Fund Kenya Africa National Union Kenya Institute for Public Policy Research and Analysis Kabaka Yekka (Uganda) Lord’s Resistance Army (Uganda) long-term development strategy Liberians United for Reconciliation and Democracy Movement for Democracy and Development (Chad) Movement for Democracy and Justice in Chad Movimento Popular para a Libertação de Angola Native Authority System (Nigeria) National Council of Nigeria and the Cameroons National Development Party (Kenya) New Partnership for African Development non-governmental organization National Islamic Front (Sudan) Northern People’s Congress (Nigeria)

xiv


NPFL NPRC NPV NRA NRM NURP OAU OLF PPT-RDA PRC PRE PRES PRS RENAMO RUF SDR SLPP SOD SP SPLA SPLM SSA SSS SSTP TICAD TPLF UN UNAMSIL UNDP UNITA UNLF UNMIL UNOMOZ UPC USA USAID USSR

National Patriotic Front of Liberia National Provisional Ruling Council (Sierra Leone) net present value National Resistance Army (Uganda) National Resistance Movement (Uganda) Northern Uganda Reconstruction Programme Organization of African Unity Oromo Liberation Front Chad Progressive Party People’s Redemption Council (Liberia) Economic Rehabilitation Programme (Mozambique) Economic and Social Rehabilitation Programme (Mozambique) poverty reduction strategy Resistência Nacional Moçambicana Revolutionary United Front (Sierra Leone) special drawing right Sierra Leone People’s Party Special Operations Division (Liberia) Sudan Policy Sudan People’s Liberation Army Sudan People’s Liberation Movement sub-Saharan Africa Special Security Service (Liberia) structural and systemic transformation programme Tokyo International Conference on African Development Tigray People’s Liberation Front United Nations United Nations Mission in Sierra Leone United Nations Development Programme National Movement for the Total Independence of Angola Uganda National Liberation Front United Nations Mission in Liberia United Nations Operation in Mozambique Uganda People’s Congress United States of America United States Agency for International Development Union of Soviet Socialist Republics

Introduction: Post-Conflict Economies in Africa Augustin Kwasi Fosu

1

Introduction

Conflict has occurred throughout history. In Africa, the partition of the continent and creation of modern states using artificial borders often divided peoples and forced ethnically diverse groups together, creating the conditions for additional conflict. Independence was often contemporaneous with ideologogical conflict, as capitalism, socialism and communism competed for supremacy. Post-independence Africa has been dominated by elite political instability (Fosu, 2003). This form of conflict is observed to have contributed significantly to the dismal economic growth and development observed on the continent (Fosu, 2001, 2002, 2003). Following independence, there remained ongoing rebel movements in several African countries. In many countries, however, the dictatorial leaders appear to have succeeded in suppressing the simmering disputes fuelled primarily by inter-ethnic rivalries. Indeed, in some instances, the forging of nation-states through the use of strong central government control probably contributed to neutralizing the pernicious elements of ethnicity (for example, in Côte d’Ivoire, Ghana, Kenya, Malawi, Tanzania and Zambia). Following global political liberalization, however, open conflict in the form of armed insurrection developed in several African countries (for example, Liberia, Sierra Leone, Burundi, Rwanda and, more recently, Côte d’Ivoire). Between 1990 and 2000 there were some nineteen major conflicts in Africa, mainly in the form of civil wars (Wallensteen and Sollenberg, 2001). In this volume the emphasis is on armed insurrections in the form of civil wars and on the economies following them. It is difficult to define a postconflict economy. Nevertheless, the term is maintained in order to stress the importance of attaining such a state, as a prelude to achieving the political stability required for economic growth and development. This volume has two themes, which complement each other. The first section consists of four chapters on thematic issues pertinent to post-conflict economies. What factors are likely to result in conflict resolution in order to achieve the post-conflict state? How can that state be maintained in terms of avoiding the resurgence of conflict? This is followed by country case studies from the sub-Saharan region: West Africa; the Horn, East and Central Africa; and Southern Africa. These cases were chosen to reflect countries where there xv

xvi

Introduction: Post-Conflict Economies in Africa

have been substantial conflicts since independence. In most instances, the countries are at the time of writing experiencing relative peace. In some cases, however, conflict is ongoing, though with some reasonable hope of resolution and with the expectation that the post-conflict state can soon be attained.

2

Thematic issues

Chapter 1: Post-Conflict Reconstruction in Africa: Some Analytical Issues Tony Addison and Mansoob Murshed model the phenomenon of likely credibility problems associated with peace agreements in post-conflict situations in the immediate aftermath of civil war, and the risks of conflict recurring. Where war provides economic gains to one side, peace is not incentive compatible and peace agreements degenerate as they become time inconsistent. The party that has something to gain from surprise warfare will first agree to peace, but then renege. The levels of conflict chosen by this group are an increasing function of greed and capturable rents, but decrease with the group’s perceived cost of war. Furthermore, in a multiple time framework, uncertainty about the group type that may renege on peace generates greater fighting. In particular, higher rates of time preference associated with the impatience to consume immediately engender greater conflict. In this context, action by external powers could lower the risk of resurgence. For example, sanctions, aid and direct intervention can eliminate conflict, as well as helping in devising ‘commitment technologies’ for peace. This analysis is relevant to the recent cases of Angola, Burundi, Liberia and Sierra Leone. Chapter 2: Economic and Political Consequences of Conflict and Implications for Post-Conflict Recovery in Africa Ibrahim Elbadawi and Njuguna Ndung’u explore empirically the channels through which risk is transmitted to conflict and post-conflict economies. Estimating a system of equations that take into account multi-causal structure in the variables, a number of results emerge. First, these ratify and strengthen two salient results in the aid-effectiveness literature, based on single-equation estimation, that a sound policy environment enhances the effectiveness of aid in augmenting growth and reducing poverty, and that the impact of aid is subject to diminishing returns (Collier and Dollar, 2002). Second, that the non-monotonic impact of variables such as aid effectiveness and social fractionalization on the risk of conflict (Collier and Dollar, 2002; Collier and Hoeffler, 1998), and debt on growth,1 are supportable. Third, that public investment is the main channel through which the deleterious effect of conflict is transmitted to the economy. Fourth, that the absence of political rights in socially fractionalized societies exposes such societies to higher risks of conflict. And, finally, that exports tend to increase the risk of

Fosu

xvii

conflict by augmenting state revenues, thus providing increased rent-seeking opportunities.2 Chapter 3: Economic Policy in Post-Conflict Societies Paul Collier details policies which reduce two sets of risks characterizing post-conflict economies: those leading to the outbreak of the conflict in the first place, and those emanating from the conflict itself. Policies to resolve the first set of risks include: • raising the growth rate; • setting up transparent systems of scrutiny and controlling revenues from primary commodities in the short run; • diversifying the economy away from high primary commodity dependence in the long run; and • attenuating ethnic dominance, such as by guaranteeing the rights of minority groups, which may require the support of international intervention. Policies to reduce the second set of risks include: • providing agreed formulae for sharing government revenues – for example, between regions, and for public-sector jobs – such as between ethnic groups; • minimizing the incentives of groups preferring conflict to revert to this, such as converting rebel groups into political parties;3 and • providing a recognized positive role for the diaspora – for example, in nation rebuilding. Collier considers three specific areas likely to be differentially important in post-conflict settings: demobilization; the restoration of the rule of law; and the rehabilitation of the formal economy. Policies to address these include: • ensuring that demobilized soldiers are engaged meaningfully in the economy (such as by providing them with adequate access to land); • re-establishing property rights, followed by the enhancement of legal and judicial systems and of policing; and • the gradual formalization of the economy, including reducing the tax burden on those in the formal sector. According to Collier, external aid and good policy have special roles to play in post-conflict economies. While aid apparently has no direct impact in terms of reducing the risk of conflict recurrence, it can serve an important indirect role through its ability to increase growth, which can in turn decrease the risk as well as provide reassurance to minority groups that their rights are being protected. Further, aid is increasingly effective when the

xviii


policy environment is sound (Burnside and Dollar, 2000). Collier also underscores the importance of social policies within the policy environment, such as the inclusiveness of education and health. Similarly, the marginal product of sound policies also increases with the level of aid. Furthermore, the policy environment of embryonic post-conflict economies starts out as fragile but improves with time, so the optimal policy might then entail increasing aid levels over time for post-conflict economies. Additionally, aid could help to reduce the tax burden on those remaining in the shrunken formal sector in post-conflict economies, so that gradual formalization of the economy can proceed more effectively. Chapter 4: Ethnicity, Institutions of Governance and Conflict Avoidance Mwangi Kimenyi highlights the importance of establishing institutions that harmonize ethnic interests. Unfortunately, recent democratization reforms that have introduced political competition are not sufficient to promote conflict-avoidance. On the contrary, political party competition can intensify conflict-generating ethnic competition in highly centralized unitary states. Kimenyi proposes the establishment of decentralized institutions of governance comprising ethnically homogenous local units of collective choice possessing a high degree of autonomy. He suggests that governments based on ethnicity can be complemented by functional institutions transcending the boundaries of ethnicity within a given country.

3

Country case studies

3.1 West Africa Chapter 5: Liberia and Sierra Leone: Interwoven Civil Wars Victor Davies observes that Sierra Leone’s civil war between 1991 and 2001, and Liberia’s civil war from 1989 to 2003 are highly interwoven. He analyses common and idiosyncratic factors causing and sustaining the civil wars in the two bordering West African countries: natural resource abundance, historical legacies, growth collapse, ethnic rivalries, geography and external instigation. After discussing the social and economic consequences, the author examines Sierra Leone’s seemingly successful peace process, highlighting the domestic and external factors that aided it. He also assesses Sierra Leone’s post-conflict initiatives and challenges, especially natural resource management. He then asks why Liberia was at war for so long. Davies attributes much of the blame for the looting and quasi-criminal behaviour to the head of state, concluding that there is no simple solution to the Liberia problem, and that any hope for transition to an enduring, post-conflict peace requires continuous international attention.

Fosu xix

Chapter 6: The Nigerian Civil War: Causes and the Aftermath Based on a concept of conflict as a struggle for supremacy and strategic advantage by antithetical forces, Abdul-Ganiyu Garba and Kassey Garba address three sets of issues: the primary or root causes of the Nigerian civil war; the duration and cessation of the war; and the aftermath and risks of recurrence. Analysing the emergence and evolution of antithetical forces before and after the formation of Nigeria, the authors conclude that the seeds of civil war were sown in Nigeria’s amalgamation, the way it was governed when it was colonized, and the way its independence was arranged. Their analysis reveals that the civil war was extended by the interplay between conflicting foreign interests and actions, as well as differential domestic interests, capacities and actions, while the landlocked nature of Biafra reduced the duration of the conflict. The risk of conflict recurring has been increased by the exploitation of asymmetric institutional, political, economic and social relations by postconflict governments, as well as by the consequent adverse effects on governance, macroeconomic management and economic progress. While the 2003 elections presented both dangers and opportunities, the landmark judgements by the Nigerian Supreme Court in 2001–3 offer a basis for building a modern state on the colonial foundations if Nigeria overcomes the current challenges. Chapter 7: The Economics of Civil Conflict in Africa: The Case of Chad Nadjiounoum Djimtoingar and Djona Avocksouma argue that the Chad conflict is inherently internal, since it involved Chadian belligerents and took place within the country. Although the civil war broke out in 1979 and lasted until 1982, the genesis of the conflict can be traced back to the 1960s. The establishment in 1966 of the first politico-military movement, the National Liberation Front (FROLINAT), marked the first armed opposition to the government since independence in 1960. However, the change to a FROLINAT government worsened the hostilities. Lacking sound public management strategies, succeeding governments – from President Goukouni’s National Union Government to Presidents Habré and Déby – became embroiled in side issues far from the concerns of the northern and southern communities. Chad has not enjoyed widespread peace since President Déby seized power in 1990: pockets of resistance still operate in both the northern and southern regions, casting doubts on the credibility of the north–south, Muslim–Christian and animist, and pastoralist–farmer cleavages as the causes of the conflict. The costs of the ongoing conflict have been enormous. Mineclearing operations cost an estimated US$88 million, while demobilization and integration of combatants cost even more. The most important factor remains a lack of self-confidence among Chadians. Governments have been

xx


unable to mobilize the nation because of poor governance, institutionalized insecurity and ethnicization of public management. Several attempts have been made to resolve the conflict with the help of third parties, but these efforts have been only minimally successful. Indeed, successive governments have not accepted democratic succession, and negotiations have only succeeded in absorbing opponents into the government. Unfortunately, promises made are binding only to the signatories and conditions for transition to a peaceful, post-conflict phase are far from satisfied. Despite possible ethnic or religious undertones, bad governance is the key cause of the conflict. 3.2 The Horn, East and Central Africa Chapter 8: Conflict, Post-conflict and Economic Performance in Ethiopia Ethiopia has had a long history of relative cultural homogeneity, notwithstanding the large number of ethno-linguistic groups. Nevertheless, that long history has been punctuated by a number of conflicts, including the last religious war fought between the Christians and Muslims in the sixteenth century, and the era of the princes in the nineteenth century. According to Alemayehu Geda and Befekadu Degefe, the more recent periods of conflict have been the imperial era, which lasted until the fall of Haile Selassie in 1974; the 1974–91 Dergue period, during which a civil war took place; and the 1998–2000 Ethiopia–Eritrea war. While the conflict of the imperial period involved rivalries among the aristocracy over the benefits of control of the peasantry, the 1974–91 conflict mainly entailed rivalries among the educated elites. Similarly, the 1998 Ethiopian–Eritrean war most probably resulted from a miscalculation as each side threatened the other with military power in search of a more favourable bargaining position following the agreement on Eritrea’s secession from Ethiopia in 1993. In Ethiopia, the post-conflict periods have so far been relatively brief as there has been a pattern that the means of resolving one conflict sows the seeds of the next. This phenomenon deserves special attention, if an enduring post-conflict period is to be fostered. International intervention is also recognized as having played a salient role, either as a catalyst, or a restraint, of conflict. With the end of the Cold War the geopolitical importance of Ethiopia has diminished, thus attenuating the catalytic role of international interventionism. In the final analysis, however, Alemayehu Geda and Befekadu Degefe believe that the major cause of conflict in Ethiopia lies in the political economy and is underscored by competition for power. If so, then the apparent solution for achieving a more stable post-conflict economy is a strategy to minimize such competition. One such potential mechanism might involve the establishment of a democratic system that makes leaders more accountable to the people.

Fosu xxi

Chapter 9: Prospects for Sustainable Peace and Post-Conflict Economic Growth in the Sudan The Sudanese conflict, at the time of writing, spanning more than thirtyeight years of civil war since independence in 1956, is characterized by religious and cultural divides between north and south, according to Ali Ali and Ibrahim Elbadawi. These fundamental divides dominate the country’s tribal and regional interaction between north and south and within the regions. While the north–south divide explains external interventions, culture and religion account for the relatively high risk of conflict predicted for Sudan. The conclusion, according to the authors, is that these two realities about the Sudanese conflict are the key to understanding why the war started and why it has persisted for so long. Furthermore, these realities define several preconditions for sustainable peace, including the centrality of a genuine democratic transition in both the north and the south. The failure of the 1972 Addis Ababa peace accord, which allowed eleven years of fragile peace, shows the importance of democracy for sustainable peace and viable economic management in a divided Sudanese society. The accord finally broke down as a result of the autocratic rule of President Nimeiri, who, inter alia, attempted to impose Islamic law on the south. The authors argue that party-based democracy is the most viable instrument for promoting stable and developmental bargaining processes among potentially conflicting social groups. Unfortunately, the peace initiative under the auspices of the Intergovernmental Authority for Development (IGAD), being confined to the two main military protagonists, does not seem to accommodate this basic but crucial lesson. With the emergence of Sudan as an oil-producing country, transparent and democratic governance becomes even more critical. International evidence suggests that dictatorships (as well as unstable factional democracies) have almost always failed to harness oil resources for development. Instead, under such regimes, oil resources have been associated with deepened poverty, social cleavages and conflict. Even if peace cannot be achieved without partitioning the country, the transition to democracy in both future states would be critical. Given current conditions in the Sudan, partition without democratic transition risks replacing a civil war with an international war between two autocratic and non-representative regimes, both of which may also risk internal conflict. Chapter 10: Entrenching Peace in Post-conflict Economies: The Case of Uganda Michael Atingi-Ego and Rachel Sebudde show that, since independence in 1962, there has been at least one conflict in Uganda during each decade. Between 1964–6 the crisis was an ethnic power struggle which resulted in suspension of the Constitution, as well as detentions and killings of Bugandan civilians. In 1971, Idi Amin replaced President Milton Obote via a coup d’etat,

xxii


and then the liberation struggle between 1978 and 1980 culminated in the overthrow of President Amin in 1979. Conflict between 1981 and 1986 brought President Museveni to power, since which there have been continuing insurgencies, mainly in the north and west of the country, although Uganda has enjoyed relative peace in the post-conflict period starting from 1986. The authors argue that grievance, exacerbated by weak institutions, has been the main source of conflict. Various factors have led to improvements of the post-conflict economy since 1986: political reconciliation, including the return of some of the former leaders; the restoration of kingdoms; and improved economic conditions, such as per capita income growth and poverty reduction. Relations with Uganda’s neighbours have improved. Institution building, with international support, has been important, and has included: • • • • •

grassroots democratization; enhancement of press freedom; decentralization of power to local government; public service, market and legal reforms; a legal framework strengthening commercial jurisdiction, property rights and the settlement of business disputes; • support for the development of effective civil society groups; • the establishment of a properly functioning financial system as well as safety nets; and • demobilization of the military and the reintegration of rebels. The problem of uneven distribution of development remains, however. While the southern part of the country has benefited from the above reforms, the north has continued to lag behind. Meanwhile, the insurgency in the north continues. It is hoped that solutions will be found to achieve both economic progress and abate the insurgency in the north. Chapter 11: The Challenge of Transition from War to Peace in Burundi Since independence in 1962, Burundi has had five civil wars: the first was in 1965 and the most recent began in 1993. This latest conflict has been the longest and costliest, and at the time of writing, the situation is barely postconflict. The cycle of violence has been played out between two ethnic groups: the Hutus, and the minority (but politically and economically dominant) Tutsis. According to Janvier Nkurunziza and Floribert Ngaruko, powerful, self-interested domestic groups have manipulated the apparent ethnic rivalry to their advantage. The authors observe that Hutus and Tutsis lived together peacefully for at least four centuries before the outbreak of hostilities in the early 1960s, and shared a normative system of governance and a political system that allowed divergent interests to be reconciled, while social and economic activities allowed Hutus and Tutsis to identify themselves as members of one nation.

Fosu

xxiii

Thus, post-conflict Burundi can learn from its distant past. Justice is a prerequisite if the wounds of the recent past are to be healed. This requires an impartial administration to punish those inciting violence for their own selfish purposes. One way to proceed, the authors argue, is to use a prisoner’s dilemma-based system of amnesty and justice that induces perpetrators to tell the truth. Meanwhile, the role of the international community is crucial, not only in terms of engendering confidence for any agreement among the warring parties, but also in keeping neighbouring countries at bay. 3.3 Southern Africa Chapter 12: The Political Economy of Post-Conflict Economic Recovery: Making Peace Substantial in Mozambique Examining the political economy of post-conflict Mozambique, Laudemiro Francisco and Edwin Connerley argue that the durability of the peace and the domestic viability of current reforms are dependent on making peace substantial. This would require a comprehensive approach to reconstruction and development, including self-rule, proper macroeconomic policies, and building real foundations for social and political integration. Following the 1992 General Peace Agreement (GPA), Mozambique has experienced an impressive economic recovery. However, the GPA failed to address the grievances that motivated the combatants, and the risk of renewed conflict has therefore remained high. Thus an approach that would realistically foster economic, social and political development is still required. How the main actors continue to perceive the ability of the GPA to address their concerns is politically crucial for the viability of reconstruction and development efforts. Chapter 13: Transformation for Post-Conflict Angola Fátima Moura Roque writes that Angola was impoverished by twenty-six years of civil war. Political interference in economic management resulted in perverse policies, financial sector weakness, opaque public accounts, and corruption. The death of UNITA leader, Jonas Savimbi, in 2002 was followed by a Memorandum of Understanding which consolidated the ceasefire and the terms of peace. Though too early at the time of writing to be completely confident, the peace seems to be holding. Post-conflict, Angola must now confront widespread malnutrition; provide accelerated basic training for 80,000 demobilized soldiers and create safety nets for about 300,000 dependents; reintegrate 4.2 million internally displaced people and 500,000 refugees; and address extreme poverty. Following such a protracted conflict, a strong commitment to national reconciliation is required – that is, if the parties are to overcome so many years of resentment and social exclusion. Meanwhile, democratic governance and good economic fundamentals are essential for the post-conflict transformation.

xxiv Introduction: Post-Conflict Economies in Africa

4

Conclusion

Chapter 14 concludes the volume by synthesizing and drawing lessons from the earlier chapters. The synthesis comprises the following ten themes: colonial history; the role of neighbours; the relevance of elite political instability (EPI); the importance of rent-seeking opportunities; the role of child combatants; the proliferation of cheap weapons; justice: inter-ethnic and class inequities; justice and impartial, effective legal systems; the role of the international community; and the importance of democracy.

Notes 1 See, for example, Fosu (1996); Elbadawi, Ndulu and Ndung’u (1997). While the latter finds a non-monotonic impact of debt on growth based on a quadratic specification of the debt variable, the former employs an analysis-of-covariance model and finds that the debt effect is positive when the level of investment is low, but becomes negative at a sufficiently high level of investment, suggesting that the marginal product of capital decreases with the level of debt. 2 This result supports the finding in the literature that the prevalence of lootable commodities, especially primary exports, tends to raise the risk of civil war (Collier and Hoeffler, 1998). It also provides support for the result based on a more general concept of conflict in terms of ‘elite political instability’ (EPI). Fosu (2003), for example, finds that a higher exchange rate misalignment tends to raise the likelihood of EPI in African countries. 3 Not only might such a strategy convert, to use Mancur Olson’s words, ‘roving bandits’ into ‘stationary bandits’, but possibly to responsible non-bandits.

References Addison, T. (ed.) (2003) From Conflict to Recovery in Africa, WIDER Studies in Development Economics (Oxford: Oxford University Press). Burnside, C. and D. Dollar (2000) ‘Aid, Policies and Growth’, American Economic Review, vol. 90, no. 4, pp. 847–68. Collier, P. and D. Dollar (2002) ‘Aid Allocation and Poverty Reduction’, European Economic Review, vol. 46, issue 8, pp. 1475–500. Collier, P. and A. Hoeffler (1998) ‘On Economic Causes of Civil War’, Oxford Economic Papers, vol. 50, no. 4, pp. 563–73. Collier, P. and A. Hoeffler (2001) ‘Greed and Grievance in Civil War’, working paper no. 2355, World Bank, Washington, DC. Elbadawi, I., B. Ndulu and N. Ndung’u (1997) ‘Debt Overhang and Economic Growth in Sub-Saharan Africa’, in Z. Iqbal and R. Kanbur (eds), External Finance for Low-Income Countries (Washington, DC: International Monetary Fund). Fosu, A. K. (1996) ‘The Impact of External Debt on Economic Growth in Sub-Saharan Africa’, Journal of Economic Development, vol. 21, no. 1, pp. 93–118. Fosu, A. K. (2001) ‘Political Instability and Economic Growth in Developing Economies: Some Specification Empirics’, Economics Letters, vol. 70, no. 2, pp. 289–94.

Fosu xxv Fosu, A. K. (2002) ‘Transforming Growth to Human Development in Sub-Saharan Africa: The Role of Elite Political Instability’, Oxford Development Studies, vol. 30, no. 1, pp. 9–19. Fosu, A. K. (2003) ‘Political Instability and Export Performance in Sub-Saharan Africa’, Journal of Development Studies, vol. 39, no. 4, pp. 68–82. Wallensteen, P. and M. Sollenberg (2001) ‘Armed Conflict 1989–2000’, Journal of Peace Research, vol. 38, no. 5, pp. 629–44.


Part I Thematic Issues


1 Post-Conflict Reconstruction in Africa: Some Analytical Issues Tony Addison and S. Mansoob Murshed

1

Introduction

In contrast to much of the twentieth century, when warfare between rich states was the norm, contemporary conflict now occurs almost exclusively in poor developing countries and is mainly internal in nature (although external parties and neighbouring countries may support internal belligerents). The period 1990–2000 saw nineteen major armed conflicts in Africa, ranging from civil wars to the 1998–2000 war between Eritrea and Ethiopia (Wallensteen and Sollenberg, 2001). Although the picture is grim, there has recently been some good news: a peace agreement has been signed in Angola; the governments of the Democratic Republic of the Congo (DRC) and Rwanda have signed a peace agreement; talks have resumed between the government and the rebels in Sudan. Sadly, these positive developments need to be counterbalanced with the empirical regularity with which peace agreements break down (Walter, 2001). Ensuring that the benefits of reconstruction are broad-based rather than narrow in their benefits, it is important to minimize the chances of conflict reigniting, since grievances will otherwise fester. Accordingly, the containment and reduction of inequality, and not just a reduction in absolute poverty, may be central to broad-based reconstruction. This means that transition from war to peace may not be seen as only a political process, or one in which attention being paid to the economic dimensions of conflict should take a back seat to the political settlement (as is sometimes argued). Concomitantly, economists may not ignore the implications for social conflict of economic decisions: conflict and the potential for conflict play a strong role in retarding growth, development and poverty reduction (Murshed, 2002). Indeed, civil wars and conflict are among the major causes of development and growth failure in today’s developing world, a point increasingly emphasized by aid donors such as the UK’s Department for International Development, the World Bank, and the United Nations (DfID, 2001; World Bank, 2001; United Nations, 2000). 3

4

Thematic Issues

As societies slide into war, competing groups are formed to overcome collective action problems (in the sense of Olson, 1965). Ethnicity, whether based on language, religion or other distinctions, is often a superior basis for collective action in contemporary conflicts in poorer countries than other social divisions such as class that featured in the internal conflicts of earlytwentieth-century Europe.1 Moreover, as Azam (2001) indicates, organization or action along ethnic lines is often a product of the failure of the state to provide public goods. Thus conflicts rooted in grievances over one group’s standard of living relative to another, can take an increasingly ethnic dimension as leaders mobilize followers by appealing to ethnic differences (for example, the appeal to religious affiliation in Nigeria’s local-level conflicts). So too with greed over valuable natural resources; in the 1990s, Angola’s rebel movement – the National Movement for the Total Independence of Angola (UNITA) – increasingly emphasized ethnicity to organize support among the Ovimbundu people so as to win control over the country’s diamonds and oil. The greed versus grievance dichotomy is important to the debate on the nature of conflict (see Murshed, 2002). But for either of these forces to cause large-scale violence there must be other factors at work, specifically a weakening of dispute resolution mechanisms, which some refer to as state failure (Marshall, 1999, for example). We may view today’s African conflicts as being rooted in a partial or complete breakdown of the social contract – the agreed rules that govern the distribution of resources and obligations across society – and the concomitant mechanisms for settling differences. Furthermore, the social contract has economic and redistributive dimensions that are, in turn, embedded in fiscal institutions. The social contract, its collapse, and the centrality of its reconstruction post-conflict therefore provide the organizing themes of this chapter. Section 2 is concerned with conflict and the breakdown of the social contract, including possible reasons why a peaceful social compact or agreement might become infeasible in the pre-conflict phase. Section 3 analyses how commitments to peace, and the broad-based fiscal transfers it entails, can be sustained and enhanced. Section 4 continues on the theme of designing commitment technologies to peace, this time in the context of capturable natural resource rents. Section 5 discusses post-conflict reconstruction, and how its design can rebuild a peaceful and broad-based social contract. Finally, Section 6 concludes with a discussion of areas of major concern in post-conflict societies.

2

The collapse of the pre-conflict social contract

Within nation-states, the fiscal system will secure a workable social contract if the allocation of public expenditure and the apportionment of taxes are judged to be fair, or at least not so unfair that some groups regard taking resources by force the better option. There are many examples of conflicts

Addison and Murshed: Reconstruction

5

emerging out of fiscal disputes. Côte d’Ivoire, for example, has become increasingly unstable with the collapse of the social contract engineered by the late President Houphouët-Boigny, in which he allocated public spending across the regions to buy the loyalty of the country’s ethnic groups. Similarly, conflict in Burundi and Rwanda has had strong fiscal dimensions (Ndikumana, 2001). Disputes over the apportionment of revenues from natural resources are especially common and, as in Nigeria, they often take on ethnic and regional dimensions. Azam and Mesnard (2001) analyse the difficulties in sustaining social contracts that involve revenue sharing via fiscal transfers.2 Contemporary civil wars are more often related to the breakdown of explicit or implicit arrangements to share resources or revenues, rather than their absence. One reason that a contract to share revenues encounters difficulties is the imperfect credibility with which the side that controls the resources honours its commitment. Let there be two parties to the potential armed conflict, the government and a rebel group. The government party has access to revenues and royalties, but is threatened by the excluded rebel group, which may violently overthrow the government with probability π, related to a contest success function. On the other hand, it may choose not to fight if it receives a fiscal transfer from the government. Similarly, the government has a choice between fighting the rebels and offering them a fiscal transfer. The utility function of the government (UG) may be denoted as:

U

G

R

μF G G -⎞ ( Y – F – λG ) = ( 1 – π ) ⎛ --------------------⎝ F G + μF R⎠

(1)

where YG is the income of the government, FG represents government military expenditure, G the transfer to the rebels, λ < 1 captures the imperfect credibility with which the government transfer is made, and the probability of the government staying in power is 1 − π, which in turn is a function of the contest success function, as well as the relative fighting efficiency of the rebels vis-à-vis the government μ. The contest success function (see Hirshleifer (1995) for example) is given by: R

μF π = --------------------G R F + μF It is defined in terms of the rebels, and says that the probability of their success in the military contest with the state depends on their military expenditure, FR relative to the total military expenditure made by both sides to the conflict FG + μFR, weighted by the rebel side’s relative military efficiency, μ. The rebel utility function (UR) is given by: U R = Y R − F R + λG + π(·)(Y G − F G − λG)

(2)

6

Thematic Issues

The income of the rebel group is described by YR, and the cost of waging a war represented by FR. A great deal can be said about the sources of income and war finance for the rebel side. The rebels may obtain an income from natural resources under their control (illegal diamond mining in Sierra Leone, for example), by raising external finance via promising commercial concessions once they seize the state (the strategy of Laurent Kabila when he overthrew Mobutu Sese Seko in DRC in 1997), criminal activities such as drug trafficking, or sympathetic contributions from diasporas (Eritrea). A feasible social contract favouring peace must give the rebels as much utility via a credible transfer as they would get in the event of a (probability weighted) overthrow of the state. A social contract favouring peace in return for a transfer is infeasible if, at a critical level of π = ϕ, the probability of toppling the government by war is greater than the chances of it credibly making the transfer, λ.3 Also the social contract is less likely with regimes that prefer military expenditure F G over making a fiscal transfer to the rebels, a common outcome in countries with powerful militaries, such as Sudan. Note that, for the government, no transfer is made in the case of a civil war, and in the case of a credible transfer being possible there is a saving in military expenditure. There is therefore a trade-off between military expenditure and a credible transfer. For the rebels, there is the transfer and no fighting expenditure if there is peace. When the transfer is highly improbable, and the potential spoils are great, warfare is more likely. This captures the pre-conflict situation. We now turn to problems of enhancing commitment to peaceful behaviour, post-conflict.

3 Sustaining the post-conflict commitment to peace by the government We now consider policies that may enhance the credibility of the government’s commitment to the broad-based transfer to the rest of the society, which will increase the chances of an enduring peace. Let us postulate that peace has been established, usually through external intervention, and for the moment there is no war, such that π = 0 in Equation (1), and the government is in no imminent danger of losing power to the rebels. The price of peace is the transfer to the rest of society, including the erstwhile rebels. Will the government stick to this commitment? There are two types of government: the first type (labelled as 1) will honour its commitment, and the second type (labelled 2) may not honour its commitment. Let us consider the utility function of the first or good type of government: UG1 = f(Y G − F G − λG)

(3)


7

Its disutility from fighting (differentiating Equation 3 with respect to F G) leads to: −f1 = 0

(4)

This type of government obtains no utility from fighting the rebels and makes the transfer, hence λ = 1 in Equation (3). Turning to the second type of government, which is perhaps more realistic in relation to post-conflict situations: UG1 = f(Y G − F G − λG) + h(F G)

(5)

This type of government derives positive utility from reneging on its commitment to the transfer, captured by the parameter h, implying that it is willing to fight in order to make a smaller transfer or no transfer at all. Maximization with respect to fighting the rebels yields: f1 = h1

(6)

Thus this type of government will equate the marginal cost and benefit of reneging on pre-commitments to peace via the transfer, so future peace is not guaranteed and the problem of imperfect credibility of the commitment to the broad-based transfer remains λ < 1. The presence of imperfect credibility requires policy-makers, particularly external powers and donors (which can be crucial to ending civil war) to devise commitment technologies to peace. Consider an example, leading to an innovative term in the utility function of the ‘bad’ (type 2) government: UG1 = f(Y G − FG − λG) + h(F G − m(F G))

(7)

Here, m refers to the cost of fighting or reneging on strategic pre-commitments imposed by donors. This takes the form of a variety of sanctions, military as well as through the suspension of aid. Maximization with respect to fighting will now yield: f1 = h1(1 − m1)Λm1 ≥ 1

(8)

When compared to Equation (6), Equation (8) states that the incentive to deviate from peaceful commitments is diminished, as the marginal utility of fighting is reduced. Furthermore, as m1 → 1, the temptation to deviate from agreements is zero, as with the first type of government in Equation (4).

8

Thematic Issues

4 Natural resource rent contests and imperfect commitment among rebels Many conflicts in Africa occur amid natural resource wealth. However, it is not natural resource endowments per se but rather the type of natural resource that matters. Addison, Le Billon and Murshed (2002) distinguish between point resources, which are non-renewable and geographically concentrated (for example, diamonds and oil), and diffuse resources, which are renewable and geographically spread, such as soil and water. Conflict occurs in countries with diffuse-resource endowments when access and ownership are highly unequal, but its probability falls when asset redistribution is undertaken and is certainly less evident than in countries richly endowed in point resources such as Angola, Congo-Brazzaville, DRC and Sierra Leone. Exactly how point-resource endowments increase the chances of conflict is now debated vigorously. One factor might be the difficulty of managing the macroeconomic consequences of resource windfalls, resulting in boom and bust cycles that ultimately lower the rate of economic growth, where low or declining growth in employment and living standards is a factor in generating conflict (as found by Nafziger, Stewart and Väyrynen, 2000). Another is the incentive of political insiders to undermine state institutions (such as budgetary systems) deliberately, in order to conceal the siphoning-off of resource revenues. This leads to the collapse of the state from within, together with its ability to exercise social control and defend itself against usurpers (as in Mobutu’s Zaire). Support for the government also declines as its commitments take the form of highly imperfect contracts. Additionally, opposition groups attract the finance of unregulated international trading networks interested in capturing valuable natural resources (giving rise to merchant-capital wars, such as those in West Africa, which receive arms and finance from entrepreneurs based in the Middle East and the former Soviet Union). In this section we examine credibility problems related to the temptation to renege on peace agreements when capturable (point) natural resource rents are at stake. Again, we have two sides, whom we refer to as government and rebels.4 One side is either tired of fighting or it has an interest in peace. In our example this group is the government. The other side, the rebels, may have something to gain from the resumption of fighting, such as rents from resources (as in Angola). The roles played by the government and rebels in the games that follow can be reversed without altering the results. Both sides have entered into a peace deal. The government side derives no benefit from breaking this agreement. Consider the utility function of the rebel group (UR): UR = −(1/2)c1w2 + Bc2(w − we)

(9)


9

The first term on the right-hand side of Equation (9) is the pure cost of conflict in quadratic form, where w represents warfare or belligerent behaviour and c1 is the parameter measuring the direct cost of warfare. The negative sign before it is to indicate the cost or disutility from fighting. The quadratic form of the cost function indicates that the costs of war rise more than proportionately as the level of w rises. The second term on the right-hand side of Equation (9) indicates the gains to the rebels from reneging on a peace agreement, or the benefit from a ‘surprise’ war, where the level of actual conflict (w) exceeds the level of conflict expected in advance (we). In other words, the spoils of war can only be wrested via the ruse of peaceful intentions. The parameter c2 captures the magnitude of this effect: the higher is c2, the greater is the gain from feigning peacemaking first and looting later. It may also be viewed as a subjective measure of greed. In addition to this, the greater the abundance of lootable resources (B), or rents to be extracted, the higher is the gain from a surprise war. The rebels maximize their utility in Equation (9) subject to w, which leads to: w = Bc2/c1

(10)

This result can be interpreted in the following manner: the equilibrium choice of warfare is greater the higher is the element of pure avarice, c2, the higher the availability of lootable resources, B, and the smaller the direct cost of fighting, c1. As far as the government is concerned, its utility function could take the form: UG = −(w − we)2∀w ≥ we and = (w − we)2∀w < we

(11)

We shall focus our attention on cases where the government’s utility declines in surprise warfare when actual war is greater than expected. The opposite, when actual levels of belligerency fall below expectations (pleasant surprises) are intellectual curiosa. The disutility from surprise war arises either because the government has to engage in unforeseen military expenditure that diverts income from other types of public expenditure, or because war increases its need for foreign loans or aid. For example, following UNITA’s return to war in 1998, Angola’s government had to raise military expenditure sharply, much of it financed by oil revenues and foreign borrowing. Now assume that the rebels enjoy a first-mover advantage and can announce total peace, and then engage in surprise warfare. In this case, the actual and

10 Thematic Issues

expected levels of warfare would diverge — w = Bc2/c1 and we = 0 in Equation (9). This involves cheating on a pre-announced commitment. At this juncture, reputation becomes important. The government forms a view of the rebel announcements based on the past behaviour of the rebels. This implies that, for the rebel group, there exists a future cost of cheating in the context of a low-intensity conflict. The cost is equal to the loss of reputation for honesty, but this cost comes in the future. In summary, the socially optimal policy of zero warfare (w = 0) is time inconsistent or incentive incompatible, and thus will not be a possible outcome. The optimal policy of no conflict is infeasible, as it is inconsistent with the incentives and expectations of the parties to the game. More particularly, the government knows it is in the interests of the rebels to renege on a preannounced policy of total peace, and thus will not find any peace-offering credible. Furthermore, there will be a range of possible conflict intensities that are feasible equilibrium outcomes; thus multiple equilibria are possible. A rise in war booty raises the minimum conflict threshold. Also situations where the future is heavily discounted are likely to raise the fighting threshold associated with the best enforceable outcome. We now consider policies to reduce conflict. Much of this implies manipulating the attitudes of the rebel leadership via sanctions, arms controls, trade restrictions and foreign aid. Consider a reformulated version of the rebel utility function where we embed external conflict prevention policy parameters, and an additional cost component associated with an implicit or explicit international agreement: UR = −(1/2)c1(M1)w2 + Bc2(T,M2)(w − we) − c3(M3)(w − we)

(12)

Aid (T) may be utilized by a foreign power to reduce greedy attitudes, c2. Similarly, trade sanctions (M2) on items such as ‘conflict’ diamonds, money laundering and the activities of foreign entrepreneurs (in supplying arms and finance) might have the same effect. International controls on arms transfers (M1) and/or sympathetic assistance from non-residents could be utilized to raise the direct cost of war, c1. This lowers the fighting efficiency of the rebels, μ in Equations (1) and (2). The last term in Equation (12) represents another commitment technology by delegation, and c3 measures the costs of reneging on peace agreements as a function of sanctions (M3) imposed by other external signatories or parties to the agreement. The presence of commitment technologies, sanctions, conditional aid and controls on international trade (from looted resources) lowers the optimal level of belligerency among rebels. This can be seen by maximizing Equation (12) with respect to w, which yields:

Addison and Murshed: Reconstruction 11

w = (Bc2 − c3)/c1

(13)

This leads to a lower level of warfare when compared to Equation (10). There is a direct effect of the external sanction or outside commitment technology, c3. Then we have the indirect effects emanating from the manipulation of the behavioural parameters of the rebels, c2 and c1. Our discussion indicates that establishing peace is a tough job, as the recent experiences of Angola, the Democratic Republic of the Congo (DRC), and Guinea-Bissau show. But even if a peace agreement can be put in place, the difficulties do not end there: the form that reconstruction takes also affects the sustainability of peace. It is to these issues that we now turn.

5

Broad versus narrow reconstruction

The 1990s saw substantial donor-supported reconstruction: Angola, Ethiopia, Mozambique and Sierra Leone are African examples. To be successful, reconstruction must rebuild the social contract, otherwise conflict can reignite. Fundamental political and economic reform is therefore necessary to change the rules of the game, such innovation being akin to the deep interventions described in Dixit (2001). In principle, the temptation to renew conflict could be eliminated completely if the ‘prize’ (oil revenue, for example) is absorbed into national income (then w = 0, in Equations (10) and (13)). The gains from capture and surprise warfare will vanish in rebel utility functions, as B = 0. Input by external actors, such as linking peaceful behaviour to aid flows, will then be unnecessary (c3 = 0 in Equation 12). But promises of action are not enough. As we saw in Section 2, the social contract became unviable because the credibility of a transfer (λ) is low relative to the probability of rebel success in armed conflict. Also, as in Section 4, the commitment to peace may be subject to temptations to renege on the agreement. The credibility of the fiscal transfer may be restored through the creation of democratic institutions. But such institutional reform takes considerable time. A badly functioning democracy also implies incomplete contracts and imperfectly credible institutions of commitment. For example, the quality of fiscal institutions may be so low (as a result of destruction or neglect) that they are unable to deliver the necessary transfer even when the government is committed to making the transfer to its former opponents. The size of the transfer (G) in Equation (2) may be too small in per capita terms if the income of the excluded rebel group is low. Our discussion so far has been quite pessimistic, in part because peace is viewed as being dependent on the distribution of a ‘fixed pie’. But peace is more likely to endure if the ‘pie’ can be made to grow. Growth, by raising the tax base, increases the possibilities of redressing grievances through

12 Thematic Issues

fiscal transfers. Growth itself provides additional income to further reduce grievance (and perhaps mitigate greed) and tightens the labour market, thereby reducing the attractiveness of joining a warlord or rebel group (a point emphasized in Collier and Hoeffler, 2001). The initial years of peace may offer considerable scope for growth as the economy re-achieves its pre-war production frontier (Mozambique, whose war ended in 1992, has achieved annual real GDP growth of 6.2 per cent over the period 1990–9, with growth of over 11 per cent in the late 1990s). Just as economic growth in general can distribute its benefits broadly or narrowly across society (depending on the initial distribution of assets and skills), so too does reconstruction-led growth. Pre-war asset and skill distributions may have been highly unequal (with the resulting grievances contributing to conflict), and can worsen dramatically during wartime. Those who are already poor often lose the few assets they have, and looting adds to the number of poor (Nafziger, Stewart and Väyrynen, 2000). In contrast, warlords and their followers accumulate assets, and so while the early years of peace may see quite rapid gross domestic product (GDP) growth, it can be very narrow in its benefits, unless policies are put in place to restore the productive assets and human capital of the poor. The immediate post-conflict situation may also offer a golden opportunity for pro-poor asset redistribution (something that cannot be done easily during peace), although this can be impeded when rich ‘winners’ from war block the necessary measures. However, the ‘post-conflict’ economy will be highly distorted, and this can impose an unfortunate path-dependence on reconstruction and growth. One source of distortion is the sharp increase in transaction costs (σ) resulting from war (including the destruction of transport, the planting of land mines, and institutional collapse) that drive a wedge between producer and consumer prices. Typically, production (especially agriculture) is more vulnerable, leading to a sharper increase in its transactions costs compared to other sectors such as urban-based trade and services (as was the case in Mozambique’s sixteenyear civil war). To get a handle on this problem, consider a model with two sectors, agriculture and services, whose relative prices are PA and PS respectively, with transaction costs σA and σS. The transaction costs lower the supply price of the two goods because producers have to incur this extra cost. Given σA > σS, relative prices shift, resulting in a switch in activity from production (agriculture) to services: PS – σS PS ------------------- > -----PA – σA PA

(14)

if σA > σS This output change is additional to the increase in production costs resulting from the direct destruction of institutions, infrastructure and skills in the


two sectors (which usually affect production disproportionately, especially agriculture). Services and trade, including smuggling and other rent-seeking, dominate wartime economies and, because of the relative price effect, the collapse in production usually exceeds that resulting from destruction alone. Smallholder agriculture, the basic livelihood of Africa’s poor, is hit especially hard. For example, at the time of writing, Angola’s agricultural output is now less than 5 per cent of its pre-war level. The result is shown in Figure 1.1. Three production frontiers are indicated: PP (peace), WW (war) and RR (recovery). These have the usual properties. Assume a peace equilibrium of Ep given by the relative price (with σA , σS = 0).5 Growth will be along a ray such as Gp, as aggregate investment pushes the production frontier outwards over time. Contemporary wars generally reduce aggregate output. The wartime production frontier must therefore lie within the peace-time frontier. Agriculture (A) is hit harder than services (S), and the production frontier changes shape (to form WW reflecting the greater difficulties of A production). And the relative price shift in favour of services causes a structural shift to the new equilibrium Ew. There may be periods of growth during wartime, from a point below the peacetime production frontier, but it will be S-biased (along the ray Gw) and is therefore unlikely to reduce poverty sufficiently.

S Gw P w p Gp R Ep

W Ew

Eb

Gb

b

W Figure 1.1

War, peace and recovery

R

P

A

14 Thematic Issues

If a peace agreement is signed, then transaction costs start to fall. Accordingly, agricultural output recovers as relative prices shift back in its favour and as infrastructure is rebuilt; the economy moves through a series of reconstruction frontiers such as RR. Although this reconstruction is characterized by a more than proportionate rise in A, the main livelihood of the poor, benefits of recovery may still be distributed too narrowly because of the loss of assets and skills among the poor (see de Sousa, 2003 on Mozambique). Moreover, the economy’s pre-war structure may have been far from optimal for growth, poverty reduction or peace because of policy distortions — a history of policy bias that depresses PA relative to PS as in many African economies. Hence significant economic reform must take place alongside reconstruction if broad, rather than narrow, recovery is to take place. In Figure1.1 reform is shown as raising the return further on A relative to S activities (that is, to the relative price line b so that the economy recovers along Gb). There is wide agreement on the need for fiscal reform to raise public spending on pro-poor basic health care and primary education, although implementation is often fraught with difficulties. If peace is uneasy then there is an additional constraint on rebuilding the social contract via broad-based reconstruction and growth. To see why, consider the net present value (NPV) of a private investment project in sector i, producing an output Qi (Equation 15). In addition to raising transaction costs (σ) and production costs (C), conflict raises uncertainty about the future, and therefore affects the private discount rates of potential investors (r):6

∑ Qi ( Pi – σi ) – Ci t NPV i = --------------------------------------------t (1 + r)

(15)

The rise in r has differential sectoral effects depending on the time-profile of private benefits and costs associated with different types of activity. Thus production sectors such as agriculture require more fixed capital than trade and services (which may only need working capital), and profits from trade can be realized more quickly than from production, whose profits may only start to flow after several periods (particularly in agriculture, where tree crops have a lengthy gestation). Any rise in the private discount rate will tend to lower the NPV of investments in production relative to trade and services. This amplifies further the war-time collapse in the production sectors, and the expansion of trade, services and associated rent-seeking. Restarting activity in pro-poor production activities such as agriculture can therefore be especially hard after war when the peace is uneasy, since while transaction costs can be reduced (for example, de-mining rural roads lowers σ) and infrastructure rebuilt (lowering C), an uneasy peace keeps private discount rates (r) high. Consequently, the reconstruction-growth path may create insufficient income and employment, and its narrowness lead to frustration among large sections of the excluded population. Hence the


effectiveness of donor support to reconstruction, while lowering σ and C, will be reduced if parallel political initiatives (both domestic and international) are not undertaken to secure peace, and lower r. In summary, if grievance is the main source of conflict, then broad-based reconstruction (along Gb) is necessary to re-create a sustainable social contract. In particular, growth raises wage levels and reduces the attractiveness of predation as a livelihood (Collier and Hoeffler, 2001).

6

Conclusions

We have stressed the importance of the breakdown of the social contract, in particular its fiscal component, as contributing to civil war in Africa. We went on to examine the problems of sustaining a commitment to peace both by a government whose commitment may be tenuous, and by a group that has something to gain from the resumption of hostilities in terms of controlling resource rents. Finally, we considered the difficulties in engendering broad-based and pro-poor growth and reconstruction in post-conflict societies. This chapter has emphasized the importance of economics and rational-choice political science. This is not to underestimate the complexity of contemporary war, nor its deep roots in cultures and history (especially colonial legacies), nor to neglect the importance of other social science methodologies in addressing it. Nevertheless, there is a danger in seeing every conflict situation as being in some way culturally or historically unique, and thus perhaps intractable. The approaches discussed in this chapter do provide a framework in which to identify the fundamental constraints that must be overcome to end war, as well as the trade-offs involved. To conclude, we highlight two areas of pressing concern that need attention. The first is the issue of grievance. For a social group that has experienced years of decline in its absolute living standards (whether measured by income, human development indicators or both), a peace deal that delivers an absolute rise in its living standard may be sufficient to secure its withdrawal from violent conflict. In this case, the current donor focus on absolute poverty reduction will also be conducive to reviving the social contract. But grievances may also be expressed in relative terms; the rebel group may seek a closing of the gap in its standard of living with wealthier groups, not just a rise in its absolute well-being. In this case, governments (and donors) will need to be much more concerned with reducing inequality, especially across regions. For example, Mozambique has seen some (modest) improvements in living standards in areas dominated by the rebel group, RENAMO, since the end of the civil war in 1992. But, continuing discontent among RENAMO supporters, which flared into violence in 2001, reflects their perception that the fruits of reconstruction have so far been concentrated in regions controlled by the FRELIMO government, so that the regional gap shows no sign of narrowing, and may indeed have widened.

16 Thematic Issues

Additionally, relative differences in living standards across social groups may play a greater part in fermenting grievances in societies rich in point resources, since the monetary value of the associated flows of income is more evident than in the case of societies dependent on diffuse, renewable resources, where much of the associated output is not marketed. The expected return from controlling natural-resource-rich countries is therefore considerably higher than in controlling countries which may have rich diffuse resources, but in which there are considerable costs in organizing and profiting from production (such as in agriculture) even if a belligerent secures the territory. Indeed, in natural-resource-rich countries it is sufficient to control only that part of the territory in which the resources are mined, which may exacerbate state collapse (DRC, for example). Second, the peace dividend. Development economists typically classify military spending as unproductive. But donor efforts to reduce military spending, and free public money for development spending, are typically thwarted, by moving the spending off-budget, even in countries such as Uganda that have demonstrated commitment to poverty-reduction. Countries can have legitimate security interests, especially when they find themselves in bad neighbourhoods, and poor people place security high on their list of priorities. Defence spending will remain high, with large opportunity costs for development spending, until the international community can offer a credible regional peace, thereby preventing harm to countries that contract their defence capabilities unilaterally. Yet UN peacekeeping operations remain grossly underfunded; their cost in the year 2000 amounted to less than half of 1 per cent of the approximately US$800 billion that member states spend on national defence (United Nations, 2000). If peace were credible, then the developmental impact of releasing those resources from military spending could be immense, and the resulting broad-based development thereby in turn solidifying the social contract upon which peace is ultimately sustained. Notes 1 On the complexity of the relationship between ethnicity and violence see Bates (2000), among others. 2 The term fiscal transfer also covers broad-based social expenditure across all groups, and the award of jobs and other benefits from government contracts. It could even mean power sharing. 3 Alternatively, the ‘contract’ governing the transfer from the government to the rebels may be highly incomplete. This is likely to arise in the context of state failure, and where government actions are not transparent. 4 A more detailed version of the analysis in this section can be found in Addison and Murshed (2002). 5 The pre-war equilibrium will depend on policies that in turn affect the relative price, and thus the structure of output. As drawn, agriculture takes a larger share of pre-war output relative to services, which is typical of a low-income country when policy does not have a large bias against farmers.

Addison and Murshed: Reconstruction 17 6 Moreover, conflict increases the cost of capital (as risk premiums on loan rates rise and as the monetary base shrinks), acting as a further disincentive to invest.

References Addison, T. and S. M. Murshed (2002) ‘Credibility and Reputation in Peacemaking’, Journal of Peace Research, vol. 39, no. 4, pp. 487–501. Addison T., P. Le Billon and S. M. Murshed (2002) ‘Conflict in Africa: the Cost of Peaceful Behaviour’, Journal of African Economies, vol. 11, no. 3, pp. 365–86. Azam, J.-P. (2001) ‘The Redistributive State and Conflicts in Africa’, Journal of Peace Research, vol. 38, no. 4, pp. 429–44. Azam, J.-P. and A. Mesnard (2001) ‘Civil War and the Social Contract’, paper presented at the Royal Economic Society Annual Conference, Durham, 9–11 April. Bates, R. H. (2000) ‘Ethnicity and Development in Africa: A Reappraisal’, American Economic Review, vol. 90, no. 2, pp. 131–4. Collier, P. and A. Hoeffler (2001) ‘Aid Policy and Peace: Reducing the Risks of Civil Conflict’, paper presented at the Royal Economic Society Annual Conference, Durham, 9–11 April. de Sousa, C. (2003) ‘Rebuilding Rural Livelihoods and Social Capital in Mozambique’, in T. Addison (ed.), From Conflict to Recovery in Africa, WIDER Studies in Development Economics (Oxford University Press). DfID (2001) The Causes of Conflict in Africa, Department for International Development, London (mimeo). Dixit, A. K. (2001) ‘Some Lessons from Transaction-Cost Politics for Less-Developed Countries’, Princeton University, USA (mimeo). Hirshleifer, J. (1995) ‘Anarchy and its Breakdown’, Journal of Political Economy, vol. 103, no. 1, pp. 26–52. Marshall, M. G. (1999) Third World War: System, Process, and Conflict Dynamics (Lanham, Md.: Rowman & Littlefield). Murshed, S. M. (2002) ‘Conflict, Civil War and Underdevelopment’, Journal of Peace Research, vol. 39, no. 4, pp. 387–93. Nafziger, E. W., F. Stewart and R. Väyrynen (eds) (2000) War, Hunger and Displacement: The Origins of Humanitarian Emergencies, Volume I, WIDER Studies in Development Economics (Oxford: Oxford University Press). Ndikumana, L. (2001) ‘Fiscal Policy, Conflict, and Reconstruction in Burundi and Rwanda’, WIDER discussion paper 2001/62 (Helsinki: WIDER). See www.wider. unu.edu. Olson, M. (1965) The Logic of Collective Action (Cambridge Mass.: Harvard University Press). United Nations (2000) ‘Report of the Secretary-General on the Work of the Organization’, United Nations General Assembly, Official Records, Fifty-Fifth Session, Supplement No. 1 (A/55/1) (New York: United Nations). Wallensteen, P. and M. Sollenberg (2001) ‘Armed Conflict 1989–2000’, Journal of Peace Research, vol. 38, no. 5, pp. 629–44. Walter, B. F. (2001) Committing to Peace: The Successful Settlement of Civil Wars (Princeton, NJ: Princeton University Press). World Bank (2001) ‘Development Co-operation and Conflict’, Operational Manual OP 2.30 (Washington, DC: World Bank).

2 The Economic and Political Consequences of Conflict and Implications for Post-Conflict Recovery in Africa Ibrahim A. Elbadawi and Njuguna S. Ndung’u

1

Introduction

In sub-Saharan Africa (SSA) one in every five people is directly affected by civil war. So it is not surprising that a recent multi-agency report on the prospects of Africa’s development identifies the task of overcoming the conflict–poverty trap as the most basic and most serious challenge facing sub-Saharan Africa at the time of writing (World Bank, 2000).1 The essence of this trap is not only reflected by the bi-causal relationship between growth (and hence poverty) and the risk of conflict, but also has pivotal linkages with other fundamental variables necessary for post-conflict economic recovery. Recent studies on the causes and consequences of conflicts, as pioneered by Collier and his research associates2 have focused on economic growth and other indicators of economic performance prevailing in periods before the war (Collier and Hoeffler, 1998, 2001) or assumed that the two phenomena of conflict and growth are driven by the same process (Elbadawi and Sambanis, 2001). Elbadawi and Ndung’u (2001) developed the models to ascertain multiple causal relationships to include additional intermediate development indicators such as private investment, public investment, and monetary and fiscal policies. This chapter utilizes the results of their structural model, composed of six behavioural equations: risk of conflict; per capita growth; private investment; public investment; money demand; and government revenue.3 Based on a panel of 130 countries over eight 5-year periods between 1970 and 1999, the results, when combined with a fiscal consistency equation, allow for the solution of the reduced-form equations for the six endogenous variables, as well as government consumption. The results inform further discussion of the economic and political agenda for post-conflict economic recovery in SSA countries. 18

Elbadawi and Ndung’u: Consequences of Conflict

19

The results of the structural model complement the endogenous growth analyses in three ways. First, the replication of empirical regularities, especially the causes and consequences of conflicts and civil wars in a multi-causal structure key, illuminates key indicators, policies and outcomes. Second, the empirical regularities may have underestimated these effects by failing to take into account the bi-causal and multi-causal structure of the key indicators, policies and other intermediate development goals. Finally, in incorporating a multi-causal structure, the analysis allows us to draw conclusions regarding appropriate policies for post-conflict recovery. The results corroborate the basic predictions of the literature on the determinants of the risk of civil wars. They also lend support to two key conclusions from studies on aid: that it is effective only in the context of good overall policy and institutional environment; and that its impact is subject to diminishing returns (Collier and Dollar, 1999; Burnside and Dollar, 2000). The results also lend support to key predictions on debt overhang; our estimations are stronger than the evidence found in the literature, where single-equation models are used. Five conclusions from the structural model results underpin this chapter: • by exploiting the multiple channels of influences allowed by accounting for mutual causation among the seven pivotal variables, our results show that some popular empirical findings (such as the non-monotonic effect of social fractionalization on risk of conflicts) as well as the aid effectiveness and debt overhang predictions (in the cases of growth and private investment) in fact apply to a much wider class of performance indicators; • in addition to its direct influence on the risk of conflict, the degree of political rights also influences economic performance through the public revenue channel, suggesting that countries with high levels of political rights have a lower risk of conflict and superior economic performance, because the legitimacy of their governments allows mobilization of resources for financing larger public investments; • public investment, rather than overall economic growth, appears to be the main economic variable associated with risk of conflict, implying that specific economic policies can reduce risk of conflict; • the association between public investment and risk of conflict on the one hand and public revenue and public investment on the other, provide multiple channels for the effects on the risks of conflicts of political rights, overall economic and institutional policy environment as well as primary exports. Similar effects on private investments and growth also obtain; and • the structural model also incorporates a money-demand equation allowing fiscal and monetary policy to interact and produce a policy mix required for economic management in post-conflict economic recovery.

20 Thematic Issues

Section 2 of this chapter provides an overview of the development conditions in conflict and post-conflict countries (CPCs), in particular those in SSA, highlighting the severity of the development trap facing CPCs, including the adverse external debt burden and policy environment they inherit in the immediate post-conflict period. The empirical results from the reduced-form analysis are discussed in Section 3, providing evidence on the extent of the development trap and the potential for reversing it through more aid and debt relief (under appropriate policy and institutional environments). Section 4 concludes.

2

The development tragedy of conflict

Conflict and post-conflict countries face a development tragedy. Conflict generates further conflict and poverty; countries emerging from conflict continue to suffer poverty and lingering risks of renewed conflict.4 The Human Development Index (HDI) for CPC countries averages only 0.362, much lower than other low-income, highly-indebted poor countries (HIPCs) at 0.457 (see Table 2.1). CPCs have lower levels of well-being; they also have higher infant mortality rates, an average of 195 per 1,000 births (using 1996 figures) compared to the HIPC average of 152. Even CPCs that have benefited from the HIPC initiative (such as Uganda and Mozambique) are still precarious. Uganda has an NPV of debt to exports of 239 per cent, while Mozambique’s is 588 per cent. This shows that reconstruction has not supported or influenced export activities. Furthermore, Mozambique has Table 2.1

Indicators of debt and wellbeing of the CPCs

Angola Burundi CAR Chad Ethiopia Eritrea Liberia Mozambique Rwanda Sierra Leone Somalia Sudan Uganda HIPC average HIPC II proposal CPC Average

HDI 1997

Infant mortality 1996 (per 1,000 births)

NPV of debt to GNP (%)

NPV of debt to exports (%)

0.398 0.324 0.378 0.393 0.298 0.346 na 0.341 0.379 0.254 na 0.475 0.404 0.457

209 176 164 189 177 na na 214 205 284 na na 141 152

291 59 48 34 127 56 218 137 437 95 73 192 31 107

0.362

195

138

165 548 244 196 791 na na 588 373 778 na 1358 239 372 150 528

Source: World Bank (2001), World Development Indicators.


21

one of the highest infant mortality rates in the group, and in both cases their HDI is lower than average HIPC levels. Furthermore, most CPCs face unsustainable debt ratios. The thirteen countries considered to be CPC countries in Africa have NPV debt ratios to export of goods and services higher than those prescribed in HIPC II (150 per cent) and higher than the HIPC average (see Table 2.1). The average NPV of debt to exports is 528 per cent for CPCs, while the HIPC average stands at 372 per cent, against a benchmark of 150 per cent. These indicators imply that CPCs are a special type of HIPC. However, CPC countries also face the risk of further conflict arising out of the underlying economic constraints that differ from other HIPCs. Understanding the causes of poverty and conflict requires a better understanding of the mutual causation between growth and conflict. A structural model of growth and risk of conflict should endogenize investment and indicators of fiscal and monetary policies. Figure 2.1 provides comparisons of these indicators for conflict and non-conflict countries. Average annual per capita growth rates between 1970 and 1999 were less than −0.2 per cent compared to almost 2 per cent for non-conflict countries. Other intermediate economic performance indicators show a similar pattern. The average shares of public investment, private investment and public revenue were about

35 30

Percentage

25 20 15 10 5 0 –5 gcapGDP

Figure 2.1

priv_inv

pub_inv

pub_rev

Conflict

5 years following the end of conflict

Non-conflict

10 years following the end of conflict

Economic performance in conflict and non-conflict countries, 1970–99

Notes: gcap GDP: GDP per capita growth; priv_inv: share of private investment; pub_inv: share of public investment; pub_rev: share of public revenue. Sources: World Bank Data Base; and Small and Singer (1994).

22 Thematic Issues

7 per cent, 10 per cent and 19 per cent, respectively, compared to 10 per cent, 14 per cent, and 29 per cent in non-conflict countries. Evidence suggests that post-conflict countries achieve significant recovery during the five years immediately following the end of the conflict. However, sustaining postconflict recovery for a longer period is problematic. An enriched analysis can furnish a long-term perspective of post-conflict recovery in order for these countries to escape the cycle of conflict and poverty. 2.1 Macroeconomic policies CPCs face dilemmas establishing macroeconomic stability, promoting growth, reducing poverty and establishing credible fiscal, monetary and exchange rate policies. Given the low growth rates registered during conflicts, there is usually pressure on the government to improve the economy. Considering general government consumption, total tax revenue and overall surplus or deficit as shares of GDP, it is evident that the CPCs gain large deficits during the conflict. The majority had healthy overall surplus/deficit ratios in the early 1980s. However, in the 1990s, the deficit situation worsened. Deficit levels were high even for countries that had entered a relatively peaceful period, such as Uganda.5 High deficits during conflict and immediately afterwards can be explained by high military expenditure. General government consumption as a share of GDP is not large relative to non-conflict countries, but this is because of low revenues, as shown by the total tax revenues to GDP ratio. Government consumption is higher during conflict, yet this is when total tax revenues as a proportion of GDP are poorest. Angola, for example, had government consumption in excess of 40 per cent, even during the UN peace accord period. But for Mozambique, the share of government consumption fell by 15.9 per cent at the cessation of conflict, to 10.4 per cent in 1997. However, as noted above, most CPCs have not exhibited dramatic declines in the share of government consumption in GDP during peace. This can be attributed partly to the neglect of fiscal discipline during conflict having a lagged impact on the following period of peace. Total tax revenues, hence general government consumption, for CPCs are low, except for Angola and to some extent Eritrea. The low ratios, particularly of total tax revenues, may be attributed to the collapse or breakdown of the revenue-collection mechanisms. It is clear that revenue growth is not commensurate with GDP growth rates witnessed at corresponding periods. Clearly, CPCs take a long time to rebuild institutions such as tax-collection agencies, but also take time to reduce war-related expenditures. Consequently, there is scope during peace to improve these ratios, even though this might be counter-productive in terms of business development, since businesses would also be trying to rebuild themselves. Debt relief assistance and fiscal support, such as International Monetary Fund (IMF) Enhanced Structural Adjustment Fund (ESAF) programmes, would enhance


23

reconstruction, easing pressure on the government to levy taxes on businesses in its attempt to restore social services. To avoid stifling growth while enabling CPC governments to provide social services, there is need to consider the special circumstances of post-conflict economies. Comparisons of policy and growth performance between CPCs and SSA countries are shown in Table 2.2. Real growth improved between 1995 and 1998 for a few successful CPCs such as Uganda. For those countries still in conflict, real growth remained low or negative. The outcome of policies, as reflected by inflation, shows high inflation regimes for the CPCs compared to SSA countries. CPC inflation apparently increased in the last period, rather than stabilizing, as it did in SSA countries. CPC fiscal policy stance is consistent with expectations: high fiscal deficits, more than the SSA average. But there were dramatic improvements during 1995–8: the average deficit to GDP fell to −4.7 per cent for the CPCs compared to the −2.5 per cent SSA average. This dramatic improvement indicates that peace settlements improve macro policy formulation. CPCs have four basic characteristics. First, regarding investment and savings, evidence shows they attract little foreign direct investment (FDI), and have low rates of savings and investment. FDI flows decline during conflict. Where flows have been reversed, as in Uganda, it is the peace dividend that has helped. FDI flows and domestic investments have a dimension indicating credible domestic policies and stability. Second, GDP has been precarious but population growth (hence dependency rates) has increased Table 2.2

Growth and macroeconomic policy stance (means)

Indicator Real GDP growth 1970–9 1980–4 1985–94 1995–8

SSA Countries

CPC Countries

4.40 1.80 2.70 2.60

0.63 0.91 0.35 4.20

Inflation 1970–9 1980–4 1985–94 1995–8

10.40 12.14 9.86 9.25

5.70 9.40 16.70 18.60

Overall fiscal deficit (incl. grants) 1970–9 1980–4 1985–94 1995–8

−2.70 −7.10 −4.30 −2.50

−5.90 −9.00 −8.90 −4.70

Sources: African Economic Indicators; World Bank, various years.

24 Thematic Issues

tremendously, leading to increased poverty. Third, regarding external debt and the ability to pay, CPCs form the bulk of countries considered to be HIPC. Ability to pay is minimal because of the destruction of economic activity and a shift of available resources to military activity. In order to support postconflict reconstruction and economic policies, debt stock reduction would be an appropriate starting point to allow flexibility for economic reform policies. Finally, fiscal (monetary) policies, exchange rate management and inflation management policies are required in support of growth. Once normality has been restored, strong growth-friendly policies to support and sustain economic recovery and minimize the risk of conflict can be implemented, strengthening private-sector activities in rebuilding markets, domestic competitive production and ensuring resource flows into profitable sectors. Collier and Gunning (1995) show that resolving the conflict-uncertainty problem takes a long time, but it is necessary to stabilize markets and induce a portfolio substitution towards fixed irreversible investments.

3

Empirical analysis6

We utilize a structural model of growth and conflict anchored around two pivotal variables: the rate of growth and the risk of civil war, where the latter is defined as the probability of observing a civil war at any given point in time (Elbadawi and Sambanis, 2001). The model also includes three other endogenous variables associated with growth and risk of conflict: private and public investments; a money demand equation; and public-sector revenue (expressed as ratios to GDP). Success in the area of public investments is likely to be the trigger for the virtuous circle of growth, poverty reduction and conflict prevention. On the other hand, failure can cause countries to slip into the cycle of rising poverty and conflict. Moreover, since public investment is usually the first casualty in conflicts, post-conflict reconstruction must begin by rebuilding the physical infrastructure, as well as public institutions. In addition to foreign aid, public investments will, in turn, depend on government revenue, mainly mobilized through the taxation of primary exports. Finally, in post-conflict countries, private investment must help to sustain growth, so that a meaningful reduction in poverty can be achieved and risks of future conflict reduced sufficiently. In addition, renewed private investment is a signal for a stable polity and for the credibility of domestic economic policies, which should translate into further inflows of foreign direct investment. Collier and Gunning (1995) show that fears of insecurity encourage private agents to hold their portfolios in liquid form, resulting in unpredictable returns because of uncertainty in the economy. This unpredictability induces erratic changes in money demand, complicating monetary policy. So rates of return should be adjusted in relation to the levels of risk of conflict prevailing in the economy.


25

Following the work of Collier and Hoeffler (2001), and Elbadawi and Sambanis (2001), several explanatory variables are included in the structural model. Controlling for social characteristics (such as ethnic and religious diversity), risk of civil war is explicable in terms of fundamental economic variables (per capita GDP; economic growth; public revenues; money demand; and dependence on primary exports, especially minerals) as well as indicators of political and human rights. Aid effectiveness and debt overhang literature are the most relevant to the empirical endogenous growth model for the analysis of economic performance in conflict and post-conflict countries. Consequently, the additional explanatory variables we consider include the stock of external debt to GDP; external debt service ratio; ratio of aid to GDP; the interaction of aid with indicators of overall policy environment; and traditional growth and investment explanatory variables (initial income, educational attainment, population growth, average inflation and real exchange rates). The structural model can be expressed as follows: F[gy, Pw, Ip, Ig, Ry, DEBTY, DEBTY2, INTPYMT, RINT, AIDY, AIDY2, AIDPOL, CPIA, GIY, Y0, School0, lagged gy, lagged Ip, FINDEP, FBY, AVGINFLN, RER, POPGRWTH, LEGACY, AIDYLGCY, POLITY, PRIMEXP, ETHNIC, ETHNIC2, ETHPOLITY] = 0 where gy, Pw, Ip, Ig, Ry are the five endogenous variables (rate of real GDP growth, risk of civil wars, private and public investments, and government revenue), with the last three variables expressed as ratios to GDP. The explanatory variables include the following: DEBTY, which stands for external debt stock to GDP; DEBTY2 is external debt stock to GDP lagged one period and squared to capture the overhang effect; INTPYMT is external debt service payment as a share of GDP; RINT is real interest rate; AIDY is foreign aid to GDP; AIDY2 is foreign aid to GDP squared; AIDPOL is foreign aid to GDP interacted with an overall policy index, CPIA, which ranges from zero (bad policy) to 6 (for good policy environment). Both AIDY2 and AIDPOL reflect two of the key findings of the aid effectiveness literature, where aid was found to be associated positively and robustly with economic performance in the context of good policy environment; however, the effect of aid is usually subject to diminishing returns. GIY is aggregate investment and POPGRWTH is the rate of population growth. The initial level of development is reflected by initial income and schooling (Y0 and School0, respectively). Traditional variables reflecting overall macroeconomic stability and competitiveness include: FINDEP, FBY, AVGINFLN and RER, which are, respectively, financial depth (ratio of stock of money to GDP), fiscal balance to GDP ratio, average inflation, and real exchange rate. LEGACY is a positive indicator for post-conflict countries (0 elsewhere), which is equal to the number of months during the first decade following cessation of conflict

26 Thematic Issues

multiplied by the total length (in months) of the previous war (Collier, 1999a). Collier regards this variable as a proxy for the ‘peace dividend’ effect. Both LEGACY and AIDYLGCY (AIDY × LEGACY) are expected to be associated positively with growth. The remaining variables (POLITY, PRIMEXP, ETHNIC, ETHNIC2, ETHPOLITY) are motivated by the literature on the risk of civil war. The variables, respectively, stand for the extent of political rights, share of primary exports to GDP, degree of ethnolinguistic fractionalization, its square and its multiple with (1-POLITY). POLITY is predicted to be associated negatively with the hazard of conflict, although Gleditsch and Hegre (1997) find the effect of POLITY to be positive and non-monotonic, suggesting that the risk of conflict can rise with democratization before reducing as the democratic process matures. Primary exports (PRIMEXP) is expected to be associated positively with conflict. Social diversity (ETHNIC) has been found to be associated either positively and non-monotonically with the risk of conflict (Elbadawi and Sambanis, 2001) or associated negatively with conflict when controlling for ethnic dominance. The latter effect is found to be associated positively with the risk of conflict (Collier and Hoeffler, 2001). In this model we also add an interaction term between ETHNIC × (1-POLITY), which should be associated positively with conflicts. Such an effect can be explained by the ‘grievance’ motive of civil war, which suggests that an absence of political rights renders socially diverse societies particularly vulnerable to the risk of civil war.

4

The estimation results

The structural model of the above equation was estimated for 133 countries over eight 5-year periods covering 1960–99, giving rise to a sample space of 1,064 data points. We first estimated fixed and random-effects panel regressions, and then re-estimated the models in non-linear, multivariate and full information maximum likelihood (see Tables 2.3–2.7). The last three methods utilize the minimum distance estimators. These three estimators are consistent with the two fundamental assumptions of the structural model of Equation (1): that the five development and intermediate target variables are characterized by mutual causality, and that the relationships between the variables may be non-linear. Because of this interdependence, the minimum distance estimators dropped the variables that were not significant even though they might have been significant in the panel regressions. The empirical procedure was to utilize various estimation methods in order to get robust estimates to help us to map the channels through which these five endogenous variables affect each other. In the multivariate regression, we utilize the fact that the disturbances from the equations may be correlated, since we assume that simultaneity exits within the endogenous variables. If this is the case, the estimation method would provide us with more efficient


27

estimates compared to the panel regressions, where each equation is estimated separately. In addition, since these equations share explanatory variables, the estimation method imposes cross-equation constraints. This method uses the generalized least squares estimator and, if the residuals are multivariate normal, the estimator will converge to the maximum likelihood estimator. In order to check for the robustness of the results we have finally estimated the model in full information maximum likelihood (FIML). This is an asymptotically efficient estimator under our assumption that the disturbance terms are multivariate normal. It has been shown that even when this assumption does not hold, FIML will still give us asymptotically efficient coefficients. If there are no simultaneity and non-linearity problems among the endogenous variables, the FIML estimator will coincide with the multivariate results. The results have no marked differences between the multivariate and FIML results. They also suggest that non-linearity may not be a problem in these results, but that the endogenous variables are interdependent. Therefore, we chose the multivariate regressions for further analysis and for the subsequent simulation exercise of the next section. The structural model can be elaborated in diagram form to show the causal structure of the three pivotal variables: growth, private investment and risk of war (see Elbadawi and Ndung’u, 2001). Other variables can be added to the comparative statics to show the richness of the mutual causality among the variables and the key indicators. 4.1 The growth model The results of the growth model suggest that the risk of conflict is associated negatively with growth, but that the length of peace following long wars (LEGACY) is associated with high growth. Both results are consistent with the received literature (Collier, 1999). Also our results strongly corroborate the theoretical predictions of the aid effectiveness and debt overhang strands of the literature. The results suggest that the growth–debt relationship is located on the ‘bad’ side of the Laffer curve, where only DEBTY2 is significantly (and negatively) associated with growth. In addition, the crowding effect of interest payments (INTPYMT) is also associated negatively with growth. Our results also corroborate a modified version of the aid effectiveness literature, where – regardless of the policy environment – aid contributes to growth but is subject to diminishing returns. However, our results also show that good policy environment further adds to the effectiveness of aid. In terms of macroeconomic policy, high inflation was found to be associated negatively with growth, while more competitive real exchange rates and high aggregate investments were positively associated with growth. The rate of population growth was positively, and perhaps surprisingly, associated with growth, but the full effect on growth for this variable will turn out to be negative (see the analysis of the reduced form effects in the following section). Finally, the multivariate estimation results suggest that

28 Thematic Issues Table 2.3

Growth model results

Explanatory variable

Fixed effects model

−0.224 (−3.970) −0.148 (−5.220) 0.077 −1.700 0.525 −11.120

−9.460 (−2.830) −0.261 (−5.750) −0.157 (−6.580) 0.084 −2.190 0.583 −19.470

0.023 −2.820 0.029 −0.895

0.017 −2.270 0.005 −0.190

CONSTANT RISK OF WAR DEBTY2 LEGACY AIDY

Random effect Non-linear model estimation

AIDY2 AIDPOL GIY Lagged growth AVGINFLN INTPYMT RER POPGRWTH R2 Hausman test

−0.030 (−1.330) −0.016 (−.802) 0.017 (−0.890) 0.654 (−26.350) 0.830

−0.024 (−1.160) −0.014 (−0.750) 0.011 (−0.640) 0.634 (−32.670) 0.800 4.980

−11.480 (−3.800) −0.330 (−6.100) −1.440 (−47.200) 0.051 −0.940 0.607 22.370 −0.090 (−1.150) −0.012 (−1.470) 0.015 −0.550 0.162 (−8.060) −0.027 (−1.340) −0.019 (−1.030) 0.014 (−0.780) 0.607 (−29.090) 0.740

Multivariate FIML estimation estimation −38.880 (−39.340) −1.133 (−26.130) −1.014 (−75.700) 0.203 −4.670 0.625 −24.480 −0.114 (−2.040) 0.074 −11.690 0.065 −4.440 0.043 (−3.280) −0.041 (−3.790) −0.075 (−6.720) 0.041 (−4.270) 0.641 (−30.030) 0.570

−28.690 (−8.290) −0.784 (−3.550) −1.060 (−42.790) 0.190 −3.660 0.621 −24.380 −0.014 (−0.060) 0.042 −3.590 0.068 −2.750 0.101 (−4.420) −0.044 (−2.370) −0.051 (−2.640) 0.032 (−1.490) 0.621 (−16.410) 0.670

Note: Figures in brackets are t-ratios. Source: Computed from estimated equations.

‘growth begets growth’; that is, past growth is likely to generate more growth in the future. This is important in explaining the episodes in CPC countries where risk of conflict depresses growth not only in the current period but also in subsequent years. 4.2 Private investment Table 2.4 shows the results from estimating the private investment equation in the same way. Public investment is associated positively with private investment, while the risk of conflict is inversely related to private investment. This is a channel through which the risk of civil war further reinforces a negative association with growth, which suggests that single equation


29

models are likely to understate the extent of this association. Public investment and risk of conflict are by far the most important determinants of private investments. The debt overhang does not have a direct effect, although, as expected, the real interest rate (RINT) has a negative and significant effect. On the other hand, aid is associated positively with private investment. Moreover, this direct effect of aid is not subject to diminishing returns. However, the net reduced-form effect would restore the predictions of the aid effectiveness literature (see section 5). Finally, the results suggest that Table 2.4

Private investment model results


Fixed effects Random effect Non-linear Multivariate FIML estimation model model estimation estimation

CONSTANT ATWAR

−0.002 (−0.405)

4.587 −4.686 −0.003 (−0.300)

DEBTY DEBTY 2 RER FINDEP AIDY AIDY2 AIDYLGCY PUINV RINT Lagged private Investment CPIA FBY AVGINFLN R2 Hausman test

−0.009 (−1.820) −0.006 (−0.500) 0.007 −0.454 0.002 −0.490 0.0005 (−0.170) 0.00008 −1.350 0.989 −24.300 −0.023 (−1.640) 0.067 −1.530 −0.0005 (−0.160) 0.0336 −1.310 0.011 −1.740 0.970

−0.008 (−1.960)

0.002 −0.310 0.003 −0.310 0.00004 −0.130 0.974 −93.830 −0.027 (−2.150) 0.070 −7.290 −0.0003 (−0.100) 0.027 −2.760 0.0097 −2.400 0.970 4.350


1.880 −2.510 −0.0001 (−0.010) −0.001 (−0.050) 0.001 (−0.050) −0.011 (−2.780) 0.024 −3.650

−38.880 (−39.340) −1.050 (−121.860) 0.032 −1.540 −0.033 −1.560 −0.007 (−2.190) 0.018 −3.270

−28.690 (−8.290) −1.110 (−36.060) −0.016 (−0.400) 0.017 −0.040 −0.009 (−2.180) 0.019 −2.490

0.007 −0.640

0.201 −11.190

0.179 −9.100

0.954 −88.740 −0.024 (−2.700) 0.078 −7.740 −0.113 (−0.340)

0.480 −50.170 −0.026 (−3.840) 0.026 −3.170 0.341 −29.720

0.549 −17.360 −0.028 (−3.120) 0.021 −2.580 0.413 −11.170

0.960

0.110

0.120

30 Thematic Issues

lagged investment, financial depth (FINDEP) and overall policy stance (CPIA) were all associated positively with private investment, while the latter is affected negatively by the real exchange rate. 4.3 The risk of civil war Table 2.5 shows the results from the estimated model of the probability of war, ATWAR. The model corroborates findings from previous literature in several aspects. For example, social diversity has a positive and non-monotonic effect, while primary exports are associated with a high risk of conflicts. A new result on the effect of social diversity, so far not obtained in the literature, is that absence of political rights makes socially diverse societies particularly Table 2.5

Risk of war model results


Fixed effects model

Random effect model

Non-linear estimation

Multivariate estimation

FIML estimation

−0.018 (−1.920) 0.283 −6.510 −0.018 (−0.850) 0.081 −1.090 −0.001 (−2.090) 0.431 −3.490 0.029 −3.250

5.620 −2.180 −0.019 (−1.820) 0.242 −11.740 −0.005 (−0.230) 0.068 −1.880 −0.0009 (−3.630) 0.500 −13.920 0.031 −3.430

2.110 −0.920 0.019 −1.770 0.170 −8.950 0.009 −0.410 0.024 −0.750 −0.0002 (−0.900) 0.756 −19.990 0.013 −0.123 0.0003 −0.780

−38.880 (−39.340) 0.005 −1.890 0.010 −2.280 −0.501 (−91.870) 0.017 −2.820 −0.00014 (−3.090) 0.165 −12.380 0.320 −30.360 0.0001 −2.070

−28.69 (−8.29) 0.006 −1.25 0.006 −0.800 −0.431 (−14.780) 0.018 −1.460 −0.00012 (−1.380) 0.145 −9.280 0.370 −12.600 0.0001 −0.130

CONSTANT Lagged growth POLITY PUINV ETHINIC ETHNIC2 PRIMEEXP CPIA ETHPOLITY AVGINFLN DEBTY DEBTY2 ETHPOL R2 Hausman test

0.043 −3.950 0.174 −1.610 −0.176 (−1.640) 0.144 −2.440 0.890

0.047 −4.09 0.158 −2.33 −0.158 (−2.33) 0.001 −3.67 0.87 43.03


0.780

0.210

0.190


31

vulnerable to the risk of conflict. Another new, and potentially very important, result is that public investment, rather than economic growth, was the economic variable associated most strongly and robustly with low risk of conflict. Finally, both POLITY and CPIA have had a perverse association with high risk of conflict. However, again the net reduced form effects restore the theoretically consistent predictions of the literature. Therefore, unlike single-equation analysis, this structural model allows for multiple channels of influence, which should prove valuable for developing policy instruments for post-conflict reconstruction as well as conflict prevention. 4.4 Public investment Table 2.6 shows the results for the public investment equation. The evidence suggests that growth and government revenues are associated with high public investments. Two important results that strongly corroborate the aid effectiveness literature are that a high level of aid tends to reduce public investment, but if delivered in a good policy environment it can augment such investment. LEGACY and population growth tend to reduce public investment. One explanation for the LEGACY effect, which also obtains in Table 2.6 Public investment model results Explanatory variable

Fixed effects Random model effect model

0.183 −3.650 0.028 −1.230

−75.490 (−24.950) 0.208 −9.370 0.029 −1.380

−0.008 (−1.440)

−0.006 (−1.140)

−0.039 (−2.600) 0.047 −1.640 0.025 −1.050 −0.0002 (−0.700)

−0.041 (−2.820) 0.038 −1.690 0.013 −0.450 −0.00008 (−0.300)

CONSTANT Revenue to GDP AIDY Growth AIDPOL LEGACY POPGROW POLITY ETHNIC ETHNIC2 Lagged per capita income

0.019 −1.210

0.019 −1.320

Non-linear Multivariate FIML estimation estimation estimation −66.160 (−23.820) 0.299 −13.020 0.051 −2.230 0.013 −0.510 0.003 −0.460 0.017 −0.690 −0.043 (−1.870)

−38.880 (−39.340) 0.743 −79.910 −0.307 (−21.090) 0.477 −50.360 0.070 −13.620 −0.122 (−5.760) −0.331 (−24.760)

−28.690 −8.290 0.899 −9.290 −0.170 −4.580 0.266 −5.990 0.082 −7.600 0.060 (−1.320) −0.178 (−4.670)

32 Thematic Issues Table 2.6

(continued)

Explanatory variable ATWAR DEBTY2 AVGINFLN AIDYLGCY R2 Hausman test

Fixed effects Random Non-linear model effect model estimation −0.002 (−0.090) 0.003 −0.152 −0.030 (−1.640) −0.003 (−0.970) 0.610

0.022 −0.580 0.010 −0.540 −0.040 (−2.570) −0.0005 (−0.368) 0.550 12.930

0.160

Multivariate FIML estimation estimation

0.970

0.130


the reduced-form case, is that, following long wars, government revenues are shifted increasingly from physical investment towards other social sectors as peace becomes established. 4.5 Government revenue Finally, we present the estimation results of the government revenue model (see Table 2.7). The results are important as they show that the risk of conflict has a strong and negative effect on government revenues. The other variable that affects government revenue negatively is external debt service payment. Two other important results that may be linked to aid effectiveness literature are that good policy (CPIA) and aid in a good policy environment (AIDPOL) are both associated with high public revenues, as are growth and primary exports. The latter provide for multiple channels for the effect of primary exports on the risk of conflict. Operating through public investment, the effect of primary exports on revenue would have a negative influence on the risk of conflict. This effect, however, is more than outweighed by its direct impact on the risk of conflict. Finally, the positive effect for POLITY suggests that political legitimacy is important for attaining high public revenues. This result has profound policy importance, especially in terms of highlighting the interdependence between economic and political development. Our results corroborate the basic predictions of the literature on the determinants of the risk of civil wars, and the aid effectiveness and debt overhang literature (the growth and private investment models). The results of our system estimations are much stronger than the evidence of either of the individual strands of the literature, which estimate single equation models.

Elbadawi and Ndung’u: Consequences of Conflict Table 2.7

33

Government revenue model results


Fixed effects model

Random effect model

Non-linear estimation

Multivariate estimation

FIML estimation

−0.079 (−1.380) −0.024 (−3.040) 0.013 −0.550 0.039 −2.080 0.016 −0.540 −13.950 (−5.360)

−83.350 (−31.990) −0.074 (−1.280) −0.023 (−3.350) 0.031 −1.110 0.050 −2.490 0.016 −0.190 −14.890 (−12.060)

−75.290 (−14.950) −0.083 (−1.600) −0.010 (−0.580) 0.056 −2.260 0.067 −3.730 0.035 −0.650 −17.350 (−13.130) 0.046 −1.250

−38.880 (−39.340) −1.590 (−56.030) 0.117 −24.170 0.042 −3.630 0.015 −2.280 0.414 −11.260 −3.640 (−6.580) 0.718 −30.630

−28.690 (−8.290) −2.260 (−8.730) 0.121 −8.130 0.042 −1.440 0.012 −0.640 0.435 −3.540 −1.460 (−1.950) 0.871 −9.890

CONSTANT ATWAR AIDPOL POLITY Growth PRIMEXP INTRY CPIA DEBTY DEBTY2 AIDY2 AVGINFLN LEGACY AIDY R2 Hausman test

0.074 −1.770 −0.065 (−1.580) 0.453 −4.020 0.003 −0.140 −0.360 (−3.620) −0.030 (−1.100) 0.580

0.047 −0.400 −0.035 (−0.300) 0.440 −4.150 −0.014 (−0.690) −0.360 (−6.950) −0.029 (−1.110) 0.520 0.630

0.17

0.210

0.250


Moreover, our estimation provides some further insight. For example, public investment, rather than growth, appears to be a main economic variable associated strongly (and negatively) with the risk of conflict. Moreover, the association between public investment and risk of conflict on the one hand, and public revenue and public investment on the other, provide multiple channels for the effects on the risk of conflict of POLITY and CPIA, as well as PRIMEXP. Similar effects on private investments and growth also obtain. A more detailed discussion of these reduced-form effects follows.

34 Thematic Issues

5

Reduced-form analysis

The empirical estimation started with panel estimation to assess the validity of the arguments. But, from the argument of mutual causality, it was felt that simultaneity would bias the results as well as the possibility of non-linear relationships in the model. So the next stage in the estimation was to re-estimate the model in non-linear form, and non-linearity was not a problem. The end result was to compare FIML estimation and multivariate estimation, which was preferred, since multivariate regression accommodates non-linear relationships in the data as well as simultaneity, otherwise addressed individually by non-linear and FIML estimations. But there arose a problem of identification, which was solved by reducing the model and retaining only the statistically significant variables in the multivariate regressions. This helped in identifying the reduced-form equations used in the discussion. Thus the results from the multivariate regressions have been used to construct a system of equations for growth; risk of war; private investment; public investment; money demand; and government revenue. In addition, we add a fiscal account identity, which allows derivation of an expression for government consumption (Gy). The reduced-form results for the six endogenous variables, shown in Table 2.8 were analysed for seven categories of exogenous variable: debt overhang; aid effectiveness; macroeconomic environment; structure of exports; levels of political rights; social diversity; and the peace dividend following end of civil war (legacy) and monetary policy.

Table 2.8

DEBTY DEBTY2 AIDY AIDY2 LEGAY AIDPOL GIY INF INTRY RER POP FINDEP RINT CPIA

Reduced-form results8 gy

Ip

Pw

Ig

Ry

Gy

0 −3.158 0.6167 −0.355 0.1038 0.9102 0.2024 −0.1277 −11.9483 0.1277 0.5626 0 0 −0.4559

0.032 −3.833 −0.0147 −0.2262 −0.1759 1.4821 0.2436 −0.1536 −21.0438 0.1466 −0.1389 0.018 −0.026 −0.1604

0 1.8923 0.0073 0.2127 0.0876 −0.7381 −0.1213 0.0765 10.4795 −0.0765 0.0692 0 0 0.4024

0 −3.7771 −0.0146 −0.4246 −0.1748 1.4732 0.2421 −0.1527 −20.9173 0.1527 −0.1381 0 0 −0.1644

0 −3.0562 −0.0024 −0.3436 −0.1377 1.3042 0.1959 −0.1236 −20.4817 0.1236 −0.1016 0 0 0.0714

1.000 0.7209 1.0122 0.0811 0.0371 −0.169 −0.0462 0.0292 −0.5644 −0.0292 0.0365 0 0 0.2358


Table 2.8

(Continued)

POLITY POLITY2 ETH ETH2 PEXP ETHPOL LGY PIYL DEBTYt−1 CONST.

0.0487 0 −0.147 0.0012 −0.0942 −0.0009 0.0907 0 0 −78.4795

0.0959 0 −0.2442 0.002 −0.0088 −0.0014 0.0893 0.026 0 −146.3149

−0.043 0 0.1297 −0.0011 0.0831 0.0008 −0.0421 0 0 34.951

0.1058 0 −0.225 0.0019 0.1635 −0.0013 0.094 0 0 −147.3673

0.1111 0 −0.2085 0.0017 0.2804 −0.0012 0.0683 0 0 −95.6293

35

0.0053 0 0.0165 −0.0001 0.117 0.0001 −0.0257 0 −1.000 51.738

Source: Computed from estimated equations.

5.1 Debt overhang Increases in external debt lead to a rise in government consumption, while foreign debt has a negative and a quadratic effect on growth, and private and public investment as well as money demand, but has no effect on public revenues (see Figures 2.2 and 2.3). Debt flows reduce the risk of conflict but the stock of debt increases the risk (see Figure 2.4). External debt seems to find its way into financing government consumption (see Figure 2.5) but, interestingly, it affects public revenues in the same way. However, the stock of external debt does not affect public revenues (see Figure 2.6) but has a negative effect on public investments. 5.2 Aid effectiveness Abstracting from the quality of the policy environment, the effect of aid (AIDY, AIDY2) has the usual inverted U-shape in the case of growth, which corroborates the previous evidence on diminishing returns to aid. However, aid seems to increase private and public investments but raises the risk of conflicts and government consumption quite substantially. Taking into consideration the effect of the overall policy environment (CPIA),7 the monotonic effect on growth disappears and the positive effects on private and public investment are strengthened, the effects on public revenues and government consumption are significantly negative, and the risk of conflict is reduced significantly. Therefore, accounting for policy, the overall net effects of aid on growth, private savings, public savings and public revenues produce a family of non-monotonic, inverted U-shaped (and, at times, mixed) relationships, reflected generally by Figure 2.7. Moreover, accounting for policy, aid is associated negatively and non-monotonically with the risk of conflicts and with government consumption (see Figure 2.8).

Growth

36

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.1

0.12

0.14

0.1

0.12

0.14

External debt

Growth and external debt

Private investment

Figure 2.2

0

0.02

0.04

0.06

0.08

External debt

Private investment and external debt

Public investment

Figure 2.3

0

Figure 2.4

0.02

0.04

0.06 0.08 External debt

Public investment and external debt

37

Revenue


0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

External debt Government consumption and external debt

Government consumption

Figure 2.5

0

Figure 2.6

0.02

0.04

0.06 0.08 External debt

0.1

0.12

0.14

Revenue and external debt

These are important results. First, they corroborate the two main theses of the aid-effectiveness literature: that aid is effective in a good policy environment, but is subject to diminishing returns. Second, whereas previous analyses focused on growth, investments and savings, these results corroborate the aid-effectiveness theses for a wider class of economic performance indicators as well as for the risk of conflict. Third, they show the dire consequences of provision of aid in the absence of a good policy and institutional environment (in terms of high risk of conflict, low public and private investment, low public revenue, high government consumption, and possibly an inherently inefficient patronage system). In the absence of a good policy environment aid flows increase the risk of conflict, but in a positive policy environment they reduce the risk of conflict and also have a positive effect on private and public investment.

38

80

Gy, Iy, Ig, Ry (%)

60 40 20 0 –20 –40

0

2

4

6

8

10

12

14

16

18

Ratio of aid to GDP PPP Gy Figure 2.7

Ig

Iy

Ry

Aid effectiveness

Risk of war, government revenue (%)

70 60 50 40 30 20 10 0

0

5

10

15

Aid as a ratio of GDP PPP Pw Figure 2.8

Aid effectiveness

G

20

25


39

5.3 Macroeconomic environment Three macroeconomic policy variables (INF, DNER, FITRINT) are considered. Real interest rate (FITRINT), adjusted for risk of conflict, is associated negatively with private investment and money demand, but has no effect on the other variables. Inflation is represented by its level and a crisis dummy (INFL40) when inflation level is consistently above 40 per cent. The results seem to show that inflation reduces growth, private investment, public investment and money demand, but is associated with a higher risk of conflict, public revenue and government consumption. The change in nominal exchange rate (DNER) has no effect on the other variables except a weak negative effect on money demand. Exports The results suggest that a high share of primary exports (XGDP) is associated with lower growth and private investment. The literature suggests that economies dominated by primary exports tend to have higher public investment, public revenues and government consumption, as well as a higher risk of conflict. It is only the risk of conflict that appears to be significant in the results we report here, but, as the literature suggests, in economies dependent on a precious non-renewable natural resource base (such as oil), these effects tend to be much stronger. However, this version of the model fails to pick up the non-monotonic effect discussed in the economics of conflicts literature (for example, Collier and Hoeffler, 2001), implying that countries very rich in such resources are likely to be safe from the risk of civil war because they have the capacity to build sufficient deterrence, such as high public investment. Political rights The results suggest that high standards of political rights are associated with lower risks of conflicts, as predicted in the literature (see Elbadawi and Sambanis, 2001), but also with high levels of growth, private investment, public investments and public revenue as well as government consumption. The implication of this is that countries with high standards of political rights have superior economic performance and a lower risk of conflicts because the legitimacy of their governments allows the mobilization of resources for financing larger public investments.9 However, these results fail to pick up the effects of POLITY that would run from public revenues to growth, private investment and public investments that would strengthen the indirect channels. Social fractionalization Consistent with some findings in the literature on the determinants of civil conflicts,10 these results find a strong positive but non-monotonic relationship

40 Thematic Issues

between social fractionalization and risk of conflict. Moreover, there is a family of non-monotonic relationships between social fractionalization and economic performance. For example, growth, private investment and public investment, as well as public revenue, all have a U-shaped relationship with social fractionalization (see Figure 2.9). On the other hand, as in the case of the risk of conflicts there exists an inverted U-shaped relationship between government consumption and social fractionalization (see Figure 2.10). The figures reflect the net effect of social fractionalization, which also account for the incremental influence of the combination of low standards of political rights (1-POLITY) and social diversity. This latter effect brings home the fact that, in the absence of political rights, socially fractionalized societies are particularly vulnerable to the higher risk of conflict (as can also be seen with risk of conflict increasing with the number of ethnic groups) and economic development failures. The results suggest that socially diverse societies with high standards of political rights and other favourable initial conditions do well, but socially polarized societies are predicted to have a much harder time. The peace dividend The extent of the peace dividend following civil wars is modelled by Collier (1999) through the effect of LEGACY on economic growth, where he defines

25

Ig, Gy, Iy, Ry

20

15

10

5

0

0

20

40

60

80

Ethnic fractionalization Gy Figure 2.9

Iy

The consequences of social diversity

Ig

Ry

100

120


41

Risk of conflict, government revenue

6 5 4 3 2 1 0

0

20

40

60

80

100

120

Ethnic fractionalization Pw Figure 2.10

G

Consequences of social diversity

the former as the number of months during the first decade following the cessation of conflict multiplied by the total length (in months) of the previous war. Contrary to Collier (1999), we find LEGACY to be associated positively with private investment. However, when LEGACY is crossed with AIDY, it is found to be associated negatively with private investments and has no effect on other variables. This suggests that, following long wars, growth is driven primarily by capital accumulation efficiency, as one would expect. We have included other variables which define the risk of conflict for CPC countries; for example, we have POSTC, which depicts countries considered to be post conflict, and WLENGTH to capture the effects of prolonged conflict. The variables, as expected, have the effect of increasing conflict and have negative effects on growth and private investment. Money demand and fiscal policy implications Money demand is driven by the standard variables, but in addition, we have all the other variables in the system entering the model. The results show that aid increases money demand but, in a poor policy environment, money demand is affected negatively. External debt and debt stock reduce money demand. What is interesting is that the variables that seem to increase the risk of conflict enter the money demand equation and seem to support Collier and Gunning’s (1995) thesis that portfolio substitution,

42 Thematic Issues

high liquidity and fiscal demand after the end of conflict make money demand erratic. On the fiscal side, the success of government revenue generation is important in financing government consumption and predicting future public revenues. Thus lagged effects of public revenues have significant consequences for current revenues, risk of conflict, private investment, government consumption and money demand, reflecting the complicated web of policies required for post-conflict economic recovery.

6

Conclusions

This chapter has attempted to replicate and strengthen the results of the recent literature on growth and poverty eradication in conflict and post-conflict countries, and to try to show new direct and indirect transmission channels through which risk of conflict is mediated in CPC countries. This strengthens the discussion of the recovery agenda in CPCs. There are three sets of conclusions. First, we have corroborated the evidence of previous studies even when we have calculated sets of equations that take into account multi-causal structures in the variables. The results strongly support the conclusions of the recent literature – for example, that a good policy environment will make aid more effective, and that the effect of aid is subject to diminishing returns. But the results in this chapter provide stronger evidence, suggesting that single-equation estimation may have underestimated these effects. Second, we have shown that single-equation models estimated in the recent literature may have underestimated the strength of the transmission channels from growth to conflict via other intermediate variables and policies. This may also imply that the policy discourse from these sets of results can include a variety of options that will lead to an effective postconflict recovery. Finally, there are new channels of transmission that are likely to be important in economic policies and the development path of post-conflict recovery: • The non-monotonic effect of, for example social fractionalization, on risk of conflict, aid effectiveness and debt overhang on growth and private investment, found in the recent analyses, apply not only to these variables but can be generalized to a wider class of performance indicators. • The degree of political rights has a stronger effect on economic performance through its effect on public revenues. • Public investment, rather than overall growth, is the main channel through which the impact of conflict is mediated to the rest of the economy. This insight provides a useful handle for reducing risk of conflict and revamping growth and investment.


43

• Public investment, risk of conflict and public revenues jointly provide multiple channels through which the effects of risk of conflict on political rights, on the overall economic and institutional policy environment and on primary export performance, are transmitted. • Absence of political rights in socially fractionalized societies makes them particularly vulnerable to a higher risk of conflict, compounds the development failures of poverty, and the collapse of law and order. The impact is further compounded when an absence of political rights and the number of ethnic groups have a combined effect on increasing the risk of conflict. • Exports are associated with high public revenues. The results show that exports seem to increase the risk of conflict: exports of primary extraction owned by the state increase public revenues, providing multiple channels where the effect of primary exports on risk of conflict is compounded by its effect on public revenues and public investment. Notes 1 The World Bank publishes the book on behalf of five development institutions, which include in addition to the Bank, the African Development Bank, the Economic Commission for Africa, the African Economic Research Consortium and the Global Coalition for Africa. 2 See, among others, Collier and Hoeffler (1998, 2001), Collier (1999) and Collier, Hoeffler and Pattillo (1999), to mention just a few. 3 This model allows a choice of policy mix in the development agenda of post-conflict recovery. 4 In the period 1960–99, sixty-six countries experienced at least one episode of civil war. 5 Data for CPC countries in the SSA region until 1997 are available on request from the authors. These include indicators of total external debt to GDP ratio; overall surplus/deficit as a share of GDP; general government consumption as a share of GDP; FDI as a percentage of GDP; gross domestic investment as a percentage of GDP; gross domestic savings as a share of GDP; and real growth of GDP at market prices. 6 An earlier draft of this chapter presented a model consisting of six behavioural equations and one identity, and the reduced-form equations derived from these. The model and the reduced-form equations, together with further details of the endogenous and exogenous variables, are available from the authors. 7 On the other hand, the influence of CPIA through aid dominates its independent effect. At least in the cases of low income countries, this result appears plausible. 8 t-ratios are available from the authors. 9 Azam (2001) notes that, under President Houphouët-Boigny, the Ivorian government bought peace by engaging in an active policy of visible redistribution across regions via public investment. For example, he notes that the taxation of cocoa and coffee was used to fund, among other public investments, the development of the San Pedro harbour in the south-west region, the area most hostile to national unity. 10 Collier and Hoeffler (1998) provide the arguments for the positive but nonmonotonic association between social fractionalization and the risk of civil wars (see also Elbadawi and Sambanis, 2001).

44 Thematic Issues

References Azam, J.-P. (2001) ‘The Redistributive State and Conflicts in Africa’, Journal of Peace Research, vol. 38, no. 4, pp. 429–44. Burnside, C. and D. Dollar (2000) ‘Aid, Policy and Growth’, American Economic Review, vol. 90, no. 4, pp. 847–68. Collier, P. (1999) ‘On the Economic Consequences of Civil War’, Oxford Economic Papers, vol. 51, no. 1, pp. 168–83. Collier, P. and D. Dollar (1999) ‘Aid Allocation and Poverty Reduction’, World Bank Policy Research working paper no. 2041, World Bank, Washington, DC. Collier, P. and J. W. Gunning (1995) ‘War, Peace and Private Portfolios’, World Development, vol. 23, no. 2, pp. 233–41. Collier, P. and A. Hoeffler (1998) ‘On the Economic Causes of Civil War’; Oxford Economic Papers, vol. 50, no. 4, pp. 563–73. Collier, P. and A. Hoeffler (2001) ‘Greed and Grievance in Civil war’, working paper no. 2355, World Bank, Washington, DC. Collier, P., A. Hoeffler and C. Pattillo (1999) ‘Flight Capital as a Portfolio Choice’, World Bank Policy Research, paper no. 2066, World Bank, Washington, DC. Collier, P., A. Hoeffler and M. Söderbom (2002) ‘On the Duration of Civil War’, World Bank, Washington, DC (mimeo). Elbadawi, I. A. and N. Ndung’u (2000) ‘External Indebtedness, Growth and Investment in Conflict and Post-Conflict Countries’, paper presented to World Bank Conference ‘The Economics of Political Violence’. Elbadawi, I. A. and N. Ndung’u (2001) ‘The Economics of Civil Wars and PostConflict Recovery’, paper presented at World Bank Conference ‘Civil Wars and Post-Conflict Transitions’. Elbadawi, I. A. and N. Sambanis (2001) ‘How Much War Will We See? Explaining the Prevalence of Civil War’, Journal of Conflict Resolution, vol. 46, no. 3, pp. 307–34. Elbadawi, I. A., B. Ndulu and N. Ndung’u (1997) ‘Debt Overhang and Economic Growth in Sub-Saharan Africa’ in Z. Iqbal and R. Kanbur (eds), External Finance for Low-Income Countries (Washington DC: International Monetary Fund). Gleditsch, N. P. and H. Hegre (1997) ‘Peace and Democracy: Three Levels of Analysis’, Journal of Conflict Resolution, vol. 41, no. 2, pp. 283–310. Small, M and J. D. Singer (1994) ‘Correlates of War Project: International and Civil War Data 1816–1992, Inter-University Consortium for Political and Social Research, Ann Arbor, Ml. World Bank (2000) Can Africa Claim the 21st Century? (Washington, DC: World Bank). World Bank (2001) World Development Indicators (Washington, DC: World Bank).

3 Economic Policy in Post-Conflict Societies Paul Collier

1

Introduction

Once a civil war has ended, what economic policies should a society adopt? Should institutions such as the World Bank, which provide policy advice, recommend policies that are distinctive because the society is post-conflict? Should donors modify their own policies – notably their rules of aid allocation – so as to make special provision for countries that are post-conflict? At some general level, the objectives of policy should be common across societies. Regardless of whether societies have recently come out of conflict, governments, and the donors that support them, should presumably seek to maintain peace and to promote growth. Although objectives may be common, policies will legitimately differ. The policies that are successful in one country are likely to bear a family resemblance to those that are successful in others, but they will differ somewhat according to country circumstances. Hence, the policies appropriate for a particular post-conflict society will inevitably depend in part on characteristics specific to that country at that time. The question posed here is not whether policy needs to be tailored to the circumstances of an individual country, but whether policy should differ systematically for a country which has recently come out of civil conflict. Post-conflict policies may need to be distinctive, for three reasons. First, peace may need to be given an unusually high priority in post-conflict societies, and so policy will need to be shaped disproportionately by the needs of peace-building. Second, post-conflict societies may have systematically distinctive economic problems that change the detail of which policy configurations are most effective for growth. Third, the political economy of post-conflict societies may be systematically distinctive, so that the attainment of policy reform in general, and of some reforms in particular, is made easier or more difficult than in other societies. I deal with each of these in turn. Section 2 discusses which government and donor policies are likely to be important for sustaining post-conflict peace, Section 3 discusses which government and donor policies are likely to be important in increasing growth 45

46 Thematic Issues

rates in the post-conflict phase, and Section 4 considers the political economy of policy reform in post-conflict situations. Section 5 concludes.

2

Prioritizing peace

The typical post-conflict society has a high risk of reverting to conflict: a typical society emerging from conflict has a 50 per cent chance of reverting to conflict within five years. This alarmingly high risk is made up of two approximately equal components. One is the risks that generated the previous conflict; usually, these risks will have remained. The other component is made up of those risks that were generated during the previous conflict. If peace is maintained, this second group of risks gradually recede, so that after around twenty years the society recovers to the level of risk it had prior to the conflict. The risks that caused the original conflict will differ, society by society. However, Hoeffler and I found that four risk factors are usually important: (i)

countries with a low per capita income face a much higher risk of conflict – and a doubling of per capita income approximately halves the risk. This is why conflict is common in low-income countries; (ii) countries with slow or negative growth face increased risks. In any five-year period, an additional percentage point on the growth rate reduces the risk of conflict by around one percentage point. For example, the catastrophic economic decline in Zaire during the 1990s may have contributed substantially to the civil war of the late 1990s; (iii) high rents from primary commodity exports increase the risk of conflict substantially. For example, the rents from oil and diamonds in Angola evidently contributed to its civil war; (iv) the structure of the society; countries in which the largest ethnic group constitutes a majority of the population (somewhere in the range 45–90 per cent) have around double the risk of conflict compared to countries with either more ethnic diversity or ethnic homogeneity. Evidently, in such circumstances the largest ethnic group can, in a democracy, be permanently dominant. Such ethnic dominance is less common in Africa than in other developing regions, but some countries, such as Rwanda and Burundi, are characterized in this way. Hence, part of post-conflict policy for peace is to reduce these risks that were inherent prior to the conflict. Since three of these risks concern economic performance, economic policies are directly important for risk reduction. Raising the growth rate will reduce risk directly, and further, will reduce risk cumulatively by raising the level of income. The policies appropriate for rapid growth in post-conflict societies are discussed below.

Collier: Economic Policy in Post-Conflict Societies

47

Additionally, policy may need to be aimed at managing the risks from high rents from primary commodities. In the short term, the best strategy is to reduce popular suspicions that such rents are misappropriated by setting up transparent systems of scrutiny and control. The involvement of civil society in the scrutiny of the revenues from the Chad–Cameroon pipeline is a pioneering example of such an attempt. In the long term, it is desirable to diversify the economy away from high primary commodity dependence. Research has shown that the same economic policies that promote growth also facilitate diversification (Collier and Hoeffler, 2002a). The remaining risk factor, ethnic dominance, cannot be dealt with simply by economic policies. Evidently, in such situations minorities require credible guarantees that their rights will be respected. In post-conflict situations, constitutional guarantees or other means of government commitment may have limited credibility to minority groups, and there may be a need for international intervention to provide reassurance. I now turn to the risks that are generated during the conflict. These arise partly because of the build-up of hatreds, and partly because of the build-up of interests. It is unsurprising that civil war polarizes society, making cooperative behaviour more difficult. Various strategies can be used to make coexistence easier, recognizing that the starting position is usually an extreme lack of trust. One approach is to agree on sharing formulas for government revenue between regions, and for public-sector jobs between ethnic groups. For example, following the Biafra war, Nigeria adopted a revenue-sharing formula that has been successful in avoiding further secessionist conflict. Even Switzerland has an explicit formula for sharing civil service jobs among the three main ethno-linguistic groups in Swiss society. The other way in which risks are generated during conflict is that some participants in the conflict come to have an interest in continuing that conflict and so may make active attempts to undermine any peace settlement. Although the divided society as a whole loses from conflict, some participants do well out of war. Further, they accumulate skills and armaments that have few uses other than in the context of conflict. There is often only a fine line between the activities of rebel groups and of organized criminal extortion rackets. One of the central tasks of a post-conflict government is to provide ‘carrot and stick’ incentives for such groups to integrate into the political and economic opportunities of a peaceful society, rather than to revert to violence. A good example of political integration was the change in RENAMO from being a rebel military organization in Mozambique into a democratic political party. A common source of post-conflict risk comes from diasporas, sometimes in neighbouring countries and sometimes in developed countries. Diasporas may harbour grievances and help to finance extremist organizations. A useful approach to reducing this risk is to provide the diaspora, and in particular diaspora organizations, with a recognized and positive role in

48 Thematic Issues

rebuilding the society. For example, in the current rebuilding of Afghanistan, the Internet has been used to create a network of the Afghani diaspora, classified by skills. The diaspora is often a major resource for building business networks. By giving it a formal role in the peace process it is more likely to be harnessed for economic recovery as opposed to subversion. In addition to policies that are targeted at a specific component of risk, such as ethnic dominance or diasporas, some policies may have overall effects on risk. Here I consider military expenditure and aid. Military expenditure is initially high following a civil war. Globally, during a civil war military expenditure rises by around 1.9 percentage points of GDP, compared to its usual level of around 3.3 per cent. Following conflict this is a considerable fiscal burden, and an important question is whether this level is appropriate, given the high risk of further conflict. It might be that continued high military expenditure is an effective way of maintaining peace. Governments not only increase military expenditure during a civil war, but also in response to a risk of civil war. However, such deterrent expenditure appears to be ineffective. A high level of military expenditure does not, in general, reduce the risk that a civil war will be initiated (Collier and Hoeffler, 2002b). There are various reasons why this might be the case. Much military expenditure is often highly inefficient: the military is the least scrutinized part of the public sector. Further, there may simply be insufficient intelligence information to make good use of military forces in the inception stages of a rebel organization. Finally, counter-insurgency operations are extremely difficult. The most common military error is to over-react, harming the civilian population and creating resentments that assist rebel recruitment. Most armies lack the tight discipline necessary to avoid such behaviour. Hence, even when government forces engage with incipient rebel forces they may be as likely to increase the problem as to reduce it. Evidently, if rebel forces escalate, then the government has no alternative but to counter the threat with military force. However, military deterrence appears to be an ineffective strategy for securing post-conflict peace. Economic and political strategies are likely to be more effective. Aid can affect the risk of conflict through various routes. It has effects on growth (discussed below) and thus has indirect effects on the risk of conflict. Here I consider whether aid has a direct effect on the risk of conflict, over and above these growth effects. Some analysts have suggested that aid might increase the risk of conflict directly by increasing the resources to be contested (Grossman, 1992). Others suggest that aid can enable the government to placate the opposition (Azam, 1995). Collier and Hoeffler (2002a) introduce aid directly into their model of conflict risk, controlling for economic outcomes, and find that there is no significant effect one way or the other. Hence, on the evidence, aid does not have strong, direct peacebuilding effects other than via its effect on economic performance, but equally, nor does it increase the risk of conflict.


49

Concern has also been expressed as to whether aid inadvertently finances military expenditure. Collier and Hoeffler (2002b) investigate this by introducing aid into a military expenditure function. Military expenditure varies considerably between countries, and these differences are well-explained by factors such as the previous history of conflict, and the military expenditure of neighbours. However, aid does not influence military expenditure significantly. This is surprising, since it might be expected that aid would be sufficiently ‘fungible’ so that some of it would finance indirectly the military component of the budget. This is especially the case since military expenditure is income-elastic, rising more rapidly than national per capita income as a whole. Evidently, donor pressure to moderate military spending is normally sufficient to counter this tendency. The most important conclusion from this discussion of post-conflict risks to peace is that rapid economic recovery is likely to be important in sustaining peace. Although in some situations there may be trade-offs between growth and peace, in most cases, policies that augment growth will tend to build peace. Aid does not generally have adverse consequences for peace and so can be determined sensibly by the ability of the economy to absorb it productively in a growth process, a matter discussed in Section 3. Military expenditure retards growth without deterring rebellion and so the high levels found in post-conflict societies are likely to be dysfunctional.

3

Rebuilding the economy

I now turn from the objective of maintaining peace to that of rebuilding the economy. During a civil war the economy typically contracts, either absolutely or at least relative to its potential. It also changes its structure: commonly there is a retreat into subsistence. All forms of capital are reduced: private investment collapses and is replaced by capital flight; public capital, such as infrastructure, is destroyed; and ‘social capital’, such as trust and reputation, is lost. The post-conflict society is thus short of both private and public capital, and is characterized by opportunistic behaviour. Post-conflict economies have wide variations in performance. Some have recovered very rapidly, while others have continued to decline virtually as if the war were continuing. A critical aspect of performance is whether people come to believe that peace will be maintained. Confidence in peace is selffulfilling, since it encourages private investment, and the resulting growth tends to consolidate the peace. A useful indicator of success in restoring confidence is whether flight capital returns. Typically, in post-conflict societies a large part of private wealth has been taken out of the economy. For example, in Uganda, by the end of the period of social disruption (1986) over 60 per cent of private wealth was held outside the country. By taking strong action to restore confidence in property rights and other reforms, such as restoring a competitive exchange rate, by the early 1990s this flight

50 Thematic Issues

capital was starting to return to the economy. In some years, capital repatriation was larger than exports. Globally, both policy reform and aid are important motors of economic growth. They are complementary – aid is more effective in a better policy environment, and policy reform is more effective in conjunction with large aid inflows (Collier and Dollar, 2002). In post-conflict situations, the complementary role of aid and policy reform is even more important (Collier and Hoeffler, 2002c). Other than in post-conflict situations, the relationship between aid and growth is estimated to be: dG/dA = 0.127P − 0.056A

(1)

where G = the growth rate of GDP; A = aid as a share of GDP, measured at purchasing power parity exchange rates; and P = the World Bank index of policy, the ‘Country Policy and Institutional Assessment’, with 1 as very poor policy and 5 as very good policy. The relationship described in Equation (1) tells us that there are diminishing returns to aid, but that aid is more effective in a better policy environment. For example, a fairly typical policy score is 3, and a fairly typical aid inflow is 3 per cent of GDP at purchasing power parity prices (equivalent to around 9 per cent of GDP at market exchange rates). For such a country, the marginal percentage point of aid raises the growth rate by around 0.2 percentage points. Hence, the marginal rate of return upon aid is around 20 per cent. For a country during the first decade of post-conflict, the relationship between aid, policy and growth is much more favourable, being: dG/dA = 0.313P − 0.056A

(2)

Hence, for the same values of the policy score and aid, the marginal contribution of aid to the growth rate would be around 0.77 percentage points. Therefore the marginal return on this volume of aid in a post-conflict situation would be around 77 per cent. An implication of this is that donors should allocate aid disproportionately to post-conflict situations. In practice, historically, this has not been the case: donors have given about the same amount of aid to post-conflict countries as to other countries. The appropriate amount of aid a postconflict country should receive, however, is a difficult issue. But there is a reasonable presumption that the lower bound for an allocation is that in which the marginal contribution to growth would be approximately equal to that of aid in non-conflict countries. This is a lower bound because, in post-conflict settings, aid may have a peace-stabilizing effect over and


51

above its effect on growth, although, as discussed above, we did not find evidence for such an effect. On the criterion of growth alone, aid as a share of GDP in post-conflict situations should be approximately double that in non-conflict situations. We can derive this from Equations (1) and (2) by introducing the concept of the saturation point where the society has so much aid that diminishing returns have driven the marginal contribution of aid to growth to zero. The saturation point is determined by setting dG/dA = 0, and solving for A in terms of P. For non-conflict situations, the saturation point is around 2.5P, whereas for post-conflict situations it is around 5.6P. Thus, for a common level of policy, the maximum volume of aid that is useful in a post-conflict setting is over twice the normal level. In the typical post-conflict setting, policy is initially considerably worse than normal, and this qualifies the amount of aid that can usefully be absorbed. However, even once we introduce a discount to correct for worse policy, aid should still be approximately double its normal level. This suggests that the donor community could improve its performance by allocating more aid to post-conflict countries. The same analysis also provides some guidance as to how these additional flows of aid should be timed. For the first three or four years after the end of conflict, the economy does not have any above-normal capacity to absorb aid. The extra scope for aid is approximately from the fourth or fifth year through to the end of the first decade. This pattern may be a result of the general disarray of public institutions in the initial post-conflict years which makes it difficult to absorb large volumes of aid productively. Hence, post-conflict aid should taper in gradually over a decade. In practice, aid flows have had precisely the opposite time profile. There is typically a surge of aid during the immediate post-conflict period, followed by a rapid decline after the first three years: aid tapers out just as it should be tapering in. Of course, there may be good reasons other than growth for aid to be high in the early post-conflict years. However, on the evidence, the present temporal pattern of aid is inappropriate to the problem of restoring post-conflict societies. The appropriate response is probably not to forgo the initial surge of aid. Politically, this is the best opportunity to get higher levels of aid commitment. Rather, donors should couple these large commitments with an understanding that the lag between commitment and disbursement will be much longer than in normal situations. Just as the relationship between aid, policy and growth can be investigated to determine the marginal impact of aid on growth, so it can be used to investigate the contribution of policy improvement to growth. Other than in post-conflict situations, the relationship between policy and growth is estimated to be: dG/dP = 1.017P + 0.127A

(3)

52 Thematic Issues

In post-conflict situations this becomes: dG/dP = 1.017P + 0.313A

(4)

An implication is that good policy is even more important for growth in post-conflict situations than in other contexts, and that the extent of this difference depends on the amount of aid received. Suppose that donors were to get aid inflows ‘right’ in the sense discussed above, so that they were approximately double in post-conflict situations to what they were in non-conflict situations – say 6 per cent of GDP (at purchasing power parity exchange rates) versus 3 per cent. In a non-conflict country a one-point improvement in the five-point policy scale used by the World Bank would raise the growth rate by around 1.5 percentage points. By contrast, in a post-conflict society, the same one-point improvement would raise the growth rate by around 3 percentage points. Thus, if donors get aid policy right in post-conflict societies, policy improvement will have double the effect on growth that it has in normal environments. The conjunction of both effective aid and policy in post-conflict situations usually leads to substantially above-average rates of growth during the first decade of peace, typically by a little over one percentage point. All this extra growth occurs from around the fourth year of peace through to the end of the first decade, with a peak around the sixth or seventh year. The policy scoring system of the World Bank is an average of four components of policy: macro, structural, social and governance. At the time of writing, twenty different aspects of these policies are considered. On the whole, good policies in one area tend to be associated with good policies in another, although analysis of the components suggests that many of these policies matter for the growth process. I now turn to the question of whether any of these policies are particularly important for growth in a post-conflict situation. That is, how might policy priorities differ if a society is post-conflict and growth is the objective? The World Bank policy index can be used to investigate this question. However, the governance component is too recent an addition to the index for it to be used in the analysis of growth. For the other three components – macro, structural and social – there are consistent data through the decade of the 1990s that can be used to investigate whether any component is differentially important in post-conflict situations. Collier and Hoeffler (2002c) find that of the three, social policies are differentially important in post-conflict societies, whereas macroeconomic policies are differentially less important. This does not mean that macroeconomic policies do not matter, or even that in some absolute sense they are less important than social policies. What the result shows is that, using as a benchmark the relative importance of the three policies in normal situations, in post-conflict societies social policies are more important and macro policies less important. Thus,


53

for example, it may be the case that in all situations macro policies are in some sense more important than other policies. The World Bank’s main indicators of social policies are the inclusiveness of education and health policies, and the provision of social protection. Superficially, it is surprising that these policies are important for growth in post-conflict situations. Evidently, the effect of education on growth is only long term, whereas the results suggest that social policies have an effect on growth within four years. What might be the channel by which social expenditures affect the growth rate? The most likely route is, perhaps, that inclusive social policies act as a signal to the population regarding the government’s intentions. If the government signals inclusion, then the population recognizes that peace is more likely to be maintained and is therefore more willing to invest. Recall that the restoration of confidence is likely to be extremely important in post-conflict situations. During a successful restoration, like that in Uganda, the investor risk rating has improved from only 5 on a scale of 0–100 in 1992 (at the time the lowest in all Africa) to 23 by 2002, at which time it was well above the African average. The move to universal primary education during the first peaceful and democratic presidency in the country’s history may have played a part in this remarkable recovery.

4

The political economy of policy reform

Post-conflict societies come out of conflict with significantly worse policies than the average for low-income countries. Using the World Bank’s policy scale, the average for all low-income countries during the 1990s was around 3.00, while the average for post-conflict countries during the first three years of peace was only 2.41. A one-point difference in the policy scale indicates a substantial difference in policies, so this 0.59 difference matters, although the fact that countries coming out of conflict have poor policies is not surprising. Should governments attempt rapid reform, or is the political situation so delicate that reform risks a resumption of conflict? Or should they accelerate reform because of its atypically important role in the growth process? Obviously, such delicate matters can only be decided by a careful analysis of the particular situation. However, the global pattern of policy reform post-conflict provides some useful guidance. First, I consider whether policy reform appears to increase the risk of conflict. Introducing both policy and the rate of change of policy into the previously discussed model of conflict risk, we find that neither the level of policy nor an improvement in policy have significant direct effects on the risk of conflict. All their important effects work through their impact on economic performance. As discussed above, policy improvement raises growth, and growth reduces the risk of conflict. Hence, policy reform should not, in general, be delayed from fear that it will lead to renewed conflict.

54 Thematic Issues

The actual behaviour of post-conflict governments bears this out. By around the end of the first decade of peace, policy improves dramatically to 3.05. Hence, during the first decade, those post-conflict governments that had managed to sustain peace had also managed to improve policies from the very poor initial level of 2.41 to being a little better than the average for low-income countries. This is a far more rapid rate of policy improvement than is found on average in non-conflict situations. Hence we can conclude that the first decade after conflict is politically a relatively easy time for a government to introduce substantial policy reforms. This is indeed consistent with theoretical analysis in political science, which suggests that the vested interests that normally block reform are weakened by periods of civil war. Further, since growth is typically more rapid in this period, this in itself may make reform easier. However, while overall policy improvement is impressive, the composition of that improvement shows no signs of being skewed towards social policies, which our previous analysis suggests should be differentially a priority. Rather, all three groups of policies tend to improve in step with each other. Finally, I turn from the level of the three broad policy groups to consideration of three specific policies that are likely to be differentially important in post-conflict settings. These are demobilization, the restoration of the rule of law, and the restoration of a formal economy. Demobilization is desirable in that, typically, the government and the rebel army will be merged, producing a very large army at a time when the need for a large army is reduced by the return to peace. However, demobilization can potentially pose a threat to both the security of ordinary households, if demobilized soldiers resort to crime, and to the security of the state if they become a recruitment pool for subsequent rebellions. With regard to the impact of demobilization on crime, there is some evidence from the effect of the Ugandan demobilization of the early 1990s (Collier, 1994). Despite widespread fears that this demobilization would lead to an escalation in crime, the detailed evidence suggested that it had no effect, with one exception: where soldiers were demobilized into districts where there was a severe land shortage, so that many of the soldiers lacked access to land, then crime in those districts escalated. Hence, it is probably important to ensure that demobilized soldiers have adequate access to land. On the threat to the security of the state there is no systematic evidence. While there may be some risk of recruitment into rebel organizations, the latter more typically target youth. Further, there may be more of a threat to the state from an excessively large regular army that is consequently badly paid and underemployed. Given the lack of evidence for a deterrence effect from military expenditure, there is on balance a case for early and substantial demobilization. The restoration of the rule of law typically involves early action on property rights, followed by a gradual improvement in the functioning of the courts and the police. Civil wars generally leave a legacy of confused property


55

rights as different vintages of powerful groups lay claim to the same property without a chain of legal transfer of title. In this environment there are often also opportunistic claims to property. While property rights are confused, investment in physical assets will be curtailed because investors cannot be sure that their titles will be upheld. A major success of the Ugandan recovery was the handling of these disputed claims. The key step was to establish a credible process for the adjudication of claims and the restitution of property, but one that was strictly time-limited. All claims had to be filed with the court by a set date to be considered, and decisions were rapid. By this means the Ugandan government restored the market in property quite rapidly while at the same time building a good image with the foreign investor community. The highly opportunistic environment typically following civil wars makes it both important and difficult to restore the courts. For example, until the banks are able to foreclose reliably on assets pledged as collateral, default rates are likely to be too high for the banking system to be able to lend profitably. Commonly, the police force will have been run down during the conflict. The rebuilding of an honest police force is important, both for the functioning of the courts and for the curtailment of opportunism. However, it is itself made difficult by the prevalence of opportunism. Police work offers many opportunities for corrupt income, and a police force can easily become part of the problem that it is designed to reduce. During a civil war the economy becomes less formal. This has consequences for the taxable base – most economic activity is not taxable. This is one reason why aid is so important in the early post-conflict years. Without aid, the government is faced with the choice between extremely low levels of expenditure and extremely high rates of taxation on the small part of the economy that is taxable. High tax rates are likely both to limit the growth of formal sector enterprises and to retard the repositioning of enterprises from the informal sector into the formal sector. Thus, in post-conflict societies, aid can legitimately be used to reduce the burden of taxation. During this phase the taxable base of the economy recovers so that when taxes are eventually imposed revenue will be sufficient to meet the public expenditure needs of the society. Premature attempts to tax the fragile and small formal economy risk delaying fiscal sustainability.

5

Conclusion

Post-conflict societies are distinctive. They require different policies for the maintenance of peace, which needs to be a much higher priority than in countries without a recent history of conflict. They require different donor policies and different government policies for the restoration of growth. To date, there is evidence that both donor and government policies have often been inefficient. Military expenditure has been too high, whereas aid has been too low. If this donor behaviour is based on a fear that aid will leak into

56 Thematic Issues

military expenditure, such fears seem on average to have been misplaced. Aid has tapered out too early in the post-conflict process. Governments have been correct to implement rapid reform, but these reforms may not have been sufficiently tailored to the special needs of post-conflict societies. References Azam, J.-P. (1995) ‘How to Pay for the Peace? A Theoretical Framework with References to African Countries’, Public Choice, vol. 83, issue 1–2, pp. 173–84. Collier, P. (1994) ‘Demobilization and Insecurity: A Study in the Economics of the Transition from War to Peace’, Journal of International Development, vol. 6, no. 3. Collier, P. and D. Dollar (2002) ‘Aid Allocation and Poverty Reduction’, European Economic Review, vol. 46, no. 8, pp. 1475–500. Collier, P. and A. Hoeffler (2002a) ‘Aid, Policy and Peace: Reducing the Risk of Conflict’, Defence and Peace Economics, vol. 13, no. 6, pp. 435–50. Collier, P. and A. Hoeffler (2002b) ‘Military Expenditure: Threats, Aid and Arms Races’, Development Research Group, World Bank, Washington, DC (mimeo). Collier, P. and A. Hoeffler (2002c) ‘Aid, Policy and Growth in Post-Conflict Societies’, Development Research Group, World Bank DC (mimeo). Grossman, H. I. (1992) ‘Foreign Aid and Insurrection’, Defence Economics, vol. 3, pp. 275–88.

4 Ethnicity, Institutions of Governance and Conflict Avoidance Mwangi S. Kimenyi

1

Introduction

Of the many challenges facing Africans, the establishment of durable peace is probably the most critical, yet the most difficult to achieve. According to some estimates, one-fifth of the people on the continent live in countries involved in conflict. It is now well established that conflicts in Africa have been a major hindrance to the improvement of the wellbeing of Africans (World Bank, 2000). In addition to costs such as loss of life, injuries and the general disruption of economic and social activities (Stewart, 2000; and Elbadawi and Ndungu, 2001), investment flows are damaged as a result of the increased risk premium. They have an impact on human development by disrupting investment in human capital and the provision of basic services, and increasing mortality rates. Civil wars are also associated with a prevalence of child labour and the disruption of social networks, and they weaken governments’ capacity to deliver crucial services. Conflicts are also associated with disease outbreaks that have a negative impact on the quality of life. In addition, conflicts involve collateral damage to neighbouring countries as a result of the inflow of refugees and guns. To the extent that conflicts disrupt investment in human capital, their impact on development tends to be long term because generations of young people miss education opportunities, as has been the case in both Angola and Sudan. Furthermore, conflicts have a major impact on investment flows because of increased risk premiums. The various development targets, such as those articulated in the Tokyo International Conference on African Development (TICAD-II), or the more recent Millennium African Renaissance Programme, are unlikely to be achieved unless there is a marked reduction in the incidence, severity and duration of civil wars. Post-independence conflicts have been relatively more intense and have entailed enormous costs; it is estimated that conflict lowers economic growth by two percentage points every year. Additionally, conflicts divert huge resources from development purposes. In Central and West Africa, such 57

58 Thematic Issues

resources have been estimated at US$1 billion and more than US$800 million a year, respectively (World Bank, 2000). These estimates exclude the indirect costs of conflict, such as costs for neighbouring countries generated by refugee flows, and costs arising from impeded communication routes and disruption in economic activity. Post-independence conflicts have generally been between different ethnic groups within a country. Virtually every African conflict has an ethno-regional dimension to it. Even conflicts that are apparently free of ethnic considerations involve factions and alliances built around ethnic lines (Deng, 1997). One of the first post-independence conflicts was the civil war in Nigeria. The conflict arose when the Ibo in the eastern region, fearing domination by other ethnic groups in the country, wanted to secede and form their own Biafran nation. This was prevented by the use of military force. In Rwanda and Burundi, where the Hutu constitute a numerical majority while the minority Tutsi dominate the countries’ economic and political life, interethnic conflicts have been reported as some of the most intense in the developing world; each ethnic group has attempted genocide aimed at the eradication of the other. In Ethiopia, ethnic conflict lasted for hundreds of years as the Eritrean people sought independence from Ethiopia. Similarly, members of the Somali ethnic group sought independence from the Amhara who have for years dominated Ethiopian politics and the economy. Other examples of interethnic conflict deteriorating into wars that have claimed the lives of hundred of thousands of people include Sudan, Mozambique, Somalia, Angola and the Democratic Republic of the Congo (DRC). Clearly, to understand conflicts in Africa necessarily requires an appreciation of ethnic dynamics. In particular, it is important to understand why ethnic groups engage in conflict, and which features of ethnicity make Africa prone to conflict. A number of causes have been advanced for the occurrence and persistence of conflicts. These include natural resource endowment, ethnic and religious hatreds, economic inequalities, lack of political rights and economic mismanagement of the state, among many others (Collier and Hoeffler, 1998 and 2001; Collier, Hoeffler and Söderbom, 1999). While it is a fact that ethnicity does play a part in triggering conflict, it has been established that it is not the mere existence of ethnic diversity that causes conflict but rather that much depends on the institutions of governance. Institutional arrangements play a critical part, not only in determining economic outcomes but also in establishing peace within nations. It is, for example, evident that the economic and political crises that characterize post-independent Africa are often an outcome of poorly-designed institutional frameworks. Specifically, political instability, civil strife and economic stagnation are to a large extent the products of institutions that are not capable of dealing with heterogeneous populations. The failure of political institutions to accommodate diverse interests (ethnic, religious, linguistic) has generated conflict situations that have an adverse

Kimenyi: Ethnicity and Conflict Avoidance 59

effect on political and economic outcomes. It is becoming increasingly clear that African countries are unlikely to achieve stability and economic growth in the absence of major institutional reform. In particular, the greatest challenge facing African countries is the establishment of institutional arrangements that can accommodate ethnic diversity effectively and facilitate the peaceful coexistence of population groups (Kimenyi and Mbaku, 1995). If institutions determine outcomes, then Africa’s problems should not be attributed merely to the existence of extremely heterogeneous population groups but must be seen as the result of institutions that are not well placed to harmonize ethnic claims and result in various undesirable outcomes. Thus the problem is not diversity but rather the existing institutional arrangements. This chapter suggests that the centralized unitary states of Africa have failed to deal effectively with diversity because the institutions have been unable to achieve cooperative agreements among the various groups. Rather, unitary centralized states tended to encourage ethnic competition for resources and power, which in turn led to conflict. The basic argument advanced in this chapter is that ethnic groups are important institutions that organize for the production of ‘ethnic goods’ in order to generate a surplus for their members. This surplus can be increased either through the creation of value (production and specialization) or by seeking transfers from other ethnic groups. Since there are substantial benefits that accrue from associating and identifying with an ethnic group, there is an interest among members in identifying with their group. The ability of an ethnic group to ‘force’ transfers from other groups is largely determined by the institutions of governance. Ethnic conflict is more likely to arise in institutions that make it easy for one group to improve its welfare by transferring surplus from other groups. This chapter contends that the optimal institutional arrangement in nations characterized by strong ethnic identification is one that allows for the establishment of governments based on local ethnic jurisdictions, each possessing a fair degree of autonomy, permitting groups to retain their surplus. In such institutions, groups can increase their wellbeing by improving efficiency of production or by engaging in mutually beneficial exchange with other groups. Section 2 discusses post-independence African institutions and policies that have tended to trigger conflicts. Section 3 outlines various features of ethnic groups in Africa and stresses that they are important institutions that can be used to organize collective activity. Section 4 introduces the idea of functional governments to complement ethnic governments, and Section 5 concludes.

2

Post-independence African institutions and ethnic conflicts

Probably the most important event explaining contemporary African institutions is European colonialism. Colonialists partitioned Africa among

60 Thematic Issues

themselves in a manner that reflected their spheres of influence, paying little regard to ethnic heterogeneity or hitherto existing institutions of governance. As a result, different ethnic groups, each with their own language, culture, religion, traditions, customs and political and economic systems, were brought together to form a single political and administrative unit. For example, in Nigeria, 400 ethnic groups were combined to form one country. The partition of Africa also resulted in situations whereby members of the same ethnic group were separated and placed in different countries. Members of the Somali community, for example, found their members dispersed between Somalia, Ethiopia and Kenya. The Yoruba and the Aja were divided between Nigeria, Benin and Togo; the Wolof and the Serers between Senegal and the Gambia. The political units that were created were not only arbitrary but also lacked any strong unifying factors. Partition disrupted long-established kingdoms and tribal governments, and disrupted existing trade networks and other inter-ethnic linkages. Furthermore, colonial rule did little to develop democratic institutions in Africa, the colonialists being interested mainly in maximizing economic gain. Members of certain tribes were favoured over others – a policy of divide and rule that weakened the opposition to colonial rule. The ethnic divisions thus created were manifested during the struggles for independence and continue to be important today. In fact, ethnic differences in levels of education and income that exist today are a legacy, in large part, of the colonial experience. The post-colonial African leaders sought to retain the unitary governments bequeathed by colonialists. They argued that for the various ethnic groups to coexist peacefully, it was necessary to homogenize otherwise heterogeneous groups. Unifying the groups into one national state was considered to be crucial for political stability. Additionally, the centralization of power allowed for the adoption of uniform policies and the balancing of economic resources. Similar arguments were advanced for the preference for singleparty political systems where all ethnic groups would be joined together in one party. Leaders warned that unity would be compromised if political party competition were to be introduced, because different political parties would be dominated by particular tribal and linguistic groups, essentially promoting tribalism. The leaders therefore sought ways for the diverse ethnic groups within each country to come together in a unitary state. The approaches to coexistence that were adopted involved suppressing expressions of ethnic identity, ostensibly to promote national unity and integration. Thus the common response to diversity in many African countries has been the adoption of institutional arrangements that seek to unify heterogeneous populations by limiting expressions of group differences (Kimenyi, 1997a). African leaders have therefore been eager to weaken the tribe as a unit of collective choice, as the tribe is considered to be harmful to peaceful coexistence. Determined to preserve national unity, leaders sought to establish unitary states with power concentrated at the centre. The result was the


creation of institutions where the head of state held the power over all public policies without constitutional constraints on the central authority in its exercise of power. As a result, in most African countries, the state wields immense power over the economy and the allocation of resources. Thus, for the most part, post-independence African institutions were characterized by one-party rule and an excessive centralization of power. The experiment with one-party centralized government has shown that such institutional arrangements are not suited to, or capable of, harmonizing the interests of heterogeneous groups. Since the state is highly centralized, the leader of the ethnic group that captures the state has control of enormous resources. To retain control of the instruments of wealth transfers, those in power have an interest in erecting barriers to entry into political and economic markets. There is ample evidence that many African governments have maintained their control over governance structures by redistributing resources in favour of groups such as the military, urban dwellers and specific ethnic groups. Instead of unifying the various ethnic groups, centralized unitary states have provided leaders with opportunities to extend benefits to some groups at the expense of others (Brough and Kimenyi, 1986; Kimenyi, 1989; Kimenyi and Shughart, 1989). Discriminatory allocation of resources creates dissatisfaction among groups that are not in the ruling coalition, which then becomes a primary cause of ethnic conflict. Thus ethnic conflicts on the continent can be viewed as a reflection of sub-optimal institutional arrangements that make it possible for governments to use discriminatory policies to transfer benefits from one group to another. The ethnic groups that are not part of the ruling coalition are excluded from full and effective participation in economic and political markets. Many turn to violence as a way to minimize further marginalization and to improve their participation in these markets. Given that control of the government has potentially large benefits for members of the ruling ethnic group, they have an interest in taking control of the leadership. Since there are many barriers that make it difficult to remove leaders through democratic processes, one common approach to attaining power has been through military coups. Conclusive evidence of the ethnic orientation of most military coups is revealed by changes in the composition of members of the cabinet and senior civil servants before and after the coup. Between 1960 and 1982 some forty of the forty-five independent countries in sub-Saharan Africa experienced a military coup an attempted coup or a plot to overthrow the government. In almost all cases military leaders awarded top government positions to members of their own ethnic groups. For example, before the 1966 coup in Central African Republic that deposed President Dacko (a member of the Baya group), no member of the Mbaka ethnic group was in the cabinet. After the coup, the composition of the cabinet changed so that 23 per cent of cabinet members were Mbaka.

62 Thematic Issues

Similarly, when Kwame Nkrumah was president of Ghana, 71 per cent of cabinet members were from Nkrumah’s Akan ethnic group. Following the coup organized by members of the Ga and Ewe ethnic groups, representation of Akan people in the cabinet dropped to 25 per cent, and that of the Ga increased from 7.7 per cent to 38 per cent (Kimenyi, 1997a). Such changes in the ethnic composition of the cabinet have also been evident after peaceful transitions of government (Kimenyi and Shughart, 1989). These examples demonstrate the notoriety of ethnic groups of imposing costs on nonmembers. The same ethnic unit that advances the wellbeing of its members can also undermine the liberty of others. In a centralized state, ethnic groups engage in conflict primarily because they are in a struggle to capture the state; they mobilize to control the instruments of government as part of the competition for resources. Control of governance structures allows the controlling group to redistribute income and wealth in favour of its members and supporters (Collier, 1999). Recently, there have been political and economic reforms in most African countries. These reforms have generally taken the form of opening up political markets to allow opposing political parties, thus replacing single-party and military dictatorships. There is little doubt that political party competition is likely to make political leadership more responsive to a wider cross-section of the population. However, the problems that plague the continent are unlikely to be resolved by such reforms. It is possible that political party competition might widen ethnic division unless it is accompanied by other far-reaching institutional reforms. This is because, under a multi-party system, issues take on an increasing ethnic character as power brokers prey on people’s ethnic background to drum up political support. The limited experience of political party competition in Africa provides an example. Political parties have more often than not turned out to be tribal factions that have no clear ideology or plan of action other than transferring benefits to members of their coalition. Following the introduction of multi-party politics in Kenya, political parties tended to form along ethnic lines: the ruling Kenya Africa National Union (KANU) became the party of the Rift Valley; the National Development Party (NDP) evolved as the party of the Luo/ Nyanza tribes; the FORD-Kenya became the party dominated by Luhyas in Western Kenya; while the Democratic Party (DP) and FORD Asili were largely dominated by the Kikuyus of Central Province (Murungi, 1995). The legalization of opposition parties has not resulted in truly competitive political markets that would otherwise serve as effective vehicles in the design of viable constitutional contracts guaranteeing protection of individual and group rights. In almost all countries that have recently entered the era of multi-partyism, the structures of governance have remained much as they were during the single-party era. As a result, African states continue to suffer from the single-party legacy of over-centralization, weakly developed civil societies and a lack of the independent institutions necessary for proper


governance. Clearly, conflicts among groups are not likely to be eliminated by political party competition alone and it is conceivable that such conflicts may be intensified by its introduction. In short, democratization as represented by political party competition is an important institutional reform but not sufficient for harmonizing ethnic interests and not necessarily conflictreducing. Thus alternative forms of institutional design are required.

3

Ethnic units of collective choice

The centralized African state has clearly been a failure. Our argument above is that marginal reforms, such as the introduction of political party competition, are unlikely to result in significant changes in terms of the peaceful coexistence of ethnic groups. Instead, we suggest that more radical reforms providing for greater decentralization of political power are necessary. Decentralization must necessarily involve the creation of autonomous ethnic governments. To appreciate the thrust of the argument advanced here, particularly why the ethnic unit is a suitable unit of collective choice, it is important first to appreciate the importance of the ethnic group as an institution as well as the complexity of ethnicity in Africa. One of the most important features of sub-Saharan Africa is the high degree of ethnic heterogeneity. Countries in this region are home to over 2,000 distinct ethnic groups, each having its own language, culture and traditions. Sudan has over 170 distinct linguistic groups, while in Zaire there are over 250 ethnic groups and in Nigeria between 200 and 400 distinct linguistic groups. The size of each tribe varies, some having as few as 100,000 members, while others have millions. One important feature of African societies that makes ethnic units suitable as units of collective choice is geoethnicity: in many cases each group associates with a particular territory. The various ethnic groups have what they refer to as their territory and in many respects members of other ethnic groups are considered to be outsiders. Although the boundaries that demarcate ethnic territories are not always clearly specified, they nevertheless exist and to a large extent are respected by the groups that live next to each other. For example, although the Democratic Republic of Congo (formerly Zaire) consists of over 250 ethnic groups, they can be grouped into 23 distinct units that are closely related in terms of culture and language. As is evident, each of the groups is confined to a clearly defined ethnic territory. Cobbah (1988, p. 73) observes that: In Africa, this ethnic identity is above all other things a territorial identity. Nothing defines the ethnic group better than its ‘standing place’. Thus the term geoethnicity has been used to describe the African ethnic phenomenon. Geoethnicity as opposed to non-territorial ethnic identification involves the historic identification of an ethnic group with a given

64 Thematic Issues

territory, an attachment to a particular place, a sense of place as a symbol of being and identity. Thus, by and large, ethnic units make up what could be called ‘ethnic nations’, which have organized and served their members for generations. Because they occupy a specific territory does not necessarily qualify ethnic groups as appropriate units of collective choice. However, associated with diversity is the fact that ethnic groups have through the years continued to remain distinct, with members identifying strongly with their own tribes. Unfortunately, tribal identification is usually viewed negatively, and African leaders and intellectuals have engaged in highly publicized efforts to denounce this identification as tribalism. Identifying with a particular tribe is considered to be regressive. Tribalism, many claim, has no place in the modernizing nations of Africa. At the national level, identifying with a tribal group is considered to be inconsistent with development. The truth, however, is that tribal attachment is alive and well throughout Africa. By and large, identifying with one’s tribe is highly valued, and resources are devoted to make certain that members of a group continue to identify with that group. Children are taught to identify and be proud of their tribe, ethnic identity is strongly felt, and behaviour based on ethnicity is normatively sanctioned (Horowitz, 1985). The result of strong ethnic identification is that most voluntary and business organizations form along ethnic lines. The majority of urban social welfare organizations, urban soccer leagues and even trade unions are organized along tribal lines. It has been observed that tribally-orientated associations rate among the most important voluntary associations in West Africa (Little, 1957). More telling evidence of the degree of ethnic identification in politics is revealed by voting patterns. No matter which country one looks at, it is clear that people vote very much along ethnic lines; political ideology in the Western sense rarely plays a significant role. The fact that we observe high degrees of tribal identification implies that the expected benefits of identifying with a tribal group exceed the costs of such identification. Because there are benefits associated with identifying with a particular tribe, the tribal leadership has an interest in investing in tribal identification. Kimenyi (1998) presents a simple model of tribal identification based on costs and benefits, as shown in Figure 4.1. The horizontal axis measures the degree to which individuals identify with their tribe, and the vertical axis measures the marginal costs and benefits of ethnic identification. While there are benefits from identifying with one’s tribe, they tend to decrease the more one identifies with it. This is represented by the demand (TI) for tribal identification. Notice that even if the costs of tribal identification were zero, it would not pay to identify completely with the tribe. MC represents the cost of identifying with one’s tribe. Point A is the optimal level of tribal identification (50 per cent in this case) and TAM is the tribal surplus.


T

Marginal cost and marginal benefit

C

A

M

0

50 Tribal identification %

80

100 I

Figure 4.1 The market for tribal identification Source: Kimenyi (1998).

This point is optimal because identifying more with one’s ethnic group beyond this point implies that the additional benefits are less than the costs associated with such identification. In essence, beyond point A, the benefits of associating with one’s tribe are less than those that would accrue from identifying with members of other tribes. The idea of maximizing tribal surplus assumes that the main focus is on group welfare and thus benefits are shared by all members. This does not imply that members share the benefits equally, but rather that the group is better off in general. Strong ties within ethnic groups make them ideally suited as units for organizing collective provision. This is because ethnicity economizes on organizational costs. This is consistent with the suggestion by Addison and Murshed in Chapter 1 of this volume, that the primary reason why many civil wars have an ethnic dimension is precisely because ethnicity is efficient at overcoming collective action problems. In particular, ethnic-based institutions have a comparative advantage in solving prisoner’s dilemma problems. It can be argued that the outcomes observed in centralized African states are the result of the fact that the costs of arriving at cooperative outcomes are very high, so that countries ultimately suffer Pareto-inferior non-cooperative

66 Thematic Issues

outcomes. Since members of ethnic groups have a long-term attachment to their own group (for example, through blood lines or past memories), they are more likely to have continuous dealings with members of their ethnic group than with members of other ethnic groups. Continuous dealing reduces cheating in prisoner’s dilemma situations and as a result ethnic groups may be more efficient than the state in the provision of public goods (Roback, 1991). The free-rider problem is likely to be less prevalent when the group is made up of a single ethnic group than when several ethnic groups are involved. If we define an optimal unit of collective choice as one that minimizes the costs of achieving cooperative agreements, we would then expect jurisdictions to vary in both size and composition. Since transaction costs associated with achieving cooperative agreements depend on the composition and size of the decision-making group, those groups composed of homogeneous members face lower transaction costs of achieving agreements than groups whose members are heterogeneous, ceteris paribus. The other important characteristic is that ethnic groups act as interest groups in competing for transfers, maximizing the wellbeing of their members by competing for wealth transfers from other groups. However, they are distinct from other interest groups such as labour unions or producer groups in that membership of an ethnic group is permanent, since entry and exit are not possible (Kimenyi, 1989). The interest group theory of government teaches that public policy outcomes reflect the interplay between demanders and suppliers of wealth transfers. Transfers may involve taxes, subsidies, regulations, public projects or even government jobs. In Africa, though many interest groups compete for wealth transfers, ethnicity represents the most important group. Kimenyi and Mbaku (1995) identify various African interest groups competing for wealth transfers, including farmers and urban dwellers, but they note that the most important characteristic of these groups in Africa is ethnicity. Permanent interest groups attempt to maximize their share of wealth transfers through the political process. To ensure that it maximizes its share of wealth transfers, a group needs to control the instruments of wealth transfers, primarily political leadership. It follows that it is more profitable for groups in African countries to invest to influence government policy than it is for them to invest their scarce resources in the market, where the returns are, in comparison, lower. Not surprisingly, the competition to control the instruments of wealth transfer has been intense in most African countries, and has been one of the underlying causes of ethnic conflict. Since governments do not create wealth, such transfers must be taken from other groups, which essentially implies that the transfer-seeking behaviour by permanent interest groups is at best a zero sum game. This analysis helps to explain the continued state of conflict and political instability in Africa, where there are so many tribal groups attempting to control governments, and why the interest-group behaviour of ethnic groups is more pronounced in centralized unitary states.


Efficient production increases the producer and consumer surplus, essentially a measure of an ethnic group’s wellbeing. Different ethnic groups within the same country are likely to be vastly different in terms of their ability to generate surplus, but each can increase its ‘ethnic’ or ‘tribal’ surplus by entering into mutually beneficial exchanges with other ethnic groups. Thus, under conditions whereby ethnic groups are able to organize production freely and to enter into market exchange with other groups if they are able to retain their tribal surplus, then each ethnic group would seek to be efficient in production in order to increase its surplus. However, efficiency in production is not the only way that an ethnic group can increase its surplus. One attractive alternative is to use the political process to capture the surplus generated by other groups through rent-seeking. Under a system of government where taxing and spending decisions are centralized, rulers are able to redistribute surplus from some ethnic groups to others. Given that ethnic groups occupy different regions, governments are able to direct the resources to specific groups. Such ethnic rent-seeking can be minimized under a system of decentralization. The discussion has so far emphasized that ethnic groups possess specific advantages in organizing for the provision of ‘ethnic’ goods. However, in unitary centralized states, ethnic groups can use their organizational capacity to undermine the wellbeing of others. As such, the first reforms that can help to reduce ethnic conflicts are those providing ethnic units with a fair degree of autonomy. Ideally, decentralized federal states would be best suited to deal with diversity. Of the various strategies aimed at resolving conflicts in societies characterized by diversity, federalism is the most prescribed system of governance. Decentralization has the potential to resolve the institutional problems affecting African countries, by increasing the efficiency and responsiveness of the government in two ways. First, compared to central government officials in distant offices, local leaders have superior information about local needs. Devolving resource allocation decisions to locally elected leaders therefore helps to improve the match between the preferences of local people and the government’s output. Second, local leaders are more accessible to local people and can be held more accountable. This results in improved management of resources through improvement in performance. Decentralization also encourages innovation in the delivery of services, as individuals have a greater incentive to participate in all aspects of community life and to seek solutions to individual and collective choices. When considering the institutional problems in Africa, however, one can argue that the primary advantage of decentralization is that it is probably the only system of organizing collective activities that protects groups from oppression by others as well as accommodating diversity. It provides mechanisms for bringing rival groups into a formal bargaining process (Burki, Perry and Dillinger, 1999), thus reducing conflicts. Decentralization

68 Thematic Issues

reduces the points of inter-ethnic contacts that result in conflict, and instead increases the contacts that are mutually beneficial. In such a decentralized system, each autonomous political jurisdiction has greater latitude in policy design and implementation, and can also determine the nature and scope of interaction with other groups. Under constitutional federalism, individuals can migrate freely and cheaply to competing political jurisdictions. This reduces significantly the ability of the state to coerce. Moreover, this ability of citizens to vote with their feet enhances competition among units and results in improved service delivery. While there is consensus about the advantages of decentralized states, the use of ethnic units of collective choice is controversial. We suggest that ethnic units have distinct efficiency characteristics that should be exploited in designing the units of collective choice. Given geoethnicity, effective decentralization naturally implies that the local governments will be composed primarily of ethnically homogenous groups. The use of ethnic groups as units of collective choice has been advanced by Kimenyi (1998). Kimenyi compares ethnic groups to voluntary associations or clubs whose members have fairly homogenous preferences. Using the model of clubs as developed by Buchanan (1965), the optimal size of the club is determined by the membership at which marginal cost is equal to marginal benefit (see Figure 4.2). In this model, costs increase and benefits decrease steadily as membership increases. Figures 4.3 and 4.4 extend the model to cases of ethnic units of collective choice. Based on previous discussion, the organization of collective activities

MB, MC MC

MB

N Number of club members Figure 4.2

Optimal size of clubs


MB, MC MC

MB

Members of same ethnic groups Figure 4.3

Members of other ethnic groups

Optimal ethnic club size: kinked marginal cost

by one ethnic group is expected to resemble the club models: costs increase and benefits decrease steadily as membership increases. This is because the ethnic group possesses various advantages that economize on organizational costs. However, increasing the membership beyond a certain level results in a situation where the benefits of organizing the activity are lower than the costs. This would imply that the benefits of ethnic provision are less than the costs (Landa, 1994). The benefits and costs change more dramatically when more than one ethnic group is involved. Organizing across different ethnic groups is more complicated, and adding members from other ethnic groups creates a discontinuity in the costs and benefits. These discontinuities result from difficulties in communicating across ethnic and linguistic boundaries as well as differences in preferences, taste and reduced trust across ethnic groups rather than within the same ethnic group. The implication of the foregoing analysis is that the optimal units of collective choice are ones that are ethnically homogeneous. Establishment of autonomous ethnic governments does not preclude collaboration across ethnic groups. However, this collaboration will only be acceptable if all

70 Thematic Issues

MB, MC MC

MB

Members of same ethnic groups Figure 4.4

Members of other ethnic groups

Optimal ethnic club size: kinked marginal benefit

groups benefit from the association. In essence, a system of ethnic governments would help to reduce ethnic rent-seeking that often translates into ethnic conflicts.

4

Overlapping functional and ethnic jurisdictions

The central theme of this chapter is that ethnic governments have desirable features that potentially can reduce conflict. The proposal, then, is for the establishment of decentralized federalism with autonomous ethnically homogenous local governments. However, this should not be seen as totally severing linkages between ethnic groups. Ethnic governments, while retaining a significant degree of autonomy, must necessarily associate with other ethnic governments. In particular, it is important to note that such governments will only be able to provide a limited number of ‘local’ public goods. Thus each unit has a desire to cooperate with other units in the provision of public goods that cannot be provided efficiently within the confines of the individual units. Such joint provision would promote cooperation across ethnic units based on mutual interest, discouraging conflicts and encouraging peaceful coexistence. In addition to the ethnic governments, various other complementary functional governments for the provision of various services are proposed (see Frey and Eichenberger, 1999). These governments would cover different regions, some transcending the boundaries of the ethnic governments. In


addition, in some cases, the functional governments would overlap because there are many goods that are best provided by more than one ethnic community. For example, the provision of water may require communities to join together to build a dam or to distribute water to their communities. Note that the communities cooperate out of self-interest. In essence, the establishment of ethnic governments can be an important cornerstone for defining the nature of cooperation among ethnic groups. Ethnic governments can be inefficient, particularly in cases where individuals do not have any mobility option and thus are not able to ‘vote with their feet’. This is quite often the case in Africa, where migration from one ethnic community to another is quite difficult. However, the establishment of overlapping functional governments such as for water government acts as a good substitute for ‘voting with one’s feet’ and thus enhances the efficient provision of services. The point to emphasize is that the establishment of ethnic governments does not preclude cooperation between ethnic groups. Instead, such institutions form the basis for groups to enter into mutually beneficial arrangements in the provision of public goods.

5

Conclusion

This chapter highlights the importance of establishing institutions that harmonize ethnic interests in Africa. It is argued that the recent democratization reforms involving the introduction of political competition are not enough to deal with the issue of conflict avoidance. In fact, it is argued that political party competition can intensify conflict-generating ethnic competition in the current highly centralized unitary states. In this chapter I have proposed the establishment of decentralized institutions of governance that involve ethnically homogeneous local units of collective choice possessing a fair degree of autonomy. In addition, the chapter suggests that ethnic governments might be complemented by functional governments that transcend the boundaries of ethnic governments within a given country. The chapter does not attempt to discuss institutional issues relating to conflicts between countries or wars of secession. These topics are discussed in Kimenyi 1997b, 1997c and 1999. References Borner, S. and M. Paldam (1998) The Political Dimension of Economic Growth, IEA Conference vol. 119 (London: Macmillan). Brough, W. T. and M. S. Kimenyi (1986) ‘On the Inefficient Extraction of Rents by Dictators’, Public Choice, vol. 48, pp. 37–48. Buchanan, J. M. (1965) ‘An Economic Theory of Clubs’, Economica, vol. 32, pp. S1–4. Burki, S., G. Perry and W. Dillinger (1999) Beyond the Centre: Decentralizing the State (Washington, DC: World Bank).

72 Thematic Issues Cobbah, J. A. M. (1988) ‘Toward a Geography of Peace in Africa: Redefining Sub State Self-Determination Rights’ in R. J. Johnson, D. B. Knight and E. Kofman (eds), Nationalism, Self-Determination, and Political Geography (London: Croom Helm). Collier, P. (1999) ‘Doing Well Out of War’, Paper prepared for the Conference on Economic Agendas in Civil Wars, London, 26–27 April. Collier, P. and A. Hoeffler (1998) ‘On Economic Causes of Civil War’, Oxford Economic Papers, vol. 50, pp. 563–73. Collier, P. and A. Hoeffler (2001) Greed and Grievance in Civil War (Washington, DC: World Bank). Collier, P., A. Hoeffler and M. Söderbom (1999) On the Duration of Civil War (Washington, DC: World Bank). Deng, F. M. (1997) ‘Ethnicity – An African Predicament’, Brookings Review, vol. 15, no. 3, pp. 28–31 (Washington, DC: The Brookings Institution). Elbadawi, I. A. and N. S. Ndungu (2001) ‘The Economics of Civil Wars and PostConflict Recovery’, Paper presented at the Development Economic Research Group Conference, University of California, May (mimeo). Frey, B. and R. Eichenberger (1999) ‘A New Proposal for Federalism and Democracy in Developing Countries’, in M. S. Kimenyi and J. M. Mbaku (eds), Institutions and Collective Choice in Developing Countries (Aldershot: Ashgate). Horowitz, D. (1985) Ethnic Groups in Conflict (Berkeley, Calif.: University of California Press). Kimenyi, M. S. (1989) ‘Interest Groups, Transfer Seeking and Democratization: Competition for the Benefits of Government Power May Explain African Political Instability’, American Journal of Economics and Sociology, vol. 48, pp. 339–49. Kimenyi, M. S. (1997a) Ethnic Diversity, Liberty and the State: The African Dilemma, The Shaftesbury Papers No. 12 (Cheltenham: Edward Elgar). Kimenyi, M. S. (1997b) ‘Redrawing Africa’s Borders: Logical Foundations for Consensual Units of Collective Choice’, Department of Economics, University of Connecticut. (mimeo). Kimenyi, M. S. (1997c) ‘Ethnic Rent-Seeking and Optimal Integration’, Department of Economics, University of Connecticut (mimeo). Kimenyi, M. S. (1998) ‘Harmonizing Ethnic Claims in Africa. A Proposal for Ethnic Based Federalism’, Cato Journal, vol. 18, no.1, pp. 43–63. Kimenyi, M. S. (1999) ‘Spatial Competition, Ethnicity, and the Optimal Size and Composition of Units of Collective Choice’, in M. S. Kimenyi and J. M. Mbaku (eds) Institutions and Collective Choice in Developing Countries (Aldershot: Ashgate). Kimenyi, M. S. (2001) ‘Institution of Governance and Ethnic Conflict in Africa: A Positive View of Ethnic Governments’, Development Economic Research Group Conference, University of California, May 2001 (mimeo). Kimenyi, M. S. and J. M. Mbaku (1995) ‘Democratization in Africa: The Continuing Struggle’, Coexistence, vol. 32, pp. 119–36. Kimenyi, M. S. and W. F. Shughart (1989) ‘Political Successions and the Growth of Government’, Public Choice, vol. 62, pp. 173–9. Landa, J. T. (1994) Trust, Ethnicity and Identity (Ann Arbor, Mich.: University of Michigan Press). Little, K. (1957) ‘The Role of Voluntary Associations in West African Urbanization’, in P. L. Van den Berghe (ed.), Africa: Social Problems of Change and Conflict (San Francisco: Chandler).

Kimenyi: Ethnicity and Conflict Avoidance 73 Murungi, K. (1995) Ethnicity and Multi-Partysim in Kenya (Nairobi: Kenya Human Rights Commission). Roback, J. (1991) ‘Plural but Equal: Group Identity and Voluntary Integration’, Social Philosophy and Policy, vol. 8, no. 2, pp. 60–80. Stewart, F. (2000) ‘Civil Wars in Sub-Saharan Africa: Counting the Economic and Social Cost’ in D. Ghai (ed.) Renewing Social and Economic Progress in Africa (London: Macmillan). World Bank (2000) Can Africa Claim the 21st Century? (Washington, DC: World Bank).


Part II Case Studies: West Africa


5 Liberia and Sierra Leone: Interwoven Civil Wars Victor A. B. Davies

1

Introduction

Liberia and Sierra Leone have common borders and interwoven civil war experiences. Rebels from the ongoing Liberian civil war ignited the 1991–2001 Sierra Leone war. Combatants subsequently moved between them with ease. This chapter addresses the following questions. What factors have caused and sustained war in the two countries? Why in 2002 was Liberia still at war? What lessons can be learned from Sierra Leone’s peace process? What are the prospects for peace for the two countries?

2

Context and background

2.1 Liberia The Liberian civil war began in December 1989 when Charles Taylor’s National Patriotic Front of Liberia (NPFL) invaded Liberia from Côte d’Ivoire (See Map 5.1). Within months, the rebels were on the outskirts of Monrovia, the national capital. In August 1990, the Economic Community of West African States (ECOWAS) despatched to Liberia a Nigerian-led intervention force known as the Economic Community of West Africa Monitoring Group (ECOMOG). Using bases in Sierra Leone, ECOMOG prevented the NPFL’s imminent capture of Monrovia, marking the beginning of a long military stand-off between ECOMOG, which held Monrovia and the surrounding regions, and the NPFL, which held the remaining 90 per cent of Liberia. In 1997, after nearly ten failed peace accords and cease-fire agreements, and several interim governments, elections were conducted under the 1996 Abuja Peace Accord. Charles Taylor won and was inaugurated in August 1997. However, once in office, Taylor flagrantly violated the Abuja Peace Accord by refusing to allow ECOMOG to restructure the national army to absorb combatants from the warring factions and promote ethnic balance. This caused ECOMOG’s premature withdrawal in 1998. Taylor became increasingly 77

78 Case Studies: West Africa

15° w

5° w

10° w

15° N SENEGAL

Dakar

GAMBIA MALI

Banjul

Bamako

GUINEA BISSAU Bissau GUINEA 10° N Conakry SIERRA LEONE

Freetown Atlantic Ocean

CÔTE D’IVOIRE LIBERIA Monrovia

5° N

0 0

15° w

Map 5.1

100

200 miles

Abidjan

100 200 300 km

10° w

5° w

Liberia and Sierra Leone in West Africa

repressive, also supporting rebels in Sierra Leone. By early 1998, small rebel groups were again active in Liberia. The Liberians United for Reconciliation and Democracy (LURD) appeared in July 2000, and, although participating in the peace agreement of August 2003, failed to disarm as required. The conflict has claimed over 150,000 lives and displaced more than 1 million people out of a pre-war 1998 population of 2.4 million. It has been factious, spawning ten ethnically-based warring factions in addition to the dominant NPFL. 2.2 Sierra Leone In March 1991, Sierra Leonean and Liberian rebels fighting for the NPFL in Liberia attacked south-eastern Sierra Leone, triggering the Sierra Leone civil

Davies: Liberia and Sierra Leone

79

war. Their leader, Foday Sankoh, announced the intention of the Revolutionary United Front (RUF) to overthrow the kleptocratic All People’s Congress (APC) government. ECOMOG, based in Sierra Leone and Liberia, supported the government in the war against the RUF. In 1992, young soldiers from the war front overthrew the APC, setting up the National Provisional Ruling Council (NPRC). The war escalated under the NPRC, coming to within twenty miles of Freetown, the national capital, in 1995. The NPRC organized elections in 1996 and handed over to the victorious Tejan Kabbah and his Sierra Leone People’s Party (SLPP). Kabbah signed the Abidjan Peace Accord with the RUF in 1996 but was overthrown by the army in collaboration with the RUF in May 1997. ECOMOG intervened to reinstate Kabbah in February 1998, driving the rebels to the countryside. Kabbah was almost overthrown again in January 1999 when the rebels captured parts of the capital before ECOMOG repulsed them again. The Lomé Peace Accord was signed in July 1999 which, though frequently jeopardized, provided a framework for ending the war, with the completion of the disarmament and demobilization of combatants in January 2002. In May 2002, general elections were held, and won by the incumbent President Kabbah and his SLPP. The RUF failed to win a single parliamentary seat but, together with other contesting parties, accepted the election results. The conflict claimed over 25,000 lives and displaced 2.1 million people, nearly half of the country’s population. The RUF had about 20,000 combatants, and pro-government forces of over 25,000, in addition to ECOMOG. Both sides used mercenaries. The rebels committed widespread atrocities, including mutilations, rape, arson and looting. This was partly an RUF military strategy, otherwise lacking an ideology to motivate and guide its campaign. The atrocities bonded combatants to the rebel organization and subjugated civilians. Recruits were drawn largely from among socially-deprived youths and children, often forcibly, and made to commit acts of terror against family and community members, ensuring exclusion from their community so they would subsequently not desert. The rebels were also drugged to induce fearlessness and to dehumanize them. Civilians and new recruits were used as human shields, positioning them in the firing line, with hard-core rebels at the rear. Both sides used child soldiers, totalling about 10,000.

3

Factors causing and sustaining civil war

3.1 Causes Modern Liberia emerged when freed slaves from the USA established settlements around the coast of Monrovia from the early 1820s. These AmericoLiberians declared independence in 1847. Constituting less than 5 per cent of the population, the settlers monopolized power, disenfranchising indigenous people. In 1931, the League of Nations established that Americo-Liberians


were using indigenous Africans as forced labour, an act of slavery. Enduring ethnic aristocracy foreclosed all peaceful means for the majority indigenous population to access power, leading to Samuel Doe’s coup of 1980. During the 1980s, Doe’s policies resulted in the most immediate causes of the civil war. Doe used ethnicity as a survival strategy, practising nepotism for his minority Krahn ethnic group, which constituted only 2 per cent of the population. He executed powerful non-Krahn fellow coup-plotters and members of the ruling People’s Redemption Council, notably Thomas Weh Sien, his second in command, and Thomas Quiwonkpa. Consequently, Doe alienated his executed colleagues’ ethnic groups, particularly Quiwonkpa’s Gios. In 1985, Doe’s security forces massacred hundreds of Gios in Nimba County. That same year, Doe stage-managed presidential elections to become the elected president. Corruption and mismanagement led to declining economic performance. Per capita income growth fell from 0.5 per cent during 1965–80 to minus 5.2 per cent during 1980–7. Thus, when Charles Taylor launched his rebellion in Nimba County in 1989, he gained widespread support among Liberians, particularly Gios and Manos from Nimba County, enabling the rebellion to spread rapidly. Diasporas, mostly USA-based, featured prominently in the conflict, some allegedly financing the NPFL, and expecting power in return. Like Liberia, Sierra Leone’s modern period began when its present capital, Freetown, was chosen as a home for freed slaves from Europe and America. However, unlike the situation in Liberia, the settlers never ruled Sierra Leone. Although ethnic differences emerged between the settlers and indigenous ethnic groups, after independence in 1961 these rivalries were superseded by ethnic rivalries among the indigenous population. The domestic causes of war in Sierra Leone can be traced to the APC government of 1968–92. The APC practised kleptocracy and political repression as a survival strategy through a patrimonial system of asset-stripping, including the allocation of access rights to natural resources; rents from pricing interventions; income redistribution measures such as exchange rate overvaluation; and taxation of rural agriculture to subsidize urban consumption. These measures were inimical to economic growth and development. The APC suppressed or bought out any opposition. Consequently, by the late 1980s, under a one-party dictatorship, Sierra Leone was heading for economic and political implosion. Living standards were among the lowest in the world, as reflected in a rank of 174th out of 174 on the UNDP Human Development Index for 1991. Basic amenities such as electricity and running water had virtually collapsed even in the national capital. Per capita income growth was negative. Youths, hardest hit by unemployment, bore the brunt of the economic and political atrophy of the 1980s and became the instruments of choice for political violence under the APC. Drugged and used as thugs during elections, and abandoned subsequently, they became increasingly rebellious, retreating


81

into crime and drug-taking. By the late 1980s, Sierra Leone’s unemployed youths constituted a large pool of potential rebels. University students spearheading the anti-APC campaign recruited volunteers from this pool for revolutionary training in Libya. External support helps to explain why war broke out in Sierra Leone and Liberia rather than other African countries with similar risk factors. Taylor and Foday Sankoh received training and military support from Libya as part of Colonel Gaddafi’s grand plan to install satellite regimes in West Africa. Côte d’Ivoire and Burkina Faso also supported the rebellions. Côte d’Ivoire hosted Foday Sankoh and offered Charles Taylor a base; Taylor, in turn, offered Sankoh a base and troops to launch civil war in Sierra Leone. Presidents Houphouët-Boigny of Côte d’Ivoire and Blaise Campoaré of Burkina Faso apparently supported Charles Taylor for personal reasons: in the coup of 1980, Doe had executed Liberian President Tolbert’s son who, like Blaise Compoaré, was Houphouët-Boigny’s son-in-law (Reno, 1998, p. 81). While there is no obvious reason why Côte d’Ivoire and Burkina Faso supported the RUF, such support may have been obtained through Charles Taylor. 3.2 Factors sustaining civil war The conjunction of abundant natural resources and the porosity of national borders has been a major factor sustaining rebellion in the two countries. Pre-war legal and illicit exports of natural resources were equivalent to more than 30 per cent of GDP in Liberia, and some 25–40 per cent in Sierra Leone. Timber and rubber resources helped to finance Liberia’s civil war; diamonds financed the rebellion in Sierra Leone and produced a warprolonging congruence of interests (Davies, 2000). Exports of conflict diamonds for 1999 were estimated at US$68.5 million, while government defence expenditure during the war averaged less than US$25 million a year. Pro-government combatants also mined diamonds, sometimes in collaboration with the rebels. The illicit trade of conflict diamonds to Europe via Liberia partly explains why Taylor, at the centre of a lucrative arms-for-diamonds trade, chose the path of warlordism, even after winning the 1997 presidential elections. Table 5.1 shows that Liberia ‘exported’ diamonds far in excess of domestic output. Maximum annual production capacity is approximately 150,000 carats, while recorded imports of ‘Liberian’ diamonds in Belgium exceeded Table 5.1

Illicit diamond traffic 1990–8 (in 000s carats)

Year

1990

1991

1992

1993

1994

1995

1996

1997

1998

Production Exports

100 5,523

100 658

150 1,909

150 5,006

100 3,268

150 10,677

150 12,320

150 5,803

150 2,558

Source:

Smillie, Gberie and Hazleton (2000).


12 million carats in 1996. A report by a UN panel of experts states that the value of Belgian imports of ‘Liberian’ diamonds was US$217 million for 1998, while official exports recorded in Liberia were valued at only US$800,000 (United Nations, 2000, p. 10). Ethnic rivalries sustain civil war in both countries, more strongly in Liberia because of the social and political divide between Americo-Liberians and the indigenous population, and Samuel Doe’s ethnically-divisive rule in the 1980s. Gios and Manos from Nimba County whose ethnic groups were massacred by Doe’s army harbour deep hatred for the Krahns, Doe’s ethnic group. Such divides tend to break down national unity and make it easy to recruit within ethnic groups by appealing to ethnic emotions. There has been no enduring social divide among the population in Sierra Leone. Intermarriage, migration and other forms of socialization have gradually blurred the Creole-Protectorate divide. Although the APC did not engage in systematic ethnic massacres, ethnic undercurrents did affect the war. The APC, the first government that fought the rebellion, was often accused of using the war to disenfranchise the south-eastern regions where the war started. People from these regions initially joined the rebel movement voluntarily out of disillusionment with the APC. Northerners distrusted the war effort of the south-eastern based SLPP, especially with regard to the prominent role of the ethnic Mende Kamajor civil defence militias. Both conflicts created a substantial supply of small arms and a generation of young men knowledgeable in the use of them. The risks of cross-border contagion were high; and combatants redeployed across countries with relative ease. The RUF field commander, General Mosquito, allegedly commanded some of Taylor’s troops. Worse still, in 2000, an insurgency in Guinea, which shares borders with both Liberia and Sierra Leone, was aided by RUF rebels and, allegedly, the Liberian government. Also, it is relatively easy to acquire small arms: West Africa is awash with weapons despite an ECOWAS moratorium on shipments to the region (United Nations, 2000, p. 1). In Liberia, deep mistrust among the protagonists makes it difficult for them to cooperate, which is one reason for the failure of past peace accords and interim governments. Peace is constrained by a severe time-consistency problem, but a major factor for disruption was removed by Taylor’s flight into exile in August 2003. The two countries’ geography favours warlordism. Much of the hinterland is heavily forested, allowing rebels to operate with ease. Unlike Sierra Leone, with a relatively high population density of 65 people per square kilometre, Liberia has a low population density of 26 people per square kilometre, which makes it more difficult to defend. The rural population density is much lower in both countries. Continued external support of the rebels, particularly from Libya, and Taylor’s support for the RUF, helped to sustain the two civil wars. On the other hand, apathy by the international community has been detrimental.


83

The belated international efforts in Sierra Leone demonstrate the potency of such support. Apart from ECOWAS, it was not until February 2004 that a UN donors’ conference put together a US$ 500 million aid package for Liberia.

4

The consequences of civil war

4.1 Liberia Liberia’s pre-war per capita income of US$485 in 1987 was higher than the SSA average of US$440 and US$237 in Sierra Leone (see Table 5.2). The 1990 Human Development Index rank of 26/130 on an ascending scale of human achievement put Liberia above twenty other African states, with 70 per cent of the labour force engaged in agriculture and forestry. Leading exports included iron ore, rubber and timber. Diamonds and gold were exported in small quantities. The effects of war have been devastating: total official annual exports averaging over US$400 million before the war plummeted to less than US$60 million in the period 1997–2000. Much of Liberia’s infrastructure was destroyed; and at the time of writing there is still no running water or electricity supply. The war has claimed over 150,000 lives (equivalent to more than 6 per cent of the population) and experienced massive emigration, especially of skilled human capital. Table 5.2 shows that per capita income fell from US$485 in 1987 to US$177 in 2000. In the 1990s employment decreased by 50 per cent in the formal sector, which accounted for about 30 per cent of total employment in the mid-1980s. Alluvial diamond mining, which employed some 60,000 Liberians prior to the war, employed only 6,000 in 2001, following UN sanctions on Liberian diamonds (United Nations, 2001). Support for Sierra Leone’s rebel RUF led to Liberia’s international isolation. After May 2001 UN sanctions banned President Taylor and his government from international travel. The export of rough diamonds was

Table 5.2

Socio-economic indicators in Liberia and Sierra Leone Period

Per capita income (US$)

Per capita income growth (%)

Incidence of poverty (%) (

Post-Conflict Economies in Africa (International Economic Association)

Structural Reform and Economic Policy (International Economic Association)

Institutional Change and Economic Behaviour (International Economic Association)

Differential Information Economies (Studies in Economic Theory)

Keynes's General Theory After Seventy Years (International Economic Association)

Documents in International Economic Law

Africa and Asia in Comparative Economic Perspective

Economic reform in sub-Saharan Africa

Latin American Economic Crises: Trade and Labour (International Economic Association Conference Volumes)

Institutions, Human Development and Economic Growth in Transition Economies

Advances in Macroeconomic Theory (International Economic Association Conference Volume No. 133)

Redefining Sovereignty in International Economic Law

Conditions for Partnership in International Economic Management

Research Handbook in International Economic Law

In Africa

Local Economies in Turmoil (International Political Economy Series)

International Finance and the Developing Economies

Dynamics, Economic Growth, and International Trade (Studies in International Economics)

Africa. An Encyclopedia for Students. Abidjan - Economic

Political and Economic Dictionary of Africa, A

Foreign Aid, Self-Reliance, and Economic Development in West Africa:

Economic Recovery In Africa: The Paradox of Financial Flows

Economic Diversification in Africa: A Review of Selected Countries

Economic Recovery In Africa: The Paradox of Financial Flows

Association Mapping in Plants

Assessing Regional Integration in Africa: Rationalizing Regional Economic Communities

Economic Crisis in Africa: Perspectives on Policy Responses

Economic Development in Africa: Debt Sustainability - Oasis or Mirage?

Post-Conflict Economies in Africa (International Economic Association)

Structural Reform and Economic Policy (International Economic Association)

Institutional Change and Economic Behaviour (International Economic Association)

Differential Information Economies (Studies in Economic Theory)

Keynes's General Theory After Seventy Years (International Economic Association)

Documents in International Economic Law

Africa and Asia in Comparative Economic Perspective

Economic reform in sub-Saharan Africa

Latin American Economic Crises: Trade and Labour (International Economic Association Conference Volumes)

Institutions, Human Development and Economic Growth in Transition Economies

Advances in Macroeconomic Theory (International Economic Association Conference Volume No. 133)

Redefining Sovereignty in International Economic Law

Conditions for Partnership in International Economic Management

Research Handbook in International Economic Law

In Africa

Local Economies in Turmoil (International Political Economy Series)

International Finance and the Developing Economies

Dynamics, Economic Growth, and International Trade (Studies in International Economics)

Africa. An Encyclopedia for Students. Abidjan - Economic

Political and Economic Dictionary of Africa, A

Foreign Aid, Self-Reliance, and Economic Development in West Africa:

Economic Recovery In Africa: The Paradox of Financial Flows

Economic Diversification in Africa: A Review of Selected Countries

Economic Recovery In Africa: The Paradox of Financial Flows

Association Mapping in Plants

Assessing Regional Integration in Africa: Rationalizing Regional Economic Communities

Economic Crisis in Africa: Perspectives on Policy Responses

Economic Development in Africa: Debt Sustainability - Oasis or Mirage?

Recommend Documents