(Conflict, Inequality and Ethnicity) Yoichi Mine, Frances Stewart, Sakiko Fukuda-Parr, Thandika Mkandawire (eds.) - Preventing Violent Conflict in Africa_ Inequalities, Perceptions and Institutions-Pa

Preventing Violent Conflict in Africa
Conflict, Inequality and Ethnicity

Series Editor: Frances Stewart, Emeritus Professor of Development Economics and Director,
CRISE, University of Oxford, UK
The series focuses on the relationships between inequality, human security and ethnicity
emanating from the work of The Centre for Research on Inequality, Human Security and
Ethnicity (CRISE), UK. Some books cover general issues such as mobilization, affirmative
action, the role and management of natural resources, and post conflict policies; others are
rooted in the experience of different regions, for example, South East Asia, West Africa and
Latin America.
The series is unique since it is united by interest in an unusual perspective – that of exploring
these issues in relation to horizontal inequalities or inequalities among groups. While each
book stands alone, they also have unity of approach and share some basic ideas. The pub-
lications from the Conflict, Inequality and Ethnicity series will be highly relevant to students,
academics and policy makers working on questions of contemporary conflict.
Titles include:
Graham Brown, Arnim Langer and Frances Stewart (editors)

POLICY AND DIVERSITY
Experiences and Challenges of Affirmative Action in an International Context
Yvan Guichaoua (editor)
UNDERSTANDING COLLECTIVE POLITICAL VIOLENCE
Arnim Langer, Frances Stewart and Rajesh Venugopal (editors)
HORIZONTAL INEQUALITIES AND POST-CONFLICT RESOLUTION
Yoichi Mine, Frances Stewart, Sakiko Fukuda-Parr and Thandika Mkandawire (editors)
PREVENTING VIOLENT CONFLICT IN AFRICA
Inequalities, Perceptions and Institutions
Frances Stewart (editor)
HORIZONTAL INEQUALITIES AND CONFLICT
Understanding Group Violence in Multiethnic Societies
Rosemary Thorp, Stefania Battistelli, Yvan Guichaoua, Maritza Paredes and Jose
Carlos Orihuela
THE DEVELOPMENTAL CHALLENGES OF MINING AND OIL
Lessons from Africa and Latin America
Rosemary Thorp, Maritza Paredes
ETHNICITY AND THE PERSISTENCE OF INEQUALITY
The Case of Peru
Conflict, Inequality and Ethnicity

Series Standing Order ISBN 978–0–230–24608–9 (hardback) and
978–0–230–24708–6 (paperback)
(outside North America only)
You can receive future titles in this series as they are published by placing a standing order.
Please contact your bookseller or, in case of difficulty, write to us at the address below with
your name and address, the title of the series and the ISBN quoted above.
Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke,
Hampshire RG21 6XS, England
Preventing Violent Conflict
in Africa
Inequalities, Perceptions and Institutions
Edited by
Yoichi Mine
Professor, Graduate School of Global Studies, Doshisha University, Japan
Frances Stewart
Emeritus Professor of Development Economics and Director, Centre for Research on
Inequality, Human Security and Ethnicity (CRISE), University of Oxford, UK
Sakiko Fukuda-Parr
Professor, The New School, USA
and
Thandika Mkandawire
Professor, London School of Economics and Political Science, UK
Foreword by Sadako Ogata

Editorial matter and selection © Yoichi Mine, Frances Stewart, Sakiko
Fukuda-Parr and Thandika Mkandawire 2013
Individual chapters © Respective authors 2013
Foreword © Sadako Ogata 2013
Softcover reprint of the hardcover 1st edition 2013 978-1-137-32969-1
All rights reserved. No reproduction, copy or transmission of this
publication may be made without written permission.
No portion 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,
Saffron House, 6–10 Kirby Street, London EC1N 8TS.
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 2013 by
PALGRAVE MACMILLAN
Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited,
registered in England, company number 785998, of Houndmills, Basingstoke,
Hampshire RG21 6XS.
Palgrave Macmillan in the US is a division of St Martin’s Press LLC,
175 Fifth Avenue, New York, NY 10010.
Palgrave Macmillan is the global academic imprint of the above companies
and has companies and representatives throughout the world.
Palgrave® and Macmillan® are registered trademarks in the United States,
the United Kingdom, Europe and other countries.
ISBN 978-1-349-46068-7 ISBN 978-1-137-32970-7 (eBook)
DOI 10.1057/9781137329707
This book is printed on paper suitable for recycling and made from fully
managed and sustained forest sources. Logging, pulping and manufacturing
processes are expected to conform to the environmental regulations of the
country of origin.
A catalogue record for this book is available from the British Library.
A catalog record for this book is available from the Library of Congress.
Contents
List of Figures and Tables viii

Foreword by Sadako Ogata xii
Preface xiv
Acknowledgements xv
Notes on Contributors xvi
1 Introduction: Disentangling the Linkages between

Horizontal Inequalities and Political Institutions 1
Sakiko Fukuda-Parr, Arnim Langer and Yoichi Mine
1.1 Introduction 1
1.2 HIs and violent conflict 3
1.3 HIs, perceptions and political institutions:
A framework of analysis 5
1.4 Structure of this book 7
2 Comparing Political Institutions: Institutional Choice and

Conflict Prevention in Africa 10
Yoichi Mine, Mari Katayanagi and Satoru Mikami
2.1 Introduction 10
2.2 Defining political institutions 11
2.3 Elements of PD and PC institutions 14
2.4 Mapping political institutions in Africa 17
2.5 Plural pathways towards sustainable democracy? 33
2.6 Conclusion 37
3 ‘Twin Countries’ with Contrasting Institutions:

Post-Conflict State-Building in Rwanda and Burundi 40
Shinichi Takeuchi
3.1 Introduction 40
3.2 Configuration of identity groups 42
3.3 Horizontal inequalities 45
3.4 Post-conflict institutional choices and their backgrounds 48
v
vi Contents
3.5 Institutions and political power: How do

the institutions work? 52
3.6 Popular perceptions 55
3.7 Conclusion: Implications for conflict prevention 57
4 Horizontal Inequalities, Ethnic Politics and Violent

Conflict: The Contrasting Experiences of Ghana and
Côte d’Ivoire 66
Arnim Langer
4.1 Introduction 66
4.2 Côte d’Ivoire: From Ivorian miracle to civil war 69
4.3 Ghana: Living with diversity 81
4.4 Some conclusions 91
5 Beyond Ad hoc Power-Sharing: Comparing South Africa

and Zimbabwe 95
Yoichi Mine
5.1 Introduction 95
5.2 Historical evolution of HIs in South Africa and
Zimbabwe 96
5.3 Contemporary political transitions in South Africa
and Zimbabwe 105
5.4 Popular perceptions in South Africa and Zimbabwe 115
5.5 Conclusion: Institutional choice and development 121
6 Is Ethnic Autonomy Compatible with a Unitary State? The

Case of Uganda and Tanzania 126
Yuichi Sasaoka and Julius E. Nyang’oro
6.1 Introduction 126
6.2 Formation of identity groups in Uganda and Tanzania 127
6.3 Socioeconomic horizontal inequalities 134
6.4 Political processes and the choice of political institutions 138
6.5 What interview surveys reveal 144
6.6 Conclusion 150
7 The Politics of Identity, Horizontal Inequalities and

Conflict in Kenya 153
Mwangi S. Kimenyi
7.2 Configuration of Kenyan society 154
7.3 Horizontal inequalities 156
7.4 Politics: Processes, institutions and policies 162
7.5 The 2008 post-election violence 166
7.6 Perceptions of ethnicity and inequality in Kenya 168
Contents vii
7.7 The new constitution and prospects

for preventing future conflicts 175
8 Managing Horizontal Inequalities and Violent Conflicts in
Nigeria 178
Ukoha Ukiwo
8.2 Configurations of identity 179
8.3 Horizontal inequalities between identity groups 182
8.4 Political process and institutional choice 192
8.5 How political institutions and policies work 194
8.6 Popular perceptions 203
8.7 Concluding remarks 205
9 The Relationship between Objective and Subjective
Horizontal Inequalities: Evidence from Five African
Countries 208
Arnim Langer and Satoru Mikami
9.2 Why objective and subjective inequalities may differ 210
9.3 Assessing objective socioeconomic horizontal
inequalities 212
9.4 Explaining the observed objective horizontal inequalities 215
9.5 Ethnic differences in risk factors associated
with individual socioeconomic development 226
9.6 Objective versus subjective horizontal inequalities 230
9.7 Some conclusions 245
Appendix 1: Operationalisation 246
Appendix 2: Descriptive statistics 249
JICA Survey 249
Afrobarometer 250
10 Findings and Implications: The Role of Development
Cooperation 252
Frances Stewart, Thandika Mkandawire and Mari Katayanagi
10.2 Chapter summaries 253
10.3 General findings 260
10.4 Conclusion 272
References 275
Author Index 295
Subject Index 299
Figures and Tables
Figures
1.1 Horizontal inequalities, political institutions and perceptions 7

1.2 Ten African countries chosen for case studies 8
2.1 Distribution of PD/PC institutions 22
2.2 Configurations of four major institutional categories 22
2.3 Population size and degrees of decentralisation 23
2.4 Ethnic fractionalisation and degrees of decentralisation 24
2.5 Colonial factors 25
2.6 Shifting positions of the countries in the case studies 29
2.7 Ibrahim Index of African Governance 30
2.8 Freedom House scores (excluding countries recently faced
with violent conflict) 32
2.9 Three paths of democratisation 34
5.1 Homelands of South Africa 98
5.2 Racial shares of total income 100
5.3 Racial breakdown of the South African population 100
5.4 Communal lands of Zimbabwe 102
5.5 Provincial HDI in South Africa and Zimbabwe, 2001 105
5.6 Regional distribution of cabinet posts in Zimbabwe 119
6.1 Uganda and Tanzania 132
7.1 Map of major ethnic groups by province 155
7.2 Distribution of population in the highest/lowest wealth
quintile by province (% within province) 159
7.3 Educational attainment by province 160
7.4 Under-five mortality by province 160
7.5 Access to piped water/public tap, sanitation and electricity
by province 161
7.6 Ownership of refrigerator and TV by province 161
7.7 Ethnic versus national identity (N = 907) 169
7.8 Would object to marriage with a specific ethnic group
(N = 321) 169
7.9 Would feel uncomfortable working with a specific ethnic
group (N = 253) 170
7.10 Trust of different groups (N = 907) 171
7.11 Perceptions of people from other ethnic groups (N = 907) 172
7.12 Perceptions of future inter-ethnic relations (N = 907) 172
7.13 Ethnicity affects a person’s chances of getting access
to . . . (N = 907) 173
viii
List of Figures and Tables ix
7.14 Has ethnicity become more important in Kenyan politics

since the introduction of multi-party elections in 1992?
(N = 907) 173
7.15 Perceptions about today’s socioeconomic conditions of own
ethnic group compared to the past (N = 907) 174
7.16 Which ethnic group do you feel is most favoured by the
government? (N = 907) 174
8.1 Political map of Nigeria, showing states and geopolitical
zones 181
9.1 Ethnic inequalities according to the BHN-index and the
Assets-index 213
9.2 Group averages of benefit from infrastructural development 228
9.3 Group averages of probability to have lower educational
attainment 229
9.4 Perceived socioeconomic horizontal inequalities 230
9.5 Perceived political horizontal inequalities (PPHIs) 242
9.6 Correlations between Perceived Political HI (PPHI) and
Perceived Socioeconomic HI (PSEHI) in Nigeria 244
Tables
2.1 Ratings of eight variables in sub-Saharan African countries 19

2.2 Change of ratings in the case-study countries 27
3.1 The ethnic power-sharing mechanisms defined in the
Burundian 2005 Constitution 51
3.2 Ethnic profile in Burundi’s National Assembly 55
4.1 Ethnic composition of the Ivorian population,
1975–1998 (%) 70
4.2 Countries of origin of the foreign population in Côte d’Ivoire 71
4.3 Religious composition of Côte d’Ivoire in 1998 (%) 71
4.4 Religion of different ethnic groups in Côte d’Ivoire in
1998 (%) 72
4.5 Income per capita in 1965 and 1975 (constant 1965 CFA
franc) 74
4.6 Ethnic composition of government, 1980–2003 79
4.7 Gross value added, per capita and per square mile, by region,
1960 88
4.8 Poverty and infant mortality rates, 1998–2006 89
4.9 Relative representation of ethnic groups in cabinet,
1954–2005 90
5.1 Home language in South Africa 97
5.2 Gini coefficients by race and place 101
5.3 Shifts in formal political systems in South Africa and
Zimbabwe 106
x List of Figures and Tables
5.4 National Assembly election results in South Africa 108

5.5 Parliamentary election results in Zimbabwe, 2000 and 2005 113
5.6 Parliamentary election results in Zimbabwe, 2008 114
5.7 National and ethnic identification of Zimbabwean city
dwellers (%) 120
6.1 Ethnic groups in Uganda (%) 128
6.2 Major ethnic groups in Tanzania 130
6.3 Major ethnic (racial) groups in Zanzibar 132
6.4 National and regional poverty rates in Uganda,
1999/2000–2009/2010 (%) 135
6.5 Household members with shoes (%) 135
6.6 Primary school net enrolment rates by region (%) 136
6.7 Poverty headcount in mainland Tanzania (%) 136
6.8 Primary school net enrolment rates (%) 137
6.9 National and ethnic identification of Ugandans (%) 149
6.10 National and ethnic identification of Tanzanians (%) 149
6.11 Multidimensional HIs and power-dispersing institutions 151
7.1 Ethnic composition of Kenya (% of total population) 154
7.2 Geo-ethnic composition 156
7.3 Cabinet representation by ethnic group 157
7.4 Political transition in Kenya 158
7.5 Compared to other ethnic groups, do you think your ethnic
group currently gets its fair share? (% of group) 175
8.1 Evolution of Nigeria’s state structure 180
8.2 Heads of three arms of government by geopolitical zone,
1960–2010 183
8.3 Composition of federal cabinets by zone, 1984–2010 (%) 184
8.4 Distribution of Senate seats, 2008 185
8.5 Distribution of seats in the House of Representatives, 2008 185
8.6 Composition of the federal civil service, 2000 186
8.7 National Assembly staff, December 2007 186
8.8 School enrolment by zone, 2008 187
8.9 Literacy status of ages 6 and above by zone, 2006 188
8.10 Nigerian university applications and admissions, 2007–2008
session 188
8.11 Educational backgrounds of members of state legislatures,
1999–2007 188
8.12 Incidence of poverty by zone, 1980–2004 189
8.13 Human development summary statistics by zone 190
8.14 Measures of standard of living by zone, 2008 (% within zone) 190
8.15 Distribution of population by wealth quintiles by zone, 2008
(% within zone) 190
8.16 Ownership of health care facilities by zone, 2004 191
8.17 Distribution of Senate seats by political party 196
List of Figures and Tables xi
8.18 Distribution of federal constituency seats by political party 196

8.19 Three most important elements of identity (% within group) 204
8.20 Feelings about ethnic and national identity (%) 204
9.1 Overview of survey locations and number of interviews 209
9.2 Determinants of living standards 217
9.3 Determinants of perceived socioeconomic HIs 234
10.1 Horizontal inequalities and stability 263
10.2 Modalities of national and national–local power
configurations 267
10.3 Categories of institutions and stability 269
Foreword
While I served as the UN High Commissioner for Refugees (UNHCR) through

the 1990s, the world witnessed horrendous violence breaking out in many
parts of sub-Saharan Africa. Civil wars ravaged through Liberia, Sierra Leone,
Somalia, Sudan, the Democratic Republic of the Congo and other countries.
Donors had continued to give aid to the Rwandan government until just
before it resorted to genocide in 1994. In many of these cases, firearms pro-
liferated, countless innocent civilians lost their lives and frightened people
crossed the borders in droves, with the refugee population in Africa reaching
the peak of nearly 7 million in the middle of the 1990s.
After leaving UNHCR in 2000, I co-chaired the Commission on Human
Security with Amartya Sen, the Nobel laureate in economic sciences, and
then assumed the presidency of the Japan International Cooperation Agency
(JICA) in 2003. During the period of my presidency, JICA started to put pri-
ority on issues confronting contemporary Africa. Although many African
countries were getting on the track of economic growth and political stabil-
isation, the risks of conflict were still rife as demonstrated in Kenya, Côte
d’Ivoire and Zimbabwe. Currently, development practitioners in JICA and
other agencies are trying hard to firmly integrate the agenda of conflict pre-
vention into their daily activities. It is imperative to activate these efforts
with consistent policy frameworks.
This book has its origin in the policy conference titled ‘Conflict Prevention
and Development Cooperation in Africa’, co-hosted by JICA and the United
Nations Development Programme (UNDP) at Wilton Park, UK, in Novem-
ber 2007. Professors Sakiko Fukuda-Parr and Robert Picciotto played leading
roles in organising this conference. In its keynote speech, I underlined the
importance of the human security concept. My message was also explicit
about the necessity on the part of development agencies to understand his-
torical backgrounds of violent conflicts. We must address the root causes of
violence, such as deep-seated poverty, inequality and social exclusion, and
pay proper attention to political processes in which those factors turn into
violence, when triggered by sudden downturns.
In the following year, we directed the JICA Research Institute (JICA-RI)
to produce cutting-edge research outcomes on global issues and promoted
active cooperation between scholars and practitioners of development. I am
delighted to see that the intensive discussion on conflict prevention ini-
tiated at Wilton Park in 2007 was carried into the research agenda at
JICA-RI under the dynamic directorship of Professor Keiichi Tsunekawa, and
reached fruition in the form of this publication. It was also fascinating
xii
Foreword xiii
for me to watch the close collaboration between political scientists and

development economists emerge through the process of research.
This research project combines three perspectives: multiple inequalities
between groups, people’s subjective perceptions of such inequalities and
modalities of political participation. This is an ambitious project that aims at
capturing the structural factors of violent conflicts and the political aspects
of development in African contexts, thereby contributing to practical pol-
icy recommendations. The method of the comparative study of country
pairs, as well as the historical analysis of constitutional engineering, added
great value to the case studies. To strengthen the quality of research, the
research team conducted original perception surveys in seven African coun-
tries. As any effort towards development cooperation and peace-building
should be based on real feelings of ordinary people in the field, I highly
commend the richness of the grassroots evidence contained in this book.
The conclusion framed in the last chapter of this book endorses the sig-
nificance of addressing material inequalities, as well as fear and hostilities
between groups. It also emphasises the relevance of assisting conflict-torn
nations to achieve inclusive governance and democratic power-sharing,
thereby preventing African countries from being trapped in vicious circles
of violence. Indeed, conflict prevention is the realisation of human security
which combines freedom from want, freedom from fear and freedom to live
in dignity.
Finally, I would like to express my deepest gratitude to three distinguished
scholars, Professor Frances Stewart, Professor Sakiko Fukuda-Parr and Profes-
sor Thandika Mkandawire, who have given invaluable advice in the research
process. I am also grateful to Professor Yoichi Mine, leader of the Japanese
research team. I sincerely hope that this book will be widely read and con-
sulted by policymakers, researchers, development practitioners and students
in Africa and worldwide. I expect that the policy implications of this research
will be applied substantially in the daily operations of development agencies,
including JICA.
Sadako Ogata
Former UN High Commissioner for Refugees
Former President of Japan International Cooperation Agency
Preface
This book is the result of close research collaboration between the Centre for
Research on Inequality, Human Security and Ethnicity (CRISE), University of
Oxford, and the JICA Research Institute (JICA-RI). We thank all scholars and
development practitioners at CRISE and JICA who have provided invaluable
academic and practical insights into this project.
We are very grateful for the continuous encouragement given by Madam
Sadako Ogata, who has contributed a gracious Foreword to this book,
Dr. Keiichi Tsunekawa, the founding director of JICA-RI, and Mr. Hiroshi
Kato, its present director. JICA-RI has organised several international confer-
ences and seminars in Tokyo and London to discuss research plans, progress
and outcomes, as well as to disseminate our findings. The discussion in
this book takes the lively feedback from the audience at those events into
account.
The perceptions surveys in seven African countries were made possible
by the active engagement of local contracting agencies: Datadesk Con-
sulting Services (South Africa), Mass Public Opinion Institute (Zimbabwe),
Synovate Tanzania (Tanzania), the Institute of Development Studies, Uni-
versity of Nairobi (Kenya), Makerere University (Uganda), Practical Sampling
International (Nigeria) and the Institute of Statistical, Social and Economic
Research, University of Ghana (Ghana). The surveys were conducted in par-
ticipatory ways; questions were modified and improved at various levels
reflecting local realities. We thank each agency as well as all enumerators
who knocked on the doors from house to house.
Finally, we would like to express our cordial gratitude to Dr. Mari
Katayanagi for her wonderful work in coordinating these efforts.
Yoichi Mine
Frances Stewart
Sakiko Fukuda-Parr
and
Thandika Mkandawire
xiv
Acknowledgements
Table 4.6 is reprinted with permission of Palgrave Macmillan from Langer,

A. 2008. Horizontal Inequalities and Violent Conflict: A Comparative Study
of Ghana and Côte d’Ivoire. In Stewart, F., ed., Horizontal Inequalities and
Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke:
Palgrave Macmillan.
Table 4.7 is reprinted with permission of Taylor & Francis from
Szereszewski, R. 1966. Regional Aspects of the Structure of the Economy.
In Birmingham, W.B., Neustadt, I. and Omaboe, E.N., eds, A Structure of
Contemporary Ghana. London: George Allen and Unwin.
Figure 5.2 is reprinted with permission of the Department of Economics,
University of Stellenbosch, from Van der Berg, S., Louw, M. and du Toit,
L. 2009. Poverty Trends Since the Transition: What We Know. Stellenbosch
Economic Working Papers 19(9). Matieland: Department of Economics, Uni-
versity of Stellenbosch.
Table 5.2 is reprinted with permission of OECD from Leibbrandt, M.,
Woolard, I., Finn, A. and Argent, J. 2010. Trends in South African Income
Distribution and Poverty since the Fall of Apartheid. OECD Social, Employ-
ment and Migration Working Papers, No. 101. Paris: OECD Publishing.
Tables 5.5 and 5.6 are reprinted with permission of Zimbabwe Election
Network from Zimbabwe Election Support Network. 2009. Possible Electoral
Systems for Zimbabwe. Harare: Zimbabwe Election Support Network.
Tables 7.1 and 7.2 are adapted with permission of Society for International
Development from Kanyinga, K. 2007. Governance Institutions and Inequal-
ity. In Society for International Development (SID), Readings on Inequality in
Kenya. Nairobi: SID.
Table 7.4 is reprinted with permission of Springer Science & Business
Media from Kimenyi, M.S. and Shughart, W.F. 1989. Political Successions
and the Growth of Government. Public Choice 62(2): 173–9.
Tables 8.3 and 8.6 are adapted with permission of the United Nations
Research Institute for Social Development (UNRISD) from Mustapha, A.R.
2006. Ethnic Structure, Inequality and Governance of the Public Sector in
Nigeria. Democracy, Governance and Human Rights Programme 24. Geneva:
UNRISD.
Every effort has been made to trace all copyright holders and the editors
would be happy to correct any errors or omissions.
xv
Contributors
Sakiko Fukuda-Parr is Professor of International Affairs at the New School,

New York. She is a development economist who has worked on a broad range
of issues of international development, including global poverty, gender,
technology, violent conflict and human rights. She is currently working on
the Millennium Development Goals (MDGs), human rights measurement
and the right to food. She was a research fellow at the Kennedy School
of Government and was lead author and director of the UNDP Human
Development Reports from 1995 to 2004. She serves on the boards of the
International Association for Feminist Economics, Centre for Economic and
Social Rights and Knowledge Ecology International.
Mari Katayanagi is Research Fellow at JICA Research Institute, Tokyo. She

served the Office of the High Representative in Bosnia and Herzegovina as a
political advisor. Her earlier positions include Senior Advisor and First Secre-
tary of the Embassy of Japan in Bosnia and Herzegovina and Special Advisor
for JICA. Her research interests are peace-building, conflict prevention and
international law, particularly international human rights law. She is the
author of Human Rights Functions of United Nations Peacekeeping Operations
(2002).
Mwangi S. Kimenyi is Senior Fellow and Director of the Africa Growth

Initiative in the Global Economy and Development programme of the
Brookings Institution in Washington, DC. He is the Founding Executive
Director of the Kenya Institute for Public Policy Research and Analy-
sis (KIPPRA) in Nairobi and has been a faculty member of the Depart-
ment of Economics at the University of Mississippi and the University of
Connecticut. He has also served as a member of the Board of Directors of
Equity Bank, Kenya.
Arnim Langer is Director of the Centre for Research on Peace and Devel-
opment (CRPD) and University Lecturer in International Relations at the
University of Leuven, Belgium. His research areas include causes and con-
sequences of conflict, post-conflict economic reconstruction, post-conflict
integration and nation-building, multidimensional group inequalities,
group behaviour and identity formation. His latest works include the
following co-edited books: Horizontal Inequalities and Post-Conflict Devel-
opment (2011), Affirmative Action in Plural Societies: International Experi-
ences (2012) and The Elgar Handbook of Civil Wars and Fragile States
(2012).
xvi
Notes on Contributors xvii
Satoru Mikami is Research Fellow at JICA Research Institute. He was Assis-

tant Professor, Faculty of Political Science and Economics and Research
Assistant, Institute for Research in Contemporary Political and Economic
Affairs, both at Waseda University. His research focuses on mass political
behaviour and attitudes, cross-national research of political institutions and
democratic theory.
Yoichi Mine is Professor at Graduate School of Global Studies, Doshisha

University, and has headed the research project on ‘Prevention of Violent
Conflicts in Africa’ as Visiting Fellow at JICA Research Institute. His research
interests include human security, development economics and African area
study. He is Secretary General of the Japan Association for Human Security
Studies.
Thandika Mkandawire is a former director of the United Nations Research

Institute for Social Development and was the first person to take on the posi-
tion of Chair in African Development at the London School of Economics
and Political Science (LSE). He was formerly Director of the Council for
the Development of Social Science Research in Africa (CODESRIA), a senior
research fellow at the Centre for Development Research in Copenhagen and
has taught at the Universities of Stockholm and Zimbabwe. His research
interests are mostly in development theory, economic policy and develop-
ment and social policy in developing countries and political economy of
development in Africa.
Julius E. Nyang’oro taught African politics and development at the Uni-

versity of North Carolina until his retirement in June 2012. His publications
include JK: A Political Biography of Jakaya Mrisho Kikwete, President of Tanzania
and Capitalist Development in Africa. His current research is in the areas of
African security, governance and development.
Yuichi Sasaoka is Professor of International Relations at the Graduate

School of Governance Studies, Meiji University. Prior to his current position,
he was Senior Research Fellow at JICA Research Institute. He has worked
on numerous development programmes in sub-Saharan Africa, Asia and
Latin America. He has also served as an advisor for the Institute for Interna-
tional Cooperation. His research interest lies in the linkage between global
governance and local governance.
Frances Stewart is Emeritus Professor of Development Economics, Emeri-

tus Fellow of Somerville College and Director of the Centre for Research
on Inequality, Human Security and Ethnicity (CRISE) at the University of
Oxford. Her research has focused on appropriate technology; basic needs;
adjustment and poverty; development during conflict; group behaviour;
xviii Notes on Contributors
horizontal inequalities; and human development. Books include, as author,

Technology and Underdevelopment (1976) and Planning to Meet Basic Needs
(1985); as co-author, UNICEF’s influential study Adjustment with a Human
Face (1987) and War and Underdevelopment (2001); as editor, Horizontal
Inequalities and Conflict: Understanding Group Violence in Multiethnic Societies
(2008).
Shinichi Takeuchi is Senior Research Fellow and Director of African Studies

Group at the Institute of Developing Economies–JETRO, Japan. He is also
Visiting Research Fellow at JICA Research Institute. Having specialised in
political economy of Central African countries, his current interest lies in the
process of state-building in the area, with particular focus on land problems.
Ukoha Ukiwo is a senior lecturer at the Department of Political and

Administrative Studies, University of Port Harcourt, Nigeria. He earned his
doctorate from St Cross College, Oxford University. He has held visiting
scholar positions at the Institute of International Studies, University of
California, Berkeley and Department of Political Science and International
Relations, Catholic University of Leuven, Belgium. His research interests
include democratisation, conflict and the political economy of development
in Africa.
1
Introduction: Disentangling the
Linkages between Horizontal
Inequalities and Political
Institutions
Sakiko Fukuda-Parr, Arnim Langer and Yoichi Mine
1.1 Introduction
In the period following the end of the Cold War, violent conflicts have
marred the African continent; since 1980, more than half of the countries
of sub-Saharan Africa1 have experienced armed conflict, sometimes multiple
conflicts taking place simultaneously in different parts of the country and
sometimes lasting for decades (Fukuda-Parr et al., 2008). Over the last decade,
the number of active violent intra-state conflicts has declined from 17 in
2001 to 10 in 2010 according to the Uppsala Conflict Data Program (UCDP).2
Despite this modest progress, political unrest and violent conflicts continue
to wreak havoc and delay (or even reverse) development in countries, includ-
ing, for example, Nigeria, Sudan, the Democratic Republic of the Congo
(DRC), Guinea and Mali.3 Most contemporary violent conflicts are identity-
based and have been labelled ‘ethnic’ (Stewart and Brown, 2007). They tend
to have a history of conflict recurrence, which is hardly surprising given
that countries emerging from conflict typically remain ‘very fragile in the
short-to-medium term, with a high risk of return to conflict or of its trans-
formation into other forms of social violence (such as crime or domestic
violence)’ (Langer et al., eds, 2011: 3).4
Yet all of this is a historical anomaly. Africa has largely been a continent of
peace. As Julius Nyerere commented, the ‘surprising thing is not that there
has been so much political instability in Africa but that there has been so
much stability, although this fact is less publicized internationally’ (Nyerere,
2000: xiv). Indeed, while the history of East Asia and Europe are charac-
terised by successive wars, violent revolutions and aggressive state-building
1
2 Introduction
(Tilly, 1992; Wong, 1997), most African states and empires emerged peace-
fully and vanished quietly in pre-colonial times (Herbst, 2000; Connah,
2001; Iliffe, 2007).
Violent conflicts are not inevitable. Thanks to an increasing amount of
research on the causes of intra-state violent conflicts, in particular by polit-
ical scientists and economists, over the last two decades, we now have a
much better understanding of the origins and risk factors contributing to
the emergence of these conflicts, and the ways in which they can be pre-
vented. An important challenge for most African governments and their
development partners is to design and implement development interven-
tions and policies that enhance the prospects for durable peace instead of
aggravating the risks of violent conflict. It is for this reason that the Japan
International Cooperation Agency (JICA) has commissioned this research
project on horizontal inequalities (HIs), focusing in particular on the role of
political institutions and people’s subjective perceptions of their situation.
The aim of this study is to contribute to an understanding of the forms of
political and economic governance that would assist in the prevention of
political violence that can undermine or reverse development.
This study explores the relationship between group inequalities and polit-
ical institutions. It focuses in particular on the role and interlinkages among
three factors: subjective or perceived inequalities among identity groups,
objective inequalities and the participation of groups in political institu-
tions. A major finding of recent research on sources of violent conflicts
is that one of the most important underlying causes is the presence of
HIs or group-based inequalities in access to political power and economic
resources. Severe HIs can provoke feelings of frustration and discontent that
may instigate violent group mobilisation and conflict along ethnic, religious
or regional lines (Stewart, 2000, 2002). While much has been learned about
the nature of HIs, two critical issues that remain unclear are how HIs are
perceived and what are the main influences on people’s perceptions, and the
way in which political institutions affect the relationship between HIs and the
emergence of violent conflicts. In Section 1.3 we will discuss these issues in
more detail.
The findings of the study have important policy implications in relation to
the strengthening of economic and political governance and the prevention
of violent conflicts. One objective of the project is therefore to translate any
new insights into actionable policy recommendations. It should be noted
in this respect that these policy recommendations are relevant not only for
‘conflict’ countries but also for other countries that want to address their
objective and subjective HIs constructively, thereby reducing the risk that
they may become a cause for group mobilisation, violent or otherwise, at a
later stage. Before discussing our framework of analysis, we will look more
closely at the concept of HIs and the evidence linking them to the emergence
of violent conflicts.
Sakiko Fukuda-Parr, Arnim Langer and Yoichi Mine 3
1.2 HIs and violent conflict
As mentioned in the previous section, research has shown that the presence
of severe HIs or inequalities between ‘culturally defined’ groups such as
ethnic, religious or regional groups in terms of access to political power and
economic resources is often an important contributing factor or cause for
the emergence of violent intra-state conflicts.5 In contrast to the relative
deprivation theory (see Gurr, 1970), the HI hypothesis points to the fact
that it can be the relatively rich, and not just the relatively deprived, who
initiate conflict (Stewart, 2002). One should not assume that it is always
the socioeconomically disadvantaged groups that cause political instability
and initiate violent conflict. While this appears to have been the case in
many disputes and conflicts around the world, sometimes relatively privi-
leged groups have attacked and violently suppressed underprivileged groups
in order to safeguard their political and economic power.
The concept of HIs and its theorised relationship with political instability
and violent conflict was first developed by the Oxford-based development
economist Frances Stewart in the early 2000s and was subsequently exten-
sively researched by the Centre for Research on Inequality, Human Security
and Ethnicity (CRISE) based at the University of Oxford.6 The concept of
HIs differs from the ‘normal’ definition of inequality, which Stewart (2000)
has termed ‘vertical inequality’, because the latter type lines individuals or
households up vertically and measures inequality over the range of indi-
viduals rather than groups. HIs, in contrast, are inequalities among groups
of people that share a common identity (for example, ethnic, religious and
regional identities). The HI literature broadly takes a constructivist approach
to ethnicity, which considers ethnic groups to be constructed socially and
historically according to various identity markers, such as ancestry, language
and region of origin. Although the constructivist view on ethnicity conjec-
tures that ethnic boundaries are malleable and fluid, it also assumes that
the reshaping of group identities is costly and that group boundaries there-
fore tend to have some continuity over time (see Stewart, 2002; Bates, 2004;
Stewart ed., 2008). And ‘it is because of [these] continuities, which go along
with the limited choices most people have to switch identities, that inequal-
ities among groups [can] become a source of unhappiness and resentment,
and a cause of social instability’ (Stewart, 2002: 8).
HIs are conceived of as inherently multidimensional and encompass
economic, social, cultural status and political dimensions:
• Economic HIs include inequalities in access to and ownership of assets –

financial, human, natural resource-based and social. In addition, they
comprise inequalities in income levels and employment opportuni-
ties, which depend on such assets and the general conditions of the
economy.
4 Introduction
• Social HIs include inequalities in access to a range of services, such as

education, health care and housing, as well as in educational and health
status.
• Political HIs include inequalities in the distribution of political opportu-
nities and power among groups, including control over the presidency,
the cabinet, parliamentary assemblies, the bureaucracy, local and regional
governments, the army and the police. They also encompass inequali-
ties in people’s capabilities to participate politically and to express their
needs.
• Cultural status HIs include disparities in the recognition and standing of
different groups’ languages, customs, norms and practices.7
The relevance of any element depends on whether or not it is an important

source of income or well-being in a particular society. For instance, the dis-
tribution of housing (an important source of discord between Catholics and
Protestants in Northern Ireland in the 1970s) is likely to be more relevant
in an industrialised country than in a country where people build their own
homes.
Evidence in support for the relationship between HIs and violent con-
flict is drawn from both case studies and econometric studies. Case studies
research has been conducted in a range of countries, including, among
others, Côte d’Ivoire (Langer, 2005), Burundi (Ndikumana, 2005), Kenya
(Muhula, 2009; Stewart, 2010a), Peru (Paredes, 2007), Ghana (Langer, 2008),
Nigeria (Langer and Ukiwo, 2008), Malaysia (Brown, 2005) and Guatemala
(Thorp et al., 2006). An important conclusion that can be drawn from these
case studies is that violent conflicts are more likely to arise in regions or
countries where economic, social, political and cultural status inequalities
occur simultaneously and where some groups are deprived across every
dimension (see, in particular, Langer, 2005, 2008; Langer and Ukiwo, 2008).
As argued by Langer (2005), the simultaneous presence of severe political
HIs and socioeconomic HIs forms an extremely explosive sociopolitical sit-
uation, because in these situations not only do the excluded political elites
have strong incentives to mobilise their supporters for violent conflict along
ethnic lines but they are also likely to find it relatively easy to gain support
among their ethnic constituencies.
Despite important data limitations, a substantial amount of quantita-
tive cross-country and within-country research has found a significant link
between the presence of HIs and the emergence of violent conflicts (see,
for example, Murshed and Gates, 2005; Brown, 2008; Mancini, 2008; Østby,
2008; Cederman et al., 2011). Thus, for instance, Østby’s (2008) analysis
of 55 countries from 1986 to 2003 shows that countries with severe eco-
nomic and social HIs have a significant higher probability of experiencing
violent intra-state conflict. Similarly, Cederman et al. (2010: 24) show that
‘groups with wealth levels far from the country average are indeed more
likely to experience civil war’. This result appears to hold regardless of the
fact whether the group is richer or poorer than the average. Alternatively,
Brown’s (2008) study of the determinants of separatist conflict has found
that the likelihood of separatist conflict increases the richer or poorer a
region is in terms of GDP per capita, compared to the national average.
Due to the emerging consensus in academic circles regarding the causal
relationship between HIs and violent conflicts, in recent years the concept
and theory have also gained purchase in donor agencies and the broader
international development community, particularly in the context of spe-
cific countries undergoing or having recently emerged from violent conflicts
in which such inequalities appear to have played an important role (Brown
and Langer, 2010). While this is a positive evolution, there remain important
issues yet to be analysed and fully understood.
1.3 HIs, perceptions and political institutions:

A framework of analysis
Despite the increased understanding and recognition of the causal link-

ages between the presence of HIs and violent conflicts, there are at least
two important issues that deserve closer attention. A first issue that requires
more detailed analysis concerns how HIs are perceived by ‘the people in the
street’ and to what extent ‘objective’ HIs and perceived or ‘subjective’ HIs are
indeed the same. The latter issue is a particularly important empirical issue
to be investigated. Brown and Langer (2010: 30) argue in this respect:
If groups do not perceive the prevailing socioeconomic and political

inequalities as unfair or unjust, severe objective horizontal inequalities
might not provoke conflict. The converse situation in which for instance
political horizontal inequality is perceived to be severe, yet, objectively
political exclusion appears to be relatively small or non-existent can also
occur. In order to gain political support among their ethnic constituents,
opportunistic politicians or ‘ethnic entrepreneurs’ sometimes actively
promote these misperceptions. To what extent people’s perceptions of
horizontal inequality accurately reflect a country’s objective political and
economic reality is an important empirical question.
A second issue that has so far received insufficient systematic analysis is

the question of how political institutions interact with various dimensions
of HIs and to what extent they affect the likelihood of violent conflicts.
Political institutions condition the modalities of political representation of
identity groups and thereby determine the structure of political HIs in a
country. While the direct link between exclusionary political institutions
and violent conflict has been extensively analysed and is well documented
(see, for example, Stewart ed., 2008; Cederman et al., 2011), the effects
6 Introduction
of different types of political institutions on conflict proneness and peo-

ple’s perceptions on HIs have received much less attention. While during
the 1990s most African countries adopted one or other form of multi-party
system (Takeuchi, 2007), it is important to emphasise that Africa’s political
institutions are still in flux, and it is this very fluidity that opens up a new
research field about the comparative advantages of different types of formal
and informal institutions in conflict prevention.
This book sets out to investigate the complex linkages between objective
and subjective HIs, political institutions and violent conflicts. The central
objective of the book is to analyse the following three parameters and to
improve our understanding of the interaction between them:
1. Objective HIs. How severe are the prevailing objective HIs in terms of
access to political power and state institutions, levels of socioeconomic
development and access to economic resources? Have they resulted
in group mobilisation, violent or otherwise, along ethnic, religious or
regional lines?
2. Subjective perceptions of HIs. How do people perceive the prevailing polit-
ical, socioeconomic and cultural status inequalities in their country?
To what extent do these perceptions accurately reflect the ‘objective’
situation?
3. Political institutions. What types of formal and informal political insti-
tutions can be distinguished in our case-study countries? How do these
formal and informal political institutions affect the management of the
prevailing objective and subjective HIs? Are certain types of political insti-
tutions more conflict-prone, or, conversely, conducive to sociopolitical
stability?
Improving our understanding of the interaction between these three factors

and the emergence of violent conflicts is crucial. Our working hypothe-
sis with respect to the interaction between our three main parameters can
be captured as follows: a particular type of political institution results in a
particular set of policies and may increase or decrease the prevailing objec-
tive HIs in every dimension; these objective HIs subsequently feed into
people’s perceptions of the prevailing inequalities, which can induce peo-
ple to mobilise along ‘ethnic’ lines and provoke political instability, which
may potentially escalate into a violent conflict. The escalation into violent
conflict is a complex process, which requires the analysis of various process-
related factors, sometimes exogenous and idiosyncratic, which may trigger
and fuel deadly conflict. Such variables include the decisions of political
leaders, the behaviours of political parties and armed forces, the spillover of
violent conflicts from neighbouring countries, a sudden change of economic
environment, as well as functions and dysfunctions of conflict-mediating
Political institutions
Socioeconomic HIs
and policies
Political processes Perceptions Structural risk factors
(Conflict) Historical
Stability/Instability
settings
Figure 1.1 Horizontal inequalities, political institutions and perceptions
mechanisms in societies. The hypothesised relationship is shown graphically

in Figure 1.1. In Chapter 10, we will examine to what extent our research
supports the hypothesised relationship between our three central parameters
and the emergence of sociopolitical instability and violent conflict.
1.4 Structure of this book
This study presents case studies of Rwanda, Burundi, Ghana, Côte d’Ivoire,
South Africa, Zimbabwe, Tanzania, Uganda, Nigeria and Kenya (Figure 1.2).
This group includes both ‘peaceful’ countries that have managed their
HIs in a non-violent way (albeit not necessarily in a very constructive way)
and ‘conflict’ countries that are currently confronted with serious political
tensions or active violent conflicts. Large-scale perceptions surveys were con-
ducted in all but three of the case-study countries. It was not possible to
conduct surveys in Côte d’Ivoire because of the ongoing violent conflict
at the time, and in Burundi and Rwanda because of the extremely sensi-
tive nature of the issues involved and the impermissibility of conducting
a survey into ‘ethnic’ issues. While the surveys were not nationally repre-
sentative, they were sufficiently large to ensure that all the major ethnic
and religious groups were included in our survey samples (see Chapter 9 for
more details on these surveys). Consequently, as Langer and Mikami argue
in Chapter 9, the ‘results are therefore only statistically representative for
the selected survey locations, but we can draw wider inferences based on
the assumption that the surveyed areas are qualitatively representative of a
larger part of society’.
The book is organised as follows. Chapter 2 provides a classification of
Africa’s political institutions and analyses whether certain types of for-
mal political institutions lower the risk of experiencing political instability
and/or violent conflicts. Chapters 3–6 present the findings of the case studies
that contrast pairs of neighbouring countries with similar HI structures but
8 Introduction
TUNISIA
MOROCCO
ALGERIA
WESTERN LIBYA EGYPT
SAHARA
MAURITANIA
CAPE VERDE MALI
NIGER
SUDAN ERITREA
SENEGAL CHAD
GAMBIA DJIBOUTI
BURKINA
GUINEA-BISSAU FASO
GUINEA
GHANA NIGERIA SOUTH
SIERRA LEONE ETHIOPIA
CÔTE CENTRAL SUDAN
LIBERIA D'IVOIRE AFRICAN REPUBLIC
CAMEROON
SOMALIA
BENIN
TOGO UGANDA
CONGO RWANDA
GABON
KENYA
DEMOCRATIC
REPUBLIC SEYCHELLES
EQUATORIAL GUINEA OF THE
CONGO
SAO TOME AND PRINCIPE
BURUNDI TANZANIA
COMOROS
MALAWI
ANGOLA
ZAMBIA
ZIMBABWE MADAGASCAR
NAMIBIA
MOZAMBIQUE MAURITIUS
BOTSWANA
SWAZILAND
SOUTH LESOTHO
AFRICA
Figure 1.2 Ten African countries chosen for case studies
contrasting political outcomes, including Rwanda and Burundi (Chapter 3),

Ghana and Côte d’Ivoire (Chapter 4), and South Africa and Zimbabwe
(Chapter 5). Chapter 6 deals with the cases of Tanzania and Uganda: two
countries with different social characteristics that cause their respective for-
mal political institutions to function very differently. Chapters 7 (Nigeria)
and 8 (Kenya) present studies of the history of constitutional engineering to
cope with serious HIs in large ethnically heterogeneous countries. Chapter 9
analyses the survey data we have collected as part of this research project
and examines the extent to which objective and subjective HIs are the
same. Chapter 10 draws the main conclusions and provides a set of policy
recommendations.
Notes
1. Throughout this book, when we use the term ‘Africa’, we refer to the countries in
sub-Saharan Africa if not otherwise stated.
2. For more information on the Uppsala Conflict Data Program (UCDP), visit http://
www.pcr.uu.se/research/ucdp/program_overview/ (last accessed 1 May 2013).
3. The World Bank has strikingly described the devastating consequences of violent
conflicts as ‘development in reverse’ (see Collier et al., 2003).
4. Illustratively, Collier et al. (2008) estimate that 40 per cent of all post-conflict
countries return to conflict within a decade of the cessation of hostilities.
5. This section draws on Stewart ed. (2008) and Brown and Langer (2010), who
extensively discuss the concept of HIs and critically review the empirical evidence
linking HIs and violent conflicts.
6. The Centre for Research on Inequality, Human Security and Ethnicity (CRISE) at
the University of Oxford was a Development Research Centre (DRC) funded by the
UK Department for International Development. For more information on CRISE,
visit http://www.crise.ox.ac.uk (last accessed 1 May 2013).
7. The cultural status dimension was first developed by Langer and Brown (2008). See
also foundational work on the relationship between cultural identity and exclusion
by Kymlicka (1995), UNDP (2004) and Sen (2007).
2
Comparing Political Institutions:
Institutional Choice and Conflict
Prevention in Africa
Yoichi Mine, Mari Katayanagi and Satoru Mikami
2.1 Introduction
When a group of people feel they are excluded from decision-making

processes, they are likely to question the legitimacy of such processes. Col-
lective feelings of alienation may lead to violent confrontations, especially
when people perceive the exclusion of their group from the political arena
to be the major cause of the deterioration in their living conditions and/or
the deprivation of their cultural dignity (see Chapter 1). While addressing
socioeconomic horizontal inequalities (HIs) may take a considerable period
of time, political HIs, at the core of which are uneven modes of group repre-
sentation in multiple tiers of government, can change drastically at the time
of regime transformation. This is the major reason why political HIs are of
critical importance in our research on conflict prevention in today’s Africa,
which is going through volatile transformation of political regimes.
Following the end of the Cold War, a number of African countries expe-
rienced a series of serious violent conflicts. Since that time most of the
continent’s countries have attained fragile peace, and a majority of African
nations have parted with one-party rules and adopted, at least formally, var-
ious patterns of multi-party democracy. However, new political institutions
sometimes emerged out of the ad hoc power-sharing of warring parties, and
are still generally changeable without firm rootage in society. In the mean-
time, the African continent began to experience rapid economic growth,
more than 5 per cent on average during the first decade of the twenty-
first century, occasioned by the strong demand in Asian economies for
African primary products. While the economic expansion provides people
with potential means for the betterment of their living conditions, the same
process tends to polarise African society into growth poles and stagnant
peripheries, intensifying multidimensional HIs between advantageous and
disadvantageous groups and regions. It is this combination of the fluidness
10
Yoichi Mine, Mari Katayanagi and Satoru Mikami 11
of political institutions and the rising inequalities in Africa that necessi-

tates us to take a closer look at the effects of institutional choice on social
inclusion and sociopolitical stability.
As a preliminary effort to justify our selection of countries for case study,
we categorise formal political institutions of African countries into two
broad classifications: power-concentrating (PC) and power-dispersing (PD)
institutions. In this chapter, we begin by defining and classifying these insti-
tutions (Sections 2.2 and 2.3), consider the factors that may have affected
institutional choices and evaluate their relative performances (Section 2.4).
Finally we present a set of hypotheses regarding plural paths to consolidate
peace and stability in Africa. Our basic proposition is that PD institu-
tions composed of power-sharing and decentralisation are relatively suitable
for conflict prevention, combined with a warning that any attempt at
institution-building must be effectively supplemented by policies to address
all dimensions of HIs (Section 2.5). Section 2.6 concludes the chapter.
2.2 Defining political institutions
Let us begin with a theoretical review. In contemporary political sci-

ence, the most influential theory of comparative democracy and consti-
tutional engineering is the theory of consociational democracy elaborated
by the renowned American political scientist Arend Lijphart. As repeat-
edly acknowledged by Lijphart himself, it was the development economist
W. Arthur Lewis who ‘invented’ the prototypical consociational theory
through his investigation of West African politics in the 1960s.1 Tracing
back still further, the intellectual origins of Lewis’s conception of plural
democracy can be found in John Stuart Mill’s Considerations on Representa-
tive Government: ‘In a really equal democracy, every or any section would be
represented, not disproportionately, but proportionately’ (Mill, 1861: 146).
Echoing this, Lewis stated that democracy had two meanings: ‘Its primary
meaning is that all who are affected by a decision should have the chance to
participate in making that decision . . . Its secondary meaning is that the will
of the majority shall prevail’ (Lewis, 1965: 64).
According to Lewis, the two-party system based on this secondary prin-
ciple of ‘the winner takes all’ was developed in the context of European
societies characterised by ‘vertical’ class divisions. In stark contrast, African
society was characterised by ethnic, cultural, religious and geographical
divides between groups, which he called ‘horizontal’ cleavages, upon which
serious economic inequalities are superimposed (ibid.: 18–19). In post-
independence West Africa, the ‘economic differences between the provinces
of a single state are much wider than is known inside any European country’
(ibid.: 24). Lewis maintained that the transplantation of the Westminster
system into such a plural society would bring about a dangerous zero-sum
game, in which an ethnic group that voted for the ‘wrong’ party would be
12 Comparing Political Institutions
harshly penalised. Group differences could be exaggerated in an irretriev-

able way under the majoritarian, ‘first-past-the-post’ voting system (ibid.:
76). His contention was that the conflict of economic interests rooted in
manifest cultural differences could not be resolved appropriately through
the pure logic of class struggle, in a Manichean dualism in which a group
representing justice overwhelms a group representing evil.
Subsequently, as a prescription for West African countries after inde-
pendence, and drawing inspiration from the works of Theodore Riker
and Maurice Duverger, Lewis advocated a unique type of political insti-
tution composed of three elements: proportional representation (or single
transferable vote), constitutionally guaranteed coalitions of major parties
and federalism (or provincial devolution) (ibid.: 49–55, 64–84).2 Although
Lewis’s proposition did not attract the attention of contemporary political
scientists, Lijphart added mutual veto to the three former elements, labelled
this new model ‘consociational democracy’ and presented it in a generic,
powerfully stylised way (Lijphart, 1977, 1985). More recently, citing Ted Gurr
(1993: 290–292), Lijphart has reformulated the notion of consociational
democracy into two dimensions: ‘power-sharing’ and ‘autonomy’ (Lijphart,
2002: 39). Lijphart underscored the advantage of consociational institutions
over majoritarian institutions, a finding that has been reaffirmed by a series
of contemporary works, including Norris (2008).
In the past two decades of ‘managed ethnic heterogeneity’ (Gurr, 2002),
Lijphart’s theory has become increasingly influential. However, as the
consociational theory achieved the status of a modern classic, specialists in
constitutional engineering began to mount a number of criticisms against
his prescription; the points in question include the fixation of group bound-
aries and identities, the lack of competitive opposition politics and the
tendency towards gridlock and disintegration, which can be the direct
outcomes of adopting consociational, decentralised, power-sharing institu-
tions.3 In particular, Donald Horowitz, a major critic of Lijphart, emphasised
the necessity of providing incentives to make political leaders behave mod-
erately towards others and proposed an alternative, hybrid combination of
a vote-pooling system, a presidency and federalism, drawing on the experi-
ences of Nigeria and South Africa (Horowitz, 1985: 568–576; 1991: chapters 5
and 6; Reilly, 2001).4 In addition, despite the coherence and attraction of
consociation as a ‘grand theory’, Robert Dahl cautiously noted that ‘there are
no general solutions to the problems of culturally divided countries’ (Dahl,
1998: 195; see also Sartori, 1997: 69–75). The (self-)criticism by Arthur Lewis
that the federal option, which is ‘in many places the best that we can do
for the time being’, might aggravate existing economic inequalities between
autonomous units (Lewis, 1985: 25) is also worth attention.
Taking into consideration the ongoing debates between the proponents
of consociational democracy and their opponents, in this chapter, we opt
instead for descriptive terms: our analysis focuses around the concepts of
power-dispersing (PD) and power-concentrating (PC) institutions. We shall

provide a working definition of these twin concepts, before providing a gen-
eral mapping of political institutions and examining the relative advantage
of these institutions more closely in African contexts.
The PD design of political institutions conforms to the consensual model
developed by Lewis and Lijphart, most typically the combination of propor-
tionality, coalition and federalism, which guarantees ethnic groups their access
to political power, at least according to their population sizes, while grant-
ing them considerable rights of territorial and/or cultural self-determination.
When individual groups are allowed to raise an effective voice, a negoti-
ated social contract is likely to take on a decentralised, pluralist character,
where national loyalty may be gradually built up on top of ethnic loyalties.
Even though the PD solution may not be a universal panacea, power-sharing
and power-dividing (federalist) institutions have been adopted as a standard
practice in the process of negotiated settlements of violent conflicts in many
places in Africa and elsewhere (Hartzell and Hoddie, 2007). Dahl also empha-
sised that the most favourable situation for a competitive democracy was a
pluralistic social order, with dispersed access to violence and socioeconomic
sanctions (Dahl, 1971: chapter 4; Bangura, 2006: 24).
On the other hand, a PC institution is a majoritarian system, most typ-
ically characterised by first-past-the-post voting, a majority party executive and
a unitary state, and is the opposite of a consensual PD institution in every
respect. A PC institution exhibits the features of the so-called Westminster
system, which formed the basis of the political institutions of today’s
Commonwealth nations and has also been adopted in a number of non-
Anglo-Saxon countries. The Westminster system in the developed nations
is considered to be a legitimate system with strong leadership, accountabil-
ity and efficiency, though some argue that such positive traits are nothing
but myths promoted by elite actors (Rhodes et al., 2009). In any case, it
is important to distinguish a PC institution from an autocracy, in that the
former is an established version of multi-party democracy with a variety of
historical experiments. It must be emphasised that the PC institutions in the
classification below are assumed to be variants of democracy, even though
some of them retain some taint of a past undemocratic de jure one-party
system.
On an abstract plane, the distinction between PD and PC institutions
can be redefined as follows. A PD institution is grounded on the principle
of proportional ‘influence’ exercised by different groups retaining their dis-
crete powers, entering an arena of intergroup negotiation and shifting policy
coalitions, while a PC institution relies on the principle of majoritarian ‘con-
trol’, in which successive teams of political leaders with concentrated powers
are rewarded or punished by the people as the sovereign on the basis of ret-
rospective responsibility (Powell, 2000: 4–7). The former is well suited to
the ‘accommodation’ of social cleavages and multiple identities, while the
latter suits the ‘integration’ of conflicting groups into a single public (and
national) identity group (McGarry et al., 2008).
2.3 Elements of PD and PC institutions
Following on from the conceptual definition outlined above, this section

aims to classify political institutions along the PD/PC spectrum for all
African countries that conduct regular elections under the formal framework
of multi-party democracy. Our baseline for comparison is the classic work
of Lijphart, in which he formulated ten variables to distinguish between
consensus and majoritarian democracies for 36 countries, including only
two African countries, Botswana and Mauritius (Lijphart, 1999: 3–4).5 Build-
ing on this work, we present a set of eight variables and subdivide them
into two dimensions with four variables each to classify formal politi-
cal institutions in African countries. The variables from A to D form the
‘majoritarian/power-sharing dimension’, which reflects the degrees of con-
centration/dispersion of power within the executive and legislative space
in the central governments, while those from E to H form the ‘centralised/
decentralised dimension’, which reflects the degrees of geographical division
of state power and the self-determination of cultural groups. The distinction
is whether ‘power is dispersed to political actors operating together within
the same political institutions or dispersed to separate political institutions’
(Lijphart, 1999: 185).6
Due to the salience of informality and the relative newness of formal
democracy in African politics, it is not very easy to gather all appropriate
quantitative data for the classification of African political institutions along
the PD/PC spectrum. Given this constraint, in our attempt, such variables as
corporatism, a rigid/soft constitution, judicial review and the independence
of the central bank have been dropped from Lijphart’s original formula and
the implementation of local elections and the recognition of group culture
added.
Each element is evaluated by the present authors on a rough three-point
scale, assigning a higher value to more PD institutions. Although institutions
with higher total scores are closer to the ‘ideal’ type of PD institution, the
high figures attached to a PD institution are not premised on its superiority.
Descriptions of each variable follow below.
[1] Majoritarian/Power-Sharing Dimension
(A) Electoral System
The ratings range from 1 to 3, as follows: 1, majoritarian; 2, combined/

intermediary; and 3, proportional representation. We consider the electoral
system of the lower chamber. There is no direct election for an upper house
in today’s Africa, except for Liberia and Zimbabwe, which, as a matter of fact,
have the same electoral systems for both houses. A majoritarian electoral
system is typically a single-member constituency system based on the first-
past-the-post voting method, while proportional representation typically
adopts a party list system.
(B) Party System
The ratings range from 1 to 3, as follows: 1, unipolar system; 2, bipolar

system; and 3, multipolar system. The effective numbers of parties are cal-
culated from the seat allocation in the parliament (the lower chamber in
case of bicameralism) after the latest general election in each country,7 using
the method developed by Laakso and Taagepera (1979) and adopted by
Lijphart (1999: 68). For the purpose of a provisional rating, the effective
party numbers for a unipolar system are considered to be 1.0 and above (and
less than 1.5), those of a bipolar system 1.5 and above (and less than 2.5)
and those of a multipolar system 2.5 and above. Although a unipolar party
system with a high degree of intra-party democracy could theoretically be
as inclusive as other systems, we consider having more parties to be more
power-dispersing.8
(C) Composition of the Executive
The ratings range from 1 to 3, as follows: 1, single-party majority; 2, coali-

tion; and 3, grand coalition. The first is the ‘normal’ shape of an executive
controlled by a majority ruling party. The second is the intermediate cat-
egory, which is typically a minimal winning coalition. The third is an
inclusive coalition in which all major parties participate, a form most typ-
ified by the South African Government of National Unity established in
1994.
(D) Head of Government
The ratings range from 1 to 3, as follows: 1, president; 2, prime minister;

and 3, coexistence of president and prime minister. The president is elected
directly from a nationwide single member constituency, while the prime
minister is elected in the legislature. In this instance a head of government
called ‘president’ but elected in the legislature is deemed to be a prime min-
ister. Constitutional monarchy as well as the case of a symbolic president
nominated in the parliament also falls into the category of prime minister.
When a president who holds office based on direct election shares executive
power with a prime minister in charge of administration, such a system is
considered to be the most power-dispersing in this variable.
In theory, the basic chain of causation is that the choice of electoral

system (A) is expected more or less to affect the configuration of political
parties (B), which then has a considerable effect on the composition of the
executive (C). The procedure to choose the head of the government (D) is
independently introduced into constitutional laws.
[2] Centralised/Decentralised Dimension
(E) Forms of Government
The ratings range from 1 to 3, as follows: 1, unitary state; 2, hybrid unions;

and 3, federation. According to Norris (2008: 170), hybrid unions are coun-
tries ‘where some independent powers are constitutionally recognised for
certain constituent territorial units’ within the unitary framework. We accept
this classification with additional reference to the constitutional documents
of each country.9
(F) Local Elections
The ratings range from 1 to 3, as follows: 1, non-implementation; 2, imper-

fect implementation; and 3, perfect implementation. This element is meant
to measure the extent of decentralisation in practice, as an addition to the
above formal element of government structure. Given that degrees of admin-
istrative and fiscal decentralisation are difficult to gauge in most African
countries, this element focuses on the political control of local governments
by local people. Local elections can be held in varying geographical dimen-
sions such as districts, cities, provinces, as well as mayoral polls. In cases
where local elections involve manifest intervention by the centre, or do not
cover all territories, the country is classified as being in the second category.
(G) Division of Legislature
The ratings range from 1 to 3, as follows: 1, unicameralism without quota;

2, upper house for representatives of non-regional/non-ethnic groups such
as women (or seat quota in the single parliament to this effect); and 3,
upper house for regional/ethnic representatives (or seat quota in the single
parliament to this effect).
(H) Recognition of Group Culture
The ratings range from 1 to 3, as follows: 1, negative reference in favour

of national unity; 2, no provision or a mixture of 3 and 1; and 3, outright
recognition. This aspect is evaluated on the basis of formal expressions in
the constitutional documents about customary laws, traditional authorities,
minority rights, cultural diversity and ‘tribalism’. As discussed below, it is

important to note that communal practices may operate effectively without
any formal recognition of cultural rights, although constitutional stipula-
tions can be an important milestone.
As Lijphart mentioned, ‘we shall simply have to live with concepts that have
very important theoretical and policy significance but that cannot be mea-
sured precisely’ (Lijphart, 2002: 46). Other than the above-listed variables,
we could conceive of several crucial elements of PD and PC institutions in
the African context, which have been dropped because of the paucity of
data and/or the extremely qualitative nature of the information. Some such
limitations are noted below.
First, certain ethnic groups having a monopoly of military power and
bureaucratic positions can be a source of serious ethnic grievances, but it
is very difficult to discover reliable continent-wide data about this subject.
Second, although the measurement of the relative size of the local budget
may be crucial in assessing the practical extent of decentralisation in a uni-
tary state, such data are only available for a handful of African countries
(see also Chapter 10, Table 10.2). Third, some unique qualities of ‘hybrid’
arrangements may not be captured in a quantitative way. For example,
the geographical voting spread requirements in Nigeria should make the
electoral system of the country more power-dispersing (Chapter 8). Finally,
constitutional recognition and everyday practices of traditions are different
matters. It is commonplace in Africa for people never to sell their communal
land in spite of legal provisions to the effect that they can do so.10
Despite all these qualifications, the classification of formal political insti-
tutions based on available data is of importance as a starting point. It is
misleading to assume that formal institutions are eventually swallowed by a
sea of informal practices, for there can also be a very strong influence in the
opposite direction. Formal political institutions like electoral systems often
delineate contours of group boundaries through ethnic coalition-building,
as has been vividly demonstrated in the study of modern ethnic politics in
Zambia by Posner (2005), and thereby shape individuals’ identity choices.
Although people’s identities are historical constructs, people may kill oth-
ers based on their own identity perceptions at a given time. Identity-based
actions are too real, and can also be framed by emergent political institutions
and policies in the formal sphere (see Chapter 1, Figure 1.1). The interac-
tions between formality and informality are discussed fully in the country
case studies in the following chapters.
2.4 Mapping political institutions in Africa
2.4.1 Four categories of institutions and their backgrounds

Applying the definitions and variables of PD and PC institutions given in
the previous sections, we provide a tentative mapping of formal political
institutions in sub-Saharan Africa. In order to represent relative differences

between institutions in a balanced way, each of the country ratings are
standardised separately and summated along each dimension, so that we
have the standardised total scores of A + B + C + D for the majoritarian/
power-sharing dimension and those of E + F + G + H for the centralised/
decentralised dimension. The resulting scores indicate the relative positions
of institutions along the PD/PC spectrum for each dimension, with zero as
the neutral point (Table 2.1). Based on these scores, we chart a distribu-
tion in a coordinate mapping of all countries for which relevant data are
available, setting the majoritarian/power-sharing dimension along the hor-
izontal axis, and the centralised/decentralised dimension along the vertical
axis (Figure 2.1).
In this map, countries located in the upper right are called PD insti-
tutions, in that their total degrees of power dispersion are more than
zero, while, conversely, those in the lower left are called PC institu-
tions. Next, we subdivide each category based on the relative contribu-
tion of two dimensions: all PD-DEC (power-dispersing and decentralised)
institutions are characterised by their orientations towards federalism but
their ratings in the majoritarian/power-sharing dimension vary, while
all PD-PS (power-dispersing and power-sharing) institutions are char-
acterised by their orientations towards power-sharing but their rat-
ings in the centralised/decentralised dimension vary. Likewise, PC-MJ
(power-concentrating and majoritarian) institutions are classified as being
majoritarian chiefly due to their common majoritarian character, while
PC-CEN (power-concentrating and centralised) institutions are classified
as being centralised due to their common character of centralisation of
national governance.11
One constraint pertaining to this two-dimensional mapping is that the
actual configurations of the eight variables are not expressly depicted.
To make up for this limitation, the average shapes of four categories are
shown in Figure 2.2. By and large, the most power-concentrating category
is PC-CEN, which is characterised by the very low degree of decentralisation
and the defectiveness of local elections. The configurations of PC-MJ and PD-
PS in the centralised/decentralised dimension are quite similar, but they dif-
fer substantially in the majoritarian/power-sharing dimension; the electoral
system and the executive of PC-MJ are noticeably majoritarian, while PD-PS
is the most power-sharing in all elements of the majoritarian/power-sharing
dimension. The average shape of PD-DEC is exemplarily federal.
The classification leads us to the next question, about the degrees to
which these categories were ‘naturally’ chosen. If there are strong correla-
tions between the types of institutions and geographical, demographic and
historical attributes of countries, the room for institutional choice must be
limited. After testing, we have found no correlation between the types of
institutions on the one hand and geographical areas and population size
Table 2.1 Ratings of eight variables in sub-Saharan African countries
Country ISO code Majoritarian/power-sharing dimension Centralised/decentralised dimension Total
A B C D Sum of standardised E F G H Sum of standardised

scores scores
Angola AGO 3 1 1 1 −2.56 1 1 1 2 −3.87 −6.43

Benin BEN 3 2 2 1 0.50 1 3 1 3 0.14 0.64
Botswana BWA 1 2 1 2 −2.48 1 3 2 3 1.23 −1.25
Burkina-Faso BFA 3 2 2 3 2.69 1 3 1 2 −1.42 1.28
Burundi BDI 3 1 3 1 0.93 1 3 3 2 0.77 1.70
Cameroon CMR 2 1 2 3 0.21 2 3 1 3 1.67 1.88
Cape Verde CPV 3 2 1 3 0.95 1 3 1 3 0.14 1.09
Central African Rep. CAF 1 1 2 3 −0.96 1 1 1 2 −3.87 −4.82
Chad TCD 2 1 2 3 0.21 1 1 1 2 −3.87 −3.66
Comoros COM 1 1 2 1 −3.15 3 3 3 2 3.85 0.70
Congo COG 1 3 2 1 −0.51 1 2 3 1 −2.00 −2.52
Congo (DRC) COD 2 3 2 3 2.85 3 1 3 2 1.40 4.25
Côte d’Ivoire CIV 1 3 2 3 1.68 1 3 1 2 −1.42 0.26
Djibouti DJI 1 1 1 3 −2.70 1 3 1 2 −1.42 −4.12
Equatorial Guinea GNQ 3 1 2 3 1.38 1 2 1 1 −4.19 −2.82
Eritrea ERI – – – 1 − 1 3 1 2 −1.42 −
Ethiopia ETH 1 1 2 2 −2.05 3 3 3 3 5.40 3.35
19
20
Table 2.1 (Continued)
Country ISO code Majoritarian/power-sharing dimension Centralised/decentralised dimension Total
A B C D Sum of standardised E F G H Sum of standardised

scores scores
Gabon GAB 1 1 1 3 −2.70 1 3 3 1 −0.78 −3.48

Gambia GMB 1 1 1 1 −4.89 1 3 2 2 −0.32 −5.21
Ghana GHA 1 2 2 1 −1.83 2 3 1 2 0.12 −1.71
Guinea GIN 2 2 2 3 1.53 1 3 1 2 −1.42 0.11
Guinea-Bissau GNB 3 2 1 3 0.95 1 1 1 2 −3.87 −2.92
Kenya KEN 1 3 3 3 3.42 1 3 2 3 1.23 4.66
Lesotho LSO 2 3 2 2 1.75 1 3 3 3 2.33 4.08
Liberia LBR 1 3 2 1 −0.51 1 1 3 1 −3.23 −3.74
Madagascar MDG 3 3 2 3 4.01 1 1 3 2 −1.67 2.34
Malawi MWI 1 3 2 1 −0.51 1 1 1 3 −2.31 −2.83
Mali MLI 1 3 2 3 1.68 1 3 2 2 −0.32 1.36
Mauritania MRT 2 3 2 3 2.85 1 3 3 2 0.77 3.62
Mauritius MUS 1 2 2 2 −0.73 1 3 3 3 2.33 1.59
Mozambique MOZ 3 2 1 3 0.95 1 3 1 3 0.14 1.09
Namibia NAM 3 2 1 3 0.95 1 3 3 3 2.33 3.28
Niger NER 2 3 2 3 2.85 1 3 1 2 −1.42 1.43
Nigeria NGA 1 3 2 1 −0.51 3 2 3 2 2.62 2.11
Rwanda RWA 3 2 1 3 0.95 1 3 2 2 −0.32 0.63
São Tomé and Príncipe STP 3 3 2 3 4.01 1 3 1 2 −1.42 2.60
Senegal SEN 2 1 2 3 0.21 1 3 3 2 0.77 0.98
Seychelles SYC 2 2 1 1 −2.41 1 3 1 3 0.14 −2.27
Sierra Leone LSE 1 2 2 1 −1.83 1 3 3 2 0.77 −1.06
Somalia SOM – – – – − – – – – − −
South Africa ZAF 3 2 1 2 −0.15 3 3 3 3 5.40 5.26
South Sudan SSD – – – 1 −1.33 2 1 2 3 0.32 −1.01
Sudan SDN 2 2 2 1 −0.67 2 3 3 3 3.86 3.20
Swaziland SWZ 1 – – 2 − 1 3 2 3 1.23 −
Tanzania TZA 1 2 1 3 −1.38 2 3 3 1 0.76 −0.62
Togo TGO 1 2 2 3 0.36 1 1 3 2 −1.67 −1.31
Uganda UGA 1 2 1 3 −1.38 1 3 2 3 1.23 −0.15
Zambia ZMB 1 3 1 1 −2.26 1 3 1 3 0.14 −2.12
Zimbabwe ZWE 1 2 3 3 2.11 1 3 3 3 1.10 3.21
Note: The ratings are based on the data as of the end of 2011. The outcome of the elections in Gabon on 17 December 2011 is included, while those in
the DRC on 28 November are excluded due to the serious disputes over the election results at the time. The classification of electoral systems is based on
a detailed dataset provided by the International IDEA (International Institute for Democracy and Electoral Assistance), http://www.idea.int/ (last accessed
1 May 2013). Resource-rich almanacs such as Africa South of the Sahara and Political Handbook of the World: Governments and Intergovernmental Organizations
published in various years have been consulted as major references for general features of political institutions and political processes. As for the election
outcomes and the forms of legislature, we have used the data compiled by Nohlen et al. eds (1999) and various contemporary information released by local
electoral commissions and the Inter-Parliamentary Union, http://www.ipu.org (last accessed 1 May 2013). Various articles in journals such as Africa Research
Bulletin: Political, Social and Cultural Series and Jeune Afrique, as well as the constitutional documents of African countries, have also been consulted for the
rating procedure. Eritrea, Somalia and Swaziland are excluded in the following figures due to the lack of relevant data. It must be noted that the ratings in
the centralised/decentralised dimension are based principally on constitutional provisions. In Cameroon and Malawi, however, the upper house has yet to
be established, despite the provisions in their respective constitutions. In such a case, the reality counts. As the standardised scores of each dimension are
rounded off to the second decimal place, the totals on the right-hand side do not necessarily match the simple sums of the two values.
21
22
6 PD-DEC
ETH ZAF
4 COM SDN
NGA
Centralised/Decentralised
NAM LSO
MUS
2 CMR
COD
UGA ZWE
BWA SEN KEN
LSE BDI
ZMB TZA MOZ MRT
0 PC-MJ SYC GHA
CPV PD-PS
BEN MLI
GMB GAB RWA
GIN NER STP
TGO CIV
–2 DJI COG BFA
MDG
MWI
LBR
–4 GNB
AGO CAF TCD GNQ
PC-CEN
–6 –4 –2 0 2 4
Majoritarian/Power-sharing
Figure 2.1 Distribution of PD/PC institutions
Electoral system
3
Recognition of 2.5 Party system

group culture
2
1.5
Division of Composition of the
1
legislature executive
Head of
Local elections
government
Forms of
government
PD-DEC PD-PS PC-MJ PC-CEN
Figure 2.2 Configurations of four major institutional categories

ZAF ETH
5
COM SDN
NGA
NAM
LSO
MUS CMR
KEN COD
BWA SWZ
LSE BDI ZWE TZA
MRT SEN UGA
SYC BEN ZMB MOZ
0 GHA
CPV
GMB
GAB RWA MLI
ERI GIN BFA
DJI CIV
STP NER MDG
TGO
COG
LBR MWI
GNB CAF AGO
TCD
GNQ
–5
0 50000 100000 150000

Total population, both sexes combined,
as of 1 July 2010 (thousand)
Figure 2.3 Population size and degrees of decentralisation

Source: Calculated from UN population statistics.
on the other, except that the top eight most populous countries, Nigeria,
Ethiopia, the Democratic Republic of the Congo (DRC), South Africa, Sudan,
Tanzania, Kenya and Uganda, signify higher degrees of decentralisation
(Figure 2.3). Of these countries, Ethiopia, South Africa, Sudan and Nigeria
present strongly federal features, being classified as PD-DEC (Figure 2.1).
Figure 2.4 shows ethnic fractionalisation based on the dataset developed
in Alesina et al. (2003), indicating that ethnically multipolar societies
tilt towards power-concentration especially in the centralised/decentralised
dimension.12 Although the combination of ethnic diversity and PC insti-
tutions is counterintuitive, it makes sense that multipolar societies can be
stable and thus compatible with democratic centralisation of power. As a
general rule, ethnic conflict is inflamed more easily in bipolar and tripolar
societies where zero-sum situations make it difficult to construct peaceful
ethnic alliances (Bangura, 2006; see also Chapter 6).
Colonial factors also affect institutional choice to some extent. While
Acemoglu et al. (2001) differentiated between extractive and settler insti-
tutions in an attempt to gauge their impacts on economic growth in Africa,
we are more concerned about the degrees to which the patterns of politi-
cal governance originated in colonial centres have been domesticated into
local realities through time. Figure 2.5 indicates that Anglophone Africa
tends to be majoritarian and decentralised, while Francophone Africa tends
ZAF
5 ETH
COM SDN
NGA
LSO MUS NAM

CMR
BWA KEN COD
SWZ UGA
SEN
BDI ZWE MRT TZA LSE
ZMB
SYC CPV GHA MOZ BEN
0
RWA MLI GMB
GAB
BFA
ERI NER CIV
GIN DJI MDG
TGO COG
MWI
LBR
CAF
TCD
GNQ AGO GNB
–5
0 .2 .4 .6 .8 1
Ethnic fractionalisation
Figure 2.4 Ethnic fractionalisation and degrees of decentralisation

Source: The fractionalisation dataset can be downloaded from: http://www.nsd.uib.no/macrodata
guide/set.html?id=16&sub=1 (last accessed 1 May 2013).
to be power-sharing and centralised. This largely conforms to the major

characteristics of the political traditions of the Anglo-Saxon sphere and con-
tinental Europe: the British adhesion to majoritarian rule and the formal
recognition of cultural pluralism in the former, and the historical traditions
of proportionality and the French aversion to intermediate associations in
the latter. It must be noted, however, that the external factors discussed in
this section do not determine everything in the longer term. The dispersity
of the political institutions in these figures shows that most African countries
have enjoyed considerable liberty to choose their institutions in the course
of time.
2.4.2 Temporal shifts

In fact, the shapes of political institutions in several African countries have
transformed substantially in recent years, sometimes crossing the lines that
separate different categories. Table 2.2 and Figure 2.6 show positional shifts
by ten countries, all of which are the targets of our case studies, indicated as
arrows of various directions and lengths.13 It is obvious that the radical shifts
of such countries as Burundi (from PC-CEN to PD-PS), Kenya (from PC-MJ
to PD-PS) and Zimbabwe (again from PC-MJ to PD-PS) are the results of the
25
4 MDG STP
KEN
COD MRT
NER
BFA
2 ZWE
LSO CIV MLI
GIN CPV GNQ
NAM
RWA BDI GNB MOZ
BEN TGO
ZAF CMR SEN
0 LBR TCD
NGA COG
MWI MUS
CAF
TZA SDN
UGA
GHA
LSE
–2 ETH
ZMB
BWA DJI AGO SYC
GAB
COM
–4
GMB
Anglophone Francophone Lusophone Others
4
–2
–4
–6
Figure 2.5 Colonial factors

26
ZAF ETH
5
SDN COM
NGA
NAM MUS
LSO
BWA CMR
KEN UGA COD
SWZ MRT
ZWE BDI SEN
LSE TZA
GHA BEN CPV MOZ SYC
0
GMB ZMB MLI
RWA GAB
BFA NER
CIV GIN STP ERI
DJI TGO
MDG COG
MWI
LBR
CAF TCD AGO GNB
GNQ
–5
4
–2
–4
Figure 2.5 (Continued)

Table 2.2 Change of ratings in the case-study countries
ISO code Majoritarian/power- Centralised/ Sum Major turning points

sharing dimension decentralised dimension total
A B C D Total E F G H Total
RWA1 2 1 3 3 1.95 1 3 1 2 −1.42 0.54 Adoption of new constitution (2003)

RWA2 3 2 1 3 0.95 1 3 2 2 −0.32 0.63
BDI 1 3 1 1 1 −2.56 1 1 1 1 −5.42 −7.98 General election (2005)
BDI 2 3 1 3 1 0.93 1 3 3 2 0.77 1.70
GHA 1 1 1 1 1 −4.89 2 2 1 2 −1.10 −6.00 Two-party counterbalance (2000)
GHA 2 1 2 2 1 −1.83 2 3 1 2 0.12 −1.71
CIV 1 1 3 3 3 3.42 1 1 1 2 −3.87 −0.44 Presidential election (2010)
CIV 2 1 3 2 3 1.68 1 3 1 2 −1.42 0.26
ZAF 1 3 2 3 2 3.34 3 3 3 3 5.40 8.74 End of major coalition (1996)
ZAF 2 3 2 1 2 −0.15 3 3 3 3 5.40 5.26
ZWE 1 1 1 1 1 −4.89 1 3 3 3 2.33 −2.57 Coalition (2008/9)
ZWE 2 1 2 3 3 2.11 1 3 3 3 1.10 3.21
UGA 1 1 1 1 3 −2.70 1 3 2 3 1.23 −1.47 From no-party to multi-party democracy (2005)
UGA 2 1 2 1 3 −1.38 1 3 2 3 1.23 −0.15
TZA1 1 1 1 3 −2.70 2 3 3 1 0.76 −1.94 The thirteenth amendment to the constitution (2000)
TZA2 1 2 1 3 −1.38 2 3 3 1 0.76 −0.62
NGA 1 2 2 2 1 −0.67 3 3 3 2 3.85 3.18 General election (2011)
NGA 2 1 3 2 1 −0.51 3 2 3 2 2.62 2.11
KEN 1 1 3 1 2 −1.16 1 3 1 3 0.13 −1.02 Coalition and nationwide party requirements (2008)
KEN 2 1 3 3 2 3.42 1 3 2 3 1.23 4.66
Note: The first row of each country shows the ratings of the period before the ‘major turning points’ specified in the last column, while the second row
indicates the latest ratings as of the end of 2011. In Kenya, although the referendum held on 4 August 2010 resulted in the adoption of a new constitution,
the dates when its provisions enter into force vary, and hence our ratings are based on the previous constitution. The new constitution is meant to limit
executive power and institute a system of checks and balances, thus becoming the culmination of its constitution-making exercise, which had lasted almost
27
two decades (Chitere et al., 2006; Cottrell and Ghai, 2007; see also Chapter 7).
power-sharing arrangements seeking refuge from cumulative political ten-

sion. The major leftward shifts of South Africa and Côte d’Ivoire are caused
by their political development after periods of solid (the former) and fragile
(the latter) power-sharing.14
Some elements of political institutions are susceptible to radical change,
even though factors that facilitate democratic development transform only
slowly. ‘Often, the electoral system may be changed by legislation alone,
even without a constitutional amendment’ (Diamond and Plattner eds,
2006: iv). It is true that the electoral system and the status of the head of
government can be changed by legislation alone, but the most changeable
aspect of the majoritarian/power-sharing dimension is the party system (B),
which reflects the latest election results, as well as the composition of the
executive (C), which is determined by the decision of major parties to form
or dissolve a coalition government. While three constitutional variables in
the centralised/decentralised dimension (E, G and H) tend to stay constant,
decentralisation is making progress in many African countries with the aid of
donors, so that a gradual movement of PC-CEN countries to its adjoining cat-
egories is expected to take place in the future through the implementation
of local elections (F).
Based on the discussion so far, we put forward a justification for the selec-
tion of countries for the case studies in the following chapters, which is
twofold: the studies of country pairs to evaluate the effects of institutional
choice and the studies of single countries to trace institutional evolution
(see Chapter 1, Figure 1.2). Each set of country pairs consists of relatively
stable and unstable countries sharing a common border. The comparison
of neighbours that have similar HI situations but whose positions and/or
directions of institutional change are contrastingly different will illuminate
the effects of institutional choice on political stability (Chapter 3: Rwanda
and Burundi; Chapter 4: Ghana and Côte d’Ivoire; Chapter 5: South Africa
and Zimbabwe). On the other hand, the comparison of those countries that
have adopted political institutions in the same category and yet shoulder dif-
ferent degrees of political instability will shed light on the roles of societal
features, as well as on key differences in institutions that are not precisely
reflected in our classification (Chapter 6: Tanzania and Uganda). The nature
of the historical evolution of institutions and their locally specific condi-
tions will be closely investigated in the single-country case studies of two big
nations with high degrees of ethnic diversity (Chapter 7: Kenya; Chapter 8:
Nigeria).
2.4.3 Performance evaluation

As discussed earlier, there has been a wide range of academic debates over
the relative advantages of PC and PD political institutions. Can we then
present some evidence to indicate real advantages of a certain category of
institutions over others in African contexts? Figure 2.7 is a boxplot of the
PD-DEC
ZAF
5
NGA
ZWE
UGA
TZA
0 PC-MJ KEN
PD-PS
GHA
RWA
CIV
–5
BDI
PC-CEN
–6 –4 –2 0 2 4
Figure 2.6 Shifting positions of the countries in the case studies
2010 Ibrahim Index of African Governance, which is an integrated measure

of no less than 88 indicators on safety and the rule of law, participation and
human rights, and sustainable economic opportunity and human develop-
ment.15 This testing is crucial because our ratings of eight variables are based
only on formal features of institutions and contain little information about
the quality of governance itself.
Two features are discernible from this figure. First, the PC-CEN category
is outperformed by the PC-MJ and PD-DEC categories, and the differences
are statistically significant. The relative weakness of a PC regime with a cen-
tralised power structure is evident. Second, both the PD-DEC and the PD-PS
category exhibit wide variances in performance. In this regard, it is critically
important not to confuse the causation; the instability of a country may be
the cause, rather than the outcome, of the choice of a particular political
institution. If a number of unstable, conflict-prone countries seek PD solu-
tions, such shelter situations will drag down the average performance of the
PD categories. Therefore, the low averages of the performance of PD insti-
tutions should not necessarily be construed as a weakness of the PD option
but rather as its attractiveness as a short-term refuge.
2010 Ibrahim index of African governance MUS

80 SYC
BWA CPV
ZAF
NAM
GHA
60 LSO STP
TZA BEN
SEN
ZMB
GMB MOZ MLI
GAB MWI BFA KEN
MDG
DJI UGA COM RWA
LSE LBR BDI MRT
CMR NGA
COG TGO NER
40 AGO GNB ETH
CIV
GNQ GIN
CAF SDN ZWE
TCD COD
20
PC-MJ PC-CEN PD-DEC PD-PS
80
2010 Ibrahim index of African governance
70
60
50
40
30
Figure 2.7 Ibrahim Index of African Governance

Source: Calculated from the 2010 Ibrahim Index. http://www.moibrahimfoundation.org (last
accessed 1 May 2013).
In order to check longer-term correlations, let us now attempt another

testing by setting two conditions. First, we set aside a group of unstable coun-
tries in which political institutions could be chosen as emergency shelters
as described above, or that are reeling from serious questions being raised
about their legitimacy. Tentatively, from the sample, we exclude 12 countries

that have experienced violent conflicts with more than 1,000 cumulative
deaths in the past decade, from 2001 to 2010. According to the datasets of
the Uppsala Conflict Data Program, those countries are Burundi, Ethiopia,
Nigeria and Sudan in the PD-DEC group; the DRC and Kenya in PD-PS;
Uganda in PC-MJ; Angola, Chad, Côte d’Ivoire and Liberia in PC-CEN; and
Somalia, which is non-classifiable. The estimated death toll ranges from
1,582 in Kenya to 29,006 in Sudan.16
Second, we drop development-related indicators, and instead make use of
Freedom House Index 2011 (standardised country scores for political rights
and civil liberty). We have tested correlations between political institutions
and development performance, but could not find any significant result
with regard to the human development index and its components, espe-
cially longevity and per capita income. This is understandable for several
reasons. The recent economic growth in Africa has been accelerated by the
unprecedented commodity boom, especially in oil exports, but the discovery
of natural resources has little to do with the types of political institutions.
Similarly, HIV/AIDS, which substantially reduces life expectancy, is a pan-
demic that has spread region-wide, and, for example, a country with an
exemplary democracy like Botswana shows a very high adult prevalence rate:
24.8 per cent in 2009 (UNAIDS, 2010).
We have thus focused on the relationship between institutional choice
and political freedom, assuming that ‘good governance’ forms part of the
potential that may contribute to more equitable development in the future.
The result is shown in Figure 2.8, which indicates high performance by PD-
DEC countries with a statistically significant difference from PC-CEN.
As stated in Chapter 1, this book aims to combine the three perspec-
tives of inequalities, perceptions and institutions. The relationships between
inequalities and perceptions as well as between inequalities and institutions
are fully discussed throughout this book (see also Chapter 9). To fill out
the triangle, however, we also have to examine the relationship between
institutions and perceptions. Therefore, making use of Afrobarometer sur-
veys and our own perceptions surveys, we have tried to test correlations
between the types of political institutions and ethnic perceptions, but
have not found any meaningful correlation between the types of institu-
tions and the proportion of people who thought their ethnic group was
treated unfairly by the government (the fourth round Afrobarometer sur-
veys conducted in 2008). Similarly, we have found no correlation between
the types of institutions and the degrees of optimism about future eth-
nic relations (our own perceptions surveys conducted in cities and towns
in seven African countries).17 However, such a comparison should be
attempted again, as we do not have appropriate information about changes
in perceptions before and after changes of institutions. This is a task for the
future.
32
1 MUS CPV
GHA ZAF BEN
STP
Freedom House sub-category score
BWA NAM
SEN MLI
SYC LSO
total 2010 (standardised)
TZA
0 LSE MWI MOZ
ZMB
COM
BFA
GNB
TGO NER
GMB MDG
CAF GIN
GAB
–1 DJI MRT
COG
RWA
CMR
ZWE
GNQ
–2
1
Freedom House sub-category score
total 2010 (standardised)
–1
–2
Figure 2.8 Freedom House scores (excluding countries recently faced with violent
conflict)
Source: Based on the standardised sums of Freedom House 2010 sub-category scores, http://www.
freedomhouse.org/ (last accessed 1 May 2013).
2.5 Plural pathways towards sustainable democracy?
In the 2010s, dozens of African nations will celebrate 50 years since inde-
pendence, and African countries are continuing their attempts at institu-
tional reforms to accommodate political tensions and multidimensional HIs.
In 2010, Angola, Guinea, Kenya, Madagascar and Niger promulgated new
constitutions. Constitutional referendums were held in Liberia in 2011 and
in Zimbabwe in 2013. Along with such contract-making events, regular elec-
tions are taking root in many African countries. Of course, these attempts all
differ in terms of their content and orientation, and are sometimes tainted
by rioting, suppression and oppositional boycotts, testifying to the fact that
the choice of government and political institutions is a part of the dynamic
political processes in contemporary Africa.
According to Dahl, there are three paths from closed hegemonies
to polyarchies, through competitive oligarchies (liberalisation preced-
ing inclusiveness), through inclusive hegemonies (inclusiveness preceding
liberalisation) and through direct routes (‘the large middle area’), and Dahl
indicated that the first path, by way of competitive oligarchies, could be
the most stable (Dahl, 1971: chapters 1 and 3). These historical sequences
formulated by Dahl imply that there can be plural paths to a univer-
sally desirable state of democratic governance, the transition to which
may not be a simple, unilinear evolution towards modernity. The major-
ity of political institutions in contemporary Africa are, at least formally,
neither oligarchies nor hegemonies but multi-party democracies based on
popular elections, however insecure they may be. PD, PC and interme-
diary or mixed institutions in their ideal forms may be regarded as the
three distinctive pathways towards more stable democracies that would be
geared to contain future violent conflict in Africa (Figure 2.9; see also Dahl,
1971: 7, 34).18
Given such a plurality of pathways to democratisation, let us present
several hypothetical conclusions of this chapter. First, the PD option can be
the best solution for preventing violent conflict as well as putting an end to vio-
lent ruptures, by providing short-term refuge from cumulative tensions in countries
with serious HIs. Despite Dahl’s preference of competitiveness, we endorse the
path of inclusive accommodation of the PD type. On the one hand, a num-
ber of African countries have deliberately opted for ad hoc power-sharing
in the national government to appease worsening political situations. The
term ‘Government of National Unity’, used first in South Africa in 1994,
was subsequently adopted in a number of countries, including the DRC,
Côte d’Ivoire, Burundi, Sudan, Kenya and Zimbabwe (all of which are PD-PS
countries with the exception of Sudan, a PD-DEC country). The attractive-
ness of ad hoc power-sharing in conflict situations is obvious, as the principle
of proportionality is likely to be accepted by all major parties as a second-
best solution. On the other hand, the relatively big countries that have
PC path
Competitive integration
Hybrid evolution
PD path
One-party
system
Inclusive accommodation
Figure 2.9 Three paths of democratisation
experienced serious conflict in the past, like Nigeria, Tanzania, Ethiopia and
South Africa (all are PD-DEC countries except for Tanzania, which is a PC-MJ
country), have elaborated federal or semi-federal systems and succeeded
in pulling the nation together. In the longer term, these two dimensions,
power-sharing and autonomy, are expected to reinforce each other for the
achievement of sustainable peace in post-conflict situations (Wolff, 2011).
The preference for PD institutions in heterogeneous countries can be natu-
ral, in that such a regime is often an extension of a compromise agreement
among antagonistic parties. PD arrangements can be superior to PC arrange-
ments in terms of conflict resolution, in that a military victory over an
enemy that entails civilian casualties can never be the best way of effecting
regime change, as long as different means to achieve the same ends are avail-
able. PC institutions have sometimes been chosen as an outcome of the end
of war, bringing about the marginalisation of the defeated (see Chapter 3).19
In the long run, however, it would be misleading to endorse the PD option
as having an absolute advantage as a means of conflict prevention. On the
one hand, a limitless continuance of a grand coalition may consolidate the
centralised power of ethnic elites, leaving the masses alienated, as in the case
of the joint rule by Christian and Muslim notables in Lebanon before the
civil war broke out in 1975. On the other hand, the unit autonomy in federal
institutions may exacerbate socioeconomic HIs, unless the regime is supple-
mented by effective mechanisms of resource redistribution.20 At present, this
issue is being treated seriously in many African countries as new discoveries
of oil and minerals in one particular region arouse feelings of jealousy in
other regions.
Second, in order to strengthen national cohesion, it may be of benefit for African
countries in relatively stable stages to take developmental paths, being assisted
by some PC elements of political institutions such as a stable majority in the

parliament and a popular president. A set of policies to redress inequalities can
be adopted more easily by a government with a strong leadership, enjoy-
ing support from an impoverished majority in the countryside, than is the
case in please-everyone PD regimes.21 In this regard, it is interesting to note
that the PC-MJ category, a group of relatively good performers in the PC cat-
egory (Figure 2.7), is generally more majoritarian than all other categories
in the power-sharing/majoritarian dimension (Figure 2.2). The political elite
in places like Tanzania and Uganda (PC-MJ countries), as well as Rwanda
and Mozambique (PD-PS countries), had been historically immersed in
political beliefs related to the integrationist cause of national liberation,
which had an ideological affinity with a PC institution. Democratic PC
regimes in Africa may eventually succeed in achieving political stability and
equitable growth, provided that these emerging nations are navigated by
well-disciplined teams of policymakers equipped with sound economic and
social policies. The feasibility of this path should not be rejected, given
the successful attempts in several African countries such as Botswana and
Mauritius to create developmental regimes after independence (Mkandawire,
2001).
Having said that, the dangers and conflict-proneness intrinsic in PC insti-
tutions should be remembered. On the one hand, competitive elections
based on the majoritarian, first-past-the-post voting system have very often
led to election-related violence in Africa, as discussed closely in the chapters
to follow that deal with the elections in Côte d’Ivoire, Zimbabwe, Nigeria
and Kenya. On the other hand, the centralised structure of national gover-
nance in PC institutions tends to undermine the aspirations of distinctive
groups under the name of national integration. In fact, governance perfor-
mance is relatively low in PC-CEN countries (Figures 2.7 and 2.8), which are
characterised by the outright concentration of power at the centre to the
detriment of group autonomy and cultural diversity (Figure 2.2). Depend-
ing on the shape of societies and their ethnic configurations, a PC regime
may face substantial opposition from rich minorities and/or poor minori-
ties, without giving them an opportunity to raise their voices. To avoid the
accumulation of HI-based grievances among minorities, the power of the
majority should be checked effectively by a set of democratic mechanisms
and decentralised governance.
A stable developmentalist path requires sufficient state revenue that can
be based on the broad taxation of peasant production, rather than para-
sitic taxation on rents from the mining sector, especially because a regime
based on the latter has tended to provoke violent conflicts (Mkandawire,
2008: 118–120). A system like a poll tax capturing small-scale producers,
however, is perceived as a remnant of colonial coercion and can be strongly
resisted under democracy, as shown in the cases of Tanzania and Uganda
(Fjeldstad and Therkildsen, 2008). As people are loath to pay tax to rulers
in remote centres, it is worth noting that decentralised fiscal systems are
required, especially in relatively big countries with PC institutions.
Taking the example of transport development projects in post-
independence Nigeria, A.O. Hirschman distinguished between trait-taking
and trait-making project designs. A trait-taking project accepts local traits,
such as HI-related ethnic cleavages, as given (a highway system open to
many small ethnic trucking firms), while a trait-making project is more
ambitious, presupposing certain changes of traits, such as the swift creation
of national solidarity, for the success of a project (a nationwide railway
system operated from Lagos). The former is feasible but tends to preserve
or even reinforce undesirable traits, while the latter is laudable but may
become ruinous before desirable traits are created (Hirschman, 1967: xi,
128–159). Clearly, the inclusive, accommodative PD path and the competi-
tive, integrationist PC path as two contrastive modalities of nation-building
are trait-taking and trait-making respectively, and the argument of this
chapter is to endorse the relative advantage and practicability of the trait-
taking PD path of gradual nation-building. In order to stabilise a PC regime
with multiple HIs, those inequalities should be rendered visible and recog-
nised as urgent policy issues. In case of a PD regime, it is taken for granted
in its founding contract that group differences should be addressed through
political efforts.
Third, recognising the contrasting features of PD and PC institutions, as well
as the relative advantage of PD institutions for conflict prevention, the diversity of
political institutions in Africa should be fully appreciated. Institutional experi-
ments conditioned by complex relationships between state and society are
to continue in African countries. The discussion of formal institutions in
this chapter has shed light on only the tip of the iceberg, in the sense that
violent conflicts can be triggered by intergroup animosities and then sub-
dued by virtue of a revival of societal norms, irrespective of the types of
institutional gimmicks. However, rather than looking at state and society
separately, it would be more fruitful for us to pay attention to the inter-
actions between the formal and the informal, as incongruity between the
design of formal institutions and the informal configurations of society may
cause serious discontent. In this regard, the past and present attempts at
constitutional reforms in relatively big countries with heterogeneous pop-
ulations, like Nigeria and Kenya, abound with historical lessons on how
to design formal institutions, by trial and error, to make them accommo-
date social cleavages in gradual ways (Chapters 7 and 8). Some countries
have called upon the opposite elements to counteract undesirable effects
of the major elements; for instance, Seychelles, Guinea and Lesotho have
adopted a combination of first-past-the-post and proportionality systems for
parliamentary elections. There is room to consider the merits of this kind of
‘hybrid’ system.
2.6 Conclusion
Post-conflict transitional periods offer valuable opportunities for the coun-

tries involved to reach for enduring social contracts as the frames of public
policy, beyond transitory treaty-making and bargaining. Unpleasant expe-
riences and the desire not to repeat them are at the root of the formation
of social contracts and innovative constitution-making (Horowitz, 1991:
149–151). Immediate post-conflict periods are critical junctures for many
African societies, where the memories of violence are still fresh and yet
people are heading towards a better future.
Once a political institution gets on a stable path, the new regime, with
a newly established incentive system and formal/informal rules, tends to
be resistant to further changes. In this sense, PD and PC institutions may
reach equilibria and become polarised in the course of events. On the other
hand, there are possibilities for PD and PC institutions to draw nearer to
each other, adopting the merits of their counterpart, and to tilt towards a
convergence, due to mutual learning among the political units in Africa.
In any case, drastic power transitions concomitant with radical changes in
political systems are likely to make the situations of conflict-prone coun-
tries more volatile, especially when shifts in domestic power relations are
caused by exogenous factors, as witnessed in the past experiments of top-
down democratisation in Africa before and after the fall of the Berlin Wall
(Gurr, 1993, 1994).
There is no free trade in constitutional innovation (Horowitz, 2002: 32).
Local and international constitutional engineers are inclined to look at
conspicuous, successful examples in distant places, or are subject to the his-
torical influence of former colonial powers and to the bias of the home
country experiences of the advisors. One advantage of the country-pair
studies in the following chapters is that we can pay due attention to local
specificities and avoid the unnecessary bias of idealising distant cases, given
that these neighbouring pairs are supposed to share historical and socio-
geographical affinities like fraternal twins. The principle of local ownership
of reform processes must be recognised by all stakeholders beyond the
rhetoric.
Notes
1. See Lewis (1965) as well as Lijphart (1971; 1977: 143–147; 1985: 97–98; 1999: 31,
308; 2002: 45; 2008: 278–279).
2. W. Arthur Lewis, a Caribbean of African descent, was the first ‘non-European’
Nobel laureate in economics (the second was Amartya Sen). Lewis (1965) was
harshly criticised for his unreserved censure of Pan-Africanist leaders, and since
then, in the field of African studies, there has been no serious attempt to reap-
praise this seminal work on constitutional engineering for West Africa (Mine,
2006).
3. For the recent development in the debate, see the following collections of dis-
cussions: Reynolds ed. (2002), O’Flynn and Russell eds (2005), O’Leary (2005),
Diamond and Plattner eds (2006) and Choudhry ed. (2008).
4. The centripetalist theory of Donald Horowitz is said to be ‘positioned towards
the integrative end’ of the accommodative spectrum in McGarry et al. (2008:
53–56, 69).
5. A similar and thorough attempt to appraise the electoral systems in five democ-
racies in Southern Africa, South Africa, Namibia, Zimbabwe, Zambia and Malawi,
has been made by Reynolds (1999), the criteria of which is based on Lijphart
(1994), an earlier study of 27 countries including no African cases.
6. See also the discussion of power-sharing through the two channels, federal
arrangements and consociation, in Human Development Report 2004 (UNDP, 2004:
51–54). Although this chapter defines power-sharing in a narrow sense, this
concept can be expanded to include vertical power-sharing between the centre
and local units. See the framework of ‘complex power-sharing’ (Wolff, 2009).
7. For comparison, we do not use an average of the election results over time, given
that multi-party elections have begun to take root in many African countries
only recently. Note that our classification only provides a snapshot of changeable
situations in contemporary Africa, rather than a comprehensive evaluation of
historically established institutions over time.
8. A unipolar party system is an outcome of (at least formally) democratic elections
and is different from a de jure one-party system. We consider having plural parties
to be better than having one dominant party, as long as all human societies are
accompanied by conflicts of multiple interests, ideologies and affections. Such
conflicts can be mitigated and transformed effectively through the activities of
political parties not ‘as whole’ but ‘as part’ (Sartori, 1976).
9. According to Watts (2008: 59–61), transitional countries such as the DRC and
Sudan are nominal federations and not yet functioning well as full-fledged federal
countries.
10. Policies towards political HIs should be discussed in terms of the three branches
of government, both in the central and local tiers, paying due attention to the
bureaucracy and the security sector as well as to the nexus between formality and
informality (Stewart, Brown, and Langer, 2008: 306–310). As the present chapter
concentrates on the formal aspects of institutions for cross-country comparison,
many crucial elements are left for the country-specific discussion in the following
chapters.
11. For a given PD country, when the difference after deducting the simple sum of
the majoritarian/power-sharing ratings from that of the centralised/decentralised
ratings is positive, the country can be classified as PD-DEC, because its relatively
high scores due to its decentralised characteristics form the major contributor
to push up this country to the PD position. The same can be applied to other
sub-classifications.
12. Fractionalisation is defined as the probability of two randomly chosen individuals
not being members of the same group. The fractionalisation index of a purely
bipolar society is 0.5. An intrinsic problem with this calculation is that ethnic
boundaries are extremely fluid, as identities are constantly chosen by individuals.
13. These past ratings are weighted using the means and standard deviations of
present ratings.
14. Compared to Figure 2.6, a similar mapping of 26 countries around the world pre-
sented in Lijphart (1999: 255) shows gentler shifts. This difference is reasonable,
as the latter is based on average data over a much longer time period, namely, 50
years. See also note 7.
15. The 2010 Ibrahim Index of African Governance is based on data from 2008 to
2009, http://www.moibrahimfoundation.org/en/media/get/20101020_2010-ibra
him-index.xls (last accessed 1 May 2013). The scores of the 2010 African Gov-
ernance Index for our target countries are as follows: Rwanda, 47.2 and Burundi,
44.7; Ghana, 64.6 and Côte d’Ivoire, 36.8; South Africa, 71.5 and Zimbabwe, 32.7;
Tanzania, 55.0 and Uganda, 50.8; and finally Kenya, 50.5 and Nigeria, 43.3.
16. UCDP Georeferenced Event Dataset v.1.0-2011, Uppsala Conflict Data Program,
Uppsala University, http://www.pcr.uu.se/research/ucdp/datasets/ucdp_ged/ (last
accessed 1 May 2013). The data are the sum total of the deaths caused by wars
and battles, one-sided violence from governments and non-state civil conflicts.
Political instability sometimes looms large with fewer casualties. The number of
conflict-related deaths in Zimbabwe during the same period is 253 in the same
dataset (all these figures of casualties are more or less speculative).
17. The question was: ‘Do you think that the relations between the different ethnic
groups in your country will improve or worsen in the future? Or will there be no
change?’
18. Steven Levitsky and Lucan A. Way argue that, since 1990, competitive author-
itarian regimes have taken three divergent paths, transforming themselves into
democracy, unstable authoritarianism or stable authoritarianism, and they anal-
ysed trajectories followed by 14 African countries according to this perspective.
In their study, the persistent dominance of authoritarianism in Africa is attributed
to the strength of the organisational power of the incumbents, the relative weak-
ness of linkages with the West, as well as the rise of leverage from the West
(Levitsky and Way, 2010: 236–308). While they discuss the paths diverging into
democratic redress and authoritarian regress, our contention is that there are
different paths to a common democratic goal.
19. A PC regime may function as a nominal democracy to consolidate the rule of
a military victor. See also Paul Collier’s mixed praise of President Museveni of
Uganda (Collier, 2009: 182–184), based on an analogy between the political vio-
lence in contemporary Africa and the violent state-making process in European
history as described by Charles Tilly.
20. Discussing the longer-term costs of power-sharing, Rothchild and Roeder (2005)
suggest ‘power-dividing’ through a strategy of civil liberties, multiple majori-
ties, and checks and balances as an alternative. A PD regime may also entrench
patterns of ethnic fragmentation as demonstrated in the cases of Bosnia and
Herzegovina (Slye, 1996; Marko, 2005; Samuels, 2006), post-war Iraq (Williams
and Simpson, 2008; Younis, 2011) and other countries. In the context of Africa,
Ian S. Spears underlines that the external pressures to make political opponents
unite are weak, and that despite prominence of ‘ethnic “arithmetic” and the
creation of multiethnic coalitions’ in African politics, bridging across serious
cleavages is not that easy (Spears, 2002: 123, 125–126).
21. The existing literature on political institutions rarely raises the issues of economic
growth and redistribution, though Lijphart (1991) once argued that proportion-
ality and inclusive government were better suited to stable and steady economic
policymaking than majority government.
3
‘Twin Countries’ with
Contrasting Institutions:
Post-Conflict State-Building
in Rwanda and Burundi
Shinichi Takeuchi
3.1 Introduction
Rwanda and Burundi have a number of critical commonalities as well as

dissimilarities. On the one hand, the common features of these two adjoin-
ing countries are readily apparent: they are both tiny and densely populated
countries, lying in the high plateau of the African Great Lakes Region; the
overwhelming majority of their nationals live in rural areas; their economies
are heavily dependent on agriculture; their populations are composed of
three ethnic groups (Tutsi, Hutu and Twa);1 these three groups share the
same language and the same religions (Christianity and Islam), and have
settled in mixed communities; and, finally, in their recent histories both
countries have repeatedly experienced serious ethnic conflicts.
On the other hand, the two countries also display many differences. First
and foremost, the nature of ethnic relations in their national politics has
developed along different and often contrasting lines. Although Tutsi fami-
lies were positioned at the political centre of both traditional kingdoms, the
two countries have developed distinct political processes, especially in the
period after independence. In Rwanda, Hutu elites assumed control of state
power at the time of independence, and maintained it until the civil war
in the 1990s, which resulted in a military victory for the Tutsi-led rebels.
By contrast, in Burundi the Tutsi elites consolidated power through repeated
violent conflicts in the 1960s; the Tutsi’s monopoly of power finally resulted
in the harsh civil war of the 1990s.
As a result of these political developments, today these two post-conflict
countries have very different ethnic policies. While Burundi introduced a
rigorous ethnic power-sharing system, Rwanda denies even the existence of
ethnic differences among its nationals. This contrast warrants some serious
40
Shinichi Takeuchi 41
reflection. The fact that Rwanda and Burundi, often referred to as les pays
jumeaux (the twin countries), share a number of characteristics makes com-
parative methods quite relevant (Durkheim, 1960). This chapter examines
the role played by ethnicity in politics in the two countries, offering a com-
parison of the two countries’ ethnic relations and political institutions in
order to draw out some more general implications for conflict prevention.
Ethnic relations will be analysed through the lens of horizontal inequal-
ities (HIs), which is a useful method for understanding the causal relation-
ships between ethnicity in politics and conflict (Stewart, 2008a). In the case
of Rwanda and Burundi, which have very similar ethnic compositions, the
conclusions that can be drawn from a comparison made using this method
is particularly relevant. Moreover, disaggregating their historical processes
into the political and socioeconomic dimensions of HIs will make any such
comparisons much clearer.
In addition to historical analyses of HIs, this chapter will examine the
political institutions introduced after the recent armed conflicts. In this con-
text institutions can be understood to mean the rules of the game, incentives
and/or norms. The term refers to ‘the humanly devised constraints that
shape human interaction’ (North, 1990: 3), including both formal rules and
informal constraints. In the context of recent post-conflict state-building,
particular attention should be paid to the gaps and interactions of two sorts
of institutions. On the one hand, the international community has tried to
enhance a set of institutions based on the idea of ‘liberal democracy’ – that
is, those that enhance such norms as multi-party democracy, free markets
and human rights, in peace-building processes (Paris, 2004). These efforts
have been reflected in the formal institutions of these countries. On the
other hand, state-building is nothing more than a process of coercive power
accumulation (Tilly, 1992); actors seizing state power thus try to make use
of institutions for their own purposes. The dynamics of post-conflict poli-
tics will therefore be determined through the interactions of these different
motivations.
In comparing the historical processes behind the formation of HIs and
political institutions in Rwanda and Burundi, this chapter aims to clarify
the implications for a general approach to conflict prevention.2 Analyses
of ethnic relations and institutions shed light on the nature of post-conflict
political regimes and the prospects for conflict prevention. In Rwanda, where
revolutionary political change has been a recurrent feature of the state’s
history, authoritarian post-conflict rule led by the ethnic minority casts a
shadow over its peace-building, as the members of the ethnic majority are
likely to resent both political and socioeconomic HIs. While the introduc-
tion of the power-sharing system has drastically reduced the importance of
ethnic politics among the Burundian elites, there is still an imminent danger
of armed conflict breaking out in that country, as the institutionalisation of
the power struggle remains unresolved.
42 Rwanda and Burundi
Our analysis elucidates various historical interactions that have culmi-

nated in the contrasting post-conflict political institutions in the two coun-
tries. These institutions are undoubtedly attributable to the way in which
their wars ended; complete military victory by the former rebel Rwandan
Patriotic Front (RPF) enabled it to impose favourable institutions for its own
rule, whereas a negotiated peace agreement forced a power-sharing system
on the warring parties in Burundi. However, the way in which the war ended
is not the only factor determining post-conflict political institutions; experi-
ences since the pre-colonial period have also influenced recent post-conflict
institutional choice in each country. The contrasting institutions in the ‘twin
countries’ have their own historical grounds.
This chapter is structured as follows. Sections 3.2 and 3.3 focus on
the analyses of HIs. While the former clarifies the formation (and essen-
tialisation) process of identity groups in Rwanda and Burundi, historical
changes in the HIs will be analysed in a disaggregated manner in the latter.
Sections 3.4 and 3.5 deal with the post-conflict political institutions; follow-
ing an analysis of their formal rules in the former, their functions in real
politics will be explored in the latter. An examination of the popular per-
ception of HIs is given in Section 3.6, and, finally, Section 3.7 concludes
with an assessment of political stability in each country. This chapter is
based on the previous literature and the author’s field research in the two
countries.3
3.2 Configuration of identity groups
In both Rwanda and Burundi, the dominant identity groups in politics are
the Tutsi and Hutu, as shown in the repeated outbreaks of ethnically influ-
enced violent conflicts. In the analysis of these groups, two caveats should
be borne in mind. First, the politicisation of the two groups has taken
place largely since the colonial period;4 ethnic antagonism arose between
Tutsis and Hutus as a result of colonial policies in both countries. This
means that the social constructivist vantage point is particularly relevant in
this instance.5 Secondly, the focus should be on the similarities and differ-
ences between the two countries as reflected in their historical experiences.
Although the social characteristics of the Tutsi and Hutu have a number of
similarities, for instance, they share the same language and religions, ethnic
relations in terms of state power have been quite distinct in the two coun-
tries. The necessity and importance of a historical analysis of their group
formation are therefore quite evident.
Pre-colonial Rwanda was the most centralised state in the Great Lakes
Region, a part of the African continent characterised by the presence of
a number of states with stratified societies.6 The centre of Rwandan state
formation, the kingdom of Nyiginya, is thought to have been founded in
the latter half of the seventeenth century. At the end of the nineteenth
century, King Rwabugiri exercised control over almost the same territory as
present-day Rwanda. The king could mobilise strong military power through
his control of influential pastoral chiefs. In this context, a relatively unitary
Tutsi group identity had developed among the ruling elites.7 The political
dominance of the Tutsi had been strengthened in particular in the central
area of the kingdom to such an extent that there are records of several revolts
by Hutu agriculturalists at the end of the nineteenth century (Vansina, 2001:
177). Although there is no doubt that the group identities of the Tutsi and
Hutu were clearly shaped during the colonial period, it is clear from the evi-
dence that Rwanda’s social cleavage between these groups can be traced back
to the pre-colonial era.
By contrast, political power in pre-colonial Burundi was more decen-
tralised. Influenced by delays in state formation, Burundian dynastic history
before the nineteenth century is quite obscure, although its regional auton-
omy was very marked even before that period (D. Newbury, 2001). What
characterised pre-colonial Burundi were the diversity of the social cate-
gory ‘Tutsi’ and the existence of serious cleavage among ruling elites. Three
points deserve to be mentioned in this regard. First, the central actors in
Burundian state formation, the dynastic family, had a unique identity as
Ganwa. While the Ganwa are ethnically a part of the Tutsis, it was only
they who had access to the political power of the pre-colonial state, and
thus they had a clearly distinct identity from Tutsis in general. Secondly, the
Ganwa elites were always in conflict internally, due to the rivalry between
different lines of descent, thus hindering the consolidation of central rule.
Especially important in this context was the opposition between the Bezi
and Batare,8 which remained a source of conflict in Burundian politics until
the 1960s. Thirdly, dichotomous relationships between Tutsis and Hutus did
not exist in pre-colonial Burundi, partly because the Tutsis had never been
a monolithic group,9 and partly because the Burundian Hutus had played
more important roles than their Rwandan counterparts in the pre-colonial
kingdom.10
In the period from 1899 to 1962, Rwanda and Burundi were administered
at first by Germany, as a part of German East Africa, then after the First
World War by Belgium, as a mandated territory on behalf of the League
of Nations, and subsequently as a trust territory on behalf of the United
Nations. Under European rule, ethnic tensions were heightened, because
the colonial authorities11 systematically prioritised the Tutsis, and discrim-
inated against the Hutus in the political system. This policy was based on
a Eurocentric ideology, the so-called Hamitic hypothesis, which argued that
the Tutsi were a superior race, with a European origin, whereas the Hutu were
an inferior race of African origin (Sanders, 1969; Chrétien, 2000). Discrim-
inatory policies were justified under the pretext of stabilising ‘traditional
society’; in public schools priority was given to Tutsi children; and Hutu
chiefs were eliminated from the administration.12 As a result of this policy,
the Hutus generally lagged behind the Tutsis in terms of education as well as
their level of employment in the modern sectors of the economy. The notion
of the Tutsi as ethnically European, based on the Hamitic hypothesis, had an
enormous impact under the unequal power relations of the colonial period,
creating widespread discontent among Hutu elites.
Hutu grievances were more intense and organised in Rwanda, where the
dichotomy between Tutsi and Hutu was much clearer than in Burundi.
In consequence, the political turbulence and ethnic strife, which was
euphemistically termed the ‘social revolution’,13 broke out at the end of the
colonial era. This was Rwanda’s first experience of nationwide ethnic strife,
and resulted in the collapse of the Tutsi-led political system, a massive out-
flow of Tutsi refugees,14 and the Hutu elites gaining a monopoly on political
power after independence. In 1965 their party, the Parti du mouvement de
l’émancipation hutu (PARMEHUTU), won all seats in the national parlia-
ment (Reyntjens, 1985: 445), systematically eliminated Tutsis from political
power and prohibited Tutsi refugees from returning to the country.
The ‘social revolution’ also had a tremendous impact on Burundi, where
people tended to regard it as their possible future, which was desirable for
Hutus but a potential nightmare for Tutsis. The ethnicisation of politics obvi-
ously accelerated after the assassination of the nationalist leader Prince Louis
Rwagasore in October 1961. In spite of King Mwambutsa’s efforts to appease
ethnic tensions, distrust of the king’s political machinations finally resulted
in a coup attempt in October 1965 by a Hutu group in the army and gen-
darmerie. The attempt was severely suppressed by Tutsi groups in the army,
enabling them to seize political power and, subsequently, to overthrow the
monarchy in November 1966. This process of dethroning the king and estab-
lishing the republic resulted in considerable changes in power relations; the
Ganwa elites, who had hitherto occupied the centre of Burundian politics,
were largely replaced by non-Ganwa Tutsi officers in the army, who came
largely from the Province of Bururi. In contrast to the Rwandan Tutsis, the
Burundian (non-Ganwa) Tutsis had not been positioned at the centre of the
traditional kingdom.15
In comparison with Rwanda, where the sudden power shift took place
during the short period of the ‘social revolution’, the post-independence
consolidation of Tutsi hegemony in Burundi advanced only gradually. Their
hegemony, however, was established through bloodshed and mass killing; a
huge number of Hutus were slaughtered following the two aborted coups in
1965 and 1969, and finally in the genocide of 1972.16 It was, in fact, follow-
ing the genocide in 1972 that the Bururi Tutsi monopoly of political power
was confirmed, because the Hutu elites were almost completely eliminated
in the genocide. In 1987, only two seats out of 65 in the Central Committee
of the sole legal party, the Parti de l’union et du progrès national (UPRONA),
were held by Hutus (Lemarchand, 1994: 108).17
3.3 Horizontal inequalities
3.3.1 The political dimension

HIs have various dimensions (Stewart, 2008a). In the case of Rwanda and
Burundi, it is easier to grasp the political than the socioeconomic dimen-
sion, as their political changes have often accompanied a shift in ethnic
relations. The shift has been clearer in Rwanda; as a result of discrimina-
tory colonial policies, political power was dominated by Tutsi elites until
the outbreak of the ‘social revolution’ in 1959, which altered the power
structure completely; Hutu elites in the PARMEHUTU then succeeded in
seizing all political power at the time of independence. Although the com-
position of the elite had been considerably transformed following Juvénal
Habyarimana’s coup in 1973,18 Hutu dominance of Rwandan politics con-
tinued until 1994. In that year the military victory of the RPF resulted in a
complete change in the political power structure. Since then, members of
the former rebel group, namely, former Tutsi refugees, have occupied the
centre of state power.19
In Rwandan history there have been few attempts to achieve ethnic power-
sharing. In fact, the first republic, led by Grégoire Kayibanda, attempted
to drive Tutsis out of the political scene completely; the Tutsi commu-
nity therefore welcomed Habyarimana’s coup, which overthrew Kayibanda
(Munyarugerero, 2003: 161–163). However, the marginalisation of Tutsis
was largely unchanged under the Habyarimana regime, leading to the cre-
ation of the Tutsi-led rebel force, the RPF, in Uganda, and its subsequent
invasion of Rwanda in 1990. As a consequence of the civil war, the for-
mer rebels seized power, sweeping away the political elites of the previous
regime (Prunier, 1995). Rwanda’s history has therefore been characterised by
repeated revolutionary power shifts.
In Burundi, changes in the political dimension of the HIs have been more
gradual and ambiguous in nature. In addition to the fact that Burundian
pre-colonial political power was more decentralised than was the case in
Rwanda, and that Burundian non-Ganwa Tutsis had not occupied a central
position in the pre-colonial kingdom, leaders had often tried to strike an
ethnic balance in politics. While the Burundian Tutsis, like their Rwandan
counterparts, had been privileged during colonial times, the nationalist
leader Rwagasore adopted the principle of ethnic equivalence in his party,
UPRONA.20 When Rwasagore was assassinated, King Mwambutsa made an
effort to attenuate ethnic antagonism by appointing equal numbers of
ministers in terms of the two ethnic groups.21 It was only following the
abortive coup attempt in 1965 that new Bururi Tutsi elites took control
of the security organisations and began to systematically expunge Hutus
from the national political scene. Their dominance in politics was further
strengthened through bloodshed in 1969 and 1972.
Although the principle of ethnic power-sharing was agreed among

Burundians in the 2000 peace agreement, its precursor had already appeared
at the end of the 1980s. Following the ethnic killing that occurred in the
north of the country in August 1988, the then President Pierre Buyoya
launched several initiatives for power-sharing with Hutus, including the
establishment of a cabinet containing equal numbers of Tutsis and Hutus,
in October 1988. However, there were ample grounds for questioning his
motives,22 because the power-sharing principle was never carried over into
security organisations such as the army, gendarmerie and police, which
had always been dominated by Tutsis (Reyntjens, 1994: 68–76). Although
Buyoya subsequently accepted the introduction of multi-party elections,
which resulted in the victory of a Hutu candidate, Melchior Ndadaye, in June
1993, the ethnic imbalance in the security organisations led to the latter’s
assassination by the army just four months later. Hutu politicians partici-
pated in governments even during the civil war caused by the assassination
of Ndadaye, although the security organisations were always controlled by
Tutsi officers.23
The conclusion of the Arusha Peace and Reconciliation Agreement in 2000
did not produce an immediate change in the ethnic balance. It was only
after the signing of a power-sharing agreement in 2003 between the govern-
ment and the largest rebel group, the Conseil national pour la défense de la
démocratie–Forces pour la défense de la démocratie (CNDD-FDD), and the
subsequent integration and reform of the security organisations, that Hutu
officers began to be appointed as core members of these organisations.24 Fol-
lowing the adoption of the constitution in 2005, the principle of ethnic
parity was fully applied and this has been observed to date.
In sum, the political dimension of the HIs in the two countries has moved
in very different ways. Developments in Rwanda have been characterised
by the two abrupt and complete changes: the ‘social revolution’ just before
independence, and the RPF’s military victory in 1994. While the Tutsi elites
had dominated all of the important political posts before 1959, the situation
shifted suddenly to the dominance of Hutu elites as a result of the ‘social
revolution’. However, the RPF victory in 1994 once again gave political
supremacy to the Tutsi elites. Each of these radical shifts was accompanied
by a total replacement of political elites, including members of both the
government and also the security organisations.
In Burundi, Tutsi–Hutu power relations have been more ambiguous.
Broadly speaking, Tutsi elites had ensured their own political dominance
from pre-colonial times to the recent introduction of the ethnic power-
sharing system. Nevertheless, despite the power shift from the Ganwa to the
Bururi Tutsis, Hutus continued to exercise influence on the political scene,
apart from a period of nearly two decades following the genocide in 1972.
The Burundian ethnic power-sharing mechanism should be understood in
this historical context.
3.3.2 The socioeconomic dimension

It is not straightforward to demonstrate scientific evidence about socioeco-
nomic HIs between Tutsis and Hutus: there are no statistics based on each
group; and geographical comparisons are generally irrelevant, as the regional
concentrations of each group are not obvious. The socioeconomic dimen-
sion of HIs, however, tends to depend on the political dimension; when a
Tutsi (or Hutu) elite group gains power, a relatively small number of people
from the same ethnic group are likely to have opportunities for accumu-
lation, although the overwhelming majority of the group may obtain no
such chances. Neither of the two groups is homogenous, but as far as the
socioeconomic dimension of HIs among political elites is concerned, its
characteristics can be estimated from the nature of the political dimension.
Additional information exists for understanding the socioeconomic
dimension in Rwanda. First, its nature in formal education as well as formal
job markets can be traced roughly. In these sectors, Tutsis were generally in a
favourable position during the colonial period; the colonial authorities privi-
leged the sons of Tutsi chiefs for modern education; as a corollary, this policy
gave them a greater chance of securing jobs in the developing sectors of the
economy. The situation was basically the same in Burundi. The ‘social revo-
lution’, however, resulted in a fundamental change in the Rwandan political
power structure. Following independence, the Kaybanda and Habyarimana
regimes adopted a quota policy, according to which the number of Tutsi stu-
dents as well as teaching staff was to be limited to 9 per cent in secondary
schools and universities. Although enforcement of the policy was generally
loose, for Tutsis the threat of expulsion was always genuine; they were, in
fact, systematically expelled from schools, universities and administrative
positions in 1973.25
Following the advent of the RPF-led government, the situation changed
completely. Policy changes, such as the abolition of the quota system and
the political environment, were generally advantageous to Tutsis, enabling
many Tutsi returnees to enter schools and universities. The language pol-
icy, for example, has undoubtedly contributed to the further enrolment of
Tutsis. Before the civil war, Rwanda had two official languages: Kinyarwanda
and French. Soon after its victory in the civil war, the RPF added a third
official language, English. This assisted those Tutsi returnees from Uganda,
where the core members of the RPF were educated, to advance in educa-
tion. In late 2008, the government adopted an important policy change,
selecting English as the sole medium of instruction, with children beginning
to study it from the first grade. This change in the language of education
not only caused serious problems for teachers, who were obliged to learn
the new foreign language, but also had a significant social and political
impact on the present Rwandan context, where language is a proxy for
identity (Hintjens, 2008; Samuelson and Freedman, 2010), precisely because
English is considered to be the language of former Tutsi refugees. This policy
change clearly demonstrated the stance of the RPF-led government in main-

streaming the ‘culture’ of former Tutsi refugees, who in fact tend to ‘ “feel”
themselves to belong to the inner circles of power’ (Ingelaere, 2010: 286).
Secondly, the rural–urban divide has a special meaning in the case of
socioeconomic HIs in Rwanda. Before the civil wars in the 1990s, the urban-
isation rate was very low in both countries; subsequently, however, the
Rwandan urban population has grown rapidly whereas that in Burundi
remains relatively low.26 This growth in urban population has been mainly
due to a massive influx of Tutsis returning to their homeland following the
victory of the RPF.27 In the same period, Rwanda has achieved rapid eco-
nomic growth, but there has also been a considerable widening of the level
of economic inequality.28 In a recent report UNDP has pointed out that the
nature of the inequality is also changing: ‘it is becoming increasingly rural
and increasingly detrimental to the poorest and most vulnerable groups
in society’ (UNDP, 2007a: 18).29 These facts demonstrate strongly that the
main beneficiaries of postwar Rwandan economic growth have been Tutsi
returnees, who often have connections with RPF members.
Regarding the socioeconomic dimension of the HIs in Burundi, the pre-
vious literature has highlighted two major problems. First, there is serious
educational inequality between the ethnic groups. This situation has deep
roots going back to the colonial period, because the colonial authorities
adopted discriminatory measures in education, as we have already men-
tioned. Although this discrimination was gradually rectified in the late
colonial period and the first half of the 1960s, the seizure of power by the
new Tutsi elite had a serious effect on the ethnic balance in education. In
particular, Hutu students in higher education, along with educated Hutus
in modern sectors such as the civil service, were systematically killed during
the 1972 genocide (Gaffney, 2000: 143). The lack of higher education has
naturally led to limited opportunities for employment in modern sectors,
thus aggravating the socioeconomic position of the Hutus.
Secondly, regional imbalances appeared, as a consequence of public
investment policies during the 1970s and the 1980s that privileged Bururi
province, which was the home province of three consecutive presidents
(Micombero, Bagaza and Buyoya).30 Calculations from data collected around
2000 showed that the social infrastructure in Bururi was much better
equipped than in other provinces.31 Such advantages for Bururi, however,
seem to be lessening as a result of the political changes in the 2000s; accord-
ing to recent statistics, the figures for Bururi province were not substantially
higher than other provinces, although its figures were still better than the
national averages.32
3.4 Post-conflict institutional choices and their backgrounds
In this section, we compare the characteristics of the basic political institu-

tions of the two countries through an analysis of the recent constitutions
they have adopted after the serious armed conflicts and explore the back-
grounds of their institutional choices.
3.4.1 Rwanda
Adopted in 2003, nine years after the end of the civil war, the Rwandan
constitution provides political institutions based on a system of multi-party
democracy (Republic of Rwanda, 2003). It can be classified as a semi-
presidential system, in which a president (the head of the state) and a prime
minister (the head of the cabinet) are both active participants in the admin-
istration of the state. As a multi-party democracy, political organisations are
permitted to be formed and to operate freely. However, they are required
not to destabilise national unity (Article 52), and are prohibited from basing
themselves on ‘divisions’ such as race, ethnic group and region (Article 54).
In essence, the Rwandan constitution is shaped by the shadow of the
1994 genocide. Determination not to repeat the genocide and to promote
national unity is expressed repeatedly in its preamble and the text, thereby
emphasising the importance of ‘unity’ and the danger of ‘division’. In fact,
commitments to ‘fighting the ideology of genocide and all its manifesta-
tions’ as well as ‘eradication of ethnic, regional and other divisions and
promotion of national unity’ are stipulated as two of the six fundamen-
tal principles of the state (Article 9). These principles are based on the
idea that the genocide was derived from ethnic discrimination and division
among nationals. The prevention of genocide is pursued further by abolish-
ing a period of limitation for the crime of genocide, and by stipulating that
its ‘revisionism, negationism and trivialisation’ are punishable (Article 13).
In addition, the constitution states not only that ‘discrimination of whatever
kind’ is ‘prohibited and punished by law’ (Article 11), but also that any form
of ‘division’, such as on ethnic, regional and racial lines, is also punishable
(Article 33).
In this context, ethnic power-sharing is unthinkable, because the very
existence of ethnic diversity is formally denied in the constitution. More-
over, as shown in its fundamental principles, the constitution considers
ethnic difference as something to be eradicated; those who emphasise dif-
ference will be regarded as wrongdoers promoting ‘division’. Under the rule
of the RPF, terms such as ‘division’, ‘divisionism’ and ‘genocide ideology’ are
often utilised when the government criticises its opponents. For instance,
before the first post-conflict election, in 2003, the biggest Hutu opposition
party (Mouvement démocratique républicain, MDR) was ordered to dissolve
itself because of its ‘divisive ideology’.33 In 2010 April, a Hutu woman, who
had declared her intention to be a candidate in the presidential election, was
arrested and charged with ‘propagating the Genocide Ideology, Revisionism
and Ethnic Division’ (The New Times, 23 April 2010).
While detailed analysis of the Rwandan constitution has already revealed
a number of problems (Reyntjens, 2003), two points deserve to be men-
tioned here. First, some of its articles constitute institutional arrangements
that operate to the advantage of RPF rule. One example is the composition
of the legislature, because a limit is placed on the proportion of members
to be elected by universal suffrage (Article 76). In the case of the Cham-
ber of Deputies (the lower house), 53 of its 80 members are elected by
universal suffrage with a secret ballot. The remaining 27 seats are reserved
for representatives of women (24 seats),34 youth (2 seats) and the disabled
(1 seat). These representatives are selected by members of local adminis-
trations and/or related official councils, on which the RPF-led government
can exert a dominant influence (Reyntjens, 2003: 77).35 Considering the fact
that the RPF’s core supporters are an ethnic minority, these measures lim-
iting universal suffrage were likely devised in order to maintain its rule.
Secondly, important civil rights such as freedom of thought, opinion and
religion (Article 33), as well as freedom of the press and information (Arti-
cle 34), are defined as being ‘guaranteed by the State in accordance with
conditions determined by law’. The text indicates the over-presence of the
state, because these fundamental human rights should include freedom from
the state. These clauses demonstrate the intention of the Rwandan state to
control the societal sphere for its own purposes.
The 2003 Rwandan constitution was based on the RPF’s political ide-
ology; against the backdrop of the victory in the civil war, it attempted,
through the constitution, to legitimise the revolutionary change and to insti-
tutionalise its gains, for example, by emphasising national unity, denying
‘division’ and promoting state control over politics and civil society. Since
the RPF’s victory, its elites have virtually monopolised important political
posts, while condemning the ‘bad governance’ of previous regimes as hav-
ing been responsible for the genocide, and have adopted a series of radical
reforms in such fields as local administration, agriculture, education and
health care. In some of these reforms, a tendency towards elitism as well as
radical social engineering has been observable (Ansoms, 2008, 2009). While
it is clear that these political behaviours derive directly from the RPF’s com-
plete victory in the civil war, they have other roots, such as its members’
origins as guerrilla fighters, their ethnic affiliation as minority Tutsis and the
influence of other ‘African new leaders’ (Ottaway, 1999).36 In addition, the
revolutionary ideology has its roots in Rwandan history itself. Monopolies of
power and revolutionary changes in power relations have occurred repeat-
edly in Rwandan history. In this sense the RPF’s victory in 1994 was very
similar to the ‘social revolution’ that had occurred 30 years before.
3.4.2 Burundi
Following the conclusion of the Arusha Peace and Reconciliation Agree-
ment in 2000 and the subsequent transition period, Burundians approved
their constitution in a referendum held in 2005 (République du Burundi,
2005). Like Rwanda, multi-party democracy was chosen as a political system
(Article 75). A president is the head of state, but Burundi does not have a
prime minister, and instead established two vice-presidents, which is clearly

a power-sharing device.
In fact, Burundi’s political institutions as stipulated in the 2005 consti-
tution are characterised by a rigorous system of power-sharing between
ethnic groups. Unlike in Rwanda, the Burundian Constitution recognises
ethnic diversity, in its first Article, and defines a series of regulations for
power-sharing. As shown in detail in Table 3.1, there are strict definitions of
Table 3.1 The ethnic power-sharing mechanisms defined in the Burundian 2005
Constitution
Function Rules of power-sharing
Vice- The two Vice-Presidents shall belong to different ethnic groups and
President different political parties. (Article 124)
Cabinet The cabinet must include a maximum of 60 per cent Hutu Ministers
and Vice-Ministers, and a maximum of 40 per cent Tutsi Ministers
and Vice-Ministers. (Article 129)
Security The Minister in charge of the National Defence Force should not
organisations belong to the same ethnic group as the Minister in charge of the
National Police. (Article 130)
Members of the same ethnic group should not account for more
than 50 per cent of the members of the Defence and Security
organisations. (Article 257)
Public Ethnic representation in public enterprises is assigned as a
enterprises maximum of 60 per cent for Hutus and a maximum of 40 per cent
for Tutsis. (Article 143)
National The National Assembly is composed of at least 100 members, with
Assembly 60 per cent being Hutu and 40 per cent Tutsi. (Article 164)
The election of members of parliament is carried out through a
proportional representation system with closed lists. Party lists must
have a multi-ethnic character, and take gender equality into
account. From every three candidates in order on a list, only two
can belong to the same ethnic group, and at least one of every four
must be female. (Article 168)
Senate The Senate is composed of: (1) two delegates of each Province,
belonging to different ethnic groups; (2) three persons from the Twa
ethnic group; and (3) the former Heads of the State. (Article 180)
Judiciary The composition of members of the Magistrate Upper Council
should take ethnic, regional and gender balance into account.
(Article 217)
Commune Neither of the principal ethnic groups should be represented beyond
67 per cent of the national total of Commune Administrators.
(Article 266)
Source: République du Burundi (2005).

the compositions of the two main ethnic groups (Tutsi and Hutu) in organ-
isations related to state power. The system, designed along consociational
lines,37 is based on the idea of sharing state power among political elites.
In this power-sharing system, ethnic quotas are calculated on an individ-
ual basis; that is, Tutsi members from Tutsi-led parties (like UPRONA) but
also from Hutu-led parties (like CNDD-FDD) are counted as Tutsis.38 This
arrangement, together with the effect of the electoral rule obliging parties to
put candidates from different ethnic groups on their lists (Article 168), has
shaped the pattern of power struggle in Burundian politics.
The reasons for the particular institutional choice in Burundi deserve some
consideration. There is no doubt that the immediate cause of the power-
sharing agreement was the military stalemate in the civil war. Although
Tutsi-dominated government forces (Forces armées burundaises: FAB) con-
trolled the capital with overwhelming forces and equipment during the civil
war, they could not impose security on rural areas, where the Hutu rebels,
the CNDD-FDD and Forces nationales de libération- Parti pour la libération
du peuple hutu (FNL-Palipehutu), prevailed. The importance of mediating
efforts offered by Julius Nyerere and Nelson Mandela (naturally with the
cooperation of their governments) should be recognised, but the fact that
neither side could win the war by military means was the most fundamental
reason for the local parties to accept a power-sharing arrangement. In addi-
tion to this immediate reason, however, Burundian historical experience
should be considered as another non-negligible factor affecting the choice
of institutions. Looking back on Burundi’s modern history, power-sharing
arrangements were often adopted to appease political tensions; to deal with
political difficulties Burundian political leaders, from Rwagasore to Buyoya,
have resorted to such arrangements. The idea of ethnic equality in political
institutions was a familiar element of the Burundian experience.
3.5 Institutions and political power: How do

the institutions work?
3.5.1 Formality and reality of the institutions

In Chapter 2, the political institutions of African countries are examined
in terms of whether they are PD (power-dispersing) or PC (power-
concentrating) institutions, for the purpose of classifying a typology of
political power. According to that analysis, the positions of Rwanda
and Burundi are not considerably different in the diagram indicating
the relationship between ‘centralised/decentralised and majoritarian/power-
sharing’ (Figure 2.1). This result appears counter-intuitive, as it is clear that
the political institutions of the two countries are quite contrasting. This
reminds us of the need to consider both formal institutions and their man-
agement in practice in order to understand the exercise of political power.
Among eight comparison variables, two elements need additional expla-

nation. The first concerns the party system rating (B), in which Rwanda is
classified as a bipolar and Burundi as a unipolar system, according to the
effective party numbers.39 However, it is highly debatable whether the oppo-
sition parties in the parliament are really ‘oppositional’ in Rwanda, because
these parties voted for the incumbent Paul Kagame, the former rebel com-
mander, in the past presidential elections. Although parties other than the
RPF exist in the Rwandan parliament, their behaviour is fairly similar to that
of the ruling party. In the case of Burundi, the effective party number tends
to reflect the reality of political power in a parliament. Its unipolar nature
was a result of the particular conditions involved: the boycott by the main
opposition parties in the 2010 election. In fact, a calculation based on the
results of the 2005 election shows a much higher score (2.42).
The second concerns the rating for the recognition of group culture (H);
although Rwanda and Burundi have no significant differences in respect of
the cultural dimension of minority rights, they diverge in terms of their
political dimension. The two countries do not admit the existence of tradi-
tional authorities; both derive from pre-colonial kingdoms, which had been
transformed into republics in the 1960s following the dethronement of their
kings. Moreover, both countries recognise special political rights for certain
minority groups; the Burundian constitution clearly ensures political posts
for the ethnic minorities, the Tutsi and Twa, because of its system of ethnic
power-sharing; in Rwanda, women and the disabled have quotas for repre-
sentation (24 seats and 1 seat respectively) in the Lower House (Article 76),
while ‘historically marginalised communities’ are provided with eight seats
in the Senate.40 However, Rwanda does not recognise any political right in
terms of ethnic groups; indeed their very existence is denied. Ethnic diver-
sity is concealed in Rwanda precisely because it is the minority group that
controls power; defining who the majority and the minority are is dangerous
for the power-holders under a system of multi-party democracy.41 The crit-
ical difference between the two countries regarding the position of ethnic
minorities in politics should be emphasised.
3.5.2 Political power and its performance

In the post-civil war period, Rwandan politics has been virtually controlled
by the RPF, whose basis of power has been nothing more than military
might. While the victory in the civil war in 1994 enabled it to dominate the
overarching political system, the military operation in the eastern part of
the Democratic Republic of the Congo (DRC) since 1996 (Reyntjens, 2009),
as well as the counter-insurgency operations in western Rwanda in 1997
and 1998,42 contributed to the establishment of an RPF-led political order in
the area, including the eastern DRC. Despite the reduction in the numbers
of military staff, as well as the integration of former civil war combatants
into the national army (Rwandan Defence Forces, RDF) under the auspices
of the Rwanda Demobilisation and Reintegration Commission (RDRC), the

power structure in the Rwandan military did not change; the core members
of the military have been always former Rwandan Patriotic Army (RPA)43 offi-
cers; that is, former guerrilla fighters.44 Although the RPF has not changed
its Tutsi-centred nature, the RPF’s power, based on the military as well as
the self-advantageous institutional arrangements, has enabled it to dom-
inate politics under the multi-party democratic formalities (Dorsey, 2000;
Reyntjens, 2004, 2011; Beswick, 2010).
Under the RPF-centred political system, post-conflict Rwanda has achieved
steady economic growth. The average annual GDP growth rate was as high
as 4.39 per cent between 2000 and 2010 (World Development Indicators).
One of the crucial factors contributing to the high level of economic growth
is the export of mineral resources,45 the overwhelming majority of which
come from the eastern part of the DRC. The increase in exports of min-
eral resources was clearly attributable to Rwandan military intervention in
the eastern DRC and its subsequent dominance of that region.46 On this
point, the former editor of the UNDP National Human Development Report
(UNDP, 2007a) argued that economic growth based on military intervention
led to high economic disparity (Silva-Leander, 2012).
While Burundian ethnic policy contrasts with that in Rwanda, the features
of its political power have been rather similar. Since its victory in the 2005
elections, the former rebel-cum-ruling party, the CNDD-FDD, has concen-
trated its power in politics to the detriment of other parties (Vandeginste,
2011). The boycott strategy of the main opposition parties in the 2010 elec-
tions resulted in extending the political domination of the ruling party.
At present, political antagonism has emerged not in the appearance of an
ethnic problem, but in terms of sheer power struggle.
The Burundian ethnic power-sharing system has been well observed to
date, with a broad consensus among stakeholders. Because of the institu-
tional constraints, political parties tend to be multiethnic. As shown in
Table 3.2, in the past two legislative elections, more than 30 per cent
of CNDD-FDD members in the National Assembly were Tutsi, despite the
party’s origins in the Hutu rebel movement. This phenomenon was the result
of the abovementioned institutional arrangements. Although the core mem-
bers of the CNDD-FDD remain Hutu males, the party has the appearance of
being multiethnic and multi-sex because of this rule.
The ethnically and sexually mixed appearance of the ruling party, how-
ever, does not necessarily result in a dispersed power structure. On the
contrary, just like the RPF, the CNDD-FDD has often been accused of concen-
trating too much power, of intimidating opponents, and of manipulating
political institutions (ICG, 2006, 2011; Human Rights Watch, 2010). Fol-
lowing the ceasefire and power-sharing agreement in 2003, the CNDD-FDD
has exerted a strong influence over the security organisations (the National
Defence Force and the National Police and National Intelligence Service) in
Table 3.2 Ethnic profile in Burundi’s National Assembly
Hutu Tutsi Twa Total
2005 CNDD-FDD 43 21 0 64
Frodebu 23 7 0 30
Uprona 0 15 0 15
CNDD 3 1 0 4
MRC 0 2 0 2
Twa 0 0 3 3
Total 69 46 3 118
2010 CNDD-FDD 54 27 0 81
Frodebu-Nyakuri 3 2 0 5
Uprona 5 12 0 17
Twa 0 0 3 3
Total 62 41 3 106
Note: Results of the elections in 2005 and 2010.

Source: Election in 2005: Reyntjens (2006). Election in 2010: CENI (2010a).
occupying key posts in these organisations. The intimidation of opposition

supporters and journalists has therefore been carried out via the security
organisations, especially the National Police and National Intelligence Ser-
vice.47 Although the strategy of boycotting the elections has often been
criticised not only by the Burundian government but also by external actors
such as donor countries, its background, in which the ruling party had relied
on authoritarian measures, should be taken into account (ICG, 2010, 2011;
Human Rights Watch, 2010).
While the dominance of the CNDD-FDD has been becoming increasingly
established in Burundi, its economic performance has been rather disap-
pointing. The average annual GDP growth rate between 2000 and 2010 was
only 1.35 per cent – less than the annual population growth rate during the
same period. The following points are important as factors explaining this
poor macroeconomic performance. First, Burundi’s economic governance
has remained fragile in comparison with Rwanda’s, as is shown by several
indicators.48 Secondly, Burundi does not have an engine of growth compa-
rable to Rwanda’s exports of its mineral resources. The reason why Burundi
does not export mineral resources is undoubtedly attributable to the extent
of military intervention in the eastern DRC; Burundi could not organise a
network for mineral exports, as it has not been as deeply and systematically
involved in the Congo War as Rwanda.49
3.6 Popular perceptions50
In spite of its official negation, ethnicity has been popularly perceived as

playing a critical role in post-conflict Rwanda (Ingelaere, 2010). Ironically,
the more the RPF negates ethnicity, the more it tends to be regarded through
the lens of ethnicity; its emphasis on national unity as well as its refusal to
recognise ethnicity have been interpreted from an ethnicity-centred logic, as
the RPF’s agenda for maintaining political power. This logic asserts that the
RPF does not want ethnicity to enter into politics, because it is supported by
the ethnic minority: the Tutsi.
With regard to post-conflict Rwanda, the dangers of social exclusion have
often been pointed out. There are, in fact, three related problems. The first
concerns state patronage; in the circumstances in which the political arena
has been dominated by the RPF, the Tutsis have had greater chances to be
the beneficiaries of state service provision, including opportunities for edu-
cation and employment in the sector, because of their patronage network
with political leaders. Since its seizure of power, the RPF has exerted its influ-
ence not only on the central bureaucracy but also on local administrations.
Today, leaders of local administrations are, almost without exception, ardent
supporters of the RPF.51 While the state patronage network also includes the
Hutus, as both Tutsi and Hutu are far from monolithic groups, it is neverthe-
less beyond doubt that the overwhelming majority of those who have been
excluded from the network are Hutus.
The second problem relates to economic disparities. As we have already
mentioned, rapid economic growth in Rwanda has widened economic dis-
parities, in a setting in which Tutsi returnees have tended to be the winners.
The third danger is related to Gacaca: a popular, participatory transitional
justice process for punishing genocide perpetrators. Its social impact has
been enormous, as the number of delivered judgements has reached around
1.4 million.52 The previous literature disagrees in its assessment of Gacaca,
but researchers who have recently observed the realities on the ground tend
to be critical of the practice;53 one of these studies has argued that, as a conse-
quence of the trials, Rwandan society has suffered from ‘a tension worsening
social cohesion and attitudes towards the “other group” ’ (Ingelaere, 2009).
The reasons for this were mainly twofold; on the one hand, in the huge
number of Gacaca trials, the relationship between accusers and accused has
always been ethnically fixed – the accusers were Tutsi and the accused were
Hutu; on the other hand, although RPF soldiers also committed atrocities
during the civil war, they have rarely been judged or punished (Human
Rights Watch, 2008). In spite of the Gacaca’s stated objectives of reconcil-
iation and national unity, it may in fact have rather widened the ethnic
divide.
It is highly possible that these three layers of exclusion have resulted in the
marginalisation of the Hutus. In addition to the fact that Hutus are in the
majority in the rural population, which makes it generally difficult for them
to benefit from macro-level economic growth, they have a smaller chance
of being included in the patronage network of political leaders. Moreover, it
is rare for a Hutu to have no family members who stood accused in Gacaca
trials. In short, the Hutus are quite likely to perceive themselves as suffer-
ing from HIs with regard to both the political as well as the socioeconomic
dimensions.54
In the case of examining perceived HIs in Burundi, it is better to sepa-
rate the perception among political elites from that among ordinary people,
because the ethnic power-sharing system has drastically transformed pat-
terns of political antagonism. The institutional change prohibited politicians
from relying on ethnic logic in organising a political party. As a result, Tutsi
elites chose to enter into Hutu-led parties in pursuing chances to become
MPs, and vice versa. Political elites ceased to use ethnicity as a tool for
mobilising the population in power struggles.
Although the Burundian power-sharing system to date has succeeded in
containing ethnic mobilisation, it fails to institutionalise power struggles.
Since 2010 the country has once again been faced with a rebel movement,
because after the election boycott the FNL rearmed and launched sporadic
attacks. Although they once accepted the ceasefire and the integration, they
defected from the security organisations and returned to the bush to fight.55
Political antagonism between two Hutu-led parties is a new development,
but the outlaw features of the power struggle, such as political violence,
oppression and terrorism, have often been witnessed in modern Burundian
history.
Among ordinary Burundians, CNDD-FDD has remained relatively popu-
lar since its election victory in 2005. The results of the Communal election
in 2010, in which CNDD-FDD gained 64 per cent of the members of the
Communal Council, were an indication of its popularity, especially in rural
areas.56 The main reasons for this popularity seemed to be the improvements
in living standards; in addition to the mere fact of the end of war, policies
ensuring free-of-charge access to primary school and health care have par-
ticularly served to garner popular support.57 It is safe to say that the actual
popularity of the CNDD-FDD is not based solely on ethnic preferences.
Nevertheless, the ethnic problem has remained unsolved among the popu-
lation. The most visible evidence for this is the camps for internally displaced
people (IDPs) that are located throughout the country.58 Burundian IDPs are
mainly Tutsis, who were chased out by their Hutu neighbours during the
civil war in the 1990s. Although the fighting has now ended, the IDPs prefer
to stay in the camps for fear of persecution. In everyday life, the legacies
of the civil war always remain a reality;59 ordinary people continue to hold
memories of fear in terms of ethnicity. It may be possible that such fear could
be manipulated for another mobilisation.60
3.7 Conclusion: Implications for conflict prevention
Despite their socioeconomic and geographical similarities and their com-

mon experiences of repeated ethnic conflicts, Rwanda and Burundi adopted
contrasting policies, in terms of ethnicity and power-sharing, after their dev-

astating recent civil wars. From the analyses of HIs, political institutions and
recent political developments, some conclusions on the nature of political
power as well as conflict prevention can be drawn.
In post-conflict Rwanda, political power has been dominated by the
Tutsi-led RPF: the victor in the civil war. Although some appearances of
power-sharing can be found in formal institutions, a number of formal as
well as informal devices serve to maintain the political dominance of the
RPF. Under the power-concentrating system, the RPF has so far succeeded in
guaranteeing security and in promoting rapid economic growth. This suc-
cess, however, has been accompanied by social exclusion, in a way that many
Hutus are likely to resent in terms of both political and socioeconomic HIs.
Overlap in two dimensions of HIs is a dangerous signal for the outbreak of
conflict (Stewart, 2010a). Despite the appearance of stability, a third revolu-
tionary political change in Rwanda’s history would not be unimaginable, if
the power concentration in the hands of the RPF were to deepen the grudge
held by the ethnic majority.
Following the introduction of the ethnic power-sharing system, post-
conflict Burundi has seen a clear change in the nature of its power struggle.
Ethnicity is no longer the determinant of antagonism among political elites,
which is a remarkable and positive change. Nevertheless, during the same
period, Burundi has also seen a concentration of power in the former Hutu-
led rebels, the CNDD-FDD, as well as economic stagnation with rampant
corruption, and the rearmament of the FNL. Although the HIs are tending
to diminish with the new institutional arrangement, especially among elites,
Burundi is always faced with a real threat of political violence; the power
struggle has not yet been institutionalised.
The case of Burundi highlights the possibilities and limitations of inter-
national intervention in conflict prevention. The introduction of ethnic
power-sharing has reduced the danger of ethnic mobilisation. Even if an
institution is introduced on the initiative of external actors, it could play a
decisive role in changing the rules of the game in national politics in a posi-
tive fashion. Nevertheless, it has not yet been able to transform the essential
nature of the politics; the institutionalisation of the power struggle remains
unsolved in Burundi.
Analyses of the two countries show that their institutions reflect long-
term historical experiences. Although the way in which recent wars end
is undoubtedly an important factor for the determination of post-conflict
political institutions, it is not the only factor. In neither country were the
post-conflict political institutions merely the results of the preceding civil
wars. The sustainability and resilience of the institutions depend on the
extent to which they can resonate with endogenous initiatives. It was pos-
sible to introduce the ethnic power-sharing system in Burundi, which is
a country with a long history of implementing similar mechanisms. In a
similar vein, the RPF’s mode of governance reflects not only its total victory
in the war but also Rwanda’s modern history, in which revolutionary polit-
ical change has taken place repeatedly. Institutions for conflict prevention
should therefore be designed taking the historical background into account.
Notes
1. Although no census on the ethnic groups has been carried out in either country,
the approximate proportions of the three groups among their total populations
are said to be almost the same: the Tutsi account for around 15 per cent and the
Hutu 85 per cent, while the proportion of the Twa is less than 1 per cent.
2. As comparative studies between Rwanda and Burundi, Lemarchand’s works (1970,
2006, 2009) are the most notable. Uvin (2010) is also worth mentioning in terms
of research on conflict prevention.
3. The author has continuously conducted field studies since 1999 in Rwanda, and
visited Burundi for research during three months in total, in 2010, 2011 and 2012.
4. The origins of the Tutsi and Hutu have been one of the hottest issues in the
history of the two countries. Although we do not have enough space to fol-
low the debates, it should be stressed that the thesis attributing the origin of
the traditional kingdoms to the migration of and conquest by the Tutsi is seri-
ously questioned today. Recent studies have clarified that the group identities
of Tutsi and Hutu were ambiguous, and their border was blurred in the pre-
colonial period. As representative studies, see C. Newbury (1988), Schoenbrun
(1993, 1998), Chrétien (2000) and D. Newbury (2001).
5. For the importance of the social constructivist view in the analysis of HIs, see
Stewart (2008a: 9–12).
6. The Great Lakes region, including south-western Uganda, Rwanda, Burundi,
north-western Tanzania and part of eastern DRC, was the site of a number of
pre-colonial kingdoms, among which only Rwanda and Burundi became modern
sovereign states. These kingdoms shared a stratified social composition, partic-
ularly between pastoralists and agriculturalists (d’Hertefelt et al., 1962; Maquet,
1971); in the case of Rwanda and Burundi, pastoralist Tutsis took the domi-
nant position over the agriculturalist Hutus, although it should be noted that
the distinction between the two groups has been ambiguous.
7. The concept was characterised by elitism (Vansina, 2001), connected with state
power, as well as pastoralism, their main socioeconomic activity (Nkurikiyimfura,
1994).
8. Batare refers to family members who were descendants of King Ntare Rugamba
(reigned from 1795 to 1852), while Bezi were descended from Mwezi Gisabo
(1852–1908). The Burundian state expanded considerably during the reign of
Ntare, who, in order to consolidate his territorial gains, appointed his sons to
administer newly acquired provinces. As a consequence of this practice, con-
necting dynastic families systematically with politically important positions, the
Ganwa as a group began to play a decisive role in politics. King Mwezi, on the
other hand, tried to remove his predecessor’s influence and appoint his own sons
to politically important positions, thus exacerbating the rivalry between the two
groups (Lemarchand, 1970: 311).
9. In addition to the Ganwa, the ethnic identity of the Hima seemed to be much
more strongly perceived in Burundi than in Rwanda. While the Hima is a group
related to the Tutsis, living in Uganda, Rwanda and Burundi, its uniqueness and
independence as a group have been different depending on the political as well
as regional context. Comparing the description of ‘Hima’ in Dorsey (1994) and
that in Eggers (1997), the latter clearly emphasises its uniqueness. For the Hima
in Burundi, see also Lemarchand (1994: 81–82).
10. A good example on this point is the role of the Bashingantahe, elders who are
in charge of justice in local communities (Laely, 1992; Naniwe-Kaburahe, 2008).
‘Although commonly (though not exclusively) Hutu, they were fully recog-
nised within the Burundi political system in a way unknown – even adamantly
opposed – in Rwanda under the Nyiginya dynasty’ (D. Newbury, 2001: 275).
11. Strictly speaking, the Belgian administration on behalf of the League of Nations
and the United Nations is different from colonial rule. In this chapter, how-
ever, the term ‘colonial’ will be used because of the similarity of power relations
between foreigners and natives.
12. The elimination of Hutus from the administration proceeded following the 1920s
in the process of administration reform, which considerably reduced the number
of chiefdoms and sub-chiefdoms under the name of administrative rationalisa-
tion. In Burundi, all 27 posts for Hutu chiefs were struck off between 1929 and
1945, as a result of rationalising chiefdoms from 133 to 35 (Gahama, 2001: 104).
No chiefdom among the 45 that existed in Rwanda as of 1 November 1959 was
headed by a Hutu chief; only 10 sub-chiefdoms among 559 had Hutu sub-chiefs
(Reyntjens, 1985: 269).
13. The term ‘social revolution’ implies a connotation of social progress, moving from
feudalism (rule by the Tutsi minority) to democracy (rule by the Hutu major-
ity). This terminology has therefore been preferred by Hutu elites, who took
power after independence. In contrast, Tutsi elites tend to avoid the term. The
present Tutsi-led government calls the incident ‘the first massacres of Batutsi’,
website of Government of Rwanda, http://www.gov.rw/page.php?id_article=56
(last accessed 1 May 2013). In this chapter, the term is used within quotation
marks.
14. For detailed studies of the ‘social revolution’, see Lemarchand (1970) and
Reynjens (1985).
15. Michel Micombero, the first president after the overthrow of the monarchy,
was of mixed Tutsi–Hima origin and from a family that did not rank high in
traditional prestige (Eggers, 1997: 84). See also Lemarchand (1970).
16. For the Burundian genocide in 1972, see Chrétien and Dupaquier (2007),
Lemarchand (1994).
17. Examples of the extreme ethnic disparity in this period were shown in Nkurunziza
and Ngaruko (2008: 76); the Tutsi monopolised almost all posts in the public sec-
tor, such as ministers, provincial governors, ambassadors, army members (from
officers to the rank and file), policemen, state-owned company directors and
magistrates.
18. Elites from central Rwanda monopolised the core of political power under the first
president (G. Kaybanda), who was originally from Gitarama (Reyntjens, 1985).
Following the coup, Rwandan politics tended to be dominated by political elites
from the north-western part of the country, which was where Habyarimana and
his wife had come from (Prunier, 1995).
19. Data on Rwandan and Burundian political elites are available in series of annuals
‘L’Afrique des Grands Lacs’ issued by the Centre d’étude de la région des Grands
Lacs d’Afrique. Every annual lists the names of members of the political elite
such as cabinet members, governors, top officers in the army and ambassadors,
with their political party, ethnic affiliation and regions of origin. The data clearly
show that the Tutsis who had been refugees have occupied important posts in
the government. For instance, top officer positions in the Rwandan armed forces
have always been monopolised by the former Tutsi refugees of the RPF, namely,
the former guerrilla fighters.
20. Despite the deepening Tutsi–Hutu divide during the colonial period, Rwagasore
succeeded in including Tutsis as well as the Hutu population in the UPRONA, thus
making it a nationalistic mass party. In addition to his charismatic leadership,
the principle of incorporating equivalent numbers of Hutu and Tutsi members at
every level of the party organisation was crucial to gaining support from ordinary
Hutus (Lemarchand, 1970: 330).
21. Each of the five governments between 1963 and 1965 comprised an almost even
proportion of Hutu and Tutsi ministers (Lemarchand, 1970: 368).
22. Unlike the case of the mass killing in 1972, the international community fiercely
reacted against the incident in 1988 and demanded that the Buyoya government
improve ethnic relations. The government was thus obliged to show some results
of ‘reconciliation’ for outsiders (Lemarchand, 1994: 128–130).
23. For example, among 48 high-ranking members (ministers and cabinet directors)
in the Buyoya cabinet in 1998, 19 were Hutu, 19 were Tutsi and 10 others were
of ethnically unknown origin. In the same year, all top officers in the Burundian
armed forces and the gendarmerie were Tutsi (Marysse and Reyntjens dir. 1999:
385–389).
24. According to the data from the ‘L’Afrique des Grands Lacs’ annuals, ethnic equiv-
alence in the army seemed to have been achieved around the period of the general
elections in 2005.
25. Before the systematic expulsion, nearly half the students in secondary schools
and universities were Tutsi (Munyarugerero, 2003: 134). For the Tutsi expulsion
in 1973, see also Reyntjens (1985: 501–504).
26. At the beginning of the 1990s, both Rwanda and Burundi were among the least
urbanised countries in sub-Saharan Africa. Burundi was in the same position in
2008, with almost 90 per cent of its total population living in rural areas. How-
ever, rapid urbanisation was seen in Rwanda in the same period, as its urban
population rate increased from 5.4 per cent in 1990 to 18.34 per cent in 2008
(World Development Indicators).
27. A tremendous number of Tutsi former refugees returned after the end of the civil
war. Although there have been various estimates of the number of returnees, it
is safe to say that 600,000 to 700,000 former refugees returned (Huggins, 2009:
69; Bruce, 2009: 112). Ansoms (2009) and Silva-Leander (2012) pointed out the
concentration of Tutsi returnees in urban areas. The RPF-led government, which
is eager to involve the Rwandan Diaspora in the development of the national
economy, has adopted several measures to promote this, such as permission
for dual citizenship (Plaza and Ratha eds, 2011). It is highly probable that the
part of the Rwandan Diaspora that has returned to the homeland since 1994 is
overwhelmingly Tutsi.
28. This fact is clearly shown in the evolution of the percentage share of GDP per
quintile of population. While the share of the top 20 per cent quintile has sharply
enlarged, reaching 51.4 per cent in 2000 from 39.1 per cent in 1985, the bottom
20 per cent shrank from 9.7 per cent to 5.4 per cent in the same period. This
means that ‘almost all the growth generated in the last few years has gone to
the top quintile’ (UNDP, 2007a: 19), leaving the average income of the other
four quintiles virtually unchanged since 2003. Rwanda’s Gini coefficient in 2000
was 0.468, considerably aggravated from 0.289 in the mid-1980s (UNDP, 2006,
2007b); it further increased to 0.510 in 2006 (Republic of Rwanda, 2007: 13).
29. According to data from the census held in Rwanda in 2000/01, the proportion
of the population whose annual income or expenditure for consumption was
under the national poverty line (64,000 Rwanda Francs, corresponding to roughly
US$150) was over five times higher in rural areas than that in the capital, Kigali
(République rwandaise, 2002: 33).
30. In the Burundian army, members from Bururi had outnumbered those from other
regions since the 1960s. After the coup in 1966, the army established the Conseil
national de la révolution (CNR) as a supreme decision-making body. In 1968,
among seventeen officers included in the CNR, eight were from the Bururi (five
Tutsi, two Hutu and one Hima) and seven were non-Tutsi (three Hutu, two Hima,
one Ganwa and one so-called ‘Swahili’) (Lemarchand, 1994: 79).
31. Compared with the national average, the school enrolment ratio in Bururi was
1.5 times higher, its number of teachers per classroom was twice as large, and
its population per hospital was only 107,000 in comparison with the national
average of 266,000 (Nkurunziza and Ngaruko, 2008: 73).
32. Bururi’s privileged position in public health was no more obvious in recent
statistics; the population per hospital in Bururi province was 129,663 in 2007
(ISTEEBU, 2009: 132); this figure was the fifth-highest rank among 17 provinces.
However, Bururi seemed to maintain its prominence in secondary schools; its
enrolment rate (21.6 per cent) – the second-highest figure, after only Bujumbura
city (37.7 per cent) – was remarkably higher than other rural provinces, among
which the highest (Makamba) was only 13.1 per cent (République du Burundi,
2006: 62–64).
33. IRIN, 23 May 2003. ‘Rwanda, Government to Start Legal Proceedings against
Banned Party’.
34. For the position of women in post-conflict Rwandan politics, see Burnet (2008).
35. In the case of the Senate, all 26 members are either elected from among or
appointed by the Executive Committees of local administrations, the President
of the Republic, the Forum of Political Organisations, and academia.
36. In Ottaway’s analysis, Eritrea, Ethiopia and Uganda are typical countries under
the rule of the ‘African New Leaders’, and Rwanda and the Democratic Repub-
lic of the Congo are their possible followers. Tendencies towards authoritarian
rule and social engineering can be commonly observed among these countries.
In addition, the ideological influence of Museveni’s Uganda on Kagame’s Rwanda
has been well known (Prunier, 1998).
37. Lijphart (1977). For analyses of the Burundian power-sharing system from the
viewpoint of consociationalism, see Sullivan (2005), Lemarchand (2006) and
Vandeginste (2006, 2008, 2009).
38. This arrangement was in accordance with the argument of Hutu parties dur-
ing the negotiations, whereas Tutsi parties had insisted on counting through
the party base, which means that only members from Tutsi parties should be
considered as representatives of the Tutsi ethnic group.
39. The effective party numbers calculated from the results of the latest elections
are 1.53 for Rwanda and 1.49 for Burundi. As a consequence of Rwanda’s leg-
islative election in 2008, the RPF gained 47 of the 53 seats in the Lower House,
PSD (Parti social-démocratique) 7, and PL (Parti libéral) 4. As mentioned earlier,
in addition to these seats elected through universal suffrage, 27 members were

selected through various administrative organisations. Burundi’s legislative elec-
tion in 2010 resulted in a landslide victory for the CNDD-FDD, accounting for 81
of 104 seats, because the main oppositional parties had boycotted it (see infra.).
For the method of calculating the effective party numbers, see Lijphart (1999: 68).
40. Article 82. Although no mention is made in ethnic terms, the ‘historically
marginalised communities’ clearly include the Twa people.
41. The policy of the ethnic ban was not the RPF’s invention. Burundians had also
seen the same policy under the Bagaza regime (1976–1987). See Lemarchand
(1994: 108).
42. In 1997/98, government forces attacked Hutu militias who had returned to
Rwanda from the DRC in the guise of civilian returnees. During this operation, a
number of Hutu civilians were allegedly slaughtered. An international NGO esti-
mates that at least 6,000 civilians were killed between January and August 1997
(Amnesty International, 1997).
43. During the civil war, the RPA constituted the military wing of the RPF.
44. On the one hand, the RDRC programme, especially its second phase which started
in December 2001, has been generally appreciated by the international com-
munity; the number of members of the Rwandan military was reduced from
80,000 in 2002 to 35,000 in 2008; the assistance for former combatants was
equally distributed without respect to their former affiliation. On the other hand,
the power structure of the Rwandan military has not changed since the end of
the civil war; its core high officer positions have been virtually monopolised by
former guerrilla force (RPA) members. See Takeuchi (2011) for details.
45. Whereas Rwandan traditional export items were almost entirely limited to coffee
and tea, the export of mineral resources such as niobium, tantalum, and vana-
dium has rapidly increased recently. In 2006, they made up 18.2 per cent of total
exports, thus constituting the second-largest export item after coffee (World Bank,
2009: 85). The dependency of Rwandan economy on mineral exports is deepen-
ing, as tin became the largest export item in 2011, accounting for 24.4 per cent
of total exports (Republic of Rwanda, 2012a: 32).
46. Regarding Rwanda’s military interference in the DRC and its illegal exploitation
of mineral resources, see, for example, UNSC (2001, 2002). Marysse and André
(2001) estimated that the benefit that had been acquired by Rwanda through
illegal exports of minerals (coltan, diamond and gold) was equivalent to 7 per cent
and 8 per cent of its GDP in 1999 and 2000, respectively. For recent developments,
see Global Witness (2011).
47. Key posts in the security organisations, such as the Chief of Staff of the Army, the
Chief of the National Intelligence Service, and the Vice-Director of the National
Police, are held by former CNDD-FDD officers. During the author’s field research
in October and November 2010, members of local NGOs all pointed out that the
National Police and the National Intelligence Service were virtually controlled by
the CNDD-FDD. See also Human Rights Watch (2010), Vandeginste (2011) and
‘Burundi: Political violence worsens’, Africa Confidential 52(22): 10.
48. According to the Worldwide Governance Indicators, Burundi’s ‘Control of Cor-
ruption’ indicator remained low during 1998 (−1. 24) and 2009 (−1. 12), while
Rwanda markedly improved from −0. 84 to 0.126 in the same period. Another
NGO survey positioned Burundi as the most corrupt of five East African countries
(Transparency International, 2010). In the Doing Business ranking in 2010, which
ranked Rwanda 58th in the world (the fourth-highest position among African
countries), Burundi was in the 181st position among 183 countries, better than
only the Central African Republic and Chad. In addition, during the author’s field
research, a number of interviewees from the civil society pointed out corruption
has intensified under the rule of the CNDD-FDD, and attributed the fact to the
behaviour of new elites, who were busy profiting from their opportunities for
‘eating’.
49. Marysse et al. (2006) contrasts Rwanda’s ‘aid darling’ status with the DRC’s and
Burundi’s ‘aid orphan’ status. Although the contrast between Rwanda and the
DRC is convincing, that between Rwanda and Burundi is not. Comparing the
average net ODA received per capita from 2000 to 2009, Burundi received US$
44.0, which counted for 72 per cent of Rwanda’s US$ 60.7 (data from WDI). Both
of these figures are not far from the average for sub-Saharan countries during the
same period (US$ 58.5).
50. Due to the sensitiveness of questions concerning ethnicity, large-scale social
inquiries were not carried out in Rwanda and Burundi in the research project.
Therefore, unlike other chapters, the perceptions of HIs will be estimated in this
chapter on the basis of analyses made thus far on the history, institutions, and
recent political processes in each country.
51. See Ingelaere (2010). The voting system through queues has confirmed this ten-
dency. In local elections in Rwanda, voters are requested to stand in a queue
behind their preferred candidates.
52. ‘Open Justice: Q&A with Minister of Justice, Tharcisse Karugarama’ Official
Website of the Republic of Rwanda, http://www.gov.rw/OPEN-JUSTICE-Q-A-with-
Minister-of-Justice-Tharcisse-Karugarama (last accessed 1 May 2013). According
to an official document distributed at the closing ceremony (Republic of Rwanda,
2012b), the total number of cases tried by Gacaca court was 1,958,634.
53. See, for example, Waldorf (2006), Drumbl (2007), Rettig (2008), Ingelaere (2009),
Thomson (2011).
54. Ingelaere (2010) supports this point. According to his survey, subjective political
representation rankings are contrasting between Tutsis and Hutus. Tutsis gener-
ally feel that they were more politically represented after the civil war of the 1990s
than before, but Hutus tend to consider themselves to be in the opposite position.
55. A UN expert panel revealed that FNL had entered the DRC for remobilisation. See
UNSC (2010: paras 113–119).
56. Although the main opposition groups such as FNL and Sahwanya–Frodebu
accused the ruling party of carrying out massive fraud and withdrew from
subsequent elections, the general popularity of the CNDD-FDD seemed to be
undeniable except in several provinces, such as Bujumbura Rural, Bururi, and
Bujumbura Marie. The popularity of FNL was particularly strong in the Province
of Bujumbura Rural (CENI, 2010b).
57. Considering that the execution of these policies will result in overloading gov-
ernmental finance, it is uncertain how long these policies and therefore the
popularity of the ruling party will be sustainable. During the author’s field visit in
2010, the diplomatic corps, international agencies, and the civil society were all
worried about the danger to the national finances that the free-of-charge policies
would create.
58. As of 2009, Burundi had 137 IDP camps, with camps in all 17 provinces; the total
number of IDPs amounted to 157,167 (Rwabahungu and Nintunze, 2009: 9).
59. A civil society activist pointed out that, as a result of the civil war, contacts
with different ethnic groups have considerably diminished, even on ceremonial
occasions such as weddings and funerals (Author’s interview, November 2010,

Bujumbura).
60. Land problems are among the most dangerous issues that can heighten ethnic
tensions. In addition to the land problems of the IDPs, land conflicts caused by
the return of Hutu refugees are also serious. The end of the civil war encouraged
a huge number of Burundian refugees, who escaped the country in the 1960s and
1970s, to return to their homeland. Many of them, however, found that their
land had been occupied (ICG, 2003). It is said that there would be considerable
numbers of Tutsi migrants among the occupiers, who had been sent there under
the one-party regime in the 1970s and 1980s. There is a danger that the land
conflicts between Hutu returnees and Tutsi occupiers have the potential to be
ethnically manipulated.
4
Horizontal Inequalities, Ethnic
Politics and Violent Conflict: The
Contrasting Experiences of Ghana
and Côte d’Ivoire
Arnim Langer
4.1 Introduction
Although the number of violent conflicts and civil wars has been on the
wane since the mid-1990s, today there are still around 30 ongoing violent
intra-state conflicts around the world. The majority of these consist of fight-
ing between groups who are united by a common ethnic or religious identity.
Since 1945, the identity basis of conflicts has become increasingly explicit,
with the proportion of all conflicts that are labelled as ‘ethnic’ increasing
from 15 per cent in 1953 to nearly 60 per cent by 2005 (Stewart and Brown,
2007). In the 1990s, politicians, journalists and intellectuals propagated the
popular view that the breakdown of multiethnic societies and the emergence
of ethnic conflicts, in both Africa and Eastern Europe, was due to the ‘erup-
tion of ancient and irrational tribal antagonisms’ (Turton, 1997: 80). Since
then, and despite the fact that a substantial amount of research has shown
the fallibility of these claims, these views remain prominent in non-scholarly
circles of society.
Indeed, ‘primordial’ ethnic hatred or, for that matter, cultural differences
are generally an insufficient explanation for the emergence of violent con-
flicts. As Abner Cohen aptly argued: ‘Men may and do certainly joke about or
ridicule the strange and bizarre customs of men from other ethnic groups,
because these customs are different from their own. But they do not fight
over such differences alone. When men do, on the other hand, fight across
ethnic lines it is nearly always the case that they fight over some funda-
mental issues concerning the distribution and exercise of power, whether
economic, political, or both’ (Cohen, 1974: 94). In the same vein, Ted Gurr
argued in his well-known book Why Men Rebel (see Gurr, 1970) that a com-
bination of deprivation-induced discontent and a sense of group cultural
66
Arnim Langer 67
identity is the primary determinant for political mobilisation, both violent

and otherwise.
In recent years, Gurr’s relative deprivation theory has been redeveloped
by Frances Stewart (see Stewart, 2000, 2002). She argues that the existence
of severe political and/or economic inequalities between culturally defined
groups, such as ethnic, religious or racial groups – which she calls hor-
izontal inequalities (HIs) – might cause a range of political disturbances,
including violent conflict and civil war. Yet, although it is argued that the
presence of severe ethnic inequalities makes countries more vulnerable to
violent conflict, some countries with severe social and economic inequali-
ties between different ethnic groups have nonetheless remained relatively
peaceful and stable. This raises a fundamental question of the causal mech-
anism between the presence of HIs and the emergence of violent conflicts.
Moreover, although HIs may be associated with higher levels of conflict, we
need a more specific theory of the mechanisms that lead from HIs to conflict
in some cases, and not in others.
However, before exploring in more detail how and when the presence of
HIs may ‘cause’ or contribute to the emergence of violent conflict, it is first
crucial to discuss the ‘problematic nature of ethnicity’ (Nagel, 1974: 154).
While there is no universally accepted definition of the concept, ethnicity
is generally characterised as a sense of group belonging, based on ideas of
common origins, history, culture, language, experience and values (see, for
example, Glazer et al., 1976; Anderson, 1983; Horowitz, 1985; Bates, 2004).
In addition to real or putative common descent, most definitions of eth-
nicity emphasise ‘the sharing of a “culture”, the most notable aspect of
which is language’ (Bates, 2004: 5). Since the early 1980s, the construc-
tivist perspective has become the most influential perspective in the study
of ethnicity. As Robert Bates argues, the constructivist approach to ethnicity
combines elements of both the primordial and instrumentalist approaches:
‘In keeping with the primordialists, constructivists view ethnic identities as
a cultural endowment; but in keeping with instrumentalists, they view eth-
nic identities as malleable. Distinguishing their position is the belief that
while identities can be reshaped, they can be altered only at significant cost’
(Bates, 2004: 167). Although ethnic identities may be constructed, once cre-
ated, they tend to endure and can arouse deep attachments from the people
involved.
While some scholars, most notably anthropologists and historians, are
sceptical about using ethnicity as an explanatory variable, an increasing
number of political scientists and economists have started to incorporate
ethnic groups into their (empirical) analyses or use ethnic diversity as an
independent variable for explaining a range of social and economic out-
comes, most notably relating to violent group mobilisation, secessionist
movements and economic growth differences (Brown and Langer, 2010).
The main points of disagreement between these two loosely defined groups
68 Ghana and Côte d’Ivoire
of academics relate to the fluidity of ethnic group boundaries and the

extent to which an ethnic group can be essentialised by a number of
observable ‘group’ characteristics (ibid.). While acknowledging the histori-
cal constructed-ness and fluidity of ethnic groups, I argue that at any point
in time ethnic group differences are often sufficiently stable and clear to the
people involved to enable us to identify inequalities across ethnic groups.
Let us now turn to the hypothesised relationship between HIs and violent
conflicts. An increasing amount of empirical research has found a signifi-
cant association between the presence of HIs and the risk of violent conflict
(see, for example, Brown, 2008; Østby, 2008; Stewart ed., 2008; Cederman
et al., 2010). There are, however, numerous cases where sharp socioeco-
nomic inequalities have not resulted in ‘ethnic’ violence or other serious
political disturbances.1 Starting from this observation, in previous work (see
Langer, 2005) I developed a theory of the causal mechanism between HIs and
violent conflict, which required a more nuanced analysis of HIs. In particu-
lar, I argued that ‘ethnic’ violence is more likely to occur when ‘elites’ and
‘masses’ both have strong incentives to mobilise along ethnic lines, and this
is likely to be the case in situations where there are both severe political and
socioeconomic HIs.
However, further theoretical refinement is needed to explain the process
through which political and economic HIs may – or may not – result in
violent conflict. In this respect I argue that the main ‘intervening variable’
linking HIs and violent conflict is the nature – or changing nature – of pol-
itics. My main argument here is that HIs may contribute directly to the
ethnicisation of the political environment or the emergence of ‘ethnic pol-
itics’ (defined here as a more or less explicit mobilisation on the basis of
ethnic group distinctions), which in turn increases the risk of violent con-
flict, as argued by a range of political scientists (see, for example, Horowitz,
1985; Fearon and Laitin, 1996; Glickman, 1998; Varshney, 2002). By focus-
ing on the intervening variable of ethnic politics, the relationship between
HIs and violent conflict becomes more complex and the possible outcome
scenarios more diverse. Instead of just violent conflict or no violent conflict,
as a result of the reconceptualisation of the relationship between HIs and
violent conflict, there are now at least three possible outcome scenarios to
be considered: that is, the first scenario is where the presence of HIs does not
result in a notable ethnicisation of politics; the second scenario is where the
presence of HIs results in a marked ethnicisation of politics, but no violent
escalation occurs as a result; and the third scenario is where the presence
of HIs leads to a serious ethnicisation of politics which contributes subse-
quently to the emergence of a violent conflict along ‘ethnic’ lines. So far
this complex link between HIs and the emergence of ethnic politics has
been neglected in the literature, despite the fact that it may be crucial for
understanding when and why HIs lead to violent group mobilisation along
ethnic lines.
Arnim Langer 69
In order to explore the hypothesised relationship between HIs and vio-

lent conflict via the intervening variable of ethnic politics, this chapter aims
to analyse and compare the contrasting sociopolitical experiences of Ghana
and Côte d’Ivoire. While both countries are confronted with severe socioeco-
nomic HIs (in particular, between their northern and southern regions and
ethnic groups), only Côte d’Ivoire was confronted with the outbreak of a vio-
lent conflict, in 2002, with a clear north–south dimension. In Ghana, on the
other hand, despite the coincidence of major ethnic, social and economic
inequalities and differences between the northern and southern regions,
‘there has been no development of “Northernness” as a basis for politi-
cal cohesion, and no north versus south patterning of political alignments’
(Brown, 1982: 42). While the north–south divide has had a very limited
impact on Ghanaian politics, its postcolonial political history is nonetheless
characterised by political instability and repeated non-constitutional regime
changes, and also, at times, serious ethnic tensions. These ethnic tensions,
however, have never resulted in any serious violence at the national level.
A Ghana–Côte d’Ivoire study offers a strong foundation for comparison
because of the large number of structural similarities between these two
countries, including population size, location, geography and climate, vari-
ety and distribution of ethnic groups, regional developmental inequalities,
economic structure and level of development; there are, however, a num-
ber of important differences, including the relative weight of the northern
and Muslim populations in both countries, the number of foreigners who
live in each country, both countries’ postcolonial economic performance
records and their different colonial histories. When analysing and compar-
ing Ghana and Côte d’Ivoire’s different sociopolitical outcomes, we need
to carefully take into account both countries’ historical, political and social
idiosyncrasies.
The chapter proceeds as follows. Before analysing how Ghana was able to
peacefully manage its ethnic inequalities and keep ethnic politics in check at
the national level, I will analyse the ethnicisation of Côte d’Ivoire’s political
system in the 1990s and the subsequent descent into violent conflict. In the
last section I will draw some conclusions regarding the linkages between HIs,
ethnic politics and violent conflict in Ghana and Côte d’Ivoire.
4.2 Côte d’Ivoire: From Ivorian miracle to civil war
In light of its economic and political achievements, international observers

often referred to Côte d’Ivoire in the 1960s and 1970s as an ‘oasis of peace’
and an ‘African miracle’.2 However, at the beginning of the 1980s, the
Ivorian success ‘model’ slowly started to unravel, first economically, and
then politically. The lowest point of Côte d’Ivoire’s political and economic
demise was the outbreak of a violent conflict in September 2002 with clear
ethno-regional undercurrents. This conflict began on 19 September 2002
when a group of soldiers with predominantly northern origins attempted

to overthrow the regime of President Laurent Gbagbo. Although the coup
d’état failed to remove the Gbagbo regime, by the end of September 2002
the insurgents were firmly in control of the northern part of the coun-
try and referred to themselves as the Mouvement Patriotique pour la Côte
d’Ivoire (MPCI). Before analysing the underlying causes of Côte d’Ivoire’s
violent disintegration, I will briefly discuss the country’s ethno-religious
composition.
4.2.1 Côte d’Ivoire’s ethno-religious demography

Côte d’Ivoire is a multiethnic country with approximately 40 different
ethnic groups that can be grouped into five larger sociocultural or ethno-
linguistic groups: Akan, Krou, Northern Mandé, Southern Mandé and
Voltaic. Whereas the largest ethnic group is the Akan, with approximately
42 per cent of the population, the two northern ethnic groups, Northern
Mandé and Voltaic, together constitute about 34 per cent of the popula-
tion (see Table 4.1). Although the latter two ethnic groups originate from
Côte d’Ivoire’s northern regions, due to extensive north–south migration
in both the colonial and postcolonial periods, many people belonging to
these groups now live in the southern regions. In addition to these inter-
nal migration flows, Côte d’Ivoire has also received a very large number of
international migrants on its territory. This process started in the colonial
era when France transferred large numbers of people from the Upper Volta,
today’s Burkina Faso, to Côte d’Ivoire, where they were forced to work on
the cocoa and coffee plantations in the southern parts of the country.
Although the French Assembly had abolished forced labour in 1946,
Côte d’Ivoire continued to attract large numbers of migrants from neigh-
bouring countries. The country’s first president, Félix Houphouët-Boigny,
promoted the influx of foreign workers by introducing liberal landownership
laws, under the slogan ‘the land belongs to those that develop it’ (Gonin,
Table 4.1 Ethnic composition of the Ivorian population,

1975–1998 (%)
1975 1988 1998
Akan 42.5 41.8 42.1

Krou 15.9 14.6 12.7
Southern Mandé 12.0 10.7 10.0
Northern Mandé 13.6 15.9 16.5
Voltaic 15.4 16.3 17.6
Note: The ethnic composition shown here only takes into account
Ivorian nationals.
Source: République de Côte d’Ivoire (2001: 68).
Arnim Langer 71
Table 4.2 Countries of origin of the foreign population in Côte d’Ivoire
Country of origin 1988 1998
Population % Population %
Burkina Faso 1,564,650 51.5 2,238,548 56.0

Mali 711,135 23.4 792,258 19.8
Guinea 225,845 7.4 230,387 5.7
Ghana 167,783 5.5 133,221 3.3
Other countries 369,624 12.1 605,633 15.1
Total 3,039,037 100.0 4,000,047 100.0
1998: 174). As a result, the origin of a large proportion of the people in

Côte d’Ivoire, in both the current and previous generations, is from out-
side the country. In 1998, such ‘foreigners’ accounted for over four million
people, or roughly 25 per cent of the population (see Table 4.2). However,
about 50 per cent of these ‘foreigners’ or ‘non-Ivorians’ were born in Côte
d’Ivoire. Importantly, the ethno-cultural and religious background of these
non-Ivorians is very similar to that of the northern ethnic groups.
Under the Houphouët-Boigny regime (1960–1993), foreigners had the
right to vote and most of them strongly supported the ruling party, the
Parti Démocratique de la Côte d’Ivoire (PDCI), because of their gratitude
and loyalty to the president: the man who had welcomed them to Côte
d’Ivoire. Given the absence of competitive elections until 1990, the ‘for-
eign factor’ played a relatively small role in Ivorian politics; however, as
will be discussed later, this changed dramatically after the reintroduction
of multi-party elections in 1990.
Religion is another important dividing line in Côte d’Ivoire. The two
largest religious groups (Muslims and Christians) are largely comparable in
size (see Table 4.3). Among Ivorian nationals, Christians form the largest
religious group (with about 34 per cent of the population), while 28 per cent
of the Ivorian population is Muslim. However, because the vast majority of
Table 4.3 Religious composition of Côte d’Ivoire in 1998 (%)
Only Ivorian nationals Including foreigners
Christians 33.9 30.4

Muslims 27.5 38.6
Traditionalists 15.3 11.9
No religion 20.7 16.7
Source: Data based on République de Côte d’Ivoire (2001: 60).

Table 4.4 Religion of different ethnic groups in Côte d’Ivoire in 1998 (%)
Religion Akan Krou N. Mandé S. Mandé Voltaic
Christians 51.5 52.2 1.4 23.1 16.4

Muslims 5.0 2.6 95.6 5.6 44.7
Traditionalists 18.7 7.5 0.6 30.1 19.2
Without religion 21.7 31.7 1.9 38.4 18.7
Note: The data shown here only take into account Ivorian nationals.
non-Ivorians are Muslim (about 70 per cent), their presence in Côte d’Ivoire
tilts the religious balance in favour of Muslims at the national level. The
Muslim/Christian balance becomes 39/30 if foreigners are included.
Religious differences appear to reinforce ethno-regional differences.
Table 4.4 shows the religious composition of the five major ethnic groups.
While the Akan and Krou ethnic groups are predominantly Christian, the
northern ethnic groups, Voltaic and Northern Mandé, are mostly Muslim.
Consequently, almost 50 per cent of the people in the northern part of the
country (comprising the Savanes, Vallée du Bandama, Zanzan, Denguélé,
Worodougou and Bafing regions) are Muslim. This percentage increases to
63 per cent if one excludes the most southerly northern region, Vallée
du Bandama. Although the northern population is predominantly Muslim,
about 70 per cent of all Muslims live in the southern regions of the country.
In 1998, if one includes foreign nationals, Muslims were the largest reli-
gious group in the southern part of the country with about 35 per cent of
the population, compared to 34 per cent who are recorded as Christians.
Excluding foreign nationals, however, Christians were the largest religious
group. Although the general perception in Côte d’Ivoire and in the interna-
tional media is that the north is basically Muslim and the south is Christian,
these figures indicate that the actual situation is somewhat more nuanced
and complicated.
4.2.2 The Houphouët-Boigny era (1960–1993)

When Côte d’Ivoire became independent in August 1960, it adopted a one-
party system. The PDCI was founded by the Baoulé tribal chief Houphouët-
Boigny in 1946 and exercised de facto control of the Ivorian political
system between 1960 and 1999. Houphouët-Boigny was elected the first
president of Côte d’Ivoire and he remained in power until his death in
December 1993. During the first 20 years of his presidency, Côte d’Ivoire
achieved remarkable economic growth, with real annual GDP growth rates
of more than 7 per cent. In addition to its strong economic progress, Côte
d’Ivoire also benefited from a relatively stable political environment in
these years. As mentioned earlier, because of its impressive economic and
Arnim Langer 73
political achievements, Côte d’Ivoire was often referred to as ‘Le Miracle

Africain’.
While the favourable economic environment contributed heavily to Côte
d’Ivoire’s relatively stable political environment, other factors also played
a crucial role. Some scholars have stressed the importance of Houphouët-
Boigny’s approach to politics, which was characterised by a culture of
dialogue, compromise, rewards, punishment, forgiveness and reintegration
(see, for example, Akindès, 2004). Another crucial aspect of Houphouët-
Boigny’s approach – or what Akindès (2004) has termed ‘Le modèle
Houphouétiste’ – was his use of economic incentives to co-opt actual and
prospective political challengers into the system (Zartman and Delgado,
1984). The robustness of the economy provided sufficient resources for
Houphouët-Boigny’s patronage system to defuse most sources of discon-
tent (Gyimah-Boadi and Daddieh, 1999). Houphouët-Boigny was, however,
uncompromising about the need to maintain order and stability in order to
secure national economic development (ibid.). His willingness to use force
and repression in order to secure such order and stability was demonstrated
on several occasions, most notably during the secessionist revolt of the
Sanwi King in December 1969 and also during the Guébié crisis in November
1970. Another factor that contributed to maintaining political stability and
keeping ethnic politics in check was Houphouët-Boigny’s ‘system of ethnic
quotas’, which was aimed at establishing a balance between different regions
and ethnic groups within the main state institutions (Bakary, 1984).
Côte d’Ivoire’s outward-oriented agricultural development strategy (with
cocoa and coffee as its two main export crops) produced impressive eco-
nomic results in the 1960s, but the ‘model’ had an endogenous tendency
to favour the southern areas over the northern areas: whereas most plan-
tations and other natural resources were located in the south, subsistence
farming was the main economic activity in the north. From the late 1960s,
however, the Ivorian government started to promote commercial food pro-
duction in the north in order to reduce food imports, which had constituted
a serious drain on the country’s foreign currency reserves (Hinderink and
Tempelman, 1979). The more active role of the Ivorian state in the north-
ern economy helped to reduce the socioeconomic inequalities between the
north and south in the period from 1965 to 1975 (see Table 4.5).
Despite the reduction of inequalities between the north and south, the
north–south divide remained severe. In 1975, for instance, income per capita
in the north was about 22 per cent below the national average and as much
as 65 per cent below the level in Abidjan. The continuation of these sharp
inequalities increasingly began to threaten Côte d’Ivoire’s ethno-regional
harmony (Gyimah-Boadi and Daddieh, 1999).
In response to the increasing discontent of the people in the northern
regions regarding their relative socioeconomic situation, Houphouët-Boigny
made several highly publicised visits to the north in 1974. During these
Table 4.5 Income per capita in 1965 and 1975 (constant 1965 CFA franc)
Monetary income per capita Total income (including

auto-consumption)
1965 1975 % change 1965 1975 % change
Abidjan 33,600 47,100 40.2 60,500 66,000 9.1

South 19,500 26,000 33.3 31,500 40,100 27.3
Central West 10,000 18,300 83.0 19,000 31,700 66.8
West 5,800 9,000 55.2 13,400 17,900 33.6
North 3,800 8,900 134.2 16,000 22,800 42.5
Centre 12,800 13,600 6.3 28,100 30,100 7.1
East 10,800 9,400 −13.0 23,400 24,300 3.8
Southwest 6,500 8,800 35.4 15,300 17,800 16.3
Côte d’Ivoire 11,000 15,300 39.1 23,200 29,200 25.9
Source: Bresson (1980: 78).
visits, he promised the local population increased public investment in order

to attain equality with the south. The president fulfilled his promise by ini-
tiating the ‘Programme du Nord’, which allocated about CFAF20 billion to
investment programmes in the northern and central regions (Den Tuinder,
1978). Another measure to mitigate the ethno-regional imbalances in the
distribution of social services was ‘to alternate Ivorian independence festiv-
ities between Abidjan and the different prefecture capitals’ (Gyimah-Boadi
and Daddieh, 1999: 137). The massive facelifts that these capitals would
undergo in preparation for this event created a considerable number of jobs.
Thus, economic redistribution, both actual and symbolic, was an important
strategy to mitigate ethnic discontent and keep a lid on ethnic politics in
this period.
However, the increase in public investment in the northern regions
quickly dried up with the deteriorating economic environment at the end of
the 1970s. The sharp decline in the prices of commodities such as coffee and
cocoa clearly exposed Côte d’Ivoire’s vulnerability to international markets.
Throughout the 1980s, the economy was stagnant and the socioeconomic
north–south divide remained as severe as in the mid-1970s.
The negative economic environment in the 1980s not only substantially
reduced the standard of living in the country but also had important polit-
ical consequences. Due to the sharp decline in government revenue, the
Houphouët-Boigny regime was no longer able to provide cosy state jobs
to large numbers of university students, which in turn led to serious stu-
dent protests. In other words, Houphouët-Boigny’s patronage system was
basically undermined by a lack of resources. The economic problems also
exacerbated tensions between locals and foreign migrants, as well as between
internal migrants from the north and locals in the southern regions. As most
Arnim Langer 75
internal and international migrants belonged to the northern ethnic groups

and were Muslim, the communal tensions were increasingly perceived as a
conflict between north and south (Dembélé, 2003). As Dembélé (2003: 36,
my translation) argues, ‘the communal conflict between north and south
was mainly related to land issues and the presence of too many migrants
from the centre and north in the rural economy in the south-western regions
and the urban economy in the south’.
In April 1990, the economic crisis resulted in major demonstrations by
the still officially illegal political opposition. In an attempt to restore social
and political stability, in May 1990 Houphouët-Boigny decided to abandon
one-party rule and legalise opposition parties. The first competitive presi-
dential elections took place in October of the same year. Houphouët-Boigny
won the elections by a considerable margin against the main opposition
party candidate, Laurent Gbagbo. However, the most significant aspect of
these elections was the introduction of ethno-nationalism and xenophobia
into Côte d’Ivoire’s electoral politics. In particular, during the 1990 elections,
Côte d’Ivoire’s main opposition party, Front Populaire Ivoirien (FPI, Ivorian
Popular Front), initiated a political campaign around the message that ‘the
PDCI was a partial regime which had systematically favoured the interests of
particular Ivorian ethnic groups – Baoulé and groups from the north – and of
foreigners’ (Crook, 1997: 222). While ethnicity had occasionally been used
in the Houphouët-Boigny era as a way of naming and shaming certain indi-
viduals and groups that were going against the will of the regime in charge,
it had never been used openly and explicitly to mobilise people in an elec-
toral contest. This episode is further noteworthy because it shows that the
introduction of perceptions of HIs into the political discourse can directly
lead to an ethnicisation of politics and increased ethnic tensions, as argued
in Section 4.1.
One important change that took place in the wake of these elections was
the appointment of Alassane Ouattara – a Malinké, a subgroup of the North-
ern Mandé ethnic group – to the newly created position of prime minister.
A former African director at the International Monetary Fund (IMF) and Gov-
ernor of the Central Bank for West African States (BCEAO), Ouattara was
chosen mainly for his economic management skills and international rep-
utation. However, by appointing Ouattara as prime minister, ‘the conflicts
between the forest people from the south and the northerners in the land
and economic sphere shifted to the political sphere’ (Dembélé, 2003: 36, my
translation). Although hardly surprising given his old age (he was 93 years
old), the death of Houphouët-Boigny on 7 December 1993 was a watershed
moment in Ivorian history, because it essentially unleashed a power struggle
between the new political leaders that contributed substantially to the dis-
integration of the Ivorian state a decade later. In accordance with the 1960
Constitution, Henri Konan Bédié, a Baoulé, succeeded Houphouët-Boigny
for the remainder of his presidential term.
4.2.3 The politicisation of ethnicity and the descent into civil war
Although the combination of several years of structural reforms and the
1994 CFA franc devaluation led to a significant recovery in economic
growth, most people did not benefit from it (see, for example, Azam, 2004).
Côte d’Ivoire also continued to experience serious political confrontations
and ethnic tensions, especially in the lead-up to the October 1995 pres-
idential elections. A growing northern consciousness, combined with an
increased dissatisfaction among the northerners towards their (perceived)
position in society, was an important change that contributed to the esca-
lation of ethnic tensions at the beginning of the 1990s. The distribution
of an anonymous document called ‘Le Charte du Grand Nord’ (Charter
of the North) in 1992 illustrated the changed attitudes of the northerners
regarding the sociopolitical system in general and the Baoulé group in par-
ticular. The Charter ‘called for fuller recognition of the Muslim religion [ . . . ],
more efforts to reduce regional inequalities, greater political recognition of
the north’s political loyalty during the upheavals of the 1980s and [ . . . ] an
end to Baoulé nepotism in recruitment to public jobs’ (quoted in Crook,
1997: 226). Northern grievances and dissatisfaction were not limited to the
economic and political sphere, but also had a ‘cultural status’ or religious
dimension. The call for greater recognition of the Muslim religion in Côte
d’Ivoire clearly illustrates this. Côte d’Ivoire’s 1960 constitution was secular
in nature, but as a direct consequence of Houphouët-Boigny’s long stay in
power as head of state, there was a growing blurring in perceptions of the
separation of religion and state. Although the political inclusion of people
from the north in various state institutions helped to counter this impres-
sion, the creation of an immense Basilica in Yamoussoukro by Houphouët-
Boigny in the late 1980s was perceived by many Muslims as a clear indication
of the superior position allocated to Christianity (particularly Catholicism)
in Côte d’Ivoire.
The emergence of a new opposition party, Rassemblement des républi-
cains (RDR), in 1994, reflected a further split in Côte d’Ivoire’s political
elite. The RDR aimed to draw support from people with a northern and/or
Muslim background, predominantly found among the Voltaic and Northern
Mandé ethnic groups. Alassane Ouattara – who in 1994 was again working
in Washington – would soon become their political leader. The emergence of
this new party confronted President Bédié with a serious challenge because
the RDR was likely to reduce PDCI’s electoral support in the northern regions
(Crook, 1997). In response, as Richard Crook notes, ‘Bédié’s initial strategy
was familiar to any student of electoral politics: he stole the opposition’s
clothes, and adopted a policy of Ivorian nationalism, under the slogan of
the promotion of “Ivoirité” (Ivorianness)’ (ibid.: 227).
Although Bédié claimed that the sole aim of the concept of Ivoirité was to
create a sense of cultural unity among all the people living in the territory of
Arnim Langer 77
Côte d’Ivoire, it is widely recognised that it was introduced for specific polit-
ical reasons. First, in contrast to Houphouët-Boigny, Bédié could not count
on the loyalty of the foreign migrants, and he was therefore very concerned
that they might switch their allegiance to the RDR because of their ethno-
cultural and religious affinities with its putative leader, Alassane Ouattara.
Bédié therefore withdrew foreigners’ right to vote once the RDR emerged on
the political scene in October 1994.
Secondly, Bédié also used the concept of Ivoirité to prevent Alassane
Ouattara (the RDR leader with northern origins) from participating in the
presidential elections in 1995. Indeed, in line with the ideology of Ivoirité,
the Electoral Code was amended in December 1994. From then on, individu-
als were only allowed to stand in presidential and parliamentary elections if
both their parents were Ivorian by birth. The revised 1995 Electoral Code fur-
ther stipulated that the candidate himself must have lived in the country in
the five years preceding the elections, and never have used another nation-
ality than that of Côte d’Ivoire. On the basis of the revised Electoral Code,
Bédié and his supporters claimed that Ouattara was ineligible to participate
in the presidential elections because his father was a Burkinabe. Ouattara,
however, countered that he was Ivorian and that he was a descendant of
the royal family of the ancient Kingdom of Kong (situated in the north-
eastern part of Côte d’Ivoire). Ouattara was also accused of having possessed
a Burkinabe passport when he was Governor of the Central Bank for West
African States (BCEAO). In response to this allegation, Ouattara stated that
he only possessed a diplomatic passport and that he never had Burkinabe
nationality. Ouattara was also ineligible according Bédié supporters because
he had lived in the United States since his resignation as Prime Minister
in 1993. However, the Supreme Court (which was responsible for judging
the eligibility of presidential candidates) did not have to proclaim Ouattara’s
non-eligibility officially, as he announced in September 1995 that he would
not participate in the presidential elections.
However, the introduction of the ideology of Ivoirité had an impact far
beyond the political sphere, because it led to a general erosion of northern
Ivorians’ social standing and cultural status, making them de facto second-
class citizens in Côte d’Ivoire.
As a result of Ouattara’s exclusion, the RDR boycotted the October 1995
presidential elections. FPI leader Laurent Gbagbo also decided to boycott the
elections, claiming that the electoral process had been manipulated. Due
to the absence of his main rivals, Bédié achieved a landslide victory in the
October 1995 elections. Until the coup d’état in December 1999, the two
opposition parties, RDR and FPI, together formed the Front Républicain.
In sharp contrast to Houphouët-Boigny, Bédié largely stopped the efforts to
balance the different ethno-regional interests and parties, and increasingly
started to favour people from his own ethnic group, the Baoulé (Dozon,
2000). Importantly, the ‘baoulisation’ of the Ivorian state institutions was
both a cause for and consequence of the increased ethnicisation of the polit-
ical system under Bédié. The ‘baoulisation’ of the Ivorian state in the 1990s
is illustrated in Table 4.6, which shows the ethnic composition of Ivorian
governments in the period 1980–2003. Each group’s relative representation
is calculated by dividing its percentage proportion of government minis-
ters by its percentage size in the entire population. Consequently, unity
means proportional representation; figures higher than one point to over-
representation and less than one to underrepresentation. The data show that
Bédié clearly favoured his own group, the Baoulé, with respect to ministerial
appointments.
In addition to the baoulisation of the political-administrative sector, Bédié
also began to change the ethnic composition of the military forces in favour
of his own ethnic group (Contamin and Losch, 2000). The ethnic tensions
that stemmed from increased favouritism towards the Baoulé were com-
pounded by general discontent in the armed forces due to a gradual decline
in their status during the 1990s, mainly arising from reduced expenditures
due to the precarious financial situation (Kieffer, 2000).
The grievances within the armed forces triggered a coup d’état in
December 1999, initiated by a group of non-commissioned officers who
claimed they were owed financial compensation for their participation in
an international peacekeeping mission in Central Africa (Kieffer, 2000). This
protest movement quickly developed into a large-scale mutiny, at which
stage more senior officers got involved. Although the coup d’état appears
to have initially originated in individual grievances, these grievances and
fears of exclusion cannot be separated from what was happening in the rest
of society.
As Kieffer argues, the opposition parties’ discourse of exclusion and Baoulé
domination of the Ivorian state is likely to have had an important impact on
the attitudes of the young non-commissioned officers involved in the coup
d’état (Kieffer, 2000). Importantly, at the time of the coup d’état in December
1999, both the Baoulé overrepresentation and the underrepresentation of
the northerners in government were the most severe in Côte d’Ivoire’s
postcolonial history up to that point (see Table 4.6; and, Langer, 2005).
Following Bédié’s removal from power, the military forces established the
Comité national de salut public (CNSP), headed by General Gueï, who was
a Yacouba, one of the ethnic groups belonging to the Southern Mandé eth-
nic group. In line with ‘Le modèle Houphouétiste’, Gueï initially promoted
the ideals of national integration and reconciliation, and openly opposed
the ideology of Ivoirité (Akindès, 2004). After negotiations between the dif-
ferent political parties and the military junta, a transitional government was
installed on 4 January 2000. As Table 4.6 shows, this transitional government
had a very inclusive character with a reasonably fair distribution of power
among the major ethno-regional groups. However, after several months in
office, Gueï’s political objectives and strategy changed drastically. In contrast
Table 4.6 Ethnic composition of government, 1980–2003
Félix Houphouët-Boigny Henri Konan Bédié Robert Gueï Laurent Gbagbo
Nov-80 Jul-86 Oct-89 Nov-91 Dec-93 Jan-96 Aug-98 Jan-00 May-00 Jan-01 Aug-02 Sep-03
% RRa % RR % RR % RR % RR % RR % RR % RR % RR % RR % RR % RR
Akan 0.49 1.15 0.41 0.98 0.47 1.12 0.61 1.46 0.52 1.24 0.52 1.24 0.59 1.40 0.50 1.19 0.30 0.71 0.46 1.09 0.52 1.24 0.40 0.95
Baoulé 0.22 1.32 0.24 1.44 0.20 1.20 0.17 1.02 0.24 1.44 0.28 1.68 0.31 1.86 0.13 0.78 0.04 0.24 0.11 0.66 0.13 0.78 0.09 0.54
Krou 0.19 1.19 0.20 1.37 0.20 1.37 0.17 1.16 0.24 1.89 0.21 1.65 0.16 1.26 0.13 1.02 0.22 1.73 0.29 2.28 0.19 1.50 0.23 1.81
S. Mandé 0.05 0.42 0.10 0.93 0.13 1.21 0.04 0.37 0.04 0.40 0.10 1.00 0.06 0.60 0.08 0.80 0.17 1.70 0.18 1.80 0.16 1.60 0.07 0.70
N. Mandé 0.08 0.59 0.17 1.07 0.13 0.82 0.09 0.57 0.08 0.48 0.07 0.42 0.03 0.18 0.17 1.03 0.17 1.03 0.07 0.42 0.13 0.79 0.19 1.15
Voltaic 0.14 0.91 0.10 0.61 0.03 0.18 0.09 0.55 0.12 0.68 0.10 0.57 0.13 0.74 0.13 0.74 0.08 0.45 0.00 0.00 0.00 0.00 0.12 0.68
Southernersb 1.04 1.06 1.19 1.22 1.23 1.28 1.25 1.10 1.06 1.44 1.34 1.08
Northerners 0.76 0.84 0.50 0.56 0.59 0.50 0.47 0.88 0.73 0.21 0.38 0.91
Northerners, 0.56 0.69 0.41 0.46 0.49 0.42 0.39 0.74 0.62 0.17 0.32 0.76
incl. ‘Foreign
Ivorians’
Ratio N/S 0.73 0.79 0.42 0.46 0.48 0.39 0.38 0.80 0.69 0.14 0.28 0.84
Total N = 37 N = 41 N = 30 N = 23 N = 25 N = 29 N = 32 N = 24 N = 23 N = 28 N = 31 N = 43
Notes: Government: president of the republic, ministers of state and regular ministers. Deputy ministers were not included in the calculations.
a. Relative Representation (RR) is calculated by dividing an ethnic group’s relative proportion in government by its relative size in the entire population.
RR figures with respect to the November 1980 government were based on the ethnic data from the 1975 census. From July 1986 until November 1991, the
ethnic demography data from the 1988 census were used, and from then onwards we have used the 1998 census data to calculate the RR figures.
b. The Northern Mandé and Voltaic ethnic groups together make up the group of ‘northerners’, while the three other groups (the Akan, Krou and Southern
Mandé) constitute the ‘southerners’. In the period from November 1980 to November 1991, ‘northerners’ and ‘foreign Ivorians’ accounted for 4.2 million
people or 39.2 per cent of the population; in the period from November 1991, this was 6.2 million people or 40.6 per cent of the population.
Source: Langer (2008: 173).
79
to his earlier statements, Gueï decided to participate in the next presidential

elections after all. Furthermore, without explicitly using the term, he also
began to use the ideology of ‘Ivoirité’ in order to gain political support and
exclude political opponents, in particular Alassane Ouattara and his RDR
party (Akindès, 2004).
The presidential elections of October 2000 were marked by chaos and vio-
lence. When the Minister of the Interior Grena Mouassi proclaimed that
Gueï had won the elections, this sparked off massive street demonstrations
by FPI supporters as well as members of the security forces. The military
forces supporting these demonstrations, which eventually forced Gueï to
leave the country, were mainly of northern origin (Banégas and Losch,
2002). The official results proclaimed by the national electoral commission
stated that Laurent Gbagbo had won the elections with 59.4 per cent of the
votes (Le Pape, 2002). Following the exclusion of their presidential candi-
date Alassane Ouattara for reasons of ‘nationalité douteuse’ (nationality in
doubt), the RDR refused to recognise the legality of the results and demanded
new elections. To support their demands, RDR supporters started to organise
large-scale street protests, which led to violent confrontations with both the
FPI supporters and security forces.
Laurent Gbagbo originates from the western town of Gagnoa. He is a Bété,
one of the ethnic groups of the Krou family. In line with his anti-Ouattara
and anti-RDR, and therefore de facto anti-northern rhetoric, Gbagbo allo-
cated most government positions in the January 2001 government to his
own party, the FPI; northerners were largely excluded from his January 2001
government (see Table 4.6). Moreover, the northern underrepresentation in
this government was considerably worse than in any government under
Bédié (Langer, 2005). This obviously aggravated feelings of political exclu-
sion among RDR supporters. Paradoxically, the same military forces that
had chased away Robert Gueï, and thereby helped Gbagbo become presi-
dent of Côte d’Ivoire, attempted to overthrow Gbagbo’s regime in January
2001 (Banégas and Losch, 2002). However, the coup d’état failed and the
military forces involved were forced into exile.
Like his two predecessors, Bédié and Gueï, Gbagbo wanted to change
the ethnic composition of the military forces in favour of his own ethnic
group. In order to achieve this, Gbagbo planned to demobilise two con-
tingents that consisted predominantly of soldiers who had been recruited
during the brief reign of Gueï (Banégas and Losch, 2002). In response to the
planned demobilisation, however, these soldiers supported a mutiny which
quickly turned into a more organised rebellion, led by officers that had gone
into exile either because of the military purges during the Gueï regime or
because of their involvement in the failed coup d’état in January 2001 (ibid.).
The violent conflict in Côte d’Ivoire started with simultaneous attacks
against the military installations of Abidjan, Bouaké and Korhogo on 19
September 2002. By the end of September, the rebels firmly controlled
Arnim Langer 81
the northern part of the country and were referring to themselves as the
Mouvement Patriotique pour la Côte d’Ivoire (MPCI). The main grievances
put forward by the insurgents related to the land ownership laws, the cri-
teria of eligibility for presidential elections, the question of identity cards
and the political domination of the northerners by southerners (Dembélé,
2003). Although the vast majority of its forces had a northern background,
the MPCI claimed to have no specific ethnic, regional or religious affiliation.
Although the northern rebellion failed to overturn the Gbagbo regime,
due to the military intervention by France at the end of September 2002,
by then, MPCI firmly controlled the northern part of the country. France’s
intervention led to a situation in which there was no clear winner or loser.
This forced the conflict parties to seek a negotiated solution to their differ-
ences, which, once agreed, proved extremely difficult to implement (Langer,
2010a). After a string of failed peace agreements the conflict parties even-
tually agreed to hold presidential elections in October 2010 in order to
advance the peace process, but the outcome of these elections was chaos
and violence.
The main responsibility for the electoral debacle lay with Laurent Gbagbo,
who refused to accept his defeat in the presidential elections, which inter-
national observers had uniformly declared free and fair. In response to
Gbagbo’s refusal to hand over power to the newly elected President Alassane
Ouattara, the northern insurgents took up arms again and started march-
ing on Abidjan, the largest city in the country and the seat of government.
The forces loyal to Ouattara quickly gained control of the whole country,
apart from Abidjan itself, where a fierce battle between the Ouattara forces
and those forces that remained loyal to Gbagbo emerged and lasted for
several weeks. The fighting eventually ended on 12 April 2011, when the
Ouattara forces (with considerable help from French and UN peacekeeping
forces) were able to enter the Presidential Residence and arrest Gbagbo and
his closest associates. In the wake of Gbagbo’s arrest, Ouattara was officially
inaugurated as president of Côte d’Ivoire.
One important task for President Ouattara in the short term will be to reas-
sure and convince the substantial proportion of the population who voted
for Gbagbo that he will be a president for all Ivorians, regardless of ethnicity,
religion, region of origin or political affiliation. This is a formidable task in
a conflict-ravaged and ethnically polarised society like Côte d’Ivoire. In the
long term, however, a durable peace in Côte d’Ivoire will depend ultimately
on the political elites agreeing on a new set of ‘rules of the (political) game’.
4.3 Ghana: Living with diversity
Ghana has had a turbulent postcolonial political history, with long periods
of political instability, repeated non-constitutional regime changes and, at
times, ethnic tensions.3 Despite this, since the introduction of the Fourth
Republic a functioning multi-party democracy has emerged. While ethnic

conflicts have occurred at the local level (in particular in the northern
regions), at the national level ethnic tensions and conflicts have never
boiled over into serious violence. Interestingly, like Côte d’Ivoire, Ghana
is confronted with major ethnic, social and economic cleavages between
its northern and southern regions; but ‘there has been no development
of “Northernness” as a basis for political cohesion, and no north versus
south patterning of political alignments’ has occurred (Brown, 1982: 42).
Before analysing the main strategies and measures that successive Ghanaian
regimes have undertaken to keep ethno-regional tensions and mobilisation
in check, including the north–south cleavage, I will first discuss the coun-
try’s ethno-religious composition and the role that ethnicity and ethnic
mobilisation has played in post-colonial politics.
4.3.1 Ghana’s ethno-religious demography

Ghana is a multiethnic country of about 23 million people. The four main
ethno-cultural groups, together comprising around 86 per cent of the popu-
lation, are the Akan, Mole/Dagbani, Ewe and Ga/Dangbe.4 While all regions
have a sizeable number of migrants or people considered to be ‘strangers’
in their region, there is a rough coincidence of ethnicity and administrative
regions in Ghana (Gyimah-Boadi, 2003).
The Akans are by far the largest ethnic group, accounting for approxi-
mately 49 per cent of the population, and they form the majority of the
population in five of the ten regions in the southern part of the country
(namely the Western, Central, Eastern, Ashanti and Brong-Ahafo Regions).
The Akans comprise around 20 smaller ethnic sub-groups (of which the
Ashantis and Fantis are the largest) that share important cultural, social
and political institutions and customs. For instance, a cultural trait that the
Akan groups have in common and which differentiates them from other
ethnic groups is their matrilineal line of inheritance. Most Akan sub-groups
also speak closely related languages. However, while the Akan sub-groups
share important linguistic and cultural traits, they do not always behave as
a cohesive political unity (Gyimah-Boadi and Asante, 2006).
The second-largest ethno-cultural group are the Mole/Dagbanis. They con-
stitute around 17 per cent of the population and are found predominantly
in the northern regions. Together with other smaller ethnic groups from the
north, particularly the Grusi and Gurma, the northern ethnic groups consti-
tute around 23 per cent of the population. The Mole/Dagbani group is, how-
ever, a very loose ethno-cultural grouping which consists of a considerable
number of relatively small ethnic sub-groups, of which the following are the
most important: Builsa (Kagyaga or Kanja), 0.7 per cent of the population;
Dagarte (Dagaba), 3.7 per cent; Dagomba, 4.3 per cent; Kusasi, 2.2 per cent;
Mamprusi, 1.1 per cent; Namnam (Nandom), 2.4 per cent; Nankansi and
Gurense 0.5 per cent; Walba (Wala), 1.0 per cent; and Nanumba, 0.5 per cent.
Arnim Langer 83
While these sub-groups share certain social and cultural institutions,

they have very different histories, customs and traditions (Brukum, 1995).
Indeed, a number of them speak languages that are mutually unintelligi-
ble (Frempong, 2001). Another reason why it is anomalous to consider the
Mole/Dagbani as a single homogenous ethnic group is the conflictual and
occasionally violent interaction between the different ethnic sub-groups.
These ethnic conflicts ‘arise from several years of relegation of certain ethnic
groups, so-called “minority” groups, to “second-rate citizens” in the tradi-
tional and political administration of the region’ (Brukum, 1995: 153). It is,
however, important to note that these ethnic conflicts and clashes were only
significant at the local level and did not have any far-reaching consequences
at the national level (Agyeman, 1998). The relatively diverse ethnic demog-
raphy of the northern regions and conflictual inter-ethnic relations are
important factors militating against the mobilisation of the north as a group.
The third-largest ethnic group in the country, with about 13 per cent of the
population, are the Ewes. They are usually portrayed as the most homoge-
nous ethnic group mainly because of the lack of sharp subdivisions as well
as the perception that they speak a single language. Nonetheless, a consid-
erable number of Ewe dialects exist. The Ewes are found predominantly in
the southern part of the Volta Region in the eastern part of Ghana. The
northern part of the Volta Region is mainly inhabited by people belonging
to the Gurma and Guan ethnic groups. The fourth-largest ethnic group, with
around 8 per cent of the population, are the Ga/Dangbes. They are mainly
found in the Accra area and constitute about 30 per cent of the population
in the Greater Accra Region.
These ethno-regional differences are partly reinforced by religious dif-
ferences. While Ghana is a predominantly Christian country, with almost
70 per cent of the population adhering to one of the Christian denomina-
tions (Catholic, Protestant or Pentecostal/ Charismatic), Muslims constitute
around 16 per cent of the population. Muslims form a particularly important
proportion of the population in the northern regions. In the largest of the
three northern regions (itself called the ‘Northern Region’), the majority of
the population (about 56 per cent) is Muslim. However, this religious north–
south divide is mitigated by the fact that Christians and traditionalists also
constitute a significant proportion of the population in the three northern
regions, in particular in the Upper West and Upper East regions. Overall,
only 42 per cent of the population in the country’s northern regions are
Muslim, while 25 per cent are Christian and 29 per cent have a traditional
religion. Compared with Côte d’Ivoire, where the Muslims constitute around
60 per cent of the northern population, the Muslim religion is much less of a
unifying factor among the people living in Ghana’s northern regions. Rather
than ethnicity or religion, the most important common feature of the people
in the northern regions is their relatively deprived socioeconomic situation
compared to the people in the southern regions.
4.3.2 Ethnicity and politics in Ghana

Since the Second World War, the importance of ethnicity and ethno-regional
tensions in Ghanaian politics has waxed and waned to a remarkable degree
(Lentz and Nugent, 2000). Ethno-regional tensions escalated especially dur-
ing the short-lived episodes of multi-party democracy (Gyimah-Boadi, 2003).
For instance, the first competitive multi-party electoral contests in the 1950s
saw the emergence of several ethno-regional political parties, including the
Northern People’s Party, the Ewe-based Togoland Congress Party and the
Asante-based National Liberation Movement (NLM) (ibid.).
The ethno-regional tensions and mobilisation around the time of inde-
pendence in 1957 was followed by a period of comparative ethnic quiescence
in Ghana’s First Republic (1960–1966) under Kwame Nkrumah, who made
national unity a major objective. Ghana’s First Republic came to an end
with Nkrumah’s removal from power by a military coup d’état in 1966.
The new military regime, known as the National Liberation Council (NLC),
was soon confronted with serious ethnic tensions. Drawing on Hutchful
(1973), Frempong (2001: 145) asserts that ‘by the time the NLC handed
over power in October 1969, the military regime had split into factions
with the Ashantis and Ewes poles apart’. These ethno-regional tensions
were subsequently transferred into Ghana’s Second Republic (1969–1972)
under Abrefa Busia. Since then, the perceived rivalry between the Ewe and
Ashanti/Akan political elites has repeatedly stirred up the issue of ethnic-
ity (see, for example, Agyeman, 1998; Frempong, 2001; Gyimah-Boadi and
Asante, 2006).
The Second Republic was, however, short-lived, and came to an end
with Colonel Acheampong’s military intervention in February 1972. While
Acheampong and his NRC were initially able to dissipate the prevailing eth-
nic tensions in the political sphere, the lull in ethnic politics was very brief.
Soon Acheampong himself started using ethnicity as a means of mobilising
support behind his regime (Brown, 1982). The appearance of Flight Lieu-
tenant Jerry Rawlings on the political scene at the end of the 1970s saw the
pendulum swing back once more. Rawlings’ coup d’état of December 1981
was ‘particularly noteworthy for the manner in which ethnic claims were
subordinated to the language of class interest’ (Lentz and Nugent, 2000: 22).
Illustratively, the main conflicts and internal disputes within the Provisional
National Defence Council (PNDC), the new governing body after the 1981
coup d’état, had an ideological rather than an ethnic character. However, the
issue of ‘tribalism’ and ethno-regional favouritism was catapulted back into
the public sphere following a lecture by the eminent history Professor Adu
Boahen in 1988. In this lecture, he criticised Rawlings and his PNDC for fan-
ning ‘tribalism’ and appointing a disproportionately high number of Ewes to
influential political-administrative positions (see Boahen, 1992). The subse-
quent path towards ‘the return to constitutional rule in the Fourth Republic
was strewn with a replay of the Ashanti–Ewe rivalry’ (Frempong, 2001: 147).
Arnim Langer 85
Moreover, the Adu Boahen ‘incident’ is noteworthy because it shows once

more that the introduction of perceptions of HIs into the political discourse
can contribute directly to the ethnicisation of a country’s political system
and induce the emergence of ethnic politics. An important reason why eth-
nic politics did not escalate in Ghana in the early 1990s was the fact that
the prevailing ‘objective’ political HIs were actually rather moderate in this
period. In particular, as shown later in Table 4.9, the ethnic composition of
Rawlings’ 1988 and 1993 governments shows that the Akans (Adu Boahen’s
larger ethnic group) were actually overrepresented in relation to their demo-
graphic size in the population as a whole. More than 50 per cent of the
ministers in these governments had an Akan background. In contrast, the
Ewes went from being overrepresented in relation to their relative demo-
graphic size in the population as a whole in the 1988 government to being
somewhat underrepresented in the 1993 government.
While Ghana has made substantial progress towards becoming a consoli-
dated democracy since the reintroduction of multi-party elections in 1992,
regional voting patterns seem to suggest that ethnicity, while not the only
factor, remains important for determining political allegiances and elec-
tion results in the Fourth Republic, particularly in the Ashanti and Volta
Regions. In contrast to other regions, the Volta and Ashanti Regions have
voted overwhelmingly for their perceived home-based parties in successive
general elections. Despite the fact that the hierarchies of the two main polit-
ical parties continued to show a high degree of ethnic heterogeneity, ‘the
National Democratic Congress (NDC) and the NPP, are largely perceived
as Ewe and Ashanti/Akan based respectively’ (Gyimah-Boadi and Asante,
2006: 248). Ewe–Ashanti exceptionalism in voting patterns is seen by some
scholars as an indication of the persistence of the Ewe–Ashanti cleavage
in Ghana’s Fourth Republic (see, for example, Agyeman, 1998; Frempong,
2001; Gyimah-Boadi and Asante, 2006).
Yet, while there have been political tensions over the years, these have
not escalated into serious political violence at the national level, nor even
threatened to do so.
4.3.3 Containing ethnic politics

Kwame Nkrumah, the first national Ghanaian leader, was confronted with
serious ethno-regional tensions and mobilisation in the immediate pre-
independence period. Nkrumah’s strategies to deal keep ethnic politics in
check and promote national integration have to some extent become institu-
tionalised, both formally and informally. An important aspect of Nkrumah’s
strategy to contain ethnic politics was the adoption of the Avoidance of Dis-
crimination Act in December 1957. Under this law, the formation of political
parties along ethnic, regional or religious lines was prohibited. The Avoid-
ance of Discrimination Act was strongly opposed by the opposition at the
time, but since then successive Ghanaian political elites have recognised
the centrifugal potential of ethnic, religious and/or regional political parties.

Consequently, the 1969, 1979 and 1992 Constitutions and the 2000 Polit-
ical Parties Act all contain provisions aimed at curbing ethnic politics and
ensuring that political parties are national in character (Gyimah-Boadi and
Asante, 2006). The Ghanaian body politic also agreed in more informal ways
to avoid using ethnicity as a means of gaining electoral support. For exam-
ple, as part of the 2004 Political Parties Code of Conduct, which is a legally
non-binding document, the political parties agreed to refrain from using
ethnicity in their political campaigning. Despite the formal and informal
agreements and rules, as well as the existence of strong norms against the use
of ethnicity as a means of political action, both government and opposition
parties/politicians have occasionally played the ethnic card. Furthermore,
ethno-regional voting patterns (especially in the Ashanti and Volta Regions)
as well as survey research suggest that ethnicity remains (or at least is per-
ceived to be) an important factor in the public/political sphere (Langer and
Ukiwo, 2008).
Nkrumah and his Convention People’s Party (CPP) are also widely
regarded as having promoted national integration by denouncing commu-
nalism and by the ‘ethnically blind’ stance of Nkrumah himself (Brown,
1982: 41). Illustratively, as early as 1946 Nkrumah wrote that ‘all provin-
cial and tribal differences should be broken down completely’ (Nkrumah,
1962 quoted in: Smock and Smock, 1975: 227). In its place Nkrumah pro-
moted a Ghanaian identity and culture, which he projected as an amalgam
of different ethnic cultures within Ghana (Hagan, 1992). For instance, one
way in which Nkrumah demonstrated the pluralistic nature of the Ghanaian
state and identity was by wearing different traditional attire (such as the
northern fugu smock or the Ashanti kente) at official ceremonies. In many
ways, Nkrumah aimed to establish what Young (1976) has termed ‘a cul-
turally neutral state’ or a situation of ‘cultural status equality’ (Langer and
Brown, 2008), which entails that no particular ethnic culture or religion is
given more status and/or recognition in and by the state. An indication
of Nkrumah’s success in terms of nation-building is the fact that many
Ghanaians think back to Nkrumah’s time as Ghanaians rather than as
Ashantis, Ewes, Mole/Dagbanis, or Gas (Hagan, 1992).
Most of Nkrumah’s successors continued to promote the Ghanaian state’s
cultural inclusiveness by undertaking a similar range of measures in the
legal/institutional, policy-oriented and symbolic spheres. Thus, for instance,
Nkrumah’s practice of alternating modes of dress on public occasions was
continued by most heads of state. Other examples from among these cul-
turally inclusive practices and measures include: the persistent refusal by
consecutive Ghanaian governments to promote a particular local language
(especially Twi/Akan) as the country’s national language; the active state sup-
port for the study and teaching of the country’s major local languages; the
incorporation by institutions such as the Ghana Dance Ensemble of songs
Arnim Langer 87
and dances from all major ethnic groups (Lentz and Nugent, 2000); the con-
scious effort to ensure that radio and television programmes are broadcast
in all major languages (ibid.); and the custom that representatives from the
government attend the most important ethnic and/or traditional festivals
and durbars throughout the country on a regular basis. Similarly, examples
of some (symbolic) actions and practices which demonstrate the political
elites’ commitment to promoting and sustaining religious status equality
and inclusiveness include: the practice that representatives from all major
religions are present at official state functions; the state’s active organisa-
tional support for the annual Hajj pilgrimage to the Muslim holy sites in
Saudi Arabia; and the introduction of a new public holiday on the Muslim
festival of Eid-al-Adha in 1996.
Nkrumah also attempted to reduce the prevailing socioeconomic inequal-
ities, in particular between the north and the south. Ghana’s socioeconomic
north–south divide arises from a combination of circumstances and poli-
cies. Three main factors (which are, to a large extent, also applicable in the
case of Côte d’Ivoire) can be identified, the first of which is the geographi-
cal concentration of most agricultural activities and natural resources in the
southern regions; the northern regions predominantly rely on subsistence
farming. A second factor arises from the British colonial policy of investing
more heavily in those regions where exploitable resources such as gold, dia-
monds, timber and cocoa were available, or readily produced, and cheapest
to export. A third factor, which is important for understanding the persis-
tence of socioeconomic inequalities between the north and south, relates to
postcolonial development strategies and investment patterns.
The economic disadvantage of the Northern Region and, to a lesser extent,
the Volta Region compared with the rest of the country (especially with
regard to the Greater Accra Region) at the time of independence is illustrated
in Table 4.7. Despite Nkrumah’s efforts to reduce the north–south divide
and integrate the north into Ghana’s economy, the north remained gener-
ally much poorer in terms of income, infrastructure, education and medical
services at the end of the 1970s. The main reason for this was that the
postcolonial economy had an ‘endogenous’ tendency to favour the south
over the north in terms of the location of economic activities and invest-
ment, much like the situation in Côte d’Ivoire. Indeed, as Charles Anyinam
notes: ‘The economic policies and planning strategies adopted after inde-
pendence failed to alter the colonial legacy of lopsided development. The
development plans implemented between 1957 and 1980 tended to per-
petuate and reinforce the inherited inequalities in Ghana’s space economy’
(Anyinam, 1993: 449).
However, successive Ghanaian regimes did attempt to mitigate the
socioeconomic north–south divide. Thus, for instance, the Rawlings regime
(1983–2001) undertook a range of specific projects in the northern regions,
including the extension of the national electricity grid, the establishment
Table 4.7 Gross value added, per capita and per square mile, by
region, 1960
Region Gross value added Gross value added

per capita (£) per square mile (£)
Accra CD 176 87,374

Western 68 7,110
Eastern 53 7,410
Volta 43 4,187
Ashanti 68 7,753
Brong-Ahafo 61 2,389
Northern 30 1,032
All regions 63 4,566
Source: Szereszewski (1966: 92).
of the University of Development Studies in Tamale, the rehabilitation of

north–south roads and greater expenditure on education in the north. Sim-
ilarly, the Kufuor government (2001–2009) also publicly acknowledged the
problematic nature of the existing regional developmental disparities (see,
for example, Government of Ghana, 2003) and introduced several mecha-
nisms to redress the seriously adverse conditions in the northern regions,
especially in relation to health, social infrastructure, education and eco-
nomic infrastructure. For instance, the funds that were freed up as part of the
highly indebted poor country (HIPC) initiative were earmarked to benefit the
north disproportionately (Shepherd et al., 2005). Notwithstanding these and
other measures, Ghana’s socioeconomic north–south divide remains severe.
This is illustrated in Table 4.8. Yet, despite the limited success of the strategies
of economic redistribution in closing the north–south gap over the decades,
the impact on reducing the political salience of the north–south divide is
likely to have been much more substantial.
In addition to attempting to reduce the north–south inequalities and pre-
serve the cultural and religious ‘neutrality’ of the state, an arguably even
more important strategy to contain ethnic politics was aimed at maintaining
certain ethno-regional balances in the political sphere. This has remained a
largely informal ‘policy’ or convention among the country’s political elites.
Table 4.9 shows the relative representation of the Akans, Ewes, ‘northerners’
and ‘southerners’ among the government ministers of consecutive Ghanaian
governments in the period 1954–2005.
Throughout the post-independence period, the southerners persistently
controlled most of the ministerial positions and were actually somewhat
overrepresented as a group in proportion to their relative demographic
size. The Akans were persistently the largest ethnic group in government,
which was unsurprising given their demographic size in the population as a
whole. Yet, in the majority of governments, they were also overrepresented
Arnim Langer 89
Table 4.8 Poverty and infant mortality rates, 1998–2006
Incidence of poverty Infant mortalitya
1998/1999 2005/2006 2003 2006
% RRb % RRb % RRb % RRb
Western 27.3 0.69 18.4 0.65 66 0.67 45 0.63

Central 48.4 1.23 19.9 0.70 142.1 1.44 69 0.97
Greater Accra 5.2 0.13 11.8 0.41 62 0.63 60 0.85
Volta 37.7 0.95 31.4 1.10 98 0.99 57 0.80
Eastern 43.7 1.11 15.1 0.53 89.1 0.90 61 0.86
Ashanti 27.7 0.70 20.3 0.71 78.2 0.79 72 1.01
Brong Ahafo 35.8 0.91 29.5 1.04 128.7 1.30 88 1.24
Northern 69.2 1.75 52.3 1.84 171.3 1.73 83 1.17
Upper East 88.2 2.23 70.4 2.47 155.6 1.57 68 0.96
Upper West 83.9 2.12 87.9 3.08 155.3 1.57 114 1.61
National 39.5 1.00 28.5 1.00 99 1.00 71 1.00
Notes:
a. UNDP (2007c: 35).
b. Ratio to the national average; calculated by this author.
Source: Data derived from the Ghana Statistical Service. Available at: http://www.ghanainfo.org/
(last accessed October 15, 2008).
in proportion to their relative demographic size, which further heightened

their dominance. The Ewes were reasonably well represented in most govern-
ments in the post-independence period, except for the Busia governments of
1969 and 1971. In more recent years, under the New Patriotic Party (NPP)
government led by President Kufuor (who is an Ashanti from the Ashanti
Region), they were also somewhat underrepresented in proportion to their
demographic size.
In order to explain the non-mobilisation of the north, it is important
to note that northerners were generally included in government through-
out the post-independence period. Although northerners were somewhat
underrepresented in proportion to their demographic size in the period
1954–1979, when Jerry Rawlings assumed power in 1981, they actually
became slightly overrepresented. Under Kufuor (2001–2009), they again
became somewhat underrepresented (see Table 4.8). However, Kufuor com-
pensated for the northern underrepresentation among government minis-
ters by appointing a more than proportionate number of deputy ministers
from among the northern ethnic groups. In the government of February
2005, for instance, about 17 per cent of all ministers (including the deputy
ministers) had a northern background, which corresponded to a relative
representation of about 0.75. The fact that the position of vice presi-
dent, the second-most important position in Ghana’s 1992 Constitution,
was occupied by a northerner, Alhaji Aliu Mahama, further contributed to
90
Table 4.9 Relative representation of ethnic groups in cabinet, 1954–2005
Nkrumah-CPP Ankrah-NLC Busia-PP Acheampong-

NRC
1954 1956 1960 1965 1966 1969 1971 1972
% RR % RR % RR % RR % RR % RR % RR % RR
Akan 0.55 1.24 0.62 1.40 0.62 1.40 0.64 1.46 0.33 0.76 0.74 1.67 0.76 1.73 0.50 1.13
Ewe 0.09 0.70 0.08 0.59 0.23 1.78 0.07 0.55 0.33 2.56 0.00 0.00 0.00 0.00 0.29 2.20
Ga/Dangbe 0.09 1.10 0.08 0.93 0.15 1.85 0.14 1.72 0.22 2.68 0.05 0.63 0.06 0.71 0.07 0.86
Southerners 0.73 1.11 0.77 1.18 1.00 1.53 0.86 1.31 0.89 1.36 0.79 1.21 0.82 1.26 0.86 1.31
Northerners 0.18 0.84 0.15 0.71 0.00 0.00 0.14 0.66 0.11 0.51 0.16 0.73 0.18 0.82 0.14 0.66
No. N = 11 N = 13 N = 13 N = 14 N=9 N = 19 N = 17 N = 14
Acheampong- Limann-PNP Rawlings-PNDC Rawlings-NDC Kufuor-NPP

SMC
1975 1979 1981 1988 1993 1997 2002 2005
% RR % RR % RR % RR % RR % RR % RR % RR
Akan 0.50 1.13 0.57 1.30 0.43 0.87 0.50 1.02 0.65 1.32 0.53 1.08 0.79 1.61 0.61 1.24
Ewe 0.38 2.88 0.14 1.10 0.14 1.12 0.30 2.36 0.12 0.93 0.12 0.93 0.05 0.42 0.56 0.44
Ga/Dangbe 0.13 1.51 0.14 1.72 0.14 1.79 0.10 1.25 0.06 0.74 0.06 0.74 0.11 1.31 0.22 2.78
Southerners 1.00 1.53 0.86 1.31 0.71 1.02 0.90 1.29 0.83 1.19 0.71 1.02 0.95 1.36 0.89 1.29
Northerners 0.00 0.00 0.14 0.66 0.29 1.23 0.10 0.43 0.18 0.72 0.29 1.21 0.05 0.22 0.11 0.46
No. N=8 N = 14 N=7 N = 10 N = 17 N = 17 N = 19 N = 18
Note: ‘Northerners’ include individuals with a Mole/Dagbani, Gurma or Grusi background. ‘Southerners’ include individuals with an Akan, Ewe and
Ga/Dangbe background. Other smaller ethnic groups (such as the Mandé-Busanga and Guan) are excluded from these calculations. For the period 1954–
1979, the ethnic composition data are based on the 1960 census, while for the period 1981–2005 the 2000 Housing and Population census was used.
Source: Langer (2009: 543).
Arnim Langer 91
reducing the political salience of the northern underrepresentation among

government ministers. Moreover, the political inclusion of different ethnic
groups and interests by successive post-colonial regimes is arguably the most
important factor for understanding why ethnic politics (when it became
rife at times) did not escalate into a violent conflict at the national level
in Ghana.
4.4 Some conclusions
In both academic and policy circles, the concept of HIs is increasingly recog-
nised as an important factor for understanding the emergence of violent
conflicts in plural societies. While a substantial amount of empirical research
has found a significant association between the presence of HIs and the
risk of violent conflict, the causal mechanism linking these two phenom-
ena continues to be insufficiently understood. As argued in the introduction
of this chapter, further theoretical refinement is especially needed to explain
why HIs sometimes result in ‘ethnic’ violence, while in other cases similar
inequalities do not translate into violent group mobilisation and conflicts.
The cases of Ghana and Côte d’Ivoire clearly demonstrate that ethno-
regional or horizontal inequalities are in and by themselves insufficient to
produce violent conflict.
In previous work I highlighted the importance of the configuration of
multidimensional HIs as an important factor for understanding when ‘eth-
nic’ inequalities are more likely to escalate into violent conflicts (see Langer
2005). In particular, I argued that ‘ethnic’ violence is more likely to occur in
situations where political, social and economic HIs run in the same direc-
tion, namely, situations in which the same ethnic groups are politically
excluded and socioeconomically disadvantaged. In this chapter I attempt
to deepen our understanding of the process through which HIs may or may
not result in violent conflict. The chapter emphasises the importance of the
intervening variable of ethnic politics in understanding when HIs are more
likely to result in violent conflict; a factor so far ignored in the literature on
HIs. In Chapter 1, I argued that the presence of HIs may directly lead to an
ethnicisation of politics, which in turn may or may not escalate into violent
conflict. The analysis of Ghana and Côte d’Ivoire has shown that both coun-
tries have indeed known episodes where perceptions of HIs were introduced
into the political discourse and this subsequently contributed to the ethni-
cisation of politics and increased ethnic tensions. Yet while both countries
were confronted at different points in time with an ethnicisation of their
respective political systems and discourses, only in Côte d’Ivoire did ethnic
politics escalate into a violent conflict at the national level, which is an issue
we will turn to now.
In Côte d’Ivoire, while ethnicity and ethnic inequalities had an impor-
tant impact on Ivorian politics and occasionally engendered serious ethnic
tensions during Houphouët-Boigny’s presidency (1960–1993), overall ethnic

politics was kept in check relatively successfully in this period. An important
reason for why ethnic politics and tensions (especially those associated with
the sharp north–south divide) did not escalate under Houphouët-Boigny
was his strategy of diffusing the salience of the north–south cleavage by
including northern politicians and interests in the main political institutions
and by undertaking economic redistribution towards the deprived north-
ern regions. Importantly, even though the actual redistribution effect of the
implemented measures was shown to be rather limited, the symbolic impact
of these inequality-reducing measures and policies was undoubtedly much
more substantial. While the strategy of mitigating the prevailing political
and economic HIs was important for keeping ethnic discontent and politics
in check, the risk of violent escalation was further reduced by severely lim-
iting the political space of different political actors and interests. In other
words ethnic politics was kept in check through repression.
The descent of Côte d’Ivoire into violent conflict in the early 2000s, a
process initiated by the death of Houphouët-Boigny in 1993, shows not
only that the referral to and the manipulation of perceptions of HIs can
directly contribute to an ethnicisation of politics, but also that ethnic pol-
itics and mobilisation can rapidly spiral out of control and lead to violent
reactions. The analysis of the evolution of political HIs in combination with
the wider political discourse on HIs and exclusion, initiated by the opposi-
tion parties in the early 1990s and especially targeted at the regime of Henri
Konan Bédié, points to a perverse dynamic whereby the political discourse
of inequality and exclusion led to an increased reliance of the Bédié regime
on its own ethnic group, the Baoulé, and thus a sharp increase in polit-
ical HIs. Indeed, the distribution of the ministerial positions under Bédié
clearly showed that he increasingly favoured his own group. The main losers
in this new political environment were the northerners, who were increas-
ingly excluded from political power, and their leader – Alassane Ouattara –
was repeatedly disenfranchised. Given the pervasive ethnicisation of the
political system and discourse throughout the 1990s, the coup d’état in
December 1999, which removed Bédié from power, should not have come
as a surprise.
Similarly, when Laurent Gbagbo, who is a Bété, one of the ethnic groups of
the Krou family, took over power in the wake of the December 2000 elections
(elections which Alassane Ouattara, the main northern political leader, was
banned from participating in) and continued to adopt strategies of political
monopolisation by and favouritism towards his own ethnic group (as was
clearly illustrated by the increased representation of the Krou and the almost
complete exclusion of the northern ethnic groups in Gbagbo’s government
of January 2001), this induced a renewed attempt by predominantly north-
ern elements in the Ivorian armed forces to remove a sitting president from
Arnim Langer 93
power by a coup d’état. The failed coup d’état ultimately evolved into a pro-
tracted violent conflict, which arguably only ended in April 2011 with the
arrest of Laurent Gbagbo and the official inauguration of Alassane Ouattara
as President of Côte d’Ivoire.
The case of Ghana is interesting because it shows that successive Ghanaian
regimes have used similar strategies in managing their prevailing HIs (espe-
cially those between the northern and southern regions) and keeping ethnic
politics in check. Illustratively, successive Ghanaian regimes included north-
ern politicians and interests in the main political institutions and undertook
economic redistribution towards the deprived northern regions in order to
diffuse the sharp north–south cleavage. However, as in the case of Côte
d’Ivoire, the actual impact of the economic redistribution measures on
reducing the economic north–south divide was rather limited at best. While
keeping political and economic inequalities between the north and south in
check was an important factor for why Ghana did not experience a north
versus south mobilisation as we saw in Côte d’Ivoire, there are a number of
‘structural’ impediments to the mobilisation of the north as a group which
makes this inherently less likely to occur in the Ghanaian context, includ-
ing ‘the presence of a considerable number of relatively small ethnic groups,
which not only differ greatly regarding their cultural traditions, histories and
institutions but have also been in conflict with one and other at different
points in time’ (Langer, 2009: 544).
The successful management of political HIs throughout nearly the whole
postcolonial period appears to be a very important factor for understanding
why ethnic politics never did escalate into violent conflict in Ghana. More-
over, the ethnicisation of the political system that occurred following Adu
Boahen’s public accusations of ethnic favouritism and overrepresentation of
the Ewes under the Rawlings regime in the late 1980s is particularly useful to
illustrate this point. Indeed, the increased ethnicisation and ethnic tensions
which occurred in the wake of his accusations were basically kept in check
and eventually petered out relatively quickly mainly because the ‘objective’
HIs were rather moderate, as the analysis showed. Thus, in conclusion, it
appears that Ghana was not necessarily better at managing and reducing its
socioeconomic HIs than Côte d’Ivoire, but Ghana appears to have been bet-
ter at managing its political HIs, thereby reducing the appeal and mobilising
‘power’ of ethnicity and ethnic inequalities, which in turn helped to keep
ethnic politics in check.
Notes
1. The labelling of conflicts or violence as being ‘ethnic’ can be misleading, because
it could be seen to suggest that ethnic differences are the cause of these conflicts
(Alexander et al., 2000a). Instead of indicating a causal link between ethnicity and
conflict, in this paper the term ‘ethnic conflict’ solely refers to a conflict where the
different conflict parties have mobilised people to a significant extent along ethnic
lines.
2. This section draws on Langer (2005), (2008) and (2010a).
3. This section draws on Langer (2009) and (2010b).
4. The population data discussed in this section are based on Ghana’s 2000 Housing
and Population Census.
5
Beyond Ad hoc Power-Sharing:
Comparing South Africa and
Zimbabwe
Yoichi Mine
5.1 Introduction
There is an abundance of didactic stories featuring a contrasting pair of

brothers: one virtuous, the other vicious. In the biblical world, the rela-
tionship between Cain and Abel in the Old Testament is well known, as
is the story of the prodigal son in Luke’s gospel. In today’s Africa, Nelson
Mandela of South Africa and Robert Mugabe of Zimbabwe are often cast in
similar terms, something reflected in the extensive biographical literature
about the pair. Mandela has been acclaimed as a saint for having taken a
decisive leadership role to save his country from imminent civil war, before
retiring gracefully, while Mugabe has been demonised for having ruined his
country in the course of three decades of autocratic rule, driving his country
to the very bottom, the 169th place on the country listing of the Human
Development Index (HDI) (UNDP, 2010).
Mandela and Mugabe, together with the current political stability and
instability of their respective countries, are thus contrasted as extremes,
though these two elder leaders must have shared a dream while they
were aspiring young nationalists. Both studied at the University of Fort
Hare, South Africa, which was originally established by a Scottish mission-
ary to raise a black African elite, and which eventually produced a wide
array of respected leaders such as Julius Nyerere, Seretse Khama, Kenneth
Kaunda, Oliver Tambo and Desmond Tutu. As a country pair, the close
neighbours South Africa, with a population of 50 million, and Zimbabwe,
with 12.5 million, have a lot of common traits. The regionwide history
of African migration shaped the original configurations of both societies;
subsequently, given the temperate and tropical semi-arid climates and rich
mineral resources, the land of today’s South Africa and Zimbabwe attracted
a substantial number of Europeans, creating rigid racial horizontal inequali-
ties (HIs). After achieving liberation in 1980 (Zimbabwe) and in 1994 (South
95
96 South Africa and Zimbabwe
Africa), both countries have experienced widening inequalities within the

majority African population.
Every nation has its own unique features and distinctive historical charac-
ter. However, it is the shared experiences and the sense of common destiny
which allow the drawing of meaningful lessons from the study of their tra-
jectories. Based on the recognition of historical parallels in South Africa and
Zimbabwe, this chapter tries to understand how and why these two countries
followed such divergent paths in the past two decades, paying special atten-
tion to the effects of institutional engineering, to obtain practical lessons for
the prevention of violent conflict in Africa’s future.
5.2 Historical evolution of HIs in South Africa and Zimbabwe
5.2.1 Apartheid, post-apartheid and new inequalities in South Africa

The first part of this chapter discusses the historical settings, tracing the evo-
lution of multiple HIs in each country. The most arresting feature of South
African society is the substantial presence of people of European origin. Set-
tlement by continental Europeans commenced as early as 1652, and their
direct descendants were to mobilise themselves as a white African ethnic
group, the Afrikaners. In the early nineteenth century, the English-speaking
community also started to take shape, and the immigration process acceler-
ated after large deposits of diamonds and gold were found inland in 1867
and 1886, respectively. The hostility between the two white groups culmi-
nated in the South African War (1899–1902), and yet eventually gave rise
to institutionalised racism based on the postwar unity of the whites (Marx,
1998). Today, white South Africans form a substantial minority, accounting
for nearly 10 per cent of the population.
The majority of the South African population, about 80 per cent, are black
Africans, whose forebears had lived in today’s South Africa for centuries
before the arrival of the Europeans. This group consists of diverse ethnic
groups, including the Zulu, Xhosa, Sotho, Tswana Tsonga, Swati, Venda,
Ndebele and others (Table 5.1). The present ethnic configuration is the direct
outcome of the Mfecane, the regionwide warfare, migration and attempts at
nation-building by various groups of Africans triggered by the emergence of
the Zulu Kingdom in the early nineteenth century. Today, the Zulu and the
Xhosa are concentrated in the eastern and in southeastern areas of the coun-
try respectively, while the African inhabitants of major inland cities such as
Johannesburg are more mixed. Interposed between whites and Africans are
Coloureds (about 9 per cent of the population), largely the mixed descen-
dants of indigenous Khoisan people, slaves from the tropics and the early
white settlers, as well as Indian South Africans (about 2.5 per cent), largely
the descendants of indentured workers brought from South Asia in the late
nineteenth century.
Sharing the colonial pluralist character of Southeast Asian society
(Furnivall, 1948: 303–312), South African society also took on a racist
Yoichi Mine 97
Table 5.1 Home language in South Africa
Language Speakers (%)
Zulu 23.8
Xhosa 17.6
Afrikaans 13.3
Northern Sotho (Pedi) 9.4
Tswana 8.2
English 8.2
Sotho 7.9
Tsonga 4.4
Swati 2.7
Venda 2.3
Ndebele 1.6
Other languages 0.5
Total 100.0
Note: This is the list of the languages most often spo-

ken in South African households. A good number of
Coloureds speak Afrikaans at home, while most Indian
South Africans speak English.
Source: Statistics South Africa (2003).
character like the American South, with variegated cultural groups becoming
polarised into two distinct camps: the victims and the beneficiaries of racial
segregation. After 1948, the ruling National Party (NP) intensified institu-
tional HIs between racial groups under the name of apartheid; under this
system Africans, Coloureds and Indians were systematically disenfranchised
and deprived of economic and social entitlements as well as basic human
rights, while European cultures and languages were elevated to a dominant
position. Such institutional retrogression was consolidated in defiance of the
waves of decolonisation in Asia, Africa and the Caribbean in the latter half
of the twentieth century.
In fact, the most striking HIs in South Africa evolved around the land issue.
The conquest of autonomous African polities by white settlers was com-
pleted by the late nineteenth century. Although African peasants actively
responded to the rising demand for foodstuffs in mining cities, their aspi-
rations were crushed by forcible land appropriation (Bundy, 1979). Africans
were supposed to live in homelands, formerly called native reserves, which
constituted only 13 per cent of the South African territory (Figure 5.1). Those
who forged their livelihoods in white areas were vigorously relocated to their
rural ethnic homelands or segregated urban townships, and the total num-
ber of victims of forced removals is estimated to have reached 3.5 million
between 1960 and 1983 (Platzky and Walker, 1985: 10). However, as the
impoverished homelands could not sustain the livelihoods of rural Africans,
migrant workers seeking cash incomes flocked into the white areas, where
ZIMBABWE
MOZAMBIQUE
BOTSWANA
NAMIBIA
Pretoria
Johannesburg SWAZILAND
Bloemfontein
LESOTHO
Durban
Cape Town
Homelands
Kruger National Park
Figure 5.1 Homelands of South Africa

Source: Union of South Africa (1955: Appendix).
African labour was in much demand in the mining, manufacturing and agri-
cultural industries, especially in the large cities. The proportion of the urban
population in South Africa grew from 32 per cent in 1960 to 43.3 per cent in
1996 in spite of strict measures against urbanisation of Africans (Thompson,
2001: 298).1
The injustice of the apartheid regime functioned as a great leveller of the
oppressed. The Black Consciousness movement strived patiently to bridge
over African, Coloured and Indian politics to form a unified agency of the
‘Black people’. The exile leadership of the African National Congress (ANC)
and the South African Communist Party (SACP), including radical white
activists, made efforts to transcend racial and ethnic boundaries despite occa-
sional tensions (Ellis and Sechaba, 1992). In the popular protest in urban
townships, potential ethnic divides among Africans did not become salient,
and progressive Christians, Muslims and Jews took to the streets hand in
hand. The urban political dynamism as well as the successful containment
of cultural and ideological rivalry appears to have been the most significant
feature of the liberation movement in South Africa.
Such unity was further consolidated as a reaction to the divisive ‘reforms’
of the National Party (NP) government, which experimented with top-down
social engineering to divide and conquer the people. First, in 1983 the gov-
ernment tried to set up two houses of parliament with limited power for
Yoichi Mine 99
Coloureds and Indians, only to trigger a mass boycott organised by the

United Democratic Front (UDF) (Seekings, 2000: 91–119). Secondly, strate-
gic attempts were made to promote an African urban ‘middle’ class as a
buffer stratum, only to bolster the position of organised labour such as the
Congress of South African Trade Unions (COSATU) (Lipton, 1986: 59–75;
Baskin, 1991). Thirdly, based on the principle of ‘separate development’ in
which all racial and ethnic groups were expected to develop along their own
distinctive lines, the government tried to cut off homelands from white terri-
tory, granting sham independence to some in the 1970s. In the homelands,
ethnic traditions and customary laws were preserved, but the collusion of
autocratic traditional leaders with the white government constantly eroded
their legitimacy. The most stubborn leader was Mangosuthu Buthelezi, the
founder of a Zulu nationalist movement, the Inkatha Freedom Party (IFP),
which expanded its influence among Zulu migrant workers in cities as well
as in the KwaZulu homeland.
In 1994, the apartheid regime was buried forever with the first non-racial
general elections and the historic watershed of the inauguration of Nelson
Mandela as president. The political and cultural-status HIs of the apartheid
regime gave way formally to the principles of individual rights and of promo-
tion of the status of historically disadvantaged groups. Under the schemes
of Black Economic Empowerment (BEE), non-white people with relatively
higher educational backgrounds are now being actively recruited to man-
agerial positions in the public and private sectors (Iheduru, 2004; Freund,
2007). A process of land reform is also in progress, slowly but irreversibly, as
South African stakeholders in land issues recognise the danger of extra-legal
solutions like those in Zimbabwe.
In South Africa, the average economic HIs between racial groups in
terms of their share of total income is still wide, but has tended to nar-
row after apartheid (Figure 5.2) partly due to the relative shrinkage of
the white population (Figure 5.3). However, a decade after the first non-
racial elections in 1994, the unemployment rate of the African population
was hovering around 40 per cent, as non-agricultural employment has
decreased from 5.2 million in 1996 to 4.7 million in 2002, eroding the
support base of trade unionism. South Africa’s Gini coefficient was also
approaching the 0.7 level (Seekings and Nattrass, 2005: 303–304, 315–316,
317–319).2 Even though extreme poverty is somewhat cushioned by gov-
ernment schemes such as old-age pensions as well as disability and child
grants, the inequalities in material living conditions are widening within
each racial group, as shown in the recent trends of intra-group Gini coef-
ficients (Table 5.2). In today’s South Africa, there has been a widening of
the income gaps between the urban rich and urban poor, as well as between
urban and rural residents. Thus, the polarisation of the multiracial rich and
the African poor is the most striking feature of South African society in the
post-apartheid era.
100%
90%
80%
50 48
70% 55
60%
50% 4 5
4 8 8
40% 8
30%
20% 38 39
33
10%
0%
1995 2000 2004
African Coloured Indian White
Figure 5.2 Racial shares of total income

Source: Van der Berg et al. (2009: 27).
100%
13 11 10
90% 3 2
3 9
80% 9 9
70%
60%
50%
40% 76 78 79
30%
20%
10%
0%
1995 2000 2004
African Coloured Indian White
Figure 5.3 Racial breakdown of the South African population

Source: Mid-year population estimates, Statistics South Africa, http://www.statssa.gov.za (last
accessed 1 May 2013).
5.2.2 Liberation, violence and multiple inequalities in Zimbabwe

Zimbabwe and South Africa have been intertwined in the regional history
of Southern Africa. Like South Africa, Zimbabwean society could also be
categorised as a colonial settler society. One important difference, how-
ever, is that the presence of the white population in colonial Zimbabwe,
Yoichi Mine 101
Table 5.2 Gini coefficients by race and place
1993 2000 2008
African 0.54 0.60 0.62

Coloured 0.44 0.53 0.54
Indian 0.47 0.51 0.61
White 0.43 0.47 0.50
Rural 0.58 0.62 0.56
Urban 0.61 0.64 0.67
South Africa 0.66 0.68 0.70
Source: Leibbrandt et al. (2010: 32).
the former Southern Rhodesia, had been weaker than that of their South
African counterparts, even though it was more powerful than in other British
colonies in the tropics. Given that the proportion of the white population
was more or less 5 per cent at its peak in the 1970s,3 ‘it was the size of
Rhodesia’s White population – too small for grand apartheid but too large for
an easy exit – that accounted for the peculiar twists and particular tragedies
in the country’s history’ (Herbst, 1990: 14).
The majority of Zimbabweans are black Africans, and their present-day
ethnic configuration was also shaped by the migration processes associ-
ated with the Mfecane in the early nineteenth century. The Ndebele people,
originally a small Zulu faction lead by King Mzilikazi, settled in the South
African Highveld around today’s Pretoria; subsequently, after being attacked
by Afrikaners, they crossed the Limpopo River and started to consolidate
their permanent polity by assimilating the locals in the late 1830s. In the
north, by the late nineteenth century, the Shona people emerged from an
assemblage of diverse groups, some of which had built up historic kingdoms
and empires on the plateau (Beach, 1994; Mazarire, 2009; Ndlovu-Gatsheni,
2009a). The Shona are concentrated in the northeastern half of the coun-
try, with the Ndebele in the southwestern half. At present, the Shona are
said to account for about 80 per cent of the population, consisting of eth-
nic clusters such as Korekore, Zezuru, Manyika, Ndau, Karanga, Kalanga
and Rozvi, while the Ndebele make up about 10 per cent. This particular
ethnic configuration of majority Shona and minority Ndebele, in con-
trast with the more fragmented, multipolar ethnic configuration in South
Africa (Table 5.1), was to foment serious antagonism within the liberation
movement in Zimbabwe.
The prototypical imperialist, Cecil Rhodes, established the British South
Africa Company (BSAC) in 1889, seeking mining interests and territorial
expansion north of the Limpopo River. While serving as Prime Minister
of the Cape Colony, Rhodes deployed armed forces to crush the peasant
rebellion, the so-called First Chimurenga, in Matabeleland and Mashonaland
in 1896–1897. As the mineral resources found in Rhodesia were not as

rich as those in South Africa, however, the interest of white settlers shifted
towards the occupation of prime farmland to engage in commercial farm-
ing. They gradually started to produce tobacco, beef, sugar, cotton and
maize on large-scale farms, relocating African peasants to native reserves,
later called communal lands: the Zimbabwean counterparts of the South
African homelands. Communal lands occupied 42 per cent of the total land
of Zimbabwe at the time of independence (Figure 5.4).
In the face of the repressive demarcation of land and the scarcity of
job opportunities in the urban space, Zimbabwean peasants continued to
aspire to reverse the process of land expropriation from rural footholds
of struggle, which were not systematically destroyed as they were in the
case of South Africa. Against this background, the national liberation
movement in Zimbabwe in the latter half of the twentieth century was
MOZAMBIQUE
ZAMBIA
NAMIBIA
Harare
Bulawayo
BOTSWANA
MOZAMBIQUE
Lands for Africans
SOUTH AFRICA
Forests and Parks
Figure 5.4 Communal lands of Zimbabwe

Source: Government of Rhodesia, Map of Land Tenure, 1974.
Yoichi Mine 103
organised principally in the countryside, based on rural ‘land hunger’ as the

prime source of grievances. The guerrilla warfare against the settler regime
waged by Mugabe’s Zimbabwe African National Union (ZANU) and Joshua
Nkomo’s Zimbabwe African People’s Union (ZAPU) was dubbed the Sec-
ond Chimurenga, though these two forces competed fiercely with each other
seeking for hegemony.4 In stark contrast with the case of South Africa, the
political rivalry as well as the rural-based nature of mobilisation appears
to have been the most noticeable feature of the liberation movement in
Zimbabwe.
The new political dispensation took on its distinctive shape when,
through the mediation of the British government, the Lancaster House
Agreement was negotiated in 1979 bringing about a lasting ceasefire between
the guerrillas and the Rhodesian army. Under this experimental power-
dispersing arrangement, the interest of the white population was protected
by means of a 20-seat reservation for white citizens in the 100-member
parliament, which was to last until 1987, as well as the moderate ‘willing-
seller, willing-buyer’ principle of land reform. Although the political HIs in
Zimbabwe were redressed radically through the expansion of the elective
franchise to ordinary Africans, the remnants of racial HIs were still visible,
as whites, who accounted for only 2 per cent of the population, contin-
ued to dominate 37 per cent of the country’s income in 1983. Large-scale
commercial farms owned by white farmers occupied 39 per cent of newly
independent Zimbabwe’s territory in 1980, and the ratio was still no less
than 28 per cent in 1997. The pace of reform was perceived to be slower
than expected (Stoneman, 2000: 51; Kanyenze, 2003: 37).
While Mugabe’s practical compromise at that time was largely praised by
western countries, it is important to note that new forms of HIs started
to emerge after the liberation of Zimbabwe. In addition to the persistent
racial HIs between black and white Zimbabweans, political HIs between the
majority Shona and the minority Ndebele, as well as socioeconomic inequal-
ities between the new elite and the poor, were becoming salient during the
1980s.
On the one hand, the ruling ZANU terrorised Ndebele citizens to consol-
idate single-party domination. In 1980, Zimbabwe’s first popular elections
were conducted under a system of proportional representation (PR), the first
such electoral experiment in a former British colony in Africa (Reynolds,
1999: 162–163). In that poll, Mugabe’s ZANU won 63.0 per cent of the
votes and Nkomo’s ZAPU won 24.1 per cent. Unsatisfied with this absolute
majority, however, in 1983 Mugabe dispatched the Fifth Brigade, trained
by North Koreans, to the Ndebele areas, Matabeleland and Midlands, to
crush the support base of the main opposition ZAPU. This brutal repres-
sion, which continued for several years and claimed the lives of more than
20,000 Ndebele civilians, was given the Shona term Gukurahundi (the early
rain which washes away the chaff), and left an unforgetable memory of
victimisation in the minds of the Ndebele, thereby hardening the eth-

nic division in the country (Alexander et al., 2000b: 217–224; CCJP/LRF,
2008;). In 1987, the battered ZAPU was absorbed by ZANU when the two
parties merged to form ZANU(PF), and the official nationalist discourse in
Zimbabwe increasingly took on an air of Shona nationalism. In the same
year Prime Minister Mugabe was inaugurated as executive president of the
country.
On the other hand, the social and economic policies adopted by the ZANU
government in the 1980s were of a social welfarist nature, however difficult
this may be to believe now. The expansion of mass education was especially
remarkable, as the level of primary school enrolment increased dramatically
from only 0.82 million in 1979 to 2.05 million in 1983, and the adult liter-
acy rate rose from 55 per cent in 1970 to 85 per cent in 1994. Despite this
achievement, the employment level in formal industrial sectors remained
stagnant, real earnings in those sectors declined sharply (Dashwood, 2000:
40–49) and most school leavers, around a hundred thousand annually, could
not find jobs in the formal labour market (Chung, 1988: 129–131). At the
same time, the bureaucracy was Africanised, and various measures of affir-
mative action enabled a proportion of black Zimbabweans to climb up the
ladder in business as well as in the public sector. In 1990, Zimbabwe’s Gini
coefficient reached 56.8, much higher than the average for sub-Saharan
Africa, 44.7 (UNDP, 1998: 49).
Post-independence Zimbabwe was thus haunted by old and new inequali-
ties. In addition to the remnants of racial HIs that, in hindsight, could have
flared up easily, the Ndebele people were violently deprived of their politi-
cal voice and cultural dignity. On top of these, socioeconomic inequalities
within the black Zimbabwean population were becoming pronounced in a
way that foretold the post-apartheid situation in South Africa (Stoneman,
1988: 367). This polarisation of haves and have-nots within the African pop-
ulation is probably the most conspicuous commonality of two countries
today. Following the beginning of the land invasion in 2000, all of these
contradictions were thrown into a crucible, entailing a spate of political
violence, institutional decay and economic collapse.
Figure 5.5 gives a snapshot of the geographical patterns of inequalities in
Zimbabwe at the turn of the century, combined with corresponding figures
for South Africa. The provincial Human Development Index indicated that
Matabeleland tended to outperform rural Shona areas by very small margins,
while the disparity between urban and rural provinces was much more con-
spicuous. It is surmised that the political grievances of the Ndebele have been
occasioned by factors other than economic deprivation. In the case of South
Africa, those provinces that accommodate former rural homelands tended to
lag behind. Limpopo Province, for example, is on the border with Zimbabwe
and is mainly inhabited by the Northern Sotho and Tsonga. North West is
predominantly Tswana, Eastern Cape is Xhosa and KwaZulu-Natal is Zulu.
Yoichi Mine 105
0.8
0.7 South Africa
0.6
0.5 Zimbabwe
0.4
0.3
0.2
0.1
0
or a pe
F C g
ric e S e
M Av tate
aZ ala ge
st -N a
N n C al
th pe
el Ave uth
e
as na dl h
m t
aw o
m ela a o
na nd ds
M d W st
ic est
la sv d
C go
l
tra
M ab nd are
Li es
rn n
Af re ap
Ea ulu ng
an rag
M ho Mi ort
ul op
Zi el H ay
na Ma lan
er at
n a
ho la an
a
he te
Kw um era
or a
nd in
la E
W
at we So
en
C
a
u
rn
d
G
an
te
B
a
el
es
p
t
W
ab
ab
b
as
ho
h
at
ut
as
M
So
M
Figure 5.5 Provincial HDI in South Africa and Zimbabwe, 2001
Source: UNDP (2003a: 282; 2003b: 28).
Provinces with a high HDI are urbanised places where there are relatively
high proportions of whites: 19.9 per cent for Gauteng, which accommo-
dates Johannesburg and Pretoria, and 18.4 per cent for Western Cape, which
accommodates Cape Town (Statistics South Africa, 2003). Uprooting poverty
in the rural areas was, and still is, an urgent issue in both countries as wealth
concentrates in cities.
5.3 Contemporary political transitions in South Africa

and Zimbabwe
5.3.1 Power-sharing and beyond in South Africa

Given the above historical depiction of social cleavages and HIs in both
countries, closer attention is now directed to recent political processes first
in South Africa and then in Zimbabwe. These countries have taken con-
trasting paths over the past two decades; the formal political system in
South Africa has transformed itself gradually from the exemplary power-
sharing of major political parties to the majoritarian rule of the ANC, while
Zimbabwe witnessed a radical switchover from the one-party domination
of ZANU(PF) to power-sharing by ZANU(PF) and the Movement for Demo-
cratic Change (MDC). In other words, in the 1990s and the 2000s, South
Africa has experienced a gradual shift from a power-dispersing (PD) towards
a more power-concentrating (PC) position, while Zimbabwe jumped in the
opposite direction, seeking a temporary refuge (Table 5.3; see also Chapter 2,
Table 5.3 Shifts in formal political systems in South Africa and Zimbabwe
Country Sequence
South Africa Last stage of ⇒ Ad hoc ⇒ Majoritarian

apartheid power-sharing politics
(1994)
Zimbabwe Ad hoc ⇒ Majoritarian ⇒ Ad hoc
power-sharing politics power-sharing
(1980) (2009)
Figure 2.6). Let us trace the processes in these countries a little more
closely.
With the demise of the Cold War, President F.W. de Klerk released Nelson
Mandela in 1990 after 27 years of imprisonment, and political foes entered
multi-party negotiations in South Africa. The talks were often derailed by
political violence, including bombings and killings performed by both the
right and the left, and yet the most serious threat to the process was posed
by Zulu nationalists. Soon after the negotiations commenced, IFP support-
ers began to attack ANC supporters in townships and villages and incited
retaliation, pushing up the casualties of political violence to 16,022 from
1990 to 1994 (SAIRR, 1996: 52). In urban townships, ethnically integrated
African communities were assaulted by militant Zulu migrants, against the
backdrop of the bifurcation between urban civil society and the world of
customary law preserved in rural homelands (Mamdani, 1996: 218–284).
Unlike the Gukurahundi in Zimbabwe, the ‘black-on-black’ bloodshed in
South Africa took the form not of military repression but of seemingly vol-
untary ethnic mobilisation. However, the government commission chaired
by Judge Richard Goldstone later established that the IFP violence had been
orchestrated by the ‘Third Force’: a covert network of the apartheid security
forces.
As political fissures loomed large both inside and outside political parties,
South Africans were terrified of their country being on the brink of civil war,
as had occurred in Yugoslavia in the same period. It was at this critical junc-
ture that Mandela perfectly played the unifying role required by all groups
that wished to avoid the horrible consequences of each group promoting
its own self-interest. Mandela embodied the characters of both a seasoned,
dignified African chief and a sophisticated, telegenic urban politician, and
made energetic efforts to bridge racial and ethnic divides, having frequent
talks with leaders of rival political forces. In addition, his long imprisonment
had bolstered his even-handed position within the liberation movement.
The role he played in the transitional period can be understood as a modern
version of that played by the ‘leopard-skin chief’ depicted in the classic of
Evans-Pritchard, The Nuer (1940).
Yoichi Mine 107
After the tenacious multi-party negotiations (Sparks, 1995; Waldmeir,

1997), the political stakeholders in South Africa succeeded in hammering
out a provisional blueprint for non-racial democracy, which was enshrined
in the Interim Constitution adopted in 1993. The principles of the Gov-
ernment of National Unity (GNU) formed after the first non-racial general
elections in April 1994 represented distinctly power-dispersing features. First,
the members of the national and provincial legislatures were elected under
the PR system. Secondly, any party that gained 20 seats (5 per cent) or more
of the 400 seats in the lower house was entitled to be represented proportion-
ally in the cabinet, and any party that gained 80 seats (20 per cent) or more
could share the deputy presidency with the majority party. Thirdly, the prin-
ciple of the devolution of executive power to nine provincial governments
was secured. An upper house made up of delegates from the provinces was
also established.
Compared with Zimbabwe, the new political dispensation in South Africa
was more home-grown,5 and its credibility was enhanced by the adoption
of the democratic Constitution in 1996 complete with an independent
Constitutional Court as well as various watchdog institutions such as the
Public Protector (Ombudsman). The Bill of Rights, Chapter 2 of the Consti-
tution, is a comprehensive list of the ‘three generations’ of human rights,
and has often been praised as the most progressive constitutional statement
of human rights in the contemporary world. In parallel with the power-
ful recognition of individual rights, the Constitution also guarantees a set
of cultural rights of groups; no fewer than 11 official languages were des-
ignated in Chapter 1 of the Constitution. The new political dispensation
of South Africa exhibits a strongly federal character, with all variables of
the unitary/federal dimension being rated as 3 (Chapter 2 of this book,
Table 2.1).
The South African GNU was designed to provide all racial and ethnic
groups with room for self-expression by means of power-sharing. The PR
system enabled small minority voices to be represented nationally, and yet
a large group could still be represented proportionally. In the 1994 general
elections, the ANC won 62.7 per cent, the NP won 20.4 per cent and the IFP
won 10.5 per cent of the total votes (Table 5.4). Mandela was inaugurated as
president6 with Thabo Mbeki and De Klerk as his deputies, and all major par-
ties, including the IFP, sat in the cabinet. The term ‘Government of National
Unity’ became a mantra for post-conflict coalition governments elsewhere,
and has been adopted by African countries such as the DRC, Côte d’Ivoire,
Sudan, Zimbabwe and Kenya since 1994. Mandela stepped down in 1999
and has served as an influential conflict mediator in Burundi and the Middle
East (Bentley and Southall, 2005). The integration of the army was relatively
smooth, not least because the scale of past belligerence between government
soldiers and guerrillas had been limited in South Africa compared to the case
of Zimbabwe.
108
Table 5.4 National Assembly election results in South Africa
1994 1999 2004 2009
Votes (%) Seats Votes (%) Seats Votes (%) Seats Votes (%) Seats
ANC: African National 62.7 252 66.4 266 69.7 279 65.9 264
Congress
NP: National Party (NNP: 20.4 82 6.9 28 1.6 7 — —
New National Party)
IFP: Inkatha Freedom Party 10.5 43 8.6 34 7.0 28 4.6 18
Freedom Front (Freedom 2.2 9 0.8 3 0.9 4 0.8 4
Front Plus)
DP: Democratic Party (DA: 1.7 7 9.6 38 12.4 50 16.7 67
Democratic Alliance)
PAC: Pan Africanist 1.3 5 0.7 3 0.7 3 0.3 1
Congress
ACDP: African Christian 0.5 2 1.4 6 1.6 7 0.8 3
Democratic Party
UDM: United Democratic — — 3.4 14 2.3 9 0.9 4
Movement
ID: Independent Democrats — — — — 1.7 7 0.9 4
COPE: Congress of the — — — — — — 7.4 30
People
Other Small Parties 0.8 0 2.3 8 2.1 6 1.8 5
Total 100.0 400 100.0 400 100.0 400 100.0 400
Note: New names of parties are given in parentheses.

Source: Electoral Commission of South Africa, http://www.elections.org.za (last accessed 1 May 2013).
Yoichi Mine 109
With the establishment of the GNU, the racial HIs in political representa-
tion and cultural status consolidated during the apartheid era were radically
dismantled. Past injustices committed by the apartheid government and
the liberation movement were the focus of the Truth and Reconciliation
Commission (TRC). On the economic front, the ANC’s policy circle had
worked out a Keynesian-oriented policy prescription of a mixed economy,
the Reconstruction and Development Programme (RDP), which was used as
the ANC’s manifesto for the 1994 elections (MERG, 1993). However, Thabo
Mbeki, who took over the presidency from Mandela in 1999, set aside the
RDP line and instead began to implement a more market-oriented policy
package called Growth, Employment and Redistribution (GEAR). The eco-
nomic power of the white population was largely preserved as a result of
the ANC’s conversion to economic realism. The adherence of the ANC to
the principle of the market economy seems to be the essential reason why
the vote-based political dominance of the ANC was practically supported
by the capitalist sector and grudgingly accepted by the white population
at large.
Meanwhile, the NP withdrew from the GNU in 1996, and the IFP dis-
engaged from the government in 2004. The ANC, as the ruling majority
party, was thus placed in a position to formulate national policy single-
handedly, accommodating critical voices raised by its historical partners in
the anti-apartheid struggle, COSATU and SACP. Reflecting the dissolution
of the grand coalition, the rating of variable C (Composition of the Exec-
utive) in South Africa changed from 3 to 1 (Chapter 2, Table 2.2). As the
influence of the NP and IFP waned, the Democratic Alliance (DA) began to
advocate the value of liberal democracy and emerged as the leading oppo-
sition party. The ANC was placed in a hegemonic position, adopting policy
elements of the left and right in a selective way, but its governing practice
was like walking on a tightrope.
The fiscal discipline espoused by the Mbeki government contradicted
the high expectations of the disadvantaged masses, mainly black South
Africans.7 The growth of employment in the formal sector remained very
slow, and the delay of social service delivery has induced violent protest in
several townships since the mid-2000s. Under the presidency of Mbeki, the
Africanisation of bureaucracy was accelerated, and wealthy Africans, dubbed
‘Black Diamonds’, bolstered their presence, but inequalities within Africans
continued to widen. Counterbalancing such realities, Mbeki often resorted
to racial discourse, for example, blaming white prejudice for the prevalence
of HIV/AIDS, to which his government was reluctant to react with effective
policy measures. As described aptly in MacDonald (2006: 133), ‘emphasizing
racial identities represents poor Africans symbolically, while putting them
off economically’. In the 2004 general elections, the ANC obtained more
than a two-thirds majority, though the voter turnout was relatively low
(Schulz-Herzenberg, 2009).
COSATU and SACP perceived that they were sidelined by the neoliberal
Mbeki mainstream faction in the ruling alliance. The left-wing faction then
began to fight back within the party structure of ANC, enabling Jacob Zuma
to be selected as the party president at the national conference in 2007,
and succeeded in forcing Mbeki to step down. As a possible prelude to a two-
party system, Mbeki loyalists broke away from the ANC to form the Congress
of the People (COPE), which challenged the ANC in the 2009 general elec-
tions. The ANC eventually won a majority with 65.9 per cent of votes, while
DA and COPE secured the position of the loyal opposition.
While politics in South Africa is thus tilting towards ‘normal’ majoritarian
politics, the practice of informal power-sharing still seems to be alive.
After the dissolution of the GNU, the ANC government has always tried
to recruit a few cabinet members from minority parties, including the
Afrikaner nationalist Freedom Front. Moreover, Zuma’s ethnic identifica-
tion as a ‘100% Zulu Boy’ (in the words on the T-shirts of his supporters)
has enabled the ANC to rally Zulu voters with no relapse of violent con-
flict between the ANC and IFP. All in all, the 2009 elections marked a shift
of the power bloc, but the change occurred through peaceful popular vot-
ing, a practice that has firmly taken root in the political life of South Africa
since 1994 (Friedman, 2009). In addition, the multicultural principle of the
Constitution is accepted by the general public, the corporate leadership is
increasingly integrated in terms of race, and the National Economic Devel-
opment and Labour Council (NEDLAC) provides a corporatist framework for
business and labour. Beneath the surface of formal majoritarian politics, the
spirit of informal power-sharing has become an accepted norm in the South
African political life.
5.3.2 Elections as conflict triggers in Zimbabwe

By 1990, Zimbabwean society was plagued with multiple socioeconomic
inequalities. As long as the economy was growing, people suffering rel-
ative deprivations could forbear from organising themselves for a radical
change. However, the dramatic deterioration in people’s living conditions
under the Economic Structural Adjustment Programme (ESAP) introduced
in 1991 ignited a fuse in Zimbabwe. In order to curtail budget deficits and
external debt, more than 20,000 posts were abolished in the public sector,
and the incidence of extreme poverty rose from 16.7 per cent to 35.7 per cent
from 1990/91 to 1995/96 (Kanyenze, 2003: 56–68). A series of droughts in
the early 1990s added to their plight, which intensified the pressure for land
reform. It is undeniable that the ESAP, impelled by a group of technocrats
in the government (Dashwood, 2000; Bond and Manyanya, 2002; Davies,
2004), was the first direct ‘trigger factor’ in the whole conflict sequence
of the past two decades in Zimbabwe, although ZANU(PF) was confident
at the time of its ruling power, having gained a 97 per cent monopoly of
parliamentary seats in the 1990 elections.
Yoichi Mine 111
Intriguingly, in the initial stage, grievances about their economic plight

were expressed by the black business and middle class as a strong voice
calling for racial redress, sometimes in the form of violent racial hatred
against white citizens (Muzondidya, 2009: 190–194). Then, faced with a fail-
ure in what would now be called the process of disarmament, demobilisation
and reintegration (DDR), outraged war veterans were becoming increasingly
vocal about the necessity of radical land reform. Finally, unionised workers
raised a disciplined voice through a wave of successful strikes staged by the
Zimbabwe Congress of Trade Unions (ZCTU) in 1996 and subsequent years
(Raftopoulos and Schikonye eds, 2001). As the Zimbabwean deployment
of troops in the Democratic Republic of the Congo (DRC) in 1998 further
exacerbated the crisis in government finance, the Movement for Democratic
Change (MDC), the most influential opposition party since independence,
was launched in 1999.
In 2000, the situation in Zimbabwe entered a critical stage. A draft con-
stitution, which would have given the government the power to seize white
farms, was rejected in a referendum, with ‘No’ votes reaching 54.7 per cent,
as a result of a campaign organised by the opposition-based National Con-
stitutional Assembly (NCA).8 Soon after that, groups of war veterans started
to invade white farms, and Mugabe, who was very keen on regime survival,
decided to cement an alliance with the outraged veterans.9 Even though the
land invasion has radically transformed the pattern of land ownership in
Zimbabwe, the method was completely illegitimate, and the prevalence of
violence and lawlessness drove the economy to the brink of total collapse,
leaving the living conditions of ordinary Zimbabweans sharply deteriorated.
While the average real GDP growth rates from 2001 to 2009 were 5.3 per
cent for all African countries and 3.6 per cent for South Africa, Zimbabwe’s
figure in the same period was minus 5.4 per cent (OECD, 2010). The cover-
age of basic vaccinations in Zimbabwe, which had reached 80.1 per cent of
children in 1994, declined to 74.8 per cent in 1999 and shrank to 52.6 in
2005/06 (CSO and Macro International, 2007: 140).
Mugabe’s despotism intensified the alienation of the masses. In 2005, the
government mounted Operation Murambatsvina (clean-up), whose aim was
to violently bulldoze urban informal settlements as retaliation for the sup-
port of city dwellers for MDC in that year’s elections (Vambe ed., 2008).
The autocracy and misrule of Mugabe appear to have functioned as a
great leveller that united people against ZANU(PF) beyond class and ethnic
boundaries. In the 2008 general elections, implemented in a vortex of hyper-
inflation, the opposition parties, MDC-T headed by Morgan Tsvangirai and
MDC-M headed by Arthur Mutambara, secured a majority in parliament.
In the presidential election held at the same time, Tsvangirai gained 47.9 per
cent of votes, while Mugabe gained 43.2 per cent, and they were to hold a
runoff. As the disgraced Mugabe organised violent intimidation against vot-
ers, Tsvangirai decided to withdraw to minimise the victims, and Mugabe
was eventually ‘re-elected’ in the second round with no rival (Masunungure

ed., 2009).
The new Mugabe presidency lacked legitimacy. In the following year,
the intervention of the Southern African Development Community (SADC)
led major stakeholders to sign the Global Political Agreement (GPA), and,
accordingly, the Zimbabwean version of a GNU, an inclusive coalition gov-
ernment, was formed by President Mugabe and Prime Minister Tsvangirai
(Matlosa, 2009; Zondi, 2011). As shown in our perceptions surveys in
Zimbabwe, the formation of the GNU was welcomed by a majority of urban
citizens who had been choked by the perpetuation of violence in the past
decade, even though this should not be misconstrued as people’s satisfaction
with the regime survival of the ZANU(PF) through coalition manoeuvring.
Mugabeism is a ‘Janus-faced’ phenomenon with contradictions, represent-
ing Africanist populism, leftist nationalism, and patriarchal ruthlessness.
As such, it might be unrealistic to expect that its presence will wither away
in a moment of time (Ndlovu-Gatsheni, 2009b: 234–298).
Throughout the decade of the 2000s every referendum and election held
in Zimbabwe, where the electoral system is designed to promote majoritarian
rule based on the ‘winner-takes-all’ principle, has resulted in outbreaks of
violence. In spite of the radical shift of variables C (Composition of the Exec-
utive) and D (Head of Government) towards a PD orientation, variable A
(Electoral System) in Zimbabwe has stayed simply majoritarian (Chapter 2,
Table 2.2). Besides, the ZANU(PF) government has continuously tampered
with political institutions, installing the senate and setting up the quota
for political appointees and traditional chiefs, as well as gerrymandering
(Booysen and Toulou, 2009).10
However, ironically, it is the very majoritarian institution developed by
ZANU(PF) that posed the threat of its downfall. Tables 5.5 and 5.6 show
the results of three major elections in the 2000s, with the actual results
of the first-past-the-post (FPTP) small constituency system contrasted with
a simulation of the parallel results under a party list PR system, which
takes the provinces as large constituencies. It is apparent that the present
FPTP system tends to overemphasise regional differences, as substantial
votes for MDC in Mashonaland and Midlands as well as substantial votes
for ZANU(PF) in Harare and Manicaland were ‘wasted’. In the 2008 elec-
tions, for example, MDC-T won all of the seats in 12 constituencies in
Bulawayo with 40 per cent of the total votes, while ZANU(PF) and MDC-
M obtained no seats, even though the sum of the votes obtained by these
two parties also reached nearly 40 per cent (Table 5.6). At a critical junc-
ture, the majoritarian system sometimes leads to a landslide victory and a
total defeat of certain political parties, even with just a slight shift in the bal-
ance of their popularity. This is probably one reason, understandable albeit
not justifiable, why ZANU(PF) organised punitive violent intimidations,
Yoichi Mine 113
Table 5.5 Parliamentary election results in Zimbabwe, 2000 and 2005
2000 2005
Votes (000) PR FPTP Votes (000) PR FPTP
Bulawayo ZANU(PF) 22.4 1 0 22.6 2 0

MDC 142.4 7 8 85.5 5 7
Harare ZANU(PF) 85.0 4 0 112.1 6 1
MDC 296.1 15 19 234.2 12 17
Manicaland ZANU(PF) 117.2 7 7 191.6 8 13
MDC 125.8 7 7 147.2 7 2
Mashonaland ZANU(PF) 189.0 8 10 229.3 8 10
Central MDC 47.5 2 0 42.9 2 0
East MDC 65.0 3 1 69.7 3 0
West MDC 79.0 4 2 77.9 4 1
Masvingo ZANU(PF) 163.0 9 12 214.5 10 13
MDC 92.2 5 2 96.2 4 1
Matabeleland ZANU(PF) 36.7 2 1 51.2 3 1
North MDC 98.9 5 6 85.9 4 5
Matabeleland ZANU(PF) 50.7 3 1 70.8 4 3
South MDC 97.1 5 7 70.0 3 4
Midlands ZANU(PF) 193.8 10 11 230.9 10 12
MDC 126.1 6 5 125.8 6 4
Total ZANU(PF) 1,207.0 61 63 1,557.0 70 78
MDC 1,169.9 59 57 1,035.3 50 41
Note: In the 2005 elections, an independent candidate won one seat. This is why the sum of the
real seats (FPTP) in 2005 is one seat fewer than the alternative calculation (PR).
Source: Zimbabwe Election Support Network (2009: 26–27).
especially in its own traditional strongholds, Mashonaland and Masvingo,

in 2008.
With the political and economic turmoil described in this section, the
number of Zimbabweans who crossed the Limpopo River seeking refuge in
neighbouring South Africa increased in the 1990s and has risen dramat-
ically in the 2000s, so that more than one million Zimbabweans, many
of whom are skilled and qualified people and a majority of whom are
Ndebele, are estimated to have settled in South Africa by 2007 (Makina,
2010: 228). Such waves of migration have created new HIs between South
African citizens and outcast foreign Africans in South Africa, derogatorily
called the Makwerekwere, resulting in the eruption of xenophobic violence
in May 2008, in which 62 were killed and countless people injured in
attacks by angry, jobless black South Africans (Hassim et al. eds, 2008). As a
Table 5.6 Parliamentary election results in Zimbabwe, 2008
2008
Votes (000) PR FPTP
Bulawayo ZANU(PF) 11.4 2 0

MDC-T 40.1 6 12
MDC-M 27.6 4 0
Harare ZANU(PF) 69.7 7 1
MDC-T 214.0 20 28
MDC-M 20.7 2 0
Manicaland ZANU(PF) 147.5 11 6
MDC-T 195.7 14 20
MDC-M 10.7 1 0
Mashonaland Central ZANU(PF) 150.2 12 16
MDC-T 64.3 5 2
MDC-M 7.5 1 0
Mashonaland East ZANU(PF) 169.5 14 19
MDC-T 101.4 8 4
MDC-M 10.6 1 0
Mashonaland West ZANU(PF) 138.5 13 16
MDC-T 89.1 8 6
MDC-M 13.2 1 0
Masvingo ZANU(PF) 164.2 14 12
MDC-T 131.4 11 14
MDC-M 9.9 1 0
Matabeleland North ZANU(PF) 47.7 4 4
MDC-T 55.1 5 5
MDC-M 39.7 4 3
Matabeleland South ZANU(PF) 45.7 5 4
MDC-T 24.6 3 2
MDC-M 37.8 5 7
Midlands ZANU(PF) 168.4 15 21
MDC-T 123.0 11 7
MDC-M 25.3 2 0
Total ZANU(PF) 1,112.8 97 99
MDC-T 1,038.5 91 100
MDC-M 203.1 22 10
Note: Regarding the FPTP seats in 2008, see the note to Table 5.5.
Source: Zimbabwe Election Support Network (2009: 28).
practical solution, it is imperative to place the ‘failing’ African states, includ-

ing Zimbabwe, back on the right track of equitable development, so as to put
an end to the chronic drain of human resources from those countries. At a
deeper level, it must also be noted that national citizenship is fully in play to
stiffen the hierarchical boundaries between insiders and outsiders in today’s
nation-states in Southern Africa (Nyamnjoh, 2006). In Zimbabwe itself, the
Yoichi Mine 115
positions of foreign farm workers during and after the land invasion sharply
deteriorated (Rutherford, 2003).
5.4 Popular perceptions in South Africa and Zimbabwe
5.4.1 Race and class in neoliberal Cape Town

In the following sections, inequalities and political processes in South Africa
and Zimbabwe are reconsidered in the light of the results of our recent
perceptions surveys. The survey in South Africa was conducted in Septem-
ber 2010 in Cape Town, the provincial capital of Western Cape. Given
the strong presence of affluent whites, the city elite has pursued a neolib-
eral development policy since 1994, leaving the human landscape of the
city with an extremely unequal, ghettoised character (McDonald, 2008). As
Western Cape is the wealthiest province in South Africa (Figure 5.5), it must
be emphasised that the samples are far from representative of the nation-
wide demography. However, given that Cape Town is the place in which
the economic difference between the rich and the poor is most palpable
in the country, the survey may provide a good indicator of the grievances
related to racial HIs among Africans, most of whom continue to reside
in segregated townships in the same way as they did during the time of
apartheid.11
According to the 2001 census, whites accounted for 18.8 per cent,
Coloureds for 48.1 per cent and black Africans for 30.0 per cent of the pop-
ulation of Cape Town. The total number of our samples is 600 (200 samples
respectively for traditionally African, Coloured and white residential areas).
Each of the white and Coloured communities is composed of Afrikaans-
speaking and English-speaking segments, and a part of the Coloured com-
munity is Muslim. Africans are predominantly Xhosa-speaking, including
both city-born citizens and migrants from Eastern Cape, mostly from the
former homelands assigned to the Xhosa people: Transkei and Ciskei. Of the
total respondents, 60.0 per cent of Africans, 26.5 per cent of Coloureds and
3.6 per cent of whites answered that they had experienced shortages of cash
income in the past year. About one in four African respondents lived in infor-
mal shelters, highlighting a sharp contrast in the living conditions among
the respondents.
The survey outcome reveals that race is still an important issue in Cape
Town and that people sometimes hold contradictory perceptions about
racial HIs. No less than 86 per cent of Africans, 72 per cent of Coloureds and
20 per cent of whites answered that their racial identity was very important.
The ‘liberal’ attitude of whites is somewhat compromised by their objection
to mixed marriage: 45.8 per cent would object if their daughter or sister
were to marry an African, while only 7.3 per cent of Africans said that they
would object to such a mixed marriage. Regarding the validity of govern-
ment interventions to redress racial inequalities, very polarised opinions
were expressed; 89.8 per cent of Africans, 64.5 per cent of Coloureds and
20.1 per cent of whites agreed that affirmative action should continue for
a considerable period. Africans believe that whites benefit most from the
current economic system; 85.9 per cent of Africans and 24.0 per cent of
Coloureds thought that the economic situation of the white population had
become much better than it was ten years ago, while only 1.0 per cent of
whites thought this was the case.
In reality, a majority of the white population lives in relatively rich
suburbs. Their lives are generally well protected from threats from the out-
side world, and even daily contact with neighbours seems to be limited.12
Although the media outside South Africa tend to depict affluent whites as
falling victim to violent crimes committed by thugs and rogues, the main
sufferers are in fact Africans; 60.5 per cent of them answered that they
had personally experienced violent crime in the past year, compared with
26.5 per cent of Coloureds and 11.5 per cent of whites. At the same time, new
forms of inequalities were clearly recognised, especially by African respon-
dents; 82.0 per cent of Africans, 71.0 per cent of Coloureds and 23.4 per cent
of whites strongly agreed with the statement that the class division between
the rich and the poor was becoming more serious than racial discrimination
(50.5 per cent of whites agreed, but not strongly). In a similar vein, 84.9 per
cent of Africans, 73.5 per cent of Coloureds and 40.1 per cent of whites
agreed strongly that corruption rather than racism is the principal problem
in this country.
Power-sharing as a means of political representation is highly valued
across racial groups, to the extent that 81.5 per cent of all respondents felt
positive about an inclusive government like the GNU of 1994; 75.7 per
cent of them added that they would continue supporting their political
parties even after the parties entered into a coalition. Of all respondents,
71.7 per cent of Africans supported the ANC, the ruling party in the central
government, while 68.0 per cent of Coloureds and 85.4 per cent of whites
supported DA, the main opposition in national politics but the ruling party
in the provincial government. As a way to solve the issues in the govern-
ment, no less than 73.7 per cent of Africans, the majority of whom must be
loyal ANC supporters, preferred having a strong opposition. Though it is not
clear whether they were thinking of the central or the provincial govern-
ment, this nonetheless indicates a mature commitment of majority Africans
to multi-party democracy.
People in Cape Town seem to be optimistic about the future of their
nation. Of the total, 84.9 per cent of Africans, 78.1 per cent of whites and
71.5 per cent of Coloureds thought that South Africans were now thinking of
themselves as South Africans more than they did ten years ago. Furthermore
it seems that less impoverished Africans tend to prioritise racial identity
over national consciousness. Cultural events can be immensely instrumen-
tal in nation-building; 80.4 per cent of the total respondents answered that
Yoichi Mine 117
the 2010 FIFA World Cup had strengthened the unity of South Africans.
Defying global trends after the incidents of ‘September 11’, the relation-
ship between Muslim Coloureds and Christians of all racial groups is quite
amicable. More than 80 per cent of Christians and Muslims felt they were
comfortable working with each other.
As for the role of South Africa in conflict mediation, Africans tended to
express more pronounced views; 54.6 per cent of Africans strongly agreed
that their government should take a leadership role in settling political prob-
lems in other countries like Zimbabwe, but 25.9 per cent strongly disagreed
with such initiatives. On the other hand, 16.1 per cent of white respondents
strongly agreed, while 14.6 per cent of them strongly disagreed. Africans
were more pessimistic about xenophobia; 44.4 per cent of Africans (49.5 per
cent for the Africans from Eastern Cape and 40.8 per cent for the Africans
born in Cape Town and its surroundings), 27.0 per cent of Coloureds and
22.9 per cent of whites thought that the relations between South Africans
and foreigners from other African countries would worsen in the future.
5.4.2 Transient peace in Harare and Bulawayo

The surveys in Zimbabwe were conducted in March 2010 in the capital
city Harare, which is thought to be inhabited mainly by the Shona, and
the second-largest city Bulawayo, located in the heartland of the Ndebele
region. The total number of samples is 402 (Shona, 162; Ndebele, 64; others,
176), with sample households and individuals being randomly selected in
Harare (294) and Bulawayo (108). The fact that about a half of the Shona
speakers did not identify with Shona but with their own minority groups
(categorised as ‘others’)13 is indicative of the fluidity of ethnic identities in
Zimbabwe. Historically, a majority of African inhabitants in Salisbury, the
former name for Harare, had been foreign migrants from colonial Malawi,
Mozambique and Zambia, and the local Zimbabweans started to wield signif-
icant influence on the city life only after the Second World War (Yoshikuni,
1999).
The political turmoil and economic breakdown have clearly had negative
effects on the livelihoods of respondents. In the past year, 94.6 per cent
of Harare residents and 88.9 per cent of Bulawayo residents suffered some
degree of cash income shortages, and 66.3 per cent and 60.2 per cent of res-
idents in the respective cities suffered food shortages. Although outsiders
may question the reliability of this sort of opinion survey in Zimbabwe,
the respondents were eager to express their ideas, with no indication of
physical threats affecting their attitudes in the survey process. However, cau-
tious interpretation of the results is required when we deal with historically
sensitive questions related to violence and ethnicity.
Popular perceptions about the legacy of racial HIs and the governance
failure of the ruling ZANU(PF) were quite mixed. As the drastic land reform
became an accomplished fact, it seems that the aversion to colonial tradi-

tions has receded substantially, though not disappeared, in the minds of
urban Zimbabweans; 48.0 per cent of all respondents answered that the
British colonial legacy had more or less contributed to the country’s current
problems, while 35.6 per cent answered that this was not the case. However,
51.5 per cent of all respondents expected that relations between black and
white Zimbabweans would be better in future. People also shared pride in
the history of the liberation struggle; 82.7 per cent of Shona, 65.6 per cent
of Ndebele and 83.5 per cent of other respondents answered that Zimbabwe’s
liberation struggle should be remembered in an appropriate way. At the
same time, half of the respondents pointed their fingers at the ruling party;
42.6 per cent of Shona, 50 per cent of Ndebele and 38.6 per cent of oth-
ers answered that ZANU(PF) was chiefly responsible for the current political
problems, while the rest indicated other factors or did not give answers.
There was no significant generational difference in the responses to these
questions.
Given the demise of the historical privilege of white Zimbabweans, there
remain two major issues that may bring about political ruptures in the
future: ethnic divisions and class differentiations within the population of
black Zimbabweans. Regarding the ethnic divisions, it is obvious that the
Ndebele had a strong sense of marginalisation relative to the majority Shona
group. As a matter of fact, from 1980 to 2000, major provinces including
Matabeleland and Midlands, the stronghold of Ndebele people, were almost
constantly represented in the cabinet. The proportion of ministers from
Matabeleland even increased during the decade after Gukurahundi was over
(Laakso, 2003: 124). However, the ministers from the Ndebele regions were
largely those politicians who were co-opted by ZANU(PF). As shown in
Figure 5.6, during the early 2000s when the land occupation spread, the
representation of the Ndebele regions was relatively weak, and more than
one-third of the cabinet posts were captured by the MPs who were appointed
by President Mugabe with no constituency.
With the establishment of the Government of National Unity in 2009,
many MDC politicians joined the government, and the regional representa-
tion of the Zimbabwean cabinet ministers changed drastically. A substantial
number of MPs elected in the two main cities, Harare and Bulawayo, joined
the cabinet, and there were also increases in the number of ministers from
the Ndebele region. However, given that the government was under draco-
nian control of Mugabe, in our survey, Ndebele people were still sceptical
about the inclusiveness of the new political dispensation; 68.8 per cent of
Ndebele, 11.7 per cent of Shona and 25.6 per cent of others thought that
their own ethnic group was underrepresented in the national government,
and 54.7 per cent, 8.6 per cent and 22.7 per cent of the respective groups
felt that their group was underrepresented in the armed forces. As shown
in Table 5.7, more than 20 per cent of the Ndebele placed more weight
Yoichi Mine 119
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Matabeleland North Matabeleland South Mildlands Bulawayo

Harare Manicaland Masvingo Mashonaland West
Mashonaland East Mashonaland Central Non Constituency
Figure 5.6 Regional distribution of cabinet posts in Zimbabwe

Note: Excluding President, Prime Minister, Provincial Governors and Deputy Ministers.
Source: Zimbabwe Parliamentary Debates, various years (December).
on their Ndebele consciousness than on the national consciousness, even

though nearly 30 per cent thought they were only Zimbabweans, indicating
that their loyalty is torn between nationhood and ethnicity.
Despite some presence of Ndebele particularism and their sense of vic-
timisation, Zimbabwean city dwellers, including the Ndebele, did not reveal
strong ethnic hostility towards each other; 96.0 per cent of all respondents
answered that they were comfortable working with people from any other
ethnic group, and 76.4 per cent said that they would not object if their
daughter or sister were to marry a person from another ethnic group. Eth-
nic relations in Zimbabwe were considered good by 64.2 per cent of Shona,
42.2 per cent of Ndebele and 61.9 per cent of others, while they were con-
sidered bad by 16.7 per cent of Shona, 26.6 per cent of Ndebele and 16.5
per cent of others.
Questions about economic inequalities were also asked; 40.1 per cent of
Shona, 67.2 per cent of Ndebele and 47.2 per cent of others agreed that
120
Table 5.7 National and ethnic identification of Zimbabwean city dwellers (%)
Group I feel only I feel more I feel I feel more I feel only I don’t know Total
Zimbabwean Zimbabwean equally [ethnic [ethnic
than [ethnic Zimbabwean identity] identity]
identity] and [ethnic than
identity] Zimbabwean
Shona 36.4 15.4 41.4 4.9 1.9 0 100.0

Ndebele 29.7 7.8 40.6 12.5 9.4 0 100.0
Others 26.7 14.8 46.0 8.5 2.8 1.1 100.0
All respondents 31.1 13.9 43.3 7.7 3.5 0.5 100.0
Source: JICA surveys.

Yoichi Mine 121
the government should give extra economic assistance to poorer ethnic

groups. In terms of their group’s socioeconomic situation, 46.9 per cent
of Ndebele thought that it was worse than that of other groups, while
only 9.9 per cent of Shona thought that their situation was worse than
that of others. In respect of the class factor, 21.6 per cent of Shona, 50.0
per cent of Ndebele and 22.7 per cent of others agreed with the statement
that class division was more important than ethnic division in Zimbabwe.
An overwhelming majority, 93.8 per cent of all respondents, agreed that
corruption had become a more serious problem in the past ten years.
A relative majority of all respondents felt closest to MDC-T; 46.9 per cent
of Shona, 39.1 per cent of Ndebele and 47.2 per cent of others sup-
ported MDC-T, while 11.7 per cent of Shona, 1.6 per cent of Ndebele
and 9.1 per cent of others gave their support to ZANU(PF), indicating
that MDC-T’s support base extends across ethnicities in urban areas. Peo-
ple with a higher educational background tended to have higher lev-
els of support for MDC-T. Very positive perceptions of the coalition
government, the Zimbabwean GNU, were expressed regardless of party
support; 81.3 per cent of all respondents answered that the opposition
entering into a coalition with ZANU(PF) had been a necessary decision.
Although 27.6 per cent of all respondents preferred that the inclusive
government should be disbanded due to its ineffectiveness, 51.2 per cent
were negative about such an immediate change at the time of this
survey.
By and large, Zimbabweans appear to be forward-looking in spite of
their daunting political troubles; 73.5 per cent of Shona, 59.4 per cent
of Ndebele and 59.1 per cent of others agreed that the emergence of a
Zimbabwean nationality had been successful, and 73.1 of all respondents
were very or somewhat positive about the future of Zimbabwe. With regard
to xenophobia, 53.5 per cent of respondents said that the relations of
Zimbabweans and foreign Africans in Zimbabwe were good, while 19.2 per
cent indicated the relationship was bad – though those who thought they
were worse off tended to be more pessimistic. Regarding South Africa’s
involvement in resolving the political problems in Zimbabwe, 69.9 per cent
of all respondents were positive, while 15.9 per cent expressed negative
opinions.
5.5 Conclusion: Institutional choice and development
Several practical lessons can be drawn from the discussion in this chapter.
First, power-sharing by all major political parties appears to be useful as
a short-term refuge from cumulative violence, especially when there is a
pressing need to accommodate losing parties still capable of using vio-
lence to destabilise the regime. The security apparatuses developed by the
white minorities in Zimbabwe and South Africa were neutralised effectively
through the emergence of the Lancaster House regime in 1980 and the for-
mation of the South African GNU in 1994 in the respective countries. The
important feature of the South African GNU would be that the coalition
government was formed after a consensus was hammered out through a
period of informal and formal negotiations among local stakeholders. The
1994 elections in South Africa were thus realised as the occasion of the con-
summation of sharing power, rather than as a fighting arena that would
necessitate power-sharing as an ex post solution. The modalities of informal
power-sharing also seem to have taken root in South African politics and
institutions, and contributed greatly to the political stability of the country
after apartheid. One important defect of a grand coalition, however, is that
political processes tend to be dominated by elite talks behind closed doors
to the exclusion of popular participation. In this regard, it is important to
recognise the temporal nature of a grand coalition and to specify its time
boundary. Indefinite extension of a GNU presided over by Robert Mugabe
would be a nightmare for Zimbabwe.
The second lesson is that a ‘winner-takes-all’ principle that is contingent
on a small constituency system could be dangerous in culturally divided
societies, as demonstrated in the successive FPTP elections in Zimbabwe.
When the choice between ‘yes’ and ‘no’ is a close game, ‘wasted’ votes may
amount to nearly half of the votes cast, and the electorate tends to remain
deeply divided even after the campaign is over. The same applies to national
referendums and presidential elections, which are fought in a nationwide
single constituency. This is why elections based on a PR system, in which
political opinions are represented proportionally, are considered superior in
countries in volatile transition. The PR system may also contribute to a pro-
gressive change in the rules of national politics beyond vested interests. For
example, after the 2009 elections in South Africa, 172 of the 400 seats in the
national assembly were held by women due to the gender quota introduced
by the ANC to its party list of candidates, pushing South Africa up to the
fourth in the global ranking of women’s equitable representation in parlia-
ment. This kind of quota system can more easily be adopted in a PR system
than under a FPTP system.
The negative side of a PR system is that the accountability of representa-
tives to the electorate tends to be weak and indirect, as the MPs are inclined
to turn their faces not to the voters but to the party leaders. In our sur-
veys in Zimbabwe where the FPTP system is adopted, 77.6 per cent of all
respondents answered that consideration of the personal qualities of candi-
dates was important when they cast their vote, and 65.7 per cent answered
that the political party was important, while 22.6 per cent said that ethnic
considerations were important. A PR system reduces such opportunities for
the electorate to examine the individual qualities of representatives. This
weakness of the PR system is the very point emphasised by Mandela in
1999 in his farewell speech to the South African Parliament on the day of
Yoichi Mine 123
his retirement. A government team then submitted a report that recom-

mended a multiple-seat constituency system for 300 seats combined with
a PR system for 100 seats as an alternative electoral system, though the
actual shape of a future electoral system is still a moot issue in the coun-
try (Electoral Task Team, 2003; Southall, 2004).14 As there is no universally
best electoral system, every country is justified in developing a hybrid sys-
tem by trial and error. It would also be useful to elaborate a time sequence;
for example, after a PR system has allowed political stability to take root,
elements of the majoritarian system can be carefully introduced to pro-
ceed towards a more dynamic democracy. In this regard, it is evident that
the overnight, top-down shift from a PR to an FPTP system in Zimbabwe
after independence undermined the democratic quality of governance in a
dangerous way.
Last but not least, it must be emphasised that the quality of develop-
ment is the key to political stability in both countries in the longer term.
A society in which more than one-third of the working-age population is
deprived of proper job opportunities, as is the case in today’s South Africa
and Zimbabwe, can never be sustainable, irrespective of the short-term
effects of the choice of political institutions. In South Africa, the ANC now
espouses the notion of a democratic developmental state, which draws on
the success of the emerging East Asian economies and the Scandinavian wel-
fare states and places more value on state intervention to mobilise resources
for sustained growth, combined with democratic accountability (Edigheji
ed., 2010). The issue of designing appropriate political institutions should
be firmly located in such a broader framework.
However, in any part of the world, developmentalist initiatives would
not work without being supplemented by bottom-up initiatives of capital
accumulation: not those meant for the acquisition of personalised wealth,
but geared to the expansion of production capacity. In contrast with the
stagnant entrepreneurship conditioned by the historical legacy of apartheid
in South African townships, it has been reported that Zimbabwean small
farmers in new resettlement areas are gaining ground in successful agricul-
ture, engaging in various off-farm activities, connecting to local markets and
investing in land and livestock, in the face of the political and economic tur-
moil of the 2000s (Scoones et al., 2010). While we can learn much from the
experience of democratic governance in South Africa, the troubled history of
modern Zimbabwe also abounds in practical lessons. If the legitimacy of the
Zimbabwean state is restored at some point in the future, through the adop-
tion of a truly inclusive constitution as a landmark event of the country’s
‘return home’,15 grassroots peasant initiatives with the necessary support of
the state may turn out to be a powerful engine of development for the entire
Southern African region. Institutional engineering is critically important for
peace and stability in any country, but this is only one part of the overall
processes to come.
Notes
1. The urbanisation rate of the total South African population reached 61.7 per cent
in 2010, much higher than 38.3 per cent in Zimbabwe and the average figure for
sub-Saharan Africa, 37.0 per cent (UNDP, 2009a).
2. For the general trend, see Seekings and Nattrass (2005: 300–339). Even the official
unemployment rates excluding the chronically unemployed are near 30 per cent.
3. This proportion was less than half that of the white population in South Africa.
In addition, in Zimbabwe, a sizeable ‘buffer’ population like the Coloureds and
Indians in South Africa is absent. The proportion of white Zimbabweans has
probably declined to less than 0.5 per cent by now, although exact population
statistics are non-existent for the moment.
4. For a detailed description of the complex ethnic power struggles within the lib-
eration movement in Zimbabwe, see Sithole (1999). The vibrant urban politics
during the period up to the 1960s was overshadowed by the sectarian liberation
politics (Scarnecchia, 2008).
5. In Zimbabwe, the Lancaster House Agreement in 1979 was mediated by the
British government, while the formation of the GNU in 2009 was mediated by
the Southern African Development Community (SADC). In South Africa, theo-
retical blueprints for a new political dispensation proposed by American political
scientists, Lijphart (1985) and Horowitz (1991), did play positive roles. Although
the 1994 regime in South Africa was closer to the consociational prescriptions of
Lijphart, one of the major four components of his proposal, mutual veto, was
rejected in the course of negotiations.
6. The head of the state of South Africa is called the president, but is elected in
parliament like a prime minister in other countries. Therefore, variable D (Head
of Government) in South Africa is rated as 2 instead of 1 (Chapter 2, Table 2.1).
7. Respected Afrikaner economic historian Sampie Terreblanche (2002: 95–124) and
leading Canadian anti-globalist Naomi Klein (2007: 245–274) expressed the same
view that the ANC’s compromise with the business sector during the backroom
negotiations in the early 1990s was the critical turning point at which the former
became trapped in free-market orthodoxy.
8. The draft constitution was a cocktail of strong poison and potentially effective
medicine. In addition to the clause on land reform, in which compensation to
white farmers was to be paid out by the British government, the document con-
tained a Bill of Rights, electoral reform including the partial introduction of a PR
system, the establishment of a senate, and the inauguration of the post of prime
minister. The presidency was to be restricted to two terms, though this would
be applied only after the enforcement of a new constitution, enabling Mugabe
to stay in power. For the tension between the top-down approach of the gov-
ernment and the bottom-up approach of civil society in the aborted process of
constitutional reform in the 1990s, see Sachikonye (2004). The new constitution
approved in the referendum of March 2013 is downloadable from: http://www.
copac.org.zw/ (last accessed 1 May 2013).
9. The war veterans have always been relatively autonomous from the ruling ZANU
elite as demonstrated in Kriger (2003) and Sadomba (2011) from different angles.
10. Although the introduction of a quota system and bicameralism contributes to
making Zimbabwe fall within the category of a power-dispersing (PD) regime
(Chapter 2, Table 2.2), the reforms initiated by the ruling party were clearly meant
to concentrate the power base of ZANU(PF). The credibility of the election results
Yoichi Mine 125
has also been questionable, but non-partisan observers agreed that the May 2008
elections (Table 5.6) were conducted in a relatively free and fair environment
(Masunungure ed., 2009: 73–75), to the effect that the result could be a good ‘cen-
sus’ of people’s preferences of political parties. Terrible violence occurred before
the subsequent presidential run-off in the same year.
11. For a detailed record of racial incidents in South Africa since 1994 as well as a
collection of interviews with opinion leaders on the topic of racism, see Holborn
(2010).
12. Of the respondents, 73.9 per cent of Coloureds, 73.2 per cent of Africans and
27.1 per cent of whites answered that their neighbourhood was a very important
part of their identity. In this survey, the refusal rate was exceptionally high for the
residents of white suburbs, 65 per cent, laying a heavy burden on enumerators.
13. In Harare, 45.2 per cent of respondents identified themselves as Shona, 28.2 as
Zezuru and 10.5 as Karanga. In Bulawayo, 54.6 per cent identified themselves as
Ndebele, 26.9 per cent as Shona and 7.4 per cent as Karanga. Although Shona-
speaking peoples like the Zezuru and Karanga are often classified as being part
of the Shona, the ethnic classification in this survey is based on their own self-
identification.
14. The proposed multiple-seat system with a single non-transferrable vote is similar
to the former electoral system of the Japanese lower house, which has now been
converted to a combination of a small constituency system and a PR system.
The same kind of medium-sized constituency system is currently in use in a few
countries such as post-conflict Afghanistan.
15. For the prime importance of state legitimacy, see Moyo (2010). Given what
Mugabe symbolises, it seems to be difficult for the government of Zimbabwe to
restore full legitimacy in the eyes of the Western countries as long as Mugabe
himself is part of the leadership (Ndlovu-Gatsheni, 2009b: 343). In any case, the
‘Mugabe factor’ is only a part of the history, and ‘there is blood on many hands’
(Bourne, 2011: 236).
6
Is Ethnic Autonomy Compatible
with a Unitary State? The Case
of Uganda and Tanzania
Yuichi Sasaoka and Julius E. Nyang’oro
6.1 Introduction
Uganda and Tanzania, two countries in East Africa, have experienced sharply
contrasting conditions of political instability and stability. On the one hand,
it is estimated that more than half a million Ugandans lost their lives
between 1971 and 1986, during the period of Idi Amin’s dictatorship and
of the subsequent ‘liberation war’ waged by the National Resistance Army
(NRA) led by the current president, Yoweri Museveni. After the victory of
Museveni’s army, a serious armed resistance continued in the country’s
Northern region, and Uganda is prone to potentially violent ethno-regional
conflict even today. On the other hand, Tanzania has been largely peaceful
for many years, going back to the days of the late president, Julius Nyerere.
As evidenced in our perceptions surveys, ethnic identity and national iden-
tity compete on even terms in Uganda, while common identity as a single
nation has firmly taken root in Tanzania.
Despite such contrasting situations, Uganda and Tanzania share several
common traits. Both countries have had to deal with influential regional
separatist movements that sought political and cultural autonomy: the
Buganda Kingdom in Uganda and the islands of Zanzibar in Tanzania. These
movements posed serious challenges to the post-independence legitimacy
of the governments of these countries. Furthermore, prominent leaders in
these countries, President Nyerere of Tanzania and President Museveni of
Uganda, were ardent proponents of African socialism and integrationist
nation-building, which were meant to go beyond ethnic particularism
and secessionism. More recently, these countries have accepted multi-party
democracy and started to implement decentralisation policies as a form of
governance reform.
The objective of this chapter is to explain the reasons why these coun-
tries, neighbours on opposite sides of Lake Victoria, have taken remarkably
126
Yuichi Sasaoka and Julius E. Nyang’oro 127
divergent paths in terms of conflict, in the presence of these similarities

and common challenges. A useful starting point is the fact that societal
differences can be a powerful factor. Ugandan society has been ethnically
polarised for a long time, and several ethnic polities, including the Buganda
Kingdom, have vied for power, sometimes forming fragile alliances. In con-
trast, ethnic identity in Tanzanian society is understood to be extremely
fragmented, and, with the exception of Zanzibar, is thus characterised by
an impressive lack of ethnic antagonism. Recognising that such contrast-
ing features of society explain many of the political configurations in the
two countries, this chapter, however, places the emphasis on the critical
importance of political norms and leadership styles, both formal and infor-
mal. Even when the formal features of the political institutions of the two
countries resemble each other, different norms and conventions would make
similar institutions function in starkly different ways. In recent decades, the
contrasting behaviours of political leaders in these countries seem to have
aggravated ethnic tensions in Uganda, but contributed to the consolidation
of peace in Tanzania.
In this chapter we present our argument as follows. First, in Section 6.2
we describe the formation of identity groups in these countries, contrast-
ing the tenacity of ethnic antagonism in Uganda with the absence of such
conflict in Mainland Tanzania (Tanganyika). Next, in Section 6.3, based on
the discussion of the historical configurations of ethnic groups, contempo-
rary mappings of the social and economic horizontal inequalities (HIs) in
the respective countries will be provided. Section 6.4, the main part of this
chapter, deals with the question of political processes, leadership and party
structure, comparing the political norms as well as the actual functions of
formal and informal institutions in the two countries. Section 6.5 examines
the outcome of perceptions surveys to illuminate the group consciousness
and political and social attitudes of ordinary people in both countries.
Section 6.6 concludes.
6.2 Formation of identity groups in Uganda and Tanzania
6.2.1 History of group antagonism in Uganda

Uganda has a population of 35.6 million, divided into 21 major ethnic
groups. Of these groups, the Baganda (plural) who live in Uganda’s Central
region adjoining Lake Victoria are the most numerous. The word ‘Baganda’
refers to the people, ‘Muganda’ (singular) to a person, ‘Buganda’ to the
kingdom, and ‘Luganda’ to the language; the King of Buganda is called
the ‘Kabaka’. The Buganda territory includes Kampala, the capital and the
largest city in the country, and the name of the country, Uganda, derives
from the name of this influential kingdom. Other influential ethnic groups
include the Banyankole, Basoga, Bakiga, Iteso, Langi, Banyoro, Batoro and
Acholi (see Table 6.1). These ethnic groups tend to live in separate areas, and
128 Uganda and Tanzania
Table 6.1 Ethnic groups in Uganda (%)
Group name Area Year
1991 2002
Baganda Central (Southern) 18.0 16.9

Banyankole Western 9.9 9.5
Basoga Eastern 8.2 8.4
Bakiga Western 8.3 6.9
Iteso Eastern 6.0 6.4
Langi Northern 5.9 6.1
Acholi Northern 4.4 4.7
Others 39.3 41.1
Source: UBOS (1995, 2005).
Uganda’s major cities and towns are inhabited largely by single dominant
ethnic groups. For example, about 80 per cent of the inhabitants of Gulu
in Northern Uganda are thought to be Acholi, and the same percentage of
the inhabitants of Mbarara in the Western region to be Banyankole (UBOS,
1995; Moses 2007).
The Buganda Kingdom, having hierarchical authority and a sophisticated
social order, was established as early as the fourteenth century, reached its
zenith in the eighteenth and nineteenth centuries, and then entered a period
of ‘modernising autocracy’ (Apter, 1961). At the time of Uganda’s indepen-
dence, ethnic cleavage was palpable between the subjects of the Buganda
Kingdom and the rest of the country. During both the colonial period
and immediately after independence, Buganda was granted a ‘special status’
within Uganda. It is worth recalling that soon after independence (when
Queen Elizabeth of the UK relinquished the post of head of state), King
Mutesa II of Buganda, supported by the Kabaka Yekka (KY: King Only) party,
became the first (non-executive) president of the country and the region
was offered semi-autonomous status under the 1962 Constitution. Mean-
while, a conflict between the Baganda elite and the central government led
by Prime Minister Milton Obote, a politician from the Langi ethnic group in
the North, emerged over various constitutional issues, especially on the posi-
tion of Buganda within the now independent, unitary state of Uganda. The
conflict escalated into violence when two counties of Buganda were trans-
ferred to Bunyoro after a referendum in 1964. The ‘Buganda Crisis’ of 1966
followed, and eventually Obote, who represented the ruling Uganda People’s
Congress (UPC), gave an order to arrest Mutesa and to abolish all traditional
kingdoms in Uganda including Buganda, Ankole, Toro, and Bunyoro. The
Busoga chieftaincy was also abolished in 1967.
The series of political struggles that took place in the 1960s was a harbinger
of subsequent conflicts between the people of socioeconomically advanced
Buganda and the relatively deprived Northern ethnic groups, particularly

the Langi and the Acholi. During the period of British colonial rule, the
Northern groups had been dissatisfied with the colonial regime’s use of the
Baganda elite to carry out the policy of indirect rule, a system that was per-
ceived by the Northerners as having given the Baganda license to abuse
them. After independence, the situation was thrown into reverse; Obote used
Northerners as military officers to contain the influence of the Baganda elite.
However, in 1971, Obote was deposed in a coup d’état plotted by his military
commander, Idi Amin, who also had a Northern background. Often cari-
catured as the prototypical ‘African dictator’, Amin slaughtered dissidents,
devastated the economy, and eventually fled into exile after the Uganda–
Tanzania War (1978–1979), one of the rare cases of inter-state war in Africa.
The political vacuum in Uganda following the victory of the Tanzanian army
and Ugandan anti-Amin forces was filled by the reinstatement of Obote, who
had been living in exile in Tanzania but was elected as president in 1980
upon his return.
With the new Obote regime, people felt that they had been returned to
the troubled past, and the Baganda were clearly frustrated by the continua-
tion of rule by Northerners. The NRA, led by Yoweri Museveni, a charismatic
young activist from the Ankole pastoral area, refused to accept the elec-
tion results, arguing that they were fundamentally flawed. He proceeded
to organise guerrilla warfare from within the Baganda stronghold against
the Obote regime and then against the government of Obote’s successor,
Tito Okello from Acholi. In 1986 guerrilla forces based on an alliance of
Baganda (Central) and Banyankole (Western) triumphantly entered the city
of Kampala, and President Museveni, as the leader of the National Resistance
Movement (NRM), launched what he called a ‘broad-based government’.
However, the Museveni government seems to have failed to build up a
truly inclusive national government in two main ways. First, apprehensive of
being alienated and punished by the government now controlled by Central
and Western people, some militant Northerners, especially Acholi followers
of the Okello government, started to arm themselves against the NRM gov-
ernment. As a result of a prolonged guerrilla war and human rights abuses
committed by a Northern guerrilla army known as the Lord’s Resistance
Army (LRA), Northern Uganda was physically devastated and continued to
lag behind in terms of economic and social development (Behrend, 1999;
Doom and Vlassenroot, 1999; Van Acker, 2004).
Secondly, over time, a fissure started to emerge between the Baganda and
the Banyankole. During the conflict against Obote in the early 1980s, the
Baganda seemed to have expected that they would be granted autonomous
status after the victory in exchange for their costly support of the NRM.
When Museveni and the NRM failed to live up to this expectation, politi-
cal tension ensued between the two ethnic groups (Green, 2006; Mutibwa,
2008a). At present, there seems to be a widespread perception that people
in the West, the home region of President Museveni, are favoured by the
government. Thus, the potential conflict between the Baganda and the West-
erners, as well as the question of the relative deprivation of the Northerners,
is of great significance in present-day ethnic relations in Uganda.
As the Buganda Kingdom has historically been the centre of political and
economic power in Uganda, many Baganda aspire for the political auton-
omy of Buganda within the country. When President Museveni acquiesced
to the reintroduction of monarchies in the early 1990s, there was strong dis-
satisfaction within Buganda at the limited nature of the authority granted to
these monarchies. Baganda royalists had hoped that the new dispensation
would essentially be a return to the status quo ante of the pre-1966 period,
when Buganda was almost a country within a country. The call for a federal
system in Uganda is thus driven mainly by Baganda royalists who seek the
restoration of the political autonomy of the Buganda Kingdom.
6.2.2 Tanzania, with a focus on Zanzibar

According to the 2000 census, Tanzania had a population of 33.7 million,
and this was projected to rise to about 40 million by 2010. Table 6.2
shows the major ethnic groups in Tanzania. There are roughly 120 eth-
nic groups in the country, but in terms of political influence there is no
one dominant group comparable to the Baganda in Uganda. The largest
group, the Sukuma, are said to constitute no more than 10 per cent of the
population; other numerically significant ethnic groups include the Haya
(northwest), Nyakyusa (southwest), Nyamwezi (central), Chagga (northeast)
and Makonde (south). It must be noted that this ethnic breakdown of the
Tanzanian population is only guesswork, because Tanzania abandoned the
practice of identifying people on the basis of ethnicity four decades ago.
Furthermore, in terms of geographical residence, there is sufficient mobil-
ity within the population as to make ethnic and geographic identity a very
tenuous proposition (Nyang’oro, 2006).
The ethnic configurations of Uganda and Tanzania appear to be quite
different. Because of the unifying ideology of African nationalism and the
lack of ethnic mobilisation, antagonism amongst ethnic groups in Tanzania
Table 6.2 Major ethnic groups in Tanzania
Group Name Area
Sukuma North (South of Lake Victoria)

Makonde South-East Coastal
Chagga North (South of Mt. Kilimanjaro)
Haya North (West of Lake Victoria)
Nyamwezi North-Western and Central
is virtually non-existent (Young, 1976; Tripp, 1994; Klugman et al., 1999;

Nyang’oro, 2006). Conflicts are limited to localised economic and envi-
ronmental issues such as access to grazing land. It must be noted that
the numerical power of a certain group does not necessarily matter when
we consider ethnic configurations. The proportions of the population rep-
resented by the Sukuma in Tanzania and by the Baganda in Uganda are
not fundamentally different. A similarly important role was played by both
groups at the time of independence in terms of support for independence
in their respective countries. What has distinguished their political weight
in the post-independence era would rather be the strongly loyal Buganda
identity on the one hand and the lack of any such constructed and politi-
cised identity on the part of the Sukuma on the other. The emergence
of the centralised polity in the heartland of Uganda and the more open,
decentralised nature of Tanzanian society may be at least partly explicable
in terms of population pressure. At present, the total populations of these
countries are very similar, at around 34 to 36 million, while the size of the
national territory of Tanzania is almost four times larger than that of Uganda
(Figure 6.1).
However, one exception to the relative homogeneity of Tanzanian society
can be found in Zanzibar, an archipelago in the Indian Ocean well known for
its rich history of trade networks and flourishing Swahili culture. Zanzibar is
located off the northeast coast of Tanzania and has a population of about
1.5 million: about 3 per cent of the total population of Tanzania. Zanzibar
consists of two major islands, Unguja and Pemba, as well as numerous
small islands. In ethnic terms, the population of Zanzibar is highly mixed.
Everyone in Zanzibar speaks Kiswahili and ethnic identification bears little
relation to political mobilisation. However, in terms of cultural identity, our
own survey indicates that the majority of Zanzibaris, especially on Pemba
Island, refer to themselves as Shirazis, the descendants of ancient immigrants
from the city of Shiraz in Persia (Table 6.3). As the label ‘Arab’ has a conno-
tation of ‘non-African’ (as does ‘Shirazi’ to some extent), the ethnic lines in
Zanzibar are sometimes regarded as being ‘racial’, even though most of the
ancestors of today’s Zanzibaris are thought to have migrated from Mainland
Tanzania. Almost everyone in Zanzibar has been linguistically ‘detribalised’,
including the original permanent residents (the Hadimu and Tumbatu) who
seem to have arrived on the islands from the continental mainland hundreds
of years ago.
Despite its cultural particularity, Zanzibar has long been closely con-
nected with Mainland Tanzania. The Shirazi people are a part of the larger
Swahili people, who inhabit the coast of East Africa and the islands of
Zanzibar and Comoros. The Swahili language is spoken widely in East Africa
and has been ordained to be the official language of Tanzania. Therefore,
the question of Zanzibar’s autonomy is not simply a question of defining
132
SOUTH SUDAN ETHIOPIA
Gulu
UGANDA
SOMALIA
Hoima
Buganda Kampala
Kingdom KENYA
Mbarara
Lake
Victoria
RWANDA
BURUNDI
DEMOCRATIC Pemba Island

REPUBLIC (Zanzibar)
OF THE CONGO
Lake TANZANIA Unguja Island
Tanganyika
Dar es Salaam (Zanzibar)
INDIAN
OCEAN
ZAMBIA
MALAWI
MOZAMBIQUE
Figure 6.1 Uganda and Tanzania
Table 6.3 Major ethnic (racial) groups in Zanzibar
Major ethnic Share %

(racial) group
Unguja Shirazi 54.7

Tumbatu 10.8
Hadimu 8.8
Zaramo 6.8
Others 18.9
Pemba Shirazi 85.5
Swahili 5.3
Sukuma 2.6
Arabs 2.6
Others 4.0

borders and ensuring minority rights, but also has something to do with the
national identity of the Republic of Tanzania. When we discuss the issue
of Zanzibar, the political division between the two major islands is also of
great significance, because the pros and cons of autonomy have been hotly
debated within Zanzibar itself, sometimes with polarised opinions. Broadly
speaking, the CCM (Chama cha Mapinduzi: Party of the Revolution), which
is the ruling party in Mainland Tanzania, has taken root in Unguja, while
the Civic United Front (CUF), the leading opposition in Mainland, has its
stronghold in Pemba (Bakari, 2001; Killian, 2008; Burgess, 2009). In a sense,
the ethnic politics within Zanzibar is a microcosm of Tanzanian politics as a
whole, and this justifies our looking into what has been happening in that
compact space in some detail.
Mainland Tanzania gained independence from Britain as the Republic
of Tanganyika in December 1961, while Zanzibar gained independence as
a constitutional monarchy in December 1963. However, soon after inde-
pendence, the Sultan of Zanzibar was deposed in a violent revolution in
January 1964 (Yeager, 1989; Petterson, 2002), in which about 5,000 ‘Arabs
and Indians’ on the islands were massacred by ‘African revolutionaries’. The
massacre, which was a precursor to the deportation of Indian Ugandans by
Idi Amin, was based on a perception of HIs between the rich Asian minority
and the poor African majority. April 1964 saw the formation of the United
Republic of Tanganyika and Zanzibar, which was then renamed the United
Republic of Tanzania. Given the political sensitivity of prior sovereignty, the
Mainland and Zanzibar formed a power-sharing arrangement. The arrange-
ment was that if the president of Tanzania were to be elected from the
Mainland, then the vice president would come from Zanzibar, and vice versa.
This has been the political arrangement for almost five decades. Zanzibar
was also given a disproportionate number of seats in the National Assembly.
In more recent times, a power-sharing arrangement has been instituted in
Zanzibar where the ruling party shares power with the opposition, as we will
see shortly.
Uganda and Tanzania are somewhat similar with regard to religious
identification in that Christians (Catholics, Protestants) and Muslims are
competing over new converts. In Tanzania, the Muslim population (just
below 40 per cent) is larger than it is in Uganda (12 per cent), and it is
predominant in Zanzibar (98 per cent) and the coastal areas.1 African Tra-
ditional Religions (ATRs) are also practised in both countries. For example,
many Baganda practice traditional rituals, as does almost every other eth-
nic group. In terms of language, there is a major contrast between the two
countries: virtually everyone in Tanzania speaks Kiswahili as the official and
national language while in Uganda, Kiswahili is mostly spoken in the East
(towards Kenya) and in the North (as the language of trade and the military).
In Uganda, the teaching language at the secondary school level and beyond
is English.
6.3 Socioeconomic horizontal inequalities
As we discussed, in summary, there are three poles in the ethnic mapping of

Uganda: the Central, North and West regions. It is surmised that the Baganda
in the Central region feel that they are victimised in terms of political and
cultural-status HIs, and that the Northerners suffer in terms of all dimen-
sions of HIs. In contrast, Tanzanian society has been generally peaceful with
no serious ethnic antagonism, except for the issue of Zanzibar (mostly in
Pemba) where a quest for political and cultural autonomy comparable to the
case of Buganda has been noticeable. Based on this rough sketch of ethnic
relations, this section tries to shed light on the socioeconomic HIs in each
country.
The Ugandan economy collapsed between the mid-1970s and the mid-
1980s, but grew 6.5 per cent in the 1990s and over 7 per cent in the
2000s.2 The Tanzanian economy expanded at a somewhat slower pace in
the 1990s, but the growth rate has been around 5 to 7 per cent since 2000.
In spite of such achievement and recent poverty reduction efforts, Uganda
and Tanzania still rank 143rd and 148th, respectively, of 169 countries on the
Human Development Index (UNDP, 2010). With regard to inequalities, the
Gini coefficient of Uganda in the period 1992–2007 was 0.43, while that of
Tanzania for the same period was no more than 0.35, indicating the egalitar-
ian nature of Tanzanian society (UNDP, 2009a). Tanzania’s Gini coefficient
seems to have been constant over the past 20 years: 0.34 in 1991/92 and
0.35 in 2007 (NBS Tanzania, 1999; 2009). In Uganda, in contrast, there is
evidence that inequalities between regions are widening: the Northern and
Eastern regions lag behind the Central and Western (UBOS, 2010). Below,
we proceed to examine the situations of socioeconomic inequalities between
groups or regions by country.
6.3.1 Regional disparity in Uganda

In Uganda, the overall poverty rate has declined from 56 per cent in 1992
to 31 per cent in 2005/6, and further dropped to 24.5 per cent in 2009/10
(UBOS, 2010). Poverty in the North is still serious largely due to prolonged
armed conflict and low investment in the social sector (Regional Forecasts,
2007). Although the poverty rate in the North declined remarkably from
60.7 per cent in 2005/06 to 46.2 per cent in 2009/10, it is still almost double
the national average (Table 6.4).
With respect to the question of material welfare, almost six in every
ten households in Uganda reported that every member of their household
owned at least one pair of shoes (Table 6.5). This proportion has generally
increased over the three survey periods, especially in the Eastern region.
The most serious situation is found in the Northern region where the figure
is still only 32 per cent despite the recent economic growth. The level of
material welfare in the Western region has become closer to that of the
Table 6.4 National and regional poverty rates in Uganda, 1999/2000–2009/2010 (%)
1999/2000 2002/2003 2005/2006 2009/2010 Major ethnic group
National 33.8 37.7 31.1 24.5

Rural 37.4 41.7 34.2 27.2
Urban 9.6 12.2 13.7 9.1
Central 19.7 22.3 16.4 10.7 Baganda
East 35.0 46.0 35.9 24.3 Basoga, Iteso
West 26.2 31.4 20.5 21.8 Banyankole, Bakiga
North 63.7 63.3 60.7 46.2 Langi, Acholi
Note: The data for 1999/2000 and 2002/2003 exclude some districts.
Source: UBOS (2003, 2006, 2010).
Table 6.5 Household members with shoes (%)
2002/2003 2005/2006 2009/2010
National 44.7 49.7 58.1

Rural 37.0 43.1 50.9
Urban 82.2 81.0 89.3
Kampala 94.3 92.5 97.4
Central 60.9 69.5 76.2
Eastern 25.2 34.1 45.7
Northern 22.3 21.4 31.8
Western 50.6 54.9 62.7
Source: UBOS (2010).
Central region. The Universal Primary Education policy introduced in 1997

has had a significant impact on school enrolment across Uganda, as shown
in Table 6.6. There are still disparities between the regions, but the Central
region is no longer in a dominant position in terms of student enrolment.
The performance of the Northern region has improved remarkably in this
regard but still lags behind other regions. The highest performance has
been achieved in the Eastern region where the enrolment rate is close to
90 per cent.
Since the end of the 2000s and the defeat of the LRA, a Northern rebel
group led by Joseph Kony, the Acholi region has started to be reintegrated
into the rest of Uganda, and the social conditions in the North are get-
ting much better with the assistance of international donors. However, the
potential for future conflict will remain if lingering issues to do with poverty
and cultural indignity are not addressed sufficiently. Meanwhile, the sense
of mistrust by the Baganda towards Westerners also seems to be increas-
ing, even though clear evidence of favourable treatment of Westerners at
large is not discernible in the above statistics. The perception of Westerners’
Table 6.6 Primary school net enrolment rates by region (%)
Girls Boys
1992/1993 2002/2003 1992/1993 2002/2003
National 58.9 85.0 63.6 84.2

Central 69.1 85.3 69.0 85.1
Eastern 62.0 89.7 66.7 89.4
Northern 39.7 72.5 54.4 73.5
Western 60.4 88.0 62.3 84.3
Source: World Bank (2006).
domination at the elite level, however, is not groundless. In 2008, 74 per cent
of Ugandan military commanders were from the West, 17 per cent from the
Central region, and 9 per cent from the North (Tumushabe, 2009: 19). Fol-
lowing the overthrow of the Amin and the Obote regimes, Northerners who
had occupied high military positions were replaced by Westerners, especially
the Banyankole. The Baganda continue to be excluded from the military, as
has always been the case since colonial times.
6.3.2 Trends in Tanzania

Over the past decade, the pace of poverty reduction in Tanzania has been
very slow. Table 6.7 shows the trend in Mainland Tanzania; the poverty rate
declined from 35.7 per cent in 2000 to 33.6 per cent in 2005/6. Over this
period, because of increases in the population, there has actually been a rise
in the absolute numbers of poor. The regions with high rates of poverty
include Mara in the north and Singida in the central region, while poverty
has been much alleviated in urban areas such as Dar es Salaam, the cap-
ital city, and Arusha. Poverty reduction in Dar es Salaam has been much
more effective than in other areas in the past two decades. It must be noted,
however, that inequalities between urban and rural areas are not necessarily
perceived as inequalities between cultural or ethnic groups.
Table 6.7 Poverty headcount in mainland Tanzania (%)
Population share Poverty headcount
2000/2001 2007 1991/1992 2000/2001 2007
Dar es Salaam 5.8 7.5 28.1 17.6 16.4

Other urban 13.8 17.7 28.7 25.8 24.1
Rural 80.4 74.5 40.8 38.7 37.6
Mainland Tanzania 100 100 38.6 35.7 33.6
Source: NBS Tanzania (1999, 2002, 2009).

Table 6.8 Primary school net enrolment rates (%)
Male/Female Total
1999 2005
Mainland 53.1 73.2

Urban 70.7 85.0
Rural 49.1 69.9
Zanzibar 66.9 71.4
Unguja 71.5 75.3
Pemba 62.5 64.4
National 53.5 73.1
Source: NBS Tanzania (1999); NBS Tanzania and ORC Macro

(2005).
Due to the paucity of statistical data, it is not easy to make a comparison

of economic situations between Mainland Tanzania and Zanzibar. Table 6.8
illustrates the uneven impact on school enrolment rates of the Universal Pri-
mary Education Policy that recommenced in 2002. Zanzibar is said to have
enthusiastically promoted a policy for universal education from an earlier
period (Burgess, 2009). The islands have made a good effort in the education
sector, having adopted compulsory and free education for both the primary
and junior secondary levels, while the Mainland has maintained compul-
sory education for primary schools only. Likewise, access to mass media in
Zanzibar is thought to be better than on the Mainland, possibly due to a
better literacy rate and the small size of the islands (NBS Tanzania, 1999;
NBS Tanzania and ORC Macro, 2005). Between 1991 and 2008, Zanzibar
achieved an average economic growth rate of 6.4 per cent, which is compara-
ble to the growth performance of the Mainland (UNDP, 2009b: 34). However,
Table 6.8 shows that the improvement in the primary enrolment rates was
much slower in Zanzibar than on the Mainland in the early 2000s. More-
over, in 2004–2005, the unemployment rate in Zanzibar, at 32.9 per cent,
was twice as high as that on the Mainland, which was 16 per cent (NBS
Tanzania and ORC Macro, 2005: 39). Economic stagnation has the poten-
tial to be a source of strong grievances on the part of the people of Zanzibar
towards the people on the Mainland.
There is also an evident gap in economic performance between Unguja
and Pemba. For example, the per capita income in the Micheweni District of
Pemba is around half that of the Mjini District of Unguja. At the time of inde-
pendence, Pemba is said to have enjoyed higher income levels than Unguja
due to the export of cloves, but the clove industry declined steadily in the
1970s (Chachage, 2000). At the moment, the tourist industry in Unguja is
developing rapidly, thereby aggravating the income gap between the two
islands (UNDP, 2009b).
6.4 Political processes and the choice of political institutions
However different the HI situations in Uganda and Tanzania, these coun-

tries share certain commonalities, as stated at the outset of this chapter. As a
young student activist, Museveni and his comrades chose to study not at
Makerere University in Kampala but at the University of Dar es Salaam,
because of their admiration for Nyerere: they were attracted by ‘Julius
Nyerere’s distinctly more positive and progressive leadership’ (Museveni,
1997: 19), although Museveni placed a greater emphasis on viewing pol-
itics from the perspective of class struggle than nationalism. Museveni
and his comrades received military training from Frente de Libertação de
Moçambique (FRELIMO), the then Marxist Mozambican liberation move-
ment, which operated from bases in neighbouring Tanzania with the support
of Nyerere’s government. After Museveni came to power in 1986, the famous
no-party ‘movement’ system was introduced in Ugandan politics. In that sys-
tem, candidates were prohibited from engaging in election campaigns under
the name of specific political parties, and voters made open queues behind
their preferred candidates instead of using ballot boxes. The no-party system
is said to have been modelled on the one-party electoral system introduced
under Ujamaa socialism in Tanzania (Kasfir, 1998); however, the Tanzanian
system did not require the crude lining up behind preferred candidates.
Tanzania adopted the multi-party system in 1992, and Uganda followed suit
after a referendum in 2005. In both countries, the ruling parties have main-
tained political power for a long period of time. When we look only at formal
institutions, the positions of the two countries are still very close to each
other (Chapter 2, Figure 2.6).
However, these countries project dissimilar images of stability and instabil-
ity, and in this chapter, such asymmetry is to be explained on the grounds of
interactions between formal and informal political institutions, rather than
a simple juxtaposition of them. On the one hand, we may have a situa-
tion in which informal institutions and practices prevail to the detriment of
formal rules and practices. As an example, we take up ethnic politics as an
influential modus operandi in Uganda and the relative freedom from sectarian
politics in Tanzania. On the other hand, informal practices may cause formal
institutions and policies to function in ways completely different from ordi-
nary assumptions, sometimes bringing about adverse effects. In this regard,
we take notice of the nature of leadership and party structure, as well as
the benign effects of power-sharing and federal arrangements in Tanzania in
contrast with the controversial political consequences of decentralisation in
Uganda. Let us discuss these points more in detail.
6.4.1 Significance of ethnic politics

In the early 1980s, the Baganda people, who had suffered under the Amin
and Obote regimes, threw their decisive support behind the NRM. With their
tradition of a historic kingdom, the Baganda have seen themselves as dis-

tinct from other ethnic groups, and this has occasionally elicited feelings
of discomfort among other groups in Uganda. In the initial years of NRM
rule, President Museveni sought to broaden its political base by extend-
ing patronage to, or sharing rent with, a variety of groups in the country
(Blum, 2006; Tripp, 2010). However, the greatest obstacle to the agenda of
the NRM regime remained in the North, where people were fearful that
they could be punished and excluded from the new regime, and eventu-
ally took up arms against the central government. The emergence of the
LRA, the Acholi-based religious movement that turned into a sectarian ban-
dit organisation, could be attributed largely to their feelings of possible
victimisation.
Over the years, grievances have also come to the surface in places other
than the North. As Westerners seemed to be favoured in the allocation of
key government jobs as well as investment, the Baganda and other peo-
ple in the Central and Eastern regions began to feel that they were being
treated unfairly by the new regime. Some Baganda have sharpened their
opposition to the regime in the face of the reluctance of the government to
grant autonomy to Buganda. This is the background to the riots in Kampala
in September 2009, which occurred when the government rejected a pro-
posed visit by Buganda’s King Mutebi to Kayunga, a district where there
was a group seeking secession from Buganda. It is reported that, during the
riot, young Baganda went on a rampage around the town, attacking people
who appeared to be Banyankole, of the same ethnic group as the president
(Lindemann, 2010). However, those Baganda who are not very committed to
the federalist agenda, including Vice-President Gilbert Bukenya (from 2003
to 2011), supported Museveni in suppressing the riots.3 Internal divisions
within Buganda on the basis of social class and religion, as well as differ-
ent views on the status of the symbolic monarchy, account for the lack of a
cohesive ‘Buganda Project’.
President Museveni may have skilfully exploited these divisions, or have
been racked by the vicissitudes of group consciousness beyond his control,
or both. In any case, the central fact is that, beneath the formal discourse
of national integration, ethnic politics has always been rife in Ugandan
national politics. Baganda’s support of Museveni was based on the enormous
sacrifice of the Banyankole and the Baganda during the 1981–1986 guerrilla
war, the so-called ‘Bush War’, against the continuation of the Northerners’
rule. The major battlefield between the government army and the NRA was
the Luwero Triangle, the heartland of the Buganda region, meaning that
Baganda civilians were killed during the war. However, as the memory of
the liberation war gradually recedes, the Baganda support of the NRM is
now far from guaranteed, and the shape of a post-Museveni regime is unpre-
dictable. The ethnic relationships in Uganda have been deeply affected by
multidimensional HIs fraught with security dilemmas.
In contrast, ethnicity has been the least important feature of politics in

Tanzania. This fact has been thoroughly documented (Nyang’oro, 2006).
From its inception the Tanganyika African National Union (TANU), the pre-
decessor of the present ruling party CCM, was regarded as a truly broad-based
nationalist party from the beginning. There was no single dominant ethnic
group within the party; indeed, the majority ethnic group, the Sukuma, was
closely affiliated with TANU. In Tanzania, one sure way of a politician losing
stature as a national leader is to invoke ethnicity as a variable in any public
discourse. As Therkildsen and Tidemand (2007) have noted recently, ethnic
linkage or affiliation has less importance in the recruitment of public sector
personnel in Tanzania. By and large, public discourse in Tanzania focuses
around issues of governance, corruption and the like, and not about HIs on
the basis of ethnic identification. These practices seem to have permeated
into Zanzibar, where there are some degrees of HIs with the Mainland as
well as between the two major islands.
The internal politics of Zanzibar has always been complicated by the mul-
tiple identities of the islanders. The political unity of Mainland Tanzania
and Zanzibar has blurred some of these differentiations, as the ideology
of Ujamaa socialism and integrationist nation-building was promoted by
Nyerere. The introduction of a multi-party system in 1992 revived some his-
torical animosities in Zanzibar, with the people of Unguja being relatively
supportive of the ruling CCM, and the people of Pemba backing the opposi-
tion CUF. The intervention of national CCM leaders, particularly the current
president, Jakaya Mrisho Kikwete, to resolve the crisis has led to the forma-
tion of a government of national unity within Zanzibar, effectively bringing
an end to the conflict (Nyang’oro, 2010). Given that both parties are more or
less at par in Zanzibar, the CCM leaders recognised that they had to coexist
with the CUF for their mutual benefit. In addition, with rumours of an oil
discovery in the offing in Zanzibar, there is an expectation that in the near
future there will be sufficient resources for the development of Zanzibar as a
whole. In spite of the relative historical marginalisation of Pemba, the cur-
rent leadership of Zanzibar is Pemba-heavy: the president of Zanzibar is Ali
Mohammed Shein (CCM) from Pemba, and the first vice-president is Seif
Sharif Hamad (CUF), also from Pemba.
The crisis resolution process in Zanzibar reflects one crucial feature of
the political culture of Tanzania, namely, the style of informal power-
sharing. During the period of TANU/CCM domination, the political leaders
of Tanzania have developed informal channels of conflict resolution through
elite bargaining. Even the launch of the governmental union of Mainland
Tanzania and Zanzibar in 1964 was made possible on the basis of a personal
understanding between Julius Nyerere, president of Tanganyika, and Sheikh
Abeid Amani Karume, president of Zanzibar. The recent intervention by
Mainland political leaders, nearly half a century later, to resolve the conflict
between the CCM and the CUF in Zanzibar was thus no accident. Likewise,
the secret talks held in 2009 between President Amani Abeid Karume, the
son of the Zanzibar’s first president, and Seif Shariff Hamad, the CUF sec-
retary general, were informally arranged but instrumental in helping bring
about the agreement of a coalition government in Zanzibar. The considera-
tion of mutual trust between political leaders thus leads us to the next topic
of our discussion, the question of leadership.
6.4.2 Leadership and party structure

Both Uganda and Tanzania have adopted a presidential system with strong
power being vested in the executive branch of the government (Chapter 2,
variable D). In the domain of formal institutions, this is one of the fea-
tures of majoritarian, power-concentrating (PC) institutions as opposed to
power-sharing, power-dispersing (PD) institutions. However, if the executive
president as leader of the nation tries to avoid top-down decision-making,
seeks a consensus and takes a strong initiative to resolve conflict through
dialogue, such a practice will surely contribute to the prevention of violent
conflict through informal power-sharing. In a similar vein, we may have a
paradoxical situation in which a unipolar party system or even a single-party
system (Chapter 2, variables B and C) can effectively be power-dispersing, if
a dominant, majority party is committed to decentralised decision-making.
Such situations in which a majoritarian institution functions in a power-
sharing way can be illuminated through a comparison of political processes
in Uganda and Tanzania.
Since being inaugurated as president in 1986, Museveni has continuously
served as president of Uganda and is now one of the longest-serving incum-
bent heads of state in postcolonial Africa. His role in reinstating stability and
raising hope among the people in the 1980s and 1990s is undeniable. How-
ever, his decision to run for a fourth term in 2011 elicited strong opposition
in many quarters of the country, even though he won the election with a
clear margin. The NRM is sometimes seen as a personal vehicle for the career
success of the political clique surrounding Museveni, and one caustic critic
has commented that Museveni has succeeded in winning successive elec-
tions through both political skill and the advantages of incumbency, rather
than as a result of the NRM’s popularity in the country (Mwenda, 2007).
In 2003, Museveni convinced his party cadres that the NRM was moving
towards democratisation, but then started to push to abolish the presiden-
tial re-election limit stipulated in the constitution. It is said that he managed
this process by essentially ‘threatening’ and ‘bribing’ people within his party
(Tripp, 2010), and, by now, it appears that he has effectively silenced oppo-
sition voices within the NRM. In the 2011 presidential election, he secured
a landslide victory over Kizza Besigye, the opposition candidate from the
FDC (Forum for Democratic Change), achieving 68 per cent of the vote
compared with his opponent’s 26 per cent. In the previous election, in 2006,
the same two candidates had garnered 59 per cent and 37 per cent of the
votes, respectively. Potentially, the Baganda royalists may be in a position

to pose an effective challenge to the rule of Museveni, but they are frus-
trated by the prohibition of political activities in the monarchy (Makara
et al., 2009). Museveni has been regarded as having many faces. He was
a ‘freedom fighter’, and was sometimes called a ‘cowboy’ because of his
trademark cowboy hat. By contrast, during his lifetime Nyerere was given
the endearing Swahili label ‘Mwalimu’ (teacher) by his fellow countrymen.
Nyerere sought deliberately to build an egalitarian, integrated nation on the
basis of Tanzanian national identity and African dignity. Ethnic parochial-
ism was decried by him at every opportunity, particularly as he mobilised
the country against colonial rule. He continued with the same themes even
after independence (Nyerere, 1966, 1968, 1973).
There is always some interaction between the leadership style and the
operational principles of a political party. Let us first consider the question
of the inclusiveness of the party support base. Looking back on the history
of party formation in Tanzania, TANU was established by Nyerere in 1954,
expanded its influence over the whole territory and then merged with the
ruling party of Zanzibar, the Afro-Shirazi Party, to form CCM in 1977. Mem-
bership of CCM has been open to every citizen. In Uganda, on the other
hand, the origins of NRM are in the liberation war against the dictatorship of
the Northerners – Amin, Obote and Okello – and the organisational principle
of the ‘liberation movement’ influenced the practices of NRM as a political
party in the 1980s and after. President Museveni symbolised the ‘freedom
fighter’ who fights against the ‘enemy’. In contemporary Ugandan politics,
the opposition is now the ‘enemy’.
Secondly, the relationship between the party and the army is of great sig-
nificance. The Ugandan army is increasingly under the personal control of
Museveni (Lindemann, 2010: 35), while the Tanzanian army is now strictly
a government institution after the establishment of a multi-party system in
1992. Prior to this time, the Tanzanian army had been under the heavy influ-
ence of the CCM, as were most institutions in the country. Museveni has
consistently been a military leader and many of his family members belong
to political circles and the army. In Tanzania, in contrast, the government
sought to integrate the institution of the military into the normal political
life of the country in the aftermath of the 1964 mutiny. Retired army officers
are regularly appointed to administrative positions in the government, and
no favour to politically influential soldiers has been visible in Tanzania since
1964 (Lupogo, 2001).
Thirdly, a political party can be a unifying, bridging force in a divided soci-
ety. Zanzibar was formally granted a semi-federal status, and until 1992 the
constitution made the president of Zanzibar automatically the vice-president
of Tanzania. In a normal federal system, several units with equal power
would form a central government above those units. In the case of Tanzania,
the ‘big’ Mainland and ‘small’ Zanzibar are not on even terms, and the CCM
in Zanzibar is clearly a sub-organisation of the national party. As a national

party CCM has successfully utilised both the formal and the informal hier-
archy within the party to continuously mobilise its members in Zanzibar.
This strategy has worked well. As a result, CCM as a party always enjoys sup-
port from both the Mainland and Zanzibar, and CUF, with a relatively small
support base in Pemba, has also been recognised as a national party.4
6.4.3 Decentralisation and federal solutions

The governments of Uganda and Tanzania have made serious efforts to
implement decentralisation policies. In principle, decentralisation has been
considered by both governments to be a useful way to alleviate poverty,
facilitate people’s democratic participation at local levels and absorb a sub-
stantial amount of foreign aid. However, in the case of Uganda, it seems
that politicians have capitalised upon decentralisation policy for political
reasons. Since 2000, the NRM government has dramatically increased the
number of districts to enhance the NRM’s patronage networks. The num-
ber rose from 39 in 1996 to 80 in 2008 (Green, 2008; Mukunya, 2009),
increasing to as many as 111 in February 2011. With a much larger terri-
tory, Tanzania has almost as many units. Mainland Tanzania had 20 regions
and 111 Local Government Authorities (LGAs) in 1998, and 21 regions and
117 LGAs in 2011.
Uganda’s policy of creating many districts is called ‘districtisation’, which
has had a clear intention and effect of dividing kingdoms into smaller
political and administrative units. Although decentralisation was supported
by donors, the NRM government exploited this process with the ‘hidden
agenda’ of weakening the autonomy of political units such as the Buganda
Kingdom (Englebert, 2002; Mutibwa, 2008b). In addition, decentralisation
was used, at least implicitly, to link central ‘big men’ and local small elites
and to maintain patronage networks with the aim of winning elections. An
open question is whether decentralisation has also contributed to corruption
in Uganda as well as in Tanzania, as is often claimed by critics (Mwenda and
Tangri, 2005).5 In 2009, the Ugandan Parliament passed the Regional Gov-
ernment Bill so as to put into effect the regional tier arrangement, which
recognises traditional kingdoms as organisational entities by establishing 48
regional governments, and yet sets limits on the roles of these kingdoms
by prohibiting political activities. Buganda and Busoga opposed the bill on
the ground that the arrangement seemed to deny the kingdoms wider roles,
confining them to the realm of culture, and devoid of substantive powers.
The Tanzanian government, on the other hand, introduced multi-party
politics in 1992 through a constitutional amendment, which prohibited the
formation of political parties based on ethnicity, religion or race. In Uganda,
in contrast, the NRM government provided for the cultural recognition of
ethnic groups, as the 1995 Constitution recognised 56 ‘indigenous’ commu-
nities, and its amendment in 2005 added nine other communities to the list.
Accordingly, in the classification of formal political institutions, the vari-

able H on cultural recognition is rated as 1 in Tanzania and 3 in Uganda
(Chapter 2, Table 2.2). Although this gives the impression that Uganda is a
multicultural country and that Tanzania is integrationist, it is obvious that
open recognition of distinctiveness for too many cultures militates against
the interests of a big group such as the Baganda, whose members consider
it to be indivisible. In short, in Uganda, pluralistic cultural recognition has
been used purposefully to suppress a federalist movement.
In order to contain group-based grievances effectively, it might be instru-
mental for Uganda to depart from the unitary framework and choose a
federal or semi-federal option as in Tanzania (Chapter 2, Table 2.2, vari-
able E). However, complete federalism in Uganda may pose a difficult
question about the extent to which an influential group such as the Buganda
Kingdom can be allowed to enjoy political autonomy. First, in contrast to
the islands of Zanzibar, Buganda is located at the very centre of the coun-
try in terms of geography, economy and government. It is hard to envisage
a situation in which a vast area surrounding the capital city of Uganda is
granted political autonomy and governed by a king, and yet coexists with
the national framework of Uganda (see Figure 6.1). As there can be a hybrid
form of the state along the spectrum from a federal to a unitary state, it might
be practical to devolve more power to key regions, under the framework of
a unitary state, than is assumed in the present regional tier arrangement.
At any rate, having so many districts is not financially sustainable in
today’s Uganda, and therein lies some of the justification for the estab-
lishment of a federal framework. With regard to budget allocation in
Uganda, the ratio of local government expenditure to the national bud-
get has increased dramatically in recent years. Before the decentralisation
programme started, the ratio was 14 per cent in 1994/95, then doubled to
28 per cent in 1997/98, increased to 34 per cent in 2000/01 and reached
37 per cent in 2009/10 (MOFPED, 2004, 2010). In Tanzania, the ratio of
the local government budget to the general budget was 19 per cent in
2003/4 (PEFAR Team, 2006). In both countries, the budget of local govern-
ments has expanded for the purpose of poverty reduction at the local level
and the related financial management and monitoring systems have been
strengthened.
6.5 What interview surveys reveal
The sites of the interview surveys were selected on the basis of their regional
representativeness. In Uganda, major towns in the Northern, Central and
Western regions were chosen. In Tanzania, the locations were Dar es Salaam,
the capital on the Mainland, and Unguja and Pemba, the two Zanzibar
islands. Ethnic concentration in each town is evident in Uganda. In our
survey, the Acholi constitute 89 per cent of the population of Gulu in the
Northern region, the Banyoro are 70 per cent of the population of Hoima in
the Western region, the Banyankole account for 58 per cent of the popula-
tion of Mbarara in the Western/Southern region and the Baganda account
for 44 per cent in Kampala, the Central region. The ethnic groups in Dar es
Salaam are diverse, with the largest ethnic group, the Zaramo, accounting
for only 13 per cent, while the Shirazi form 86 per cent of the residents of
Pemba and 55 per cent of Unguja residents.
Our Ugandan survey consisted of 500 residents. The regional breakdown is
as follows: Kampala, 200 respondents; Hoima, 100; Mbarara, 100; and Gulu,
100. Kampala is the largest city, attracting many immigrants from the sur-
rounding rural areas. Kampala represents the opinion of the Central region,
but the sample size had to be doubled in order to obtain an appropriate
weight of ethnic groups. The interview survey was conducted just before
the presidential election that took place in January and February 2011. In
Tanzania, the number of respondents was 600. The breakdown is as fol-
lows: Dar es Salaam, 300 respondents; Unguja, 148; and Pemba, 152. The
perceptions survey was conducted in March and April 2011. In both coun-
tries, a stratified sampling method was applied to obtain a collective sample
of the same quality from each group.
6.5.1 Uganda
According to our survey, the living standards of the people in the Western
region appeared to have improved and to be approaching those in the
Central region. This trend was consistent with the results of recent house-
hold surveys. In our surveys, in Kampala and Mbarara, the ratio of those
who had completed secondary education or above was high (85 per cent
and 49 per cent, respectively), and many households possessed refrigera-
tors (35 per cent and 36 per cent), although there were many people who
had experienced problems of not having enough water, food and money
in the previous 12 months. The North represents the poorest region, and
Northerners recognised their disadvantaged status (71 per cent replied that
their status was ‘worse’ or ‘much worse’ than other ethnic groups), and
tended to perceive that their situations had been worse five years ago than 10
or 20 years ago. Ninety per cent of respondents in all three places shared the
view that the North had been the least favoured by the government. There
was also a dominant view, even in remote Gulu, that the Banyankole of the
West were the richest group.6 This must reflect the discovery of a promising
oil field in the Western region in 2009, as well as the ethnicity of President
Museveni. Our survey town in the West, Hoima, is close to the oil field in
the Albertine Basin, bordering the DRC.
The recognition of their own poor status by Northern people seems to
have made them feel alienated. The support rate for the opposition FDC
was 28 per cent in Gulu, which was the highest in all places, even though
33 per cent in the same place supported the ruling NRM. Northern people
used to have an aversion solely to Central people, but they now seem to have
the same level of discomfort in working with Western people. The answer
that the ethnic factor was important when they voted for the national par-
liament and local councils was preponderant in Gulu and Kampala where
people thought that politicians of other groups utilised ‘tribalism’ to garner
political support (43 per cent of respondents in Gulu and 40 per cent in
Kampala replied that politicians did this ‘often’ or ‘always’). In addition, the
survey results indicated that the Baganda and Banyankole did not always
like each other. In the question asking if they would accept a daughter’s or
a sister’s marriage with a member of another group, a significant number
of people in Mbarara in the West (23 per cent) opposed marriage with a
Muganda husband.
Northern people perceived their underrepresentation in the government,
police and other institutions and felt that the Banyankole were the most
prosperous. In Gulu, 52 per cent considered Banyankole to be the most
prosperous, while 38 per cent replied that the richest were the Baganda.
In contrast, Western people do not share such perceptions. In Mbarara, in
the West, the willingness to give extra economic assistance to poorer groups
was thinner than in other places: those people who ‘strongly agree’ with this
accounted for only 49 per cent in Mbarara. Concurrence with the opinion
that some portion of government ministers should be allocated to disadvan-
taged groups was also weak in Mbarara. A possible explanation for this is that
the Banyankole, the influential group in the West, have become relatively
rich only recently and may still think that their group should be prioritised
by the government.
The ruling NRM was supported by 69 per cent of the respondents in
Hoima, 57 per cent in Mbarara, 48 per cent in Kampala and 33 per cent
in Gulu. This is consistent with the results of the election conducted in
February 2011, signifying the end of the long-term decline of the NRM.
The view that the president should not have too much power was agreed
to by 86 per cent of people in Gulu, 68 per cent in Kampala, 39 per cent in
Hoima and 36 per cent in Mbarara. Even though the living standards in the
North and Central regions were very different, people in Gulu and Kampala
share a much more cautious view on presidential power than do people in
the West.
The statement that ‘giving semi-autonomous status to Buganda is fair and
justified’ was supported by a majority of the Baganda (54 per cent ‘strongly
agree’ or ‘agree’), while only 23 per cent of Northern people supported this
view. The approval of the statement that ‘the former kingdom should be
given more political power’ showed similar results. There was discontent
regarding regional resource distribution in Gulu and Kampala. Regarding the
question asking if ‘oil revenue will be distributed fairly among the different
regions in future’, the respondents who ‘strongly disagree’ were 74 per cent
in Gulu, 58 per cent in Kampala, 36 per cent in Mbarara and 32 per cent
in Hoima. In the North and Central regions, people tended to perceive that
the current decentralisation policy exacerbated ethnic tension at the local
level (68 per cent ‘strongly agree’ or ‘agree’ in Gulu, 56 per cent in Kampala,
35 per cent in Hoima and 28 per cent in Mbarara).
6.5.2 Tanzania
The living standards were generally higher in Dar es Salaam than in Unguja
and Pemba. As Dar es Salaam was the only place surveyed on the Mainland,
the survey result must contain an urban bias. In Zanzibar, Unguja was gen-
erally richer than Pemba, but it should also be noted that there were more
people who replied that they did not have enough clean water and food over
the past year in Unguja than in Pemba. Forty-eight per cent of the respon-
dents in Dar es Salaam answered that the living conditions of their own
ethnic group were considered to be the same as other groups. On the other
hand, in Unguja (Pemba), 46 per cent (58 per cent) answered ‘much better’
to the same question. This high rate of satisfaction indicates that grievances
related to economic HIs were not very strong in Zanzibar.
With regard to the question asking whether or not ‘you feel uncom-
fortable to work with other groups’, 20 per cent of all respondents said
‘yes’. In Zanzibar, the proportion of respondents who felt uncomfortable
working with Christians was higher than in Dar es Salaam. Regarding the
question about marriage, reluctance was more evident in case of different
religious groups than in cases of different ethnic groups.7 While the num-
bers of Muslims and Christians were almost even in Dar es Salaam, Muslims
were dominant in Zanzibar. As for social trust, people of Pemba tended to
attach the highest value to relatives, neighbours and co-religionists, but
this was less so in Unguja and Dar es Salaam.8 Regarding future prospects,
the respondents showed general optimism that ethnic (racial) relation-
ships would improve rather than worsen in the future, with the people of
Pemba the most positive about such a view.9 In Dar es Salaam, the rela-
tionship between different religious groups seems to have worsened, as the
respondents who answered ‘worsen a lot’ accounted for 26 per cent, while
23 per cent answered ‘stay the same’.
The idea that a specific identity group was favoured in government jobs
was shared by a majority of the people in Pemba, followed by Unguja, and
least in Dar es Salaam. In Pemba, 55 per cent of people replied that eth-
nic (racial) background became ‘much more important’ or ‘somewhat more
important’ after the introduction of the multi-party system in 1992. About
the statement that ‘politicians tend to use ethnicity for mobilising political
support’, 48 per cent of people in Dar es Salaam and 43 per cent in Pemba
answered ‘sometimes’, while 42 per cent in Unguja said ‘never’. The opti-
mism of people in Pemba towards the future of Tanzania turned out to be
stronger than in other places, indicating their satisfaction with the newly
established power-sharing arrangement on the island.
The support rates for the ruling CCM were 56 per cent in Dar es Salaam
and 60 per cent in Unguja, while in Pemba, CCM and the opposition CUF
seemed to be in close competition with each other, and 28 per cent refused to
answer, reflecting the sensitivity of party support in Pemba. Regarding voting
in presidential elections, 62 per cent of people in Unguja and 64 per cent
in Pemba answered that religion was more important than other elements,
while only 28 per cent in Dar es Salaam stated this to be the case. As it
happens, the current president of the country, Jakaya Mrisho Kikwete, is a
Muslim who previously served as chief political instructor in the military.
In contrast, 72 per cent of respondents in Dar es Salaam thought that the
‘policy program and proposed actions are much more important’ at the time
of elections. The ethnic origin of a president was not the main issue in all
three places, where the view that a political party should not be formed in
line with ethnicity was widely shared in Mainland (agreed to by 90 per cent
in Dar es Salaam, 44 per cent in Unguja and 61 per cent in Pemba).
People were generally happy with the formation of the unity govern-
ment in Zanzibar. The respondents who answered either ‘strongly agree’ or
‘agree’ regarding the formation of the unity government were 85 per cent in
Unguja and no less than 94 per cent in Pemba. In Pemba, expectations were
especially high for the positive effects of the unity government on poverty
reduction. The majority of the respondents in the three places expressed the
view that the president had too much power and agreed that the current
decentralisation policy that devolved power to the district and community
levels was useful for the improvement of public service provision (agreed
to by 81 per cent in Dar es Salaam, 85 per cent in Pemba and 53 per cent
in Unguja). Moreover, the statement that ‘more autonomy and indepen-
dence given to Zanzibar is desirable’ was supported by many in both Unguja
(82 per cent) and Pemba (93 per cent).
6.5.3 National and ethnic consciousness in contrast

When comparing the overall survey results in Uganda and in Tanzania, a
number of contrasting points emerge. First, the majority of respondents in
Tanzania, 60 per cent in Dar es Salaam and nearly 90 per cent in Pemba,
acknowledged that the formation of Tanzanian nationhood had been suc-
cessful. Eighty-six per cent of people in Dar es Salaam, 82 per cent in
Unguja and 83 per cent in Pemba agreed that the country’s political sta-
bility was important in strengthening the Tanzanian nationality, while such
perception was weaker and divided in Uganda.
In Tanzania, people tend to attach more weight to nationality and religion,
as subjectively essential elements of identity, rather than to ethnicity.10 This
is in sharp contrast with the Ugandan situation in which ethnic conscious-
ness always vies with national consciousness. As demonstrated in Tables 6.9
and 6.10, the proportion of the respondents who answered ‘I feel only
Tanzanian’ was the highest in Pemba (88 per cent), followed by Unguja
Table 6.9 National and ethnic identification of Ugandans (%)
Group I feel I feel I feel I feel I feel I don’t Total

only more equally more only know /
Ugandan Ugandan Ugandan [ethnic [ethnic Refused
than and identity] identity] to answer
[ethnic [ethnic than
identity] identity] Ugandan
Kampala 10.0 17.5 49.5 13.0 9.5 0.5 100.0

Gulu 5.0 3.0 75.0 12.0 4.0 1.0 100.0
Mbarara 14.0 17.0 57.0 10.0 1.0 1.0 100.0
Hoima 6.0 9.0 74.0 5.0 6.0 0.0 100.0
All respondents 9.0 12.8 61.0 10.6 6.0 0.6 100.0
Table 6.10 National and ethnic identification of Tanzanians (%)
Group I feel only I feel I feel I feel I feel I don’t Total

Tanzanian more equally more only know/
Tanzanian Tanzanian [ethnic [ethnic Refused
than and identity] identity] to
[ethnic [ethnic than answer
identity] identity] Tanzanian
Dar es Salaam 60.0 23.3 5.0 5.7 5.0 1.0 100.0

Unguja 74.3 6.1 12.2 3.4 2.7 1.4 100.0
Pemba 87.5 0.7 11.8 0.0 0.0 0.0 100.0
All respondents 70.5 13.3 8.5 3.7 3.2 0.8 100.0
(74 per cent) and Dar es Salaam (60 per cent). In Uganda, the equilibrium of
perceptions seems to lie in dual identity: ‘I feel equally Ugandan and [own
ethnicity]’ (61 per cent on average).11
There are also contrasting perceptions about decentralisation policy.
Respondents in Uganda tended to think that resource allocation is biased
towards a certain region, while such a perception was not strong in Tanzania.
People in Gulu and Kampala who think they are disadvantaged in terms of
several or all dimensions of HIs find many problems in the present decentral-
isation policy and do not regard it as a useful means of resource allocation.
In contrast, in Tanzania, people are generally satisfied with the decentrali-
sation policy in their own country. People in Zanzibar share the view that
social services in Pemba is inferior to that of Unguja, and such a perception
of socioeconomic HIs should form part of the expectations held by local
islanders in respect of what the government of national unity in Zanzibar
can achieve in the future. In Uganda, there is a widely shared perception
that the Northern region is still the poorest in the country. Tackling the
regional disparity of development will continue to be an overriding task for

the government of Uganda.
6.6 Conclusion
Throughout its existence as an independent country, Uganda has always

been ethnically polarised, and the major pole has historically been the
Buganda Kingdom in the Central region. After independence, Northerners
were used by politicians as the mainstay of the Ugandan army to counter-
act Buganda influence. The insurgent movement, the NRM, led by Museveni
was based on the alliance between the Baganda and the Banyankole in the
West and secured an emphatic victory in 1986. Subsequently, the armed
conflict between the NRM government army and the Northern Acholi rebels
has continued for more than two decades, devastating livelihoods in the
North. In contrast, the society of Mainland Tanzania has been relatively
homogeneous with no conspicuous political unit characterised by a hier-
archical structure. The major exception was Zanzibar, where the fact of the
people belonging to Tanzania as a nation-state was not taken for granted.
In the past decades a cleavage has emerged between the two major islands
of Zanzibar: Unguja, which is politically closer to the Mainland, and Pemba,
which is a stronghold of the opposition.
At the societal level, heterogeneous Uganda and homogeneous (Mainland)
Tanzania appear to be mirror opposites. However, at the political level, these
neighbouring countries also share several common traits, such as the influ-
ence of modern African socialism and communitarianism, donor-sponsored
reforms involving radical decentralisation and poverty reduction policies,
and also the presence of historically powerful groups – Buganda in Uganda
and Zanzibar in Tanzania – seeking political and cultural autonomy. It is such
commonality that justifies this country pair study of Uganda and Tanzania.
Differences in ethnic configurations do not in themselves lead automat-
ically to the presence or absence of violent conflict. Rather, it is peoples’
informal conventions and practices, influenced by such configurations,
which play a key role in determining the future paths of nations’ politi-
cal development. It must be noted that the norms and practices of political
leaders often substantiate the functions of institutions and give actual direc-
tions to policies and that leadership styles and the nature of party politics are
critically important. The modern history of Uganda has been characterised
by dynamic, volatile ethnic politics. President Museveni tried to impose dis-
cipline on the whole nation by punishing separatist groups. On the other
hand, Tanzania seems to have succeeded in developing a political culture
characterised with sharing power. President Nyerere was symbolically called
a teacher, and a teacher is a person who is expected to treat all pupils with
equal attention.
Table 6.11 Multidimensional HIs and power-dispersing institutions
Political Economic Power- Federalism

and cultural HIs sharing
status HIs
Tanzania Mainland Low Low Informal No

Between Potentially Some Formal and Yes
Mainland high evidence informal
and Zanzibar
Uganda High High No No
The contrast of HI situations and institutional features in these two coun-

tries are summarised in Table 6.11. No country is predestined to be peaceful
or violent. Tanzania could have easily had more violence in its modern his-
tory. For instance, the bloodshed in the Zanzibar Revolution might have
been a predictor of prolonged political turbulence, and in the 2000s, peo-
ple were horrified to see Tanzanian riot police shooting into the crowds and
killing dozens of people in Zanzibar after the heavily contested elections. The
political stakeholders have ridden out the storm through processes of power-
sharing, both formal and informal. Pemba’s optimism about the future of
Tanzania as a nation as shown in our perceptions survey demonstrates that
power-sharing combined with semi-federal autonomy has been genuinely
appreciated by opposition forces.
Integrationist policies seem to have worked out well in Mainland
Tanzania, where ethnicity is extremely fragmented. Faced with a difficult
problem in Zanzibar, however, Tanzanians decided to apply a categorically
different, accommodative principle to the islands, seeking a semi-federal
and power-sharing solution. The question is whether this kind of solution
can also be applied to Uganda. As discussed, in contrast with the islands of
Zanzibar, Buganda is located at the very centre of the country, and it seems
that this geographical position makes it rather problematic for Buganda to
obtain full political and administrative autonomy from the rest of the coun-
try. Furthermore, the democratic nature of a monarchy may be debatable in
contemporary Africa. The South African case is exemplary. In that country,
the role of traditional leaders, including the Zulu King, has been confined
to the realm of cultural symbolism and consultation. For the moment, the
Ugandan government is trying to divide the Baganda people by recognising
a multitude of cultural rights and promoting smaller political and admin-
istrative units, but this policy direction has not necessarily contributed to
solidifying a PD polity in Uganda. It also appears to be costly in terms of
finance.
The future of Uganda is unpredictable, and all the more so in light of
recent events such as the independence of the new Republic of South Sudan
in the north and the discovery of oil in the Western region. Some have
argued that the arena of conflict in Uganda was shifting from the national to
the local level due to the progress of decentralisation (Green, 2008), but this
proposition is only speculative as long as the central government holds the
reins on the process. As power concentrates, counteracting poles are likely
to emerge. The failure of President Museveni to promote a reliable successor
also clouds the vision of the country’s political future. The Ugandan people
are to deal with these uncertainties to reach their own solution.
Notes
1. Muslims make up 35 per cent of the population of Mainland Tanzania and more
than 99 per cent in Zanzibar, and 12 per cent of the total population in Uganda.
(The 2002 Uganda Population and Housing Census. UBOS, 2005.)
2. World Bank Country Results Profile, http://data.worldbank.org/country/uganda
(last accessed 1 May 2013).
3. ‘Bukenya lauds Buganda crackdown, CBS closure’, The Observer (Kampala), 13
September 2009, http://www.observer.ug/index.php?option=com_content&task=
view&id=5119&Itemid=59 (last accessed 1 May 2013).
4. In Tanzania as a whole, in 2010, Chadema (Party for Democracy and Progress)
was the largest opposition party, and the CUF was the second opposition party.
5. BBC News, ‘Tanzania PM to Resign over Graft’, 7 February 2008, http://news.bbc.
co.uk/2/hi/africa/7232141.stm (last accessed 1 May 2013).
6. Twenty-nine per cent pointed to the Banyankole, while 28 per cent indicated the
Baganda. In Mbarara, 31 per cent of the respondents replied that the Bahima,
a subgroup of the Banyankole, were the richest and 22 per cent said that the
Baganda were the richest.
7. Fifty-two per cent of respondents in Unguja, 55 per cent in Pemba and 21 per cent
in Dar es Salaam revealed their reluctance to work with Catholic people.
8. Ninety-six per cent of respondents in Pemba and 91 per cent in Unguja said that
they trust their relatives, while 76 per cent in Dar es Salaam said they did so.
9. Sixty-three per cent of respondents in Pemba, 45 per cent in Unguja and
42 per cent in Dar es Salaam said that the ethnic relationship would improve
in the future.
10. Eighty-eight per cent of respondents in Dar es Salaam answered that national-
ity was very important, while 76 per cent in Unguja and 64 per cent in Pemba
answered thus. On the other hand, 86 per cent in Dar es Salaam answered that
religion was very important, whereas this ratio was 85 per cent in Unguja and
95 per cent in Pemba.
11. People in Uganda, including Baganda, often seem to strategically select a more
advantageous identity in a given situation.
7
The Politics of Identity, Horizontal
Inequalities and Conflict in Kenya
Mwangi S. Kimenyi
7.1 Introduction
Prior to the post-election violence in 2008, Kenya was commonly considered

a model of peace and stability in sub-Saharan Africa, but there were clear
signs that conflict was likely. In fact, since the reintroduction of competitive
politics in 1992, the country has experienced isolated ethnic clashes that
peaked around general election cycles, notably in 1992 and 1997. In a study
entitled Sporadic Ethnic Violence: Why Has Kenya Not Experienced a Full-Blown
Civil War? Kimenyi and Ndung’u (2005) provided a detailed analysis of a
number of likely triggers of conflict in the country and suggested that the
likelihood of conflict was high. The authors considered factors such as the
politicisation of ethnicity, grievances associated with land ownership, espe-
cially in the Rift Valley, and poverty as triggers that could lead to conflict
among the various groups in the country.
While the factors considered by Kimenyi and Ndung’u are indeed rel-
evant, it is becoming increasingly evident that systematic inequalities –
more specifically, horizontal inequalities (HIs) – which refer to inequalities
across identifiable social groups, are a more important trigger for conflict.
In Kenya, the most common axis of self-identification is a person’s ethnic
group. HIs could be a serious source of grievance that can trigger conflict as
some groups who are marginalised in terms of access to economic opportu-
nities, resources such as land, and public services, and who are also excluded
from political participation, use violence to express their grievances. As will
be apparent, there are systematic inequalities across regions in Kenya.
Because ethnic groups are to a large extent associated with particular regions,
regional inequality necessarily implies ethnic inequality. These regional and
ethnic inequalities represent a most serious manifestation of HIs that has
the potential for triggering violent conflict. The likelihood of such conflict is
magnified by the fact that political mobilisation has increasingly been along
ethnic and regional lines.
153
154 Kenya
In this chapter, I focus on socioeconomic and political HIs in Kenya and

evaluate the role of such inequalities in triggering violence. In Section 7.2,
I start by looking at the configuration of Kenyan society. Here I focus on
its ethnic composition and also on the regional distribution of the various
groups. In Section 7.3, I provide information on the prevailing HIs in Kenya
focusing on both political representation and various dimensions of well-
being. Section 7.4 focuses on the drivers of political processes and political
institutions. The subsequent section analyses the 2008 post-election violence
in Kenya and examines some of the root causes of the conflict. Section 7.6
reports the results of surveys on popular perceptions, which were conducted
in three towns in Kenya: Nairobi, Mombasa and Nakuru. Section 7.7 con-
cludes with a review of the new constitution and prospects for preventing
future conflicts in Kenya.
7.2 Configuration of Kenyan society
The most important dimension of Kenyan society that is particularly rel-

evant to HI is ethnic identification. Other important dimensions of the
configuration of the society include the rural and urban populations, reli-
gion and the age distribution of the society. All of these dimensions could
have some bearing on HIs; but it is inequalities amongst ethnic groups that
seem to be most important in explaining conflict.
Kenya has a diverse mix of ethnic groups; indeed, no one group is particu-
larly dominant. The largest is the Kikuyu, which accounts for 22 per cent of
the population, followed by the Luhya (14 per cent), the Luo (13 per cent)
and the Kalenjin (12 per cent). Table 7.1 shows the population of Kenyan
ethnic groups based on 1969 and 2010 data. Most of the groups not
shown have relatively small populations. Although there has been marked
Table 7.1 Ethnic composition of Kenya (% of

total population)
Ethnic group 1969 Present estimation
Kikuyu 20.1 22
Luhya 13.3 14
Luo 13.9 13
Kalenjin 10.9 12
Kamba n.a. 11
Kisii n.a. 6
Meru n.a. 6
Other African n.a. 15
Non-African n.a. 1
Source: Kanyinga (2007) and CIA (2011).

Mwangi S. Kimenyi 155
urbanisation since independence, the largest part of the population remains

in the rural areas, and each of the ethnic groups is concentrated in specific
regions of the country. The Kikuyu live in Central Province and some parts
of the Rift Valley, the Luhya are in Western Province, the Luo in Nyanza
Province and the Kalenjin in the Rift Valley. Other smaller ethnic groups
also occupy specific areas of the country. In essence, ethnicity is tied to geo-
graphical region, which is referred to as geo-ethnicity (see Figure 7.1 and
Table 7.2).
SOUTH SUDAN ETHIOPIA
UGANDA
SOMALIA
5
4
2
Lake
Lake 3 1
Victoria
Victoria
6
7
TANZANIA
1 Kikuyu majority : Central

2 Luhya majority : Western
3 Luo majority : Nyanza
4 Kalenjin majority : Rift Valley
5 Somali majority : Eastern (North) and North Eastern
6 Meru, Embu and Kamba : Eastern (South)
7 Mijikenda, Tata-Taveta and Swahili-speaking : Coast
Figure 7.1 Map of major ethnic groups by province

Note: Geographic boundaries are approximate.
Source: Author.
156 Kenya
Table 7.2 Geo-ethnic composition
Main ethnic Percentage Percentage of Percentage of

group of total population of major Rift Valley
population location (region) population
Kikuyu 22 91.8 (Central) 21.2

Luhya 14 83.8 (Western) 11.2
Luo 13 53.4 (Nyanza) 2.3
Kalenjin 12 41.4 (Rift Valley) 41.4
Source: Kanyinga (2007) and CIA (2011).
Ethnicity forms the most important cleavage in Kenya. Of the various

political parties that have emerged since the introduction of multi-party
politics in 1991, all have an ethnic base and each is dominated by mem-
bers of one or a few ethnic groups. Likewise, voting patterns have by and
large followed ethnic lines. Thus ethnicity is not only a social cleavage but
also the main axis of political mobilisation.
As noted, most of the Kenyan population lives in rural areas and there
are marked differences in access to public services between rural and urban
residents. The majority of the Kenyan population is Christian although there
is a significant population of Muslims primarily in Coast Province and North
Eastern Province.
7.3 Horizontal inequalities
For Kenyan society, there seem to be two important dimensions through

which HIs are manifested. The first is political inequality, which is evidenced
by differences in the relative political power and representation in govern-
ment of different ethnic groups. The other important manifestation of HIs is
evidenced by the wide differences in levels of well-being and access to public
services across regions. To the extent that ethnic groups are concentrated in
some particular regions, regional inequality means ethnic inequality. This
section provides evidence of various dimensions of HIs in Kenya.
To some extent, today’s political and socioeconomic HIs in Kenya have
their roots in the country’s colonial history. The British created Kenya’s
political and administrative units along ethnic boundaries with the Kikuyu
primarily in Central Province, the Luhya in Western Province, the Luo
in Nyanza and a mix of the Kalenjin and Masai in the Rift Valley. Fur-
thermore, during colonisation, the British invested primarily in areas that
were inhabited by settlers, which resulted in disproportionate infrastructure
development. Central Province and the Rift Valley, where a large propor-
tion of settlers were based, benefitted from higher levels of infrastructure
investment than the rest of the country. These inequalities persisted after
independence.
(a) Political horizontal inequality

Probably one of the most obvious pieces of evidence for political HI is
demonstrated by the distribution of senior government positions. Since
independence, successive leaders have tended to favour members of their
own ethnic groups with regard to appointments to senior positions. In his
first term of office, Kenya’s first president, Jomo Kenyatta, seemingly sought
to create an ethnically diverse government. He also tried to invest in increas-
ing access to health and education across the country. However, even in
those early days, Kenyatta’s own group, the Kikuyu, were already overrep-
resented in government compared to their share of the total population.
Kikuyu accounted for 29 per cent of all cabinet positions under Kenyatta at a
time when they accounted for only 20 per cent of the population (Table 7.3).
Upon taking office in 1978, President Daniel Arap Moi proceeded to
increase the number of Kalenjin, his own tribe, in government, from
12 per cent in 1979 to 18 per cent by the end of his term in 2001. The
number of Kikuyu in the cabinet fell from 31 per cent when Moi entered
office to 4 per cent in 2001. The Luhya representation in cabinet increased
from 12 per cent to 14 per cent. Like the Kikuyu, Luo representation in cab-
inet also decreased, from 12 per cent in 1979 to 7 per cent in 2001. For the
Kikuyu and Luo, the cost of not having one of their ‘own’ in office quickly
became clear (see also Table 7.4).
The disproportionate ethnic representation in the cabinet is also apparent
when one considers other senior positions such as the Provincial Commis-
sioners, Permanent Secretaries, heads of parastatals and so on. The story
that emerges when one looks at the ethnic representation in government
Table 7.3 Cabinet representation by ethnic group
Percentage in Kenyatta Moi Kibaki

cabinet (Kikuyu) (Kalenjin) (Kikuyu)
1966 1978 1979 2001 2003 2005
Kikuyu 28.6 28.6 30.8 3.6 16.0 18.2

Luhya 9.5 4.8 11.5 14.3 16.0 21.2
Luo 14.3 14.3 11.5 7.1 16.0 3.0
Kalenjin 4.8 4.8 11.5 17.9 8.0 6.1
Total number of 21 21 26 28 25 33
cabinet members
Source: Stewart (2008b).

158 Kenya
Table 7.4 Political transition in Kenya
Kenyatta’s last Moi’s first Moi’s second

cabinet (1978) cabinet (1982) cabinet (1988)
Number of cabinet members 22 28 34

Members of Kenyatta’s – 35% 14%
cabinet retained
Members of Kenyatta’s tribe – 14% 3%
retained
Source: Kimenyi and Shughart (1989).
is that the president’s ethnic group benefits disproportionately in terms of

high-level appointments.
(b) Socioeconomic HIs by region

Although there are several factors that could contribute to socioeconomic
inequalities, it is true that political inequalities do play a major role. This
is because political influence concentrated amongst some ethnic groups is
often used to direct resources to their specific communities. This was quite
evident during President Moi’s rule when most development projects were
in the Rift Valley and there was neglect of projects in Central Province and
other parts of the country that were considered to be opposition zones.
Thus, access to various public goods such as education, health, water and
other physical infrastructure projects tended to follow patterns of access to
political power.
However, it is important to note that many existing inequalities are deter-
mined by various other factors, such as geography. In particular, the quality
of agricultural land and rainfall impact on the productivity of land and con-
sequently on incomes. Thus, to a large extent, geography has a close bearing
on the inequalities that we observe, and they are not necessarily the result
of the political allocation of resources. As such, we note that differences in
well-being across regions are in part due to geographical factors. In essence,
it is not necessary that inequalities be attributable to the political alloca-
tion of resources for conflicts between groups to arise. Nevertheless, the fact
that systematic differences exist across regions and ethnic groups does create
conditions that could trigger conflict. As long as the systematic inequalities
exist, the probability of conflict is real.
Figure 7.2 shows the state of wealth distribution. The data reveal that there
are marked regional imbalances, with the majority of the country’s poor
being concentrated in North Eastern Province, the Rift Valley and Coast
Province. Over 70 per cent of the population in North Eastern Province
falls within the lowest quintile. On the other hand, Kenya’s wealth is
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Nairobi Coast Central Rift Valley Nyanza North Eastern Western
Eastern
Lowest Highest
Figure 7.2 Distribution of population in the highest/lowest wealth quintile by

province (% within province)
Source: KNBS and ICF Macro (2010: 26).
concentrated in urban areas, and especially in Nairobi. Over 90 per cent of

the population in Nairobi falls within the highest quintile.
An important manifestation of HI is in terms of educational attainment.
Differences in levels of average educational attainment can have major
implications by perpetuating inequalities over time. The data reveal that
there is significant geographical inequality in access to education in Kenya.1
As can be seen in Figure 7.3, in the 2008/09 national health survey, nearly
70 per cent of the population in North Eastern Province had no formal
education, compared with around 20 per cent in Central Province and
Nairobi.
An important indicator of the quality of life in a given country is the infant
mortality rate. Although Kenya has made great strides in reducing infant
mortality, which has fallen from 77 deaths per 1,000 in the 2003 national
health survey to 52 deaths per 1,000 in the 2008/09 national health survey,
there are significant differences across regions. Four provinces, namely North
Eastern, Coast, Western and Nyanza, still have high infant mortality rates
that are far above the national average (Figure 7.4). These outcomes reflect
differences in resources and also in access to medical facilities.
Probably the clearest evidence of regional inequalities comes from look-
ing at differences in various public investments in infrastructure. By and
large, investment in infrastructure in Kenya has been concentrated in urban
160 Kenya
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
rn
bi
st
za
rn
rn
tra
lle
ro
oa
e
e
te
n
st
en
st
Va
ya
ai
es
C
Ea
Ea
N
W
ift
th
R
or
N
University or college Secondary or post-primary/vocational

Primary or incomplete primary No formal education
Figure 7.3 Educational attainment by province

Source: Kenya Demographic and Health Surveys, 2008–2009, http://www.measuredhs.com/data/
dataset/Kenya_Standard-DHS_2008.cfm?flag=0 (last accessed 1 May 2013).
160
Deaths per 1,000 live births
140
120
100
80
60
40
20
0
Central Eastern Rift Nairobi North Coast Western Nyanza
Valley Eastern
Figure 7.4 Under-five mortality by province

Source: KNBS and ICF Macro (2010: 107).
areas, and especially in Nairobi and Mombasa and the adjoining areas. This
has resulted in wide differences in access to water, electricity and sanita-
tion. Figure 7.5 shows the regional inequalities in terms of access to different
types of infrastructure. Nyanza, North Eastern and Western provinces have
the lowest access to running water and sanitation. Similarly, Nyanza, Eastern
and Western provinces have the lowest access to electricity.
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Nairobi Coast Central Rift Nyanza North Eastern Western
Valley Eastern
Water supply Sanitation Electricity
Figure 7.5 Access to piped water/public tap, sanitation and electricity by province
Note: Access to sanitation means access to flush toilet or flush pit latrine.
As would be expected, ownership of assets that reflect quality of life is

influenced by incomes and also by access to infrastructure and, especially,
electricity. Thus, ownership of assets such as a refrigerator or a television is
also concentrated in Nairobi and Coast and Central Province. The remaining
provinces have noticeably lower ownership of these assets by comparison.
Western Province reported the lowest ownership of refrigerators and North
Eastern Province reported the lowest ownership of televisions (Figure 7.6).
This information is indicative of the trend in well-being and reflects the
general character of regional HIs.
80%
70%
60%
50%
40%
30%
20%
10%
0%
Nairobi Coast Central Rift Nyanza North Eastern Western
Valley Eastern
Refrigerator TV
Figure 7.6 Ownership of refrigerator and TV by province

162 Kenya
In summary, the data provided above show that Kenya is a country that
is characterised by significant regional inequalities in terms of quality of
life and access to public goods and services. We have observed that these
regional inequalities do in fact reflect ethnic inequalities due to the fact that
ethnic groups occupy distinct areas of the country. Although the political
inequalities may not necessarily explain socioeconomic inequalities, they
do exacerbate the inequalities. Nonetheless, the origin of the inequalities is
not important in terms of the potential for the inequalities to trigger con-
flict. Of importance is how inequalities interact with political processes and
institutions.
7.4 Politics: Processes, institutions and policies
During the first few years of independence, political mobilisation in Kenya

was primarily on an ideological basis. Thus the contests between the Kenya
African National Union (KANU) and the Kenya African Democratic Union
(KADU), and later between KANU and the Kenya People’s Union (KPU), were
driven primarily by differences in ideology. At the time, the political parties
adopted distinct ideological positions, with KANU taking a more Western
capitalist stance, KADU a federalist stance and KPU leaning towards Soviet-
style socialism. However, this era of ideology-driven politics was short-lived,
and since the late 1960s Kenyan politics has been influenced primarily
by ethnicity as the primary axis of political mobilisation, accompanied by
a high concentration of power in the executive. By and large, the post-
independence constitution has concentrated power in the executive branch,
with a weak set of checks and balances on the executive. The combination
of a strong presidency and the ethnicisation of politics in Kenya resulted
in a ‘winner-takes-all’ system in which the loser was left with little role in
policymaking (Kniss, 2010: 10). Although there have been some reforms
that have served to disperse power over time, concentration of power has
been the norm, marked by what has been referred to as an ‘imperial’ presi-
dency. These two aspects of Kenyan politics have impacted the relationship
between ethnic groups, as evidenced by popular perceptions, and also may
have impacted HIs. Jeni Klugman explains that the distribution of public
goods such as education facilities, health, water and physical infrastructure
tends to follow patterns of access to political power (Klugman, 2000: 300).
Although at independence the structures of governance were centralised,
there were several features of the political system that served to disperse
power. In particular, the independence constitution provided for a multi-
party system, a dual executive, a bicameral legislature and a quasi-federal
system. In addition, the constitution provided for effective checks on the
executive and the devolved and dispersed powers were relatively difficult
to amend. In addition, the constitution provided for the protection of
all civil and political rights enshrined in a Bill of Rights. However, these
structures that could have served to disperse power among several points
were systematically eliminated, and power was eventually concentrated in
the presidency. Not only was the bicameral legislature done away with, but
the limited powers of local governments were eventually transferred to the
central government. Furthermore, and probably most important, the coun-
try slowly gravitated to a de facto single-party state, then de jure single-party
rule, before returning to multi-party system in the early 1990s. It has only
been since the enactment of the new constitution in 2010 that meaning-
ful constitutional dispersion of political power has now come to define the
political landscape. Of course, the implementation of the new constitution
is still in the very early stages. Below I discuss the various phases that seem
to define Kenyan political institutions and processes, and the relationship
with ethnicity and HIs.
De facto and de jure single-party politics

As noted, during the first few years of independence, political parties com-
peted for leadership and adopted divergent policy stances. Two national par-
ties were established in 1960 just prior to independence: the KANU and the
KADU. These two parties had different political stances and also were dom-
inated by different groups: KANU by the more urbanised population and
KADU by more rural, pastoral populations. While KANU advocated a highly
centralised political system, KADU feared that such a centralised system
would lead to domination by the people of Central Province and Nairobi.
As such, KADU advocated a decentralised system with a quasi-federal struc-
ture commonly referred to as majimbo (regions). The competition between
these two parties, however, ended after the 1963 elections when KANU won
the majority of seats in both houses and KADU dissolved itself and joined
KANU in 1964. After the 1963 elections, Jomo Kenyatta became the prime
minister, but the Queen remained the head of the state. In 1964, Kenya
became a republic with Kenyatta as its president. This marked the begin-
ning of not only the dominance of KANU in Kenyan politics but also the
increasing concentration of power in the executive.
By 1966, growing discontent with Kenyatta’s leadership culminated in the
resignation from the party of 30 KANU members led by Oginga Odinga
and the establishment of a new opposition party, the Kenya People’s Union
(KPU), a socialist-leaning party. Odinga and his colleagues cited the increas-
ing dominance of members of Kenyatta’s tribe – the Kikuyu – in the running
of the government and the exclusion of other Kenyans as the primary source
of their grievances. After by-elections, only 9 of the 30 were re-elected back
into parliament and thus KANU did not face any significant opposition.
Between this time and the death of Kenyatta, the powers of the executive
were expanded and opposition both within and outside the party was pun-
ished severely, including the jailing of political opponents. In fact, KPU was
164 Kenya
banned in 1969 and its leader, Odinga, was jailed. With the banning of KPU,
Kenya became in essence a de facto single-party state. Over time, several other
important power-dispersing institutions were all but extinguished. For exam-
ple, a series of constitutional amendments allowed the president to appoint
high court judges, a law to detain opponents without trial was enacted,
the two legislative houses were merged to form the National Assembly and
independent candidates were barred from contesting seats in the National
Assembly. Finally, in 1982, following an attempted coup, Section 2A of the
constitution was introduced, effectively turning the country into a de jure
one-party state. Essentially, by 1982, the country was mainly in a civilian
dictatorship.
In brief, as noted in Kibara (2003: 1), after independence, the ruling party
gradually undermined the checks on executive power to create a one-party
state. The ruling party harassed and co-opted political opposition, removed
effective powers of the judiciary and parliament and tampered with the
constitution to consolidate presidential power. From 1964 to 1990, thirty
amendments were made to the constitution, depriving the original values
and principles of the constitutions. As a result, highly personalised rules
emerged from the wreckage of the dreams of Kenya’s founding fathers.
In addition, the period was also marked by a persistent neutralisation of
non-state actors. The civil society was weakened through intimidation, the
use of force and detention without trial, denial of permits and outright
de-registration of non-governmental organisations. Thus the period was not
only one marked by the negation of constitutionalism, but was also marked
by widespread and far-reaching abuses of human rights.
What is particularly significant to note about the era of single-party
rule was that the president was extremely powerful and was able to solely
determine who held positions of authority, which groups were rewarded
and which were punished. It was through these wide discretionary pow-
ers that both Presidents Kenyatta and Moi were able to discriminate in
favour of members of their communities for government appointments,
contracts, land and the concentration of public services. As already observed,
the Kikuyu benefited disproportionately during Kenyatta’s rule while the
Kalenjin were favoured by Moi. The allocation of land to Kikuyus in Kenya’s
Rift Valley remains one of the key issues that have divided the Kikuyus and
Kalenjins, and is the primary source of grievances that have triggered recent
conflicts between these two groups. The patronage system perfected during
the long rule of President Moi, who discriminated against the Kikuyu and
other ethnic groups and awarded preferences to the Kalenjin, had the effect
of widening the divisions in Kenyan society.
Multi-party politics and the rise of ethnic politics

During the 1980s, opposition to the increasingly oppressive KANU regime
gained steam as church groups, civil society, political activists and also
the international community worked together to demand change. In 1991

Section 2A of the constitution was repealed, thus effectively ending de jure
one-party rule in Kenya. With the repeal of section 2A, citizens could now
establish political parties to contest elections in competition with KANU.
Within a short period following the repeal of Section 2A numerous political
parties were registered and permitted to contest the 1992 general elections.
However, opposition unity proved to be illusive, and even parties that were
multiethnic at the start split along ethnic lines; thus, multi-party politics
was also marked by intensified ethnic mobilisation. Although the introduc-
tion of political party competition would have otherwise acted to disperse
power, on the contrary, the result was that the country was even more
divided, allowing then President Moi to remain in office for another ten
years. Even worse, political mobilisation along ethnic lines was also associ-
ated with ethnic clashes as politicians sought to weaken the voting strength
of their opponents. Although there were several reforms that sought to
weaken the power of the executive, power remained concentrated with the
president.
However, before the 1997 general elections, far-reaching reforms were
instituted by the Inter-Parties Parliamentary Group (IPPG). These reforms
sought to create a more level playing field in the contest for elections.
The reforms focused on trimming those advantages that the incumbent
party was using to remain in power. These included the removal of the
requirement that political parties obtain licenses to address public meet-
ings, the requirement that the state media – the Kenya Broadcasting
Corporation – give equal treatment to all parties, the barring of public
officers from engaging in partisan politics and so on. Although the rul-
ing party KANU won the presidency, the political terrain had changed
considerably.
In 2002, a grand coalition of parties came together to form the National
Rainbow Coalition (NARC) in opposition to the incumbent party KANU.
Under the leadership of Mwai Kibaki, NARC won the elections that year and
formed what was then an inclusive government with broad representation
of most of the major ethnic groups, except Kalenjins, who remained firmly
in KANU. In coming together to form the grand coalition, leaders of the
constituent parties, the National Alliance Party of Kenya (NAK) and the Lib-
eral Democratic Party (LDP), had agreed to share positions in government,
as was supposedly detailed in a memorandum of understanding (MOU).
In addition, it had been agreed that a new constitution would be enacted
that would create a new power structure that divided executive authority
between a prime minister and a president. However, the coalition collapsed
in 2005 after disagreements over the constitutional referendum. The presi-
dent dropped his opponents in the coalition, who were led by Raila Odinga
and Kalonzo Musyoka, both of whom opposed President Kibaki during the
2007 elections. Voting in the election was primarily along ethnic lines, and
it ended in dispute and violence.
166 Kenya
7.5 The 2008 post-election violence
Prior to the 2008 post-election violence, Kenya was broadly considered to

be one of the best examples of peace and stability on the African continent.
Indeed, the speed and scale of the violence that occurred in the wake of the
2007 election caught many in the international community by surprise.
There had been past instances of violence in Kenya; in 1991 a group of
arsonists calling themselves the ‘Kalenjin warriors’ terrorised Luo, Luhyia,
Kikuyu and Kisii communities in the Rift Valley. They targeted farms occu-
pied by these ethnic groups and committed various acts of violence against
them until many of the occupants fled their homes. Similarly, in 1997, in
Coastal Province, local Mijikenda tribes targeted Kikuyus and Luos, which
led to the displacement of about 10,000 people and the death of 65 people,
including 13 police officers (Kimenyi and Ndungu, 2005). Despite these past
conflicts, Kenya was largely considered to be a peaceful and stable country,
although it is important to note that other instances of violence in 1992,
1997 and 2002 seemed to occur before or after elections.
The December 2007 elections were no exception. The president, Mwai
Kibaki, led the Party of National Unity (PNU), a political party composed
primarily of the Kikuyus. The opposition Orange Democratic Movement
(ODM) and its splinter group (ODM-K) backed Raila Odinga: a Luo. The
ODM consisted primarily of an alliance between the Kalenjin, Luhyas and
Luos. Both candidates launched mass media and public rally campaigns
inciting ethnic cleavages to gain loyalty and support. The state-owned Kenya
Broadcasting Corporation gave the ruling party 76 per cent of radio cover-
age compared with only 13 per cent and 6 per cent for ODM and ODM-K,
respectively. Similarly KBC gave 71 per cent of its television coverage to PNU
and only 11 per cent and 5 per cent to ODM and ODM-K. As PNU was
dominated primarily by Kikuyus, this gave the impression that the govern-
ment favoured that group over others (Kniss, 2010). The private media in
the country gave better coverage to the opposition, although PNU still gar-
nered a larger share of that coverage as well. This provided fodder for ODM
charges of electoral fraud by PNU.
Despite the tension in the run-up to the election, there was general calm
and order on election day. An estimated 9.9 million Kenyans voted on
December 27: a voter turnout of about 69 per cent. Three days later, despite
delays in reporting constituency returns, the Electoral Commission of Kenya
(ECK) announced that Kibaki had won the election, though by a small mar-
gin. ODM cried foul, citing gains in the parliamentary vote and irregularities
during the election. Odinga refused to concede the vote, and the two leaders
found themselves in a deadlock.
In the two months following the election, violence erupted through-
out the country, and by the time an agreement was reached, about 1,000
Kenyans had been killed and hundreds of thousands were displaced.
Allegations of corruption by both parties sparked the violence. Much of

the violence occurred in the Rift Valley, a historical hotbed for inter-ethnic
conflict, as protestors in ODM-dominated areas launched attacks directed
primarily at the Kikuyus, who also launched revenge attacks, especially in
Central Province. In the capital city, gangs utilised their networks to launch
attacks against other ethnic groups (Kniss, 2010). Local politicians and some
business people exploited youth frustration and unemployment to launch
attacks against their opponents. Police largely failed to contain the violence
and in some instances were rumoured to have taken part in attacks (Alston,
2009).
By February 2008, there was strong international pressure to end the vio-
lence. Under the mediation of former UN Secretary General Kofi Annan and
other prominent personalities a deal was brokered that saw the formation
of a government of national unity with Kibaki as president and Odinga as
prime minister. Even though no other major changes were implemented, the
power-sharing arrangement dispersed a significant amount of power from
the president.
Roots of the crisis

The report of the Kenya Commission of Inquiry into Post-Election Violence,
known as the ‘Waki report’, points to three main factors that contributed to
the violence. The first is the politicisation of violence in Kenya: politicians
have deliberately used violence to gain power since the country insti-
tuted multi-party democracy in 1991. They have at times exploited ethnic
differences in order to win elections.
The second factor the Waki report points to is the imperial presidency
noted above: the ‘winner-takes-all’ system which gave significant power to
the presidency put pressure on groups to have one of their ‘own’ in power
lest they be excluded from access to resources and patronage. Lastly, the
report points to:
a feeling among certain ethnic groups of historical marginalization, aris-

ing from perceived inequities concerning the allocation of land and other
national resources as well as access to public goods and services. This feel-
ing has been tapped by politicians to articulate grievances about historical
injustices which resonate with certain sections of the public. This has cre-
ated an underlying climate of tension and hate, and the potential for
violence, waiting to be ignited and to explode.
(CIPEV, 2008)
In isolation, each of these factors was insufficient to cause the violence

that occurred in Kenya in 2008; however, in combination, they created
an explosive environment. The previous section showed that there are
168 Kenya
socioeconomic HIs that exist in Kenya. Resources do not seem to have

been allocated equitably, and this helped to fuel the conflict. In agreement
with the Waki report, this chapter asserts that ‘perceived inequities’ in com-
bination with other factors ignited the 2008 post-election violence that
occurred in Kenya. Accordingly, the next section examines the perceptions
of ethnicity and inequality in Kenya. How do Kenyans view themselves rel-
ative to their neighbours? Understanding these perceptions is important for
developing strategies to prevent conflict in the future.
7.6 Perceptions of ethnicity and inequality in Kenya
The post-election violence in 2008 revealed how vulnerable Kenya is to eth-

nic conflict. In this chapter, I have discussed the remarkable HIs that define
Kenyan society. I have also highlighted the political institutions and polit-
ical processes that could impact conflict between groups. In particular, it
is suggestive that the high concentration of power in the executive and
the patronage that has defined Kenyan politics could have created condi-
tions that divide the Kenyan people along the lines of social identity groups.
Although the existing inequalities may be the result of a number of fac-
tors, including political exclusion and discriminatory allocation of resources,
the fact remains that these inequalities are reflected in how different ethnic
groups perceive each other. These perceptions are important in that they
signal differences and attitudes that may trigger conflict.
This section examines perceptions of ethnicity and inequality among
Kenyans. Indeed, perhaps more so than actual inequality, it is perceptions of
inequality that can be the drivers of conflict. Using data from a perceptions
survey of 907 respondents in three towns, namely Nairobi, Mombasa and
Nakuru, conducted by the Japan International Corporation Agency (JICA),
this chapter will now examine perceptions of inequality among adult
Kenyans.
Figure 7.7 shows how Kenyans identify themselves with regard to ethnic
versus national identity. The majority of Kenyans, about 57 per cent, identify
themselves more as Kenyans than as their ethnic group. Interestingly, only
9 per cent of Kenyans identified themselves more with their ethnic group or
only in terms of their ethnic group. The rest, about 32 per cent, felt equally
Kenyan and their ethnic group.
Respondents were then asked whether or not they would object to their
sister/daughter marrying someone from another ethnic group. Once again,
the majority (62 per cent) of respondents would have no objections to such
marriage with another ethnic group. About 35 per cent of respondents said
that they would object only to a specific ethnic group and 2 per cent would
object to all ethnic groups but their own. The respondents that said they
would only object to marriage with specific groups were then asked which
group they would object to. As shown in Figure 7.8, the Luo (111 out of
I feel only my own

ethnic group, 14 Don't know/
(2%) refused to answer,
I feel more of my 14 (2%)
ethnic group than
Kenyan, 66 (7%)
I feel only Kenyan,

239 (26%)
I feel equally
Kenyan and my
own ethnic group,
289 (32%)
I feel more Kenyan

than my own ethnic
group, 285 (31%)
Figure 7.7 Ethnic versus national identity (N = 907)

Source: JICA Surveys.
120 111
100
100
80
67 64
60 53
39
40 31 29
25 23
19 16 15
20 13 12 12 10
0
M a
R ria
ba
so
bu
a
Em a
Ba e
Ka ru
o
yu
in
Tu kot
sa
ni
Lu ii
Ka ali
an
hy
l
Lu
di
nj
Te
ju
Ku
m
ra
ku
Ki
aa
m
Po
M
en
rk
le
Bo
So
Ki
Figure 7.8 Would object to marriage with a specific ethnic group (N = 321)
321 respondents: 35 per cent), Kikuyu (100 out of 321: 31 per cent), Somali
(67 out of 321: 21 per cent) and Kalenjin (64 out of 321: 20 per cent)
were the most cited groups that respondents would object to having their
sister/daughter marry.
Respondents were not asked why they objected to these specific groups.
Was the objection based on a past unpleasant encounter or just a general dis-
like of the other group? In conducting future research it might be of interest
to find out why respondents objected to these groups.
170 Kenya
When asked whether they would object to their sister/daughter marrying

someone based on that person’s religion, 58 per cent of respondents said
that they would have no objections, 4 per cent said they would object to
all religions but their own and 38 per cent of respondents said that they
would object to specific religions. When asked which specific religion they
would object to, about 56 per cent of the respondents said they would object
to their sister/daughter marrying someone who was Muslim (17 per cent
against believers of traditional religions, 12 per cent against Catholics and
11 per cent against Protestants). This is not necessarily surprising, as about
80 per cent of Kenyans identify themselves as being Christian (KNBS, 2010).
To further examine inter-group trust, respondents were asked whether
they would feel uncomfortable working with people from other ethnic
groups. The overwhelming majority of respondents, 71 per cent, said they
would not be uncomfortable working with people from other ethnic groups.
However, about 28 per cent of respondents would be uncomfortable working
with people from a specific ethnic group and less than 1 per cent of respon-
dents would be uncomfortable working with any ethnic group other than
their own. Of those respondents who reported that they would feel uncom-
fortable working with members of a specific ethnic group, the Kikuyu (72
out of 253 respondents: 28 per cent) were the most commonly cited group
that respondents felt uncomfortable working with, followed by Luo (66 out
of 253: 26 per cent), Somali (46 out of 253: 18 per cent) and Kalenjin (36 out
of 253: 14 per cent) (Figure 7.9).
Respondents were asked how much they trusted different groups (see
Figure 7.10). To this, they could respond ‘not at all’, ‘just a little’, ‘I trust them
somewhat’ or ‘I trust them a lot’. The most trusted groups – ‘somewhat’ or
‘a lot’ – were relatives (66 per cent), neighbours (50 per cent) and people of
the same religion (49 per cent). The least trusted groups – ‘not at all’ or ‘just
80 72
70 66
60
50 46
40 36
30 22 19 18
20 14 14
11 7
10 6 6 5 5 3
4
0
yu
na
Bo o
ba
Tu ria
o
le
u
bu
in
R a
M ot
Ba i
Ka ali
ni
Ka ii
sa
s
hy
Lu
er
an
s
di
nj
ku
ra
Te
ju
m
Ku
m
Em
Ki
aa
Po
M
en
Lu
le
rk
Ki
So
Figure 7.9 Would feel uncomfortable working with a specific ethnic group (N = 253)
Relatives
Neighbours
People of own ethnic group
People from other ethnic groups
People of same religion
People of other religions
People from other provinces
People from other African countries

0%
%
%
%
%
%
%
0%
10
20
30
40
50
60
70
80
90
10
I trust them a lot I trust them somewhat Just a little
Not at all Don't know/refused to answer
Figure 7.10 Trust of different groups (N = 907)

a little’ – were people from other ethnic groups (63 per cent), people from
other religions (60 per cent) and people from other countries (60 per cent).
The high level of distrust towards other ethnic groups is somewhat surprising
given the respondents’ responses to previous questions about marriage and
working together. These results suggest that other factors, such as religion
and proximity (that is, being neighbours), are also important factors in how
Kenyans perceive other groups.
The majority of respondents, 63 per cent, believe that inter-ethnic rela-
tions in Kenya are either positive or very positive (Figure 7.11). Only a small
number of respondents, 4 per cent, believe that inter-ethnic relations are
negative or very negative. However, it appears that Kenyans have somewhat
polarised opinions about future inter-ethnic relations (Figure 7.12); about
14 per cent of respondents expect things to worsen somewhat or a lot, while
13 per cent expect a lot of improvement.
The study then sought to examine respondents’ perceptions of whether
ethnicity affects access to public goods and services (Figure 7.13). The over-
whelming majority of respondents believed that ethnicity affects a person’s
chances of getting a government job (80 per cent) or a government contract
(79 per cent). Ethnicity plays a much larger role in access to government jobs
than access to private sector jobs (66 per cent), though it plays a large role
in getting access to a job in both sectors. Historically, land ownership has
172 Kenya
0% 20% 40% 60% 80% 100%
Very positive Positive Neither positive nor negative

Negative Very negative Don't know/refused to answer
Figure 7.11 Perceptions of people from other ethnic groups (N = 907)

0% 20% 40% 60% 80% 100%
Improve a lot Improve somewhat Stay the same

Worsen somewhat Worsen a lot Don't know/refused to answer
Figure 7.12 Perceptions of future inter-ethnic relations (N = 907)

been a source of tension in Kenya. It is therefore a concern that 48 per cent

of respondents felt that ethnicity affects a person’s chances of owning land.
Regarding the ethnicisation of Kenyan politics, respondents were asked
whether they thought that ethnicity has become more important in Kenyan
politics since the introduction of multi-party elections in 1992. About 71
per cent of respondents thought that ethnicity has become somewhat or
much more important in Kenyan politics (Figure 7.14).
Respondents were asked whether today’s socioeconomic conditions of
their ethnic group is better, the same or worse off than under the differ-
ent administrations (Figure 7.15). About 38 per cent of respondents thought
that their ethnic group was better off at the time of interview than it had
been under Kibaki before 2007, compared with 28 per cent of respondents
who felt that they were better off than they had been under Moi before 2002
or KANU before 1992. Overall, it appears that respondents felt that they had
been worse off under Kibaki (between 2002 and 2007) than under Moi or
KANU (before 1992).
Public sector jobs
Public sector contracts
Private sector formal jobs
Public housing
Land ownership
University education
Pre-university education
Private loans
0% 20% 40% 60% 80% 100%
Figure 7.13 Ethnicity affects a person’s chances of getting access to . . . (N = 907)

Don't know/
Much less
refused to
important, 27
answer, 53
(3%)
(6%)
Somewhat less
important, 54
(6%) Much more
important, 342
No change, (38%)
130 (14%)
Somewhat
more
important, 301
(33%)
Figure 7.14 Has ethnicity become more important in Kenyan politics since the
introduction of multi-party elections in 1992? (N = 907)
Approximately 40 per cent of respondents felt that Kikuyus are most

favoured by the government. This was followed by the Luo (21 per cent)
(Figure 7.16). The perception that one group is significantly favoured by the
government is a cause for concern, as it may present grounds for other ethnic
174 Kenya
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
Today compared Today Compared Today Compared
to Kibaki to Moi from to KANU
before 2007 1992 to 2002 before 1992
Better and much better Same Worse and much worse

Don't know/refused to answer
Figure 7.15 Perceptions about today’s socioeconomic conditions of own ethnic group
compared to the past (N = 907)
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
a
Ka o
in
a
ba
R na
le
so
hy
u
Bo ia
yu
So u
t
M ii
Tu ali
ni
Em i
ko
sa
Lu
an
er
b
nj
di
r
ju
m
Te
ra
Ki
Lu
ku
Ku
m
M
Po
aa
le
en
rk
Ba
Ka
Ki
Figure 7.16 Which ethnic group do you feel is most favoured by the government?
(N = 907)
groups to target the ‘favoured’ group during times of conflict. It is therefore

important that the Kenyan government implements policies that facilitate
more equitable resource allocation.
When asked whether their ethnic group received its fair share of govern-
ment revenue, projects, or access to universities, the majority of respondents
belonging to one of the smaller ethnic groups felt that their group received
Table 7.5 Compared to other ethnic groups, do you think your ethnic group
currently gets its fair share? (% of group)
Group Revenue from Government Access to public

the central projects university
government
No Yes No Yes No Yes
Kikuyu 18.3 81.7 17.7 82.3 11.9 88.1

Luo 36.5 63.5 35.3 64.7 16.7 83.3
Luhya 46.2 53.8 49.5 50.5 38.1 61.9
Kamba 68.8 31.3 62.0 38.0 61.1 38.9
Meru 40.9 59.1 50.0 50.0 30.4 69.6
Kisii 54.5 45.5 52.2 47.8 34.0 66.0
Kalenjin 25.8 74.2 28.6 71.4 16.4 83.6
Maasai/Samburu 87.5 12.5 87.5 12.5 60.0 40.0
Bajuni 100.0 0.0 100.0 0.0 100.0 0.0
Somali 88.9 11.1 86.3 13.7 80.6 19.4
All other 78.9 21.1 82.1 17.9 78.9 21.1
less than their fair share. Indeed, feeling excluded from access to government
resources with no opportunity to redress this issue could fuel violent action
against the government and against those seen to be garnering the lion’s
share of benefits (Table 7.5).
7.7 The new constitution and prospects

for preventing future conflicts
Behind these perceptions of inequality, there is actual evidence of unequal

resource allocation in Kenya (see Section 7.3). Changing perceptions is per-
haps a daunting task; nonetheless, it is important that Kenya implement
policies to address some of the underlying causes that led to the 2008 post-
election violence. This suggests that the country must implement measures
to allocate resources more equitability as a necessary starting point for Kenya
and enforce strategies to address the political environment that precipitated
previous conflicts. Kenya’s new constitution attempts to address some of
these underlying factors.
The devolved constitution: 2010

On August 4, 2010, Kenyans voted for a new constitution: one that devolves
power to local administrative units in an attempt to address some of the
political challenges that prompted previous conflicts. By devolving power to
localised units, the new constitution also seeks to facilitate more equitable
176 Kenya
resource allocation, because resource allocation decisions will not rest solely
on the presidency.
Although the new constitution retains a presidential system, it has sev-
eral features that disperse political power to many centres and positions and
also restrains the power of the executive. In this new constitution, the pres-
ident is still powerful, but there are various constraints on the powers of
the executive as compared to the previous system. For example, although
the president will still be able to appoint the cabinet and senior govern-
ment secretaries, the appointees must be vetted by various organs of the
legislature and other independent commissions. Furthermore, there are sev-
eral requirements regarding the composition of the cabinet and other senior
appointments that guarantee broad representation of the Kenyan people.
Thus, while the new constitution retains a presidential system, it is no longer
the same as the ‘imperial’ presidency that has dominated Kenyan politics.
Another feature of the new constitution that will serve to disperse power
is the establishment of devolved structures of governance. The constitution
calls for the establishment of 47 devolved governments called ‘counties’.
These counties will be the focal points for the delivery of services to the
people. Each country will have its own government headed by a governor
and a legislature. Although the counties will receive resource transfers from
the central government, counties will have the power to determine their
own policies and development programmes, including raising some forms of
revenue. The devolved governments will serve to increase the participation
of the people in determining their own destinies and will ensure broader
inclusion in decision-making than has been the case under the centralised
system.
A key power-dispersing feature of the new constitution brings back a
bicameral legislature with a parliament and a senate. Members of parliament
will be elected as they were in the past, based on electoral constituencies
much like they are now. The members of parliament will be responsi-
ble for making broad policies at the central government level. Members
of the senate will be elected to represent county interests in the central
government.
Looking ahead
As noted, for most of the post-independence era, Kenyan politics has been
dominated by an ‘imperial’ presidency. The new constitution radically alters
Kenyan politics by placing checks on the executive and dispersing power to
many centres. Although a presidential system was retained, there are various
checks on executive power. However, it is still not clear how the estab-
lishment of county governments will impact ethnic relations and also HIs.
Finally, the commencement of cases by the International Criminal Court
(ICC) against six suspects related to the post-election violence in Kenya is
also creating uncertainty about the future and will most likely redefine the
politics and ethnic alliances in a major way.
The 2008 post-election violence in Kenya revealed how easily the coun-
try could disintegrate into chaos. Although violence was triggered by
perceptions that the election had been stolen by President Kibaki, it is
apparent that there are underlying grievances that have divided the Kenyan
people. In this chapter, I have focused on HIs and have suggested that
these inequalities could be important triggers for conflict. I have also sug-
gested that, to a certain extent, power-concentrating institutions have in
part contributed to the existing HIs. The new constitution radically changed
the institutions of governance to provide for wide dispersion of power and
effective checks on the executive.
However, the information reported concerning perceptions of ethnicity
and inequality reveals that negative perceptions about other ethnic groups
are widespread. To the extent that such perceptions do influence the way
people act, the probability of inter-ethnic conflict in Kenya remains high.
Just because a new constitution has been adopted may not resolve the
differences unless serious steps are taken to address these differences.
Note
1. See also Alwy and Schech (2004).
8
Managing Horizontal Inequalities
and Violent Conflicts in Nigeria
Ukoha Ukiwo
8.1 Introduction
Nigeria is a country of very startling paradoxes. It is one of Africa’s major

states, holding the continent’s largest population (150 million) and its
largest oil reserves (Bach, 2004). It is one of the claimants for Africa’s
presumed slot in the proposed enlargement in the number of permanent
seats on the United Nations Security Council. One of the justifications for
this claim is the indefatigable ‘big brother’ role Nigeria has played in sta-
bilising the continent. It has contributed more troops to both UN and
regional peacekeeping missions and aid to needy countries than any other
country on the continent and has the fourth-largest peacekeeping mis-
sion in the world. In fact, Nigeria initiated and bankrolled the world’s first
regional peacekeeping and peace enforcement initiative (Adebajo, 2008).
Still Nigerians are desperately poor, as 60 per cent of Nigerians live below
the minimum subsistence level of US$1 per day.
The country not only manifests symptoms of state fragility, but is also
periodically haunted by the prospects of failure and collapse. There have
been heated discussions across the country on the possibility of the ‘Arab
spring’ spreading to Nigeria, with sombre suggestions about whether the
crises will follow the Egyptian or the Libyan course. In 2007, the US National
Intelligence Committee published a report that listed Nigeria as one of the
countries likely to disintegrate in the medium term (USAID, 2010). Although
many Nigerians disagree on the veracity of the parameters deployed for this
prediction, there is no doubt that the country is still haunted by its past civil
war. The task of keeping Nigeria together has led to a flourishing of political
institutions to address the underlying causes of conflicts.
This chapter seeks to examine the attempts of post-civil war Nigeria
to grapple with fundamental challenges to nationhood. The chapter is
divided into seven sections. After this introduction, Section 8.2 will exam-
ine the ethnic and regional configuration of the country in order to put
the competing ethno-political social forces in context. In Section 8.3, I shall
examine horizontal inequalities (HIs) in Nigeria and their impact on group
178
Ukoha Ukiwo 179
mobilisations and conflicts. Sections 8.4 and 8.5 will examine the political
institutions developed to manage conflicts and how they have functioned
over time. In Section 8.6, I will present the findings of a popular perceptions
survey conducted to gauge public views on the nature of identities, HIs,
and political institutions. Section 8.7 presents the concluding remarks of
the study.
8.2 Configurations of identity
Historically, ethnicity, region, and religion have been the major social cleav-
ages in Nigeria as well as the fault lines for most political mobilisations and
violent conflicts in the country. The probability of violent conflict tends to
increase where one social cleavage reinforces another. For instance, when
ethnic identity is coterminous with religious identity, as we have in some
parts of the North Central zone (the Middle Belt), mobilisations along emer-
gent ethno-religious identities have tended to result in frequent violent
conflicts.
Nigeria has about 350 ethnic groups (Otite, 1990). Since the colonial
period, however, some of these ethnic groups have played dominant roles in
politics. These groups, known as the ‘major ethnic groups’, historically dom-
inated the three regions formally established as centres of government in
1946. In fact, the placement of the peoples with similar cultures and related
languages in colonially drawn administrative boundaries promoted the rise
of ethnic consciousness, cohesion, and mobilisation (Nnoli, 1978). Thus,
Hausa and Fulani ethnic groups dominated the Northern Region, while
the Yoruba and Igbo were dominant in the Western Region and Eastern
Region, respectively. The other ethnic groups in the regions constituted
the minorities and were known as Northern minorities, Western minori-
ties, and Eastern minorities. The larger minority groups in the North are
the Kanuri, Angae, Berom Gwari, Kataf, Nupe, Tiv, Idoma, Ebira, Chamba
and Igala. These groups are found mostly in the North Central and North
East geopolitical zones. The Western minorities include the Edo, Urhobo,
Ijaw, Itsekiri, Ishan, Isoko, Ika Igbo, and Etsako among others, while the
Eastern minorities include the Ijaw, Efik, Ibibio, Ikwerre, Anang, Ogoni, and
Ejagham among others. Both the Eastern and Western minorities currently
belong to the South South geopolitical zone. A key feature of Nigeria’s eth-
nic configuration is the fluidity of both identities and boundaries. Against
this background, groups can adopt entirely new identities or fuse with other
groups. For example, because Igbo identity was considered a political lia-
bility after the collapse of the Igbo-led Biafran secessionist attempt, some
groups on the Igbo frontiers have privileged their sub-group identities over
Igbo identity (Ndu, 2011).
It is almost impossible to estimate the current population size of the eth-
nic groups. This is because of the disputes that threatened the fabric of the
180 Nigeria
nation over the controversial census exercises of the 1950s and 1960s. Since
the end of the civil war, successive governments have omitted ethnic and
religious identities from the censuses. Although ethnic leaders have tried
to use the state populations of successor states and provinces as proxies
of the populations of the different ethnic groups, the figures they present
are evidently exaggerated as the non-ethnic resident populations are hardly
factored into such computations.
Ethnic mobilisations in the late colonial period and early years of indepen-
dence were marked by the quest for minorities in each of the three regions
to pull out of the regions through the creation of new states. Often the three
major ethnic groups that tussled for supremacy at the federal tier of gov-
ernment were drawn into the conflicts due to the alliances they forged with
minorities in neighbouring regions in order to undermine their rival parties.
Because ethnic identity almost coincided with party affiliation, such political
alliances often crystallised as ethnic alliances (Coleman, 1958; Sklar, 1963;
Vickers 2000).
Successive state-creation exercises have, however, aggravated the politics
of marginalisation and ethnic domination. They have led to the emergence
of new major ethnic groups and new ethnic minorities in the new states.
This has resulted in endless agitations for the creation of new states. As a
result of the dynamic changes in the structure of the federation as shown
in Table 8.1, the configuration of ethnic cleavages is fluid. Throughout the
colonial and early postcolonial periods, the structure was tripolar, with the
Hausa, Igbo, and Yoruba competing for supremacy. However, this configura-
tion has evolved into a multipolar structure with the ascendancy of minority
ethnic groups following the successive creation of states, the recognition of
six geopolitical zones, and the discovery of oil and gas deposits among the
Southern minorities. The geopolitical zones have become particularly salient
since 1999 when they became the template for distribution of principal fed-
eral posts such as the president, vice president, Senate president, speaker
of the House of Representatives, and secretary to the federal government
(Figure 8.1).
Furthermore, there are situations where the most salient cleavage is
between the North and South, leading to a bipolar structure. In such
Table 8.1 Evolution of Nigeria’s state structure
1960 1963 1967 1976 1987 1991 1996
3 regions 4 regions 12 states 19 states 21 states 30 states 36 states

+ 1 FCT + 1 FCT
Note: FCT: Federal Capital Territory.

Source: Author.
Ukoha Ukiwo 181
NIGER
CHAD
Sokoto
Katsina
Jigawa
Zamfara
Kano Yobe Borno
Kebbi
BENIN Kaduna Bauchi Gombe

Niger
Kwara Adamawa
Plateau
FCT.Abuja
Nassarawa
Oyo
Taraba
Osun Ekiti Kogi
Benue
Ogun Ondo
Enugu CAMEROON
Lagos Edo
Anambra Ebonyi
Cross
Delta Imo Abia River
ATLANTIC Akwa
OCEAN Bayelsa Rivers Ibom
Zone
South South
South East
North Central
South West
North East
North West
Figure 8.1 Political map of Nigeria, showing states and geopolitical zones
circumstance, the minorities in one region align with the major ethnic
group(s) in the region against the major ethnic group(s) and minorities in
the other region(s). The situation is also dynamic in the 36 states of the
federation, depending on their ethnic composition. For instance, sub-ethnic
identities become more salient and are mobilised for political contestations
in states where there is only one indigenous ethnic group. Such situa-
tions often result in multipolarity. Multipolarity also characterises the ethnic
structure of states with a number of fairly equally sized ethnic groups.
However, in cases, where there is a dominant ethnic group, a bipolar struc-
ture often emerges, as the minority ethnic groups coalesce into a bloc to
neutralise the dominant group.
Religious cleavages are also salient in Nigeria, with Islam and Christianity
competing for influence and dominance. The historical dominance of the
religions in different regions has generated conflicts as dominant groups
react to protect their turf in response to the quest of rival religions to
make inroads. Although other religions exist in Nigeria, the configuration
182 Nigeria
is basically bipolar, as the salient cleavage is that between Christianity and

Islam, which respectively have at least 45 per cent of Nigerians as adherents.
8.3 Horizontal inequalities between identity groups
At the root of the political contestations described above is the existence or

perceived existence of HIs among ethnic, regional, and religious groups in
the country. The fault lines of social cleavages are usually the boundaries
between those who have and those who do not have, and between those
who feel included and those who feel excluded. This section presents data on
current actual HIs. These include political HIs, socioeconomic HIs, and cul-
tural status HIs. A further discussion of HIs will be presented in Section 8.6,
which focuses on perceived HIs.
8.3.1 Political horizontal inequalities

A major concern of Nigerians over the years has been how to prevent partic-
ular identity-based groups from dominating political positions. This concern
derives from the role of the state as the principal allocator of resources. Thus,
political HIs are considered to be very critical, as they are veritable drivers of
other dimensions of HIs. The aspects often included in discourses of political
HIs are ethno-regional representations in the executive council, the legisla-
ture, and the military. However, the scope of offices has been extended over
the years to include bureaucratic organisations such as the judiciary and the
civil service as well as the media: indeed, all institutions considered to be crit-
ical in the deployment of state power and authority, and in the distribution
of public resources.
Table 8.2 shows the heads of the three arms of government. It makes it
clear that the North West zone has dominated the headship of the execu-
tive branch, both among military heads of state and elected presidents and
prime ministers, producing five heads of government since independence.
It is followed closely by North Central, which has produced three military
heads of state. Thus, while the Northern zones have produced nine heads
of the executive, the Southern zones have produced four. This history is
responsible for the widespread grievance in Southern Nigeria about ‘North-
ern hegemony’. The perception of ‘Northern hegemony’ is aggravated by
the duration of Northern leadership of the executive, which comprises 38
out of 50 years of nationhood. It is against this background that there was
little support among Southerners for the emergence of a president of North-
ern origins after the death of President Umaru Yar’Adua in 2010. The North
has also featured prominently in the headship of the legislative and judi-
cial arms of government as well as the headship of the armed forces and the
police during this period. The Southern zones have fared better in producing
heads of the legislative and judicial arms of government.
Table 8.2 Heads of three arms of government by geopolitical zone, 1960–2010
Zone Military Head Prime Senate Speaker of Chief Total Rank Percentage Rank in
of State Minister/ President House of Justice of number in of total population
President Representatives the of posts posts population
Federation 2006a
North West 3 2 1 3 9 1 25.6 1

North East 1 1 2 6 13.6 4
North Central 3 2 1 6 3 13.5 5
South West 1 1 1 5 8 2 19.7 2
South East 1 2 1 4 4 11.7 6
South South 1 1 1 4 4 15.0 3
Note: a. As this column excludes FCT, the figures do not add up to 100.
Source: Author’s compilation.
183
184 Nigeria
Table 8.3 Composition of federal cabinets by zone, 1984–2010 (%)
Zone Percentage Buhari Babangida Abacha Obasanjo Yar’Adua Jonathan

of 1983–1985 1990 1993 2004 2008 2010
population
North West 25.6 30.0 33.3 22.7 21.2 19.5 20.0

North East 13.6 10.0 16.7 13.6 15.1 17.1 12.5
North 14.5 20.0 11.1 18.2 18.2 19.5 17.5
Central
(+FCT)
South West 19.7 20.0 16.7 18.2 15.1 19.5 17.5
South East 11.7 10.0 16.7 9.1 12.1 9.8 12.5
South South 15.0 10.0 5.5 18.2 18.2 14.6 20.0
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Source: Mustapha (2006); Author’s compilation.
Northern dominance is also reflected in the composition of the Federal

Executive Council since independence. As shown in Table 8.3, the Northern
zones – especially the North West, which is the zone with the largest concen-
tration of the Hausa and Fulani – have dominated successive federal cabinets
since independence, accounting for an average of 56 per cent of cabinet
members. As Mustapha (2006) has argued, the North has not only domi-
nated in actual numbers of ministerial positions but has also been favoured
in the allocation of so-called ‘juicy’ ministerial posts. These include portfo-
lios in ministries such as Defence, Petroleum Resources, Works and Housing,
Transport, Foreign Affairs, Agriculture, Internal Affairs, and Finance, among
others, which have lucrative opportunities for dispensing patronage.
The Northern zones also take the lead in the composition of both cham-
bers of the National Assembly. While the North controls 53 per cent of
Senate seats, the South holds 47 per cent. The disparity is similar in the
Federal House of Representatives, where constituencies are delimited on the
basis of population. The North produces roughly 53 per cent of members of
the lower house. Again, the dominance of the North West zone is striking,
as it controls a quarter of the seats in the House (Tables 8.4 and 8.5).
The Northern advantage is, however, mitigated by the strong showing of
the Southern zones in the composition of the federal civil service, which
has historically been skewed in favour of the South. In 2010, the 18 states
deemed to have exceeded their quota of between 2.5 per cent and 3 per cent
in the federal civil service were mostly Southern states (Mustapha, 2007).
As Table 8.6 shows, the Southern zones dominate the federal civil service,
accounting for 61 per cent of its total staff. The South West zone is the most
advantaged zone, with a quarter of all staff of the federal civil service. How-
ever, the advantage enjoyed by the Southern zones is counterbalanced by
the application of quotas for promotions and appointments to senior civil
Ukoha Ukiwo 185
Table 8.4 Distribution of Senate seats, 2008
Zone Percentage Number Percentage Rank in Rank in

of of of population representation
population Senators Senators
North West 25.6 21 19.3 1 1

North East 13.6 18 16.5 4 2
North Central 13.5 18 16.5 5 2
South West 19.7 18 16.5 2 2
South East 11.7 15 13.8 6 6
South South 15.0 18 16.5 3 2
FCT 1.0 1 0.9 7 7
Total 100.0 109 100.0
Source: Adapted from PARP (2008: 20).
Table 8.5 Distribution of seats in the House of Representatives, 2008
Zone Percentage of Number of Percentage of Rank in Rank in

population representatives representatives population representation
North West 25.6 92 25.6 1 1

North East 13.6 48 13.3 4 5
North Central 13.5 49 13.6 5 4
South West 19.7 71 19.7 3 2
South East 11.7 43 11.9 6 6
South South 15.0 55 15.3 2 3
FCT 1.0 2 0.6 7 7
Total 100.0 360 100.0
service positions. The application of such quotas has helped to bridge the
gap between the North and the South. This explains why the North West
and North East zones have a disproportionately high percentage of directors
relative to their share of staff in the service (Table 8.6).
It should be noted that consolidation of staff by geopolitical zones and the
entire public service occludes certain salient disparities. There is a tendency
for particular units of government to be dominated by persons from particu-
lar zones and states. This suggests that, though very salient, the North–South
dichotomy hides more than it reveals. For instance, while the South South
and South West zones are generally well represented in the federal service,
this cumulative representation masks the overrepresentation of states such
as Ogun in the South West and underrepresentation of states such as Bayelsa
in the South South.
Moreover, some groups tend to dominate different government depart-
ments. For instance, the staff profile of the National Assembly (Table 8.7)
186 Nigeria
Table 8.6 Composition of the federal civil service, 2000
Zone Directorate All staff Population Rank of Rank of

% % % Directorate all staff
North West 16.6 10.4 25.6 2 5

North East 12.4 8.6 13.6 6 6
North Central 16.4 18.4 13.5 3 3
South West 24.4 24.9 19.7 1 1
South East 13.4 16.0 11.7 5 4
South South 15.8 20.0 15.0 4 2
Source: Adapted from Mustapha (2006).
Table 8.7 National Assembly staff, December 2007
Zone Percentage of Number of Percentage Rank of staff

population staff of staff
North West 25.6 486 14.5 4

North East 13.6 407 12.2 6
North Central 13.5 801 24.0 1
South West 19.7 409 12.2 5
South East 11.7 583 17.5 3
South South 15.0 603 18.0 2
FCT 1.0 52 1.6 7
Total 100.0 3,341 100.0
shows the dominance of personnel from the North Central, South East and
South South zones. The fact that two of the three zones (North Central and
South East) have produced more Senate presidents and speakers of the House
of Representatives, especially in the early years of the institution, lends some
credence to suspicions that heads of governmental institutions influence
recruitment of persons from their ethnic groups.
These disparities within zones and regions have moderated the potential
of HIs triggering violent conflict, as they stifle the emergence of an undif-
ferentiated community of the deprived and, ipso facto, the possibility of
mobilisations of one group against the other. A clear case in point was the
undisguised lukewarm attitude of key leaders of the North Central zone to
mobilisations for the election of another president from the North after the
death of Alhaji Umar Musa Yar’Adua in March 2010. Another dimension
of HIs in Nigeria that has helped to stymie violent conflict between the
North and South is the fact that political HIs and socioeconomic HIs are not
mutually reinforcing. As shall be seen in the following discussion of socioe-
conomic HIs in Nigeria, there is an emergent division of powers where each
Ukoha Ukiwo 187
Table 8.8 School enrolment by zone, 2008
Zone Percentage of Primary Junior secondary Senior secondary

population
No. % No. % No. %
North West 25.6 5,575,580 26.6 475,879 13.3 464,392 16.0

North East 13.6 3,237,533 15.5 616,640 17.2 331,168 11.4
North Central 13.5 3,273,120 15.6 534,308 14.9 479,181 16.5
South West 19.7 3,726,869 17.8 908,398 25.3 795,499 27.4
South East 11.7 2,403,777 11.5 559,180 15.6 386,724 13.3
South South 15.0 2,498,978 11.9 427,060 11.9 415,655 14.3
FCT 1.0 218,697 1.0 66,314 1.8 32,535 1.1
Total 100.0 20,934,554 100.0 3,587,779 100.0 2,905,154 100.0
Source: Adapted from NBS Nigeria (2009: 59, 72–3).
zone or region tends to dominate some spheres. It is also important to note

that the application of the ‘Federal Character’ principle, which requires rep-
resentation of all states in the federal cabinet, has helped to stabilise the
polity as it ensures representation of most of the groups. Both this provi-
sion and the delimitation of legislative constituencies on the dual basis of
equality and population have contributed to managing conflicts by effec-
tively satisfying the cravings of the majority for pre-eminence as well as the
desires of the minority elements for inclusion and representation.
8.3.2 Socioeconomic inequalities

There are great regional disparities in access to education, health and other
social services in favour of the South. This is evident in the statistics pre-
sented on school enrolments. As Table 8.8 shows, the Southern zones
outpace the Northern zones in educational attainments as the level of educa-
tion progresses. Thus, while the Northern zones have roughly 58 per cent of
primary school enrolments, this drops to 44 per cent at the senior secondary
school level. Correspondingly, the percentage of schools in the North drops
from 64 per cent of primary schools to 46 per cent of secondary schools,
while the share of schools in the Southern zones increases from 35 per cent
for primary schools to 54 per cent for secondary schools (NBS Nigeria, 2009:
57, 64). It is hardly surprising, therefore, that the percentage ratio of literate
population between North and South is 39 to 60 (Table 8.9). Furthermore, as
shown in Table 8.10, the North only accounted for about 22 per cent of those
admitted into universities. Consequently, in most public institutions, there
are more Southerners than Northerners with higher degrees (Table 8.11).
The result of the educational advantage enjoyed by the Southern zones
is that the South tends to have more qualified candidates vying for posi-
tions. This naturally generates resentment among Southerners, who allege
188
Table 8.9 Literacy status of ages 6 and above by zone, 2006
Zone Percentage of Number Percentage Rank of

population of literates of literates literates
North West 25.6 15,237,837 20.5 2

North East 13.6 6,324,707 8.5 6
North Central 13.5 7,217,318 9.7 5
South West 19.7 19,188,114 25.8 1
South East 11.7 11,488,258 15.4 4
South South 15.0 14,173,592 19.0 3
FCT 1.0 864,086 1.2 7
Total 100.0 74,493,912 100.0
Source: Adapted from NBS Nigeria (2009: 8).
Table 8.10 Nigerian university applications and admissions, 2007–2008 session
Total applications Total admissions
Zone No. % No. %
North West 58,476 5.66 2,846 6.00

North East 38,439 3.72 2,606 5.49
North Central 140,625 13.60 4,776 10.06
South West 251,719 24.34 11,958 25.19
South East 274,013 26.50 14,106 29.71
South South 269,239 26.04 11,087 23.35
FCT 1,535 0.15 93 0.20
Total 1,034,046 100.00 47,472 100.00
Source: Adapted from NBS Nigeria (2009: 154–5).
Table 8.11 Educational backgrounds of members of state legislatures, 1999–2007
Zone With higher education With only diplomas or

secondary certificates
1999–2003 2003–2007 1999–2003 2003–2007
North West 35 71 99 131

North East 53 62 97 91
North Central 67 83 47 48
South West 91 96 41 48
South East 97 102 29 29
South South 105 96 19 25
Total 448 510 332 372
Source: PARP (2007: 65).

Ukoha Ukiwo 189
Table 8.12 Incidence of poverty by zone, 1980–2004
Category 1980 1985 1992 1996 2004
North West Total poor 37.7 52.1 36.5 77.2 71.2

Core poor 8.3 14.2 9.0 37.3 26.8
North East Total poor 35.6 54.9 54.0 70.1 71.2
Core poor 11.8 16.4 18.5 34.4 27.9
North Central Total poor 32.2 50.8 46.0 64.7 67.0
Core poor 5.7 16.4 14.8 28.0 29.8
South West Total poor 13.4 38.6 43.1 60.9 43.0
Core poor 2.1 9.0 15.7 27.5 18.9
South East Total poor 12.9 30.4 41.0 53.5 26.7
Core poor 2.4 9.0 15.7 18.2 7.8
South South Total poor 13.2 45.7 40.8 58.2 35.1
Core poor 3.3 9.3 13.0 23.4 17.0
Population in poverty (millions) 17.7 34.7 39.2 67.1 68.7
Source: UNDP (2009b: 64).
discrimination when they lose out in public sector employment and pro-
motions due to the application of quotas in consonance with the Federal
Character Principle.
Generally, Nigerians in the Southern zones enjoy a better standard of liv-
ing than their Northern compatriots. As Table 8.12 shows, the Northern
zones lead in the percentage rate of Nigerians classified as ‘core poor’. In the
same vein, the Northern zones perform dismally relative to the Southern
zones in most of the human development indicators (Table 8.13). Further
disaggregation of standard of living measures indicates that the Southern
zones also fare relatively better in access to improved sources of water
and electricity. However, there is a mixed record with respect to access to
improved and not-shared toilet facilities, as the North West (47.6 per cent)
and South East (39.6 per cent) zones fare better than the South West (17.8
per cent) and North Central (22.3 per cent) zones (Table 8.14). Nevertheless,
the Southerners appear to be wealthier than the Northerners. As Table 8.15
shows, a greater percentage of Southerners belong to the highest wealth
quintile, while the Northerners have a greater percentage of their population
in the lowest wealth quintile.
The pattern of socioeconomic HIs derives from a number of historical,
economic, and ecological factors, notably the level of educational attain-
ments as well as the type and volume of economic activities. Socioeconomic
HIs are not driven or sustained by any specific state policies aimed at priv-
ileging the South. On the contrary, the North appears to be favoured in
state redistribution initiatives. For instance, Table 8.16 shows that more than
9000 public health facilities are located in the North as against 5000 in the
South. However, this does not imply that the Northern zones fare better than
190 Nigeria
Table 8.13 Human development summary statistics by zone
Zones Human Human Gender Gender Inequality

development poverty development empowerment measure
index (HDI) index (HPI) measure measure (INQ)
(GDM) (GEM)
North West 0.420 44.15 0.376 0.117 0.44

North East 0.332 48.90 0.250 0.118 0.42
North Central 0.490 34.65 0.478 0.244 0.49
South West 0.523 21.50 0.507 0.285 0.48
South East 0.471 26.07 0.455 0.315 0.38
South South 0.573 26.61 0.575 0.251 0.41
Source: UNDP (2009b: 93).
Table 8.14 Measures of standard of living by zone, 2008 (% within zone)
Zone Population Population with Population with

with access to access to access to improved
improved water electricity not-shared
source sanitation facility
North West 49.8 38.5 47.6

North East 32.6 25.1 27.1
North Central 47.4 32.5 22.3
South West 68.8 70.8 17.8
South East 68.2 64.5 39.6
South South 57.8 56.8 26.2
Source: NPC of Nigeria and ICF Macro (2009: 321, 324, 327).
Table 8.15 Distribution of population by wealth quintiles by zone, 2008 (% within

zone)
Zone Lowest quintile Middle quintile Highest quintile
North West 31.9 17.2 7.6

North East 47.4 16.4 3.2
North Central 20.6 25.3 13.8
South West 4.2 15.3 44.6
South East 4.6 28.1 25.8
South South 6.7 22.9 26.0
Source: NPC of Nigeria and ICF Macro (2009: 26).
the Southern zones in health indicators. This is because a considerable pro-

portion of Southerners depend on private and third sector health services
providers. In fact, the Southern zones account for 68 per cent of privately
owned health facilities in the country.
Ukoha Ukiwo 191
Table 8.16 Ownership of health care facilities by

zone, 2004
Zone Public Private
North West 3,874 420

North East 2,225 353
North Central 3,306 1,852
South West 2,439 2,481
South East 912 2,416
South South 1,841 1,285
FCT 10 225
Total 14,607 9,032
Source: NBS Nigeria (2009: 183).
Consequently, the relative advantage enjoyed by the Southern zones in

socioeconomic HIs is autonomous of, and not directly related to, state
policies. A cursory examination of indicators such as the number of compa-
nies registered in the different zones, the number of professionals based in
the different zones, and the number of non-governmental organisations in
different zones show that there are more diverse economic and high-income-
yielding activities in the South than in the North (Mustapha, 2006). The
persistence of socioeconomic inequalities between North and South is partly
a product of the dearth of representation of the private sector and third sec-
tor in the North, and especially of the category of community self-help to
address human development challenges (Mustapha, 2007). This situation,
and the fact that Northern politico-military elites have held power for most
of the post-independence period, have significantly reduced the potential
for long-standing socioeconomic HIs to become the basis of mobilisations
against the Nigerian state.
8.3.3 Cultural status inequality

In terms of cultural status, the Nigerian constitution forbids discrimina-
tion on the basis of ethnicity, religion, and other identities. However, some
cultures have historically been more visible in the public sphere than oth-
ers. For instance, the constitution allows deliberations and presentations
in the National Assembly to be conducted in the languages of the three
major ethnic groups (Hausa, Igbo, and Yoruba) in addition to the English
language. National news broadcasts in the federal government-owned elec-
tronic media are also conducted in the languages of the major ethnic groups.
Even national dress codes privilege the cultures of the three major ethnic
groups, especially the Hausa. The dominance of the dress codes of the major
groups is, however, ameliorated by the symbolic adornment of key political
authority figures in the attires of different ethnic groups during important
192 Nigeria
events. For instance, it has become fashionable for Nigerian presidents (and
first ladies) to dress like their hosts during their tours of different states
around the country.
Moreover, the presence of state government-owned media allows full
expression to cultures of minorities in the states in which they live. The
prominence given to the major ethnic groups in the federal government-
owned media is also being neutralised by the emergence of private media
organisations. Some of the media houses, such as African Independent
Television (AIT), which is owned by a business tycoon from the South
South zone, have consciously promoted the visibility of minority cultures
in the public sphere. Finally, different states and local government areas
have distinct policies aimed at supporting the development of the lan-
guages and cultures of their communities. All these direct and indirect
forms of representation of most Nigerian cultures in the public sphere have
reduced the risks of conflict associated with cultural alienation in plural
societies.
8.4 Political process and institutional choice
The design of political institutions in Nigeria is largely informed by the

fear of ethnic mobilisation, ethnic domination, and national disintegration.
These three factors have contributed to the design of institutions that at once
seek to assuage both centrifugal and centripetal social forces. The challenge
has been in how to strike a balance between the need to grant autonomy
(and some degree of self-determination) to different identity-based groups
and the need to have a powerful government agency that is able to contain
secessionist and disintegrative tendencies.
Although the foundations of the institutional structures were laid during
the intense political negotiations that preceded the country’s independence,
the experience of a bloody civil war (1967–1970) has had an overwhelming
influence on the design of contemporary political institutions. The Nigerian
civil war was the culmination of a protracted series of political crises that
undermined the country’s First Republic (1960–1966). At the roots of the
political crises that led to military intervention on 15 January 1966 were
the quest for advantage by the political class and the organisation of the
general masses under the aegis of competing ethno-regional groups. The
crises undermined successive efforts to conduct a credible population cen-
sus and organise free and fair elections. Ethno-regional competition also
undermined the emergence of consensus on revenue allocation, delimita-
tion of constituencies, the ethnic composition of the armed forces and the
civil service, and the allocation of federal projects. The challenges of forg-
ing national consensus on the above issues were aggravated by a dearth of
political leaders whose legitimacy and popularity cut across ethno-regional
boundaries.
Ukoha Ukiwo 193
It was against this background that the military officers who seized power
sought to portray themselves as nationalists. Although the political class
had infiltrated the Nigerian military during the First Republic, the military
was one of the institutions that still had a national orientation at the time
the first coup took place (Luckham, 1971). However, certain actions and
inactions of the early military regimes, which culminated in the civil war,
eroded the autonomy of the military from social forces and its credibility as
a national institution. Successive post-war regimes have subsequently grap-
pled with the challenge of repositioning the military as a national institution
that would be instrumental in promoting national integration.
The prevailing perception that the virtual independence of the regions was
the bane of the Nigerian federation influenced the military regimes to intro-
duce institutions aimed at transforming the country from a federation with
a historically weak centre and strong federating units to a federation with a
strong centre and weak federating units. This centralising tendency was also
a product of the inherently hierarchical nature of the military as an insti-
tution. The policy instruments for the centralisation of the federation and
the enfeeblement of the federating units included dividing big regions into
smaller states and transferring more powers and resources to the federal gov-
ernment. The modality of military-engineered state-creation exercises also
factored in the need to discourage ethnic mobilisation. Consequently, the
boundaries of new states were drawn in such a way that they do not neatly
coincide with ethnic boundaries; also important was the fact that no state
was named after an ethnic group. As shown above in Table 8.1, the feder-
ating units increased from four regions to 12 states in 1967 and 19 states
in 1976.
Furthermore, the military jettisoned the Westminster cabinet system char-
acterised by shared leadership and collective responsibility in favour of a
presidential system where power is concentrated in the chief executive.
The introduction of a powerful executive at the federal and state levels
endowed with enormous powers and resources was intended to avert the
kinds of political crises that had been associated with jostling for power and
influence between the prime minister and the president in the First Repub-
lic. As midwives of Nigeria’s Second Republic (1979–1983), the military
regimes reckoned that the status of the prime minister as first-among-
equals was ill suited to a federation with strong, assertive, and centrifugal
ethno-regional-based groups.
The persistence and exacerbation of fears and allegations of ethnic dom-
ination also contributed to the attempts to further entrench power-sharing
and affirmative action policies in the 1979 Second Republic Constitution.
For instance, the Federal Character Principle, which stipulated that the com-
position of government institutions should reflect the diversity of groups
within the country, was enshrined in the 1979 Constitution. The Federal
Character Principle was strengthened further with the establishment, in
194 Nigeria
1997, of the Federal Character Commission (FCC), which was charged with
monitoring and enforcing application of the Federal Character Principle, and
the subsequent entrenchment of the Commission in the Fourth Republic’s
1999 Constitution.
Moreover, in its desire to promote national unity, the military government
introduced a number of institutions to discourage the creation of ethnic-
based parties. Since the Second Republic, successive Nigerian political parties
have been required to have a national outlook (Horowitz 1985). Successive
constitutions have stipulated that, to qualify for registration, parties should
have offices in a minimum of two-thirds of the federation’s states and that
they should also reflect the Federal Character in the selection of their offi-
cers. This stringent requirement for registration has been relaxed since 2007
when the law courts ruled that it infringed upon the fundamental rights of
Nigerian citizens to freedom of association. However, the electoral system
stipulates that in addition to securing a majority of votes, candidates for
presidential and governorship elections should also secure a quarter of the
votes in two-thirds of the states in the federation or of the local government
areas in the state, respectively, which continues to be a major incentive for
the emergence of inter-ethnic and cross-regional alliances in political party
formation.
8.5 How political institutions and policies work
In this section, I discuss the modus operandi of political institutions that

have arisen from the need to balance fears of ethnic domination and fears
of national disintegration. The discussion is structured on the template of
the two dimensions: the majoritarian/power-sharing dimension and the
centralised/decentralised dimension (see Chapter 2).
8.5.1 Majoritarian/power-sharing dimension

(A) Electoral system
Efforts to introduce an institutional mechanism for sharing political power
through the electoral system have failed in Nigeria. Despite the long-
standing criticism of the ‘winner-takes-all’ logic of the first-past-the-post
system, the country’s electoral system has remained distinctly majoritarian.
The system allows governments to be formed on the basis of simple plu-
rality, under a single-member constituency arrangement. Although some
have canvassed the introduction of proportional representation to give a
better reflection of electoral outcomes in government, arguments in favour
of the preservation of the majoritarian system have prevailed. This is princi-
pally because of fears that proportional representation would contribute to
a proliferation of parties and governmental instability.
The single-member constituency allows each political party to field one
candidate per constituency. The candidate that polls above 50 per cent in
Ukoha Ukiwo 195
an election is declared the winner of the election. However, some provisions

have been mainstreamed in the electoral system to ensure that the victo-
rious candidate enjoys the support of most segments of the constituency.
This is accomplished through the requirement enshrined in sections 133,
134 and 179 of the 1999 Constitution that candidates for presidential or
governorship elections can only be declared winners if they secure not less
than a quarter of the votes from two-thirds of the constituent states or local
government areas, respectively.
As has been observed in other political systems, the majoritarian system
works against the interest of minority parties as it does not recognise their
electoral strength in the formation of government. Consequently, the simple
plurality system has encouraged the emergence of a dominant party system
in Nigeria. Clearly, the designers of Nigeria’s electoral system have consid-
ered the eclipse of minority parties to be a lesser evil than the proliferation
of ethnic-based parties or the ethnicisation of parties.
(B) Party system

Nigeria operates a multi-party system. The failed attempt by the military to
impose a two-party system in the ill-fated Third Republic demonstrates the
craving of the diverse groups for a multi-party system. The Nigerian con-
stitution empowers the electoral commission to register and regulate the
activities of political parties. Ancillary legislation such as the Electoral Act
gives details of requirements for party registration and regulation. Since the
end of the civil war, electoral guidelines have prohibited the formation of
parties on the basis of ethnicity. Parties are forbidden from bearing names
or symbols associated with a particular group or from selecting most of their
officers from a particular ethnic group.
However, in practice, Nigeria’s party system can be characterised as a one-
party dominant system. As noted earlier, the majoritarian electoral system
has contributed to the emergence of a one-party dominant system. Since
independence the tendency has been for parties that won the formative elec-
tions to grow bigger and more powerful. As shown in Tables 8.17 and 8.18,
the People’s Democratic Party (PDP), which prides itself on being Africa’s
biggest party, has been the dominant player in Nigeria’s Fourth Republic.
The PDP has increased its holdings of seats in both chambers of the National
Assembly after three successive elections, in 1999, 2003 and 2007. While the
PDP won 68 per cent of seats in the Senate in 1999 it won 81 per cent in
2007. Likewise, the PDP’s control of the federal House of Representatives
has increased from 63 per cent in 1999 to 73 per cent in 2007. The party
also increased the number of states where it was declared the winner in
governorship elections from 21 states in 1999 to 28 states in 2007. This
emergent dominant party system has evolved from the culture of patronage
politics that favours political incumbents. Deprived of patronage resources
196 Nigeria
Table 8.17 Distribution of Senate seats by political party
Political party 1999–2003 2003–2007 2007–2011
No. % No. % No. %
PDP 74 67.9 76 69.7 88 80.7

ANPP 18 16.5 27 24.8 14 12.8
AD 17 15.6 6 5.5 – –
AC – – – – 5 4.6
PPA – – – – 1 0.9
ACCORD – – – – 1 0.9
Total 109 100.0 109 100.0 109 100.0
Table 8.18 Distribution of federal constituency seats by political party
Political party 1999–2003 2003–2007 2007–2011
No. % No. % No. %
PDP 226 62.8 228 63.3 261 72.5

ANPP 74 20.6 97 26.9 62 17.2
AD 60 16.7 32 8.9 – –
APGA – – 3 0.8 – –
AC – – – – 33 9.2
PPA – – – – 3 0.8
LP – – – – 1 0.3
Total 360 100.0 360 100.0 360 100.0
to lubricate their networks after losing elections, leading members of the

minority parties tend to decamp to dominant parties.
Since the advent of the Fourth Republic, scores of legislators elected on
the platform of opposition parties have joined the ruling parties in both
the federal and state legislatures. The executive arm of government has
not been immune from the virus of ‘cross-carpeting’ (that is, ‘crossing the
carpet’, or defecting to another party) as state governors and ministers in
the federal cabinet have moved from one party to another without losing
office. The 1999 Constitution allows for cross-carpeting in cases of factional
crises and conflicts in political parties. The constitutional provision has
enabled incumbents to undermine and weaken opposition parties and has
aggravated the parlous state of party discipline. Moreover, cross-carpeting is
driven by the lack of fundamental ideological differences between the par-
ties, which has transformed parties into mere election-winning machines.
Ukoha Ukiwo 197
The attrition rate from minority parties is, however, mitigated by the ten-
dency for parties that lose national elections to win elections in vital states
or zones. Control of state power in some states enhances prospects for the
survival of the party if the persons elected under its platform remain com-
mitted to the party and provide it with patronage resources. The survival
of the All Nigeria Peoples Party (ANPP) and the Action Congress of Nigeria
(ACN) has stemmed from their ability to retain control in some strategic
states, where loyal party members were elected as governors. Such governors
have in some cases been able to lure members of other parties, including
the dominant PDP, to their parties. The PDP has also lost members in situa-
tions where there is a conflict of interest between state-based party stalwarts
and the national leadership. For instance, the quest of some prominent Igbo
politicians to contest for the presidency contributed to the PDP’s loss of two
states in the South East zone in 2007.
(C) Composition of the Executive

The dominant party system is also driven by the constitutional provision
that does not require the inclusion of other parties in the cabinet. Against
this background, therefore, the normal shape of the executive is one in
which the single party that won the elections dominates. Thus, since 1999,
the ruling PDP has dominated the federal Executive Council. However,
because the constitution does not also forbid the inclusion of opposition
party members, victorious parties have often deemed it expedient to appoint
some ministers from the ranks of opposition parties. Such gestures have
arisen from the need to form governments of national unity in the aftermath
of contentious elections. This is the case when some parties make a very
strong showing in particular sections of the country. For instance, one of the
people President Olusegun Obasanjo appointed to his cabinet was Chief Bola
Ige, a stalwart of the Alliance for Democracy (AD), which won and controlled
all six states in the Southwest in the 1999 elections. Governments of national
unity are also considered expedient when the results of the elections are
strongly disputed. Ruling parties incorporate elements of the opposition par-
ties to mitigate the legitimacy deficits confronting them. For example, in
2007, President Umaru Musa Yar’Adua formed a government of national
unity (GNU) including a few ministers from ANPP and the Progressive
Peoples Alliance (PPA), which controlled some states of the federation.
The dominant party, however, always reaps more dividends from GNU
arrangements. The power-sharing arrangement not only helps to beef up
the legitimacy of the dominant party but also contributes to the factional-
isation of participating opposition parties, as appointed persons are easily
torn between loyalty to their own parties and loyalty to the chief exec-
utive, who pays the piper and is more inclined to seek to dictate the
tunes. Independent-minded ministers are often dropped in frequent cabinet
198 Nigeria
reshuffles. Ultimately, the dominant party dispenses of the services of such

ministers once it has addressed legitimacy challenges, as the fraudulent elec-
tions fade in the public memory. Consequently, as the experiences of the
AD, ANPP and PPA in Nigeria’s Fourth Republic have amply demonstrated,
participation in a GNU has become what Nigerian politicians have called a
‘banana peel’ (a booby trap) for opposition parties.
(D) Head of government

Since 1979, Nigeria has been operating a presidential system of government.
This system is characterised by the fusion of the posts of head of state and
head of government into one individual: the president, governor, or local
government council chairman. The chief executive does not share power,
but is unmistakably the head of the executive arm of government. Having
secured a mandate in direct elections and not through the legislature, the
chief executive is relatively insulated from legislative control, except in
the case of a breach of the constitution, where the legislature can wield
the ultimate power of impeachment. The president (governor) appoints
members of the executive council and boards of other governmental agen-
cies. Apart from controlling and directing all governmental agencies and
taking charge of the day-to-day administration, the chief executive is also
commander-in-chief of the armed forces and the ceremonial head of state.
The powers, authority, and influence of the head of the executive also
derive from the constitutional provisions that provide for the chief execu-
tive to convene the first sitting of the legislature and to assent to bills before
their passage into law. The chief executive also appoints judicial officers and
officers of the armed forces. Although the constitution stipulates that the
power to approve appointments and budgets is vested in the legislature, the
influence of the chief executive, who is also more often than not the de
facto leader of the dominant party in the legislature, guarantees easy passage
of executive bills and successful screening of ministerial nominees. How-
ever, legislative confirmation of nominees is not always automatic and all
smooth sailing. This is especially so in situations where the legislature is
controlled by an opposition party. Nevertheless, the chief executive can still
influence legislative decisions through the ingenious deployment of carrots
and sticks in the context of poor party discipline and the prevalent culture
of the politics of the belly.
Thus, by concentrating power in the institution of the executive, the pres-
idential system of government also contributes to reinforcing the tendency
towards a dominant party system. The concentration of power in the chief
executive raises the stakes for securing the post. This, therefore, makes the
position the most highly prized post for political competition. The political
contestation for the office of the chief executive, especially in the context of
simple plurality, is framed in a zero-sum context as political actors consider
it to be a ‘do or die’ affair.
Ukoha Ukiwo 199
Against the backdrop of the overconcentration of powers in the institu-

tion of the chief executive and the attendant contestations among different
ethno-regional groups for the position, some informal political institutions
have evolved to facilitate the dispersal of power. The principal objective of
this arrangement is to allow for some level of alternation in holders of the
principal political position among the constituent ethno-regional groups
and to prevent any one group from monopolising this highly coveted posi-
tion. This innovation was introduced by the National Party of Nigeria (NPN),
the dominant party in Nigeria’s Second Republic. The party applied a pol-
icy of zoning, whereby different geopolitical zones were allocated political
offices.
Although non-constitutional, the elite pact to alternate power among the
geopolitical units has contributed to the evolution of non-ethnic and cross-
ethnic parties. Politicians from minority areas in particular have found the
cross-ethnic parties irresistible and tend to pitch their tents with them. This
contributed substantially to the electoral successes of the NPN (and the PDP)
in minority areas and their eventual ascendancy as dominant parties. How-
ever, the major problem with zoning and the rotation of political offices is
that it is based on an elite pact, which is mostly non-justiciable. Political
crises have emanated from conflicting interpretations of the provisions of
such pacts and the alleged refusal of key actors to honour them in real-life
situations. For instance, the NPN was thrown into crisis over the alleged
contravention of the zoning formula by fielding incumbent President Shehu
Shagari for a second term, as some factions claimed the party zoned the
presidency to the Southwest in 1983. The crisis that rocked the party con-
tributed to the military intervention of December 1983 (see Othman 1984).
In 2003, the PDP opted to make zoning part of its constitution and to make
key players in the party sign up to the zoning formula. This arose from the
controversy surrounding the second-term ambitions of President Olusegun
Obasanjo, amid the claims of some founders and leaders of the party of the
existence of a pact that restricted Obasanjo to a single term. However, neither
the elevation of the principle of rotation of power into the party’s constitu-
tion nor the appendage of signatures of party stalwarts to resolutions on
the principle were enough to deter President Olusegun Obasanjo from nurs-
ing his third-term ambitions in 2007, nor was it enough to stop President
Goodluck Jonathan from denying the existence of power rotation in the
PDP when Yar’Adua died, after protracted ill-health, in 2010.
8.5.2 Centralised/decentralised dimension

(E) Forms of government
Nigeria is a federal republic. From a three-region federation, the country is
currently divided into 36 states, one federal capital territory and 774 local
government areas. The constitution divides functions and powers among
the three tiers of government. The rationale for adopting federalism has
200 Nigeria
remained the need to preserve unity in the midst of diversity. The fed-
eral structure was considered to be the ‘natural’ form of government in a
country with over 300 ethnic groups who want to preserve their individual
identities while seeking common protection and benefits under the Nigerian
federation.
Federalism is the overarching power-dispersing institution in Nigeria.
Nigeria’s federalism has been calibrated to address the twin fears of ethnic
domination and national disintegration. As indicated earlier, this has been
accomplished by creating centres of countervailing power at the federal,
state and local levels. Federalism, in its three-tiered format, provides a sphere
of autonomy for most identity-based groups. Groups that are demographi-
cally insignificant in the context of national politics can become salient at
the level of either state governments or local councils.
Nigerian federalism was adopted not just in recognition of the diverse
ethnic topography of the country, but also as an institutional mechanism
to address HIs that contribute to violent group mobilisation. It is apposite
at this juncture to examine the extent to which the complex federal archi-
tecture has impacted on the nature of HIs. This will be examined through
an analysis of the outcomes of some strategies designed over the years to
operationalise Nigeria’s federalism.
States creation. One of the distinct features of federalism in Nigeria is the

division of the subnational and federating units. As noted earlier, the main
driver for states creation was mobilisation in each of the three regions against
domination of the major ethnic groups. From four regions when the coun-
try became a republic in 1963, Nigeria currently has 36 states and a federal
capital territory. Apart from the quest for autonomy, states creation served
the purpose of the designers of Nigeria’s constitution, who perceived the
lopsided nature of the situation where one federating unit (the North) was
bigger in landmass and population than the two other units put together.
This tended to encourage destabilising dyadic and triadic contestations that
adversely impacted on national cohesion. The large-sized and well-endowed
regional tails wagged the national dog.
States creation has promoted national stability by diverting conflicts to
states and local councils. It has also satisfied the yearning of marginalised
groups for better representation due to the enlargement of the number of
legislative seats in the National Assembly and the establishment of new seats
at the state level. What is more, new states have enhanced the prospects of
the culture of minority ethnic groups gaining visibility in the public sphere.
Finally, creation of states has contributed to addressing political HIs by
opening new opportunities for political appointments.
Federal Character Principle. This possibility is enhanced by the entrench-

ment of the ‘Federal Character’ principle in the Nigerian constitution. The
Ukoha Ukiwo 201
principle was aimed at ensuring the spread of opportunities across the coun-
try and preventing ethnic domination. More precisely, the objectives of the
Federal Character Principle are:
1. To promote national unity;

2. To command national loyalty; and
3. To ensure that there shall be no predominance of persons from a few
states or from a few ethnic or other sectional groups in government or
any of its agencies.
The Federal Character principle is a variant of consociationalism. It has

served as an institutional framework for ensuring that the chief execu-
tive of the federal government selects members of the cabinet from the
federating states (and from local government areas in the case of state gov-
ernments). It has therefore contributed to ensuring that no state of the
country is totally neglected in political appointments. Although the prin-
ciple has not eliminated the syndrome of marginalisation, it has recorded
modest success in mitigating domination and alienation. Consequently, the
scope of the principle has been enlarged to include employments and pro-
motions in the public service and the allocation of federally funded projects.
Moreover, the justiciability of the principle has been enhanced by the estab-
lishment of the Federal Character Commission (FCC), which is charged with
documentation and monitoring of implementation.
The Federal Character Principle has virtually become a political doctrine.
It is one of the major criteria considered by the legislature in screening
nominees for public offices and considering appropriation bills. While the
legislative arm of government performs oversight to ensure compliance
by the chief executive, the FCC monitors government agencies to ensure
that recruitments do not disproportionately favour some states or ethnic
groups. For instance, the FCC, which is statutorily empowered to participate
in recruitment exercises conducted by federal government agencies, inter-
vened in 2010 to check disproportionate employment of persons from a
particular section of the country in the Independent National Electoral Com-
mission (INEC). Nigerian legislatures have often passed resolutions to revoke
appointments and promotions that have contravened the Federal Character
Principle. For instance, the Benue State House of Assembly called for the can-
cellation of a recruitment exercise of the Federal Roads Safety Corps (FRSC),
conducted in the state in 2010, because it favoured a particular section of
the state.
The intervention of the legislature and the watchdog FCC is usually driven
by vigilant social forces seeking to protect some kind of sectional interest.
This is because the principle has created a high level of consciousness and
sensitivity to domination and alienation. The media is often inundated with
202 Nigeria
reports of how particular agencies have contravened the Federal Charac-

ter Principle. For instance, Daily Trust, a newspaper that purports to speak
for the North, ran a cover-page story on how the North was marginalised
in senior management promotions in the Nigerian National Petroleum
Corporation (NNPC) conducted in 2011:
Forty five per cent of the total number of management positions in the
recent promotion exercise conducted by the Nigerian National Petroleum
Corporation (NNPC) went to officers from the South South geopoliti-
cal zone . . . The home region of both President Goodluck Jonathan and
Minister of Petroleum Resources Diezani Allison-Madueke grabbed 17
new management staff positions out of the 38 on offer. The docu-
ments indicated that six out of the top eight management positions
of Group General Manager (GGMs) are occupied by South South indi-
genes . . . . In the promotion list, Southeast came second with seven posts
equivalent to 18 per cent while Northwest came third with 5 new posts
or 13 per cent. South West and North Central zones occupied the fourth
position with 4 new posts each which or 10.5 per cent, while the North
East was left holding the hat with only one post . . . In all, the North has
only 10 new posts out of 38 and the remaining 28 went to the southern
part of the country.
(Muhammad, 2011)
The embarrassments elicited by such a public outcry serve as a restraint on

public officers that would have used their positions to unduly favour particu-
lar sections of the country. For instance, President Obasanjo was embarrassed
by the FCC publication on the composition of the Federal Civil Service in
2002, which showed that his native Ogun State exceeded its quota of jobs in
the federal civil service.
Although there are concerns that its introduction in the public service
would undermine merit, federal agencies are increasingly adopting a for-
mula that assigns fixed quotas to states but applies merit-based criteria in
the selection of persons from each state. The rationale for this is to select
the best from each identity-based group. This approach has been used by
educational institutions and scholarship agencies to select candidates. For
example, the Petroleum Technology Development Fund (PTDF) applied the
quota-cum-merit-based selection procedure in the 2010 overseas scholar-
ship programme. The best five candidates from each geopolitical zone were
selected and placed on the national merit list; two candidates were selected
on the basis of gender from the North and South, while two other candi-
dates were selected from oil-producing states. The application of quotas in
admission into universities has also significantly increased the proportion
of persons from so-called ‘educationally less-advantaged states’ who have
access to university education.
Ukoha Ukiwo 203
Another area where the application of the Federal Character Principle has
contributed to tackling HIs is in the location of federal institutions. Since
the 1970s, when it started to proactively promote university education, the
Federal Government has pursued a policy aimed at establishing federal uni-
versities in all states of the federation (Ukiwo, 2007). Although economic
recession forestalled realisation of this objective in the 1980s and 1990s,
improvements in the revenue profile of the federal government have made
the government respond to public pressure for the establishment of federal
universities. In 2011, President Goodluck Jonathan approved the establish-
ment of nine more federal universities. The result of such initiatives is that
most states now have educated elements to work in the state civil service
and fill the state’s quota in the federal civil service, thereby reducing ten-
sions that arose from the dependence of some states/regions on educated
elements from other states for the basic functioning of public services. Simi-
lar initiatives have been adopted by the federal government in the location
of tertiary health care institutions and airports, as well as in the construction
of federal highways.
Another intrinsic mechanism of the country’s federalism that tackles HIs is
the Federation Account. The 1999 Constitution allows for federally collected
revenues to be pooled into an account known as the Federation Account and
empowers the National Assembly to legislate on the formula for distribution
of the revenues. These include vertical allocation between the three tiers of
government (federal, state and local government) as well as horizontal allo-
cation among the states. Among the criteria applied for revenue allocation
are derivation, population, landmass, equality of states and ecological chal-
lenges. The criterion of equality of states has led to the provision of block
grants that have enhanced the capacity of less-endowed and indigent states
to provide basic state services. Although such equalisation transfers have not
bridged historical gaps, given disparities in resource endowment and tax col-
lection capacities, they have ensured the survival of many non-viable states,
and, ipso facto, the stability of the federation. Nevertheless, there is still no
consensus about a revenue allocation formula, and contestations over the
revenue allocation formula remain a perennial threat to national cohesion.
The contentious issues remain the share of the federal government, and the
demands of oil-producing states for increment of the proportion allocated
on the basis of derivation.
8.6 Popular perceptions
Popular perception surveys are being used increasingly in social studies

to ascertain the extent to which views expressed by government officials,
community leaders and the mass media are reflective of the position of
the average citizen. This is particularly the case in deeply divided societies
where group leaders are interested in amplifying differences and inequalities
204 Nigeria
Table 8.19 Three most important elements of identity (%

within group)
Group Nationality Religion Ethnicity Job
Hausa 40 90 26 38
Igbo 28 72 33 61
Yoruba 42 65 19 55
Note: Multiple choice question.

Table 8.20 Feelings about ethnic and national identity (%)
Group I feel only I feel more I feel equally I feel more I feel only I don’t Total
Nigerian Nigerian Nigerian and [ethnic [ethnic know
than [ethnic identity] identity]
[ethnic identity] than
identity] Nigerian
Hausa 25 8 51 8 7 100.0
Igbo 3 13 60 17 7 100.0
Yoruba 10 14 60 12 2 100.0
All respondents 11 14 58 12 5 100.0
between groups (Langer and Ukiwo, 2008). It is against this background that
a perceptions survey was conducted for the present study. As a result of
resource constraints, the study was conducted in only one location: Lagos.
The choice of Lagos – Nigeria’s capital between 1906 and 1991 – was predi-
cated on its position as the most ethnically and religiously diverse city in
the country. This was based on the need to ensure that Nigerians from
all social backgrounds had an equal opportunity of being selected for the
survey.
The study sampled 12 enumeration areas with concentrations of diverse
populations. A total of 412 respondents based on a 50/50 gender quota
were selected for the study. Based on the random sampling method adopted
for the study, as expected, most of the respondents in the survey were
from the major ethnic groups, as follows: Yoruba, 45 per cent; Igbo, 22 per
cent; Hausa, 17.5 per cent; Southern minorities, 9.7 per cent and Northern
minorities, 6 per cent.
The study confirmed the perception that Nigerians attach importance to
their religion. However, it did not confirm the perception that Nigerians con-
sider their ethnic identity to be more important than their national identity,
as seen in the responses of persons from the major ethnic groups shown in
Tables 8.19 and 8.20.
Ukoha Ukiwo 205
The study confirmed the perception and official statistics that indicate that
Southern Nigerians generally fare better than Northern Nigerians in social
and economic indicators.
There was no evidence of ethnic animosity, as there was no strong objec-
tion to intermarriage on the basis of ethnicity. Most of the respondents
who would object to their sister or daughter being married to someone
from another group would do so on religious grounds. In fact, most of the
respondents (95 per cent) indicated their willingness to work with persons
from other ethnic groups, and only a cumulative 6 per cent of respondents
claimed that their perception of people from any other ethnic group was
negative. Positive attitudes were also generally expressed about the current
and future state of ethnic relations in the country.
The respondents were, however, polarised in their perception of the
importance of ethnic identity in access to state resources such as government
jobs, government contracts and admissions to public universities. Gener-
ally, respondents from Southern ethnic groups felt ethnicity mattered a great
deal for access, while their counterparts from the North felt ethnic identity
was not important. However, most of the respondents felt that ethnicity has
become more important in the public sphere since 1999.
Interestingly, most of the respondents felt that the socioeconomic status
of their ethnic group was better than it had been 20 years ago. Signifi-
cantly, though, the Igbo felt they were underrepresented in government
and the armed forces. There was strong support for affirmative action for
marginalised groups, and most of the respondents felt the federal govern-
ment accorded recognition to their culture, even though most of them felt
the Hausa were given the greatest visibility in the public sphere.
The survey confirmed the perception among the Igbo that they are
marginalised in the Nigerian federation. Perception of marginalisation
tended to influence voting behaviour with the Igbo respondents more than
any other group, indicating they would attach importance to the ethnic
identity of candidates during presidential elections. This probably explains
the strong clamour for an Igbo presidency, as this group is the only major
ethnic group that has not produced a democratically elected president.
8.7 Concluding remarks
Since the end of the civil war in 1970, Nigeria has continued the nation-
building project by institutionalising mechanisms that would simultane-
ously protect national stability and prevent ethnic domination. The fact
that the country has not experienced any major secessionist conflict dur-
ing this 40-year period testifies to the modest progress recorded in national
integration. The country has so far defied the thesis that countries that
have experienced civil wars are likely to experience repeated implosions (see
Collier and Hoeffler, 2004). This has largely resulted from the ability of the
206 Nigeria
country to adapt the mixture of power-concentrating and power-dispersing

institutions to its fractious history.
Against the backdrop of the fear of national disintegration, the archi-
tects of Nigeria’s political system have concentrated power in certain state
institutions charged with the ultimate responsibility of ensuring national
security. These institutions include principally the presidency and the fed-
eral government. The rationale of what has been frequently described as the
‘centralised’ federalism and ‘imperial’ presidency is the need to have institu-
tions that can ultimately ensure that the federation does not fall apart. While
endowing these key state institutions with enormous powers, the crafters of
the Nigerian constitution have been minded to introduce political institu-
tions that ensure that no single group or bloc of groups monopolise control
and dominance of state institutions. This insurance policy is targeted at the
morbid fear of ethnic domination in the country and explains the flour-
ishing of power-dispersing institutions. Thus, although Nigeria ranks low
in the majoritarian/power-sharing category as shown in Chapter 2, it has a
surfeit of power-sharing arrangements. The distinctiveness of the Nigerian
system lies in the fact that the target of power-sharing is not the political
parties but ethno-regional groups. By isolating political parties in power-
sharing arrangements, the architects of the Nigerian political system have
been inspired by the need to discourage the formation and or evolution of
ethnic parties.
Although it is not flawless, given the high incidence of intermittent con-
flicts and power struggles in the country, the mixture of power-concentrating
and power-sharing arrangements have helped to prevent any major seces-
sionist conflict. The success of the institutional design is its ability to make
the Nigerian federation so attractive that no ethno-regional group in con-
temporary Nigeria seeks to opt out of the federation. What most groups seek
is better representation and more inclusion. Thanks to the elasticity of its
power-sharing institutions and the providence of oil wealth, the Nigerian
federation has continued to improve upon its capacity to meet the legitimate
desires of diverse groups for inclusion. Nigeria is thus a case par excellence
of the secessionist conflict deficit in Africa (Englebert and Hummel, 2005).
The Federal Character Principle, which is the principal power-sharing
mechanism, has been particularly successful in addressing political HIs and
mitigating the pervasive impact of socioeconomic HIs. However, the Federal
Character Principle has created a culture of hypersensitivity to domination,
and thereby promoted endless protests and mobilisations against marginali-
sation. Much of the observed instability arises from such protests and mobili-
sations. Still, the Nigerian federation has been able to mitigate the impact on
the federation of such conflicts, because power-sharing arrangements target
states and local government areas, not ethnic groups per se.
The Nigerian federation is still grappling with some of the unintended
effects of its post-civil war political architecture. These include the perennial
Ukoha Ukiwo 207
indigene versus non-indigene conflicts, the zero-sum context of electoral

politics and high incidence of electoral fraud, the concentration of power
in the presidency, and the disproportionately large portion of national rev-
enues retained by the federal government. The federation has so far shown
enough resilience to address these challenges. Its future course would depend
on deepening democracy and massive non-oil-based economic development
that tackles poverty and inequality. Democratic consolidation, especially the
strengthening of democratic institutions, would promote constitutionalism
and produce free and fair elections, which are rare in the Nigerian context.
Non-oil-based development would strengthen Nigeria’s fiscal federalism by
reducing the dependency of states on the federally managed dispensation
of oil rents, which is one of the drivers of the perennial conflicts in the
oil-producing Niger Delta region.
9
The Relationship between Objective
and Subjective Horizontal
Inequalities: Evidence from Five
African Countries
Arnim Langer and Satoru Mikami
9.1 Introduction
In recent years an increasing amount of both qualitative and quantitative

research has shown that the presence of severe inequalities between ‘cultur-
ally’ defined groups such as ethnic or religious groups – or what Stewart
(2002) has termed horizontal inequalities (HIs) – makes countries more
susceptible to a range of political disturbances, including violent conflict
and civil war (see, for example, Langer, 2005; Mancini, 2008; Østby, 2008;
Stewart eds, 2008; Cederman et al., 2011). It appears that the risk of violent
conflict increases in particular if political and socioeconomic HIs are ‘con-
sistent’ or run in the same direction: that is, a situation where an ethnic
group is both politically excluded and relatively disadvantaged in socioeco-
nomic terms (Langer, 2005, 2008; Østby, 2008).1 Most quantitative studies
that have found evidence in support of the relationship between the pres-
ence of HIs and the emergence of violent conflicts have used an objective
measure of socioeconomic HIs, such as a household asset index or a school-
ing inequality index, rather than a measure of perceived inequalities, in
their statistical models (see, for example, Mancini, 2008; Østby, 2008).2 With
respect to the assessment of political HIs, the situation is somewhat different.
Due to the lack of cross-sectional data on the evolution of objective political
HIs, most quantitative studies testing the relationship between HIs and con-
flict have included ‘semi-objective’ or ‘partly subjective’ measures of political
HI. For instance, several studies have used the ‘semi-objective’ Minorities
At Risk (MAR) dataset for assessing a country’s level of political HI (see,
for example, Østby, 2008). The MAR dataset tracks the political, economic
and cultural status and position of 282 ethno-political minorities around
the world – minorities with at least 100,000 members that are at risk of
208
Arnim Langer and Satoru Mikami 209
discrimination – by collecting and analysing a wide range of open-source

information, which is coded by MAR researchers into a limited number of
categories.3
While the quantitative studies on HIs and violent conflict have con-
tributed enormously towards establishing the relationship between these
two concepts, the operationalisation of HIs in objective terms is to some
extent problematic, because people act on the basis of their perceptions of
the world in which they live, and these perceptions may differ substantially
from the objective reality. With respect to the relationship between HIs and
group mobilisation, Frances Stewart rightly notes that: ‘People take action
because of perceived injustices rather than because of measured statistical
inequalities of which they might not be aware’ (Stewart, 2010b: 14). More-
over, most quantitative studies on HIs and conflict de facto assume that there
is consistency between objective and subjective HIs.
The question to what extent objective and subjective HIs are consistent
in practice is clearly a very important empirical question, which has been
largely ignored in the literature on HIs. This chapter aims to contribute to
filling this void, by analysing the relationship between objective and sub-
jective HIs in five African countries – Ghana, Kenya, Nigeria, Uganda and
Zimbabwe – which are all confronted with sharp socioeconomic inequali-
ties between their major ethnic groups and/or regions. In order to explore
the relationship between objective and subjective HIs, we have conducted
perceptions surveys in each of our five case-study countries.4 The surveys
were not nationally representative, but we did ensure that each of our
national survey samples contained a sufficiently large number of respon-
dents from all the major ethnic and religious groups. The results are therefore
only statistically representative for the selected survey locations, but we can
draw wider inferences based on the assumption that the surveyed areas are
qualitatively representative of a larger part of society. Table 9.1 provides an
overview of the survey locations and the number of interviews conducted in
each of our case-study countries.
In addition to our own surveys (that is, the JICA survey), we also
use the Afrobarometer Round 4 surveys,5 which cover similar topics and
issues, although these surveys do not have the same degree of detail and
Table 9.1 Overview of survey locations and number of interviews
Country Survey sites and number of interviews Total
Ghana Accra (406) 406

Nigeria Lagos (412) 412
Kenya Nairobi (300), Nakuru (303) and Mombasa (304) 907
Uganda Kampala (200), Gulu (100) Hoima (100) and Mbarara (100) 500
Zimbabwe Harare (294) and Bulawayo (108) 402
210 Objective and Subjective Horizontal Inequalities
extensiveness when it comes to issues of inequality and identity. However,

a major advantage of the Afrobarometer surveys is that they are nationally
representative. By using these two sets of surveys in a complementary way,
we will greatly enhance the robustness of our findings.
The chapter proceeds as follows. The next section will reflect on the rea-
sons why there can be a mismatch between objective and subjective HIs in
particular situations or countries. In Section 9.3 we will examine the pre-
vailing objective HIs in our five case-study countries. In Section 9.4 we will
analyse the extent to which individual ‘risk’ factors associated with lower
standards of living (such as educational attainment) can help to explain
the observed ethnic inequalities. In Section 9.5, in turn, we will analyse the
extent to which these individual ‘risk’ factors are themselves unequally dis-
tributed across different ethnic groups. In Section 9.6 we will then analyse
people’s perceptions of the prevailing HIs and analyse the extent to which
there are discrepancies between the objective and subjective situations. In
the last section we will draw some conclusions.
9.2 Why objective and subjective inequalities may differ
In this section we will examine the main reasons why people’s perceptions of
the prevailing HIs in a country may differ sharply from more objective mea-
surements or assessments of these inequalities. One issue that complicates
matters in this respect is that there may be sharp differences in the perceived
inequalities across ethnic groups. Thus, for instance, in societies with sharp
objective HIs (possibly resulting from past and/or ongoing discriminatory
practices by the state), it is not unlikely that the deprived groups will ‘cor-
rectly’ perceive that they are in a relatively disadvantaged position compared
to other groups, but their perceptions may nonetheless reflect a considerably
worse or better picture than the one that emerges from the analysis of objec-
tive data. Relatively advantaged groups in horizontally unequal societies
may also ‘correctly’ perceive their relatively privileged position, although
they may have very different views about the level of inequality compared
with disadvantaged groups and also of the causes of the prevailing inequal-
ities. Moreover, even in cases where the objective HIs are not very severe,
there may still be substantial differences in the perceived levels of HI across
different ethnic groups.
There are a number of reasons why there can be a mismatch between the
objective and subjective HIs in a particular society, which we outline below.
• Impact of objective personal situation on perceived group situation. When

asked to assess the prevailing socioeconomic HIs in a country, people
should not let their personal socioeconomic situation interfere with or
blur their perceptions. Indeed, assuming that the prevailing objective
HIs can be perceived correctly by individuals, two people from the same
ethnic group with different levels of income and welfare should in prin-
ciple have the same perceptions about their group’s situation and relative
position. However, it is not unlikely that people’s individual socioeco-
nomic background and situation may colour their perceptions of the
prevailing group inequalities.
• Manipulation of perceptions by elites or group leaders. In order to gain polit-
ical support (or pre-empt losing it), the leaders or elites of a particular
group may decide to manipulate their constituents’ perceptions of the
prevailing HIs. While elites occasionally attempt to mitigate perceptions
of inequality (for example, to pre-empt criticism that they have not done
enough to improve their group’s socioeconomic situation and relative
position), it appears to be more common that they try to exacerbate
the existing perceptions of inequality among their group members or
constituents in order to gain or maintain political support. Leaders of
relatively advantaged groups, in turn, may play down the severity of the
prevailing inequalities and concomitantly stress that the deprived groups
are themselves to blame for their relatively disadvantaged situation.
• Inaccurate media reporting. The media can play an important role in
bringing objective HIs to the attention of the population at large. Yet,
inaccurate reporting on the part of the media due to sloppy reporting, a
lack of sufficiently qualified and experienced journalists, or for political
reasons can clearly have a major impact on people’s perceptions of the
existing HIs and possibly their perceived causes.
• Lack of objective data on horizontal inequalities. Ethnically segregated
socioeconomic data are usually not readily available. Sometimes – as, for
example, in Nigeria – ethno-cultural variables are not included in surveys
because of their political sensitivity (Okolo, 1999). While language and
region can sometimes be used as proxies for ethnic groups, in a substan-
tial number of countries this might not be possible or might not provide a
sufficiently accurate picture of the prevailing HIs. The absence of accurate,
comprehensive and independent data on HIs in many multiethnic coun-
tries increases the risk that people’s perceptions might instead be based
on personal experiences, opinions and stories of friends, family and peo-
ple in positions of ‘power’ (such as politicians, community leaders and
church leaders), or even on rumours and hearsay.
• Insufficient access to information. Another reason why objective and sub-
jective inequalities may differ is because certain groups may lack access
to the necessary information and data to form a reasonably accurate pic-
ture of the prevailing HIs in their country. Thus, for instance, groups in
rural and geographically remote areas may have insufficient access to the
media or other sources of information, which in turn makes it difficult for
them to compare their own situation to that of other groups. Low mobil-
ity among the people living in rural and remote areas is another obstacle
for assessing the relative position of their own group.
• ‘Misleading’ comparisons. HI is a relational concept that essentially requires

comparing different groups’ positions to the position of a selected other
group (such as the richest group in a country) to an average measure
of performance (such as the national average), or to a group’s relative
demographic size. People’s perceptions of the prevailing inequalities are
clearly affected by the ‘yardstick’ they implicitly or explicitly use to assess
their group’s relative position vis-à-vis other groups. The government, the
media and community and church leaders are important influences on
people’s choice of yardstick. The issue of which particular socioeconomic
or political indicator individuals are using to compare their group to other
groups (for example, level of income, educational attainment, beneficia-
ries of public investment, ministers in cabinet, judges and so on) is as
important as the yardstick being used by people to form an opinion about
the prevailing inequalities. Given that the observed inequalities may
differ substantially across different indicators, this could have a major
impact on the overall perceptions of the prevailing HIs.
• Misjudgement of group size. Another important factor that may also con-
tribute to a mismatch of objective and subjective HIs is people’s inaccu-
rate views of the relative size of their own group and that of other groups.
If people have to assess whether they get a ‘fair’ share of, for example,
parliamentary seats, ministerial positions, government contracts or gov-
ernment jobs, they usually compare – either explicitly or implicitly – their
group’s share of these positions to their relative demographic size in the
country as a whole. If people believe that their group’s relative demo-
graphic size is considerably larger or smaller than it actually is in reality,
this can substantially distort their perceptions of the prevailing HIs.
• Cross-dimensional ‘contamination’. If people are politically excluded or
marginalised, this may affect or ‘blur’ their perceptions of the pre-
vailing socioeconomic inequalities, and vice versa. Moreover, it is even
possible that misperceptions with respect to the prevailing political/
socioeconomic conditions induce misperceptions with respect to the
prevailing socioeconomic/political inequalities.
To what extent these factors are at play in our five case studies, in which
specific combinations and to what effect are issues that go beyond the scope
of this chapter. However, in cases where we observe a mismatch between
objective and subjective HIs (see Section 9.5), we will examine the impact of
some of these factors in more detail.
9.3 Assessing objective socioeconomic horizontal inequalities
In this section we will analyse the prevailing objective socioeconomic HIs

in our five case-study countries. We will use the nationally representative
Afrobarometer surveys to determine different groups’ socioeconomic status
or standard of living.
In order to determine the prevailing socioeconomic inequalities across dif-

ferent ethnic groups, we have composed two ‘welfare’ indices on the basis of
data available in the Afrobarometer surveys.6 The first index – called ‘BHN’ –
aims to measure the extent to which respondents were able to secure their
basic human needs, including having enough food to eat, having access to
health care and having decent shelter. The second index – called ‘Assets’ –
is an asset wealth index based on whether or not respondents have such
things as a television, a mobile phone or a car. The index is calculated by
adding together the weighted binary scores for each of these assets. For both
indices, higher scores indicate higher standards of living.7 Figure 9.1 shows
the prevailing ethnic inequalities in our five case-study countries according
Nigeria: BHNs (not adjusted) Nigeria: assets (not adjusted)

1.5 1.1
1 1
0.5 0.9
0 0.8
–0.5 0.7
–1 0.6
–1.5 0.5
–2 0.4
Hausa/Fulani
Igbo (n = 358)
Yoruba (n = 491)
Other Nigerians
Hausa/Fulani
Igbo (n = 366)
Yoruba (n = 486)
Other Nigerians
(n = 603)
(n = 814)
(n = 578)
(n = 815)
Ghana: BHNs (not adjusted) Ghana: assets (not adjusted)
1 1.2
0 1
0.8
–1
0.6
–2 0.4
–3 0.2
–4 0
Akan (n = 588)
Ewe (n = 156)
Ga/Dangbe
MoleDagbani
Other
Ghananians
Akan (n = 581)
Ewe (n = 157)
Ga/Dangbe
Mole/Dagbani
Other
Ghananians
(n = 123)
(n = 100)
(n = 123)
(n = 103)
(n = 192)
(n = 193)
Zimbabwe: BHNs (not adjusted) Zimbabwe: assets (not adjusted)

1.4
1.2 1.2
1 1
0.8
0.6 0.8
0.4 0.6
0.2
0 0.4
Shona (n = 902)
Ndebele
Other
Zimbabweans
Shona (n = 897)
Ndebele
Other
Zimbabweans
(n = 147)
(n = 147)
(n = 122)
(n = 121)
Figure 9.1 Ethnic inequalities according to the BHN-index and the Assets-index
Source: Authors’ calculations based on the Afrobarometer R4 surveys.
Kenya: BHNs (not adjusted) Kenya: assets (not adjusted)

2 0.1
1
0 0.8
–1 0.6
–2
–3 0.4
–4
–5 0.2
Kikuyu
(n = 204)
Luo (n = 135)
Luhya
(n = 135)
Somali (n = 95)
Kamba
(n = 115)
Kalenjin
(n = 128)
Kisii (n = 66)
Other
Kenyans…
(n = 206)
Kikuyu
(n = 202)
Luo (n = 129)
Luhya
Somali (n = 93)
Kamba
Kalenjin
Kisii (n = 66)
Other
Kenyans…
(n = 193)
(n = 130)
(n = 109)
(n = 127)
Uganda: BHNs (not adjusted) Uganda: assets (not adjusted)
0 0.9
–1 0.8
0.7
–2 0.6
–3 0.5
–4 0.4
0.3
–5 0.2
Baganda
Banyoro
Banyankole
Acholi
Other
Ugandans
Baganda
(n = 523)
Banyoro
(n = 117)
Banyankole
(n = 273)
Acholi
(n = 179)
Other
Ugandans
(n = 1,289)
(n = 522)
(n = 117)
(n = 276)
(n = 177)
(n = 1,320)
to both indices. The figures depict the linear predictions of the point esti-
mates as well as the 95 per cent confidence intervals that were calculated on
the basis of the Afrobarometer surveys.
As can be seen in each plot, all countries covered here contain consider-
able gaps between the main ethnic groups according to one or both welfare
indices. In Nigeria, for instance, the Hausa/Fulani are poorer than the other
two main ethnic groups (the Yoruba and the Igbo) as well as the com-
bined group of other ethnic minorities; in Ghana, the Ga/Dangbe seem to
be significantly wealthier than other groups regardless of how we measure
living standards; in Zimbabwe, we find a difference between the Ndebele
and other ethnic minorities in terms of household assets; in Kenya, rela-
tionships are more complicated due to the greater number of major ethnic
groups, but the results still indicate that there is a significant gap between
the Kikuyu and the Somali, with rough parity between the remaining ethnic
groups; in Uganda, the Acholi and residual ethnic minorities exhibit con-
sistently lower levels of welfare compared to the three main ethnic groups
(the Buganda, Banyoro and Banyankole). Moreover, the picture that emerges
from the point estimates is much in line with other data and information
that are available on the relative socioeconomic situations of these differ-
ent ethnic groups (see, for example, for Ghana: Gyimah-Boadi and Asante,
2006; Langer, 2008; for Nigeria: Mustapha, 2006; Langer and Ukiwo, 2008;
for Kenya: Kanyinga, 2006: Stewart, 2010a).
It should be noted that the causes and origins of the prevailing socioeco-
nomic inequalities between different ethnic groups and/or regions in most
developing countries, in particular in Africa, are usually related to such fac-

tors as: ecological and climatological differences between different regions in
a country; the geographical distribution of natural resources; the differential
impact of colonialism, which Figueroa (2006) labels as a ‘foundational shock’
from which the initial inequalities between different ethnic groups and/or
regions usually originate; the extent of group discrimination and favouritism
towards particular groups by the government; and the differential impact of
economic policies on different groups and/or regions. Once HIs are in place
they tend to endure for very long periods of time, as illustrated by black–
white differentials in the United States or indigenous–Ladino differentials
in Latin America, which have been in existence for centuries (Stewart and
Langer, 2008).8 Moreover, quite often, HIs appear to persist not because of
conscious decisions by political actors, nor because of an unequal distribu-
tion of power, nor due to explicit discrimination and exclusionary policies
towards particular groups (as, for instance, in South Africa under apartheid),
but because they are the outcome of more ‘intangible’ economic forces and
mechanisms (Brown and Langer, 2010).
9.4 Explaining the observed objective horizontal inequalities
The ethnic inequalities observed in the previous section do not take into
account differences in factors that are conventionally used to explain dif-
ferences in standard of living across individuals. If we assume that the
factors conventionally associated with lower levels of individual socioeco-
nomic welfare are unrelated to ethnicity, we should be able to explain a
large proportion of the observed ethnic inequalities by means of these indi-
vidual ‘risk’ factors. In order to test this, we aim to explain as much as
possible the variances of (a) assets and (b) BHNs with the help of several
commonly used risk factors and then see if the included ethnic dummies
are significantly different from zero or not. The risk factors we include in
our regression analysis are: (1) level of educational attainment; (2) people’s
employment status; (3) level of infrastructural development of people’s liv-
ing environment; and (4) people’s experience of physical insecurity due to
violence and crime. Also incorporated are individual demographic variables,
such as sex, age, region of birth (whether a respondent is indigenous to a par-
ticular place of living) and ethnicity as well as a range of context variables,
such as the urban/rural distinction, and a series of dummies for each of the
administrative distinctions made in a country (province, region or state).
It is important to note here that not all variables are available in both
survey datasets. Thus, while Afrobarometer lacks information on people’s
region of birth, an infrastructure variable is not included in our survey.
Also, variables for insecurity and employment are differently measured (see
Appendix 1 for more details on these issues). While we can use an OLS regres-
sion for our survey, this is not appropriate for the Afrobarometer surveys
because of the different levels on which the infrastructural development

variable is measured. For the Afrobarometer surveys, we therefore employ
a mixed-effect regression analysis instead, which allows the intercept to vary
according to the primary sampling unit on which the infrastructure devel-
opment variable is rated.9 Despite these subtle nuances, we assume that the
results from both estimation methods are largely comparable. The results are
reported in Table 9.2.
First, it is interesting to note that lower levels of educational attainment
have, as expected, a largely negative impact on individuals’ standard of liv-
ing regardless of country, the proxy, dataset and estimation method. The
same appears to be true for the effects of perceived insecurity: people who
tend to be threatened or victimised are more likely to be relatively poor com-
pared those who live in more secure situations. Interestingly, the expected
positive impact on people’s standard of living of ‘being employed’ is only
confirmed in the Afrobarometer surveys, but not in our surveys. A possible
explanation for this finding could be that in our surveys there may be a
relative overrepresentation of students due to the fact that most interviews
were conducted in the largest city in the country (usually the capital, except
in Nigeria). Moreover, students are generally relatively well-off despite being
unemployed. Finally, infrastructure development also has the expected pos-
itive influence upon individuals’ standards of living, except in the case of
Ghana. Thus, the more developed an area is in terms of infrastructure, the
richer the residents who live in that area tend to be.
Having largely confirmed the expected effects of the most important indi-
vidual ‘risk’ factors, we can now examine the remaining effect of ethnic
affiliation by analysing the ethnic group dummies. In Nigeria, the Hausa/
Fulani are set as reference category. Interestingly, while the ethnic dum-
mies have no significance in our surveys, in the Afrobarometer surveys,
they do. In particular, the Igbo dummy is consistently larger than zero,
which suggests that the Igbo are inherently richer than the Hausa/Fulani.
In Ghana, where the Akan are used as the baseline group, we find that the
Mole/Dagbani dummy shows a significantly negative impact on ‘Assets’ if
we use Afrobarometer sample, while the Ga/Dangbe dummy has a signifi-
cantly positive effect on BHNs when we use the JICA sample. The Ewe show
inconsistent results in terms of BHNs. In Zimbabwe, the margin that the
Ndebele had against the Shona (that is, the baseline group) in the preced-
ing bivariate analysis now disappears completely, while the negative effect
of belonging to a residual ethnic minority persists. In Kenya, the observed
ethnic inequalities between the Kikuyu (the baseline group) and the Somali
cease to exist except for the model that uses Assets as a dependent variable
and the JICA survey as a sample, while the lower status of the Luhya newly
emerges. Finally, in Uganda, we find no remaining ethnic differences in our
surveys, while the residual ethnic minorities’ disadvantage relative to the
Banyankole (the baseline group) remains in the Afrobarometer survey.
Table 9.2 Determinants of living standards
JICA Afrobarometer JICA Afrobarometer

OLS mixed effect OLS mixed effect
assets p-value assets p-value BHNs p-value BNHs p-value
Nigeria
Female 0.006 0.913 −0.121 0.000 0.577 0.146 0.023 0.890

(0.054) (0.026) (0.396) (0.164)
Age 0.003 0.249 0.002 0.099 −0.013 0.432 0.001 0.866
(0.002) (0.001) (0.017) (0.008)
Indigene 0.007 0.906 0.678 0.104
(0.057) (0.416)
Lower Education −0.201 0.022 −0.310 0.000 −0.752 0.243 −0.128 0.558
(0.088) (0.035) (0.643) (0.218)
Insecurity −0.167 0.003 0.001 0.820 −2.188 0.000 −0.540 0.000
(0.056) (0.006) (0.407) (0.040)
Job (any) −0.090 0.161 −0.567 0.220
(0.064) (0.462)
Part-time job 0.108 0.001 −0.268 0.205
(0.034) (0.212)
Full-time job 0.237 0.000 0.088 0.682
(0.034) (0.214)
Rural −0.145 0.004 0.008 0.982
(0.051) (0.334)
Small Urban −0.122 0.045 −0.038 0.924
(0.061) (0.402)
Infrastructure 0.079 0.002 0.491 0.004
(0.026) (0.169)
217
218

Yoruba 0.130 0.104 0.142 0.102 0.073 0.900 0.965 0.079

(0.079) (0.087) (0.579) (0.550)
Igbo 0.048 0.592 0.174 0.052 −0.178 0.784 1.210 0.034
(0.090) (0.089) (0.650) (0.571)
Other Nigerians −0.027 0.777 0.095 0.135 −0.500 0.477 0.782 0.050
(0.096) (0.063) (0.701) (0.400)
Intercept 0.894 0.000 0.819 0.000 1.729 0.020 −1.497 0.037
(0.102) (0.110) (0.743) (0.718)
sd cons 0.193 1.381
(0.021) (0.129)
sd(Residual) 0.546 3.436
(0.009) (0.059)
Number of observations 399 1935 395 1950
Number of groups 236 236
F 2.560 0.007 5.480 0.000
Adjusted R2 0.034 0.0928
LR chi2
Wald chi2 552.960 0.000 398.32 0.000
Restricted-LR −1709.556 −5232.341
Ghana
Female −0.093 0.280 −0.182 0.000 0.332 0.384 0.113 0.695

(0.086) (0.039) (0.381) (0.287)
Age 0.004 0.170 −0.002 0.131 0.016 0.237 −0.007 0.441
(0.003) (0.001) (0.014) (0.009)
Indigene 0.012 0.958 −1.436 0.174
(0.238) (1.053)
Lower Education −0.368 0.000 −0.435 0.000 −0.705 0.092 −0.802 0.018
(0.094) (0.046) (0.417) (0.338)
Insecurity 0.012 0.901 0.026 0.005 −1.288 0.002 −0.395 0.000
(0.093) (0.009) (0.411) (0.070)
Job (any) −0.147 0.249 −0.673 0.233
(0.127) (0.563)
Part-time job 0.126 0.018 −0.490 0.209
(0.053) (0.390)
Full-time job 0.231 0.000 0.606 0.066
(0.044) (0.329)
Rural −0.243 0.000 −0.423 0.374
(0.065) (0.476)
Infrastructure 0.031 0.170 0.248 0.141
(0.023) (0.169)
Ewe −0.040 0.768 −0.101 0.208 1.083 0.072 −1.213 0.042
(0.136) (0.080) (0.601) (0.596)
Ga/Dangbe −0.058 0.808 0.074 0.357 2.182 0.041 −0.224 0.706
(0.240) (0.081) (1.062) (0.594)
Mole/Dagbani −0.100 0.571 −0.302 0.003 0.682 0.383 0.338 0.660
(0.176) (0.103) (0.781) (0.768)
Other Ghananians 0.078 0.538 −0.039 0.608 0.886 0.113 −0.831 0.143
(0.126) (0.076) (0.558) (0.567)
Intercept 1.193 0.000 1.526 0.000 −0.047 0.946 0.657 0.498
(0.156) (0.134) (0.693) (0.970)
219
220

sd cons 0.148 1.034

(0.029) (0.239)
(0.014) (0.103)
F 2.250 0.015 2.120 0.023
Adjusted R2 0.0373 0.0336
LR chi2
Wald chi2 346.310 0.000 184.430 0.000
Restricted-LR −991.1947 −3010.196
Zimbabwe
Female −0.072 0.382 −0.138 0.001 −0.035 0.918 −0.131 0.429

(0.083) (0.042) (0.338) (0.165)
Age 0.009 0.016 0.003 0.027 −0.016 0.284 −0.006 0.339
(0.004) (0.002) (0.015) (0.006)
Indigene 0.070 0.419 0.672 0.059
(0.086) (0.355)
Lower Education −0.457 0.000 −0.392 0.000 −0.665 0.201 −0.594 0.003
(0.127) (0.053) (0.519) (0.203)
Insecurity −0.103 0.203 0.001 0.919 −1.869 0.000 −0.241 0.000
(0.081) (0.010) (0.331) (0.039)
Job (any) −0.102 0.280 −0.298 0.442
(0.094) (0.387)
Part-time job −0.016 0.817 0.330 0.226
(0.071) (0.273)
Full-time job 0.378 0.000 0.208 0.402
(0.065) (0.249)
Rural −0.451 0.000 0.951 0.023
(0.121) (0.418)
Infrastructure 0.108 0.007 −0.171 0.218
(0.040) (0.139)
Ndebele 0.233 0.109 −0.043 0.630 −0.506 0.396 0.385 0.260
(0.145) (0.089) (0.596) (0.341)
Other Zimbabweans −0.046 0.595 −0.145 0.077 −0.987 0.006 0.199 0.524
(0.087) (0.082) (0.356) (0.312)
Intercept 1.262 0.000 0.937 0.000 2.280 0.000 0.879 0.140
(0.154) (0.184) (0.625) (0.595)
sd cons 0.119
(0.042)
sd(Residual) 0.698
(0.015)
Number of groups 65
F 3.900 0.000 7.060 0.000 7.140 0.000
Adjusted R2 0.0621 0.1222 0.0927
LR chi2
Wald chi2 432.030 0.000
Restricted-LR −1256.31
221
222

Kenya
Female −0.049 0.256 −0.084 0.008 −0.153 0.606 0.078 0.699

(0.043) (0.032) (0.297) (0.202)
Age 0.009 0.000 0.002 0.126 −0.021 0.112 0.002 0.847
(0.002) (0.001) (0.013) (0.008)
Indigene 0.023 0.645 −0.413 0.224
(0.049) (0.339)
Lower Education −0.421 0.000 −0.376 0.000 −2.809 0.000 −0.667 0.004
(0.053) (0.036) (0.365) (0.229)
Insecurity −0.004 0.931 0.005 0.488 −1.939 0.000 −0.276 0.000
(0.050) (0.008) (0.347) (0.049)
Job (any) −0.025 0.586 0.261 0.420
(0.047) (0.324)
Part-time job 0.077 0.071 0.111 0.679
(0.042) (0.269)
Full-time job 0.338 0.000 1.061 0.000
(0.044) (0.283)
Rural −0.082 0.369 0.438 0.484
(0.092) (0.626)
Infrastructure 0.109 0.000 0.321 0.065
(0.026) (0.174)
Luo −0.037 0.591 −0.114 0.244 −0.832 0.081 −0.642 0.316
(0.069) (0.098) (0.477) (0.640)
Luhya −0.007 0.926 −0.222 0.021 −1.519 0.003 −1.010 0.102
(0.072) (0.097) (0.502) (0.617)
Somali 0.181 0.019 0.235 0.153 −0.518 0.333 −1.411 0.185
(0.077) (0.164) (0.535) (1.065)
Kamba −0.082 0.284 −0.120 0.242 −0.609 0.249 −1.641 0.013
(0.076) (0.103) (0.528) (0.660)
Kalenjin 0.187 0.034 −0.053 0.546 −0.246 0.685 −0.463 0.424
(0.088) (0.088) (0.605) (0.579)
Kisii 0.065 0.501 −0.142 0.208 0.383 0.567 −0.409 0.581
(0.096) (0.113) (0.669) (0.741)
Other Kenyans 0.186 0.017 −0.173 0.035 0.308 0.568 −0.710 0.177
(0.077) (0.082) (0.539) (0.526)
Intercept 0.866 0.000 0.939 0.000 1.844 0.004 −0.825 0.391
(0.091) (0.144) (0.634) (0.961)
sd cons 0.179 1.305
(0.024) (0.157)
(0.012) (0.075)
F 10.460 0.000 11.710 0.000
Adjusted R2 0.1373 0.1542
LR chi2
Wald chi2 468.070 0.000 151.410 0.000
Restricted-LR −822.1003 −2798.129
223
224

Uganda
Female −0.171 0.044 −0.082 0.000 −0.399 0.245 0.108 0.477

(0.085) (0.023) (0.343) (0.152)
Age 0.002 0.593 −0.001 0.138 −0.056 0.000 −0.029 0.000
(0.004) (0.001) (0.015) (0.007)
Indigene −0.100 0.350 0.364 0.403
(0.107) (0.435)
Lower Education −0.440 0.000 −0.327 0.000 −1.930 0.000 −0.524 0.002
(0.091) (0.025) (0.368) (0.166)
Insecurity −0.102 0.245 0.013 0.013 −1.805 0.000 −0.347 0.000
(0.088) (0.005) (0.352) (0.035)
Job (any) 0.084 0.351 0.149 0.682
(0.090) (0.363)
Part-time job 0.098 0.001 0.970 0.000
(0.028) (0.187)
Full-time job 0.409 0.000 1.391 0.000
(0.036) (0.235)
Rural −0.154 0.004 −0.460 0.226
(0.054) (0.379)
Infrastructure 0.079 0.000 0.174 0.055
(0.013) (0.091)
Baganda 0.255 0.103 −0.102 0.104 −0.167 0.790 −1.088 0.009
(0.156) (0.063) (0.627) (0.418)
Banyoro 0.243 0.222 −0.060 0.390 −0.032 0.969 −1.265 0.007
(0.199) (0.069) (0.802) (0.471)
Acholi −0.034 0.896 −0.146 0.058 −1.353 0.200 −0.293 0.580
(0.263) (0.077) (1.055) (0.529)
Other Ugandans −0.014 0.917 −0.168 0.001 −0.523 0.346 −1.052 0.002
(0.139) (0.050) (0.554) (0.341)
Intercept 1.116 0.000 0.961 0.000 3.243 0.000 0.490 0.458
(0.178) (0.096) (0.718) (0.659)
sd cons 0.145 1.137
(0.018) (0.109)
(0.009) (0.056)
F 4.750 0.000 15.350 0.000
Adjusted R2 0.0935 0.2858
LR chi2
Wald chi2 799.300 0.000 431.670 0.000
Restricted-LR −2017.53 −6513.31
Note: Estimation results for region dummies are not shown.

Source: Authors’ calculations based on the Afrobarometer R4 and JICA surveys.
225
Regardless of whether these statistically significant effects of ethnic dum-

mies represent real ethnic idiosyncrasies or spurious correlations caused by
omitted variables, just because a significant proportion of the variance with
respect to people’s living standards can be accounted for by a range of indi-
vidual ‘risk’ factors does not mean that these risks are themselves unrelated
to ethnicity (as we have assumed in this section). Indeed, there are good rea-
sons to assume that this is actually not the case, which is the issue we will
turn to in the next section.
9.5 Ethnic differences in risk factors associated

with individual socioeconomic development
In this section we will analyse whether the risk factors associated with indi-
vidual standards of living are distributed equally among different ethnic
groups. If the probability of being exposed to these risks does not differ from
ethnic group to ethnic group, only then can these factors be considered to
contribute to individual well-beings independently of ethnic affiliations. If,
however, the probability of facing one of these risk factors varies across eth-
nic groups, the root cause of inequality is ethnicity, whose effect is mediated
through these factors.
It should be further noted that the risk factors that were analysed in the
previous section are related to each other. For example, someone’s risk of
living in a rural area depends on the region in which he/she lives, because
the proportions of rural areas differ from region to region. Moreover, the
region of residence is itself a risk factor that could affect individuals’ stan-
dard of living due to a region’s climatological and ecological characteristics.
Likewise, the risk of facing infrastructural underdevelopment depends on,
among other things, the place of residence (that is, in an urban or rural
area) as well as the region of residence. Risk of quitting school after primary
education depends on factors such as the extent to which infrastructure is
available (such as secondary schools), access to school, degree of urbani-
sation and region of residence, as well as on gender and generation. Risk
of being threatened or actually victimised by violent crime is determined
largely by similar factors. Finally, risk of being unemployed depends on edu-
cational attainment, gender, age and infrastructure. Therefore, in assessing
ethnic gaps in risk factors, we need to control for these ‘interdependencies’
accordingly.
Let us start by examining whether the infrastructural development people
enjoy systematically differs according to the group to which they belong. Of
course, infrastructure is supposed to be ‘public’, meaning everyone, at least
among the citizens of a country, should be able to use it without being dis-
criminated against. In some cases such as paved roads, even non-citizens can
benefit from infrastructural development. Indeed, one of the defining char-
acteristics of public goods is their non-excludability. However, if different
groups are segregated from each other even within the same region and if
they have very low levels of mobility, inherently ‘public’ infrastructure can
become ‘private’ (or a ‘club’ good). The individuals who benefit from the
infrastructural development of a particular area tend to be members of a
particular ethnic group because the group tend to live in that area. In such
a context, infrastructure development site or its priority might be contested
among ethnic groups.
In order to detect possible group differences resulting from ethnic com-
petition or discrimination in infrastructural development, we regress our
infrastructural development index on ethnic affiliation, controlling for
regions of residence as well as the urban/rural distinction. Figure 9.2 presents
the 95 per cent confidence intervals of the average scores of infrastructure
development for each ethnic group. In Nigeria and Zimbabwe, no signifi-
cant ethnic gap is found in terms of infrastructure development, although
average scores for the Yoruba and the Ndebele are higher than the respec-
tive reference groups. In Ghana, however, the Ga/Dangbe have statistically
significant advantages in terms of infrastructure development over the Ewe
and other residual ethnic minorities. In Kenya, too, we find significant eth-
nic gaps: groups with relative advantages are the Luhya, the Kalenjin and
the Kisii, while the Luo and the Kamba face higher risks of infrastructure
underdevelopment. Finally, in Uganda, we find a significant gap between
the Baganda and other residual ethnic minorities on the one hand and the
Banyoro on the other hand. It should be noted that the Banyankole, like
the Kikuyu in Kenya, do not necessarily benefit most from infrastructural
development, although they do not suffer most either.
Let us now turn to the ethnic gaps that might exist in education, espe-
cially in terms of the risk that one quits schooling early. Because educational
attainment usually varies across generations and gender as well as environ-
mental factors like the availability of infrastructure just examined, we need
to control for these variables when estimating ethnic influences on the risk
of ‘quitting’ education before or at completion of primary school. Therefore,
we applied a mixed-effect logit model, assigning 1 to those whose highest
level of educational achievement is at most completion of primary educa-
tion and 0 otherwise. We present the adjusted predicted probabilities (based
on fixed parts) of lower educational attainment for each ethnic group (that
is, quitting school before or at completion of primary schooling) in our case-
study countries for the case of a 35-year-old male living in an urban area
with an average infrastructure in the capital city (Figure 9.3).
The results show that in Nigeria, the Hausa/Fulani face a significantly
higher risk of having lower educational attainment compared to the Igbo
and the Yoruba, while the educational difference with the residual ethnic
minorities is not statistically significant. In Ghana, the Mole/Dagbani face
a higher risk of not having post-primary schooling when compared to the
Ewe and the Ga/Dangbe. However, these differences do not have statistical
Nigeria Ghana
1.6 3.2
1.5 3
1.4 2.8
1.3 2.6
2.4
1.2
2.2
1.1
2
1 1.8
Hausa/ Igbo Yoruba Other
an
ni
an r
ni he
Ew
gb
ba
Fulani Nigerians
Ak
s
na Ot
an
ag
D
D
a/
e/
G
ol
ha
M
G
Zimbabwe Kenya
4.1 5
4.8
4 4.6
4.4
3.9 4.2
4
3.8 3.8
3.7 3.6
3.4
3.6 3.2
3
3.5 2.8
Shona Ndebele Other
in
Ka ba
So a
yu
Lu o
ny er
Ka ali
i
hy
si
s
Lu
nj
Ke Oth
ku
an
m
Zimbabweans
Ki
le
Ki
Uganda
4.2
4
3.8
3.6
3.4
3.2
3
a
le
li
ho
nd
or
nd er
ko
s
an
ny
ga th
Ac
ga
an
O
Ba
Ba
ny
Ba
Figure 9.2 Group averages of benefit from infrastructural development

Note: Point estimates and 95 per cent CIs are plotted. Predictions are adjusted to the situation in
urban areas in the respective capital cities (in the case of Nigeria, in Lagos).
Source: Authors’ calculation based on Afrobarometer data.
significance. In Zimbabwe, belonging to a residual ethnic minority increases

the probability of lower educational attainment compared to the Shona.
However, the difference is within the margin of error in terms of predicted
probabilities. In Kenya, the Somali stand out for their higher risk of having
Nigeria Ghana
35% 60%
30%
50%
25%
40%
20%
15% 30%
10% 20%
5% 10%
0% 0%
Hausa/ Igbo Yoruba Other
an
ni
an r
ni he
Ew
gb
ba
Ak
s
na Ot
Fulani Nigerians
an
ag
D
D
a/
e/
G
ol
ha
M
G
Zimbabwe Kenya
60% 100%
50%
80%
40%
60%
30%
20% 40%
10% 20%
0% 0%
Shona Ndebele Other
in
So a
Ka ba
yu
Lu o
ny er
Ka li
i
si
hy
Lu
nj
s
a
ku
Ke Oth
m
an
Ki
m
Zimbabweans
le
Ki
Uganda
40%
35%
30%
25%
20%
15%
10%
5%
0%
da
le
ro
nd er
li
s
o
ho
an
n
ga h
nk
ga
ny
U Ot
Ac
ya
Ba
Ba
n
Ba
Figure 9.3 Group averages of probability to have lower educational attainment

Note: Point estimates and 95 per cent CIs are plotted. Predictions are adjusted to the case of a
35-year-old male living in an urban area with average infrastructure in the respective capital cities
(in the case of Nigeria, in Lagos).
Source: Authors’ calculation based on Afrobarometer data.
lower levels of education. Finally, no statistical differences exist between dif-

ferent ethnic groups in Uganda, but the control variables exert significant
influence on the probability: being female, getting older and living in rural
areas increase the risk of having lower educational attainment; conversely,
living in areas with higher levels of infrastructural development reduces the
risk of not having post-primary education. The last point largely applies to
the other countries as well (statistical table not reported).
Due to space constraint, we report here only results of group com-

parisons in terms of education and infrastructure. Yet other risk factors
associated with lower individual standards of living also indicated system-
atic differences (result not reported). In particular, the Hausa/Fulani face
a disproportionately higher risk of having lower educational attainment
and of unemployment. In the other countries as well, exposure to these
two risk factors appears to differ substantially across different ethnic groups
although most of which did not reach the statistical significance level. Hav-
ing established the presence of objective HIs, namely collective gaps in living
standards deeply rooted in ethnic difference, we can now turn to the ques-
tion whether these inequalities are indeed ‘correctly’ perceived by the people
involved.
9.6 Objective versus subjective horizontal inequalities
In order to determine how people perceived the prevailing socioeconomic

HIs, we included the following question in our survey.
Nigeria (Afrobarometer R4) Nigeria (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Hausa/Fulani Hausa/Fulani
Igbo Igbo
Yoruba Yoruba
Other Other
Nigerians Nigerians
Ghana (Afrobarometer R4) Ghana (JICA surveys)
0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Akan Akan
Ewe Ewe
Ga/Dangbe Ga/Dangbe
Mole/Dagbani Mole/Dagbani
Other Other
Ghananians Ghananians
Zimbabwe (Afrobarometer R4) Zimbabwe (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Shona Shona
Ndebele
Ndebele
Other
Zimbabweans Other
Zimbabweans
Figure 9.4 Perceived socioeconomic horizontal inequalities

Kenya (Afrobarometer R4) Kenya (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Kikuyu Kikuyu
Luo Luo
Luhya Luhya
Somali Somali
Kamba Kamba
Kalenjin Kalenjin
Kisii Kisii
Other Other
Kenyans Kenyans
Uganda (Afrobarometer R4) Uganda (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Baganda Baganda
Banyankole Banyankole
Banyoro Banyoro
Acholi Acholi
Other Other
Ugandans Ugandans
Superior Same Inferior
Think about the condition of your ethnic group. Are their economic
conditions worse, the same as, or better than other groups in this
country?
The Afrobarometer surveys included the same question. People were asked
to respond to this question on the following five-point ordinal scale: 1,
much better; 2, better; 3, same; 4, worse; and 5, much worse. We sub-
sequently consolidated people’s answers into a three-point ordinal scale:
‘superior’ (much better/better); ‘same’ (same); and ‘inferior’ (worse/much
worse). Figure 9.4 depicts the distributions of responses according to our
three-point ordinal scale for each case-study country, where the results based
on the Afrobarometer surveys are displayed on the left-hand side and those
based on our surveys on the right-hand side.
It is interesting to see that the distribution patterns of both surveys are
quite similar despite the underlying differences in the scope and timing of
both surveys. For example, in both surveys, it emerges that the most frus-
trated groups in each country are as follows: the Igbo in Nigeria, the Ewe
in Ghana, the Ndebele in Zimbabwe, the Somali in Kenya and the Acholi
in Uganda. Except for Nigeria, there also appears to be consistency between
both surveys regarding the most ‘satisfied’ ethnic groups: the Mole/Dagbani
in Ghana, the Shona in Zimbabwe, the Kikuyu and the Kisii in Kenya and
the Banyankole in Uganda.
Comparing the results of Figure 9.4 with our previous analysis, it appears
that there are some notable discrepancies between the perceived and
objective HIs in our case-study countries. Thus, for instance, in Nigeria,
while the Hausa/Fulani were ‘objectively’ the poorest ethnic group, it was
among the Igbo respondents that the highest proportion of people felt
that they had an inferior level of socioeconomic development. In Ghana,
it was the Mole/Dagbani group that was objectively most disadvantaged, yet
the Ewe respondents were the ones who held this perception most widely.
In Zimbabwe, while there was no sharp objective divide between the Shona
and the Ndebele, the latter group clearly perceived themselves to be seriously
disadvantaged. In Uganda, while the Baganda and the Banyoro were at least
objectively at the same level as the Banyankole, they do not seem to perceive
things that way. Moreover, some ethnic groups had a better than expected
view of their own socioeconomic situation compared to other ethnic groups.
Thus, for instance, the confidence in their superior status displayed by the
Kisii respondents in Kenya was largely ungrounded on the basis of the objec-
tive analysis. For some ethnic groups, the perceptions of HIs appear to offer
a relatively good reflection of the actual situation. Thus, for instance, the
Somali people in Kenya and the Acholi people in Uganda perceived them-
selves correctly to be at a disadvantage compared to other ethnic groups.
Similarly, the Kikuyu in Kenya and the Banyankole in Uganda correctly
perceive themselves to be in a relatively advantaged position.
Returning to the factors that could induce a mismatch between objec-
tive and subjective HIs discussed in Section 9.2, we will now examine the
extent to which people’s individual socioeconomic situation influenced or
blurred people’s perceptions of the prevailing objective HIs. In order to do
this, we have to control for the effects of individual attributes as well as
environmental factors affecting respondents’ lives and see if the adjusted
distributions of perceived HIs converge to the expected distributions based
on the objective HIs. It is worthwhile recalling here that if perceptions reflect
objective group differences, people from the same ethnic group should
respond with the same answer regardless of their personal socioeconomic
status. Thus, two respondents from the same ethnic group – one being
rich and the other being poor – should choose ‘inferior’, if their group is
relatively disadvantaged compared to other ethnic groups. Similarly, two
equally wealthy respondents, with one of them belonging to a relatively rich
group and the other belonging to a relatively poor group, should choose
different answers regardless of the fact that they are in the same personal
socioeconomic situation. Complicating matters in this respect is the fact
that if there is an objective difference between ethnic groups, it is possible
that individual wealth correlates with perceiving one’s group to be supe-
rior in terms of socioeconomic development, as richer individuals are more
likely to belong to the relatively richer group, and vice versa. Yet once we
include ethnic dummy variables in our model, the variables that measure
respondents’ individual wealth situation should no longer have explanatory
power, because the coefficients are adjusted to the situation where the
respondent’s ethnic group is the same.
Infrastructural development can also affect respondents’ perceptions of
the prevailing socioeconomic HIs. This is because it may have a direct
demonstration effect via its visual impact upon perceptions; but it can also
influence perceptions indirectly through the improvement of objective indi-
vidual living standards, which therefore also have to be controlled for by
individual wealth indicators. Therefore, in estimating the effects of ethnic
affiliation on respondent’s choice from our three possibilities (‘inferior’, ‘the
same’ and ‘superior’), it is necessary to control for the effects of differences
in infrastructural development as well as individual socioeconomic charac-
teristics, including education and employment. We also control for group
size by means of a dummy variable, which captures whether or not groups
are larger than 10 per cent of the total population. The logic for including
this variable is that larger groups are in a better position to ‘claim’ socioe-
conomic benefits and resources from the state than smaller groups, thereby
possibly improving their group’s position, which might be reflected in their
perceptions of the prevailing HIs.
Given that there is no mixed effect multinomial logit model, we decided
to divide the three answering options (‘inferior’, ‘the same’ and ‘superior’)
into two sets of binary variables, and used mixed-effect logit instead. When
estimating the probability of feeling inferiority, the dependent variable takes
the value 1 if a respondent chooses ‘inferior’ and 0 if the answer is ‘the same’,
excluding those who choose ‘superior’ from the analysis. When estimating
the probability of perceiving group ‘superiority’, respondents who choose
‘inferior’ are excluded from the analysis. Respondents who chose ‘superior’
are assigned the value 1, while respondents who answered ‘the same’ were
assigned 0. The results of this analysis are shown in Table 9.3, where each
case-study country has four models. The first column from the left lists the
effects of the above variables on the probability that respondents perceived
their own group to be inferior to other groups. Conversely, the third col-
umn lists the effects on the probability that respondents perceived their own
group to be superior to other groups.
As expected, most variables measuring individual socioeconomic status
have no significant effect. The exception is the BHN indices. In four of our
five case studies, it emerges that the respondents who are better able to sat-
isfy their basic human needs are also less likely to perceive their own group
to be inferior to other groups. This suggests that people’s individual socioe-
conomic situation to some extent affects their assessment of the prevailing
socioeconomic HIs, which was argued to be a possible factor for the exis-
tence of a mismatch between objective and subjective HIs in Section 9.2.
Conversely, there is little evidence to support the idea of a ‘visual’ effect
of infrastructural development. Although respondents in areas with more
infrastructural development are less likely to think their group is inferior
to other groups in Uganda, unexpectedly, the opposite effect is found in
Table 9.3 Determinants of perceived socioeconomic HIs
234
Eonomically inferior (vs. same) Economically superior (vs. same)
coefficient (SE) p-value coefficient (SE) p-value coefficient (SE) p-value coefficient (SE) p-value
Nigeria
Politically inferior 2.176 0.000
(0.162)
Politically superior 1.767 0.000
(0.142)
Assets 0.132 0.234 0.079 0.514 0.278 0.005 0.229 0.029
(0.111) (0.120) (0.100) (0.105)
BHNs −0.048 0.005 −0.041 0.027 0.032 0.035 0.018 0.258
(0.017) (0.018) (0.015) (0.016)
Lower education 0.005 0.973 −0.076 0.664 −0.007 0.966 0.022 0.888
(0.163) (0.176) (0.152) (0.160)
Part-time 0.140 0.389 0.222 0.205 −0.263 0.083 −0.290 0.069
(0.163) (0.175) (0.152) (0.160)
Full-time 0.357 0.028 0.363 0.039 −0.200 0.187 −0.150 0.351
(0.162) (0.176) (0.151) (0.161)
Rural 0.140 0.460 0.229 0.229 −0.159 0.335 −0.282 0.090
(0.190) (0.191) (0.165) (0.166)
Infrastructure 0.016 0.868 0.117 0.240 0.030 0.723 −0.024 0.777
(0.099) (0.100) (0.084) (0.085)
Groupsize10%+ −0.474 0.220 −0.226 0.570 0.181 0.589 −0.305 0.385
(0.387) (0.397) (0.335) (0.351)
Yoruba −0.114 0.653 −0.212 0.407 −0.298 0.148 0.157 0.463
(0.254) (0.256) (0.206) (0.214)
Igbo 0.509 0.059 0.029 0.918 −0.538 0.027 0.108 0.668
(0.269) (0.277) (0.242) (0.252)
Other Nigerians −0.228 0.556 −0.289 0.466 0.082 0.808 0.139 0.691
(0.387) (0.396) (0.339) (0.351)
Intercept −0.556 0.213 −1.487 0.001 −0.462 0.237 −0.878 0.030
(0.447) (0.461) (0.391) (0.404)
sd(_cons) 0.893 0.769 0.579 0.475
(0.121) (0.132) (0.104) (0.117)
Number of observation 1305 1286 1336 1319
Number of psu 235 234 231 231
Wald ch2 24.200 0.012 198.560 0.000 22.950 0.018 168.810 0.000
Log likelihood −841.692 −717.444 −885.569 −788.459
Ghana

(0.260)
(0.209)
Assets −0.182 0.261 −0.149 0.400 −0.129 0.313 −0.145 0.309
(0.162) (0.177) (0.128) (0.143)
BHNs −0.087 0.000 −0.078 0.002 −0.018 0.338 −0.014 0.500
(0.023) (0.025) (0.019) (0.021)
Lower education 0.174 0.485 0.426 0.112 −0.129 0.507 −0.115 0.588
(0.249) (0.268) (0.195) (0.213)
Part-time 0.157 0.569 0.121 0.689 0.199 0.405 0.332 0.202
(0.276) (0.303) (0.239) (0.260)
Full-time 0.064 0.781 0.123 0.628 0.021 0.913 0.007 0.972
(0.229) (0.253) (0.196) (0.217)
Rural −0.346 0.353 −0.487 0.215 −0.135 0.599 −0.089 0.747
(0.372) (0.393) (0.257) (0.275)
Infrastructure −0.124 0.332 −0.110 0.412 0.151 0.095 0.089 0.359
(0.128) (0.135) (0.090) (0.097)
235
236
Groupsize10%+ −0.873 0.193 −0.162 0.825 0.889 0.251 0.370 0.639

(0.670) (0.729) (0.774) (0.788)
Ewe 1.239 0.001 0.519 0.186 0.187 0.555 0.318 0.355
(0.363) (0.393) (0.317) (0.344)
Ga/Dangbe 0.704 0.066 0.082 0.845 −0.158 0.605 −0.034 0.920
(0.382) (0.422) (0.306) (0.341)
Mole/Dagbani −0.711 0.345 −0.372 0.648 1.221 0.134 0.857 0.302
(0.754) (0.814) (0.815) (0.831)
Other Ghananians −0.195 0.783 −0.289 0.702 0.460 0.567 0.252 0.758
(0.706) (0.757) (0.803) (0.820)
Intercept 1.006 0.231 −0.298 0.744 −0.818 0.347 −0.978 0.275
(0.840) (0.911) (0.870) (0.896)
sd(_cons) 1.068 0.979 0.401 0.357
(0.189) (0.205) (0.204) (0.277)
Number of psu 97 96 97 97
Wald ch2 39.110 0.000 103.380 0.000 15.250 0.228 82.530 0.000
Log likelihood −380.51 −317.885 −422.566 −360.552
Zimbabwe

(0.226)
(0.186)
Assets 0.202 0.135 0.187 0.202 −0.100 0.348 −0.179 0.118
(0.135) (0.147) (0.106) (0.115)
BHNs −0.232 0.000 −0.217 0.000 −0.058 0.029 −0.046 0.104
(0.040) (0.043) (0.027) (0.029)
Lower education 0.106 0.606 0.015 0.946 0.105 0.558 −0.123 0.524
(0.205) (0.224) (0.179) (0.193)
Part-time 0.171 0.587 0.198 0.558 0.201 0.402 −0.018 0.945
(0.316) (0.337) (0.240) (0.260)
Full-time −0.312 0.283 −0.168 0.592 −0.530 0.029 −0.644 0.013
(0.291) (0.314) (0.242) (0.260)
Rural 0.580 0.276 0.622 0.252 −0.780 0.076 −0.787 0.070
(0.532) (0.542) (0.440) (0.434)
Infrastructure −0.225 0.207 −0.196 0.283 −0.220 0.145 −0.256 0.089
(0.178) (0.182) (0.151) (0.151)
Groupsize10%+ −1.554 0.000 −1.398 0.000 −0.003 0.989 −0.227 0.359
(0.264) (0.279) (0.234) (0.247)
Ndebele 1.965 0.000 1.160 0.000 −0.422 0.162 0.082 0.792
(0.312) (0.330) (0.302) (0.310)
Other Zimbabweans −0.598 0.095 −1.174 0.003 −0.731 0.040 −0.456 0.217
(0.358) (0.392) (0.356) (0.369)
Intercept −0.215 0.765 −0.870 0.240 0.597 0.315 0.513 0.384
(0.720) (0.740) (0.595) (0.589)
sd(_cons) 0.696 0.559 0.591 0.464
(0.152) (0.167) (0.128) (0.147)
Number of psu 65 65 65 65
Wald ch2 88.410 0.000 135.010 0.000 23.350 0.010 95.050 0.000
Log likelihood −404.066 −345.051 −547.505 −483.927
237
238
Kenya

(0.178)
(0.239)
Assets −0.235 0.100 −0.311 0.040 0.238 0.186 0.195 0.300
(0.143) (0.152) (0.180) (0.188)
BHNs −0.028 0.194 −0.039 0.093 0.003 0.921 0.014 0.652
(0.022) (0.023) (0.030) (0.031)
Lower education 0.176 0.309 0.109 0.548 0.312 0.167 0.334 0.156
(0.173) (0.182) (0.226) (0.236)
Part-time −0.354 0.071 −0.252 0.227 0.380 0.124 0.408 0.115
(0.196) (0.209) (0.247) (0.259)
Full-time −0.082 0.695 0.085 0.701 0.054 0.841 0.061 0.828
(0.208) (0.222) (0.270) (0.279)
Rural 0.605 0.040 0.657 0.033 0.736 0.057 0.637 0.115
(0.295) (0.308) (0.387) (0.404)
Infrastructure 0.154 0.089 0.171 0.071 0.099 0.394 0.097 0.424
(0.091) (0.095) (0.116) (0.122)
Groupsize10%+ −0.962 0.000 −0.397 0.154 0.929 0.006 0.605 0.093
(0.258) (0.278) (0.337) (0.360)
Luo 1.911 0.000 2.285 0.000 −0.036 0.927 −0.154 0.715
(0.310) (0.318) (0.398) (0.423)
Luhya 0.976 0.000 0.895 0.003 −0.549 0.120 −0.268 0.471
(0.279) (0.298) (0.353) (0.371)
Somali 0.466 0.135 0.369 0.263 0.168 0.747 0.332 0.536
(0.312) (0.330) (0.522) (0.536)
Kamba 0.924 0.002 0.631 0.048 −1.527 0.001 −1.378 0.005
(0.304) (0.319) (0.476) (0.492)
Kalenjin 0.604 0.035 0.453 0.131 −0.469 0.164 −0.230 0.517
(0.287) (0.300) (0.337) (0.355)
Kisii −0.746 0.060 −0.300 0.483 1.589 0.000 1.692 0.000
(0.397) (0.427) (0.429) (0.444)
Other Kenyans (omitted) (omitted) (omitted) (omitted)
Intercept −0.175 0.695 −1.250 0.010 −2.295 0.000 −2.402 0.000
(0.447) (0.483) (0.608) (0.636)
sd(_cons) 0.129 0.000 0.392 0.435
(0.395) (0.239) (0.228) (0.234)
Number of psu 136 136 131 131
Wald ch2 67.330 0.000 138.490 0.000 36.730 0.001 53.490 0.000
Log likelihood −519.042 −468.454 −327.853 −310.181
Uganda
(0.142)
(0.190)
Assets 0.045 0.660 0.076 0.504 0.247 0.030 0.151 0.219
(0.102) (0.114) (0.114) (0.123)
BHNs −0.091 0.000 −0.055 0.002 0.074 0.000 0.062 0.005
(0.015) (0.017) (0.020) (0.022)
Lower education −0.285 0.025 −0.216 0.134 0.006 0.967 −0.075 0.657
(0.127) (0.144) (0.155) (0.169)
Part-time −0.173 0.202 −0.238 0.121 −0.479 0.005 −0.470 0.010
(0.136) (0.154) (0.169) (0.183)
239
240
Full-time −0.199 0.263 −0.260 0.199 −0.315 0.131 −0.360 0.116

(0.177) (0.203) (0.209) (0.229)
Rural −0.393 0.115 −0.288 0.276 0.280 0.273 0.251 0.355
(0.249) (0.265) (0.256) (0.271)
Infrastructure −0.239 0.000 −0.214 0.002 −0.021 0.757 0.010 0.889
(0.065) (0.069) (0.067) (0.071)
Groupsize10%+ 0.001 0.997 −0.097 0.692 −0.549 0.050 −0.573 0.057
(0.224) (0.245) (0.281) (0.300)
Baganda 2.201 0.000 1.640 0.000 0.306 0.151 0.477 0.041
(0.267) (0.296) (0.213) (0.234)
Banyoro 1.859 0.000 1.219 0.008 −0.144 0.734 0.270 0.548
(0.420) (0.461) (0.422) (0.450)
Acholi 2.749 0.000 1.909 0.000 −0.730 0.122 −0.380 0.448
(0.414) (0.453) (0.472) (0.501)
Other Ugandans 2.256 0.000 1.477 0.000 −0.602 0.046 −0.182 0.577
(0.310) (0.341) (0.301) (0.325)
Intercept −0.527 0.151 −1.178 0.007 −0.067 0.872 −0.775 0.084
(0.367) (0.440) (0.415) (0.449)
sd(_cons) 0.711 0.699 0.282 0.135
(0.097) (0.113) (0.211) (0.496)
Number of psu 299 299 273 271
Wald ch2 149.220 0.000 361.100 0.000 36.830 0.000 140.320 0.000
Log likelihood −1060.74 −847.022 −632.568 −539.919
Source: Authors’ calculations based on the Afrobarometer R4 surveys.

Kenya, which may suggest the possibility that infrastructural development

generates competition between ethnic groups, possibly inducing a sense of
frustration and dissatisfaction rather than satisfaction. It is also interesting
to note that in Kenya the group size dummy is significant in both mod-
els and has the expected signs: that is, members of larger groups are less
likely to feel inferior and more likely to feel superior. In Zimbabwe, the
group size dummy is also significant and has the expected negative sign
in the first model (namely, the probability of respondents perceiving their
group to be socioeconomically inferior), while it has no statistical signifi-
cance in the second model. In the other cases, while the group size dummies
usually have the right sign except for Uganda, these are not statistically
significant.
Despite controlling for all these possibly distorting effects, we still find
discrepancies between what people perceive and what people ‘should’ per-
ceive given the objective HIs established in Section 9.3. In Nigeria, for
instance, the Igbo are less likely to think their group to be superior to
other groups and more likely to feel inferiority when compared to the
Hausa/Fulani (that is, the baseline group), who have more reason to have
such a feeling. In Ghana, the Ga/Dangbe (in addition to the Ewe) are more
likely to perceive collective deprivation compared to the Akan (the baseline
group). Likewise, the Ndebele are more likely to be frustrated collectively
compared to the Shona (the baseline group). In Kenya, a feeling of col-
lective deprivation is shared by all groups compared to the Kikuyu (the
baseline group), except for the Kisii. Similarly, groups that are not in a par-
ticularly disadvantaged position in Uganda (that is, the Baganda and the
Banyoro) share a feeling of frustration with the Acholi, who are objectively
poorer.
These results suggest that there might be other factors that could help
to explain the observed mismatch between objective and subjective HIs.
In what follows, we will examine the extent to which cross-dimensional
contamination might have contributed to the mismatching of subjective
and objective HIs in our case studies. As argued in Section 9.2, people’s
perceptions of the prevailing socioeconomic HIs may be ‘distorted’ by their
perceptions of the prevailing political HIs. Thus, people who feel their group
is politically excluded or marginalised may also feel that their group is at
a disadvantage in terms of socioeconomic development compared to other
groups or the politically dominant group.
Before examining whether the perception of political HIs has an impact on
people’s perception of socioeconomic HIs, it is necessary to explore people’s
perceptions of the political situation first. We can do this with the help of the
following question, which was included in both the JICA and Afrobarometer
surveys.
Think about the condition of your ethnic group. Do they have less, the
same, or more influence in politics than other groups in this country?
In the same vein as for the perceived socioeconomic HIs in Figure 9.4,
Figure 9.5 shows the distribution of responses to this question for both
surveys for our five case-study countries. The results show that the
Hausa/Fulani respondents perceive themselves to be in a relatively advan-
taged position, while the Igbo feel relatively marginalised politically.
In Ghana, it is the Akan who feel relatively advantaged, while a substan-
tial proportion of Ewe respondents feel politically inferior according to
Afrobarometer. In Zimbabwe, the Ndebele feel largely excluded from polit-
ical power. The Kenyan situation is very complex; yet, it appears that the
Luo are quite satisfied with their degree of political power perhaps due to
Odinga’s rise to power, while more than 70 per cent of the Somali people
felt they were politically inferior. Lastly, in Uganda, Banyankole respondents
appear relatively satisfied with their political situation, while the Banyoro
and Acholi feel considerably disadvantaged in terms of political power.
Without going into detail as to the extent to which people’s perceptions
of the prevailing political HIs are in line with the objective situation, the
Nigeria (Afrobarometer R4) Nigeria (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Hausa/Fulani Hausa/Fulani
Igbo Igbo
Yoruba Yoruba
Other Other
Nigerians Nigerians
Ghana (Afrobarometer R4) Ghana (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Akan Akan
Ewe Ewe
Ga/Dangbe Ga/Dangbe
Mole/Dagbani Mole/Dagbani
Other Other
Ghananians Ghananians
Zimbabwe (Afrobarometer R4) Zimbabwe (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Shona Shona
Ndebele Ndebele
Other Other
Zimbabweans Zimbabweans
Figure 9.5 Perceived political horizontal inequalities (PPHIs)

Kenya (Afrobarometer R4) Kenya (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Kikuyu Kikuyu
Luo Luo
Luhya Luhya
Somali Somali
Kamba Kamba
Kalenjin Kalenjin
Kisii Kisii
Other Other
Kenyans Kenyans
Uganda (Afrobarometer R4) Uganda (JICA surveys)

0% 20% 40% 60% 80% 100% 0% 20% 40% 60% 80% 100%
Total Total
Baganda Baganda
Banyankole Banyankole
Banyoro Banyoro
Acholi Acholi
Other Other
Ugandans Ugandans
Superior Same Inferior
observed patterns of responses appear to be broadly in line with what one

would expect on the basis of a detailed political-historical analysis and
contextualisation of the evolution and nature of the prevailing objective
political HIs in each of our case-study countries (see, for example, Mustapha,
2006 for Nigeria; Langer, 2008 for Ghana; and Stewart, 2010a for Kenya).
An interesting issue to be examined is the extent to which people’s
perceptions of the prevailing political and socioeconomic HIs overlap (that
is, go in the same direction). Our analysis in this respect shows that people’s
responses to both questions are quite highly correlated, as demonstrated for
the Nigerian case in Figure 9.6. This suggests there is a substantial association
between feeling politically excluded and socioeconomically disadvantaged,
or, conversely, feeling politically included or empowered and socioeconom-
ically advantaged. There are at least two reasons why it seems likely that
the perceived political situation affects people’s perceptions with respect to
the prevailing socioeconomic HIs and not the other way around. First, polit-
ical HIs are arguably more visible, and people are more likely to have an
‘informed’ opinion about their group’s political influence and inclusion in
the state institutions. Secondly, given that the state is the most important
economic actor in most African countries (for example, the state is usually
the largest employer and investor), controlling the state or having political
influence can be very important for different groups’ economic situation
Hausa/Fulani Igbo
26.34
28.36
30 22.15 30
15.94
25 25 16.96
10.23
20 20 9.06
8.72 6.43
15 11.91 15 11.99
5.7 2.85 4.97
7.6 4.39
10 10
4.19 Politically superior Politically superior
5 2.18 5
Politically equal Politically equal
Politically inferior Politically inferior
0 0
Economi- Economi- Economi- Economi- Economi- Economi-
cally cally cally cally cally cally
superior equal inferior superior equal inferior
Yoruba Other minorities
36.53
25.32
40 30
35
25 14.56
30 16.49
17.97
20
25
8.86 6.71
6.89
20 9.39 15 6.96 4.68
12.32
15 7.31 2.92 8.99
10 5.95
10 Politically superior Politically superior
3.97 4.18
Politically equal 5 Politically equal
5
0 Politically inferior 0 Politically inferior
Economi- Economi- Economi- Economi- Economi- Economi-
cally cally cally cally cally cally
superior equal inferior superior equal inferior
Figure 9.6 Correlations between Perceived Political HI (PPHI) and Perceived

Socioeconomic HI (PSEHI) in Nigeria
Source: Authors’ calculations based on Afrobarometer data.
and progress. Thus, having political influence might influence people’s

perceptions of the prevailing socioeconomic HIs because of the associated
economic power that comes with it.
The second and fourth columns in Table 9.3 report the estimated coeffi-
cients of determinants of perceived socioeconomic HIs when perceptions of
the prevailing political HIs are included. Interestingly, we find that feelings
of political group inferiority and superiority boost the feelings of economic
group inferiority and superiority, respectively, indicating the presence of
cross-dimensional contamination. After feelings of group political inferior-
ity are controlled for, the mere fact of affiliation with the Igbo in Nigeria,
and with the Ewe, Ga/Dangbe and residual minorities in Ghana, shows no
significant effect on the feeling of group economic inferiority. The Ndebele
in Zimbabwe still tend to feel group economic deprivation, but the size of the
effect is considerably reduced compared to the result when political group
deprivation is not controlled for. Likewise, when we control for the feeling of
political superiority, all ethnic dummies in Nigeria, Ghana, and Zimbabwe
lose statistical significance.
In Kenya and Uganda, we find a considerable reduction of the positive
influence of ethnic dummies on the feeling of economic inferiority in the
models that control for political group deprivation, except for the case of the
Luo who somewhat surprisingly show a strengthened tendency to exhibit
feelings of group economic deprivation. With regard to the probability of
feeling group economic superiority, we find that the Kamba and residual
minorities in Kenya, who are less likely to feel group economic superiority
compared to the Kikuyu when political feeling is not controlled for, are still
less likely to do so, but the size of the negative impacts themselves are con-
siderably reduced. The Kisii, on the other hand, become more likely to feel
economic superiority, which means that their economic confidence is sup-
pressed to a degree by their political feelings; this could be explained by the
fact that the Kisii are one of the most economically active communities in
Kenya. In the same vein, the Baganda in Uganda, who are not economically
inferior to the ruling Banyankole in objective terms, become more likely to
feel economic group superiority after political feelings are controlled for.
9.7 Some conclusions
This chapter examined the extent to which objective and subjective HIs dif-
fered in five African countries. So far, this issue has been largely ignored in
the literature on HIs. It was established that there were significant socioe-
conomic HIs in each of our five case-study countries. We also examined
the extent to which these seemingly ethno-based gaps could be explained
by personal characteristics and/or other environmental factors commonly
used to explain differences in standards of living between individuals. While
these individual risk factors explained a significant portion of the observed
inequalities, ethnic affiliations remained important as well. It also emerged
that different ethnic groups faced different chances of experiencing these
risk factors. In other words, there were significant ethnic differences in the
distribution of these risks.
Having established the presence of objective socioeconomic HIs, we
analysed people’s perceptions of these inequalities and found a surpris-
ingly large discrepancy between our objective and subjective measures
of inequality, especially in Nigeria, Ghana and Zimbabwe. Further anal-
ysis of the mismatch between objective and subjective HIs showed, first,
that people’s individual socioeconomic situation tended to ‘distort’ their
perceptions of their group’s situation and relative position, and, secondly,
that there appeared to be cross-dimensional contamination, whereby peo-
ple’s perceptions of political HI had a significant effect on their perceptions
of the prevailing socioeconomic HIs. Moreover, given the finding that
perceptions of socioeconomic and political HIs tended to be positively cor-

related, countries where an economically inferior group takes control of
political power, as in Nigeria and Zimbabwe, are more likely to have a
discrepancy between subjective and objective socioeconomic HIs than coun-
tries where an economically dominant group takes political power, as in
Kenya, and to some extent in Uganda and Ghana.
Appendix 1: Operationalisation
Female (both Afrobarometer and JICA): respondent’s gender (1 = female,

0 = male)
Age (both Afrobarometer and JICA): respondent’s age
Indigene (JICA): whether respondent is from the city in which survey was
conducted. (1 = yes, 0 = from somewhere else)
Lower education (both Afrobarometer and JICA): a binary measure coded
1 if respondent’s highest level of educational attainment is up to primary
school, coded 0 otherwise. ‘Don’t know’ and ‘refused to answer’ are treated
as missing value.
Insecurity (JICA): a binary measure based on the following question:
Over the last 12 months, how often, if ever, has your household gone without:
Physical security?
Original responses are measured by five-point ordinal scale (0 = Never,
1 = Just once or twice, 2 = Several times, 3 = Many times, 4 = Always), which
we recoded 0 if answer is ‘never’, 1 otherwise (excluding NA/DK).
Insecurity (Afrobarometer): sum of standardised (country-wide) scores of
responses to the following questions:
Over the past year, how often, if ever, have you or anyone in your family
A. Feared crime in your own home?

B. Had something stolen from your house?
C. Been physically attacked?
Responses are five-point ordinal scale (0 = Never, 1 = Just once or twice,

2 = Several times, 3 = Many times, 4 = Always), which we treated as if they
are interval scale. ‘Don’t know’ and ‘refused to answer’ are treated as missing
value.
Job (JICA): 1 = currently respondent has a job; 0 = currently respondent does
not have any job (including students and housewives).
Employment status (Afrobarometer): three-point ordinal scale (0 = no job,
1 = part-time job, 2 = full-time job) decomposed to dummy variables, Part-
time and Full-time, respectively.
Rural (Afrobarometer): 1 if PSU is rural; 0 otherwise.
Small Urban (Afrobarometer: Nigeria): 1 if PSU is coded as ‘small urban’; 0
otherwise.
Infrastructure (Afrobarometer): sum of weighted (with proportion of zero

within the capital) scores of responses to the following nine questions:
Are the following services present in the primary sampling unit/enumeration area:
1. Electricity grid that most houses could access?
2. Piped water system that most houses could access?
3. Sewage system that most houses could access?
4. Cell phone service?
Are the following facilities present in the primary sampling unit/enumeration area,
or within easy walking distance:
5. Post-office?
6. School?
7. Police station?
8. Health clinic?
9. Market stalls (selling groceries and/or clothing)?
10. Was the road at the start point in the PSU/EA paved/ tarred/ concrete?
Groupsize10%+ (Afrobarometer): 1 if the size of ethnic group at national
level exceeds 10%; 0 otherwise.
Assets (JICA): sum of weighted (with proportion of zero within the capital)
scores of responses to the following seven questions:
Which of these things do you personally own? (1 = yes, 0 = no)
A. Radio
B. Bicycle
C. Television
D. Mobile phone
E. Refrigerator
F. Flush toilet
G. Car
Assets (Afrobarometer): sum of weighted (with proportion of zero within the
capital) scores of responses to the following seven questions:
Which of these things do you personally own? (1 = yes, 0 = no)
A. Radio
B. Television
C. Motor vehicle, car or motorcycle
Where is your main source of water for household use located? (1 = Inside the
house, 0 = Inside the compound or outside of the compound)
How often do you use:
A. A mobile phone?
B. A computer?
C. The internet?
1 = Every day, A few times a week, A few times a month or, Less than once a
month, 0 = Never.
BHNs (both Afrobarometer and JICA): sum of standardised (within the
capital) scores of responses to the following five questions:
Over the past year, how often, if ever, have you or anyone in your family gone
without:
A. Enough food to eat?

B. Enough clean water for home use?
C. Medicines or medical treatment?
D. Enough fuel to cook your food?
E. A cash income?
Responses are five-point ordinal scale (0 = Never, 1 = Just once or twice,

2 = Several times, 3 = Many times, 4 = Always), which we treated as if they
are interval scale. ‘Don’t know’ and ‘refused to answer’ are treated as missing
value.
PSEHI (perceived socioeconomic horizontal inequality) (both JICA and
Afrobarometer):
Economically inferior: 1 = Much worse/Worse; 0 = Same; otherwise treated
as missing
Economically superior: 1 = Much better/Better; 0 = Same; otherwise treated
as missing
Based on the following questions and answers:
Think about the condition of ____________ [R’s Ethnic Group]. Are their economic
conditions worse, the same as, or better than other groups in this country?
Responses: 1 = Much better, 2 = Better, 3 = Same, 4 = Worse, 5 = Much
worse, 7 = Not applicable, 9 = Don’t know, 998 = Refused to answer,
–1 = Missing data
PPHI (perceived political horizontal inequality) (both JICA and
Afrobarometer):
Politically inferior: 1 = Much less/Less; 0 = Same; otherwise treated as
missing
Politically superior: 1 = Much more/More; 0 = Same; otherwise treated as
missing
Based on the following questions and answers:
Think about the condition of ____________ [R’s Ethnic Group]. Do they have less,
the same, or more influence in politics than other groups in this country?
1 = Much more, 2 = More, 3 = Same, 4 = Less, 5 = Much Less, 7 = Not appli-
cable, 9 = Don’t know, 998 = Refused to answer, –1 = Missing data
Appendix 2: Descriptive statistics
JICA Survey
Variable Country Obs Mean Std. Dev. Min Max Variable Country Obs Mean Std. Dev. Min Max.
Assets Nigeria 406 0.869 0.534 0.000 2.586 Lower education Nigeria 411 0.129 0.336 0 1
Ghana 324 1.033 0.767 0.000 2.889 Ghana 324 0.395 0.490 0 1
Zimbabwe 400 1.406 0.802 0.000 3.127 Zimbabwe 402 0.142 0.349 0 1
Kenya 904 0.885 0.663 0.000 2.213 Kenya 902 0.213 0.410 0 1
Uganda 498 1.257 0.905 0.000 3.875 Uganda 499 0.297 0.457 0 1
BHNs Nigeria 402 0.029 3.950 −13.929 4.276 Insecurity Nigeria 408 0.593 0.492 0 1
Ghana 324 0.000 3.395 −13.432 3.386 Ghana 324 0.309 0.463 0 1
Zimbabwe 397 0.107 3.376 −7.947 7.720 Zimbabwe 397 0.587 0.493 0 1
Kenya 892 −1.313 4.611 −19.482 2.980 Kenya 902 0.220 0.414 0 1
Uganda 491 −0.572 4.045 −14.819 3.712 Uganda 494 0.287 0.453 0 1
Female Nigeria 412 0.500 0.501 0 1 Job Nigeria 410 0.688 0.464 0 1
Ghana 324 0.491 0.501 0 1 Ghana 324 0.858 0.350 0 1
Zimbabwe 402 0.498 0.501 0 1 Zimbabwe 402 0.729 0.445 0 1
Kenya 907 0.502 0.500 0 1 Kenya 906 0.670 0.470 0 1
Uganda 500 0.566 0.496 0 1 Uganda 500 0.652 0.477 0 1
Age Nigeria 412 33.049 13.013 18 84 PSEHI Nigeria 404 2.698 0.862 1 5
Ghana 323 38.774 15.688 18 99 Ghana 313 3.080 1.002 1 5
Kenya 907 32.871 11.077 18 99 Kenya 880 2.880 0.849 1 5
Uganda 499 30.826 11.639 18 100 Uganda 484 3.045 1.064 1 5
Indigene Nigeria 412 0.427 0.495 0 1 PPHI Nigeria 403 2.918 0.939 1 5
Ghana 324 0.241 0.428 0 1 Ghana 303 2.865 1.076 1 5
Kenya 905 0.259 0.438 0 1 Kenya 875 3.170 1.388 1 5
Uganda 483 0.290 0.454 0 1 Uganda 484 3.279 1.099 1 5
249
Afrobarometer
250
Variable Country Obs Mean Std. Dev. Min Max Variable Country Obs Mean Std. Dev. Min Max
Assets Nigeria 2245 0.812 0.684 0.000 2.909 Rural Nigeria 2324 0.507 0.500 0 1
Ghana 1177 0.688 0.732 0.000 3.480 Ghana 1200 0.547 0.498 0 1
Zimbabwe 1186 0.825 0.888 0.000 3.607 Zimbabwe 1200 0.633 0.482 0 1
Kenya 1062 0.640 0.661 0.000 2.770 Kenya 1104 0.775 0.418 0 1
Uganda 2381 0.525 0.676 0.000 3.646 Uganda 2431 0.799 0.401 0 1
BHNs Nigeria 2266 −0.588 4.390 −13.556 4.635 Small Nigeria 2324 0.129 0.335 0 1
Ghana 1174 −1.281 5.020 −18.695 3.258 urban Ghana
Zimbabwe 1181 0.601 2.839 −6.204 7.385 Zimbabwe
Kenya 1097 −1.188 3.774 −13.758 4.380 Kenya
Uganda 2412 −2.432 4.182 −14.787 3.690 Uganda
Female Nigeria 2324 0.499 0.500 0 1 Infra- Nigeria 2051 1.648 0.942 0.000 3.640
Ghana 1200 0.500 0.500 0 1 structure Ghana 1096 2.046 1.328 0.000 3.915
Zimbabwe 1200 0.500 0.500 0 1 Zimbabwe 1192 1.985 1.418 0.013 3.816
Kenya 1104 0.502 0.500 0 1 Kenya 1104 1.808 1.399 0.000 4.863
Uganda 2431 0.499 0.500 0 1 Uganda 2423 1.838 1.570 0.000 6.020
Age Nigeria 2316 31.303 11.410 18 86 Employ- Nigeria 2299 0.773 0.841 0 2
Ghana 1184 39.017 16.459 18 110 ment Ghana 1191 0.959 0.894 0 2
Zimbabwe 1189 36.563 15.294 18 94 status Zimbabwe 1199 0.362 0.700 0 2
Kenya 1100 35.215 12.918 18 95 Kenya 1100 0.595 0.786 0 2
Uganda 2421 33.709 12.286 18 81 Uganda 2430 0.538 0.729 0 2
Lower Nigeria 2319 0.266 0.442 0 1 PSEHI Nigeria 2245 3.017 1.005 1 5
education Ghana 1195 0.623 0.485 0 1 Ghana 1121 3.093 1.114 1 5
Kenya 1100 0.459 0.499 0 1 Kenya 1078 3.412 1.053 1 5
Uganda 2431 0.504 0.500 0 1 Uganda 2334 3.586 1.059 1 5
Insecurity Nigeria 2293 −0.009 2.334 −1.80298 12.15934 PPHI Nigeria 2239 2.926 1.062 1 5
Ghana 1194 0.001 2.046 −1.36252 14.01731 Ghana 1087 2.959 1.211 1 5
Zimbabwe 1196 0.003 2.126 −1.91515 11.36737 Zimbabwe 1097 3.027 1.067 1 5
Kenya 1096 0.000 2.195 −1.81166 12.48099 Kenya 1073 3.185 1.119 1 5
Uganda 2421 −0.001 2.236 −1.84514 10.24266 Uganda 2290 3.559 1.044 1 5
Notes
1. Langer (2005) provides a theoretical foundation for this empirical finding
by emphasising that the simultaneous presence of severe political horizontal
inequalities and socioeconomic horizontal inequalities forms an extremely explo-
sive sociopolitical situation, because in these situations the excluded political
elites not only have strong incentives to mobilise their supporters for violent
conflict along ethnic lines but also are likely to gain support among their ethnic
constituencies quite easily.
2. Please note that the word objective was put in inverted commas to indicate that
any indicator can only be an approximation of the objective reality. Furthermore,
the selection of entities, variables or indicators used to quantify objective horizon-
tal inequalities at a particular point in time is clearly to some extent an arbitrary
choice by the researchers involved.
3. For more information on the Minorities at Risk project, visit http://www.cidcm.
umd.edu/mar/ (last accessed 1 May 2013).
4. These perceptions surveys were conducted as part of the JICA-RI project ‘Preven-
tion of violent conflicts in Africa’.
5. For the Afrobarometer surveys, visit http://www.afrobarometer.org/ (last accessed
1 May 2013).
6. Appendix 1 provides a detailed description of the operationalisation of the
variables used in our analysis.
7. Please note that scores were ‘normalised’ vis-à-vis the capital of the country,
expect for Nigeria, where Lagos was used as a base.
8. Stewart and Langer (2008) propose a formal framework for understanding the
persistence of group inequalities based on the following factors: unequal rates of
accumulation; dependence of returns of one type of capital on the availability of
other types of capital; and asymmetries in social capital.
9. We used OLS for the model that examined determinants of ‘Assets’ in Zimbabwe
because the mixed-level model failed to converge.
10
Findings and Implications: The Role
of Development Cooperation
Frances Stewart, Thandika Mkandawire and Mari Katayanagi
10.1 Introduction
The aim of this book is to seek appropriate measures to prevent violent

conflict in sub-Saharan African countries, combining the triple perspectives
of objective inter-group inequalities, subjective perceptions of inequalities,
and the role of political institutions. The underlying thesis is that horizontal
inequalities (HIs) in socioeconomic, political and cultural-status dimensions
form a major part of the root causes of violent conflict. Political mobili-
sation for violence is more likely when HIs are consistent across different
dimensions, in other words, running in the same direction. These consider-
ations are especially relevant to Africa where there are multiple ethnicities
and ethnic relations are extremely diverse and complex.
Among the various dimensions of HIs, this project started by focusing
attention on political HIs, since political institutions and policies are often
in flux and changeable, especially in post-conflict situations in African
countries, as described in Chapter 1. In order to improve our understand-
ing of the ways in which political systems affect sociopolitical stability,
Chapter 2 of this book classified political systems into two broad types:
power-dispersing (PD) and power-concentrating (PC) types. A PD institu-
tion is characterised by power-sharing and decentralisation. The chapter
categorises country institutions in this way and traces changes in this
categorisation in each country in the 2000s. It is hypothesised that PD
institutions are more conducive to political stability than PC institutions,
especially where group-based grievances are rife.
However, as this classification is based only on the formal aspects of politi-
cal systems, we undertook a set of country studies to shed light on historical
trends, particular societal features and the informal conventions of individ-
ual countries. The case studies not only examine the actual functioning of
formal rules against a backdrop of rich and idiosyncratic local contexts, but
also attempt to assess prevailing multidimensional HIs in these countries.
252
Frances Stewart, Thandika Mkandawire and Mari Katayanagi 253
The authors used the available data on socioeconomic HIs but a daunting
problem was the paucity of appropriate statistics in conflict-prone countries
in Africa. The authors combined statistical evidence with a description of
the historical evolution of boundaries and inequalities between groups.
In conflict processes, people’s perceptions and views play critical roles.
Where available, therefore, case studies incorporate the results of JICA
perceptions surveys, which were designed to examine how people them-
selves see their own identities, those of others and their views of the
prevailing HIs and their country’s political system, as well as their attitudes
towards violence. Chapter 9 provides a cross-country analysis of the survey
results using polimetric methods.
In the following section, we give a brief summary of each chapter. Next, we
formulate general findings from the study as a whole, presenting some tables
that elucidate the relationships between different dimensions of HIs and
classify the political institutions of each country based on both their for-
mal and informal features. We then touch on contemporary processes in
conflicts in Africa, involving multiple actors both inside and outside the
continent. The final section concludes with a set of policy recommenda-
tions for effective conflict prevention, directed to development agencies and
other stakeholders in Africa.
10.2 Chapter summaries
The backbone of this book consists of case studies. Most of the country cases
consist of paired comparisons: Rwanda and Burundi (Chapter 3), Ghana
and Côte d’Ivoire (Chapter 4), South Africa and Zimbabwe (Chapter 5) and
Tanzania and Uganda (Chapter 6). Each of the countries in the paired com-
parisons exhibits differences in the way they manage HIs through political
institutions and policies, with contrasting political outcomes in terms of
stability and instability. In addition to these four pairs, there are two single-
country cases of large ethnically heterogeneous countries, Kenya (Chapter 7)
and Nigeria (Chapter 8), which exemplify unique approaches to managing
diversity through their political systems, as well as in changes over time.
These case studies are framed by two chapters that contain quantitative anal-
ysis on institutions (Chapter 2) and perceptions (Chapter 9). The integration
of quantitative analysis and qualitative case studies is a major feature of this
book. The following are summaries of all these chapters.
10.2.1 Institutions, inequalities and perceptions (Introduction)

In Chapter 1, Sakiko Fukuda-Parr, Arnim Langer and Yoichi Mine present our
overall motivation in undertaking this study. The aim of this book is to
contribute to an understanding of the forms of political and economic
governance that would contribute to the prevention of political violence.
254 Findings and Implications
Despite the strong potential for peace in Africa following the end of the
Cold War which had instigated and sustained a number of conflicts, the
continent faced a spate of violent conflicts in the 1990s against the back-
drop of post-Cold War globalisation. Many of these conflicts are identity
based. Previous studies, both case studies and econometric analysis, have
shown a strong relationship between HIs and violent conflict, and, con-
sequently, this study gives weight to HIs. We focus in particular on two
issues that have previously been dealt with insufficiently: the relationship
between objective and subjective HI; and the interaction between political
institutions and HIs.
10.2.2 Analysis of institutions: advantage of power-dispersing

institutions for reducing conflict propensity
In Chapter 2, Yoichi Mine, Mari Katayanagi and Satoru Mikami discuss
the comparative advantage of different political institutions in relation
to conflict prevention in today’s Africa. The political systems of 49 sub-
Saharan African countries are classified into power-dispersing (PD) and
power-concentrating (PC) institutions. PD institutions are characterised by
proportional representation, coalition government, federalism and other
power-sharing elements, while PC institutions involve a first-past-the-post
voting system, a majority party executive and a unitary state structure
among other characteristics. The choice of institutions in any country is
affected by its territorial and population size, and its ethnic configurations
and colonial background, although none of these factors on their own deter-
mine institutional choice. In order to prevent or resolve violent conflicts,
many countries have adopted new institutions and have consequently rad-
ically shifted their position in the PD/PC mapping. Given the rich history
of institutional experiments in large heterogeneous countries, the authors
predict that some countries may adopt hybrid regimes in the longer term,
combining PD and PC elements. While the design of formal institutions is
expected to influence the paths of African countries, informal conventions
and values are also relevant to the understanding of the nature of societal
resilience to violent ruptures.
The major hypothesis put forward in Chapter 2 is that PD institutions, a
combination of power-sharing and decentralisation, are generally most suit-
able for conflict prevention. Quantitative analysis in the chapter indicates
that decentralised institutions are associated with more stable situations.
In the 1990s and 2000s, a number of conflict-ridden African countries,
including South Africa, the DRC, Côte d’Ivoire, Burundi, Sudan, Kenya and
Zimbabwe, introduced ad hoc power-sharing with the intention of reducing
cumulative tensions, yet these have persisted. This may be because in efforts
to bring about sustainable peace any attempt at constitutional engineering
must be accompanied by effective policies to address socioeconomic HIs.
10.2.3 Dissimilar twins (Rwanda and Burundi)

Chapter 3 by Shinichi Takeuchi, the first chapter presenting a case study,
compares the experiences of two densely populated nations, in the Great
Lakes Region, Rwanda and Burundi, the so-called twin countries (les pays
jumeaux), where the major ethnic cleavage is to be found between the
majority Hutu and the minority Tutsi. During colonial rule, the Hutu were
sidelined, while the Tutsi were favoured by the administration due to their
dominant positions in the traditional kingdoms. Even though the patterns
of ethnic configuration and HIs have been largely similar in these countries,
these twins had different modalities in relation to pre-colonial rule in their
local kingdoms – a centralised system in Rwanda and a decentralised one in
Burundi, and this historical difference may have conditioned the choice of
institutions after independence.
Rwanda and Burundi pursued radically different paths in the postcolonial
decades. In Rwanda, the majority Hutu boosted their influence through a
so-called ‘social revolution’, a radical transformation from ‘feudalism’ in
which the minority Tutsi ruled, to ‘democracy’ dominated by the major-
ity Hutu. However, soon after the genocide in 1994, organised by Hutu
supremacists, the regime was subverted through the military victory of
Tutsi-led forces. Although the subsequent Rwandan government propagated
national unity and introduced several apparently power-dispersing mea-
sures, in practice the new regime is clearly power-concentrating and favours
Tutsi returnees. At the same time, grievances among Hutu people who
are now discriminated against in all dimensions of HIs are believed to be
growing under the surface.
In Burundi, in contrast, the Tutsi elites consolidated their power until an
intensification of violent conflict in the 1990s. After a military stalemate
and vigorous international mediation, the Burundian government decided
to introduce an extensive power-dispersing system, culminating in constitu-
tional power-sharing arrangements in 2005 based on the principle of ethnic
proportionality. The power-dispersing institutions in Burundi have proved
effective in diminishing the danger of violent ethnic mobilisation. However,
the political power struggle has not been fully institutionalised in the present
system, and Burundian politics are still prone to organised violence. In these
two countries, the contrasting choice of political institutions was a direct
outcome of different ways in which their wars ended – the decisive victory
of one side in Rwanda and the military stalemate and negotiated peace in
Burundi.
10.2.4 Significance of leadership quality (Ghana and Côte d’Ivoire)

Chapter 4 by Arnim Langer traces the contrasting paths taken by Ghana and
Côte d’Ivoire, comparable countries in West Africa. The ethnic configura-
tions and HIs of this pair of nations are basically similar in that both are
characterised by a division between a relatively rich South and a relatively

poor North. However, there are some noticeable differences in social config-
urations. In Ghana, the Akans in the South account for around half of the
population, while the Northern groups lack unity as they are segmented into
small subgroups. In both countries, the Muslim population is largely concen-
trated in the North, but Christians account for more than two-thirds of the
Ghanaian population, including many in the North, while in Côte d’Ivoire,
Christians and Muslims are comparable in size and the North is largely
Muslim and the South largely Christian. In Côte d’Ivoire, the Northern
region is more populous, and includes a majority of ‘non-Ivorian’ migrants.
Such a contrast between the social configurations of the two countries may
partly explain the relative stability of Ghana compared with Côte d’Ivoire.
Ethnic politics as the ‘intervening variable’ between HIs and violent con-
flict seems to have had a greater effect on the contrasting situations in the
2000s. Since the time of Kwame Nkrumah, successive regimes of Ghana have
promoted national inclusiveness and cultural neutrality. Nkrumah demon-
strated and celebrated the pluralistic nature of the Ghanaian identity, for
example by wearing a range of traditional modes of dress at official cer-
emonies. In today’s mature democracy in Ghana, even the majoritarian
system has not led to political instability. Ethnic balance in political rep-
resentation and some economic redistribution towards deprived Northern
regions appear to have contributed greatly to sustaining social stability.
In Côte d’Ivoire, the first president, Félix Houphouët-Boigny, promoted
investment in the Northern and Central regions, and also alternated inde-
pendence festivities between Abidjan and different prefectural capitals.
However, ethno-nationalism and xenophobia were introduced into poli-
tics during the latter days of his rule and to a greater extent after his
rule. The growth of the autochthonistic notion of Ivoirité together with
economic stagnation reinforced ethnic exclusion, eventually leading to a
military insurgency which started in 2002, and was only ‘settled’, at least
for the short term, in 2011. In this chapter, emphasis is put on the quality
of leadership and informal conventions, which have played critical roles in
supporting political stability in Ghana but fuelled ethnic violence in Côte
d’Ivoire.
10.2.5 Ad hoc power-sharing as a refuge (South Africa

and Zimbabwe)
Chapter 5 by Yoichi Mine compares South Africa and Zimbabwe in south-
ern Africa. In both countries, major HIs developed historically along racial
lines as a result of colonial immigration and policies. However, the social
configurations of black Africans were somewhat different. South Africa’s
anti-apartheid movement was characterised by a unity of urban political
forces, while its counterpart in Zimbabwe faced rivalry among guerrilla com-
manders. Both political and cultural-status HIs between racial groups were
eventually redressed with the independence of Zimbabwe in 1980 and the

liberation of South Africa in 1994. While social and economic HIs between
races were also gradually reduced after these turning points, they remain
high and economic stratification among black Africans is becoming large in
both countries.
In South Africa, the grand coalition instituted after the end of apartheid
has been dissolved, but the spirit of power-sharing still remains. In
Zimbabwe, in contrast, the government of Robert Mugabe consolidated
majoritarian power-concentrating institutions, excluding the opposition
from power, but the regime had to accept power-sharing in 2008 after a
presidential election which was fraught with violence against opposition
supporters resulting in a loss of legitimacy on the part of ‘re-elected’ Presi-
dent Mugabe. In the JICA perceptions surveys, respondents in both countries
agreed that class was becoming more important than race. They were also
generally positive about power-sharing as a way of solving political crises.
In Zimbabwe, it was found that feelings of victimisation were strong among
the largest minority group, the Ndebele. The political experiences of the two
countries indicate that ad hoc power-sharing is useful to prevent violent con-
flict. While informal power-sharing contributes to political stability, formal
power-sharing mechanisms such as proportional representation help rein-
force informal power-sharing, especially in countries with serious HIs. In the
longer term, however, it is essential to promote equitable development, as it
is impossible to attain sustainable democracy by constitutional mechanisms
alone.
10.2.6 Pros and cons of federalism (Tanzania and Uganda)

In Chapter 6, Yuichi Sasaoka and Julius Nyang’oro discuss the contrasting
ways of managing ethnic politics in Tanzania and Uganda in East Africa.
Prominent political leaders such as Julius Nyerere in Tanzania and Yoweri
Museveni in Uganda propounded African socialism and integrationist
nation-building. However, both countries had to deal with the aspira-
tions of politically influential minorities for substantial autonomy, Zanzibar
in Tanzania and the Buganda in Uganda. Recently, both these countries
accepted multi-party democracy and started to implement radical decentral-
isation policies. In terms of formal political institutions, their positions are
very close to each other (Chapter 2). Despite these commonalities, however,
these countries experienced contrasting political situations, with relative
stability in Tanzania and instability in Uganda.
One major reason for such divergent situations lies in the different
societal configurations. The ethnic identities of Tanzanians are extremely
fragmented, and people tend to prioritise their national identity over eth-
nicity, as confirmed in the JICA surveys. Moreover, the Tanzanian govern-
ment adopted semi-federal and power-sharing arrangements with regard to
Zanzibar after the unification in 1964, recognising their political and cultural
autonomy. A spirit of inclusive power-sharing seems to have permeated the

leadership style and party politics in Tanzania.
In contrast, ethnic cleavages in Uganda have been more pronounced as
is also confirmed in the JICA surveys. The Ugandan government led by
the NRM (National Resistance Movement) has tried to gain control over
the whole territory, fighting with the Northern rebels for more than two
decades. The current decentralisation empowers small administrative units
rather than large ethnic polities and the quest for political autonomy by the
Buganda Kingdom has been frustrated, but resentment remains. The cen-
tral government of Uganda may have to consider more power-dispersing
institutions as in semi-federal Tanzania, though granting more autonomy to
Buganda located in the central area surrounding the capital city would be
problematic.
10.2.7 Breaking away from ethnic politics (Kenya)

Chapter 7 by Mwangi Kimenyi argues that deeply entrenched grievances
based on ethno-regional inequalities are the most important factor behind
election-related violence in Kenya. Multidimensional HIs between ethnic
groups such as Kikuyu, Luhya, Luo, Kalenjin, Kamba and others were con-
solidated during the colonial period, and since independence all major
political parties have operated on the basis of ethnicity. By the end of the
1960s, Kenya became a power-concentrating single-party state with a strong
‘imperial’ presidency. The ethnic group of an incumbent president has been
consistently overrepresented in government and thus political HIs have
shifted in accordance with the ethnic affiliations of the presidents (with
Kikuyu advanced by Kenyatta, Kalenjin under Moi, and Kikuyu again by
Kibaki). The distribution of socioeconomic benefits has been biased towards
those groups with political power. While a multi-party system was rein-
troduced in 1991, party mobilisation along ethnic lines intensified. Since
democratisation the country has been haunted by ethnic violence each time
elections took place.
The ‘stolen’ presidential election at the end of 2007 sparked widespread
ethnic violence in the Rift Valley and elsewhere, claiming the lives of
a thousand Kenyans, and leading to the formation of a Government of
National Unity based on ad hoc power-sharing. In 2010, Kenyans voted
for a new constitution, which contains power-dispersing mechanisms such
as checks on presidential power, further decentralisation to local govern-
ments called Counties, and the introduction of bicameralism, reviving some
key clauses of the independence constitution of 1963. According to the
JICA perceptions surveys, the majority of Kenyan respondents identified
themselves more as Kenyans than as belonging to their ethnic group. Despite
signs of hope, the author cautions that given the underlying socioeconomic
HIs as well as prevalent negative perceptions about other ethnic groups, the
new constitution will not solve all problems, and that other policy steps
need to be taken to address socioeconomic HIs to prevent the recurrence of

conflict.
10.2.8 Hybrid constitutional engineering (Nigeria)

Chapter 8 by Ukoha Ukiwo is the final case-study chapter, examining the
experience of Nigeria, the most populous country in Africa with no less than
350 ethnic groups, including the Hausa and Fulani in the North, the Yoruba
in the West and the Igbo in the East. The ethnic configurations of Nigeria
are very fluid in terms of identities and geographical boundaries. After inde-
pendence, the successive creation of new states following the Biafran war
encouraged a further proliferation of identities. In terms of religion, how-
ever, Nigerian society presents a bipolar structure with the country divided
between Northern Islam and Southern Christianity. The Northerners tend
to dominate the political and military leadership but the Southerners are
substantially better off in economic terms, due to more vibrant and higher-
value economic activities in their region. Such a balance between major
groups based on the asymmetric distribution of political and economic
power may have contributed to containing local conflicts and preventing
them escalating into national turmoil in Nigeria since 1970.
The design of political institutions in Nigeria has been strongly influ-
enced by the bitter lessons of the Nigerian Civil War (1967–1970). Through
successive reforms, Nigeria has succeeded in avoiding a relapse into large-
scale conflict for more than four decades, although there have been periodic
episodes of localised violence. The success of conflict prevention seems to be
the result of adopting a hybrid system which combines power-concentrating
and power-dispersing institutions. For example, Nigeria’s electoral system is
majoritarian, and enormous power is conferred on the ‘imperial’ presidency,
due to fear of the nation’s falling apart. At the same time, ethnic-based polit-
ical parties are prohibited, and an informal principle of ethnic alternation of
key positions is respected in various institutions. As a means of addressing
persistent HIs, Nigeria has introduced the Federal Character Principle, which
ensures ethnic balance in government institutions, and also established the
Federation Account, a pool of federally collected revenues, for which the dis-
tribution formula is legislated by the National Assembly. The formula aims
to achieve a fair distribution of the substantial oil revenues across states.
The JICA perceptions survey in Lagos showed that Nigerians attached more
importance to religious than ethnic identity, and that the Igbo tended to
feel relatively marginalised. The future stability of Nigeria depends on the
deepening of democracy and support for equitable, ‘non-oil’ development.
10.2.9 Primacy of perceptions (survey analysis)

In Chapter 9, Arnim Langer and Satoru Mikami analyse the outcome of
JICA perceptions surveys in five countries, Ghana, Kenya, Nigeria, Uganda
and Zimbabwe. In addition to the original JICA surveys, this polimetric
analysis utilises Afrobarometer Round 4 surveys. In the five countries in

question, the existence of socioeconomic HIs is statistically confirmed.
Although those HIs can be partly explained in terms of individual risk fac-
tors, ethnic affiliations still matter; moreover, there are significant ethnic
differences in the distribution of these risk factors. However, the chapter
shows that subjective inequalities, or those perceived by people themselves,
can differ substantially from objective, material inequalities. This inconsis-
tency is important because people take action because of perceived injustices
rather than because of statistical inequalities that they may not be aware
of. The authors argue that such a discrepancy may, in theory, be caused
by differences in personal circumstances, elite manipulation of popular
perceptions and lack of information, as well as various other sources of
misconceptions.
In the JICA surveys, a large discrepancy between subjective and objective
socioeconomic HIs is to be observed, especially in the cases of Nigeria, Ghana
and Zimbabwe. It is also found that the ‘distortion’ of their perceptions
about socioeconomic HIs are caused not only by people’s individual socioe-
conomic situation but also by their subjective perceptions of political HIs.
In other words, when people feel they are politically discriminated against as
a group, they tend to think they are worse off in socioeconomic terms than
they actually are. Although it was not possible to contrast objective and sub-
jective political HIs because of lack of precise data on political HIs, on the
basis of known political imbalances the chapter argues that perceptions of
political HIs are broadly in line with objective political HIs. Discrepancies
between subjective and objective HIs appear to be larger in a country in
which an economically inferior group takes control of political power, like
in Nigeria and Zimbabwe, than in a country in which an economically dom-
inant group holds political power, like in Kenya, and to some extent Uganda
and Ghana.
10.3 General findings
Based on the quantitative analysis and qualitative case studies just sum-
marised, this section sets out our general findings. Tracing the flow of logic
in the case studies, we start by presenting a series of findings about eth-
nic configurations and HIs in these countries. Next, findings about popular
perceptions on political and socioeconomic HIs and their implications are
formulated. Then the advantage of different types of political institutions in
addressing overall HIs is discussed in the light of the lessons learnt from case
studies.
10.3.1 Social configurations and structural inequalities

Our studies show that the formation of salient identity groups and the emer-
gence of socioeconomic HIs generally have a long historic origin, dating back
to colonial and sometimes pre-colonial times. In the cases of South Africa

and Zimbabwe, fierce racism was the consequence of late colonialism and
remained significant even after the establishment of majority rule at the end
of the twentieth century. Elsewhere, regional, ethnic or religious differences
emerged over a long time span – often accentuated by colonialism. This was
the case in every country observed.
The historical effects include the special position of the Buganda King-
dom in Uganda (which predates colonialism), for example, and the land
settlement scheme in the years immediately after independence in Kenya.
In Nigeria, colonialism left divisions between North and South, and between
settlers and indigenes which continue to be of relevance to disputes today.
Ghana had been ruled as a single state during colonialism (as the Gold
Coast), while, in contrast, Northern and Southern Nigeria were treated
separately – and indeed there was some consideration of the possibility of
creating two countries rather than one. The historical legacy, which includes
sharp HIs between groups in every one of these cases, makes peaceful man-
agement of the countries difficult. Over time, feelings of dissatisfaction
may grow beneath the surface (for example, among Northerners in Côte
d’Ivoire, Acholi and Buganda in Uganda, Hutus in Rwanda, and Ndebele in
Zimbabwe). The HI situation of each country needs to be studied carefully in
a historical perspective and any ongoing changes should be observed. In a
fragile situation, vigilance is needed in relation to grievances of a particular
identity group caused by historical HIs.
The demographic configuration of ethnic groups in a country can also
matter for political stability. Unipolar and multipolar configurations may
make it easier to achieve political stability, given good management of ethnic
differences (Tanganyika, which was highly fragmented, is one such exam-
ple). However, when a bipolar or tripolar configuration coincides with HIs,
the risk of violent conflicts seems to increase, as indicated by the experience
of other countries, for example, Burundi and Rwanda.
Both case-study evidence and quantitative surveys show considerable
HIs in each country. To a large extent, differences in economic outcomes
are the result of historic and contemporary differences in infrastructural
development, in levels of education and access to good jobs – themselves
important dimensions of HI. However, there are HIs in outcomes, even after
one takes into account the effects of HIs in such variables (Chapter 9).
Like other studies (Stewart ed., 2008; Cederman et al., 2010), these cases
confirm that HIs in socioeconomic dimensions are more likely to lead to
conflict where there are also similar HIs in political power. This is shown, for
example, by the history of South Africa and Kenya. In each case, violence
erupted when the same group experienced deprivation in both political
and socioeconomic dimensions; violence did not occur when one group
was deprived in socioeconomic dimensions but dominated politically. Simi-
larly, in Côte d’Ivoire persistent socioeconomic HIs only led to conflict when
there was also political exclusion. This points to the importance of political
inclusion – and consequently of the political system – as well as the ways
in which people perceive their positions in such a system, for stability. It is
important to note that when it is difficult to introduce or sustain redistribu-
tive policies, possibly due to economic recession, and political exclusion
continues, the conflict risk may sharply increase (for example, Côte d’Ivoire,
Zimbabwe).
10.3.2 Popular perceptions and political horizontal inequalities

From the perspective of conflict prevention, it is important to recognise
that people take actions because of perceived HIs rather than because of
the ‘objective’ HIs shown in statistics. Moreover, the perceived HIs may
not be the same as ‘objective’ HIs. Our surveys show that there can be
considerable differences between perceptions of socioeconomic HIs and
actual HIs, with notable examples of this in Nigeria and in Ghana,
where the groups perceived as poorest are not those who are the poorest
according to social and economic statistics. This distortion of perceptions
seems to be associated with political HIs. In both of these cases, the
groups that are politically marginalised appear to regard themselves as also
being economically marginalised, even when this contradicts the objective
evidence.
According to our surveys, perceptions of political HIs seem to reflect the
realities more accurately than perceptions of socioeconomic HIs. That being
so, one way of reducing perceptions of socioeconomic HIs is to reduce polit-
ical HIs. This may avoid the situation where political and socioeconomic
HIs together reinforce a perception of inferiority. However, large socioeco-
nomic HIs remain unjust and increase the possibility of a violent conflict,
while a radical shift in political power may occur and lead to a politically
sensitive situation with high potential for conflict. Though it may sound
counterintuitive, those who have confidence in group economic superiority
can be as or more hostile to other groups than are disadvantaged groups,
especially if political HIs and public policies change suddenly to the disad-
vantage of the privileged group. Economically privileged groups may resist
redistributive measures.
Table 10.1 illustrates relationships between dimensions of HIs and
sociopolitical stability. As emphasised in the preceding section, the com-
bination of high socioeconomic HIs and high political HIs tends to cause
greater mistrust between groups, so that the stability will turn out to be
very low (as shown in the first row). Although people who are disadvan-
taged both economically and politically may be ‘accustomed’ to deep-seated
HIs and feel powerless (for example, war-weary Acholi in Northern Uganda),
it is undeniable that the concurrence of high HIs in both dimensions breeds
frustration among people, undermines social cohesion and augments con-
flict propensity of the country in the longer term (the historical experiences
in Rwanda, Burundi, Kenya and South Africa are examples).
Table 10.1 Horizontal inequalities and stability
Socioeconomic Political HIs Combination Likely sociopolitical

HIs of HIs situation
High High Consistently higha Prone to instability

High High Balancedb Can be stable
High Low, through Uneven Stable, but fragile
power-sharing/ because of underlying
decentralisation grievances
Low High Uneven Prone to instability
Low Low Even and low Most stable
a Consistent if same group is in higher (or lower) position in both socioeconomic and political
HIs.
b Balanced if one group is dominant in politics, another in socioeconomics, or conversely.
There can be cases in which the dimensions of deprivations among major

groups are not consistent (the second row). In this situation, a politically
dominant group may be relatively poor (similarly, an economically powerful
group may be politically marginalised). As mentioned, this asymmetric situ-
ation may produce a balance of power, but such equilibrium is fragile when
the extent of socioeconomic HIs is large as in Côte d’Ivoire after the settle-
ment of conflict in 2011. Taking into account the short-term effect of having
shared power, the outcome of this combination is regarded as ‘mixed’.
Political HIs can be reduced substantially in some forms of power-
dispersing institutions (the third row). In this category, though disparities in
material status between groups are still palpable, major groups are entitled
to broadly equal access to political power. However, those who lose power
under this arrangement may resent it, while the group(s) that are deprived
in material terms may also be unsatisfied in the medium or long term. Race
relations in post-apartheid South Africa provide an example.
When political power relations shift radically, groups previously in rela-
tively advantageous positions may be hostile towards those who gain power
and may take to violence to resist this loss. In spite of the need to reduce
political HIs, such reforms should be implemented with caution, taking into
account the possibility of harmful reactions of such groups. For this reason
and because a high degree of socioeconomic HIs remain, the sociopolitical
stability of this combination may be fragile.
There can be a combination of low socioeconomic HIs and high political
HIs, in which the material conditions of different groups are generally sim-
ilar, but political power is distributed unevenly among groups (the fourth
row). In such a situation, groups that are politically sidelined may wrongly
believe that their socioeconomic status is lower than that of other groups, or
than the ‘objective’ one, as is the case for the Ndebele in Zimbabwe. Without
an inclusive political framework, this situation can be highly politicised with
the perceived pattern of HIs coming closer to the first row combination. Thus
the outcome of this combination is regarded as unstable. As noted above, we

have found that the perceptions of political HIs and those of socioeconomic
HIs tend to converge, with the former conditioning the latter. Finally, the
combination of low socioeconomic HIs and low political HIs will contribute
to sociopolitical stability (the bottom row). The exemplary case of peaceful
ethnic relations in mainland Tanzania falls in this category.
In terms of conflict prevention, it should be remembered that coun-
tries with higher political HIs tend to be more unstable (the first and the
fourth rows), because political HIs are generally more visible than socioeco-
nomic HIs. While the transformation of socioeconomic HIs is a long-term
process, the political HIs can change overnight through new legislation,
constitutional amendments or a change of government after elections. This
illuminates both the significance of reducing political HIs for conflict pre-
vention and the possibility of so doing, but this may be fiercely resisted by
those who are asked to share power. Where political HIs are being addressed
(the third row), or where the country is in a situation of a fragile bal-
ance of power (the second row), it is important to take heed of the fear of
people who have lost or are losing some of their relative advantages. Lead-
ership which believes in national inclusivity, and impartial attitudes among
political leaders, may contribute to the reduction of such apprehensions.
In the seven countries where we conducted perceptions surveys, in general
people believe that their national identity is more important than their eth-
nic identity. This finding refutes the widespread view that the past attempts
at nation-building in African countries have been a complete failure. What
we find is that most people have multiple identities, including their eth-
nicity, religion, profession, gender as well as their nationality. For example,
in Kenya, the bulk of respondents (71 per cent) feel both an ethnic and
a national identity, 26 per cent feel only Kenyan and just 2 per cent feel
only ethnic (Figure 7.7). When the value of national inclusiveness is shared
by diverse peoples, this common view is expected to reduce resistance to
reforms designed to alleviate socioeconomic and political HIs.
10.3.3 Political institutions and policies

Pairs of neighbouring countries in our case studies, which share a similar
social structure, exhibit sharp contrasts in terms of stability and instability.
This highlights the important role played by political institutions and poli-
cies which in some cases exclude some groups and in other cases include
representatives of all major ethnic groups. There are many factors that bring
about distinctive institutions. In the historical process of evolution of insti-
tutions and policies, relatively small differences in initial conditions after
a time may lead to significant variance between two countries with similar
social settings. We may recall the dissimilarity in governing styles of king-
doms in pre-colonial Burundi and Rwanda, the disparity in the proportion of
the population who are Muslims in Ghana and Côte d’Ivoire, or the different
degrees of urbanisation in South Africa and Zimbabwe.
Our case studies confirm the importance of political systems that lead to
power dispersal in supporting stability. Such power-dispersal is of two types:
first, including all major groups in central government, or power-sharing at
the centre, through cooperative or coalition government (PS); and secondly,
decentralising power to regions and/or districts (DEC). Both appear to be
important. PS may be formal or informal. In Ghana, informal convention
ensures that in general all major groups participate in power. In Nigeria,
formal arrangements were introduced, as a response to the Biafran war, via
the Federal Character Principle, as well as informal arrangements – with the
convention that the presidency should alternate between North and South.
Formal PS is a feature of the Tanzanian, Burundian and South African con-
stitutions. In Kenya and Zimbabwe, coalition government was introduced
in the late 2000s in response to conflict, but has not been enshrined in
the countries’ constitutions. Rwanda, Uganda and Côte d’Ivoire currently
lack PS.
PS may reduce the likelihood of political conflict in the short run, but
if large socioeconomic HIs remain, there is a potential for group grievance
which could give rise to violent conflict at some point. That is especially the
case when PS is ad hoc, and comes to an end. Consequently, redistributive
socioeconomic policies are an important aspect of conflict prevention. Con-
versely, we should bear in mind that prolongation of ad hoc PS, especially
involving a coalition government of all major parties, may negatively affect
democratisation (as may be the case in Burundi and Zimbabwe), if such a
formal union is not supplemented by mutual trust and informal PS. It is pos-
sible, however, to combine democratic processes with PS arrangements, as in
Nigeria, Tanzania and South Africa. On the other hand, concentrating power
at the centre, based on majoritarian systems without PS, may be supported
in the belief that this will contribute to national unity (as has been argued,
for example, in Uganda and Rwanda), but this can then lead to grievances
among groups excluded from the central government, which may then have
negative effects on national unity. In such situations, close attention should
be paid to the nature of grievances which can simmer below the surface, as
well as to the democratic quality of regimes. Problems of this kind appear
in a number of our cases, including Rwanda, Côte d’Ivoire, Zimbabwe and
Uganda.
DEC, an important aspect of power-sharing in a broad sense, can be
achieved through a federal system, as in Nigeria, or by decentralisation
in a unitary state, as in Ghana and Uganda. In each case, the reality of
decentralisation – whether the arrangements involve genuine devolution of
decision-making and finance, or the centre has retained power – should be
investigated. For example, it is argued that in Uganda the centre has retained
power, despite the fact that there has been a large amount (30 per cent of
government expenditure) of finance devolved; in contrast, in Tanzania, it is

argued that decentralisation involves more real devolution of power, even
though much less finance is involved. Decentralisation which creates a large
number of local units can even be a tool for strengthening the central gov-
ernment through the maintenance of patronage networks (as is arguably the
case in Uganda, Kenya and Nigeria). Even if genuine, however, decentrali-
sation is unlikely to prevent conflict in itself, unless accompanied by PS at
the centre, formally or informally. DEC can lead to inequality between the
local units of government and needs to be accompanied by redistribution of
resources from the centre to avoid this. In Nigeria, the Federation Account
is an example of such a redistribution mechanism, although the revenue
allocation formula remains a matter of disputes among states.
In addition to the general advantages of PS and DEC, it is also important
to recognise that during its history, every country has developed its own
measures to manage ethnic diversity and prevent national disintegration.
Tenacious attempts to create hybrid institutions to contain ethnic conflict
are noticeable in some African countries (for example, Nigeria and Kenya).
While imposing institutions from above, based on a ready-made theoret-
ical blueprint, would not lead to coherence with local conditions nor to
‘ownership’ by local politicians, outsiders can contribute by emphasising the
importance of developing institutions which ensure political inclusion.
In order to obtain a general view of political institutions, we present a
comparison of the institutions of ten countries in Table 10.2. This matrix
combines both formal and informal features of political institutions, sup-
plemented by data on fiscal decentralisation. ‘Power-Sharing at Centre’ and
‘Power-Dispersal to Local Levels’ correspond to the two features of the
power-dispersing (PD) institutions discussed in Chapter 2. Table 10.3 pro-
vides a summary of relationships between institutional types and political
outcomes. While the classification of political institutions in Chapter 2 is
based only on the features of formal institutions, this table takes into consid-
eration the modalities of informal conventions and practices contextualised
in the case studies, as well as the information presented in Table 10.2. The
appraisal of political stability and instability in these tables is not necessarily
meant to evaluate the appropriateness of government policies, as they are
often constrained by historical legacy of the past conflicts.
Looking at the first row of Table 10.3, the combination of majoritarian pol-
itics and centralised power structures coincides with political instability, as
in the cases of the civil war in Côte d’Ivoire and the post-Cold War violence
in Rwanda. Some countries with majoritarian politics (for example, Uganda)
have introduced extensive DEC policies, but the processes can be politically
exploited by the centre, and thus have mixed outcomes.
In the late 2000s, a number of African countries abandoned the
majoritarian system and adopted ad hoc PS arrangements to assuage cumu-
lative political tensions, as shown in the second row of the table. The formal
Table 10.2 Modalities of national and national–local power configurations
Country Power-sharing at centre Power dispersal to lower levels Current political

tensions
Formal Informala Federalism Fiscal

decentralisation
% of locally raised
revenue in local
government revenueb
Rwanda X X X 17 (2010/2011) Politically stable with

potential tensions
Burundi + X X n.a. Tensions
Ghana X + X 31 (1997) Stable
Some degree
of local
autonomy
Côte d’Ivoire X X X n.a. Post-conflict peace
Varies; some attempts with tensions
at power-sharing
South Africa X + + (semi) 78 (2009/2010) Politically stable with
Grand coalition at the social tensions
time of the GNU
Zimbabwe + X X 65 (1997e) Tensions
Temporary coalition
adopted to solve crisis
267
268
Country Power-sharing at centre Power dispersal to lower levels Current political

tensions
Formal Informala Federalism Fiscal

decentralisation
% of locally raised
revenue in local
government revenueb
Tanzania + + + (semi) 7 (2009/10) Stable

Between Tanganyika
and Zanzibar
Uganda X X X 15 (2007/08) Conflict ended in the
North; Strong
regionalism
Kenya + + X 59 (2007/08) Stable since 2009
Temporary coalition conflict
adopted to solve crisis
Nigeria + + + 41 (2002) Stable at national level,
but tensions; some
conflict at local levels
Notes: + indicates that the characteristic is present and X indicates that characteristic is not present; number in parenthesis, e and n.a. indicate the year,
the estimate and the unavailability of the data respectively.
a. Difficult to assess.
b. Based on Commonwealth Local Government Forum Country Profiles, http://www.clgf.org.uk/country-profiles/ (last accessed 1 May 2013). For Zimbabwe, the
estimation is obtained from Dennys et al. (2000: 120, Table D9). Comparison of the sizes of local government expenditure is also of significance, but the
cross-country data is difficult to obtain.
Table 10.3 Categories of institutions and stability
Centralisation Decentralisation
Majoritarian Unstable Mixed

Ad hoc PS Unstable Mixed
Sustainable PS Stable Stable
power-sharing in Burundi, Zimbabwe and Kenya in the 2000s may have

been necessary to avoid the escalation of conflict, but such arrangements
do not lead automatically to long-term peace, which can only be realised
when the political will of antagonistic parties to cooperate is sufficiently
strong. Among these countries, Burundi and Zimbabwe are characterised by
centralisation of power, while Kenya is characterised by a set of DEC poli-
cies. Kenya may be moving from ad hoc to more sustainable PS with its
gradual implementation of the new Constitution to accommodate group
differences. However, it is too early to judge the overall effects of such
reforms on HIs in Kenya, and, besides, the outcomes of the International
Criminal Court (ICC) against people suspected of engineering the 2007/8
post-election violence may radically change future ethnic relations in Kenya
(Sing’Oei, 2010). Therefore, while the situations in Burundi and Zimbabwe
are viewed as unstable, we put Kenya into the mixed category.
All countries in the third row enjoy political stability, buttressed by inclu-
sive political leadership and strengthened by a combination of elements of
formal PS and a culture of informal PS. We call this ‘sustainable PS’. Even
when a country adopts a majoritarian electoral system, the country may
abound with formal rules and informal practices for power-sharing as is
the case in Ghana. In our case studies, the countries in this category range
from de jure unitary Ghana to semi-federal Tanzania, de facto federal South
Africa and de jure federal Nigeria. Federalism or semi-federalism seems to
be conducive to political stability and national integration, especially in the
presence of large geographically concentrated groups aspiring for autonomy.
According to Table 10.2, the degree of fiscal decentralisation varies in these
ten countries, although it is often argued that local autonomy can only be
realised when it is accompanied by financial autonomy. The share of locally
raised revenue in total local government revenue is relatively high in South
Africa, Kenya and Nigeria (and it was also high in Zimbabwe before the start
of the political crisis). However, Tanzania has achieved stability with a very
low level of fiscal decentralisation. An identity group may be satisfied if peo-
ple feel their voice is heard in central government even when the material
base for autonomy is limited.
It emerges from these concluding tables that a combination of PS
and DEC, power-dispersing institutions reinforced by informal practices,
provides favourable conditions for political stability. It should also be
emphasised that leadership styles matter for conflict prevention in the short
term. If such styles take root as long-term practices, they will contribute
greatly to bringing sustained stability to their countries. The different
personalities of Nelson Mandela and Robert Mugabe, for example, were
responsible for the differences in the PS and majoritarian paths taken by
South Africa and Zimbabwe.
10.3.4 Conflict processes and outside actors

It is evident that elections can be a major trigger for violent conflict (as
has occurred, for example, in Côte d’Ivoire, Kenya and Zimbabwe). In each
case, violence arose in the context of a winner-takes-all election, in which
the potential (or actual) losers were denied power, and then protested vio-
lently. For the purpose of conflict prevention, what matters, then, is not only
the fairness of election processes, which may be supported by technical aid
and international monitoring activities, but also the risk of violence from an
election system based on a winner-takes-all principle. The risk of violence is
particularly strong, when much power is vested in a president; a presidential
election can then become a moment for a violent split in the nation. Conse-
quently, for conflict prevention one may need to go beyond the requirement
for democratic processes, to support power-sharing democracies.
In navigating the country towards sustainable PS, the quality of political
leaders is of great significance. When a political leader is sensitive to inequal-
ities, and is determined and able to take measures to address persistent HIs,
this contributes to the stability of the country. Examples in our case stud-
ies include some of the history of Ghana and Tanzania, in particular. It is
not only important that political leaders adopt inclusive policies towards
identity groups, but that they also implement culturally inclusive policies.
Examples of where this did occur include Ghana, South Africa, Nigeria and
Tanzania. In contrast, political leaders’ favouritism towards their own eth-
nic group can trigger violent conflicts. Therefore, it is very important to
encourage those in power to act inclusively and impartially. A political leader
can also take a lead in promoting national unity and national identities, as
against ethnic loyalties. Nkrumah in Ghana and Nyerere in Tanzania were
famously successful in this regard. Nation-building policy can be applied in
various ways including education, language policy and purposeful use of the
mass media (Miguel, 2004).
Africa has historically been a continent of large-scale migration and soci-
etal fluidity. In the era of democratic politics, the issue of ‘who the citizens
are’ and who has a vote is becoming prominent in some African countries
(Geschiere, 2009; Manby, 2009). Depriving particular groups of citizen-
ship rights can cause violent conflicts (as happened, for example, in Côte
d’Ivoire, South Africa and Zimbabwe). In Côte d’Ivoire, long-term immi-
grants or descendants of immigrants were deemed to be foreigners and
such exclusion from citizenship rights was an important element trigger-

ing the armed conflict. In South Africa and Zimbabwe, the black population
was regarded as second-class citizens under the racial segregation regime
without full citizenship rights. In Nigeria, local citizenship differentiating
between settlers and indigenes is a major source of local violence. Migra-
tion, border trade and conflicts in neighbouring countries may also affect
conflict paths (this was the case, for example, in Rwanda, Zimbabwe and
South Africa). Many skilled and qualified Zimbabweans, the majority being
Ndebele, together with Mozambicans and other Africans emigrated to South
Africa from the 1990s. This created new HIs in South Africa and resulted in
the outbreak of xenophobic violence there in 2008.
In principle, policies to prevent conflict are most effectively implemented
by national governments as the legitimate units of governance. How-
ever, outside actors, including donor agencies, may also play constructive
roles. The minimum requirement is to uphold the principle of Primum
non nocere (First, do no harm), a motto familiar to medical practitioners
around the world. While this principle is thought much of by practition-
ers of peace-building in conflict areas, it must also be taken seriously by
development practitioners. When it is foreseen that a certain form of inter-
vention by development agencies, whether through project aid or through
general budgetary support, would lead to aggravation of some aspects of
multidimensional HIs, such a practice should be precluded.
Wherever possible, however, development agencies should go beyond that
principle and should be alert and responsive to significant downturns in
social, economic and political trends of conflict-prone countries (Ogata,
2010). They should also aim to support measures to reduce HIs more explic-
itly by addressing the root causes of violent conflict. This means that they
should guide their own projects towards less well-off groups, including
‘socially excluded’ groups of population. Agencies may also wish to intro-
duce the principle of equity among groups in dialogues with governments
about socioeconomic development. However, there may be instances when
the government itself appears to be accentuating HIs, and favouring par-
ticular identity groups in politically provocative and morally unjustifiable
ways. One approach to such a case is to address the grievances of disad-
vantaged groups directly and/or to provide assistance through civil society
organisations. Such assistance should be given with sensitivity, so as not to
provoke reactions that derail the processes of democratisation and equitable
development.
Although the chapters in this book have endorsed the general advantages
of power-dispersing, inclusive institutions, outside actors should refrain
from imposing any one particular type of political system in a top-down
manner because they may not know the most appropriate institutions in
local contexts and also because the shape of national institutions should
be chosen by people themselves. However, mutual learning about the
comparative advantages of alternative democratic institutions should never

be neglected, given that the question of the institutional distribution of
political power is relevant to developing and developed nations alike, and
that the experiences of every country abound with practical lessons. In addi-
tion, professionals knowledgeable about legal system reform can make
technical contributions to the process of institution-building in post-conflict
societies (Sannerholm, 2007). When warring parties seek power-sharing
arrangements to avoid the intensification of conflict, some forms of for-
eign mediation may become necessary to support the peace process. In such
cases, a local, regional or international group of wise people, or regional
organisations, are likely to play more constructive roles than big powers,
as demonstrated in recent experiences of ad hoc power-sharing in Burundi,
Kenya and Zimbabwe.
10.4 Conclusion
In closing, the major messages of this book may be put as follows. First
of all, we emphasise the necessity of mainstreaming the HI perspective in
development policy-making in Africa. As confirmed in the whole of our
study, serious HIs, political, socioeconomic and cultural, do exist in African
countries, and the HI perspective provides one (although not the only) sig-
nificant explanation of the cause of violent conflict between groups. The
recent growth of African economies is improving the absolute situation
for many people but we still need to monitor the quality of develop-
ment closely, paying due attention to shifting patterns and intensities of
multidimensional HIs. Secondly, ‘subjectivity’ matters, as demonstrated by
the quantitative survey analysis. People take actions not because of objec-
tive data on structural inequalities but because of their subjective, sometimes
emotional perceptions about such inequalities, the nature of which must be
studied regularly. Thirdly, the political dimension of HIs is of prime impor-
tance, especially given the post-Cold War fluidity of politics and political
institutions in Africa. It is interesting to note that our research suggests that
people’s perceptions of political HIs may override those of socioeconomic
HIs. Fourthly, the case studies of paired countries as well as the cross-country
comparison of political institutions have confirmed the general advantage of
power-dispersing institutions, namely, a combination of sustainable PS and
DEC institutions, in conflict resolution and prevention.
Fifthly, the significance of informal institutions also emerged in the
case studies. Inclusive leadership and informal power-sharing have played
important roles in preventing violent conflicts in all the stable countries
investigated. The case studies have also shown that, depending on historical
contexts, informal values and practices strengthen, neutralise or sometimes
betray the intentions of formal reforms. Sixthly, all case studies indicate that
the focus on the specific characteristics of each locality is critically impor-

tant. Paired country studies have shown that relatively small differences in
the initial conditions sometimes result in big differences in the outcome,
and such processes can be traced by taking fully into account idiosyncratic,
historical traits of each society. In terms of conflict prevention, we have
noticed that hybrid institutions to assuage inter-group antagonism are tak-
ing shapes in some ethnically heterogeneous countries in Africa. We can
learn much from these endogenous experiences.
Finally, we conclude this chapter with the identification of three major
areas for policy, and ten policy recommendations for development agencies,
governments of African countries and other actors, as follows:
[1] HIs are major root causes of violent conflict. The perspective of
multidimensional HIs should be mainstreamed in development policy-
making in Africa.
1. African governments and development agencies should systemati-

cally collect data on HIs and promote the study of their size and
origin.
2. Development projects and government budgets should be thor-
oughly reviewed to ensure that they contribute to alleviate exist-
ing HIs.
3. Officers or advisors should be appointed in major development agen-
cies to monitor the state of all inequalities (including vertical as well
as horizontal inequalities) and their causes and to evaluate projects
and policies in the light of their effects on inequalities and in relation
to conflict prevention.
[2] Objective and subjective HIs are not always in accord. Development
practitioners and government officials should take care to avoid pro-
moting misperceptions concerning HIs and their own fairness and
impartiality.
1. Development agencies and their counterparts should pay attention to

the regional distribution of projects to avoid a perception that certain
regions are neglected.
2. They should pay attention to the ethnic representation of those
who participate in development planning to avoid a perception that
certain groups are excluded unfairly.
3. They should give support to disadvantaged groups without alienating
the others who may think they are worse off than they really are.
4. Political HIs tend to distort people’s perceptions of socioeconomic
HIs. The nature of political and social transformations of countries
after violent conflict should be carefully observed by all stakeholders.
[3] Inclusive institutions, both formal and informal, are the foundation
of peace. Outside actors should respect local initiatives for sustainable
power-sharing and effective decentralisation.
1. In the face of escalation of violence, warring parties are encouraged
to form power-sharing arrangements with the support of regional
organisations and thereby alleviate inter-group animosities.
2. Development agencies should continue to assist efforts towards
decentralisation so as to enhance power-sharing at the local level,
while ensuring that there are redistributive measures to support local
units across regions and localities.
3. Leadership quality is of prime importance. Opportunities for mutual
learning should be provided for African young leaders to embed in
them a culture of conflict prevention and inclusive nation-building.
References
Acemoglu, D., Johnson, S., and Robinson, J.A. 2001. The colonial origins of
comparative development: An empirical investigation. The American Economic
Review 91(5): 1369–1401.
Adebajo, A. 2008. Hegemony on a shoestring: Nigeria’s post-cold war foreign policy.
In Adebayo, A., and Mustapha, A.R., eds, Gulliver’s Troubles: Nigeria’s Foreign Policy
after the Cold War. Scottsville: University of KwaZulu-Natal Press.
Agyeman, D.K. 1998. Ethnic conflicts and politics in Ghana. In Nnoli, O., ed., Ethnic
Conflicts in Africa. Dakar: CODESRIA.
Akindès, F. 2004. Racines des crises socio-politiques en Côte d’Ivoire et sens de
l’histoire. Identity, Security and the Negotiation of National Belonging in West Africa:
Reflections on the Crisis in Côte d’Ivoire. CODESRIA and Nordiska Afrika Institut
Conference, Dakar, May 15–16.
Alesina, A., Devleeschauwer, A., Easterly, W., Kurlat, S., and Wacziarg, R. 2003.
Fractionalization. Journal of Economic Growth 8(2): 155–194.
Alexander, J., McGregor, J.A., and Ranger, T. 2000a. Ethnicity and the politics of con-
flict: The case of Matabeleland. In Nafziger, W.E., and Stewart, F., eds, War, Hunger,
and Displacement: The Origins of Humanitarian Emergencies, Vol. 1: Analysis. Oxford:
Oxford University Press.
Alexander, J., McGregor, J.A., and Ranger, T. 2000b. Violence and Memory: One Hundred
Years in the ‘Dark Forests’ of Matabeleland. Oxford: James Currey.
Alston, P. 2009. Press statement by Prof. Philip Alston. UN Special Rapporteur on
Extrajudicial, Arbitrary or Summary Executions, Mission to Kenya. New York: United
Nations.
Alwy, A., and Schech, S. 2004. Ethnic inequalities in education in Kenya. International
Education Journal 5(2): 266–274.
Amnesty International. 1997. Rwanda: Ending the Silence, No. AFR 47/032/1997, http://
www.amnesty.org/en/library/info/AFR47/032/1997 (last accessed 1 May 2013).
Anderson, B. 1983. Imagined Communities: Reflections on the Origin and Spread of
Nationalism. London: Verso.
Ansoms, A. 2008. Striving for growth, bypassing the poor? A critical review of
Rwanda’s rural sector policies. Journal of Modern African Studies 46(1): 1–32.
Ansoms, A. 2009. Re-engineering rural society: The visions and ambitions of the
Rwandan elite. African Affairs 108(431): 289–309.
Anyinam, C. 1993. Spatial implications of structural adjustment programmes in
Ghana. Tijdschrift voor Economische and Sociale Geografie 85(5): 446–460.
Apter, D.E. 1961. The Political Kingdom in Uganda: A Study in Bureaucratic Nationalism.
Princeton, NJ: Princeton University Press.
Azam, J.-P. 2004. Poverty and growth in the WAEMU after the 1994 devaluation.
Journal of African Economies 13(4): 536–562.
Bach, D.C. 2004. Inching towards a country without a state: Prebendalism, violence
and state betrayal in Nigeria. In Clapham, C., Herbst, J., and Mills, G., eds, Big
African States. Johannesburg: Witwatersrand University Press.
Bakari, M.A. 2001. Democratization Process in Zanzibar: A Retarded Transition, Hamburg
African Studies 11. Hamburg African Studies 11. Hamburg: Hamburg University Press.
275
276 References
Bakary, T. 1984. Elite transformation and political succession. In Zartman, W.I., and
Delgado, C., eds, The Political Economy of Ivory Coast. New York: Praeger Publishers.
Banégas, R., and Losch, B. 2002. La Côte d’Ivoire au bord de l’implosion. Politique
Africaine 87: 139–161.
Bangura, Y. 2006. Introduction: Ethnic inequalities and public sector governance.
In Bangura, Y., ed., Ethnic Inequalities and Public Sector Governance. Basingstoke:
Palgrave Macmillan.
Baskin, J. 1991. Striking Back: A History of COSATU. Johannesburg: Ravan Press.
Bates, R. 2004. Ethnicity. In Clark, D., ed., The Elgar Companion to Development Studies.
Cheltenham: Edward Elgar Publishing.
Beach, D. 1994. The Shona and Their Neighbours. Oxford: Blackwell.
Behrend, H. 1999. Alice Lakwena and the Holy Spirits: War in Northern Uganda, 1986–
1997. Oxford: James Currey.
Bentley, K.A., and Southall, R. 2005. An African Peace Process: Mandela, South Africa,
and Burundi. Cape Town: HSRC Press.
Beswick, D. 2010. Managing dissent in a post-genocide environment: The challenge
of political space in Rwanda. Development and Change 41(2): 225–251.
Blum, S. 2006. East Africa: Cycles of violence, and the paradox of the peace. ZEF
Discussion Paper 107. Bonn: Center for Development Research, University of Bonn.
Boahen, A.A. 1992. The Ghanaian Sphinx: Reflections on the Contemporary History of
Ghana, 1972–1987. Accra: Sankofa Educational Publishers.
Bond, P., and Manyanya, M. 2002. Zimbabwe’s Plunge: Exhausted Nationalism,
Neoliberalism and the Search for Social Justice. Pietermaritzburg: University of Natal
Press.
Booysen, S., and Toulou, L. 2009. Zimbabwe. In Denis, K., and Booysen, S., eds,
Compendium of Elections in Southern Africa, 1989–2009. Johannesburg: EISA.
Bourne, R. 2011. Catastrophe: What Went Wrong in Zimbabwe? London: Zed Books.
Bresson, Y. 1980. Politique économique et revenus en Côte d’Ivoire. World Employment
Programme Research Working Papers, WEP 2-23/WP 95. Geneva: ILO (International
Labour Office).
Brown, D. 1982. Who are the tribalists? Social pluralism and political ideology in
Ghana. African Affairs 81(322): 37–69.
Brown, G.K. 2005. Balancing the risks of corrective surgery: The political economy of
horizontal inequalities and the end of the new economic policy in Malaysia. CRISE
Working Paper No. 20. Oxford: CRISE (Centre for Research on Inequality, Human
Security and Ethnicity), University of Oxford.
Brown, G.K. 2008. Horizontal inequalities and separatism in Southeast Asia: A compar-
ative perspective. In Stewart, F., ed., Horizontal Inequalities and Conflict: Understanding
Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan.
Brown, G.K., and Langer, A. 2010. Horizontal inequality and conflict: A critical review
and research agenda. Conflict, Security & Development 10(1): 27–55.
Bruce, J.W. 2009. International standards, improvisation and the role of interna-
tional humanitarian organizations in the return of land in post-conflict Rwanda.
In Pantuliano, S., ed., Uncharted Territory: Land, Conflict and Humanitarian Action.
Warwickshire: Practical Action Publishing.
Brukum, N.J.K. 1995. Ethnic conflicts in the northern region of Ghana: A study of the
conflict in the Gonja district, 1980–1994. In Oquaye, M., ed., Democracy and Conflict
Resolution in Ghana. Accra: Gold Type Press.
Bundy, C. 1979. The Rise and Fall of the South African Peasantry. Berkeley, CA: University
of California Press.
References 277
Burgess, T. 2009. Race, Revolution, and the Struggle for Human Rights in Zanzibar. Athens,
OH: Ohio University Press.
Burnet, J.E. 2008. Gender balance and the meanings of women in governance in post-
genocide Rwanda. African Affairs 107(428): 361–386.
CCJP/LRF (Catholic Commission for Justice and Peace in Zimbabwe and the Legal
Resources Foundation). 2008. Gukurahundi in Zimbabwe: A Report on the Disturbances
in Matabeleland and the Midlands, 1980/1988. New York: Columbia University Press.
Cederman, L.A., Gledditsch, K.S., and Weidmann, N.B. 2010. Horizontal inequal-
ities and ethno-nationalist civil war: A global comparison. Paper prepared for
presentation at Yale University, April.
Cederman, L.A., Gledditsch, K.S., and Weidmann, N.B. 2011. Horizontal inequali-
ties and ethno-nationalist civil war: A global comparison. American Political Science
Review 105(3): 478–495.
CENI (Commission Electorale Nationale Indépendante). 2010a. Elections législatives
2010: Liste nominative des élus. Bujumbura: CENI.
CENI (Commission Electorale Nationale Independante). 2010b. Tableau d’affectation
des administrateurs communaux aux fins des équilibres ethnique et de genre. Bujumbura:
CENI.
Chachage, S.L.S. 2000. Environment, Aid and Politics in Zanzibar. Dar es Salaam: Dar es
Salaam University Press.
Chitere, P., Chweya, L., Masya, J., Tostensen, A., and Waiganjo, K. 2006. Kenya
Constitutional Documents: A Comparative Analysis. Berge: Chr. Michelsen Institute.
Choudhry, S., ed., 2008. Constitutional Design for Divided Societies: Integration or
Accommodation? Oxford: Oxford University Press.
Chrétien, J.P. 2000. L’Afrique des Grands Lacs: Deux mille ans d’histoire. Paris: Aubier.
Chrétien, J.P., and Dupaquier, J.F. 2007. Burundi 1972 au bord des génocides. Paris:
Karthala.
Chung, F. 1988. Education: Revolution or reform. In Stoneman, C., ed., Zimbabwe’s
Prospects: Issues of Race, Class, State, and Capital in Southern Africa. London:
Macmillan.
CIA (Central Intelligence Agency). 2011. The CIA World Factbook 2012. New York:
Skyhorse Publishing.
CIPEV (Commission of Inquiry into the Post Election Violence). 2008. Report of the
Commission of Inquiry into the Post-Election Violence in Kenya (Waki Report). Nairobi:
CIPEV.
Cohen, M.A. 1974. Urban Policy and Political Conflict in Africa: The Study of the Ivory
Coast. Chicago, IL: The University of Chicago Press.
Coleman, J. 1958. Nigeria: Background to Nationalism. Berkeley, CA: University of
California Press.
Collier, P. 2009. Wars, Guns, and Votes: Democracy in Dangerous Places. New York:
HarperCollins.
Collier, P., Elliott, V.L., Hegre, H., Hoeffler, A., Reynal-Querol, M., and Sambanis,
N. 2003. Breaking the Conflict Trap: Civil War and Development Policy. Washington,
DC: World Bank.
Collier, P., and Hoeffler, A. 2004. The challenge of reducing the global incidence of
civil war. Copenhagen Consensus Challenge Paper, http://www.copenhagenconsensus.
com/sites/default/files/CP%2B-%2BConflicts%2BFINISHED.pdf (last accessed 1 May
2013).
Collier, P., Hoeffler, A., and Söderbom, M. 2008. Post-conflict risks. Journal of Peace
Research 45(4): 461–478.
278 References
Connah, G. 2001. African Civilizations: An Archaeological Perspective. 2nd edn.

Cambridge: Cambridge University Press.
Contamin, B., and Losch, B. 2000. Côte d’Ivoire: La voie étroite, Politique Africaine 77:
117–128.
Cottrell, J., and Ghai, Y. 2007. Constitution making and democratization in Kenya
(2000–2005). Democratization 14(1): 1–25.
Crook, R. 1997. Winning coalitions and ethno-regional politics: The failure of the
opposition in the 1990 and 1995 elections in Côte d’Ivoire. African Affairs 96(383):
215–242.
CSO (Central Statistical Office) and Macro International. 2007. Zimbabwe Demographic
and Health Survey, 2005–2006. Harare: CSO.
Dahl, R. 1971. Polyarchy: Participation and Opposition. New Haven, CT: Yale University
Press.
Dahl, R. 1998. On Democracy. New Haven, CT: Yale University Press.
Dashwood, H.S. 2000. Zimbabwe: The Political Economy of Transformation. Toronto:
University of Toronto Press.
Davies, R. 2004. Memories of underdevelopment: A personal interpretation of
Zimbabwe’s economic decline. In Raftopoulos, B., and Savage, T., eds, Zimbabwe:
Injustice and Political Reconciliation. Harare: Weaver Press.
Dembélé, O. 2003. Côte d’Ivoire: La fracture communautaire. Politique Africaine 89:
34–48.
Den Tuinder, B. 1978. Ivory Coast: The Challenge of Success: Report of a Mission Sent
to the Ivory Coast by the World Bank. Baltimore, MD: Johns Hopkins University
Press.
Dennys P., Steffensen, J., Trollegaard, S., Martin, R., Khumalo, C., Chando, V., and
Ncube, D. 2000. Fiscal Decentralisation and Sub-National Government Finance in
Relation to Infrastructure and Service Provision in Zimbabwe, Final Report, http://www.
consultpsi.com/Publications/Zimbabwe%20Fiscal%20Decentralisation%20Report.
pdf (last accessed 1 May 2013).
D’Hertefelt, M., Trouwborst, A.A., and Scherer, J.H. 1962. Les anciens royaumes de
la zone interlacustre méridionale: Rwanda, Burundi, Buha. Tervuren: Musée Royal de
l’Afrique Centrale.
Diamond, L., and Plattner, M.F., eds, 2006. Electoral Systems and Democracy. Baltimore,
MD: The Johns Hopkins University Press.
Doom, R., and Vlassenroot, K. 1999. Kony’s message: A new Koine? The Lord’s
Resistance Army in Northern Uganda. African Affairs 98(390): 5–36.
Dorsey, L. 1994. Historical Dictionary of Rwanda. Metuchen: The Scarecrow Press.
Dorsey, M. 2000. Violence and power-building in post-genocide Rwanda. In Doom,
R., and Gorus, J., eds, Politics of Identity and Economics of Conflict in the Great Lakes
Region. Brussels: VUB University Press.
Dozon, J.-P. 2000. La Côte d’Ivoire entre démocratie, nationalisme et ethnonationa-
lisme. Politique Africaine 78: 45–62.
Drumbl, M.A. 2007. Atrocity, Punishment, and International Law. Cambridge: Cambridge
University Press.
Durkheim, E. 1960. Les règles de la méthode sociologique. Paris: Presses Universitaires de
France.
Edigheji, O., ed., 2010. Constructing a Democratic Developmental State in South Africa:
Potentials and Challenges. Cape Town: HSRC Press.
Eggers, E.K. 1997. Historical Dictionary of Burundi. 2nd edn. Lanham, MD: The
Scarecrow Press.
References 279
Electoral Task Team. 2003. Report of the Electoral Task Team. Pretoria: Govern-
ment Printer, http://www.elections.org.za/content/DynamicDocs.aspx?id= 329&
BreadCrumbId= 329&LeftMenuId= 251&name= home (last accessed 1 May 2013).
Ellis, S., and Sechaba, T. 1992. Comrades against Apartheid: The ANC and the South
African Communist Party in Exile. London: James Currey.
Englebert, P. 2002. Born-again Buganda or the limits of traditional resurgence in
Africa. Journal of Modern African Studies 40(3): 345–368.
Englebert, P., and Hummel, R. 2005. Institutionalising ethnic representation: How
effective is the Federal Character Commission in Nigeria? African Affairs 104(416):
399–427.
Evans-Pritchard, E.E. 1940. The Nuer: A Description of the Modes of Livelihood and
Political Institutions of a Nilotic People. Oxford: Clarendon Press.
Fearon, J.D., and Laitin, D.D. 1996. Explaining interethnic co-operation. American
Political Science Review 90(4): 715–735.
Figueroa, A. 2006. Education, labour markets and inequality in Peru. Paper presented
at the CRISE Latin America Workshop, Santa Cruz, Bolivia, September 18–20.
Fjeldstad, O.H., and Therkildsen, O. 2008. Mass taxation and state-society relations
in East Africa. In Bräutigam, D., Fjeldstad, O.H., and Moore, M., eds, Taxation and
State-Building in Developing Countries. Cambridge: Cambridge University Press.
Frempong, K.D. 2001. Ghana’s election 2000: The ethnic under current. In Ayee J.,
ed., Deepening Democracy in Ghana: Politics of the 2000 Elections. Accra: Freedom
Publications.
Freund, B. 2007. South Africa: The end of Apartheid and the emergence of the ‘BEE
elite’. Review of African Political Economy 34(114): 661–678.
Friedman, S. 2009. An accidental advance? South Africa’s 2009 elections. Journal of
Democracy 20(4): 108–122.
Fukuda-Parr, S., Ashwill, M., Messineo, C., and Chiappa, E. 2008. Conflict-
development nexus: Survey of armed conflicts in sub-Saharan Africa, 1998–2006.
Journal of Peacebuilding and Development 4(1): 6–19.
Furnivall, J.S. 1948. Colonial Policy and Practice: A Comparative Study of Burma and
Netherlands India. Cambridge: Cambridge University Press.
Gaffney, P.D. 2000. Burundi: The long sombre shadow of ethnic instability.
In Nafziger, E.W., Stewart, F., and Väyrynen, R., eds, War, Hunger, and Displacement:
The Origin of Humanitarian Emergencies, Vol. 2. Oxford: Oxford University Press.
Gahama, J. 2001. Le Burundi sous administration belge: La période du mandat, 1919–1939.
Deuxième édition. Paris: Karthala.
Geschiere, P. 2009. The Perils of Belonging: Autochthony, Citizenship, and Exclusion in
Africa and Europe. Chicago, IL: University of Chicago Press.
Glazer, N., Moynihan, D.P., and Schelling, C.S. 1976. Ethnicity: Theory and Experience.
Cambridge, MA: Harvard University Press.
Glickman, H. 1998. Ethnicity, elections and constitutional democracy in Africa.
In Sisk, T.D., and Reynolds, A., eds, Elections and Conflict Management in Africa.
Washington, DC: United States Institute of Peace Press.
Global Witness. 2011. Congo’s Minerals Trade in the Balance: Opportunities and Obstacles
to Demilitarisation. London: Global Witness.
Gonin, G. 1998. Ethnicity, politics and national awareness in Côte d’Ivoire. In Nnoli,
O., ed., Ethnic Conflicts in Africa. Dakar: CODESRIA.
Government of Ghana. 2003. Ghana Poverty Reduction Strategy (GPRS), 2003–2005:
An Agenda for Growth and Prosperity. Accra: Government of Ghana.
280 References
Green, E. 2006. Ethnicity and the politics of land tenure reform in Central Uganda.
Commonwealth & Comparative Politics 44(3): 370–388.
Green, E. 2008. Decentralisation and conflict in Uganda. Conflict, Security & Develop-
ment 8(4): 427–450.
Gurr, T.R. 1970. Why Men Rebel. Princeton, NJ: Princeton University Press.
Gurr, T.R. 1993. Minorities at Risk: A Global View of Ethnopolitical Conflicts. Washington,
DC: United States Institute of Peace Press.
Gurr, T.R. 1994. Peoples against states: Ethnopolitical conflict and the changing world
system. International Studies Quarterly 38(3): 347–377.
Gurr, T.R. 2002. Peoples versus States: Minorities at Risk in the New Century. Washington,
DC: United States Institute of Peace Press.
Gyimah-Boadi, E. 2003. Ghana: The political economy of ‘successful’ ethno-regional
conflict management. In Bastian, S., and Luckham, R., eds, Can Democracy be
Designed? The Politics of Institutional Choice in Conflict-torn Societies. London: Zed
Publications.
Gyimah-Boadi, E., and Asante, R. 2006. Ethnic structure, inequality and public sec-
tor governance in Ghana. In Bangura, Y., ed., Ethnic Inequalities and Public Sector
Governance. Basingstoke: Palgrave Macmillan.
Gyimah-Boadi, E., and Daddieh, C. 1999. Economic reform and political liberalization
in Ghana and Côte d’Ivoire: A preliminary assessment of implications for nation
building. In Mengisteab, K., and Daddieh, C., eds, State Building and Democratization
in Africa: Faith, Hope and Realities. Westport, CT: Preager Publishers.
Hagan, G.P. 1992. Nkrumah’s cultural policy. In Arhin, K., ed., The Life and Work of
Kwame Nkrumah. Trenton: Africa World Press.
Hartzell, C.A., and Hoddie, M. 2007. Crafting Peace: Power-sharing Institutions and the
Negotiated Settlement of Civil Wars. University Park, PA: Pennsylvania State University
Press.
Hassim, S., Kupe, T., and Worby, E., eds, 2008. Go Home or Die Here: Violence, Xenopho-
bia and the Reinvention of Difference in South Africa. Johannesburg: Wits University
Press.
Herbst, J. 1990. State Politics in Zimbabwe. Berkeley, CA: University of California
Press.
Herbst, J. 2000. States and Power in Africa: Comparative Lessons in Authority and Control.
Hinderink, J., and Tempelman, G.J. 1979. Development policy and development prac-
tice in Ivory Coast: A miracle or a mirage? Diskussiestukken van de Vakgroep sociale
geografie van ontwikkelingslanden 4. University of Utrecht.
Hintjens, H. 2008. Post-genocide identity politics in Rwanda. Ethnicities 8(1): 5–41.
Hirschman, A.O. 1967. Development Projects Observed. Washington, DC: Brookings
Institution.
Holborn, L. 2010. The Long Shadow of Apartheid: Race in South Africa since 1994.
Johannesburg: South African Institute of Race Relations.
Horowitz, D.L. 1985. Ethnic Groups in Conflict. Berkeley, CA: University of California
Press.
Horowitz, D.L. 1991. A Democratic South Africa? Constitutional Engineering in a Divided
Society. Berkeley, CA: University of California Press.
Horowitz, D.L. 2002. Constitutional design: Proposals versus processes. In Reynolds,
A., ed., The Architecture of Democracy: Constitutional Design, Conflict Management, and
Democracy. Oxford: Oxford University Press.
References 281
Huggins, C. 2009. Land in return, reintegration and recovery processes: Some lessons
from the Great Lakes region in Africa. In Pantuliano, S., ed., Uncharted Territory:
Land, Conflict and Humanitarian Action. London: Practical Action Publishing.
Human Rights Watch. 2008. Law and Reality: Progress in Judicial Reform in Rwanda.
New York: Human Rights Watch.
Human Rights Watch. 2010. Burundi: Closing Doors? The Narrowing of Democratic Space
in Burundi. New York: Human Rights Watch.
Hutchful, E. 1973. Military disengagement in Ghana. Unpublished Ph.D. Thesis.
University of Toronto.
ICG (International Crisis Group). 2003. Réfugiés et déplacés au Burundi: Désamorcer
la bombe foncière. Rapport Afrique 70. Bruxelles: ICG.
ICG (International Crisis Group). 2006. Burundi: Democracy and peace at risk. Africa
Report 120. Bruxelles: ICG.
ICG (International Crisis Group). 2010. Burundi: Garantir un processus électoral
crédible. Rapport Afrique 155. Bruxelles: ICG.
ICG (International Crisis Group). 2011. Burundi: Du boycott électoral à l’impasse
politique. Rapport Afrique 169. Bruxelles: ICG.
Iheduru, O.C. 2004. Black economic power and nation-building in post-Apartheid
South Africa. Journal of Modern African Studies 42(1): 1–30.
Iliffe, J. 2007. Africans: The History of a Continent. 2nd edn. Cambridge: Cambridge
University Press.
Ingelaere, B. 2009. Does the truth pass across the fire without burning? Locating
the short circuit in Rwanda’s Gacaca courts. Journal of Modern African Studies 47(4):
507–528.
Ingelaere, B. 2010. Peasants, power and ethnicity: A bottom-up perspective on
Rwanda’s political transition. African Affairs 109(435): 273–292.
ISTEEBU (Institut de Statistiques et d’Etudes Economiques du Burundi). 2009. Annuaire
statistique du Burundi 2007. Bujumbura: ISTEEBU.
Kanyenze, G. 2003. The performance of the Zimbabwean economy, 1980–2000. In
Darnolf, S., and Laakso, L., eds, Twenty Years of Independence in Zimbabwe: From
Liberation to Authoritarianism. Basingstoke: Palgrave Macmillan.
Kanyinga, K. 2006. Ethnic structure, inequality and public sector governance
in Kenya. In Bangura, Y., ed., Ethnic Inequalities and Public Sector Governance.
Basingstoke: Palgrave Macmillan.
Kanyinga, K. 2007. Governance institutions and inequality. In SID (Society for
International Development), Readings on Inequality in Kenya. Nairobi: SID.
Kasfir, N. 1998. ‘No-party democracy’ in Uganda. Journal of Democracy 13(4): 49–63.
Kibara, G. 2003. The state of constitutionalism in Kenya, 2003, http://www.
kituochakatiba.org/index2.php?option= com_docman&task= doc_view&gid= 21&
Itemid= 36 (last accessed 1 May 2013).
Kieffer, G.-A. 2000. Armée ivoirienne: le refus du déclassement. Politique Africaine 78:
26–44.
Killian, B. 2008. The state and identity politics in Zanzibar: Challenges to democratic
consolidation in Tanzania. African Identities 6(2): 99–125.
Kimenyi, M.S., and Ndung’u, N.S. 2005. Sporadic ethnic violence: Why has Kenya not
experienced a full-blown war? In Collier, P., and Sambanis, N., eds, Understanding
Civil War: Evidence and Analysis, Vol. 1: Africa. Washington, DC: World Bank.
Kimenyi, M.S., and Shughart, W.F. 1989. Political successions and the growth of
government. Public Choice 62(2): 173–179.
Klein, N. 2007. The Shock Doctrine: The Rise of Disaster Capitalism. New York: Picador.
282 References
Klugman, J. 2000. Kenya: Economic decline and ethnic politics. In Nafziger, E.W.,
Stewart, F., and Väyrynen, R., eds, War, Hunger, and Displacement: The Origins of
Humanitarian Emergencies, Vol. 2: Case Studies. New York: Oxford University Press.
Klugman, J., Neyapti, B., and Stewart, F. 1999. Conflict and Growth in Africa, Vol. 2:
Kenya, Tanzania and Uganda. Paris: OECD Development Centre.
KNBS (Kenya National Bureau of Statistics). 2010. Kenya 2009 Population and Housing
Census, http://www.knbs.or.ke/population.php (last accessed 1 May 2013).
KNBS (Kenya National Bureau of Statistics) and ICF Macro. 2010. Kenya Demographic
and Health Survey (DHS), 2008–09. Calverton, MD: KNBS and ICF Macro, http://
www.measuredhs.com/pubs/pdf/FR229/FR229.pdf (last accessed 1 May 2013).
Kniss, M. 2010. Walking Kenya back from the brink: A micro-level study of horizontal
inequity and civil conflict prevention. Center for International and Security Studies at
Maryland (CISSM) Working Paper June 2010. College Park, MD: CISSM, University of
Maryland.
Kriger, N. 2003. Guerrilla Veterans in Post-War Zimbabwe: Symbolic and Violent Politics,
1980–1987. Cambridge: Cambridge University Press.
Kymlicka, W. 1995. Multicultural Citizenship: A Liberal Theory of Minority Rights. Oxford:
Clarendon Press.
Laakso, L. 2003. Regional voting and cabinet formation. In Darnolf, S., and Laakso,
L., eds, Twenty Years of Independence in Zimbabwe: From Liberation to Authoritarianism.
Laakso, M., and Taagepera, R. 1979. Effective number of parties: A measure with
application to West Europe. Comparative Political Studies 12(1): 3–27.
Laely, T. 1992. Le destin du bushingantahe transformations d’une structure locale
d’autorité au Burundi. Genève-Afrique 30(2): 75–98.
Langer, A. 2005. Horizontal inequalities and violent group mobilisation in Côte
d’Ivoire. Oxford Development Studies 33(1): 25–45.
Langer, A. 2008. Horizontal inequalities and violent conflict: A comparative study
of Ghana and Côte d’Ivoire. In Stewart, F., ed., Horizontal Inequalities and Con-
flict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave
Macmillan.
Langer, A. 2009. Living with diversity: The peaceful management of horizontal
inequalities in Ghana. Journal of International Development 21(4): 534–546.
Langer, A. 2010a. Côte d’Ivoire’s elusive quest for peace. Bath Papers in International
Development, WP11. University of Bath.
Langer, A. 2010b. The situational importance of identity in Ghana. Ethnopolitics 9(1):
9–29.
Langer, A., and Brown, G.K. 2008. Cultural status inequalities: An important dimen-
sion of group mobilization. In Stewart, F., ed., Horizontal Inequalities and Con-
flict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave
Macmillan.
Langer, A., Stewart, F., and Venugopal, R., eds, 2011. Horizontal Inequalities and Post-
Conflict Development. Basingstoke: Palgrave Macmillan.
Langer, A., and Ukiwo, U. 2008. Ethnicity, religion and the state in Ghana and
Nigeria: Perceptions from the Streets. In Stewart, F., ed, Horizontal Inequalities and
Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave
Macmillan.
Le Pape, M. 2002. Chronologie politique de la Côte d’Ivoire, du coup d’état aux élec-
tions. In Le Pape, M., and Vidal, C., eds, Côte d’Ivoire: l’Année Terrible, 1999–2000.
Paris: Karthala.
References 283
Leibbrandt, M., Woolard, I., Finn, A., and Argent, J. 2010. Trends in South African
income distribution and poverty since the fall of Apartheid. OECD Social, Employ-
ment and Migration Working Papers 101. Paris: OECD.
Lemarchand, R. 1970. Rwanda and Burundi. London: Pall Mall Press.
Lemarchand, R. 1994. Burundi: Ethnocide as Discourse and Practice. Cambridge:
Cambridge University Press.
Lemarchand, R. 2006. Consociationalism and Power Sharing in Africa: Rwanda,
Burundi, and the Democratic Republic of the Congo. African Affairs 106(422):
1–20.
Lemarchand, R. 2009. The Dynamics of Violence in Central Africa. Philadelphia, PA:
University of Pennsylvania Press.
Lentz, C., and Nugent, P. 2000. Ethnicity in Ghana: A comparative perspectives.
In Lentz C., and Nugent, P., eds, Ethnicity in Ghana: The Limits of Intervention.
Levitsky, S., and Way, L.A. 2010. Competitive Authoritarianism: Hybrid Regimes after the
Cold War. Cambridge: Cambridge University Press.
Lewis, W.A. 1965. Politics in West Africa. Toronto: Oxford University Press.
Lewis, W.A. 1985. Racial Conflict and Economic Development. Cambridge, MA: Harvard
University Press.
Lijphart, A. 1971. Cultural diversity and theories of political integration. Canadian
Journal of Political Science 4(1): 1–14.
Lijphart, A. 1977. Democracy in Plural Societies: A Comparative Exploration. New Haven,
CT: Yale University Press.
Lijphart, A. 1985. Power-Sharing in South Africa, Policy Papers in International Affairs 24.
Berkeley, CA: Institute of International Studies, University of California.
Lijphart, A. 1991. Constitutional choices for new democracies. Journal of Democracy
2(1): 72–84.
Lijphart, A. 1994. Electoral Systems and Party Systems: A Study of Twenty-Seven Democra-
cies, 1945–1990. Oxford: Oxford University Press.
Lijphart, A. 1999. Patterns of Democracy: Government Forms and Performance in Thirty-six
Countries. New Haven, CT: Yale University Press.
Lijphart, A. 2002. The wave of power-sharing democracy. In Reynolds, A., ed., The
Architecture of Democracy: Constitutional Design, Conflict Management, and Democracy.
Oxford: Oxford University Press.
Lijphart, A. 2008. Thinking about Democracy: Power Sharing and Majority Rule in Theory
and Practice. London: Routledge.
Lindemann, S. 2010. Exclusionary elite bargains and civil war onset: The case of
Uganda. CSRC Working Paper 76. London: CSRC (Crisis States Research Centre).
Lipton, M. 1986. Capitalism and Apartheid: South Africa, 1910–1986. Aldershot:
Wildwood House.
Luckham, R. 1971. The Nigerian Military: A Sociological Analysis of Authority and Revolt,
1960–1967. Cambridge: Cambridge University Press.
Lupogo, H. 2001. Tanzania: civil–military relations and political stability. African
Security Review 10(1): 75–86.
MacDonald, M. 2006. Why Race Matters in South Africa. Cambridge, MA: Harvard
University Press.
Makara, S., Rakner, L., and Savsand, L. 2009. Turn-round: The National Resistance
Movement and the reintroduction of multiparty system in Uganda. International
Political Science Review 30(2): 185–204.
284 References
Makina, D. 2010. Zimbabwe in Johannesburg. In Crush, J., and Tereva, D., eds,
Zimbabwe’s Exodus: Crisis, Migration, Survival. Kingston: Southern African Migration
Programme.
Mamdani, M. 1996. Citizen and Subject: Contemporary Africa and the Legacy of Late
Colonialism. Princeton, NJ: Princeton University Press.
Manby, B. 2009. Struggles for Citizenship in Africa. London: Zed Books.
Mancini, L. 2008. Horizontal inequality and communal violence: evidence from
Indonesian districts. In Stewart, F., ed., Horizontal Inequalities and Conflict: Under-
standing Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan.
Maquet, J. 1971. Power and Society in Africa. London: Weidenfeld & Nicolson.
Marko, J. 2005. Post-conflict reconstruction through state- and nation-building:
The case of Bosnia and Herzegovina. European Diversity and Autonomy Papers
(EDAP) 4/2005. Bolzano-Bozen: EDAP, http://www.eurac.edu/en/research/institutes/
imr/activities/Bookseries/edap/Documents/2005_edap04.pdf (last accessed 1 May
2013).
Marx, A.W. 1998. Making Race and Nation: A Comparison of the United States, South
Africa, and Brazil. Cambridge: Cambridge University Press.
Marysse, S., and André, C. 2001. Guerre et pillage économique en République
Démocratique du Congo. In Marysse, S., and Reyntjens, F., eds, L’Afrique des Grands
Lacs, Annuaire 2000–2001. Paris: L’Harmattan.
Marysse, S., Ansoms, A., and Cassimon, D. 2006. The aid ‘darling’ and ‘orphans’ of
the Great Lakes Region in Africa. IOB Discussion Paper 2006.10. Antwerp: Institute
of Development Policy and Management, University of Antwerp.
Marysse, S., and Reyntjens, F., dir., 1999. L’Afrique des Grands Lacs, Annuaire 1998–
1999. Paris: L’Harmattan.
Masunungure, E.V., ed., 2009. Defying the Winds of Change: Zimbabwe’s 2008 Elections.
Harare: Weaver Press.
Matlosa, K. 2009. The role of the Southern African Development Community in the
management of Zimbabwe’s post-election crisis. Journal of African Elections 8(2):
46–73.
Mazarire, G.C. 2009. Reflections on pre-colonial Zimbabwe, c.850–1880s. In
Raftopoulos, B., and Mlambo, A.S., eds, Becoming Zimbabwe: A History from the
Pre-Colonial Period to 2008. Harare: Weaver Press.
McDonald, D.A. 2008. World City Syndrome: Neoliberalism and Inequality in Cape Town.
London: Routledge.
McGarry, J., O’Leary, B., and Simeon, R. 2008. Integration or accommodation? The
enduring debate in conflict regulation. In Choudhry, S., ed., Constitutional Design
for Divided Societies: Integration or Accommodation? Oxford: Oxford University Press.
MERG (Macroeconomic Research Group). 1993. Making Democracy Work: A Framework
for Macroeconomic Policy in South Africa. Bellville: Centre for Development Studies,
University of the Western Cape.
Miguel, E. 2004. Tribe or nation? Nation building and public goods in Kenya versus
Tanzania. World Politics 56(3): 328–362.
Mill, J.S. 1861. Considerations on Representative Government. New York: Prometheus
(1991 edition).
Mine, Y. 2006. The political element in the works of W. Arthur Lewis: The 1954 Lewis
model and African development. Developing Economies 44(3): 329–355.
Mkandawire, T. 2001. Thinking about development states in Africa. Cambridge Journal
of Economics 25(3): 289–313.
References 285
Mkandawire, T. 2008. The terrible toll of postcolonial rebel movements: Towards an

explanation of the violence against the peasantry. In Nhema, A., and Zeleza, P.T.,
eds, The Roots of African Conflicts: The Causes and Costs. Oxford: James Currey.
MOFPED (Ministry of Finance, Planning an Economic Development of the Republic of
Uganda). 2004. The Background to the Budget 2003/04 Fiscal Year. Kampala: MOFPED.
MOFPED (Ministry of Finance, Planning an Economic Development of the Republic of
Uganda). 2010. The Background to the Budget 2010/11 Fiscal Year. Kampala: MOFPED.
Moses, T. 2007. Grassroots-led strategies and actions for changing urban governance:
A case of the Community Peace Programme in Mbarara, Uganda. Network-
Association of European Researchers on Urbanization in the South, http://www.
n-aerus.net/web/sat/workshops/2007/papers/Final_Tukwasiibwe%20Moses_paper.
pdf (last accessed 1May 2013).
Moyo, S. 2010. The Zimbabwe crisis, land reform and normalisation. In Anseeuw, W.,
and Alden, C., eds, The Struggle over Land in Africa: Conflicts, Politics and Change.
Cape Town: HSRC Press.
Muhammad, H. 2011. NNPC Promotions, S-South 45 percent, NW/NE/NC 28 percent.
Daily Trust, http://allafrica.com/stories/201102170730.html (last accessed 1 May
2013).
Muhula, R. 2009. Horizontal inequalities and ethno-regional politics in Kenya. Kenya
Studies Review 1(1): 85–105.
Mukunya, F. 2009. Uganda: Growth of districts needs to be checked. New Vision, 30
November 2009, http://www.newvision.co.ug/D/8/459/702835 (last accessed 1 May
2013).
Munyarugerero, F.-X. 2003. Réseaux, Pouvoirs, Oppositions: La compétition politique au
Rwanda. Paris: L’Harmattan.
Murshed, S., and Gates, S. 2005. Spatial-horizontal inequality and the Maoist insur-
gency in Nepal. Review of Development Economics 9(1): 121–34.
Museveni, Y.K. 1997. Sowing the Mustard Seed: The Struggle for Freedom and Democracy
in Uganda. Basingstoke: Palgrave Macmillan.
Mustapha, A.R. 2006. Ethnic structure, inequality and governance of the public sector
in Nigeria. Democracy, Governance and Human Rights Programme 24. Geneva: UNRISD
(United Nations Research Institute for Social Development).
Mustapha, A.R. 2007. Institutionalising ethnic representation: How effective is the
Federal Character Commission in Nigeria? CRISE Working Paper No. 43. Oxford:
CRISE (Centre for Research on Inequality, Human Security and Ethnicity), Univer-
sity of Oxford.
Mutibwa, P. 2008a. Uganda Since Independence. Trenton, NJ: Africa World Press.
Mutibwa, P. 2008b. The Buganda Factor in Uganda Politics. Kampala: Fountain
Publishers.
Muzondidya, J. 2009. From buoyancy to crisis, 1980–1997. In Raftopoulos, B., and
Mlambo, A.S., eds, Becoming Zimbabwe: A History from the Pre-Colonial Period to 2008.
Mwenda, A.M. 2007. Personalizing power in Uganda. Journal of Democracy 18(3):
23–37.
Mwenda, A.M., and Tangri, R. 2005. Patronage politics, donor reforms, and regime
consolidation in Uganda. African Affairs 104(416): 449–467.
Nagel, J. 1974. Inequality and discontent: A nonlinear hypothesis. World Politics 26(4):
453–472.
Naniwe-Kaburahe, A. 2008. The Institution of Bashingantahe in Burundi. In Huyse,
L., and Salter, M., eds, Transitional Justice and Reconciliation after Violent Conflict:
286 References
Learning from African Experiences. Stockholm: International Institute for Democracy

and Electoral Assistance.
NBS (National Bureau of Statistics) Tanzania. 1999. Reproductive and Child Health Survey
1999. Dar es Salaam: NBS.
NBS (National Bureau of Statistics) Tanzania. 2002. Household Budget Survey 2000/01.
Dar es Salaam: NBS.
NBS (National Bureau of Statistics) Tanzania. 2009. Household Budget Survey 2007: Final
Report. Dar es Salaam: NBS.
NBS (Tanzania National Bureau of Statistics) Tanzania and ORC Macro. 2005. Tanzania
Demographic and Health Survey, 2004–2005. Dar es Salaam: NBS and ORC Macro.
NBS (National Bureau of Statistics) Nigeria. 2009. Social Statistics in Nigeria. Abuja: NBS.
Ndikumana, L. 2005. Distributional conflict, the state and peace building in Burundi.
The Round Table 94(381): 413–427.
Ndlovu-Gatsheni, S.J. 2009a. The Ndebele Nation: Reflections on Hegemony, Memory and
Historiography. Pretoria: UNISIA (University of South Africa) Press.
Ndlovu-Gatsheni, S.J. 2009b. Do ‘Zimbabweans’ Exist? Trajectories of Nationalism,
National Identity Formation and Crisis in a Postcolonial State. Bern: Peter Lang.
Ndu, E.C. 2011. Identity politics: The case of the ‘non-Igbo’ Igbo. In Ekekwe, E. et al.,
eds, Identity Politics in Nigeria: Essays in Honour of Kimse Okoko. Port Harcourt:
University of Port Harcourt Press.
Nkrumah, C. 1988. The Cohesion of Oppression: Clientship and Ethnicity in Rwanda,
1860–1960. New York: Columbia University Press.
Newbury, D. 2001. Precolonial Burundi and Rwanda: Local loyalties, regional royalties.
The International Journal of African Historical Studies 34(2): 255–314.
Nkrumah, K. 1962. Towards Colonial Freedom. London: Heinemann.
Nkurikiyimfura, J.-N. 1994. Le gros bétail et la société rwandaise, évolution historique: dès
XIIe – XIVe siècles à 1958. Paris: L’Harmattan.
Nkurunziza, J.D., and Ngaruko, F. 2008. Why has Burundi grown so slowly? The politi-
cal economy of redistribution. In Ndulu, B.J., O’Connell, S.A., Azam, J.P., Bates, R.H.,
Fosu, A.K., Gunning, J.W., and Njinkeu, D., eds, The Political Economy of Economic
Growth in Africa, 1960–2000: Vol. 2, Country Case Studies. Cambridge: Cambridge
University Press.
Nnoli, O. 1978. Ethnic Politics in Nigeria. Enugu: Fourth Dimension Publishers.
Nohlen, D., Krennerich, M., and Thibaut, B., eds, 1999. Elections in Africa: A Data
Handbook. Oxford: Oxford University Press.
Norris, P. 2008. Driving Democracy: Do Power-Sharing Institutions Work? Cambridge:
North, D.C. 1990. Institutions, Institutional Change and Economic Performance.
NPC (National Population Commission) of Nigeria and ICF Macro. 2009. Nigeria
Demographic and Health Survey 2008. Abuja: NPC and ICF Macro.
Nyamnjoh, F.B. 2006. Insiders and Outsiders: Citizenship and Xenophobia in Contempo-
rary Southern Africa. Dakar: CODESRIA Books.
Nyang’oro, J.E. 2006. Ethnic structure, inequality and governance of the public sec-
tor in Tanzania. In Bangura, Y., Ethnic Inequalities and Public Sector Governance.
Nyang’oro, J.E. 2010. JK: A Political Biography of Jakaya Mrisho Kikwete, President of the
United Republic of Tanzania. Trenton, NJ: Africa World Press.
Nyerere, J. 1966. Freedom and Unity. Dar es Salaam: Oxford University Press.
Nyerere, J. 1968. Freedom and Socialism. Dar es Salaam: Oxford University Press.
References 287
Nyerere, J. 1973. Freedom and Development. Dar es Salaam: Oxford University Press.
Nyerere, J. 2000. Foreword. In Museveni, Y.K., What Is Africa’s Problem? Minneapolis,
MN: University of Minnesota Press.
OECD (Organisation for Economic Co-operation and Development). 2010. African
Economic Outlook. Paris: OECD.
O’Flynn, I., and Russell, D., eds, 2005. Power Sharing: New Challenges for Divided
Societies. London: Pluto Press.
Ogata, S. 2010. Development co-operation and human security. Conflict, Security &
Development 10(1): 181–188.
Okolo, A. 1999. The Nigerian census: Problem and prospects. The American Statistician
53(4): 321–325.
O’Leary, B. 2005. Debating consociational politics: Normative and explanatory
arguments. In Noel, S., ed., From Power Sharing to Democracy: Post-Conflict
Institutions in Ethnically Divided Societies. Montreal: McGill-Queen’s University
Press.
Østby, G. 2008. Horizontal inequalities, political environment and civil conflict: Evi-
dence from 55 developing countries. In Stewart, F., ed., Horizontal Inequalities and
Conflict: Understanding Group Violence in Multiethnic Societies. Basingstoke: Palgrave
Macmillan.
Othman, S. 1984. Classes, crises and coup: The demise of Shagari’s regime. African
Affairs 83(333): 441–461.
Otite, O. 1990. Ethnic Pluralism and Ethnicity in Nigeria. Ibadan: Shaneson.
Ottaway, M. 1999. Africa’s New Leaders: Democracy or State Reconstruction? Washington,
DC: Carnegie Endowment for International Peace.
Paredes, M. 2007. Fluid identities: Exploring ethnicity in Peru. CRISE Working Paper
No. 40. Oxford: CRISE (Centre for Research on Inequality, Human Security and
Ethnicity), University of Oxford.
Paris, R. 2004. At War’s End: Building Peace after Civil Conflict. Cambridge: Cambridge
University Press.
PARP (Policy Analysis and Research Project). 2007. Nigerian State Assemblies: A Study of
Legislative Activities, 1999–2005. Abuja: PARP, National Assembly of Nigeria.
PARP (Policy Analysis and Research Project). 2008. National Assembly Statistical
Information 2(1). Abuja: PARP, National Assembly of Nigeria.
PEFAR (Public Expenditure and Financial Accountability Review) Team. 2006. Local
government fiduciary assessment: Joint evaluation report. Tanzania Public Expendi-
ture Review Annual Consultation Meeting, Dar es salaam: PEFAR.
Petterson, D. 2002. Revolution in Zanzibar. Boulder, CO: Westview Press.
Platzky, L., and Walker, C. 1985. The Surplus People: Forced Removals in South Africa.
Johannesburg: Ravan Press.
Plaza, S., and Ratha, D., eds, 2011. Diaspora for Development in Africa. Washington, DC:
World Bank.
Posner, D.N. 2005. Institutions and Ethnic Politics in Africa. Cambridge: Cambridge
University Press.
Powell, G.B. Jr. 2000. Elections as Instruments of Democracy. New Haven, CT: Yale
University Press.
Prunier, G. 1995. The Rwanda Crisis: History of a Genocide, 1959–1994. London: Hurst &
Company.
Prunier, G. 1998. The Rwandan Patriotic Front. In Clapham, C., ed., African Guerrillas.
Oxford: James Currey.
288 References
Raftopoulos, B., and Sachikonye, L., eds, 2001. Striking Back: The Labour Movement and
the Post-Colonial State in Zimbabwe, 1980–2000. Harare: Weaver Press.
Regional Forecasts. 2007. Uganda Multi-Donor Group: Northern Uganda Public Expendi-
ture Review (NUPER), Final Report. Hillsborough: Regional Forecasts Limited.
Reilly, B. 2001. Democracy in Divided Societies: Electoral Engineering for Conflict Manage-
ment. Cambridge: Cambridge University Press.
Republic of Rwanda. 2003. The Constitution of the Republic of Rwanda. Kigali: Republic
of Rwanda.
Republic of Rwanda. 2007. Economic Development and Poverty Reduction Strategy, 2008–
2012. Kigali: Republic of Rwanda.
Republic of Rwanda. 2012a. Rwanda External Trade Statistics Report Fourth Quarter 2011.
Kigali: Republic of Rwanda.
Republic of Rwanda. 2012b. Summary of the Report Presented at the Closing of Gacaca
Court Activities. Kigali: Republic of Rwanda.
République de Côte d’Ivoire. 2001. Tome 1, Etat et structures de la population.
Recensement Général de la Population et de l’Habitation de 1998 (RGPH-98) IV. Abidjan:
Institut National de la Statistique.
République du Burundi. 2005. Loi portant promulgation de la constitution de la République
du Burundi. Bujumbura: République du Burundi.
République du Burundi (Ministère de la planification du développement et de la
reconstruction nationale). 2006. Enquête QUIBB 2006, Rapport final. Bujumbura:
République du Burundi.
République Rwandaise. 2002. Ministère des finances et de la planification
économique. Rapport final, Enquête intégrale sur les conditions de vie des ménages au
Rwanda, 2000–2001. Kigali: République Rwandaise.
Rettig, M. 2008. Gacaca: Truth, justice, and reconciliation in postconflict Rwanda?
African Studies Review 51(3): 25–50.
Reynolds, A. 1999. Electoral Systems and Democratization in Southern Africa. Oxford:
Reynolds, A., ed., 2002. The Architecture of Democracy: Constitutional Design, Conflict
Management, and Democracy. Oxford: Oxford University Press.
Reyntjens, F. 1985. Pouvoir et droit au Rwanda: Droit public et évolution politique, 1916–
1973. Tervuren: Musée Royal de l’Afrique Centrale.
Reyntjens, F. 1994. L’Afrique des Grands Lacs en crise: Rwanda, Burundi, 1988–1994.
Paris: Karthala.
Reyntjens, F. 2003. Les nouveaux habits de l’empereur: Analyse juridico-politique de
la constituion rwandaise de 2003. In Marysse, S., and Reyntjens, F., eds, L’Afrique des
Grands Lacs, Annuaire 2002–2003. Paris: L’Harmattan.
Reyntjens, F. 2004. Rwanda, ten years on: From genocide to dictatorship. African
Affairs 103(411): 177–210.
Reyntjens, F. 2006. Briefing: Burundi: A peaceful transition after a decade of civil war.
African Affairs 105(418): 117–135.
Reyntjens, F. 2009. The Great African War: Congo and Regional Geopolitics, 1996–2006.
Reyntjens, F. 2011. Constructing the truth, dealing with dissent, domesticat-
ing the world: Governance in post-genocide Rwanda. African Affairs 110(438):
1–34.
Rhodes, R.A.W., Wanna, J., and Weller, P. 2009. Comparing Westminster. New York:
References 289
Rothchild, D., and Roeder, G.P. 2005. Dilemmas of state-building in divided societies.
In Roeder, G.P., and Rothchild, D., eds, Sustainable Peace: Power and Democracy after
Civil Wars. Ithaca, NY: Cornell University Press.
Rutherford, B. 2003. Belonging to the farm(er): Farm workers, farmers, and the shift-
ing politics of citizenship. In Hammer, A., Raftopoulos, B., and Jensen, S., eds,
Zimbabwe’s Unfinished Business: Rethinking Land, State and Nation in the Context of
Crisis. Harare: Weaver Press.
Rwabahungu, M., and Nintunze, A. 2009. Rapport sur l’étude d’actualisation de la
base de données sur la situation des déplacés et des rapatriés encore dans les sites.
Bujumbura: Ministère de la solidarité nationale du rapatriement des réfugiés et de la
réintégration sociale.
Sachikonye, L. 2004. Zimbabwe: Constitutionalism, the electoral system and chal-
lenges for governance and stability. Journal of African Elections 3(1): 140–159.
Sadomba, Z.W. 2011. War Veterans in Zimbabwe’s Revolution: Challenging Neo-
colonialism and Settler and International Capital. London: James Currey.
SAIRR (South African Institute of Race Relations). 1996. South Africa Survey, 1995–96.
Johannesburg: SAIRR.
Samuels, K. 2006. Post-conflict peace-building and constitution-making. Chicago
Journal of International Law 6(2): 1–20.
Samuelson, B.L., and Freedman, S.W. 2010. Language policy, multilingual education,
and power in Rwanda. Language Policy 9(3): 191–215.
Sanders, E.R. 1969. The Hamitic hypothesis: Its origin and functions in time perspec-
tive. Journal of African History 10(4): 521–532.
Sannerholm, R. 2007. Legal, judicial and administrative reforms in post-conflict soci-
eties: Beyond the rule of law template. Journal of Conflict & Security Law 12(1):
65–94.
Sartori, G. 1976. Parties and Party Systems: A Framework for Analysis. Cambridge:
Sartori, G. 1997. Comparative Constitutional Engineering: An Inquiry into Structures,
Incentives and Outcomes. 2nd edn. Basingstoke: Palgrave Macmillan.
Scarnecchia, T. 2008. The Urban Roots of Democracy and Political Violence in
Zimbabwe: Harare and Highfield, 1940–1964. Rochester, NY: University of Rochester
Press.
Schoenbrun, D. 1993. We are what we eat: Ancient agriculture between the Great
Lakes. Journal of African History 34(1): 1–31.
Schoenbrun, D. 1998. A Green Place, A Good Place: Agrarian Change, Gender, and Social
Identity in the Great Lakes Region to the 15th Century. Portsmouth: Heinemann.
Schulz-Herzenberg, C. 2009. Trends in party support and voter behaviour, 1994–
2009. In Southall, R., and Daniel, J. eds, ZUNAMI! The 2009 South African Elections.
Johannesburg: Jacana.
Scoones, I., Marongwe, N., Mavedzenge, B., Murimbarimba, F., Mahenehene, J., and
Sukume, C. 2010. Zimbabwe’s Land Reform: Myths and Realities. London: James
Currey.
Seekings, J. 2000. The UDF: A History of the United Democratic Front in South Africa,
1983–1991. Cape Town: David Philip.
Seekings, J., and Nattrass, N. 2005. Class, Race, and Inequality in South Africa. New
Haven, CT: Yale University Press.
Sen, A.K. 2007. Identity and Violence: The Illusion of Destiny. New York: W.W. Norton.
Shepherd, A., Gyimah-Boadi, E., Gariba, S., Plagerson, S., and Musa, W.A. 2005.
Bridging the north south divide in Ghana, Background paper for the 2006 world
290 References
development report. In World Bank. World Development Report 2006: Equity and
Development. Washington, DC: World Bank.
Silva-Leander, S. 2012. Structural violence and conflict: Vertical and horizontal
inequality in post-genocide Rwanda. In Langer, A., Stewart, F., and Venugopal,
R., eds, Horizontal Inequalities and Post-conflict Development. Basingstoke: Palgrave
Macmillan.
Sing’Oei, A.K. 2010. The ICC as arbiter in Kenya’s post-electoral violence. Minnesota
Journal of International Law Online 19: 5–20.
Sithole, M. 1999. Zimbabwe: Struggles within the Struggle. 2nd edn. Harare: Rujeko
Publishers.
Sklar, R.L. 1963. Nigerian Political Parties: Power in an Emergent African Nation.
Slye, R.C. 1996. The Dayton peace agreement: Constitutionalism and ethnicity. Yale
Journal of International Law 21: 459–473.
Smock, D.R., and Smock, A.C. 1975. The Politics of Pluralism: A Comparative Study of
Lebanon and Ghana. New York: Elsevier.
Southall, R. 2004. Post-election South Africa: The continuing case for electoral reform.
Journal of African Elections 3(2): 154–165.
Sparks, A. 1995. Tomorrow is Another Country: The Inside Story of South Africa’s Negotiated
Revolution. London: Heinemann.
Spears, S.I. 2002. Africa: The limits of power-sharing. Journal of Democracy 13(3):
123–136.
Statistics South Africa. 2003. Census 2001: Census in Brief. Pretoria: Statistics South
Africa.
Stewart, F. 2000. The root causes of humanitarian emergencies. In Nafziger, W.E.,
Stewart, F., and Väyrynen, R., eds, War, Hunger and Displacement: The Origins of
Humanitarian Emergencies, Vol. 1: Analysis. Oxford: Oxford University Press.
Stewart, F. 2002. Horizontal inequalities: A neglected dimension of development. QEH
Working Paper Series 81. Oxford: Queen Elizabeth House, University of Oxford.
Stewart, F. 2008a. Horizontal inequalities and conflict: An introduction and some
hypotheses. In Stewart, F., ed., Horizontal Inequalities and Conflict: Understanding
Group Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan.
Stewart, F. 2008b. Note for Discussion: Kenya, Horizontal Inequalities and the Political Dis-
turbances of 2008. Oxford: CRISE (Centre for Research on Inequality, Human Security
and Ethnicity), University of Oxford.
Stewart, F. 2010a. Horizontal inequalities in Kenya and the political disturbances
of 2008: Some implications for aid policy. Conflict, Security & Development 10(1):
133–159.
Stewart, F. 2010b. Horizontal inequalities as a cause of conflict: A review of CRISE find-
ings. CRISE Overview 1. Oxford: CRISE (Centre for Research on Inequality, Human
Security and Ethnicity), Oxford University Press.
Stewart, F., ed., 2008. Horizontal Inequalities and Conflict: Understanding Group Violence
in Multiethnic Societies. Basingstoke: Palgrave Macmillan.
Stewart, F., and Brown, G.K. 2007 Motivations for conflict: Groups and individuals.
In Crocker C.A., Hampson, F.O., and Aall, P.R., eds, Leashing the Dogs of War: Conflict
Management in A Divided World. Washington, DC: USIP (United States Institute of
Peace).
Stewart, F., Brown, G., and Langer, A. 2008. Policies towards horizontal inequalities.
In Stewart, F., ed., Horizontal Inequalities and Conflict: Understanding Group Violence in
Multiethnic Societies. Basingstoke: Palgrave Macmillan.
References 291
Stewart, F., and Langer, A. 2008. Horizontal inequalities: Explaining persistence and
change. In Stewart, F., ed., Horizontal Inequalities and Conflict: Understanding Group
Violence in Multiethnic Societies. Basingstoke: Palgrave Macmillan.
Stoneman, C. 1988. Some concluding thoughts. In Stoneman, C., ed., Zimbabwe’s
Prospects: Issues of Race, Class, State and Capital in Southern Africa. Basingstoke:
Palgrave Macmillan.
Stoneman, C. 2000. Zimbabwe land policy and the land reform. In Bowyer-Bower,
T.A.S., and Stoneman, C., eds, Land Reform in Zimbabwe: Constraints and Prospects.
Aldershot: Ashgate.
Sullivan, D.P. 2005. The missing pillars: A look at the failure of peace in Burundi
through the lens of Arend Lijphart’s theory of consociational democracy. Journal of
Modern African Studies 43(1): 75–95.
Szereszewski, R. 1966. Regional aspects of the structure of the economy. In
Birmingham, W.B., Neustadt, I., and Omaboe, E.N., eds, A Structure of Contemporary
Ghana. London: George Allen and Unwin.
Takeuchi, S. 2007. Political liberalization or armed conflicts? Political changes in post-
Cold War Africa. The Developing Economies 45(2): 172–193.
Takeuchi, S. 2011. Gacaca and DDR: The disputable record of state-building in
Rwanda. JICA-RI Working Paper No. 32. Tokyo: JICA Research Institute.
Terreblanche, S. 2002. A History of Inequality in South Africa, 1652–2002.
Pietermaritzburg: University of Natal Press.
Therkildsen O., and Tidemand, P. 2007. Staff Management and Organizational Perfor-
mance in Tanzania and Uganda: Public Servant Perspectives. Copenhagen: DIIS (Danish
Institute for International Studies).
Thompson, L. 2001. A History of South Africa. 3rd edn. New Haven, CT: Yale University
Press.
Thomson, S. 2011. The darker side of transitional justice: The power dynamics behind
Rwanda’s Gacaca courts. Africa 81(3): 373–390.
Thorp, R., Caumartin, C., and Gray-Molina, G. 2006. Inequality, ethnicity, polit-
ical mobilisation and political violence in Latin America: The cases of Bolivia,
Guatemala and Peru. Bulletin of Latin American Research 25(4): 453–480.
Tilly, C. 1992. Coercion, Capital, and European States, AD 990–1992. Oxford: Blackwell.
Transparency International. 2010. East Africa Bribery Index 2010, http://probe
international.org/library/wp-content/uploads/2010/10/EAST-AFRICA-BRIBERY-
INDEX-2010.pdf (last accessed 1 May 2013).
Tripp, A.M. 1994. Gender, political participation and the transformation of associa-
tional life in Uganda and Tanzania. African Studies Review 37(1): 107–31.
Tripp, A.M. 2010. Museveni’s Uganda: Paradoxes of Power in a Hybrid Regime. Boulder,
CO: Lynne Rienner Publishers.
Tumushabe, G.W. 2009. Trends in public administration expenditure in Uganda: The
cost of the executive and its implications on poverty eradication and governance.
Advocates Coalition for Development and Environment (ACODE) Policy Research Series
No.27. Kampala: ACODE.
Turton, D. 1997. War and ethnicity: Global connections and local violence in
Northeast Africa and Former Yugoslavia. Oxford Development Studies 25(1): 77–94.
UBOS (Uganda Bureau of Statistics). 1995. The 1991 Uganda Population and Housing
Census: Analytical Report. Kampala: UBOS.
UBOS (Uganda Bureau of Statistics). 2003. Uganda National Household Survey
2002/2003: Report on the Socio-Economic Survey. Kampala: UBOS.
292 References
UBOS (Uganda Bureau of Statistics). 2005. The 2002 Uganda Population and Housing
Census: Main Report. Kampala: UBOS.
2005/2006: Report on the Socio-Economic Survey. Kampala: UBOS.
2009/2010: Socio-Economic Module, Abridged Report. Kampala: UBOS.
Ukiwo, U. 2007. Education, horizontal inequalities and ethnic relations in Nigeria.
International Journal of Educational Development 27(2): 266–281.
UNAIDS (Joint United Nations Programme on HIV/AIDS). 2010. Report on the Global
AIDS Epidemic 2010. Geneva: UNAIDS.
UNDP (United Nations Development Programme). 1998. Zimbabwe Human Develop-
ment Report 1998. Harare: Institute of Development Studies, University of Zimbabwe.
UNDP (United Nations Development Programme). 2003a. South Africa Human Devel-
opment Report 2003. Cape Town: Oxford University Press.
UNDP (United Nations Development Programme). 2003b. Summary: Zimbabwe Human
Development Report 2003. Harare: Institute of Development Studies, University of
Zimbabwe.
UNDP (United Nations Development Programme). 2004. Human Development Report
2004: Cultural Liberty in Today’s Diverse World. Oxford: Oxford University Press.
2006. New York: UNDP.
UNDP (United Nations Development Programme). 2007a. Turning Vision 2020 into
Reality: From Recovery to Sustainable Human Development, National Human Develop-
ment Report, Rwanda 2007. Kigali, Rwanda: UNDP.
UNDP (United Nations Development Programme). 2007b. Human Development Report
2007/2008. New York: UNDP.
UNDP (United Nations Development Programme). 2007c. Towards a More Inclusive
Society: Ghana Human Development Report. Accra: UNDP.
UNDP (United Nations Development Programme). 2009a. Human Development Report
2009: Overcoming Barriers: Human Mobility and Development. Basingstoke: Palgrave
Macmillan.
UNDP (United Nations Development Programme). 2009b. Zanzibar Human Develop-
ment Report 2009: Towards Pro Poor Growth. Zanzibar: UNDP Zanzibar.
UNDP (United Nations Development Programme). 2009c. Nigeria National Human
Development Report. Abuja: UNDP.
2010, 20th Anniversary Edition: The Real Wealth of Nations: Pathways to Human
Development. Basingstoke: Palgrave Macmillan.
Union of South Africa. 1955. Summary of the Report of the Commission for the Socio-
Economic Development of the Bantu Areas within the Union of South Africa, U.G.
61/1955. Pretoria: Government Printers (Tomlinson Commission Report).
UNSC (United Nations Security Council). 2001. Report of the Panel of Experts on the
Illegal Exploitation of Natural Resources and Other Forms of Wealth of the Democratic
Republic of the Congo, S/2001/357. New York: United Nations.
UNSC (United Nations Security Council). 2002. Final Report of the Panel of Experts on
the Illegal Exploitation of Natural Resources and Other Forms of Wealth of the Democratic
Republic of the Congo, S/2002/1146. New York: United Nations.
UNSC (United Nations Security Council). 2010. Final Report of the Group of Experts on
the Democratic Republic of the Congo, S/2010/596. New York: United Nations.
References 293
USAID (United States Agency for International Development). 2010. Nigeria Conflict
Assessment. Abuja: USAID.
Uvin, P. 2010. Structural causes, development co-operation and conflict prevention in
Burundi and Rwanda. Conflict, Security & Development 10(1): 161–179.
Vambe, M., ed., 2008. The Hidden Dimensions of Operation Murambatsvina in Zimbabwe.
Van Acker, F. 2004. Uganda and the Lord’s Resistance Army: The new order no one
ordered. African Affairs 103(412): 335–357.
Van der Berg, S., Louw, M., and du Toit, L. 2009. Poverty trends since the transition:
What we know. Stellenbosch Economic Working Papers 19(9). Matieland: Department
of Economics, University of Stellenbosch.
Vandeginste, S. 2006. Théorie consociative et partage du pouvoir au Burundi, Cahier
2006.04. Anvers: Institut de politique et de gestion du développement, Université
d’Anvers.
Vandeginste, S. 2008. Burundi: Entre le modèle consociatif et sa mise en oeuvre.
In Marysse, S., Reyntjens, F., and Vandeginste, S., eds, L’Afrique des Grands Lacs,
Annuaire 2007–2008. Paris: L’Harmattan.
Vandeginste, S. 2009. Power-sharing, conflict and transition in Burundi: Twenty years
of trial and error. Africa Spectrum 44(3): 63–86.
Vandeginste, S. 2011. Power-sharing as a fragile safety valve in times of electoral tur-
moil: The costs and benefits of Burundi’s 2010 elections. Journal of Modern African
Studies 49(2): 315–335.
Vansina, J. 2001. Le Rwanda ancien: Le royaume nyiginya. Paris: Karthala.
Varshney, A. 2002. Ethnic Conflict and Civic Life: Hindus and Muslims in India. New
Haven, CT: Yale University Press.
Vickers, M. 2000. Ethnicity and Sub-Nationalism in Nigeria: Movement for the Mid West
State. Oxford: Worldview.
Waldmeir, P. 1997. Anatomy of a Miracle: The End of Apartheid and the Birth of the New
South Africa. New York: W.W. Norton.
Waldorf, L. 2006. Mass justice for mass atrocity: Rethinking local justice as transitional
justice. Temple Law Review 79(1): 1–87.
Watts, R.L. 2008. Comparing Federal Systems. 3rd edn. Montreal: McGill-Queen’s
University Press.
Williams, P.R., and Simpson, M.T. 2008. Rethinking the political future: An alternative
to the ethno-sectarian division of Iraq. American University International Law Review
24(2): 191–247.
Wolff, S. 2009. Complex power-sharing and the centrality of territorial self-governance
in contemporary conflict settlements. Ethnopolitics 8(1): 27–45.
Wolff, S. 2011. Post-conflict state building: The debate on institutional choice. Third
World Quarterly 32(10): 1777–1802.
Wong, R.B. 1997. China Transformed: Historical Change and the Limits of European
Experience. Ithaca, NY: Cornell University Press.
World Bank. 2006. Uganda Poverty and Vulnerability Assessment. Washington, DC:
World Bank.
World Bank. 2009. Africa Development Indicators 2008/09: Youth and Employment in
Africa: The Potential, the Problem, the Promise. Washington, DC: World Bank.
Yeager, R. 1989. Tanzania: An African experiment. 2nd edn. Boulder, CO: Westview
Press.
Yoshikuni, T. 1999. Notes on the influence of town–country relations on African urban
history, before 1957: Experiences of Salisbury and Bulawayo. In Raftopoulos, B., and
294 References
Yoshikuni, T., eds, Sites of Struggle: Essays in Zimbabwe’s Urban History. Harare: Weaver
Press.
Young, C. 1976. The Politics of Cultural Pluralism. Madison, WI: University of
Wisconsin Press.
Younis, N. 2011. Set up to fail: Consociational political structures in post-war Iraq,
2003–2010. Contemporary Arab Affairs 4(1): 1–18.
Zartman, W.I. and Delgado, C. 1984. Introduction. In: Zartman, W.I., and Delgado,
C., eds, The Political Economy of Ivory Coast. New York: Praeger Publishers.
Zimbabwe Election Support Network. 2009. Possible Electoral Systems for Zimbabwe.
Harare: Zimbabwe Election Support Network.
Zondi, S. 2011. ZANU-PF and MDC power-sharing: Zimbabwe still at crossroads?
In Besada, H., ed., Zimbabwe: Picking Up the Pieces. Basingstoke: Palgrave Macmillan.
Author Index
Acemoglu, D., 23 Choudhry, S., 38

Adebajo, A., 178 Chrétien, J.-P., 43, 59, 60
Agyeman, D.K., 83, 84, 85 Chung, F., 104
Akindès, F., 73, 78, 80 Cohen, M.A., 66
Alesina, A., 23, 24 Coleman, J., 180
Alexander, J., 93, 104 Collier, P., 9, 39, 205
Alston, P., 167 Connah, G., 1
Alwy, A., 177 Contamin, B., 78
André, C., 63 Cottrell, J., 27
Ansoms, A., 50, 61 Crook, R., 75, 76
Anyinam, C., 87
Apter, D.E., 128 Daddieh, C., 73, 74
Asante, R., 82, 84, 85, 86, 214 Dahl, R., 12, 13, 33
Azam, J.-P., 76 Dashwood, H.S., 104, 111
Davies, R., 111
Bach, D.C., 178 Delgado, C., 73
Dembélé, O., 75, 81
Bakari, M.A., 133
Den Tuinder, B., 74
Bakary, T., 73
Dennys, P., 268
Banegas, R., 80
D’Hertefelt, M., 59
Bangura, Y., 13, 23
Diamond, L., 28, 38
Baskin, J., 99
Doom, R., 129
Bates, R., 3, 67
Dorsey, L., 60
Beach, D., 101
Dorsey, M., 54
Behrend, H., 129
Dozon, J.-P., 77
Bentley, K.A., 108
Drumbl, M.A., 64
Beswick, D., 54
Durkheim, E., 41
Blum, S., 139
Boahen, A.A., 84, 85, 93
Edigheji, O., 123
Bond, P., 111 Eggers, E.K., 60
Booysen, S., 113 Ellis, S., 98
Bourne, R., 125 Englebert, P., 143, 206
Bresson, Y., 74 Evans-Pritchard, E.E., 107
Brown, D., 69, 82, 84, 86
Brown, G.K., 1, 2, 4, 5, 9, 38, 66, 67, 68, Fjeldstad, O.H., 36
86, 215 Freedman, S.W., 47
Bruce, J.W., 61 Frempong, K.D., 83, 84, 85
Brukum, N.J.K., 83 Freund, B., 99
Bundy, C., 97 Friedman, S., 111
Burgess, T., 133, 137 Fukuda-Parr, S., 1–9, 253
Burnet, J.E., 62 Furnivall, J.S., 96
Cederman, L.A., 4, 5, 68, 208, 262 Gaffney, P.D., 48

Chachage, S.L.S., 137 Gahama, J., 60
Chitere, P., 27 Gates, S., 4
295
296 Author Index
Geschiere, P., 270 Lewis, W.A., 11, 12, 13, 37

Ghai, Y., 27 Lijphart, A., 12, 13, 14, 15, 17, 37, 38,
Gonin, G., 70–1 39, 62, 63, 124
Green, E., 129, 143, 152 Lindemann, S., 139, 142
Gurr, T.R., 3, 12, 37, 66, 67 Lipton, M., 99
Gyimah-Boadi, E., 73, 74, 82, 84, 85, 86, Losch, B., 78, 80
214 Luckham, R., 193
Lupogo, H., 142
Hagan, G.P., 86
Hartzell, C.A., 13 MacDonald, M., 110
Hassim, S., 115 Makara, S., 142
Herbst, J., 1, 101 Makina, D., 115
Hinderink, J., 73 Mamdani, M., 107
Hintjens, H., 47 Manby, B., 270
Hirschman, A.O., 36 Mancini, L., 4, 208
Hoddie, M., 13 Manyanya, M., 111
Hoeffler, A., 205 Maquet, J., 59
Holborn, L., 124 Marko, J., 39
Horowitz, D.L., 12, 37, 38, 67, 68, 124, Marx, A.W., 96
194 Marysse, S., 61, 63, 64
Huggins, C., 61 Masunungure, E.V., 112, 124
Hummel, R., 206 Matlosa, K., 112
Hutchful, E., 84 Mazarire, G.C., 101
McDonald, D.A., 115
Iheduru, O.C., 99 McGarry, J., 14, 38
Iliffe, J., 1 Miguel, E., 270
Ingelaere, B., 48, 55, 56, 64 Mill, J.S., 11
Mine, Y., 1–9, 10–39, 95–125, 253, 254,
Kanyenze, G., 103, 111 256
Kanyinga, K., 154, 156, 214 Mkandawire, T., 35, 252–74
Kasfir, N., 138 Moses, T., 128
Kieffer, G.-A., 78 Moyo, S., 125
Killian, B., 133 Muhammad, H., 202
Kimenyi, M.S., 153–77, 258 Muhula, R., 4
Klein, N., 124 Murshed, S., 4
Klugman, J., 131, 162 Museveni, Y.K., 39, 62, 126, 129, 130,
Kniss, M., 162, 166–7 138, 139, 141, 142, 145, 150, 152,
Kriger, N., 124 257
Kymlicka, W., 9 Mustapha, A.R., 184, 186, 191, 214,
243
Laakso, L., 119 Mutibwa, P., 129, 143
Laakso, M., 15 Mwenda, A.M., 141, 143
Laely, T., 60
Langer, A., 1–9, 38, 66–94, 204, 208–52, Naniwe-Kaburahe, A., 60
253, 255, 259 Nattrass, N., 99, 123
Le Pape, M., 80 Ndikumana, L., 4
Leibbrandt, M., 101 Ndlovu-Gatsheni, S.J., 101, 112, 125
Lemarchand, R., 44, 59, 60, 61, 62, 63 Ndung’u, N.S., 153
Lentz, C., 84, 87 Newbury, D., 43, 59, 60
Levitsky, S., 39 Ngaruko, F., 60, 62
Author Index 297
Nintunze, A., 64 Samuelson, B.L., 47

Nkrumah, K., 84, 85, 86, 87, 90, 256, 270 Sanders, E.R., 43
Nkurikiyimfura, J.-N., 59 Sannerholm, R., 271
Nkurunziza, J.D., 60 Sartori, G., 12, 38
Nnoli, O., 179 Scarnecchia, T., 124
Nohlen, D., 21 Schech, S., 177
Norris, P., 12, 16 Schoenbrun, D., 59
North, D.C., 41 Schulz-Herzenberg, C., 110
Nugent, P., 84, 87 Scoones, I., 123
Nyamnjoh, F.B., 115 Sechaba, T., 98
Nyang’oro, J.E., 126–52, 257 Seekings, J., 99, 123
Nyerere, J., 1, 52, 95, 126, 138, 140, 142, Sen, A.K., 9
150, 257, 270 Shepherd, A., 88
Silva-Leander, S., 54, 61
O’Flynn, I., 38 Simpson, M.T., 39
Ogata, S., 271 Sing’Oei, A.K., 269
Okolo, A., 211 Sithole, M., 124
O’Leary, B., 38 Sklar, R.L., 180
Østby, G., 4, 68, 208 Slye, R.C., 39
Othman, S., 199 Smock, A.C., 86
Otite, O., 179 Smock, D.R., 86
Ottaway, M., 50, 62 Southall, R., 108, 122
Sparks, A., 107
Paredes, M., 4 Spears, S.I., 39
Paris, R., 41 Stewart, F., 1, 2, 3, 4, 5, 9, 38, 41, 45, 58,
Petterson, D., 133 59, 66, 67, 68, 157, 208, 209, 214,
Plattner, M.F., 28, 38 215, 243, 252–74
Platzky, L., 97 Stoneman, C., 103, 104
Plaza, S., 61 Sullivan, D.P., 62
Posner, D.N., 17 Szereszewski, R., 88
Powell, G.B, Jr., 13
Prunier, G., 45, 60, 62
Taagepera, R., 15
Takeuchi, S., 6, 40–65, 255
Raftopoulos, B., 111
Tangri, R., 143
Ratha, D., 61
Tempelman, G.J., 73
Reilly, B., 12
Rettig, M., 64 Terreblanche, S., 124
Reynolds, A., 38, 103 Therkildsen, O., 36, 140
Reyntjens, F., 44, 46, 49, 50, 53, 54, 55, Thompson, L., 98
60, 61 Thomson, S., 64
Rhodes, R.A.W., 13 Thorp, R., 4
Roeder, G.P., 39 Tidemand, P., 140
Rothchild, D., 39 Tilly, C., 1, 39, 41
Russell, D., 38 Toulou, L., 113
Rutherford, B., 115 Tripp, A.M., 131, 139, 141
Rwabahungu, M., 64 Tumushabe, G.W., 136
Turton, D., 66
Sachikonye, L., 124
Sadomba, Z.W., 124 Ukiwo, U., 4, 86, 178–207, 214, 259
Samuels, K., 39, 47 Uvin, P., 59
298 Author Index
Vambe, M., 112 Way, L.A., 39

Van Acker, F., 129 Williams, P.R., 39
Van der Berg, S., 100 Wolff, S., 34, 38
Vandeginste, S., 54, 62, 63 Wong, R.B., 1
Vansina, J., 43, 59
Vickers, M., 180 Yeager, R., 133
Vlassenroot, K., 129 Yoshikuni, T., 118
Young, C., 86, 131
Waldmeir, P., 107 Younis, N., 39
Waldorf, L., 64
Walker, C., 97 Zartman, W.I., 73
Watts, R.L., 38 Zondi, S., 112
Subject Index
Abidjan, 73, 74, 80, 81, 256 Banyankole, 127, 128, 129, 135, 136,
Accra, 83, 87, 88, 89, 209 139, 145, 146, 150, 152, 214, 216,
Acheampong, Ignatius Kutu, 84 227, 228, 229, 231, 232, 242, 243,
Acholi, 127, 128, 129, 135, 139, 144, 245
150, 214, 231, 232, 241, 242, 261, Banyoro, 127, 128, 145, 214, 225, 227,
262 228, 229, 231, 232, 240, 241, 243
Action Congress of Nigeria (ACN), 197 Baoulé, 72, 75, 76, 77, 78, 79, 92
ad hoc power-sharing, 10, 33, 95–125, Basic Human Needs (BHN), 213, 215,
254, 256–7, 258, 265, 266, 269, 272, 216, 233, 248
273 Bédié, Henri Konan, 75, 92
affirmative action, 104, 116, 193, 205 Biafran War, 259, 265
African Independent Television (AIT), Black Economic Empowerment (BEE), 99
192 Boahen, Albert Adu, 84, 85, 93
African National Congress (ANC), 98, Botswana, 14, 31, 35
105, 107, 108, 109, 110, 117, 122, budget, 17, 111, 144, 198, 271, 273
123 Buganda Kingdom, 126, 127, 128, 130,
Afrikaners, 96, 101, 110, 124 143, 144, 150, 258, 261
agriculture, 40, 43, 50, 59, 73, 87, 98, 99, Bujumbura, 62, 64, 65
123, 158, 184 Bukenya, Gilbert, 139
Akan, 70, 72, 82, 84, 85, 86, 88, 216, Bulawayo, 114, 117–21, 125
241, 242, 256 Burkina Faso, 70
Alliance for Democracy (AD), 196, 197, Burundi, 4, 7, 8, 24, 28, 31, 33, 39,
198 40–65, 108, 253, 254, 255, 261, 262,
All Nigeria Peoples Party (ANPP), 196, 264, 265, 267, 269, 272
197, 198 Bururi, 44, 45, 46, 48, 62, 64
Amin, Idi, 126, 129, 133 Busoga Kingdom, 128, 143
Angola, 8, 19, 31, 33 Buyoya, Pierre, 46, 48, 52, 61
Annan, Kofi, 167
Arab Spring, 178 cabinet
Arusha Peace and Reconciliation in Burundi, 46, 51
Agreement, 46, 50 in Côte d’Ivoire, 75, 78, 108
Ashanti, 82, 84, 85, 86, 88, 89 in Ghana, 89–91
assassination, 44, 46 in Kenya, 108, 157, 158, 176
assets, 3, 161, 213, 214, 215, 216, 239, in Nigeria, 184, 187, 196, 197–8, 202
247, 248 in Rwanda, 49
autonomy, 12, 34, 35, 43, 97, 124, in South Africa, 107, 108, 110
126–52, 191, 192, 193, 200, 257, in Tanzania, 141–3
258, 267, 269 in Uganda, 141–3
Avoidance of Discrimination Act, 85 in Zimbabwe, 106, 108, 119
Cameroon, 21
Baganda, 126, 127, 128, 129, 130, 131, Cape Town, 105, 115, 116, 117
132, 133, 134, 136, 138, 139, 142, cash crops
143, 144, 145, 146, 150, 151, 152, cocoa, 73
227, 232, 241, 245, 257, 258, 261 coffee, 73
299
300 Subject Index
Catholics, 4, 76, 83, 133, 152, 170 constitutional engineering, 8, 11, 12, 37,
Central Bank for West African States, 75, 254, 259
77 constitutions
Centre for Research on Inequality, in Burundi, 46, 50–2
Human Security and Ethnicity in Côte d’Ivoire, 75, 76
(CRISE), 3, 9 in Ghana, 84, 86, 89–90
Chama cha Mapinduzi (CCM), 133, 140, in Kenya, 154, 162–3, 164–5, 175–7
142, 143, 148 in Nigeria, 191, 193–4, 195, 196, 197,
Charte du Grand Nord, 76 198, 200, 206
child mortality, 89, 159, 160 in Rwanda, 49–50
Chimurenga, 101, 103 in South Africa, 107–8, 111
Christianity, 34, 40, 71, 72, 76, 83, 98, in Tanzania, 133, 142, 143
117, 133, 147, 156, 164, 170, 181, in Uganda, 128, 141, 144
182, 211, 212, 256, 259 in Zimbabwe, 111–12
constructivism, 3, 42, 59, 67
Civic United Front (CUF), 133, 140, 141,
Convention People’s Party (CPP), 86
143, 148, 152
corruption, 58, 63, 64, 117, 121, 140,
civil society, 50, 64, 107, 124, 164, 271
143, 167
civil war, 5, 34, 40, 45, 46, 47, 48, 49, 50,
Côte d’Ivoire, 4, 7, 8, 28, 31, 33, 35, 39,
52, 53, 56, 57, 58, 61, 63, 64, 65, 66,
66–94, 108, 253, 254, 255–6, 261,
67, 69–81, 95, 107, 153, 178, 180,
262, 263, 265, 266, 267, 270
192, 193, 195, 205, 206, 208, 259,
coup d’état, 70, 77, 78, 80, 84, 92, 93,
266
129
class, 11, 12, 84, 99, 111, 112, 115, 116,
cultural status horizontal inequalities
118, 119, 138, 139, 192, 193, 257,
(HIs)
271
in Burundi, 41
Cold War, 1, 10, 106, 254, 266, 272 in Côte d’Ivoire, 76, 91
Colonialism, 23, 25, 35, 37, 42, 43, 44, in Ghana, 91
45, 47, 48, 60, 61, 69, 70, 87, 96, in Kenya, 153, 156–62
100, 118, 128, 129, 136, 142, 156, in Nigeria, 182, 191
179, 180, 215, 254, 255, 256, 258, in Rwanda, 41
261 in South Africa, 95
Coloureds, 96, 97, 98, 99, 116, 117, 123, in Tanzania, 127
124 in Uganda, 134
Comité national de salut public (CNSP), in Zimbabwe, 108
78
conflict prevention, 6, 10–39, 41, 57–9, Dahl, Robert, 12, 13, 33
253, 254, 259, 262, 264, 265, 270, Dar es Salaam, 136, 138, 144, 145, 147,
273, 274 148, 149, 152
Congress of the People (COPE), 110 decentralisation
Congress of South African Trade Unions in Burundi, 19, 22, 23, 24, 26, 43, 45,
(COSATU), 99, 110 52, 255
Conseil national pour la défense de la in Côte d’Ivoire, 19, 22, 23, 24, 26, 28,
démocratie–Forces pour la défense 31, 33
de la démocratie (CNDD-FDD), 46, in Ghana, 20, 22, 23, 24, 26
52, 54, 55, 57, 58, 63, 64 in Kenya, 20, 22, 23, 24, 26, 162–5,
consociational democracy, 11, 12 167, 176–7, 258–9
constituency, 15, 113, 119, 122, 125, in Nigeria, 20, 22, 23, 24, 26, 194,
166, 194, 195 199–203, 206
Constitutional Court, 107 in Rwanda, 20, 22, 23, 24, 26, 52
Subject Index 301
in South Africa, 21, 22, 23, 24, 26, 103, electoral systems, 14, 15, 16, 17, 18, 21,
105–11 28, 38, 112, 113, 122, 125, 138, 194,
in Tanzania, 21, 22, 23, 24, 26, 126, 195, 259, 270
131, 138, 141, 143–4, 150, 151, elite, 13, 35, 45, 47, 48, 50, 59, 60, 76,
257–8, 266 95, 103, 115, 122, 124, 128, 129,
in Uganda, 21, 22, 23, 24, 26, 126, 136, 140, 199, 260
141, 143–4, 146–7, 149, 151–2, Ethiopia, 23, 31, 34, 62
257–8 ethnic fractionalisation, 23, 24
in Zimbabwe, 21, 22, 23, 24, 26, 103, ethnicity
105–11, 113, 124 in Burundi, 41, 46, 48, 51, 53, 54, 55,
De Klerk, F.W., 106, 108 57
democracy, 10, 11, 12, 13, 14, 15, 21, 31, in Côte d’Ivoire, 69, 81
33–6, 39, 41, 49, 50, 53, 60, 82, 84, in Ghana, 82, 83, 84–5, 86
85, 107, 110, 117, 122, 126, 152, in Kenya, 153, 156, 162, 163, 168–75
167, 197, 207, 255, 256, 257, 259 in Nigeria, 179, 191
Democratic Alliance (DA), 110, 117 in Rwanda, 41, 45, 56, 57
Democratic Republic of the Congo (the in South Africa, 96, 97, 98, 101, 118
DRC), 1, 21, 23, 31, 33, 38, 53, 54, in Tanzania, 130, 140, 143, 151
62, 63, 64, 108, 111, 145, 254 in Uganda, 149
development agencies, 253, 271, 273, in Zimbabwe, 119
274 Europe, 1, 24, 66
development cooperation, 28, 64, 119, Evans-Pritchard, E.E., 107
135, 143, 146, 178, 252–74 Ewe, 82, 83, 84, 85, 90, 216, 227, 231,
disarmament, demobilisation and 232, 241, 242, 244
reintegration (DDR), 54, 111
discrimination, 48, 49, 85, 116, 189, 191,
209, 215, 227 Federal Character Commission (FCC),
194, 201, 202
distortion of perceptions, 241, 260, 262,
273 Federal Character Principle, 187, 189,
193, 194, 200, 201, 202, 203, 206,
259, 265
East Africa, 126, 131, 257 federalism, 12, 13, 18, 23, 34, 38, 108,
East Asia, 1, 123 130, 138, 142, 143, 144, 180, 184,
education, 4, 44, 47, 48, 50, 56, 87, 88, 185, 187, 189, 191, 192, 193, 194,
104, 135, 137, 145, 157, 158, 159, 195, 196, 197, 199, 200, 201, 202,
160, 162, 173, 187, 188, 202, 203, 203, 205, 206, 207, 254, 257–8, 259,
226, 227, 229, 230, 233, 261, 270 265, 267–8, 269
Egypt, 8, 178 feelings of inferiority, 233, 241, 244, 245
elections feelings of superiority, 244, 245
in Burundi, 51, 55, 57, 62, 63 FIFA World Cup, 117
in Côte d’Ivoire, 71, 75, 76, 77, 81 first-past-the-post (FPTP), 12, 13, 35, 36,
in Ghana, 85, 92 113, 114, 122, 123, 194, 254
in Kenya, 153, 163, 165, 166, 167, 172, fiscal decentralisation, 16, 266, 267–8,
173, 176, 177 269
in Nigeria, 186, 192, 194, 195 Freedom House, 31, 32
in Rwanda, 49, 53, 54, 62, 64 Frente de Libertação de Moçambique
in South Africa, 99, 107, 109, 110, 121 (FRELIMO), 138
in Tanzania, 148 Front Populaire Ivoirien (FPI), 75, 77, 80
in Uganda, 138, 141, 143 Fulani, 179, 184, 214, 216, 227, 230,
in Zimbabwe, 103, 111–15, 122, 124 232, 241, 242, 259
302 Subject Index
Gacaca, 56, 64 horizontal inequalities (HIs)

Ga/Dangbe, 82, 83, 214, 216, 227, 241, definition of, 67
244 perceptions of, 6, 64, 75, 85, 91, 92,
Ganwa, 43, 44, 45, 46, 59, 62 232, 233, 246, 260, 262–4, 272
Gbagbo, Laurent, 70, 75, 77, 80, 81, 92, and vertical inequalities (VIs), 3
93 See also cultural status HIs; political
gender empowerment, 51, 53, 122, 190, HIs; socioeconomic HIs
204 Horowitz, Donald, 12, 37, 38, 67, 68,
gender inequalities, 178–207, 226 124, 194
Genocide in Rwanda, 49, 50 Houphouët-Boigny, Félix, 70, 71, 72–5,
Ghana, 4, 7, 8, 28, 39, 66–94, 209, 214, 76, 77, 78, 92, 256
216, 227, 231, 232, 241, 242, 243, human capital, 3
244, 245, 246, 253, 255–6, 259, 260, Human Development Index (HDI), 31,
261, 262, 265, 267, 269, 270 95, 104, 105, 134
Ghana Dance Ensemble, 86 human rights, 29, 41, 50, 54, 55, 56, 63,
Gini coefficient, 62, 99, 101, 104, 134 97, 108, 129, 164
Global Political Agreement (GPA), 112 Hutus, 42, 43, 44, 45, 46, 47, 48, 56, 57,
Gold Coast, 261 58, 59, 60, 61, 64, 261
government of national unity (GNU), hybrid systems, 36, 122, 259
15, 33, 107, 108, 119, 140, 149, 167,
197, 258, 259
Great Lakes, 40, 42, 59, 256 Ibrahim Index of African Governance,
grievances, 17, 35, 44, 76, 78, 81, 103, 29, 30, 39
104, 111, 115, 137, 139, 144, 147, Igbo, 179, 180, 191, 197, 204, 205, 214,
153, 163, 164, 167, 177, 182, 252, 216, 227, 231, 232, 241, 242, 244,
255, 258, 261, 263, 265, 271 259
Gross Domestic Product (GDP), 5, 54, inclusiveness, 15, 33, 36, 39, 78, 86, 87,
55, 61, 63, 72, 112 112, 117, 119, 121, 123, 129, 142,
Guatemala, 4 165, 256, 258, 263, 269, 270, 271,
Gueï, Robert, 78, 80 272, 274
Guinea, 1, 33, 36 Indian South Africans, 96, 97
Gukurahundi, 103, 107, 119 informal institutions, 6, 127, 138, 272
Gulu, 128, 144, 145, 146, 147, 149 informal power-sharing, 110, 111, 121,
Gurr, Ted, 3, 12, 37, 66, 67 141, 257, 265, 269, 272
Inkatha Freedom Party (IFP), 99, 107,
108, 110
Habyarimana, Juvénal, 45, 47, 60 institutionalisation of power struggle,
Hamad, Seif Shariff, 140, 141 41, 57, 58
Hamitic hypothesis, 43, 44 internally displaced persons (IDPs), 57,
Harare, 114, 117–21, 125 64, 65
Hausa, 179, 180, 184, 191, 204, 205, 214, International Criminal Court (ICC), 176,
216, 227, 230, 232, 241, 242, 259 269
health, 4, 50, 57, 62, 88, 157, 158, 159, International Monetary Fund (IMF), 75
160, 161, 162, 187, 189, 190, 191, international trade, 54, 55, 63, 73, 87,
199, 203, 213, 247 131, 133, 137
HIV/AIDS, 31, 110 Islam, 34, 40, 69, 71, 72, 75, 76, 83, 87,
Hoima, 145, 146, 147 98, 116, 117, 133, 147, 148, 152,
homelands, 48, 61, 65, 97, 99, 102, 104, 156, 170, 181, 182, 256, 259, 265
107, 116 Ivoirité, 76, 77, 78, 80, 256
Subject Index 303
Japan International Cooperation Agency Kenya People’s Union (KPU), 162, 163,
(JICA), 2, 132, 149, 168, 169, 170, 164
171, 172, 173, 174, 174, 204, 209, Kenyatta, Jomo, 157, 163, 164, 258
216, 225, 230, 241, 242, 246, 247, Khama, Seretse, 95
248, 249, 253, 257, 258, 259, 260 Khoisan, 96
Jews, 98 Kibaki, Mwai, 165, 166, 167, 172, 177,
Johannesburg, 96, 105 258
Jonathan, Goodluck, 199, 202, Kigali, 62
203 Kikuyu, 154, 155, 156, 157, 163, 164,
journalism, 50, 55, 66, 211 166, 167, 169, 170, 173, 214, 216,
Judiciary 227, 231, 232, 241, 245, 258
in Burundi, 46, 51 Kikwete, Jakaya Mrisho, 140, 148
in Côte d’Ivoire, 77 King Mutebi, 139
in Ghana, 84, 85, 90 King Mutesa II, 128
in Kenya, 157–8, 162, 164, 176, 270 King Mwambutsa, 44, 45
in Nigeria, 180, 182, 184–6, 194, King Rwabugiri, 43
196–200 Kisii, 166, 227, 231, 232, 241, 245
in Rwanda, 49, 56 Krou, 70, 72, 79, 80, 92
in South Africa, 107, 110 Kufuor, John, 88, 89
in Tanzania, 142, 143–4, 146, 151
in Uganda, 130, 139, 142, 143–4, 146, Lagos, 36, 204, 228, 229, 259
151 Lake Victoria, 126, 127
in Zimbabwe, 106, 107–8, 119 Lancaster House Agreement, 103, 124
Languages, 4, 47, 82, 83, 86, 87, 97, 108,
Kabaka, 127, 128 179, 191, 192
Kagame, Paul, 53, 62 Lebanon, 34
Kalenjin, 154, 155, 156, 157, 164, 165, Legislature
166, 169, 170, 227, 258 in Burundi, 51, 54, 55
Kamba, 154, 155, 169, 170, 174, 175, in Côte d’Ivoire, 77
214, 223, 227, 228, 229, 231, 239, in Ghana, 85
243, 245, 258 in Kenya, 162, 163, 164, 176
Kampala, 127, 129, 138, 139, 145, 146, in Nigeria, 182, 184, 198, 201
147, 149, 152 in Rwanda, 50, 53
Karume, Amani Abeid, 141 in South Africa, 107, 109, 122
Karume, Sheikh Abeid Amani, in Tanzania, 133
140 in Uganda, 143, 146
Kaunda, Kenneth, 95 in Zimbabwe, 106, 111, 113, 114
Kayibanda, Grégoire, 45 Lewis, W. Arthur, 11, 12, 13, 37
Kenya, 4, 7, 8, 23, 24, 27, 28, 31, 33, 35, liberal democracy, 41, 110
36, 39, 108, 133, 153–77, 209, 214, Liberal Democratic Party (LDP), 165
216, 227, 228, 231, 232, 241, 243, Libya, 8, 178
245, 246, 253, 254, 258–9, 260, 261, Lijphart, Arend, 11, 12, 13, 14, 15, 17,
262, 264, 265, 266, 268, 269, 270, 37, 38, 39, 62, 63, 124
272 local elections, 14, 16, 18, 28, 64
Kenya African Democratic Union Local Government Authorities (LGAs),
(KADU), 162, 163 143
Kenya African National Union (KANU), Lord’s Resistance Army (LRA), 129, 135,
162, 163, 164, 165, 172, 179 139
Kenya Broadcasting Corporation (KBC), Luhya, 154, 155, 156, 157, 166, 216,
165, 166 227, 258
304 Subject Index
Luo, 154, 155, 156, 157, 166, 168–9, multi-party system, 6, 10, 13, 14, 41, 49,
170, 173, 227, 242, 245, 258 50, 53, 82, 84, 117, 126, 138, 140,
Luwero Triangle, 139 142, 147, 163, 167, 195, 257, 258
Museveni, Yoweri, 39, 62, 126, 129, 130,
138, 139, 141, 142, 145, 150, 152,
Mahama, Alhaji Aliu, 89 258
majoritarian politics, 110, 111, Musyoka, Kalonzo, 165
267
Malawi, 21, 38, 118 Nairobi, 154, 159, 160, 161, 163,
Malaysia, 4 168
Mali, 1 Nakuru, 154, 168
Mandela, Nelson, 52, 95, 99, 106, 107, National Alliance Party of Kenya (NAK),
108, 270 165
Mauritius, 14, 35 National Democratic Congress (NDC),
Mbarara, 128, 145, 146, 147, 152 85
Mbeki, Thabo, 108, 110 National Liberation Council (NLC),
Mfecane, 96, 101 84
migrant workers, 97, 99 National Liberation Movement (NLM),
military 84, 102
in Burundi, 52 National Party of Nigeria (NPN), 199
in Côte d’Ivoire, 80 National Party (NP), 97, 98, 108, 110,
in Ghana, 84 143, 199
in Kenya, 39, 133 National Patriotic Party (NPP), 85, 89
in Nigeria, 182, 192, 193, 195 National Rainbow Coalition (NRC), 84,
in Rwanda, 40, 42, 43, 45, 46 165
in South Africa, 107 National Resistance Army (NRA), 126,
129, 139
in Tanzania, 142
National Resistance Movement (NRM),
in Uganda, 133, 136
129, 138, 139, 141, 142, 143, 145,
in Zimbabwe, 107
146, 150, 258
Mill, John Stuart, 11
natural resources, 31, 73, 87, 215
minority groups, 53, 83, 118, 179, Ndadaye, Melchior, 46
258
Ndebele, 96, 101, 103, 104, 115, 117,
Minority At Risk (MAR), 208, 209 118, 119, 121, 125, 214, 216, 227,
Miracle Africain, 73 231, 232, 241, 242, 245, 257, 261,
Moi, Daniel Arap, 157, 164, 165, 172, 263, 271
259 Niger Delta, 207
Mole/Dagbani, 82, 83, 86, 90, 216, 227, Nigeria, 1, 4, 7, 8, 12, 17, 23, 28, 31, 34,
231, 232 35, 36, 39, 178–207, 209, 211, 214,
Mombasa, 154, 160, 168 216, 227, 228, 231, 232, 241, 243,
Mouvement démocratique républicain 244, 245, 246, 247, 253, 259, 260,
(MDR), 49 261, 262, 265, 266, 268, 269, 270,
Mouvement Patriotique pour la Côte 271
d’Ivoire (MPCI), 70, 81 Nigerian Civil War, 192, 259
Movement for Democratic Change See also Biafran War
(MDC), 105, 111, 112, 113, 114, Nigerian National Petroleum
119, 121 Corporation (NNPC), 202
Mugabe, Robert, 95, 103, 104, 112, Nkomo, Joshua, 103
119, 122, 124, 125, 257, Nkrumah, Kwame, 84, 85, 86, 87, 256,
270 270
Subject Index 305
Northern Mandé, 70, 72, 75, 76, 79 in Rwanda, 41, 45, 48

Nyerere, Julius, 1, 52, 95, 126, 138, 140, in South Africa, 95
142, 150, 257, 270 in Tanzania, 127
in Uganda, 134
Obasanjo, Olusegun, 197, 199, 202 in Zimbabwe, 103
objective inequalities, 2, 5, 6, 93, 210, political institutions
211, 230, 232, 241, 260, 262, 273 definition of, 2, 11
Obote, Milton, 128, 129, 136, 138, 142 formal, 7, 8, 11, 14, 17, 144, 257
Odinga, Oginga, 163 informal, 6, 138, 199
Odinga, Raila, 165, 166 power-concentrating (PC), 11, 12, 13,
Okello, Tito, 129, 142 58, 206, 252, 254, 255, 257, 259
one-party system, 13, 38 power-dispersing (PD), 11, 13, 18, 52,
Orange Democratic Movement (ODM), 206, 252, 254, 255, 258, 259, 266,
166, 167 269, 271, 272
Ouattara, Alassane, 75, 76, 77, 80, 81, polyarchy, 33
92, 93 population size, 13, 18, 23, 69, 179, 255
poverty, 62, 89, 99, 105, 111, 134, 135,
Parti de l’union et du progrès national 136, 143, 144, 148, 150, 153, 178,
(UPRONA), 44, 45, 52, 61 189, 190, 207
Parti Démocratique de la Côte d’Ivoire power-sharing
(PDCI), 71, 72, 75, 76 in Burundi, 40, 41, 46, 49, 51
Parti du mouvement de l’émancipation in Ghana, 84
hutu (PARMEHUTU), 44, 45
in Kenya, 167
Party of National Unity (PNU), 166
in South Africa, 105–11, 108, 110, 121
patronage, 56, 73, 74, 139, 143, 164,
in Tanzania, 138, 141, 147, 151
167, 168, 184, 195, 197, 266
in Zimbabwe, 106
peace-building, 41, 271
See also ad hoc power-sharing;
Pemba island, 131
informal power-sharing and
People’s Democratic Party (PDP), 195,
sustainable power-sharing
197, 199
presidency
perceived political horizontal
inequalities (PPHI), 242, 244, 249 in Burundi, 50
perceived socioeconomic horizontal in Côte d’Ivoire, 72, 81
inequalities (PSEHI), 230, 244, 248 in Ghana, 89
perceptions surveys in Kenya, 162, 165, 167, 176, 258
in Ghana, 83, 85, 91 in Nigeria, 205, 206
in Kenya, 168 in Rwanda, 49
in Nigeria, 204 in South Africa, 107, 108, 110, 112
in South Africa, 115 in Tanzania, 129, 133
in Tanzania, 127, 145, 151 in Uganda, 128, 139
in Uganda, 126, 133 in Zimbabwe, 104
in Zimbabwe, 112, 115 Progressive Peoples Alliance (PPA), 197,
Peru, 4 198
policy recommendations, 2, 8, 254, 273 proportional representation (PR), 12, 13,
political horizontal inequalities (HIs) 14, 15, 24, 33, 36, 39, 51, 78, 103,
in Burundi, 41 107, 108, 113, 122, 123, 124, 125,
in Côte d’Ivoire, 92 194, 254, 257
in Ghana, 69, 81, 85, 93 Protestants, 4, 83, 133, 170
in Kenya, 154, 258 Provisional National Defence Council
in Nigeria, 182, 186 (PNDC), 84
306 Subject Index
quota system, 47, 122, 124 Somali, 169, 170, 214, 216, 228, 231,
232, 242
Rassemblement des républicains (RDR), South Africa, 7, 8, 12, 23, 28, 33, 34, 38,
76, 77, 80 39, 95–125, 215, 253, 254, 256–7,
Rawlings, Jerry, 84, 85, 87, 89, 93 261, 262, 263, 265, 267, 269, 270,
refugees, 44, 45, 47, 48, 61, 65 271
South African Communist Party (SACP),
religions, 40, 42, 50, 71, 72, 76, 81, 83,
98, 110
86, 87, 133, 139, 143, 147, 148, 152,
154, 170, 171, 179, 181, 191, 204, South African War, 96
259, 264 Southern African Development
Community (SADC), 112, 124
See also Christianity; Islam; traditional
Southern Mandé, 70, 78, 79
religions
state-building, 1, 40–65
Rhodes, Cecil, 13, 101
structural adjustment, 111
Rift Valley, 153, 155, 156, 158, 164, 166,
subjective inequalities, 2, 5, 6, 8, 209,
167, 258
210–15, 232, 233, 241, 245, 260, 273
Rwagasore, Prince Louis, 44, 45, 52, 61
Sudan, 1, 23, 31, 33, 38, 108, 255
Rwanda, 7, 8, 28, 35, 39, 40–65, 253,
Sukuma, 130, 131, 140
255, 261, 262, 264, 265, 266, 267,
sustainable power-sharing, 269, 270,
271
272, 274
Rwanda Demobilisation and
Swahili, 62, 131, 133, 142
Reintegration Commission (RDRC),
54, 63
Tambo, Oliver, 95
Rwandan Defence Forces (RDF), 53
Tanganyika African National Union
Rwandan Patriotic Army (RPA), 54, 63
(TANU), 140, 142
Rwandan Patriotic Front (RPF), 42, 45,
Tanzania, 7, 8, 23, 28, 34, 35, 39, 59,
46, 47, 48, 49, 50, 53, 54, 56, 58, 59,
126–52, 253, 257–8, 264, 265, 266,
61, 62, 63
268, 269, 270
Togoland Congress Party (TCP), 84
secessionism, 126 traditional religions, 83, 133, 170
senate, 14, 16, 21, 51, 53, 62, 107, 113, trait-taking and trait-making, 36
124, 176, 180, 184, 185, 186, 195, Truth and Reconciliation Commission
196 (TRC), 108
Shirazi, 131, 142, 145 Tsvangirai, Morgan, 112
Shona, 101, 103, 104, 117, 118, 119, Tutsis, 42, 43, 44, 45, 46, 47, 48, 50, 51,
121, 125, 216, 228, 231, 232, 241 52, 56, 57, 59, 60, 61, 64
Social Revolution, 44, 45, 46, 47, 50, 60, Tutu, Desmond, 95
255 Twas, 40, 51, 53, 55, 59, 63
socioeconomic horizontal inequalities
(HIs) Uganda, 7, 8, 23, 28, 31, 35, 39, 45, 47,
in Burundi, 41 59, 60, 62, 126–52, 209, 214, 216,
in Côte d’Ivoire, 69, 92, 93 227, 229, 231, 232, 241, 242, 245,
in Ghana, 68, 69 246, 253, 257–8, 259, 260, 261, 262,
in Kenya, 156, 158, 168 265, 266, 268
in Nigeria, 186 Ujamaa, 138, 140
in Rwanda, 41, 48 unemployment, 99, 123, 137, 167, 230
in South Africa, 99 Unguja island, 131, 133, 137, 140, 144,
in Tanzania, 147, 149 145, 147, 148, 149, 150, 152
in Uganda, 134 unitary state, 13, 16, 17, 126–52, 254,
in Zimbabwe, 103, 104, 111 265
Subject Index 307
United Democratic Front (UDF), 99 World War I (WW I), 43

United Nations Development See also civil war; Cold War
Programme (UNDP), 9, 38, 48, 54, West Africa, 11, 37, 255
62, 89, 95, 104, 105, 123, 134, 137, Westminster system, 11, 13
189, 190 winner-takes-all, 112, 122, 162, 167,
United Nations Security Council, 178 194, 270
Universal Primary Education Policy, 135, World Bank, 9, 63, 136, 152
137
Upper Volta, 70 xenophobia, 75, 117, 121, 256
Uppsala Conflict Data Program (UCDP), Xhosa, 96, 104, 116
1, 9, 31, 39
urbanisation, 48, 61, 98, 123, 155, 226,
Yar’Adua, Umaru, 182, 186, 197, 199
265
Yoruba, 179, 180, 191, 204, 214, 227,
259
violent conflict
in Burundi, 40 Yugoslavia, 107
in Côte d’Ivoire, 80, 92
in Ghana, 69, 91, 93 Zambia, 17, 38, 118
in Kenya, 153 Zanzibar, 126, 127, 130, 131, 133, 134,
in Nigeria, 178, 179 137, 140, 141, 142, 143, 144, 147,
recurrence of, 259 148, 149, 150, 151, 152, 257, 268
in Rwanda, 42 Zimbabwe, 7, 8, 15, 24, 28, 33, 35, 38,
in South Africa, 96, 110 39, 95–125, 209, 214, 216, 227, 228,
in Tanzania, 141 231, 232, 241, 242, 245, 246, 253,
trends in Africa, 33, 96 254, 256–7, 259, 260, 261, 262, 263,
in Uganda, 141 265, 267, 268, 269, 270, 271, 272
in Zimbabwe, 104, 111, 112 Zimbabwe African National Union
Voltaic, 70, 72, 76 (ZANU), 103, 104, 105, 111, 112,
Volta Regions, 83, 85, 86, 87 113, 114, 115, 118, 119, 121, 124
Zimbabwe African People’s Union
Waki report, 167, 168 (ZAPU), 103, 104
war Zimbabwe Congress of Trade Unions
South African War, 96 (ZCTU), 111
Uganda-Tanzania War, 129 Zulu, 96, 99, 101, 104, 107, 110, 151
World War II (WW II), 118 Zuma, Jacob, 110

(Conflict, Inequality and Ethnicity) Yoichi Mine, Frances Stewart, Sakiko Fukuda-Parr, Thandika Mkandawire (eds.) - Preventing Violent Conflict in Africa_ Inequalities, Perceptions and Institutions-Pa

Uploaded by

Copyright:

Available Formats

You might also like

(Conflict, Inequality and Ethnicity) Yoichi Mine, Frances Stewart, Sakiko Fukuda-Parr, Thandika Mkandawire (eds.) - Preventing Violent Conflict in Africa_ Inequalities, Perceptions and Institutions-Pa

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Conflict, Inequality and Ethnicity) Yoichi Mine, Frances Stewart, Sakiko Fukuda-Parr, Thandika Mkandawire (eds.) - Preventing Violent Conflict in Africa_ Inequalities, Perceptions and Institutions-Pa

Uploaded by

Copyright:

Available Formats

Preventing Violent Conﬂict in Africa

Conﬂict, Inequality and Ethnicity

Graham Brown, Arnim Langer and Frances Stewart (editors)

Conﬂict, Inequality and Ethnicity

Foreword by Sadako Ogata

List of Figures and Tables viii

1 Introduction: Disentangling the Linkages between

2 Comparing Political Institutions: Institutional Choice and

3 ‘Twin Countries’ with Contrasting Institutions:

3.5 Institutions and political power: How do

4 Horizontal Inequalities, Ethnic Politics and Violent

5 Beyond Ad hoc Power-Sharing: Comparing South Africa

6 Is Ethnic Autonomy Compatible with a Unitary State? The

7 The Politics of Identity, Horizontal Inequalities and

7.7 The new constitution and prospects

1.1 Horizontal inequalities, political institutions and perceptions 7

7.14 Has ethnicity become more important in Kenyan politics

2.1 Ratings of eight variables in sub-Saharan African countries 19

5.4 National Assembly election results in South Africa 108

8.18 Distribution of federal constituency seats by political party 196

While I served as the UN High Commissioner for Refugees (UNHCR) through

for me to watch the close collaboration between political scientists and

Table 4.6 is reprinted with permission of Palgrave Macmillan from Langer,

Sakiko Fukuda-Parr is Professor of International Affairs at the New School,

Mari Katayanagi is Research Fellow at JICA Research Institute, Tokyo. She

Mwangi S. Kimenyi is Senior Fellow and Director of the Africa Growth

Satoru Mikami is Research Fellow at JICA Research Institute. He was Assis-

Yoichi Mine is Professor at Graduate School of Global Studies, Doshisha

Thandika Mkandawire is a former director of the United Nations Research

Julius E. Nyang’oro taught African politics and development at the Uni-

Yuichi Sasaoka is Professor of International Relations at the Graduate

Frances Stewart is Emeritus Professor of Development Economics, Emeri-

horizontal inequalities; and human development. Books include, as author,

Shinichi Takeuchi is Senior Research Fellow and Director of African Studies

Ukoha Ukiwo is a senior lecturer at the Department of Political and

1.2 HIs and violent conﬂict

• Economic HIs include inequalities in access to and ownership of assets –

• Social HIs include inequalities in access to a range of services, such as

The relevance of any element depends on whether or not it is an important

1.3 HIs, perceptions and political institutions:

Despite the increased understanding and recognition of the causal link-

If groups do not perceive the prevailing socioeconomic and political

A second issue that has so far received insufﬁcient systematic analysis is

of different types of political institutions on conﬂict proneness and peo-

Improving our understanding of the interaction between these three factors

Political processes Perceptions Structural risk factors

Figure 1.1 Horizontal inequalities, political institutions and perceptions

mechanisms in societies. The hypothesised relationship is shown graphically

1.4 Structure of this book

Figure 1.2 Ten African countries chosen for case studies

contrasting political outcomes, including Rwanda and Burundi (Chapter 3),

When a group of people feel they are excluded from decision-making

of political institutions and the rising inequalities in Africa that necessi-

2.2 Deﬁning political institutions

Let us begin with a theoretical review. In contemporary political sci-

harshly penalised. Group differences could be exaggerated in an irretriev-

power-dispersing (PD) and power-concentrating (PC) institutions. We shall

2.3 Elements of PD and PC institutions

Following on from the conceptual deﬁnition outlined above, this section